The determination of palladium by atomic-absorption, atomic-fluorescence and thermal-emission spectroscopy in various flames

Atrafyficd Chimica Ada

2%

Elscvicr

Publishing Company, Amsterdam. Prlnted in The Netherlands

THE DETERMINATION ATOMIC-FLUORESCENCE VARIOUS FLAMES*

OF AND

PALLADIUM BY THERMAL-EMISSION

ATOMIC-ABSORPTION, SPECTROSCOPY

IN

Atomic absorption spectroscopy (a.a.s.) of palladium was first described by kc-~:rcsI;,rrANI.) I-In~lEsi who used a low-temperature flame ancl reported a sensitivity c~f I p.pm. at the 247.6%nm line. Since tlien, several other. authors have reported palladium absorption at various analytical lines and in various flames~-4. The best sensitivity of 0.01 p.p.m. was obtained by IZu13ESKAANn S-rtrI*ARGwith the long-tube technique. Most authors 7-r.’ have found no interference from a number of elements (including noble metals) and from various acids. In spite of these facts, we agree completely with SLAvIN’s statementI”--” more work is required before the pallacliutn determination can be considered well understood”. Atomic fluorescence spectroscopy (a.f.s.) as a sensitive method for the determinati.on of metals is now well established, and it is being applied to an increasing number of elements. However, the volume of literature on palladium is relatively sma!l. OnraNE’r*ro AND Rossrl” reported a detection limit of IO p.p.m. by exciting fluorescence of the 34o.46-nm line with a cadmium vapour clischarge lamp by means of the cadmium 34o.37-nm line. DINNIN 17 found a detection limit of 0.5 “overlapping” 1J.p.m. using a hot hollow-cathode lamp as an excitation source. MANNING AND excited palladium fluorescence both with a 150-w xenon arc and a HIZNEAGE~~ shielded hollow-cathode lamp ancl reported detection limits of 50 p.p.m. and 2 p.p.m., respectively, for the 34o.46-nm line in an air-hydrogen flame. it has been shown that excellent detection In previous publications .19-21 limits in fluorescence can be obtained with high-intensity hollow-cathode lamps as the excitation source in combination with a suitable flame. In this study, palladium atomic fluorescence was excited both with a high-intensity hollow-cathode lamp and sn electrodeless discharge tube and measured in various premixed flames. The determination of palladium by flame thermal-emission spectroscopy (f.e.s.) in hot flames has also been investigated by several autllors22-27. ESHEL~XAN et al.Q~~20, for example, have reported its detection in an oxy-acetylene flame burning in a total-consumption nebuliser-burner with a detection limit down to I p.p.m. at the + Prcscntcrl to the Itltcrll:ttiot~;ll Symposium on Flame Spectroscopy, Ostr:rvn, Czcchoslovd
Tcchtron

Pty

Ltd.,

G79-687

Eqringvalc

rfnal.

CJrh.

Tioacl, North

Acta,

Springvalc.

52 (1970)

259-273

260

V. SYCHKA, I’. J- SLEVIN, J. MATOUSEK, P. SEK

363,47-nm line and free from interferences from a number of elements and acids. Similar results have been pubkshed by PARBLLADA-BKLLoD~~. Recently, it has been shown that the reducing nitrous oxide-acetylene flamei+31 and in some cases (except for elements forming refractory oxides) even the nitrous oxide-hydrogen flame32*33 provide excellent sensitivity in the determination of a, number of elements by f.e.s. PICKKTT AND J
A Techtron AA-4 atomic absorption spectrophotometer equipped with an A.S.L. palladium high-intensity hollow-cathode lamp, variable atomizer, and with a u.v.-sensitive HTV R 106 photomultiplier, was coupled to an Hitachi-Perkin-Elmer, Model 165 recorder. The lamp and detector were square-wave-modulated at 285 Hz. For emission measurements, a spherical aluminium mirror was placed in front of the entrance slit of the monochromator behind the flame. The radiation from the flame was focussed with a condensing quartz lens behind the entrance slit of the monochromator and chopped at 285 Hz with a mechanical chopper placed between the flame and the monochrornator slit to match the a-c. amplifying frequency. The apparatus was modified for atomic fluorescence measurements as described previous1~20921. The monochromator was fitted with a wavelength-scanning motor. Palladium electrodeless discharge tubes were prepared as described by ALDOUS et a1.30. The tubes were supplied from the microwave power generator Microtron 200, Mk II (2450 MC s-l) coupled to a Microtron modulator unit (Electra-Medical Supplies Ltd., London) and operated both in 210 L and 214 I, cavities. Initiation was obtained by using a Tesla vacuum tester. Source and monochromator settings used for atomic absorption measurements were: primary lamp current 8 mA, booster current 400 InA, slit-width 50 ,um, band-width 0.17 nm. For atomic fluorescence measurements, electrodeless discharge tubes were operated at 75 W, and the high-intensity hollow-cathode lamp at the maximum currents recommended by the manufacturers, i.e. 25 mA for the primary and 500 mA for the secondary discharge. The monochromator slit-width was set to its maximum value of 300 ,urn (band-width 0.9s) nm). Emission measurements were made with a slit-width of 10-25 pm (band-width o.o33-o.ogg nm) . Burners and f lames Techtron AB-51 and AB-50 slot burners were used for air-acetylene hydrogen flames, and nitrous oxide-hydrogen and nitrous oxide-acetylene Aural.Chbz. Actn, 52 (1970) 259-273

