Design Considerations for High Performance Background ...

Product Specifications

Part of Thermo Fisher Scientific

e l e m e n t a l a n a l y s i s

What is Background Absorption? The technique of atomic absorptionspectrometry depends on the formation ofan atomic vapor from a sample solution andon the measurement of element specificradiation absorbed by the vapor.

Any metallic element in atomic formwill absorb light over a narrow range ofwavelengths centred on the resonancewavelength whose value is specific to theelement. The absorption line width is about0.002 nm at normal atomizationtemperatures. In an atomic absorptionspectrometer, the light source is usually ahollow cathode lamp, which producesresonance radiation of about 0.001 nm linewidth. The lamp also generates light atwavelengths other than the resonancewavelength, and this makes it necessary topass the light through a monochromator.

The monochromator is tuned to theresonance wavelength, but will pass arange of wavelengths (typically 0.1 - 1.0 nm) -much greater than the atomic absorptionline width.

When only the analyte atomic speciesare present, the amount of light absorbedwill depend only on the concentration of theanalyte element in the light path.

Materials other than the analyte,derived from the sample matrix and theatomization system itself, may also bepresent in the light path. In somecircumstances, these can absorb or scatterthe radiation, thus increasing the totalabsorbance and hence the apparentconcentration of the analyte. This effect isknown as non-specific, or background,absorption and, if present, can reduce theaccuracy of the analysis unless steps aretaken to correct for it.

Design Considerations forHigh Performance BackgroundCorrection Systems inAtomic Absorption Spectrometry

The Thermo Scientific iCE 3000Series AA Spectrometers arefitted with highly efficientbackground correction systemsfor flame and furnace use. Theirimportant design features aredescribed in this article alongwith discussions on theadvantages and disadvantages ofother techniques.

Product Specifications

Molecular AbsorptionMany sample matrices contain halide saltsat relatively high concentrations. In theatomizer, these will form molecular vapors,resulting in molecular absorption of theradiation. Molecular absorption spectra arebroad band, with band widths of the orderof tens of nanometres, and for alkali metalhalide salts, absorption increases sharply inthe deep ultra-violet region (below 220 nm)of the spectrum. Some typical vapor phasespectra of alkali metal halide salts areshown in Figure 1 below.

Figure 1: Molecular spectra of simple halide salts

Background absorption caused bymolecules is thus particularly troublesomewhen measuring elements with their mainresonance line below 220 nm and in samplematrices containing alkali metal halides. Thisincludes many of the toxic heavy metals,such as lead, cadmium and arsenic inbiological and environmental samples, suchas natural waters, effluents and body fluids.

ScatterSome types of sample matrix are notcompletely vaporized in the atomizer, butform small particles or clusters. Thesescatter the light from the hollow cathodelamp, preventing it from reaching thedetector, and so cause an apparentabsorption signal. The degree of scatterdepends strongly on both the wavelength ofthe light and the size of the particles, andincreases markedly at low wavelengths.

Background absorption caused byscatter is, therefore, also most significantfor elements with their resonance linesbelow 220 nm, and depends both on thetotal dissolved solids content of thesamples, and the nature of the dissolvedsolids - refractory metals and those formingvery stable oxides, such as the alkalineearths, cause the greatest problems.

Structured AbsorptionOther components of the sample matrix willbe atomized, and their atomic absorptionspectra may contain lines close to the analyteresonance line. Cases of direct line overlapare very rare, but it is possible that linesderived from the matrix components will fallwithin the monochromator band pass.

In most cases, the broadbandbackground absorption is a relatively smoothfunction of wavelength, and may be regardedas virtually constant over the bandpass of thespectrometer monochromator.

Occasionally, examples are encounteredof more complex background absorptionspectra with line structure superimposed onthis relatively constant background. Anexample of these so-called structuredbackgrounds is the determination ofselenium in whole blood, where the ironcontent of the matrix exhibits such structurewithin the monochromator bandpass. Shouldsuch structure occur within themonochromator bandpass of a deuteriumbackground correction system, systematicerrors will be present in the correctedatomic absorption signal, and hence theanalytical result. Such examples arerelatively uncommon in AA analysis but,when present, may advantageously be dealtwith using the Zeeman correction technique.

