The oxygen-shielded air-acetylene flame in emission analysis

792

SHORT

COIlMLNlCATl0l.S

Summary-The ring-oven method combined with solvent extractton has been used for the separation of all six platinum metals when present in pg-amounts in a drop of solution. The separated constituents are determined by ring-calorimetry. Zusammenfaasung-Die Ringofen-Methode wurde zusammen mit der fhissig-tliissig-Extraktion zur Trennung aller sechs Platinmetalle verwendet. wenn in einem Tropfen Liisung pg-Mengen vorlagen. Die getrennten Bestandteile werden durch Ring-Colorimetrie bestimmt. R&sum&La methode du four annulaire, combinie avec l’extraction par solvant. a tte utilisee pour la separation des six mttaux de la mine du platine quand ils sont presents en quantites de l’ordre du pg dans une goutte de solution. Les constituants sipares sont doses par colorimetrie annulaire.

Talanc~~.Vol. 21. pp 192 196

Pergamon

Press

1974. Prmted in Great Br~tam.

THE OXYGEN-SHIELDED AIR-ACETYLENE IN EMISSION ANALYSIS

FLAME

Recent studies have shown flame-emission spectroscopy (FES) to be a useful complementary technique to atomic-absorption and fluorescence spectroscopy,‘-’ with good analytical sensitivity’ and ready applicability to multi-element analysis.9*Lo The present work exammes FES with an oxygen-shielded air-acetylene flame. The low burning velocity of the shielded flame permits it to be supported on a wide burner slot, to grve a low background inner zone similar to that observed with a multi-slot burner. ’ 1A simultaneous enhancement of the flame temperature occurs on shielding, because of oxygen-supported combustion m the outer diffusion zone.‘l This creates a high-temperature low-background system. found to be very suitable for FES. EXPFRIMELTAL

A Varian AA5 spectrometer was used. with the standard Ebert 0.5-m monochromator and 638 line/mm grating. giving a reciprocal dispersion of 3.3 nm/mm at the exit slit for the first order. A 285 Hz chopper was used to modulate flame emission signals. The spectrometer output was recorded on a Phillips PR 4069M/OO recorder. with a 3 set time constant. The oxygen-shielded burner was similar to that described earlier.” but with a slot reduced to 25 cm in length and increased to 1.25 mm in width, to give the low-background inner zone mentroned above. and with the water cooling omitted. The burner was mounted directly in the AA5 burner assembly. This emission burner was unable to support a C,H,-N,O flame. because of the higher burnmg velocny. and emission measurements on C,H,-N20 flames were therefore made by using the standard Varian mtrous oxide head. Detection limits in this flame were obtained at a constant N1O flow of 8.2 I./min. with fuel flow. burner position. monochromator slit and photomultiplier voltage optimized for each element. FES results for all names used were obtained by using analytrcal-reagent grade chemrcdts dissolved in etther demmerabzed distrlled water or I M hydrochloric acid. A IWO {cgjml potassmm solutron was used as an tonization suppressant where necessary, except for calcium, for which potassmm salts of sufficient purity were not avatlable. and for barium in the shielded flame. Owing to the low background of the shielded Hame. signal-enhancement due to ionization suppression was found to be more than offset by increased flame noise from the potassium ionization continuum in the spectral region observed

793

SHORT COMMUNICATIONS

RESULTS

All studies on the shielded flame were made at a constant air flow-rate of 6.2 I./mm and constant oxygen flowrate of 4.0 l.lmin. corresponding to the shielding “plateau” for this particular flame.12 For most of the elements eiammed. emlsslon signals were observed to increase with fuel Rows higher than those corresponding to the lummous hmlt of the flame. However. a simultaneous and severe increase of flame noise also occurred at this pomt. and accordmgly a flame stoichlometry Just below that of the luminous limit was adopted for optimum sensltlvlty.