and airflames,

DETERMINATION

OF PALLADIUM

BY A.A.S.,

A.F.S.

AND

F.E.S.

261

respectively, in a.a.s. measurements. Where not otherwise mentioned, measurements were taken immediately above the primary reaction zone. The solution uptake rate was 4.8 ml and 3.0 ml for air-supported and nitrous oxide-supported flames, respectively. A.f.s. measurements were performed in nitrogen-separated air-acetylene and air-hydrogen flames and oxy-hydrogen flames diluted with argon. For hydrogen flames and the separated air-acetylene flame, a specially manufactured Meker-type burner head’3 and the Techtron FE-I emission burner bead, respectively, were used. Separation of the flame was achieved as described previouslyzo~~r. The height of measurement in the flame was not a critical factor for the fluorescence signal. For all types of flame studied, measurements were taken at a height of 2.5 cm above the burner top. The sample uptake rate was set at 4.8 ml min-1 for all flames used.

TABLE BURNBR

1 OPERATING

CONDITIONS ----~

Mzlhod

h.CL.S.

A.f.s.

F.C.S.

I~lu??te

Hlia

J&-am Hz-NzO C&&-air C&I..-NaO

8.8 8.8

Ha-sir I&-02-.4r C&la--air

z.s

Hz-N20 C$&-N20

(3.6

.-_---___

c ‘2I-f‘2

Gusjlow-r~4li~

(2 mitt -.I)

.---_-

1.0

3.2

2.0 0.8

3.4

5.5 5.6

Flame emission was measured in nitrous oxide-supported flames. For the nitrous oxide-acetylene flame, a Techtron AU-50 burner head, situated perpendicularly to the optical axis was used. A nitrogen-sheathed nitrous oxide-hydrogen flame was operated in a circular Meker-type stainless steel burner head. Comparative emission measurements were also carried out in premixed nitrous oxide-acetylene and nitrous oxide-hydrogen total consumption flames in the Hetco burner (The Ditric Corporation, Waltham, Mass.) under exactly the same operating conditions as described by MOSSOTTI AND DUGGAN 20. For the premixed nitrous oxide-acetylene flame, all measurements were performed in the lower part of the red interconal zone. The optimum height of the measurement in a nitrous oxide-hydrogen flame was 1.5 cm above the burner top, and the optimal sample uptake rate was 4.0 ml min-1. Burner operating conditions for all the three methods studied are summarized in Table I. Reagents Palladium metal, prepared by dissolving

analytical grade. A xooo-p.p.m. palladium stock solution was 0.5000 g of metal in the minimum amount of aqua r&a,

262

V.

SYCHl
I’. J.

SLEVIN,

J.

MATOU:EK,

I:.

I3EK

evaporating almost to dryness, dissolving the residue in hydrochloric acid (I: +5) and diluting to 500 ml with twice-distilled water. More dilute standard solutions were made, as required, by appropriate dilution of the stock so that they contained I.SC~~ of hydrochloric acid. Solutions containing less than 5 p.p.m. of palladium should be prepared immediately before use because of the marked adsorption of the metal on the surface of bottles. Ammovaiwn pyrrolirlivte dithiocarbamate (ADPC) was synthesized by the 37. Aclueous solutions of APDC (5(yO) for use in method of MALISSA AND ‘%HOEFI’RlANN’ methyl isobutyl ketone (MIBK) extractions were always freshly prepared. All other solutions were prepared from analytical-reagent-grade cliemicals.

Analytical

working curves were obtained by spraying 0.1, 0.2, 0.5, I., 2, 5, to, 500 and TOOO p.p.m. solutions of palladium into the desired flame. For the concentration range of 0.05-x.0 p.p.m., full gain and scale expansion (5 x ) were used. For the APDC-MIBlC extraction, a 2-l separating -funnel was coated with a PTFE emulsion and annealed at 450” in a furnace in order to obtain a homogeneous protective film. The extraction procedure was carried out as follows: xooo ml of 0.005-1.0 p.p.m. solutions of palladium in a buffered medium (pr-r 2.8) were prepared directly in the separating funnel. After the addition of 5 ml of the 5% APDC solution, the chelate of palladium with APDC was extracted by 3-min shaking with 35 ml of MIBK (18 ml of MIBK is dissolved). The lower (aqueous) phase was discarded and the organic phase was transferred to a 25-ml volumetric flask, diluted to 25 ml with MIBK and sprayed directly into the flame. The duel flow-rate was reduced until a lean flame was obtained. The analytical working curves were plottecl as relative fluorescence (absorption) signal VS. concentration, taking into account the blank solution. Scale expansion (5,x ) was used.

