by flame photometry

7
J. clin. Path. (1961), 14, 463 A new principle applied to the determination of calcium in biological materials by flame photometry J. K. FAWCETT AND V. WYNN From the Surgical Unit, St. Mary's Hospital, London SYNOPSIS The effect of magnesium sulphate in releasing calcium emission from interference by phosphate and sulphate has been investigated. Samples were diluted in 10 mM MgSO4, 2 mM NaCl, giving final calcium concentrations of about 0 05 to 0-10 mM. In this diluent, galvanometer readings were proportional to calcium con- centrations up to 04 mM. The magnesium sulphate released calcium emission from depression by phosphate and sulphate. The excess sodium chloride eliminated enhancement of calcium emission by added sodium and potassium in the sample. Subtraction of background readings excluded direct interference. A 3 % correction was made for the effect of the viscosity of 1: 50 plasma dilutions. Satisfactory recoveries of added calcium were obtained from plasma, urine, and faeces using the diluent described above. Results on urine and faeces correlated closely with those obtained by an EDTA titration method. Results on plasma were consistently 2% higher by flame photometry than by EDTA titration. Other methods of calcium determination, depending on the use of radiation buffers or standard addition, were found to be unsatisfactory because of variable interference by phosphate at different calcium levels. The measurement of calcium in biological materials by flame photometry is more difficult than that of sodium and potassium because of its lower emission energy and more serious interference by contaminat- ing substances. The first difficulty can be overcome by using a high flame temperature, a prism mono- chromator, and a sensitive detector. Instruments incorporating these features permit the use of the weak calcium line at 422-7 m,u where interference is less than using the calcium oxide bands at 554 m,u and 622 m,u. Even at 422-7 m,u, however, there are three important sources of interference as follows: (I) Background (direct) interference: using a prism monochromator, this is due mainly to hetero- chromatic spectral radiation by sodium and potas- sium, only a fraction being due to scattered light. (2) Enhancement or depression of calcium emission (indirect interference): the emission by calcium itself is enhanced by sodium and potassium and depressed by certain anions, notably phosphate and sulphate. Received for publication 21 November 1960. (3) Alteration of the physical properties of the solution because the quantity and dispersal of the spray reaching the flame may be influenced by differences in viscosity and surface tension due to protein and other organic compounds. Many procedures for measuring calcium by flame photometry involve prior separation of calcium from interfering substances by precipitation of calcium oxalate (Powell, 1953; Chen and Toribara, 1953; Butterworth, 1957; Woollen and Walker, 1959) or by means of a cation exchange resin (Brabson and Wilhide, 1954; Denson, 1954; Jackson and Irwin, 1957). These preliminary steps are time-consuming and have errors of their own, thus combining dis- advantages of other methods for calcium deter- mination with those of flame photometry, while losing the simplicity of the latter. Apart from the separation of calcium from inter- fering substances, many methods have been intro- duced to control their effects, but none has been generally accepted. The complexity of interfering effects makes the calculation of correction factors 463

Upload: vanngoc

Post on 30-Jan-2017

296 views

Category:

Documents


1 download

TRANSCRIPT

J. clin. Path. (1961), 14, 463

A new principle applied to the determinationof calcium in biological materials

by flame photometryJ. K. FAWCETT AND V. WYNN

From the Surgical Unit, St. Mary's Hospital, London

SYNOPSIS The effect of magnesium sulphate in releasing calcium emission from interference byphosphate and sulphate has been investigated.

Samples were diluted in 10 mM MgSO4, 2 mM NaCl, giving final calcium concentrations ofabout 0 05 to 0-10 mM. In this diluent, galvanometer readings were proportional to calcium con-centrations up to 04 mM. The magnesium sulphate released calcium emission from depression byphosphate and sulphate. The excess sodium chloride eliminated enhancement of calcium emissionby added sodium and potassium in the sample. Subtraction of background readings excluded directinterference.A 3 % correction was made for the effect of the viscosity of 1: 50 plasma dilutions. Satisfactory

recoveries of added calcium were obtained from plasma, urine, and faeces using the diluent describedabove. Results on urine and faeces correlated closely with those obtained by an EDTA titrationmethod. Results on plasma were consistently 2% higher by flame photometry than by EDTAtitration.