Methods of Compensatingfor Background Absorption

IntroductionInstrumental methods of correcting forbackground absorption depend upon makingat least two separate measurements of thesample absorption. The first is the normalatomic absorption measurement made atthe analyte resonance line, and gives thetotal (atomic plus background) absorption. Asecond measurement is then made, in sucha way as to reduce the effect of theabsorption from the analyte atoms to anegligible amount, while not affecting thebackground absorption. This secondmeasurement gives only the backgroundabsorbance, so that subtraction of the twomeasurements will give the true atomicabsorbance.

All instrumental background correctionsystems are based on this principle. Theydiffer in the means used to reduce the atomicsignal to negligible levels for the backgroundmeasurement, and in the frequency withwhich the two measurements are made.

Continuum MethodThe second (background) measurement in thistechnique is made using radiation from asource with a broad spectral output, such as adeuterium arc lamp. The radiation intensityfrom such lamps is almost independent ofwavelength over the monochromator bandpass (i.e. over a wavelength range of about 1nm), and so ‘fills’ the monochromator bandpass. The analyte atomic absorption takesplace over a wavelength range of about 0.002nm; the amount of continuum source radiationabsorbed by the analyte atoms is, therefore,negligible. Background absorption is largely abroad band phenomenon, and is alsoreasonably independent of wavelength overthe monochromator band pass. Only thebackground absorption will, therefore, bemeasured by the continuum source. Thissituation is shown diagrammatically in figure 2.

Figure 2: Atomic absorption from a hollowcathode lamp and a continuum source.

It is possible to use continuum sourcebackground correction with sequentialmeasurements of total (line source) andbackground only (continuum source)absorbance, followed by manual subtractionto give the corrected atomic signal. It is alsopossible to design the spectrometer in such away as to automatically make bothmeasurements and subtract the results, andto do this sufficiently rapidly that many pairsof measurements can be made andsubtracted over the normal sampling period,providing fully automatic backgroundcorrection. This is the basis of almost allcommonly used background correctionsystems, and most modern atomic absorptionspectrometers are capable of makingmeasurements that are automaticallycorrected for background absorption.

Well designed background correctionsystems based upon the continuum sourcetechnique are capable of correcting themajority of non-specific absorption problemsencountered during both flame and graphitefurnace atomic absorption measurements.Only in situations in which the backgroundabsorption is NOT independent ofwavelength over the monochromator bandpass, or where direct spectral overlap of theanalyte atomic line and a narrow bandabsorption due to a matrix componentoccurs, will it give incorrect results.

Commonly used continuum sources,such as deuterium arc lamps, emit most oftheir radiation in the ultra-violet region ofthe spectrum, below 350 nm. Backgroundcorrection in the visible region is, therefore,not possible, unless an additional source isused. However, almost all backgroundabsorption phenomena occur in the ultra-violet region - scatter and molecularabsorption are usually negligible atwavelengths above 350 nm - so this is not aserious practical limitation.

Product Specifications

Self Reversal MethodThis technique (also known as the Smith-Hieftje technique) relies upon a broad bandsource to measure the background absorption.Rather than employ a separate source, thehollow cathode lamp itself is used.

Hollow cathode lamps are normallyoperated at lamp currents up to 25 mA,when they produce the narrow line emissionspectra required for atomic absorptionmeasurements. If the peak current isincreased to several hundreds of milliamps,the emission lines broaden and self reverse(Figure 3). Under these conditions, thenarrow atomic absorption line can onlyabsorb a small part of the broadenedemission line, while the broad band orcontinuous background absorption will occuras usual. Absorbance measurements madeat very high lamp currents, will be mainlydue to the non-specific absorption.

Figure 3: Hollow cathode lamp emission linebroadening at high lamp currents.