17-

16-

15-

14-

mm from ccntre of flame Fig. 1. Background emission profiles of the shIelded flame. Contours give the “/, intensity Increase relative to the mmimum Intensity at the centre of the flame. Contours were measured at 240.0 nm and a slit-width of 25 pm

hdAL

Background Spg/ml

(a)

Background 2pg/ml

Cu

Cu

Background 2pg/ml Cu (b)

(c) Fig. I!. Cu emission at 324.8 nm (a), Air-acetylene flame. monochromator sht 25 llrn. photomultlpher voltage 850 (b): Shlelded air-acetylene flame. same instrumental settings as for (a). (c): Nitrous oxide-acetylene flame. monochromator slit 25 jtm. photomultiplier voltage 450 V.

794

SHORT

COMMUSICATIOUS

Figure I shows the background notse contours of the shielded flame. measured for the luminous-hmtt Hame. The results of Fig. I were obtained by using a 0.5 mm dra. ctrcular collimator: subsequent results were obtamed by replacing this with a larger collimator. approxrmately matching the 104, contour. shown by the dotted outlme in Fig. I. The low flame background given by this system is illustrated by the results in Fig. 2. showing the Hame emrssion signals of Cu at the optimum 324.8 nm line. These results show the shielded Hame to give a background intensity little different from that of an air-acetylene flame alone, but an emission signal for Cu of similar magnitude to that from the C,Hz-N,O flame at the instrument settings used. Analytical

sensrrimty

The work of Koirtyohann and Pickettk6 among others has shown the C,H,-N,O flame to be probably the most sensitive and versatile system currently available for flame emission work. Accordmgly the emisston sensitrvtties grven by the shielded burner were compared with the corresponding results given wtth the same Instrument. by the CzH2-N20 Hame. and wtth those determined elsewhere (Table I). Although dtthculttes artse m gtvmg any exact comparison of analytical data between different laboratortes. a.’ it is felt that the AA5 spectrometer used here IS sufficiently similar to many other units in current use for the comparative data m Table I to offer some indication of the level of analytical sensitivity whrch may be expected from FES with the shielded tlame. Spectral

interferences;

calibration

curves

Spectral interferences were generally found to be of comparable magnitude m the shielded flame and in the C,H*-N,O flame. The latter flame, however, did show a clear advantage m reducmg interferences due to band emission from the alkaline-earth metal oxides and hydroxides. presumably because of Its enhanced chemical Table I. Detection limits at a stgnal to root-mean-square nm

Species

i,

Al Ba Be Be0 Ca co

3962 553.6 234.9 470.9 422.7 345.4 352.7 425.4 324.8 327.4

Cr cu

Zf Mg Mn Ni Sc

Sr Ti Tl V Zr

670.8 372.0 285.2 403.1 403.3 341.5 352.4 391.2 402.0 402.4 460.7 334.9 399.9 377.6 535.1 318.5 437.9 360. I

Shielded flame* 05 0.001t 20 0.4 OGOO2t 007 001 0.006 0.01 0.05 00JO1 0005 OGO8

0.03 2.5 OGOOl >20 0.01 @03 I.5 120

CH-N20* 0.05

oGI2 40 0.4 OWOI 0.05 006 0.05 0.03 0.06 OGOO2 0.01 0.006 0.04 0.4 OGOOl 0.2 0.08 0.05 0.1 IO

&Hz-N20L4 0.05 0.002 I.0 Iwnn2 0.03 _~

OW4 0.01 0.03 oGOOO2 0.07 OX08 0.02 0.8 oGOO5 0.2 0.1 0.02 0.1 5

noise ratio of 2: I. la, J~II C2H,-N,OJ _

Separated C-H,-NZO” 1

0.019 OGO2~ 40 @2 0.000 I 0.05 WOO5 001 0.05 oG?OQ3 0.005 oGO5 0.03 0.03 OGOO2~ 0.2 0.08t

0.02

0.01 3

-

* This study: 100 ilrn and 25 Ltm monochromator slits were used with the shielded and CIH2-N20 flames respectively. t No ionizatron suppressant used with these elements. g Detection limits for Al. Ba and Sr given as OGO5.0001 and OWOI respectively. in later publications.‘,“