Other methods of calcium determination, depending on the use of radiation buffers or standardaddition, were found to be unsatisfactory because of variable interference by phosphate at differentcalcium levels.

The measurement of calcium in biological materialsby flame photometry is more difficult than that ofsodium and potassium because of its lower emissionenergy and more serious interference by contaminat-ing substances. The first difficulty can be overcomeby using a high flame temperature, a prism mono-chromator, and a sensitive detector. Instrumentsincorporating these features permit the use of theweak calcium line at 422-7 m,u where interference isless than using the calcium oxide bands at 554 m,uand 622 m,u. Even at 422-7 m,u, however, there arethree important sources of interference as follows:

(I) Background (direct) interference: using a prismmonochromator, this is due mainly to hetero-chromatic spectral radiation by sodium and potas-sium, only a fraction being due to scattered light.(2) Enhancement or depression of calcium emission(indirect interference): the emission by calcium itselfis enhanced by sodium and potassium and depressedby certain anions, notably phosphate and sulphate.

Received for publication 21 November 1960.

(3) Alteration of the physical properties of thesolution because the quantity and dispersal of thespray reaching the flame may be influenced bydifferences in viscosity and surface tension due toprotein and other organic compounds.Many procedures for measuring calcium by flame

photometry involve prior separation of calcium frominterfering substances by precipitation of calciumoxalate (Powell, 1953; Chen and Toribara, 1953;Butterworth, 1957; Woollen and Walker, 1959) or bymeans of a cation exchange resin (Brabson andWilhide, 1954; Denson, 1954; Jackson and Irwin,1957). These preliminary steps are time-consumingand have errors of their own, thus combining dis-advantages of other methods for calcium deter-mination with those of flame photometry, whilelosing the simplicity of the latter.Apart from the separation of calcium from inter-

fering substances, many methods have been intro-duced to control their effects, but none has beengenerally accepted. The complexity of interferingeffects makes the calculation of correction factors

463

J. K. Fawcett and V. Wynn

(Severinghaus and Ferrebee, 1950) impracticable.Synthetic standards (Chen and Toribara, 1953;MacIntyre, 1954; Teloh, 1958), with a compositionsimilar to that of the solutions analysed, cannot beused when the composition of the materials is widelyvariable. Internal standards (Baker, 1955) are of onlylimited value because few sources of interferenceaffect calcium emission and that of the internalstandard proportionately.Background correction (Vallee, 1954; Margoshes

and Vallee, 1956) is made by subtracting readingstaken at wavelengths near the analytical wavelengthfrom those taken at the analytical wavelength. Thisaccounts for direct interference, but indirect inter-ference effects present much greater difficulty.

Radiation buffers (Severinghaus and Ferrebee,1950; Maclntyre, 1957), which incorporate an excessof interfering ions, can control sources of indirectinterference which have no increased effect abovecertain concentrations of the interfering substance orwhich have a plateau region over a restricted range.

In the determination of magnesium, anion inter-ference has been controlled (West and Cooke, 1960;Fawcett and Wynn, 1961) by adding a large excess ofethylene diamine tetra-acetic acid (EDTA). West andCooke (1960) found that EDTA could also be appliedin this way to the determination of calcium butthat the excess had to be very much greater, 160-foldinstead of 10-fold. They used the disodium salt ofEDTA, from which the high background emission isa serious disadvantage, particularly with instrumentsnot equipped for automatic spectral scanning.The standard addition method (Rothe and

Sapirstein, 1955) may be used when the relationshipbetween concentration and emission is linear, andsources of indirect interference have the same pro-portional effect upon standard calcium added todiluted samples as upon the calcium already present.The original concentration is thus calculated fromthe equation:calcium in sample emission by samplecalcium added increase in emission due to

calcium addedWe investigated the use of radiation buffers and

the standard addition procedure but, as reportedbelow, neither was found satisfactory for our pur-pose.

