Quantitative analysis of solid mixtures by diffuse reflectance measurements

Quantitative analysis of solid mixtures by diffuse reflectance measurements

AXALYTICA 108 QUANTITATIVE REFLECTANCE ANALYSIS OF SOLID MIXTURES I3Y CHlMICA ACTA DIFFUSE M~ASUIWMISNTS ‘I’llc factors nffcrcting the dif...

548KB Sizes 2 Downloads 140 Views

AXALYTICA

108

QUANTITATIVE REFLECTANCE

ANALYSIS

OF

SOLID

MIXTURES

I3Y

CHlMICA

ACTA

DIFFUSE

M~ASUIWMISNTS

‘I’llc factors nffcrcting the diffuse reflcctancl! spectrum of an rtb:iorl>ing substancc clilutcd x(9- to lo”-fold with a non-absorbing St;lMli~d llavc Iw31 investigatccl by I
A Unicnm SP 500 spcctrol,llotometcr littcd with a. standard diffuse reflectance attacliment was used to measure reflcctanccs in the range ~270-1000 mp. Solids

DIFFUSE

REFLECTANCE

SPECTROSCOPY

Oh’ SOLID

hi ISTURES

109

absorbing in this region were lead monoxide, which has an absorption maximum, as measured by diffuse reflectance, at 5x5 rnp, silver iodide, absorption maximum 425 m,u, and zinc oxide, absorption maximum 36s m,u; as solids not absorbing in this region magnesium oxide and silicic acid were used. Mixtures were prepared by nlechanical shaking of components passing zoo RS., or finer, mesh ; the compositions of all mixtures are expressed as percentages by weight. Matt sample surfaces were preparccl by arranging the sample in a heap in the sample-holder of the reflectance attachment and using a rubber bunhi to remove cxccss material and to lcvcl the surface, taking care not to compress the sample in the holder. All reflectanccs were measured relative to the non-absorbingmatcrial in the particular mixturewhich was taken as 100% rcflcctance. Pcrccntagc rcflcctancc hils been sllown to hc clepcndcnt on particle size”. It \vas found, however, that the cxpccted variittion of reflectance with clcg~e of grinding coulcl bc eliminated by hancl-l;!,inding of the mixtures in iLn agate mortar for about 15 min; grinding for longer periods resulted in no further change in rcflcctancc. This is illustrated in Table I for a mixture! of silver iodide and magnesium nsidc containing

ICPI’HCT

01’

1’HOI.Osr,;lit>

us

GHINI~lSG

lil:l=I,liC’rhSCI~

012

A

SII.VI’I~

IOl~lUI~--

JlhGSIfS1Uhl

OSII>IS

~lls’clfI
-

Total ‘X

_

_ grindinK

rcflcct;kncc

,_

.

-

titnc (tliin) nt ,125 III/~

0 ‘,‘).S

_ _

‘2 .I’). 8

._ _. ._,

_

. . ..

.s 5 27,” _jO.!, .. . ..-

_..

_

. .._....___

IO I.? 2*3.(> 21.7 __. . . .

_ ’ 4

~3~5 ,.

.__ . I (i

Z”.c)

. __ 20

._ 30

13.2 23.4 ..--_- ..-^-... -.

9.0”!, silver ioclide originally passing 300 mesh. During this work and work to 1~ described elscwhcrc, reflectances were measured for several cliffercnt compositions of cncli of 20 cliffcrcnt mixtures; in every case grincling for not longer than 15 min resulted in a limiting value of reflectance. The reference substance was always ground for the same lcngtl~ of time as the test material so that its packing characteristics would bc similar. The effect of unavoidable variations in surface preparation on the measured reflcctancc of a given mixture was invcstigatcd. It was found that variations in the standard surface affcctecl the reflectance of the test sample equally at all wavelengths. Hence the cffcct of such variations was reduced for all substances with a non-absorbing region in the range 270-1000 nip by expressing reflcctanccs in an absorbing region relative to the reflectance in the region of no absorption. For 3.0”/~ zinc oxidcmagnesium oxide mixtures, the standarcl deviation of the reflectance measured as described was 2.40/O. To reduce the effect of variations in the preparation of the sample surface, the reflectance of a given sample was always measured usually for 4, and never for less than 3, separately prepared surfaces and tlw mean value used. RESULTS

The variation of reflectance, at the wavelength of maximum absorption, with concentration of the absorbing component was investigated for 5 two-component Aad.