795

SHORT COMMUNICATIONS

reducmg propertress The calibratton curves obtamed wtth both flames, both in the presence and absence or spectral interferences. were also similar. In ail cases examined, calibration curves were linear over a range of more than two orders of magnitude from the detectron hmtt The results obtained show that FES with the oxygen-shielded flame is capable ofgtvmg a very similar performance to the C2H,-NIO flame for many elements. The major advantage of the former appears to be its low background. which permtts good sensitivity to be obtained, in parttcular for Cu and Tl, wtth a comparatively low-resolutton monochrochromator. Thiscontrasts wtth the rather stringent optrcal requirements for FES m the C,H2-N20 flame.“ a conclusion which ISemphasized by the experimental condttions used. where resolutton was further reduced by the use of large slit-wtdths to compensate for inadequate photomultiplier sensittvity. Ac~rlo~ledye~,~rrlr.\-lt IS a pleasure to acknowledge the helpful comments made by Professor W. F. Ptckermg of the Universttv of Newcastle, N.S.W.. Australia, during this study. We are also indebted to the Nattonal Research Councrl of Canada for a grant m support of this work. Trace Atlal~sts Research Came Depatmnertt of Chrrt~rsr~~ Dalhouste L‘ttirersrr~: Hal$tr. X.S.. Canada

K.E R. G.

CURTIS JR.* STkPHENS

STWENSON.

R.

REFERENCES 1 R. N Kmseleq. A. P D’Srlva and V A. Fassel. Anal. Cheat.. 1963, 35, 910. 2. A. P. D’Silva. R. N. Kmseley and V. A. Fassel. ibid., 1964, 36, 1287. 3. V. A. Fassel and D. W. Golightly, ibid., 1967, 39, 466. 4 E. E. Ptckett and S R. Kotrtyohann. Spectrochan. Acta. 1968, 23B, 235. 5 S. R. Kontyohann and E. E. Ptckett. ihld., 1968, 238, 673. 6. E. E. Ptckett and S. R. Koirtyohann. ibid.. 1969, 248, 325. 7 G. D Christian and F. J. Feldman. Anal. Chent., 1971. 43, 611. X J. D. Winefordner. V. Svoboda and L. J. Chne. Crmcal Rev. Anal. Chetn.. 1970. 1. 233. 9. E E. Ptckett and S. R. Kontyohann. Anal. Chetn.. 1969,41, No. 14. 28A. IO. V. A. Fassel. R. W. Slack and R. N. Kniseley, ibid.. 1971, 43, 186. I I E A. Boling. Specrrochrtn. Acta, 1966. 22, 425. 12 R Stephens. Talattta. 1973. 20, 765. 13 S. R. Korrtyohann and E. E. Pickett. Appl. Spccrrj: 1969, 23, 597. I-I G D Chrtstian and F J Feldman. ibid.. 1971. 25, 660. 15 M. S Cresser. P. B. Joshtpura and P. W Keliher. Sprcrroscop~ Letters, 1970.3, 267.

Summary-Flame-emtsston studtes have been made on 18 elements in the mner zone of an oxygenshielded atr-acetylene flame. The shrelded flame gave higher emission sensitivity that that of the C,HI -NZO flame for Cu and Tl. and comparable sensitivity for a number of other elements, but poorer sensttivtty for elements forming stable refractory oxides in flames. The inner zone of the shtelded flame has low emission-background and high flame-temperature. permrtting good analytical sensttivity to be obtained wtth relattvely low-resolution optical equipment.

Zusammenfassung-Die Flammenemissron von 18 Elementen wtude in der inneren Zone einer mit Sauerstoff abgeschirmten Luft-Acetylen Flamme untersucht. Die abgeschirmte Flamme ergab bei Cu und Tl eme hiihere Emisstonsepfindlichkeit als die &Hz-N20-Flamme, bei einer Anzahl anderer Elemente verglerchbare Empfindlichkeit, geringere Empfindlichkeit jedoch bei Elementen, dte in Flammen stabile feuerfeste Oxide bilden. Die innere Zone der abgeschinnten Flamme weist emen germgen Emissions-Untergrund und eine hohe Flammentemperatur auf; daher erlaubt sie gute analytische Empfindlichkeit bei relativ gering aufliisender optischer Ausriistung.

* Present address. Department or Chemtstr!. Memphis State Untversny. Memphis. Tennessee. U.S.A.