Mitchell and Robertson (1936) reported that highconcentrations of strontium released calcium emis-sion from the inhibitory effect of aluminium. Willis(1960), measuring calcium by atomic absorptionspectroscopy, used strontium chloride to controlphosphate interference, and we tested its use for thesame purpose in flame photometry. We also exploredthe potentialities of other compounds and achieved

the best results with magnesium sulphate, which isreadily available in a calcium-free state.

Since we completed our investigation on a methodof releasing calcium emission from anion inter-ference, Dinnin (1960) has reported results of aninvestigation on alkaline earth and rare earthelements, and iron, yttrium, and scandium asreleasing agents, and he has proposed an explanationfor their mode of action. The results of our investi-gation are reported below.

EXPERIMENTAL

INSTRUMENTAL The Unicam SP 900 flame spectro-photometer was used. It has a Meker type burner utilizingacetylene and compressed air. The latter conveys theatomized solution from a separate spray chamber. Theinstrument is equipped with a fused silica prism mono-chromator and a photomultiplier detector. A slit width of0-08 mm., corresponding to a nominal band width of 1 mp,was selected for all experiments. Air pressure was 30lb./in.2 and acetylene pressure was about 12 cm. of water,except when the effect of altering flame conditions wasbeing tested.

PREPARATION OF SAMPLES Samples were diluted to givea final calcium concentration of about 0-05 to 0 10 mM.Plasma samples were obtained from heparinized bloodand diluted 1 : 50. Twenty-four-hour urine samples werepreserved with 10 ml. of hydrochloric acid and diluted1: 50 or 1: 100, according to the expected calciumconcentration. Faeces were prepared by homogenizingwith water and transferring portions of about 2 g. tonickel crucibles of 50 mm. diameter. The contents weredried at 1050 and ashed overnight in the uncoveredcrucibles at 475 to 5250. If magnesium, as well as calcium,was to be determined, ashing was at 420 to 450° becauseof loss of magnesium at higher temperatures (Fawcettand Wynn, 1961). The ash was dissolved by carefulboiling with I ml. of I 0 N. hydrochloric acid, transferredquantitatively and diluted to 100 ml. This solution wasfurther diluted 1: 50 to 1: 10, according to the expectedcalcium concentration.

GALVANOMETER READINGS When special diluting solu-tions were used, standard calcium solutions were dilutedin the same diluent as the samples, and this diluent wasalso used for zero-setting the galvanometer. Blank andstandard readings were checked between every two orthree readings of unknowns. Each dilution was read atleast three times, and more often if any reading differedby more than I % from the mean.

INTERFERENCE BY SODIUM AND POTASSIUM The effects ofsodium were observed by adding various concentrationsof spectroscopically pure sodium chloride to deionizedwater and to solutions containing 0 05 mM and 0-10 mMcalcium chloride.Sodium interference was of two kinds: direct inter-

ference due to spectral radiation by sodium in the regionof the calcium line, and indirect interference due to

464

Determination of calcium by flame photometry

enhancement of calcium emission by sodium. It wasfound that direct interference by sodium was proportionalto its own concentration and represented about 0-04% ofthe emission from equimolecular concentrations ofcalcium, under the conditions of the investigation. TenmM sodium chloride gave identical readings at 422-7 mju,418 mz, and 428 mju, and readings at the latter two wave-lengths were unaffected by calcium emission.

In contrast to the direct interference, the enhancementinterference at any stated sodium level was directly pro-portional to calcium concentration. As the sodium con-centration was raised, the percentage enhancement roseto a maximum of 9% at 1 mM, remaining at this per-centage for sodium concentrations at least as high as10 mM. This is shown in Fig. 1.The interference effects of potassium were similar to

those of sodium, and direct interference was corrected inthe same way. In the presence of 2 mM sodium chloride,as a radiation buffer, there was no further enhancementof calcium emission by added potassium up to 5 mM.

In subsequent investigations, direct interference wasexcluded by subtracting the mean of readings at 418 mj.and 428 mi& from readings at the calcium line, 422-7 m1A.Indirect interference by sodium and potassium was con-trolled by incorporating 2 mM NaCl in all standards andsample dilutions.