The hollow cathode lamp is, run withalternating low and high current pulses.Total absorbance is measured during thelow current pulse and the backgroundabsorbance is measured during the highcurrent pulse. The two signals areautomatically subtracted to give thecorrected atomic signal.

The self-reversal technique is capableof correcting for most backgroundabsorption phenomena over the fullwavelength range of the spectrometer.Although the emission line is broadenedduring the high current pulse, it does not'fill' the monochromator band pass in thesame way as a true continuum source. Thesystem can, therefore, correct for somecases of structured background. It will not,however, correct for actual line overlap.

The disadvantages of the technique arethe effect of the high current pulses on thehollow cathode lamp - unless speciallydesigned lamps are used, the lamp lifetimewill be significantly reduced - and thebreakdown of the assumption that theatomic absorption is negligible during thehigh current measurement.

The degree to which the lines for thedifferent elements broaden and self reverse,and hence the validity of this assumption,will depend on the element and on thedetailed construction of the lamp,particularly the cathode material and the fillgas pressure. When the assumption breaksdown, the background-only signal measuredwill actually contain a significantcontribution from the atomic signal. Whenthis is subtracted from the total signal, theapparent atomic signal that results will beover-corrected. The effect is to reduceanalytical sensitivity (by as much as 70% insome cases) and increase the curvature ofthe calibration graph.

Zeeman MethodThis technique does not rely on a continuumor pseudo-continuum source for themeasurement of the background absorption.Instead, the absorption of the analyteatomic absorption line is cancelled, byapplying an intense magnetic field to theatomizer. The effect of the magnetic field onthe absorption spectra of free atoms iscomplex, but essentially causes the singlesharp line to split into a family of linesspread around the wavelength of the originalline (some examples are shown in Figure 4).

Furthermore, the split absorption linesare polarised in directions either parallel orperpendicular to the magnetic field. If thehollow cathode radiation is also polarized,by passing it through a suitable polarizer,measurements made with the magneticfield on will not contain any analyte atomicabsorption, and so will be due solely tobackground absorption taking place at theanalyte resonance wavelength.

Figure 4: Zeeman splitting of atomic lines

The Zeeman technique is the only onein which measurement of the backgroundabsorption takes place at exactly the samewavelength, and over exactly the sameband pass, as measurement of the totalsignal. It is, therefore capable of correctingfor highly structured background spectraand even cases of spectral overlap. Adisadvantage is the relatively complexmagnet and its control electronics; magneticfield strengths up to almost 1 tesla (10kilogauss) are required to obtain optimumline splitting for some elements.

Some elements, including many ofthose most commonly determined by atomicabsorption spectrometry, exhibit aphenomenon known as ‘anomalous’ Zeemansplitting. Here a component of the split lineremains at exactly the same wavelength asthe unsplit line. The atomic absorption isnot, therefore, reduced completely to zerowhen the magnetic field is on and thebackground-only measurement contains acontribution from the analyte atomicabsorption. This, when subtracted from thetotal, magnet-off, measurement, results inover-correction of the atomic signal,reducing the apparent chemical sensitivityand increasing the calibration curvature, ina similar way to the self-reversal technique.

With Zeeman background correction,the magnet must be placed around theatomizer. It is impractical, therefore, toapply this technique to flamemeasurements, and commercially availablesystems are usually limited to graphitefurnace analyses. It is also possible to applythe Zeeman effect to the radiation source,causing the emission lines, rather than theabsorption lines, to split. Measurementsmade with the split lines (for elementsexhibiting the ‘normal’ Zeeman effect, atleast), will not include a contribution fromthe analyte atomic absorption. However, thebackground absorption is not then measuredat exactly the same wavelength as theatomic absorption and, furthermore, ismeasured with a highly structured source. Ifthe background absorption is alsostructured, large errors can be introduced.Normal hollow cathode lamps will notoperate in intense magnetic fields andspecially designed sources are required.Consequently, this arrangement has onlybeen used in a limited way, and in specialpurpose instrumentation.