796

SHORT

COMMLYICATIONS

R&m&-On a elIectue des etudes d’emission de flamme sur 18 iitments dans la zone interne d’une flamme air-ac&yl&ne prottgCe de l’oxyg+ne. La flamme protegee a donne une sensibilite d’tmission plus Blev&e que celle de la flamme C,H,-N20 pour Cu et l-1. et une sensibilitk comparable pour un certain nombre d’autres Cltments. mais une sensibiliti plus faible pour les tlCments formant des oxydes rbfractaires stables dans les flammes. La zone interne de la flamme protigke an un found d’tmission bas et une haute temperature de flamme. permettant d’obtenir une bonne sensibilite analytique avec un equipement optique de risolution relativement faible.

Jhr~r~r.Vol

?I. pp 796799

Pergamon

Press. 1974 Prmted m Great Brnnm

ZUR ENTMISCHUNG

DER L&UNGSMITTEL

CHROMATOGRAPHISCHEN SORPTION

BE1 DER

TRENNUNG-V*

VON FLIESSMITTELDAMPFEN

AN CELLULOSE

(Eirtyeyartgen N,~I27. Mdr-_ 1972. Angerlornrnrn a11119. Jnmrar 1974) In einer friiheren VerGffenttichung’ wurde gezeigf da0 bei der Sorption von Alkohol-Wasser-Gemischen an Cellulose deren Massezunahme nicht allein auf eine Erhahung des Wassergehalts zuriickgefihrt werden kann. sondern, da0 neben Wasser such Alkohole an der Cellulose gebunden werden. Nun war es interessant. noch die Zeitabtingigkeit der Massezunahme der Sorption zu priifen und festzustellen. wie sich Cellulosepulver verhllt. das Liisungsmitteldiimpfen mit steigendem Wassergehalt ausgesetzt worden ist. EXPERIMENTELLER

TEIL

Die benutzten LGsungsmittel wurden durch mehrfache Destillation gereinigt. Die Sorption der LGsungsmittel an der Cellulose wurde in einer zylindrikhen chromatographischen Kammer von 10 cm Durchmesser und 25 cm Hijhe durchgefiihrt. Das Cellulosepulver wurde in emem Wlgeglas 55 x 35 mm untergebracht. wghrend das LGsungsmittel den Boden der Kammer bedeckte. Crllcdose: Whatman Cellulose Powder. Standard Grade. Der Wassergehalt wurde durch blamperometrlsche Titration mit Karl-Fischer-Liisung bestimmt. ERGEBNISSE

UND

DISKUSSION

Urn die Verhaltnisse bei der Sorption des FlieDmittels am Trlger noch von emer anderen Seite her zu erfassen. wurde die Sorption aus der Gasphase untersucht. Zertahhiingigkect

der Sorptiorl

Urn die Zeitabhlngigkeit der Massezunahme zu erfassen. wurden I,0000 g des getrockneten Cellulosepulvers I bis IO Tage in der getittlgten Atmosphlre der mijglichst wasserfreien niederen aliphatlschen Alkohole au&ewahrt. Die Resultate sind in Abb. I dargestellt. Das Diagramm zelgt. dal3 die Massezunahme in erster Nlherung mlt der Dielektrizitltskonstanten der Solventien abfiillt. Der schwankende Wassergehalt der Gemische schemt dabei die MeBwerte nur in untergeordnetem MaBe zu beeinflussen.

Zur Untersuchung des EinRusses des Wassergehalts auf die Sorption der Liisungsmitteldlmpfe durch lufttrockene Cellulose wurden weltere Proben eine Woche iiber dem Gemisch L&ungsmittel/Wasser aufbewahrt und ddnach wieder die Massezunahme bestimmt. Die erhaltenen MeDwerte sind Ielder ohne gral3eren experimentellen Aufwand nur qualitativ reproduzierbar. Die Ergebnisse sind in Abb. 2 dargestellt. Trotz der unbefriedigenden Reproduzierbarkelt kann m;Ln aus diesen Versuchsreihen wesenthche Informationen iiber die Massezunahme. iiber den EinBuR des Alkohol-Wasser-Verhlltnisses. iiber die Zusammensctzung des Sorbats usw. bekommen. * Mitteilung IV: Tulor~tu. 1974. 21, 641.