20,

50,

IOO,

nToRnc

200,

sruI~ms

RRSORPTION

.Sensitiwii?y ad

avbalytical livt.es

Both air-supportecl ancl nitrous oxide-supported hydrogen and acetylene flames were tested, under exactly the same experimental conditions, with respect to the’ sensitivity of the pallaclium cleterrnination. Table II shows that lean flames

SENSITIVITY

OF PhLLA13IUhI

&I-Ia--air I-l&-air H~--N.~O C’LHz-NnO n Unclcr min-1.

13Y A.A.S.

IN VARIOUS

FLAMES

AT THIS 244.79-11111

LINE”

O.&j 0.250

0. ‘4.5 0.105

tlw optimal

AI&d. Chif%

I3ETIERMINh’I’ION

.dCtC&,52

conditions

(1970)

for each flame

259-273

in a s-cm

X3-50

burncr;

solution

uptalcc

rntc

3.0 ml

I>ETERMINATION

OF PALLAUIUBL

BY

A.&S.,

A.F.S.

A?*‘11 F.E.S.

263

provide a higher sensitivity than fuel-rich ones, and that the highest sensitivity is achieved in an air-acetylene flame, while the sensitivity in nitrous oside-supported flames is surprisingly small. This interesting fact is very difficult to esplain, for ionization interferences are unimportant owing to the high ionization potential of palladium (8.33 eV). It is of interest to note that also for copper both nitrous osidesupported flames provide a significantly poorer detection limit than does the airacetylene flame3”.

I,im

Lower

(,rrrr)

lcvci

ClbErgy (e 1’)

--.--___

_-.-_-

Dcklcclion limil” -.. ” .._..._- --- ..__.- ..__-. _____..._.. C2H--air ,fltrirre I-I?-tcir fiali:c .-_-- -.----__--..- --._--__-_ ..___

244.79 247.G.* 276.3 I

0

0.0.3

o.I_=j

0

0.0’1

0

O.lfj

0.25 0 . 00

3.+0+

O.t.31 _I

0.15

0.X0

324.27

0

.s I *t

O.“Jj

3G3.47 3!‘o.gC_,

0.X1_& 0 .go I

0. r5

_---n SigniLl

____

INTIfI~I~III~ISNCISS

ATOMIC

ARSOHI”rION

OF 017

INOIICi,\NIC PALL,\I>IURI

~~-----..-___~_

None

O.‘@Z

AP’-

0.427 o..+ IO

Co? 1.

0.370

I\TiZ-lHI:

(a.01

111)

ACIL>S IN

AN

ANI)

O.O‘~tr

n In il s-cm AB-50 burner 13IXrrors grcatcr than 5%

017

so-IWl.1,

,\blC>UNT?j

,\[I(-ACIJ’T\‘LISNIS

--_ ..,-....-.---A 6sov/mucc .-__---._.-_Pd (50 p.P.m.)

hlcrf?ritrg .Qwcies

1.x

.i.C’ -.-._._--...--_

0. 50

I_L_._..__.___._ : noise = z : I.

CIIIPBIICAL THE

1*.5

._.-..._... -- _.__ -_._-_-_-...

_... .._-__,__.__._... _._._ Pd /50 fi.fi.rt1.) -t_ LtaCI:, (0.5’;4,)

OF

OTHlSIl

I
ON

FLAME”

. - - .___ “__._ .__._ Dij/crcuw

(O.OI

-

((;<,)

vs.

Pd f.50 P.fi.ttr.) --I. BDT.4

- ___I-

;\I)

PC1 (JO fi.fi*??l.)

0.468 0. ,I GS

O.&X

o.qGo

__r)ta

0..*68 o..+G5 o..+Go

O.‘~C>Z

-11

0.165 0. ‘165

-80

-

-20

and for the 2+.+.7pnrn line. arc’ considcrccl as an intcrfcrcnce.

In Table III, the seven most sensitive palladium analytical lines are listed in both air-supported flames. With a high-intensity hollow-cathode lamp, the 244.7% nm line provides a somewhat lower detection limit than the more often used 247.6.q nm line, owing to suppression of the nearby ionic line at 244.62 nm. The lines originating from metastable states also provide relatively low detection limits which indicate large population of these states. This fact is more apparent in atomic fluoxscence. These results are in very good agreement with those reported by SLAV~N~~. Ivtterferevtces In contrast to some reportsly‘1, several significant chemical interferences were observed in the determination of palladium in a cool air-hydrogen flame, e.g. from aluminium, nickel, cobalt, iron, gold, hydrofluoric and sulphuric acids, etc., even when the concentration of acids in the investigated solutions was matched. Table IV And.