The magnetic field may be applied aseither a DC (permanent) magnet system oras an AC (electromagnet) system. In thecase of a DC system, the magnetic field isalways present. Differentiation of total andbackground signals is obtained by switchingthe plane of polarization of the hollowcathode radiation. This is usually achievedby physically rotating the polarizer itselfwhich requires a complex, high precisionmotor driver mechanism. The AC Zeemanmethod where differentiation is obtained bymodulating the magnetic field, is preferable,as the field can be modulated by driving themagnet with a simple AC waveform.Furthermore, the higher the frequency ofmagnet modulation, the closer the twosignals of total and background absorptionare measured in time. With fast transientfurnace absorption peaks, this reducesanalytical errors.

Instrumental Design

RangeIrrespective of the physical techniqueemployed to derive the backgroundabsorbance signal, all systems have tosubtract this signal from the total signalto obtain the corrected atomic absorption.To obtain an accurate result, both the totaland the background absorbance signalmust be within the measuring range of the spectrometer.

AccuracyAs the analytical signal (from which thefinal result is derived) is the differencebetween the total and the background-only signal, it will contain errorsoriginating in both of these signals, inaddition to any errors introduced by thesignal subtraction itself. The accuracywith which the primary signals aremeasured, therefore, needs to besignificantly greater than the accuracyrequired for the analytical signal.

It can be shown from statisticalconsiderations that the error in thecorrected absorbance signal will beminimized when the errors in the totalabsorbance and background-onlyabsorbance signals are of similarmagnitude. This implies that theintensities of the radiation sources used tomeasure the absorbance signals shouldalso be similar.

Product Specifications

EmssionFlame, and particularly graphite furnaceatomizers, can produce intense emissionsof radiation as a result of the hightemperatures used.The spectrometer must be capable ofaccurately measuring the source lampintensities in the presence of the atomizeremission, in order to derive the twoabsorbance signals. Bright source lamps,and an optimized optical system are,therefore, essential.

The internal dimensions of the ThermoScientific cuvette have been considered inthe design of the spectrometer opticalsystems, to ensure that thermal emissionfrom the cuvette walls is, as far as possible,screened from the detector withoutcompromising the high optical transmissionefficiency and low noise characteristics ofthe spectrometer.

This is particularly useful whendetermining refractory elements requiringhigh atomization temperatures and withmain resonance lines above 320 nm.

Transient ResponseAtomic and background signals generated byflame atomization systems are normallystable, steady state measurements. Graphitefurnace atomization, however, can generaterapidly changing signals. A modern, fastheating design, such as the Thermo ScientificGFS35, can generate total and backgroundsignals that are varying independently atrates of up to 50 absorbance units persecond. Pairs of measurements must,therefore, be taken at a sufficiently fast rateto enable the background correction system toaccurately track these two signals andsubtract them to give an accurate, undistortedcorrected signal.

Background Correction Systems

QuadlineThe continuum source technique used in theThermo Scientific AA Series ofspectrometers offers the performancerequired by the vast majority of analysts ina cost effective, and theoretically sound,package. Detailed improvements in thesource design, the optical configuration ofthe spectrometer and the signal processingfunctions overcome many of the limitationstraditionally associated with the technique.This results in a high performance systemcapable of handling the most demandingapplications.

Source LampA high intensity deuterium arc lamp is usedas the continuum source, with a novel 4electrode design that allows the lamp to beelectronically modulated at four times themains frequency, without the use ofmechanical choppers. This modulationfrequency, together with the digital signalprocessing described below, permits accuratecorrection of signals changing at rates of upto about 70 absorbance units per second.

The novel lamp design, coupled with aninnovative power supply, allow the arccurrent, and hence theintensity of theemitted radiation, tobe varied over a widerange whilemaintaining excellentstability at anyparticular setting. Thecontinuum sourceintensity can,therefore, beaccurately matchedwith the hollowcathode lampintensity over a wide range of elements andwavelengths, ensuring that the correctedsignal is measured as accurately as possible.