INTERFERENCE BY PHOSPHATE AND SULPHATE The effectsof phosphate were observed by adding spectroscopicallypure sodium dihydrogen phosphate to the solutions, andexcluding the effects of sodium.At any stated calcium concentration between 0-02 and

0-2 mM, the depression of calcium emission by phosphatewas constant within the range 0 5 to 10 mM phosphate.The percentage depression of calcium emission byphosphate increased, however, with calcium concentra-tion. Added phosphate thus distorted the linear relation-ship between calcium concentration and emission, asshown in Fig. 2. The degree of distortion varied withflame conditions, such as the pressures of air and acety-

zo

U

w

20JZO

-i'

I U.

{f

w

zw Sw

4-

w

<.10

NoCI (aM)

FIG. 1. Enhancement of calcium emission by sodium.

Z75()2

desoX2I-0

25-

mm Ca

FIG. 2. Depression of calcium emission by phosphate.Curve (a): Calcium emission in absence of phosphate.Curve (b): Calcium emission in presence of 0 5 mMNaH2PO,. Enhancement of calcium emission by Na wascontrolled by including 2 mMNaCo in all solutions.

lene, and it was difficult to reproduce exactly the samecurve on different occasions.

Like phosphate, 10 mM sulphuric acid or 10 mMammonium sulphate caused depression of calcium emis-sion and distortion of the calibration curve. In thepresence of 10 mM sulphate, however, I mM phosphatehad no additional effect on calcium enmission. Similarly,in the presence of 0-5 mM sodium dihydrogen phosphate,as a radiation buffer, there was no further depression dueto added sulphate up to 0 5 m.M.

STANDARD ADDITION METHOD The standard additionmethod depends upon proportionality between calciumconcentration and emission. As phosphate and sulphatecaused distortion of the calcium calibration curve it wassuspected that the method might not be valid whenapplied to biological materials. This was investigated asfollows.

If the standard addition method were valid, the resultcalculated for any sample concentration should havebeen independent of the known quantity of standardcalcium added. Therefore two different quantities wereadded to portions of three plasma samples and four urinesamples diluted in water. The two results calculated foreach sample were compared. These seven experimentswere carried out on five different occasions, with slightvariations in flame conditions.Calcium concentrations were calculated as

2-5 mM x E ' (Formula I)and

5-0 mM x E- (Formula 2)where Eo = emission by sample diluted I1: 50 in water,El = emission by the diluted sample incorporating 0-05

465

J. K. Fawcett and V. Wynn

mM calcium chloride, and E2 = emission by the dilutedsample incorporating 010 mM calcium chloride.

Table I shows the results of these experiments. Thetwo results for each sample concentration usually differedby 2 to 5 %. The non-linear relationship between calciumconcentration and emission, in the presence of plasma orurine, is apparent from the last column of Table I. This

TABLE I

CALCULATION OF CALCIUM CONCENTRATION USINGTHE STANDARD ADDITION METHOD

Sample Ca Concentration of Sample(mM.)

Calculated Calculatedfrom fromFormula I Formula 2

Plasma 1Plasma 2Plasma 3Urine 1Urine 2Urine 3Urine 4

1-742522-720-320552-305-00

1-742572850340-562-385-33

Percentage El -EoDifference E,-E,

0256237

1-001-041-101.111-031-081-13

shows that the increase (E1 - EO) in emission when theconcentration was raised by 0-05 mM was usually sub-stantially greater than the additional increase (E2 - E1)on raising the concentration by a further 0 05 mM. Thesefindings were consistent with the earlier observationsthat, in the presence of phosphate, calcium emission wasno longer proportional to concentration.

CALCIUM RELEASE METHOD Phosphate interference incalcium emission was investigated with solutions dilutedin 10 mM SrCl2, 2 mM NaCl, and in 10 mM MgSO4, 2mM NaCl. Calcium emission was depressed by about 4%by 0-5 mM phosphate in the former diluent but was notdepressed at all in the latter.