Chint.. Acfa,

52 (1970)

259-273

V,

264 0.4

SYCHRA

I’.

J.

SLEVIN,

J. MATOUhK,

F.

BEK

I

0.3

0.2 E x 5 4 0.1

500 Ni, p.p.m.

750

1

500

Ni , ppm.

Pig. 2. Effect

of

on palIiKlitlIl1 absorption in an air-acclylctte flatno. Fuel-lean flntnc, height 3 mn~ (X ). 8 mtn (e), nnel I 5 mm (y) ; fuel-rich flwmz, ltcigltt of the mcwtttrc8 nun (a), uxl r5 mm (a) alxwc the bttrncr top. AEG50 lmrncr, Pcl 50 p.p.tx.,

itickct

of the tncasttrctncnt

ntcnt 3 mm ( a). 244.79 nnt. Afiat.

Chit&

AGfa,

52 (I97O)

259-273

DETERMINATION

OF

PALLADIUXL

BY

A.A.S.,

A.F.S.

AND

F.E.S.

265

shows that there still remain some interferences in the hotter air-acetylene flame. The effect of nickel on palladium absorption was investigated in detail in all four flames and in various positions in the flame. The results obtained in an air-hydrogen and an air-acetylene flame are summarized in Figs. I: and 2. These results indicate that the nickel interference is more pronounced in a fuel-rich flame and strongly depends on the heigllt of measurement in the flame. The liiglicr the position in the flame, the higher the absorption signal and the smaller is the effect of nickel. In a fuel-lean airacetylene flame and at a height of S mm above the burner top, no interference from a zo-fold amount of nickel occurs. In a fuel-rich air-acetylene flame and in both types of the air-hydrogen flame, nickel affects the palladium absorption even when the measurement is carried out at higher positions in tile flame. Tllere is no evidence of the interference of nickel in flames supported with nitrous oxide in any investigated position in the flame. The effect of cobalt on the palladium absorption shows similar clependences. From Table IV it can also be seen tliat the effect of interfering species can be completely eliminated by adding releasing agents, such as lanthanum chloride or EDTA. No interference was found from zo-fold amounts of Ru(IV), Ir(IV), Rh(II I), Os(VI), so-fold amounts of Pt(IV), Au(lII), Cu(II), Fe(lII), %n(II), Si(lV), Ca(II), I<(l), Na(l), Mg(II), Ca(II), Sr(II), W&2-, MOOA+, or from o.or-o.5 M HCl, HN03, HC1O.l, H&O.r, and I-IsPOe. Al’OJlIC

Excitation

I’LUOI
of atomic

S’I‘UlJIES

_fluorescwce

Several palladium discharge tubes made by the method of Ar.~:~orrset al.“fl were used in attempts to excite palladium fluorescence. They were operated both in the I/g-wave and 3/,$-wave cavities at ~10-60 W. The emission spectrum of palladium obtained was similar to that of the authors mentioned, the 340.46nm line being the most intense. However, even wlm~ a very long “running-in” period was usecl, the discharge was rather unstable, much weaker than espectecl, and sometimes, after a long operation time disappeared completely, probably because of insufficient recombination of palladium chloride in the tube (localized “palladium mirror”). Under these conditions, the tubes were very difficult to handle and detection limits for palladium fluorescence clown to I p.p.m. could not be achieved. A palladium high-intensity hollow-cathode lamp proved much more satisfactory and was used as excitation source. This lamp run under the conditions described above was found to be a sufficiently intense source for exciting atomic fluorescence. In order to find the optimum gas flow-rates (see Table I), the clependence of background emission of the flames and of the fluorescence signal on the fuel-tooxidant ratio was investigated. For the separated air-acetylene flame, the dependence of the fluorescence signal on the acetylene flow-rate exhibits a flat masimum. From Fig. 3 it is apparent that the fluorescence signal both in the air-hydrogen flame and hydrogen-oxygen-argon flame (at a given oxygen flow-rate) is strongly dependent on the hydrogen flow-rate (particularly for the latter flame). To choose the most sensitive fluorescence lines, the relative intensities of lines emitted by the excitation source and tile corresponding fluorescence lines in all the fInal. Chi??Z. ACfn, 52 (1970)

259-273

266

V. SYCHIIA,

.

125

P. J. SLEVIN,

J. MATOUkK,

I?. BEK

.