Optical System

Source Location and AlignmentAlthough the deuterium arc lamp has a longuseful lifetime, it will eventually requirereplacement.

For accurate operation a continuumsource background correction system, it isessential that the two sources are preciselyaligned relative to each other and to theatomizer. In the Thermo Scientific iCE 3000Series AA Spectrometers, the deuterium arclamp, once initially aligned, does not requirefurther adjustment. The hollow cathodelamps are automatically aligned to thedeuterium beam using the spectrometer'sautomatic lamp alignment facilities.

iCE 3500 Optical System LayoutThe novel echelle optical system layout ofthe Thermo Scientific iCE 3500 AA is shownin Figure 5. Front and rear beam selectormirrors act together to direct the light beamthrough either the flame or the furnaceatomizer. The optics have been optimizedfor furnace use so that the “pencil-beam”optics eliminate potential emissionproblems and half height slits (traditionallyused but with an attendant loss of energy)are not required.

Figure 5: iCE 3500 optical system showing theechelle design.

Echelle technology is the only solution toproviding high energy throughput and pencilbeam optics, without going to a very largetraditional Czerny-Turner design (which wouldneed a 500 mm focal length and an 1800lines/mm grating). The echelle reciprocallinear dispersion of 0.5 nm/mm is the best inthis class of instrument.

Signal Processing

Analogue versus Digital ProcessingAutomatic background correction systemsrequire significant amounts of signalprocessing to derive the analytical signal.The basic measurements taken at thedetector give the intensities of radiationfrom the hollow cathode lamp and thecontinuum source lamp, each with, andwithout, the presence of sample in theatomizer. From these signals, the total andbackground absorbance values are derivedby taking the logarithm of the measuredradiation intensities from each source with,and without, sample. Finally, the twoabsorbance values are subtracted to givethe corrected atomic absorbance.

Traditional instruments performed thisprocessing using analogue electronics.They inevitably have a number of variableelements which have to be set upindividually to compensate for smalldifferences in the electronic components.This adds to the initial cost of theinstrument, and means that regularmaintenance and adjustment is needed tomaintain the original performance.

With the development of digitalmicroprocessor based electronics, analternative has become available. In a digitalsystem, the values of the different quantitiesare represented by numbers, not voltages,and are manipulated by a microprocessor.The major advantage of this approach is thatthere are no variable components in thesystem that need to be set up, and that canlater go out of adjustment. The long termperformance is more stable and little, or no,maintenance is required.

Digital systems are also more versatile.Operations such as averaging successivesignal values, which are difficult to performwith analogue circuits, can be simplyachieved by programming themicroprocessor. For these reasons, the signalprocessing functions in the iCE 3000 Seriesare carried out by a digital system. Radiationintensities are measured by a photomultiplier,and converted immediately to digital formusing a device known as an Analogue-to-Digital convertor. All subsequent processing,including logarithmic conversion to give theabsorbance values and subsequentsubtraction of the background-only signal,are performed by the microprocessor.

The microprocessor also optimizes boththe gain of the photomultiplier and theintensity of the deuterium arc lamp, ensuringthat the raw intensity measurements aremade under optimum conditions.

Signal BracketingIt is not possible in a practical backgroundcorrection system to measure the total andbackground absorbance signals at exactly thesame time. If the signals are changingrapidly, the actual value of the total signalwhen the background signal is measuredcould be slightly different, resulting in anerror in the corrected signal. Consider thesequence of measurements shown in Figure6a for simple correction, where T1, B1 arethe first pair of total and backgroundabsorbance measurements, taken at times tT1

and tB1, and T2, B2 are the second pair, takenat times tT2 and tB2. Simple subtraction of B1from T1 gives the result C for the correctedatomic absorbance, underestimating the truevalue C1 by the amount that the total signalhas changed in the time from tT1 to tB1.

Figure 6a: Measurement sequence for simplecorrection.