Table II shows the release of calcium emission fromthe depressing effect of phosphate when the diluent con-tained 10 mM magnesium sulphate. When the concentra-tion of phosphate was raised above 2 mM, however,calcium emission decreased rapidly, and depression wasvery much greater than that caused by phosphate in theabsence of magnesium sulphate. The concentration ofphosphate which could be controlled varied with theconcentration of magnesium sulphate. Ten mM mag-nesium sulphate was adequate to control phosphateinterference at concentrations up to 100 mM in samplesdiluted 1: 50, or up to 200mM in samples dilutedI :100. Under the conditions of the investigation, themagnesium sulphate appeared to cause a 1% enhance-ment of the calcium emission.With 2 mM sodium chloride in the diluent, sodium

concentrations at least as high as 400 mM had no effecton calcium emission when samples were diluted 1 50.Table II also shows that hydrochloric acid depressedcalcium emission by a maximum of 3 %, but the effectwas negligible at concentrations of 1 mM or less. Thiswas about the concentration of hydrochloric acid in thefinal dilutions, when it was used as a urine preservativeor to dissolve faecal ash.

TABLE II

EFFECT OF VARIATIONS IN COMPOSITION OF DILUENTON EMISSION BY 0-10 mM CaCI2

MgSO,(mM) NaCi (mM)

0010101010101010555

202020101010101010

2222222222222220

23510

1010101010

22222

NaH,PO,(mM) GalvanometerDeflection'

0

05222-32-7350

2024000000

HCI (mM)

52050100

100851011011016845454310098451011019593101101101101101

101100999797

'As percentage of deflection with 0-10 mM CaCi, in 2 mM NaCi.

With 0-05 mM calcium, and the same concentrationsof other solutes as those recorded in Table II, galvano-meter deflections were one-half of those reported for

100'

7 5

u

um0

so-

-I

01 0-2OM Ca

FIG. 3. Calibration curve,NaCI as diluting solution.

03 04

using IOmM MgSO4, 2 mM

1 w

466

Determination of calcium by flame photometry

0-10 mM calcium. The only occasion when calcium con-centration and emission were not proportional was whenphosphate was added in' the absence of magnesium sul-phate. In the presence of magnesium sulphate, however,even when phosphate concentration was so high thatcalcium emission was depressed, the percentage depres-sion was the same at both calcium levels.When a diluting solution containing 10 mM MgSO4,

2 mM NaCI, was used, calcium emission was proportionalto concentrations at least as high as 0-4 mM, the highestconcentration which could be measured at the slit widthof 0-08 mm. The calibration curve is shown in Fig. 3.The background readings at 418 mp and 428 miz were

usually about 2% of the readings at 422-7 mHt for plasmaand urine, but were usually negligible for faeces.

PLASMA VISCOSITY The effect of the viscosity of 1: 50dilutions of plasma on the results obtained was tested bycomparing the rate of uptake of diluted plasma into theatomiser with that of standard solutions, and by com-paring calcium results obtained with and without priorashing of the plasma. The rate at which 1: 50 plasmadilutions were drawn into the atomiser was found to be3 % less than the uptake rate of standard solutions. Whenseven samples of plasma were analysed both directly andafter ashing, calcium emission from the unashed sampleswas an average of 3-6% (S.D. 1-6%) less than theemission after ashing them.The possibility that the lower surface tension of diluted

plasma, compared with standard solutions, might affectthe proportion of the aerosol reaching the flame was alsoinvestigated. A surface-active agent (Pluronic F 68) wasadded to a standard calcium solution to reduce the surfacetension well below that of diluted plasma. Calciumemission in the flame, however, was not affected.

The physical properties of plasma dilutions thusdepressed observed calcium results by 3 to 3-6%, withthe instrument used by us. A 3 % correction factor wasselected and the validity of this is supported by the resultsof the recovery experiments reported below, in which thiscorrection factor was used.