100

# I-

1

2

a

3

2

5

6

hydrogen flow-rate

,

7

0

9

l/rnin

Fig. 3. Vnriation of rolativc Iluorcsccncc intensity for pnllaclium nt 3‘+0.+6 nm in nir-hytlro~cn, :Lrgon-hyclrogcn, ant1 hycl~ogcll-oxy~:ctI-argoll Ilanics. Air--1iytlrogcn (a), argon-hydrogen ( 0). Ilydrogcrl-oxygcll-nrgon, osygcn flow o.zo I min-* (a), 0.50 1 tnin-1 ( A), o.G5 I ruin-l (A), and 0.70 1 min-1 (+).

three flames were determined. In scanning the fluorescence spectrum, a Io-p.p.m. solution of palladium was sprayed into the given flame under the optimum conditions and at a slit-width of 50 &cm.The emission spectrum of the lamp was scanned at the maximum operating currents and at a slit-width of IO ,um.

Spectral characteristics of the eight most sensitive palladium fluorescence lines are listed in Table V. Besides these lines, weak fluorescence emission was also observed at 355.3r,348.98,348.12,346.o8.344.r4~ 343.35337.30~276.31, and 247.64 nm. The highest fluorescence intensities correspond to the lines having lower electronic states at 0.814 and o.gGr eV (3D3 and 3D2 multiplet levels, respectively). Considering the higb population of the 3D multiplet levels (see atomic absorption sensitivities for corresponding lines), the measured fluorescence signals for the lines originating on these levels are mostly due to thermally-assisted resonance fluorescence38vau. In some cases, contributions from direct-line fluorescence (e.g. for the lines at 325.16, 35X.69, 330.21 m-n) or from stepwise-line fhOm!SCenCC! (e.g. for the lines at 324.27,363.47 nm) could be expected, because of the relatively low intensities of fluorescence corresponding to resonance transitions to the ISO level (see palladium atomic term diagram). It is very interesting to note that the relative fluorescence signal for the Anal. Chilot.Actrc,52 (1970)

259-273

DETEitRIINATION TABLE

OF

PALLADIUM

BY

A.A.S.,

A.F.S.

ANI)

F.E.S.

267

v

KEI_ATIVE

enIIS!?.ION

AND

FLUORESCENCE

INTENSITIES

OF

PALLADIUM

lllbcrgy (e I.-)

Liue (Tarn)

.._____

0.81.}--.}..~5‘f o.8r.p4.224

O.g_jI--.~.3cj5 7

3 2 ‘+ . z 342.12

o.HI.i-.}.fijCl 0.961-4.584 o.9rG-,(.,pG

351.69 24.1.79 330.2

o--j I

_._-

i~clrtlivc itblmsify

I~tJClS

__.______.__~_--____~ 3W.‘IG 3fi3.47 360.9cJ

.&3

I. 25‘1-5.005

.-....--

I* Cw-rcctcd

fur

__--

ct)bissimbu

__._ _..--_._-_--.--_.--.

i~~Intive/I~rore~ce~rcc~ irrIctbsily~k ___.____---.-__.._--_----_ f-f~-o:!--:I, .-I ir-l;T2 :I ii*--Cal-ia ._._...-_-._-..---_...-_

. .-.-

100 90 6 I

IOCI 98

100 8v

40

39

20

49 43 ‘10

32 19 ‘7

27 19 19

3’ IX 18

9

IO 7

‘7

‘S

12 ________._-

-

LISES

-

____.--_.-_

._.._ .__--

clctcctor rcspoiisc.

__..__-

-

100

iq

IO _______ _.__ --_---..-_-

.._.

IV . ..- .._-. I.

363.47-nm iine varies significantly witlr tire flame used (see Table 17). Tire iriglrcst absolute fluorescence signals for all fluorescence lines were obtained in the osylrydrogen flame diluted with argon, owing to the Irigller fluorescence yield factor than for other flames because of tile quenclring cross-section of argon being much smaller than for nitrogenarc. A 2.2, z.S, 2.3, 2.6, 2.2, 2.0, 1.2, and x.3-fold increase, resulting from using the hydrogen-osygen-argon flame in place of the air-lrydrogcn flame was found for the 340.46, 363.47, 3Go.96. 324.27, 342.12, 35x.69, 244.79, and 330.2x-nm lines, respectively. Awdyticnl zuovkirzg c~irves and tlctcctimt limits palladium III l?g. 4, the analytical working curves for tile most sensitive fluorescence lines in tile oxy-liydrogen-argon flame are slrown. Tile sllapc and slope

’ o!--

I 1

01

10 Pd, Wm.

Fig. ,I. An~~lytical working cx.n-vcs for palladium fhlllc. Illll.

(X) (a)

340.46 2763T

11111,(u)

363.47

11111, (I)

1000

100

360.96

lines nlll,

in the hydrogcn-osygcrl-;lr60rl

obscrvcd (-+)

351.69

11111, (A)

2<14.75,

llm,

( A)

257.6.t

lllll.

A IbiZl.

Chiltb.