Errors from this source can be reducedby using fast modulation of the sources anddetector system, so that many pairs ofmeasurements are made rapidly and theamount that the signals change betweenindividual measurements is reduced. Theversatility offered by microprocessor baseddigital signal processing allows a furtherrefinement to be made, further improving theaccuracy of the background correction (asshown in Figure 6b for bracketing correction).

Figure 6b: Measurement sequence for bracketingcorrection.

If we wait to perform the calculationuntil tT2, when the next measurement ofthe total signal T2 is taken, we can use thevalues of T1 and T2, together with theknown times tT2, tB1 and tT2, to estimatethe actual value, TE, of the totalabsorbance at time tB1.

TE = T1 + ( T2 -T1 )x(tB1 - tT1)/(tT2 - tT1)Subtracting B1 from TE will then give

a very much better estimate of the trueatomic absorbance C1.

This technique is known as signalbracketing, and only becomes practicablewhen digital signal processing is used. Incombination with the fast modulationsystem used in the iCE 3000 Series, it willreduce errors in the corrected signalscaused by timing effects, to less than 1%for signals changing at up to 70 absorbanceunits/second.

Product Specifications

Quadline Conclusions The key aspects of the QuadLine method ofbackground correction may be summarizedas follows:• highest possible measurement sensitivity

compared with alternative techniques• novel wideband source allowing

continuous intensity variation and fastmodulation frequency

• optical matching of the beam profilesfrom both the background correctionsource and HCLs

• fast modulation frequencies of 200/240 Hz• beam combining techniques ensuring

homogeneity within the beam profile• bracketing correction of both background

and dark level signals to overcome timedependence of measurement periods

• dark level correction to zero every 10 mSto improve dynamic range of ADCconversion

• intensity matching of background sourceand HCL to improve dynamicmeasurement range

Zeeman CorrectionEmploying the alternative major backgroundcompensation technique of Zeemancorrection can be shown to be of analyticalbenefit in a number of specific situations asoutlined earlier, in particular those ofstructured backgrounds and spectraloverlaps. Furthermore, some analysts mayfind the technique fundamentally less proneto operator variability, since no sourcealignment is required, beyond maximisingthe HCL energy throughput.

For these reasons, Thermo FisherScientific provide the iCE 3500 (dualatomiser), and iCE 3400 Zeeman-onlyfurnace systems. The system employ theelectromagnetic, or AC, Zeeman techniqueapplied at the atomizer. Uniquely, all thesesystems also provide the QuadLinebackground correction technique asstandard within the same spectrometer.Thus, the technique of preference may bechosen for a particular analysis by a simpleoption within the operating software.

QuadLine correction may be employedfor the highest measurement sensitivity,whereas Zeeman correction may be selectedwhen encountering a structured background.An additional benefit is the ability to employboth techniques in combination within afumace analysis. Since any modulatedZeeman system is limited to correction duringthe atomize phase only, due to potentialoverheating of the magnet system, vitallyimportant method development informationmay be lost in the earlier pre-treatmentfurnace phases due to the inability to monitorthe corrected absorption signal. By selectingthe combined mode, QuadLine correction maybe employed in all but the atomize phase,during which the Zeeman technique is used.By this unique means, much valuable methoddevelopment information may be obtained.

Zeeman TechniqueThe magnet, placed around the furnace head,is modulated at a high 100/120Hz frequencyin order to achieve minimal time separationof total and corrected signals. The magnetdrive circuit achieves a maximum fieldstrength of almost 0.9 Tesla (9 kilogauss) -optimum for all AA analytes. A trapezoidalwavefom is used for the drive current,ensuring that the measurement period isperformed totally during maximum andconstant magnetic field conditions, unlikesinusoidal field modulation techniques.

The magnet pole pieces are so designedto provide a magnetic field constant towithin 5% in the key central working volumeof the graphite cuvette in the furnace head.Due to the inhomogeneity of atomic andmolecular species within the graphitecuvette volume, such a constant magneticfield is vital in providing reliable data.