RESULTS

RECOVERY EXPERIMENTS USING CALCIUM RELEASEMETHOD Recovery experiments were carried out on16 samples of plasma, 15 of urine, and 14 of faeces.Samples were prepared as described above, incor-porating 1 ml. of a stock solution of 0-5 M MgSO4,0-1 M NaCl, in each 50 ml. dilution. Dilutions wereprepared in parallel, with the inclusion of knownquantities of calcium. In the recovery experimentson faeces, calcium was added sometimes beforeashing and sometimes afterwards, the results beingreported separately. The experimentally determinedviscosity factor was used in calculations of recoveriesfrom plasma.

Results of the recovery experiments are reportedin Tables III, IV, and V. Mean recoveries were

TABLE III

RECOVERIES OF CALCIUM ADDED TO16 DILUTED PLASMA SAMPLES

Ca Added' (mM) Ca Recovered' (mM) Percentage Recovery

1-25 1-27 101-61-25 1-27 101-61-25 1-28 102-42-50 2-43 97-22-50 2-44 97-62-50 2-45 98-02-50 2-46 98-42-50 2-46 98-42-50 2-49 99-62-50 2-50 100-02-50 2-50 100-02-50 2-50 100-02-50 2-52 100-82-50 2-53 101-22-50 2-55 102-02-50 2-56 102-4

Mean 100-1S.D. 1-9

'In terms of apparent increase in plasma Ca concentration.

TABLE IV

RECOVERIES OF CALCIUM ADDED TO15 DILUTED URINE SAMPLES

Ca Added' (mM) Ca Recovered' (mM) Percentage Recovery

2-50 2-47 98-82-50 2-51 100-42-50 2-52 100-82-50 2-54 101-65-00 4-85 97-05-00 5-00 100-05-00 5-00 100-05-00 5-05 101-05-00 5-05 101-05-00 5-05 101-0

10-00 10-05 100-510-00 10-05 100-510-00 10-10 101-010-00 10-10 101-010-00 10-25 102-5

Mean 100-5S.D. 1-3

'In terms of apparent increase in urine Ca concentration.

TABLE V

RECOVERIES OF CALCIUM ADDED TO (a) SEVEN SAMPLES OFFAECES BEFORE AND (b) SEVEN SAMPLES OF FAECES AFTER

ASHING

Ca Added' (m-mole)

10-010-010-010-010-020-020-075-0

125-0

Ca Recovered'(m-mole)

(a)9-9

9-910-010-019-4

75-0123-0

(b)

9-910-110-29-919-520-376-0

Percentage Recovery

(a)99-0

99-0100-0100-097-0

100-098-4

Mean 99-1S.D. 1-1

(b)99-0101-0102-099-097-5101-5101-3

100-21-6

'In terms of apparent increase in daily output.

467

J. K. Fawcett and V. Wynn

100-1% for plasma, 100-5% for urine, 99-1% forfaeces to which calcium was added before ashing,and 100-2% for faeces to which calcium was addedafter ashing. Standard deviations were less than 2%.

COMPARISON BETWEEN RESULTS OBTAINED BY CALCIUM

RELEASE METHOD AND BY EDTA TITRATION Theresults of 26 calcium determinations on plasma, 10on urine, and 11 on faeces obtained by flame photo-metry using the 10 mM MgSO4, 2mM NaCl, diluentwere compared with those obtained by titration withEDTA. The method for plasma was direct titrationat pH 12 with automatic detection of the end-point(Wootton and Haslam, to be published).The methodfor urine and faeces was the back-titration of a

measured excess of EDTA with a standard calciumsolution at pH 12 using a calcein-thymolphthaleinindicator.

Results of this comparison are recorded in TablesVI, VII, and VIII. Values for plasma were consis-tently 2% higher by flame photometry than bytitration, but the differences with urine and faeceswere small and unsystematic.

TABLE VI

COMPARISON OF RESULTS ON 26 PLASMA SAMPLES OF CAL-CIUM DETERMINATIONS BY FLAME PHOTOMETRY, USING THECALCIUM RELEASE METHOD, AND BY EDTA TITRATION

Calcium Concentration (mM)

By Flame Photometry"(a)

1-221-372-182-322-352-352-372-382-422-432-432-452-462-492-522-552-552-622-682-692-702-762-772-822-953-46

Difference (mM)

By EDTA(b)

1-171-322-152-272-302-302-322-332-352-322-422-392-432-442-522-522-552-602-602-622-622-722-702-752-883-42

(a-b)

+0-05+0-05+0-03+0-05+0-05+0-05+0-05+0-05+0-07+0-11+0-01+0-06+0-03+0-050

+0-030

+0-02+0-08+0-07+0-12+0-04+0-07+0-07+0-07+0-04

Mean +0-05S.D. 0-03

'3 %7 was added to each observed result by flame photometry to correctfor the effect of plasma viscosity.