AC/U,

52

(1970)

2fjg-273

SYCHRA’,

V.

268

D15’rlLCTION LlhlI’fS

FOR

PALLhI>IUhl

Idine (?WL)

ANALYTICAL

LINES

/I ir*-f-f 2 -..- ._-__. --

3.tv..tO 363.47 jGo.gO 3.}2.12

o.oat 0.05 0. LO 0.20

‘2++.7CJ

v.zfJ

35r.w 324.“7 330.2

0.30

0.5”

I

0.50

P. J- SLEVIN,

13Y

J. MATOUSEK,

P. BEK

A.P.S.

A iv-Cpi-Is -----

O.OiJ

0.08

0.12

0.18

0.20

0.30

0.35 0.25 (>.‘I5 0.25 0.50

o.,to

0.35 0.50 O.Zfj

0.70

1’ Signal : Iloisc z-2 2 : I,

ClIIShlICAL ATOMIC

(5

INT15I~li111~15NCISS

C)lr

1~I.UOIlI~SC1~NCI5

p.p.111.

i-ICI

of

n.l

OF

0.1

nfr

0.5

ivr

1-rI.;

A4

0.5

h!f

-I-

-

01’

O’TlIISIi

IONS

ON

-XI

-

nfr 0.1 n/r

--

WO.I”--

18 -t_ 18 -t_

I I

-

IOL’

-

-+-IO

7

-ti~~i~~ur~~n~I~ts

n:in-1. I1Errors grciltcr c zg-fold csccss. ” Not tlctcctecl.

wcro

than

7

N. I>. ”

-I-

I$l.t-

G

-

-6 -

“---t- 9 -t- 1-t

‘,“~.,‘,.

-

-go - 92 -I-3 0

1;; -+- ;

1\1:,-1. A I, :I+

hll

AMOUNTS

-7

-__

co’r-I. Ni 1+

‘1

IOO-I’OLI)

-

IO

--

0.01

l;&~k

AND

-

-I- 9

0 5 n!I 0.1

AC1135

talccll)

l~fzso., 0.1 lb1 I-t:Ipo.i

1NORC;ANIC

I’AI~LADIUM

~nnclc

20

3nIn

ihovc

.

tlic burner

top ancl with

sample

uptnlcc rate of dt.8 ml

5’l/o arc consitlc:rccl as ntl intcrfcrcncc. _

of these curves is in good agreement with theory‘*1*42. The curves are linear with concentration over a range of more than three orders of magnitude with curvature appearing for the resonance lines at lower values of concentrations. Table VI summarizes detection limits of the eight most sensitive palladium fluorescence lines in all the three flames used. The best detection limit of 0.04. p.p.m. was obtained for the 34o.&-nm line in the oxy-hydrogen-argon flame. As expected, this flame provides better detection limits for all the lines than the two other flames, except for the 324.27-nm line. For this line, a high noise of measurement in the oxy-

DETERblIN~TIOh’

OF

I’RLLADIUJI

l3Y A.A.S.,

A.F.S.

AND

F.E.S.

2%

hydrogen-argon flame in this spectral region causes deterioration of the detection limit compared to the other two flames. An approsimately z-fold increase in the fluorescence signal in the separated air-acetylene flame compared to that in the airhydrogen flame does not result in better detection limits, because of the much higher noise. Interferermes

Chemical interferences of inorganic acids in concentrations of 0.0x-0.5 M and of xoo-fold excess of other ions on the fluorescence signal produced by 5 p.p.m. of palladium were investigated. These are listed in Table VII for the air-hydrogen and oxy-hydrogen-argon flames. Depressive effect of acids (escept for phosphoric acid), nickel, and cobalt can he eliminated by addition of strontium (chloride) or EDTA to a final concentration of 0.27L and 0.01 M, respectively. The depressive effect of plrosplloric acid can be eliminated only by addition of EDTA. Practically the same species were found to interfere in tire separated air-acetylene flame as have been described for atomic absorption. However, tlrese interferences wcrc more pronounced in a.f.s., particularly for nickel and aluminium. 13esides these interferences, a slight depressive effect of platinum was observed. All cationic interferences in all the three flames were found to be strongly dependent on the height of the measurement above tile burner top and on tile concentration of acids in the analyzed solution. No interferences were found from a roo-fold escess of Naf, I<+, Ca”+, Mg”+, hiIn?+, Crs+, Hi”+, %n”.+, Cu”+, Rlr:j.+, MoO4”-, VOS-, OsO.r”-, RuOJ”-, SiO,i”-, and I-[NOS.