The design of the Zeeman magnetsystem must also achieve other vitallyimportant objectives. The system must bereadily demountable for routine service anduser maintenance operations. It must beextremely durable in the presence of thepotentially aggressive chemicals employedin the AA technique and, of course, mustminimize extraneous stray fields due to theirdetrimental effects on nearby, sensitiveelectronics and displays. Finally, the entiresystem must be comprehensively equippedwith safety interlocks which, together withsimilar furnace devices, ensure operator andequipment safety.

OperationOn selection of the Zeeman-basedcorrection system, measurements arealternately taken in both the magnet on andmagnet off conditions. With the magnetoff, the instrument provides a measure ofthe total (atomic + background signal). Withthe magnet on, the instrument provides ameasure of the background-only signal atthe analyte wavelength. This is achieved byeffectively “removing” the analyteabsorption profile from the measurementwavelength, as shown in Figure 7. Subtraction of these signals then providesthe displayed, corrected absorption profile.

A further benefit of the Zeemancorrection system is that effectively the twomeasurements (magnet on and magnet off)provide a referenced measurement andsource drift correct ion without thenecessity of conventional double beamoptics. Since only a single optical beam isemployed, alignment of the measurementsource and atomization unit is alsorelatively straightforward.

Figure 7: Inverse AC Zeeman method of correction.

Zeeman Conclusions The key aspects of this Zeeman method ofbackground correction may be summarizedas follows:• accurate correction of structured

backgrounds and spectral overlaps• correction at the exact analyte

wavelength• not limited to a wavelength range, as

with a deuterium source• self-referenced to reduce drift• less alignment necessary, due to single

source• efficient, optimized magnetic field

modulation ensures that the completemeasurement is at full magnetic fieldstrength

• field strength of almost 0.9 Tesla toensure complete line separation

• rapid field modulation of 100/120Hz toensure faithful tracking of fast furnacesignals

• relatively constant magnetic field strengthalong the length of the furnace cuvette toachieve efficient spatial correction

• alternative QuadLine method for thehighest sensitivity requirements andcombined QuadLine/Zeeman mode formethod development work

Correction Performance

Steady StateIt is difficult to generate large backgroundsignals using conventional flameatomization. Metal gauzes held in the lightpath are, therefore, used to generate largebackground-only signals. Since there canbe no atomic absorption in this experiment,any deviation of the background correctedresult from zero reflects the residual errorsin the system.

The results from a typical experimentperformed on an iCE 3000 Series AAspectrometer set up for nickel analysisusing the default parameters and making10 replicate measurements, each of 2seconds duration, are shown in Table 1.

They show that, even with abackground signal approaching 2absorbance units, residual backgroundcorrection errors are of a similar order tothe optical noise in the system. In a realsample measurement, these errors wouldbe lost completely in flame and atomization noise.

Emission BreakthroughThe background corrected determination ofchromium in a nitrous oxide/acetylene flameis a severe test of the backgroundcorrection systems ability to reject flameemission signals. Not only is the chromiumresonance line at 357.9 nm extremely bright,but the deuterium arc lamp intensity is alsobeginning to fall in this region of thespectrum. The optimum flame conditions forthe determination are such that an intensecyanide molecular emission, centred around355 nm, is produced from the flame. Theeffects of these phenomena on thebackground correction signal can beillustrated by comparing the chromium flamedetection limits with and without, backgroundcorrection. The results obtained from anearlier, analogue instrument are comparedwith the digital spectrometer in Table 2 .

The chromium detection limits of theearlier analogue instruments are clearlydegraded when the background correctionsystem is used, while the Thermo ScientificAA maintains its performance with nosignificant change.

GAUZE ABSORBANCE CORRECTED STANDARDABSORBANCE DEVIATION

1 0.353 0.0019 0.0006

2 0.506 0.0017 0.0009

3 0.860 0.0027 0.0007

4 1.285 0.0031 0.0028

5 1.803 0.0059 0.0044

Table 1: Background correction tests with gauzes

INSTRUMENT BC OFF BC ONDETECTION LIMIT DETECTION LIMIT(µg/mL) (µg/mL)

Analogue 1 0.033 0.115

Analogue 2 0.028 0.211

Digital Thermo Scientific AA 0.033 0.029

Table 2: Rejection of Emission effects

Product Specifications

Transient Accuracy andTransient ResponseLarge transient background signals arerelatively easy to generate by vaporizingrelatively large amounts of a typical matrixmaterial such as sodium chloride from agraphite furnace cuvette. Furthermore, if thespectrometer is set up on a non-atomicemission line, the ‘atomic’ signal must bezero, so that any errors produced by thebackground correction system will becomeimmediately apparent.