Cirirn. Actu, 33 (1965) 108-I 14

W.

LX0

P. DOYLE.

1:. FORBES

mixtures. As shown in Fig. I, for all the mixtures investigated a linear relationship was obtained by plotting log A/Z2 (R = percentage reflectance; A = x00 - R) against log W ( W is the percentage by weight of the absorbing component). The slope of the lint was, within cxpcrimental error, independent of the particular mixture and

LOO

w

ffl

-

12

-

e

PbOb.QO

C

AOl/MOO

L ZnO/Moo

io aa- t

72

err

75

zJ3 LW

OD

0.2

c.4

0

4

2

-f$

6

8

% Pbo

Fig. I. lJlot of log (cl /II‘) 11s.log IV for Yotnc two-conlponcnt irbY0lption. Fig. 2. Plot of (A/fC)~~J~:‘vs. LV for Iwcl inonoxidc-silicic

rnixliircs ;hrl

wavclcngtli

of wilximum

lnixturcu zrt il nunihfrof

wavclcngths,

ilt

tlic

the mean value of the slope was 1.383 so that the results may be rcpresentcd by the equation IV = K (A/Z<) I.383 whcrc the value of ZCclcpcnds on the particular mixture. This equation liolcls when ;hc mcnsured rcflcctnncc is in the range 3o-7oo/o; outside this range a small error in reflectance results in a compnrativcly large error in (A/Z?) l*z~*:l so that the vnliclity of the equation cannot bc assessed. Of the 20 investigated mixturcs containing different components, only zinc oxide-aluminium oxide mixtures did not conform to the equation; the results for these mixtures arc given below. The quantity (A/R)l.“H” is subsequently termed the “absorption function”. It was found that the linear relationship between the absorption function and concentration hclcl at all wavelengths and not only at the wavelength of maximum absorption. Figure 2 shows the results for 4 diffcrcnt concentrations of lead monoxide in silicic acid at a number of wavelengths. Thcsc results were applied to the quontitativc analysis of two-component mixturcs by preparing a calibration plot of absorption function VS. concentration for the particular mixture ancl interpolating the mcasurcd absorption function of the snrnplc of unknown concentration. The results for a number of indcpcndcntly prepared zinc oxide-magnesium oxide mixtures are shown in Table II. The standard deviation from the actual concentration was 6.4%. /I oral. c/JiJJJ.

Aclrr,

33

(rgG5)

rot&r

I,#

DIFFUSE

REFLECTANCE

TABLE

11

RESULTS -_

OF

%

SPECTROSCOPY

ANALYSIS

%nO

01’

Actual

-_.-__-

EFFECT

OaSfDR

0.70 0.70

TflERhfAt Tkt’.AThlf?NT OF COPPICR(It) OStDE OXIDE--IClAGNf?ZIUM OStDE MlXTURI:S -.._---.-_ .._- . ..._ -..... _-.________ _-... (..l/l~)f*3@~ -_..-.___I-.___. .__-.- . .._ (0) 750 qc 6.50 rrrp

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

. _... ._-.- ._..._____,_____

650 850 950 1000

---

MIXTUllES

0.73 OS77

III

AltZ;TURIIS 0.81

I.12

I.18

0.83

I -23

I.IL

1 II OT

COPPXX(II) ----_----. Yrclrculiwg tonpcYntlrre

-__

OXIDE--hlACNi~StUM

0.14 0.59 0.13 0.56 _-..._-_-____,._~--____.-

Found

TABLYi

ZINC

01; SOLID

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

.