Il’lter?,2al-enlissiort cltamcteristics The ntomic emission of palladium was examined both in nitrogen-sheathed premised nitrous osidc-acetylene and nitrous aside-llydrogen flames, unseparated premised nitrous oxide-acetylene flames, and in the premixed total consumption nitrous osiclc-acetylene flame”“. In Fig. 5, the variation of relative emission intensity with -fuel flow-rate in the nitrogen-sheathed nitrous oxide-hydrogen (5a) and nitrous oxide-acetylene (5b) flames is shown. Both dependences exhibit a masimum. The optimal flow-rates are listed in Table I. The best signal-to-noise ratio (approsimately equal for both fuel gases) was obtained in nitrogen-sheathed flames, as expected. Considering this fact, all other measurements were made with the nitrous oxide-hydrogen flame, owing to its much simpler and safer operation. Table VIII summarizes the relative emission intensities of the ten most sensitive emission palladium lines. The most intense atomic emission was observed from the lines which result from the 5seD j 5p3P, gs3D * ~$317, or gs3D --f 593D transitions. A very weak emission was observed for the 244.79-nm resonance line. Analytical zwork.ing tames and detection limits The most sensitive line at 363.47 nm gives a detection limit of 0.04 pp.“. which is the same as that for the 34o.4G-nm line in atomic fluorescence (see Table VII). The detection limit of the 324.27-nm line is poorer, owing to the strong background emission in this region of the spectrum. Analytical working curves for the 363.47 and A9rd.

chinl.

l~ckz,

52

(X970) 259-273

V. SYCHRA,

270

L

6

O5

,

7

Hydrogen flow -rote,

J. MATOUSEK,

1:. BEK

I 10

9

8

I’. J. SLEVIN,

llmin

75 b

I

3.25

O3.00

3.50

Acetylene

flow - rate

3.75

c

vrnin

Pig. 5. Vnri&tion of rclativc cmiusion intensity with a fuel flow-r&c for the palladium lino in nitrous oxitlc-hyclrogcn (5~) and nitrous oxiclc-:rcctylcnc (sb) flkulx%4. TABLE RIU.Al’1V.E

VI I .I JNTlZNSITIE3

SPECTROSCOPY

IN

ANI> A

DK?I7X’J.ION

SRPA1~ATE.D

NITROUS

LIhJITS

OF

P,\LLADIUM

OXIDE-HYDROCRN

OsciE Idor StYe~tglh~~

303.47 340.46

I.2

0.0‘1

2.0

0.05,

360.96

1.8

0.18

324.27 342.12 351-l-)9

I .2 1.5

346.08

o.Go

DY

Llctcctio~b li9rzi11)-(%:p.nr.) I__--.--_____

348.12

2.0

355*3r 244.79

3 ’4 0.074

2.0

0.70 0.70

1.2

* Corrcctcd for clctcctor 11Signal: noise = 2 : I.

LINES FLAME

(M 02)

r*h?

363.‘\7-nm

;:z I.5 20

response.

Aural. China. A.&z, 52 (1970)

259-273

I’LAME

TJJl:liMAL-IShIISSIOh’

DETERMINATION

1NTERFERENCE

OF

EFFECTS

PALLADIUM

OF

GTHEK

BY

IOXS

A.A.S.,

AND

A.F.S.

AND

COhIPOUXDS

F.E.S.

ON

THE

271

TH~R~IAL-EZIISSION

FLAME

OF

PALLADIUM

-

---_

r/e spccics I mlr

t kg

iielulive iuteitsity o/sigrrcil -.__-..._.___ -I’d (20 p.p.,rt.) Ide~fwcwl + irrterJercut (.-I) without -_----___.__---..363.47 340.46 363.47

340.46nm lines plotted in the coordinates tion of 1000 p.p.m. of palladium.

--._ ._

I’d

(U) _.._ - ..340.46

__---1~i’Jcr~~~c (.*I

--

U)

-__._--363.47

3.1046

log I 7)s.log C are linear up to a conccntra-

The effect of so-fold amounts of a range of cations and anions on the emission intensity produced by IO p.p.m. of palladium under optimized conditions and by using a slit-width of 25 pm was investigated. The following elements did not interfere: R113+, Pt4+, Ir3+, Au3+, A&, Ni3+, Co3+, Ala+, Znz+, Cu3+, K+, SiO4”-, hIoO43-, OsO43-, Ru043-, HCl, HN03, HC104, H~SOI and H3P04. On the other hand, Fe3*, Na+, Mg3+, and Ca3+ caused a significant increase of the palladium emission signal. Tljese interferences are listed in Table IS for the two most sensitive emission lines. It is difficult to assign the cause of these interferen
The MIBK estracts of palladium chclate with APDC gave a scale rending approximately 4.5 times greater than that of aqueous solutions of the same concentration. Following the extraction procedure described above, detection limits of O,OOO~ and 0.0003 p.p.m. were obtained for the atomic fluorescence and atomic absorption technique, respectively. In atomic fluorescence, linear analytical working curves were obtained over the concentration range of O.OOI--1.0 p,p.m. The extraction procedure in combination with the emission method was not tested. One of us (I?. J.S.) wishes to thank the British Council for a Cultural Exchange Visit to Czechoslovakia.