TABLE VII

COMPARISON OF RESULTS ON 10 URINE SAMPLES OF CAL-CIUM DETERMINATIONS BY FLAME PHOTOMETRY, USINGCALCIUM RELEASE METHOD, AND BY EDTA TITRATION

Calcium Concentration (mM)

By Flame Photometry(a)

0-320-400-573-863.934-154-984.945-017-61

Difference (mM)

By EDTA(b)

0-260-440-573-814-054-154-895-005-207-81

(a-b)

+0-06-0-040

+0-05-0-120

+0-09-0-06-0-19-0-20

Mean -0-04S.D. 0-10

TABLE VIIICOMPARISON OF RESULTS ON 1I SAMPLES OF FAECES OFCALCIUM DETERMINATIONS BY FLAME PHOTOMETRY, USINGCALCIUM RELEASE METHOD, AND BY EDTA TITRATION

Calcium Output (m-mole/day) Difference(m-mole/day)

By Flame Photometry By EDTA(a) (b) (a-b)

1-53-38-29-011-414-720-120-221-963-076-0

1-63-38-38-7

10-714-019-920-321-561-576-0

-0-10

-0-1+0-3+0-7+0-7+0-2-0-1+0.4+1-50

Mean +0-3S.D. 0-4

DISCUSSION

The proper control of interference effects in thedetermination of calcium in biological materials byflame photometry requires that each source of inter-ference should be considered individually. Allowancecan be made for differences in the physical propertiesof solutions and, when a monochromator is used, itis simple to correct for direct interference by sub-tracting background readings. Enhancement ofcalcium emission by cations can be kept constant bytaking advantage of the plateau effect. Anions, suchas phosphate and sulphate, which depress calciumemission, present a more difficult problem becauseof their variable influence at different calcium levels.The effect of phosphate in distorting the linear

relationship between calcium emission and concen-tration was noted by Jackson and Irwin (1957). Thedata of Baker and Johnson (1954) are difficult tointerpret in this respect because, while their Fig. 2

468

Determination of calcium by flame photometry

shows that percentage depression by phosphateincreased with increasing calcium concentration, atsimilar phosphate levels their Fig. 7 shows a linearrelationship between calcium emission and con-centration. Other workers (Chen and Toribara, 1953;Denson, 1954; Margoshes and Vallee, 1956;Maclntyre, 1957; Teloh, 1958) reported the effectof phosphate at a single calcium level and did notrefer to distortion of the calibration curve.

Reports on the quantitative aspects of phosphateinterference are generally conflicting. Although thismight be attributed partly to the diversity of instru-ments and conditions used, it is noteworthy thatJackson and Irwin (1957), who observed non-proportionality of calcium concentration and emis-sion in the presence of phosphate, employed asystem different from our own, namely, a BeckmannDU spectrophotometer, a hydrogen-oxygen flame,and a wavelength of 554 m,u.

If the distorting effect of phosphate has sometimesbeen overlooked, this might help to account for thelarge discrepancies which have been found betweenresults of flame photometry and of other methods,to which Hunter (1958) and Willis (1960) have drawnattention. Our findings agree well with those obtainedby EDTA titration, even the 2% discrepancy withplasma being much less than that reported by others.Of the known methods of controlling indirect

interference effects in flame photometry, the use ofradiation buffers or the addition of standard calciumto the samples, seemed to us the most hopeful. Wefound, however, that calibration curves usingradiation buffers in the presence of phosphate weredifficult to reproduce on different occasions, an effectwhich we attributed to variations in flame conditions.The standard addition method also proved unsoundbecause of the non-proportionality of calcium con-centration and emission in the presence of phosphate.Further disadvantages of this method were the super-imposition of errors because of its dependence upondifferences between large values, and the necessity toprepare a standard for each sample, even in a largeseries.