ON

AffSORPTION

FUNCTION

l’Of2

3.6%

,___.. ..__

7.00

4.20

3.40 0.61 cl.36

2.02 0.75 O..lZ

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

,____

Analysis by cliffusc reflectance is alq~lical~le to pourly crystallized materials, which could not be detcrmincd by S-ray powder diffractiorr, as well as to matcrinls of larger crystallite size. For csample, it was found that thcrc wcrc no lines in the powclcr diffraction photograph of alun~iniun~n~olybclatc hut this substance could be clctcrmined in mixtures with molybdenum trioxiclc and magnesium cbxide by diffuse reflectance mcasuremen ts. For some substances the reflectance proved to bc dcpcnclcnt on the previous history of the particular sample. For cxamplc, the effect of pm-heating coppcr(II) oxide to various temperatures for 24 h on the value of tllc absorption function at two wavelengths is shown in Table I I I, The following deviations from the behaviour described above were observed for zinc oxide -aluniinium oxide mixtures: (I) mixtures ground to a limiting reflectance absorbed less strongly than mixtures of larger particle size; (2) the variation of absorption with concentration did not conform to the equation IV = K (A/Z<)i-3”3 but was in accord with the KUI~ELKA-MUNK function as shown in Fig. 3; (3) in mixtures with a variety of non-absorbing substances zinc oxide began to absorb between 410 and 425 m,u but in zinc vxidc-aluminium oxidc mixtures there was a bathochromic shift to 450 rnp. This behaviour corresponds to that observed in mixtures where one component is a&orbed on the other1 and suggests that the zinc oxide is aclsorhecl on the aluminium oxide. Thus in general the proposed analytical method may not apply when thcrc is physical interaction between the components of the mixture. For a three-component mixture, the region of no &sorption is in general followed at shorter wavelengths by a region in which only one component absorbs and this region is followed by a region in whiclr both the first and a second component absorb; determination of the second component by rcflcctancc mcasurcmeuts iti this region requires that the absorptions of the two components bc additive. To test for additivity, the reflectance spectra of 3 lcad monoxide-zinc oxide-silicic acid mixtures of different concentrations were measured and for each mixture, at all wavclcngths the AM/.

C/&u.

,4&t,

33 (1965)

108-114

W.

I12

I’.

DOYLE,

T;. FORBES

absorption function was founcl to bc additive, as illustratecl for one of the mixtures in Fig. 4. The validity of the analytical method for three-component mixtures was investigntccl by mc;~surcments on lead monoxide-zinc oxide-silicic acid mixtures. Calibration curves were first prcparcd for lead monoxide-silicic acid mixtures at

0.0

as

10

15

2.0

25

lh Zno

Fig.

ikcitl mixture. function for zinc oxidc- lcatl rnonoxitlc-silicic :witl; A spcclrurn of o.CqLyU %nO-silicic acid ; 0 sun1 of spcctrn PbD-nilicic ;iciclXllCl o.C,.~~~~ zinc oxitlc-silicic :~cicl; l obscrvccl spcctruni of 2.020/, %nO-silicic acid.

‘1. Atltlilivity

of almorption

0 spoctrutn of 2.02’)$ Z’LO-silicic

of

2.0274 rw3-0.c,g~%,

DIFFUSE

REFLECTANCE

SPECTROSCOPY

OF

SOLID

BIIXTURES

1x3

rnp and for zinc oxide-silicic acid mixtures at 368 nq. Also the ratio of the absorption function at 368 rnp to that at 5x5 m,u was determined from measurements on several lead monoxide-silicic acid mixtures; the mean value was 0.689. To analyse a lead monoxide-zinc oxide-silicic acid mixture, the absorption function was dctermined at 5x5 v, whence the percentage of lead monoxide was derived, and at 368 rnpu; in the latter value, allowance was made for the absorption clue to lead monoxide by subtracting 0.689 times the value of the absorption function at 5x5 rnp thus giving the value of the absorption function at 368 rnp due to zinc oxide alone. The results obtained are shown in Table IV. The standarcl deviation from the actual percentage was G.GOk,for zinc oxide and 5.20/Ofor the lead monoxide. 515

RISSULTS

01’ ANALYSIS

O.SII)I’. --sII.1CIc

Ll3AlJ

hlOh’O.X11)lS--ZINC

7.15 5.69 5.22 ,).08

3.05

3.26

5.93 4.95 3.94

.Z.GO

2.G2

2.90

2.82

2.29 I.87 I,?8

S&1

I.99

Z.IG 1.72’ 1.27 O-S3

o.gs

I

2 3 4 2

01’

ACID

XlISTUI~ISS

0.80

CONCLUSIONS

Advantages of the method described are the accuracy obtainable in the analysis of two- and three-component mixtures and its applicability to poorly crystallized materials. Since reflectance may depend on the previous history of the particular sample the method is most useful in applications where the calibration curves can bc prepared from the same samples as are present in the mixture to be analysed. The method would be less useful in the analysis of an unknown mixture; in such a case the components present mi&t be identified by infrared spectroscopy or by X-ray powcler diffraction and the mixture then analysed by diffuse reflectance measurements, but the results would be of varying reliability, because of difficulties in preparin8 calibration curves.