A comprehensive investigation of the optimal experimental conditions for the determination of palladium by atomic absorption, atomic fluorescence and flame thermal-emission spectroscopy was made. All measurements were carried out with a Anal.

Chiw.

ACtU,

52

(1970)

259-273

V. SYCHIZA,

272

I’. J.

SLEVIN,

J. MhTOUhK,

F. 13EK

absorption spcctropllotomcter. A Iiigll-intensity modified Techtron AA-4 atomic hollow-cathode lamp was usecl as source both for atomic absorption and atomic fluorescence measurements. Atomic absorption was measured botll in air-supported and nitrous oxide-supported llydrogen and acetylene flames. The best detection limit of 0.03 p.p.m. was obtained in an air-acetylene flanic at the 244.79-nm line. Cationic and ariionic interferences were stucliecl in clctail. The atomic fluorescence of pallaclium separated air-acetylene and oxy-hydrogen-argon flames in premixecl air-hydrogen, The complete fluorescence spectrum was studied ; the most was also investigated. intense fluorescence line at 340.46 nm gave a detection limit of 0.04 p.p.m. in the osy-llydrogen flame diluted with argon. Flame tllertiinl emission cliaracteristics for a nitrogen-sl~cntl~ed nitrous oxide-llyclrogen flame are presented. Relative elnission intensities ancl sensitivities of ten lines are tnl~ulatccl. The best detection limit of 0.04 p.p.m. was found for the 363.47-nm line. Possible spectral interferences are discussed.

Une rechcrche est effectuce sur le dosage du palladium par absorption atomiquc, fluorescence atomique et spectroscopic cl’8mission tliermique. ‘routes les mesures ont Ctd effect&es h l’aicle d’un spectropllotom&tre par absorption atomique, Techtron AA-4 modifid. Une lampe B catllode creuse, de forte intcnsiti:, est utilisBc comme source, soit pour l’absorption atomiclue, soit pour les mesurcs de fluorescence atomique. La meilleure limite de d&tection de 0.03 p.p.m. a rX obtenue dans une flamme air-ac&hyl&ne 5 244.79 nm. On examine en detail les interferences cntioniques ct anioniques. On etudie ensuite le spectre cle fluorescence complet ; la raie la plus intense ZL340.46 nm donne une limite cle cl6tection de 0.04 p.p.m. clans une flantmc oxyhydroghne, dilude avec argon. En ce qui concerne l’&nission tllermique, on a trouvd la meilleure lirnite de detection de 0.04 p.p.m., ri 363-47 nm. On examine les interf& rences spectrales possibles.

Es wurcle eine umfassende Untersuchung clcr optimalen expcrimentcllen Uedingungen fiir die Bestimmung von Palladium clurch Atomabsorptions-, Atomf luorcszenzund l&mmen-entissionsspektroskopie durcllgefiilu-t. Alle Messungen erfolgten unit einem abgewandelten Techtron AA-4 Atomabsorptions-spektropllotometer. Als Lichtquelle fiir Atomabsorption und Atomfluoreszenz wurclc eine Holllkathodenlarnpe hoher Intensitlit verwenclet. Die Atomabsorption wurde in mit Luft und m.it Lachgas unterhaltenen Wasserstoffund Acetylenflanmmen gemessen. Die beste Nachweisgrenze von 0.03 p.p.m, wurclc in einer Luft-Acetylen-Flamme bci der z44,7g nm-Linic erhalten. Kationisclle und anionisclle Stiirungen wurden im einzclnen untersucht. Die Atomfluoreszenz wurde such in vorgemiscllten Luft-Wasserstoffsowie nicllt vorgeinischten Luft-Acetylenund Sauerstoff-Wasserstoff-ArgonHammen gemessen. Es wurcle das vollstiindige l~luoreszenzspelrtrum untersucht ; die intensivste ITluoreszenslinie bei 340.46 nm ergab eine Nacllweisgrenze von 0.04 p,p.m. in cler mit Argon verdiinnten SauerstofE-Wasscrsto,ff-Flamme. Es werden die Emissionseigenschaften in einer von Stickstoff umgebenen Lachgas-Wasserstoff-Flamme mit den relativen lGnissionsintensit!iten und Empfindlichkeiten von zehn Linien Rwal. Ckiwa. Ada,

52 (1970) 259-273

DETEIZBfINr\TIOS

OF

I’~~LLAL)IU,\I

BY

A.A.S.,

r1.F.S.

AND

F.E.S.

273

angegeben.‘.Die beste Nachweisgrenze von 0.04 p_p.m. ergab sic11 bei der 363.57 nmLink. Es werden m8gliclie spektrale Stijrungen erijrtert.