In the method presented here, calcium emissionwas released from depression by phosphate by theaddition of a large excess of magnesium sulphate.This also eliminated any effect on calcium emissionby the much smaller quantities of sulphate andmagnesium ions already in the samples. When 10mM MgSO4, 2 mM NaCl was used for zero-settingand dilution of standards and samples, and whensimple corrections were made for direct backgroundinterference and viscosity effects, the results obtainedon biological materials were independent of varia-tions in the concentration of interfering substanceslikely to be encountered.3

Dinnin (1960) has published a comprehensivereport on releasing effects in flame photometry.Although he used strontium and praseodymium andnot magnesium salts in quantitative analysis forcalcium, he drew attention to the potentialities ofmagnesium in this connexion, particularly because ofits low background emission in the region of thecalcium line.

Dinnin's explanation of the mechanism of releas-ing effects is based upon chemical equilibria in theevaporating droplets of solution, involving anexchange ofanions between calcium and the releasingcation. Two of our observations seem inconsistentwith this explanation. Although magnesium wasadded in the form of the sulphate, there was nodepression of calcium emission. In the presence of10 mM magnesium sulphate, there was a sudden andlarge depression of calcium emission when the con-centration of phosphate was raised above 2mM, andthis occurred at calcium concentrations of both0-05 mM and 0- 10 mM. This behaviour seemed moreconsistent with stoichiometric reactions involvingmagnesium and phosphate ions, in which calciumplayed no direct part, than with the displacement ofequilibria. These theoretical aspects of calciumrelease are being investigated further.

The flame spectrophotometer used in this study waskindly lent to us by Unicam Instruments Ltd.We thank Dr. P. H. Sanderson who carried out the

EDTA titrations on plasma, Mr. P. B. Wells and Mr. H.Thomas who carried out the EDTA titrations on thesamples of urine and faeces, and Mr. J. A. Corbett whodrew our attention to the work of Dr. J. B. Willis onflame absorption spectroscopy.

REFERENCES

Baker, G. L., and Johnson, L. H. (1954). Analyt. Chem., 26, 465.Baker, R. W. R. (1955). Biochem. J., 59, 566.Brabson, J. A., and Wilhide, W. D. (1954). Analyt. Chem., 26, 1060.Butterworth, E. C. (1957). J. clin. Path., 10, 379.Chen, P. S., and Toribara, T. Y. (1953). Analyt. Chem., 25, 1642.Denson, J. R. (1954). J. biol. Chem., 209, 233.Dinnin, J. I. (1960). Analyt. Chem., 32, 1475.Fawcett, J. K., and Wynn, V. (1961). J. clin. Path. 14, 403.Hunter, G. (1958). Nature (Lond.), 182, 263.Jackson, W. P. U., and Irwin, L. (1957). J. clin. Path., 10, 383.Maclntyre, I. (1954). Biochem. J., 56, xliii.

(1957). Ibid, 67, 164.Margoshes, M., and Vallee, B. L. (1956). Meth. biochem. Anal., 3, 353.Mitchell, R. L., and Robertson, I. M. (1936). J. Soc. chem. Ind.

(Lond.), 55, 269 T.Powell, F. J. N. (1953). J. clin. Path., 6, 286.Rothe, C. F., and Sapirstein, L. A. (1955). Amer. J. clin. Path., 25,

1076.Severinghaus, J. W., and Ferrebee, J. W. (1950). J. biol. Chem., 187,

621.Teloh, H. A. (1958). A.M.A. Arch. Path., 66, 474.Vallee, B. L. (1954). Nature (Lond.), 174, 1050.West, A. C. and Cooke, W. D. (1960). Analyt. Chem., 32, 1471.Willis, J. B. (1960). Nature (Lond.), 186, 249.Woollen, J. W., and Walker, P. G. (1959). J. clin. Path., 12, 149.Wootton, I. D. P., and Haslam, R. M. To be published.

469