Atomic

This work was supported in part by the Metallurgy Division, Energy Research Establishment, Harwcll, Berks. Anal.

Chim.

United Kingdom

Ada, 33 (1965)

x08-114

W.

1x4

I’.

DOYIZ,

F. FORBES

SUMMARY

The factors affecting the reflectance of an absorbing substance mixed with material not absorbing at the wavelength of measurement were investigated and it was found that the effect of particle size could be eliminated by prolonged grinding. It is shown that the percentage by weight of the absorbing substance is directly proportional to the quantity (A/K)l*3”3 (f< = percentage reflectance; A = x00 - Z?) and that this quantity is additive for z absorbing substances. A method is developed for the accurate determination of low concentrations of one and two absorbing substances mixed with non-absorbing material. The m&hod is applicable to poorly crystalli;r.ecl substances which give no characteristic X-ray powder diffraction patterns as well as to crystallir~c solids. ‘I‘he rcflcctancc of some substances is shown to be clepcnclcnt on the previous history of the particular sample. The method is not applicable if there is pllysical interaction between the components of the mixture.

Lcs autcurs ant examind lcs factcurs influensant le pouvoir r&lectcur de substances absorbantes, n&langdes h dcs procluits non-absorbants, ZIla longueur d’onde de la mesure. lJnc m&hodc est cl&veloppde pour le closagc prbcis de faibles conccntrntions cl’unc ou clc cleux substances absorbantes, m&langtZes rt cles corps nondwwbants. Ce proc&l6 pcut Gtrc qqdiclub aussi bicn h clcs substances ma1 cristallis6es clu’h clcs soliclcs cristallins. %UShMMENl:ASSUNG

Es wurden clic Fnlctorcn untcrsucht, die die Reflcktion eincr absorbicrcndcn Substanz bccinflussen, die mit Material gcmischt ist, das bci dcr Messwcllenltinge nicht absorbiert. Es xcigte sich, dass dcr Einfluss dcr Tcilchcngrbsse durch vcrl%gcrtes Mahlcn climiniert wcrden kann ; ferncr. dass clcr prozcntuale Gcwichtsanteil der absorbiercnclcn Substanz dirckt proportional ist der Griissc (A/R)l.3H3 (IZ = prozcntuale Rcflektion ; A = IOO -- Z<), und class cliesc Grijssc ftir zwci absorbicrende Substanzcn aclclitiv ist. Es wircl cinc Methoclc entwickclt ftir die genaue Bestimmung nicdriger Konzentrationcn von ciner odcr zweier absorbicrcnder Substanzcn, die mit nicht absorbicrenclem Material gcmischt sind. Die Methodc is anwcndbar sowol~l fUr schlccht kristallisierenclc Substanzen, die kcine charakteristischen Iiijntgcnbeugungsaufnahmen zcigen, nls such filr gut kristallisierenclc Fcstkijrper. An einigen Substanzen wird gezeigt, class clic Reflcktion von ihrer Vorgcschichtc abh5ngig ist. Die Methodc ist nicht anwcndbar, wenn clie Komponenten der Mischung sich physikalisch beeinflussen. .RHl’ElXENCES I G.

ICORTtiM,

2

f
\v. IJRAUN AND ct. IIERZOG , .d?rgeru.Chcr~. Iutmt. Grl. En&., J. VOGEL. f~upw. Ch)rr., 71 (1~59) 451. 3 11, w. ~7rsctxE~ AND F. VHATN~, /flurl. Chiru. /I&a, 13 (1955) 588. 4 C. A. Lmnlom AND L. 13. Rooms, Anal. Chcm, 27 (xQ55) 340. 5 iM. BILLY AND h. 13HR’L’ON, Co,,lfit. &lId., 206 (1938) 1958. G.

A9raf.

AND

Clbh. Ada,

33 (19G5)

x08-114

2 (1963)

335.