Figure 8 shows the result of atomising100 µg of sodium chloride at 1500 °C, withthe spectrometer set up on the non-absorbing lead line at 220.4 nm. Abackground signal approximately 1.4absorbance units high has been produced,but the corrected signal (shown in Figure 9)shows no significant deviation from thebaseline. Increasing the amount of sodiumchloride to 200 µg increases the backgroundpeak height to around 2.6 absorbance units(Figure 10). Expanding the corrected signal(Figure 11) reveals a significant increase innoise at the background peak maximum (a2.6 absorbance signal will reduce theintensity of light reaching the detector toapproximately 0.25 % of the value at zeroabsorbance), but, as before, there is noindication of a systematic error present.

By expanding the time axis of thesesignals, it is possible to measure the actualrate at which the background signal ischanging. At the point at which it ischanging most rapidly, on the leading edgeof the peak between 0.7 and 0.9 absorbanceunits, the measured rate of rise is slightlyover 25 absorbance units per second.

Figure 12 shows the corrected signalgenerated by atomising 10 µL of areference sea water containing less than0.2 µg L-1 of chromium, while the samesample when spiked with 4 µg L-1 gives thesignal shown in Figure 13. The chromiumsignal is clean and undistorted, and can bereadily quantified.

Figure 12: Corrected chromium signal andbackground for unspiked seawater

Figure 13: Corrected chromium signal andbackground for spiked seawater

Figure 10: Background signal from 200 µg of NaCl

Figure 11: Corrected signal under 2.6Abackground peak

Figure 8: Background signal from 100 µg of NaCl

Figure 9: Corrected signal under 1.4A backgroundpeak

Correction MethodContinuum SourceIt will normally be the case thatadvantageous measurement sensitivityimprovements can be obtained by employingthe QuadLine correction method inpreference to Zeeman. Figure 14 illustrates atypical improvement achieved in the case ofthe analysis of copper. This may often proveto be the difference between being able toactually perform a given analysis, or not,when working at the extreme trace levels.

Figure 14: Improved sensitivity for QuadLine (- - -)compared with Zeeman (----) for 200pg Cu

ZeemanWhen structured backgrounds, spectraloverlaps or correction in the visible part ofthe spectrum are encountered, Zeemancorrection will often solve the problem.Figure 15 illustrates the example of a cobaltinterferent at the gold wavelength, ananalysis that would prove impossible if onlythe continuum source method were available.

Figure 15:The improved accuracy of Zeemananalysis for Au in the presence of 150µg Co

Combination ModeSince the Zeeman technique can only beapplied during the relatively short atomizationphase of the furnace, due to magnetoverheating restrictions, adequate knowledgeof the efficiency of other pretreatmentphases, or clean-out phases, of the furnacecycle will not normally be available.

Using the unique Thermo Scientific iCE3500 and iCE 3400 Combination mode, it isnow possible to view the progress of theanalysis in all phases by employing Zeemanfor the main measurement phase and usingthe Quadline technique in all other phases.

SummaryThe QuadLine and Zeeman backgroundcorrection systems designed for the ThermoScientific iCE 3000 Series range of atomicabsorption spectrometers are capable ofcorrecting for background signals up to at leasttwo absorbance units, which can be changing atrates exceeding 25 absorbance units per second.They provide a level of performance likely tomeet the needs of all analysts performing flameor furnace atomic absorption measurements.

Product Specifications

©2010 Thermo Fisher Scientific Inc. All rights reserved. All trademarks are the property of Thermo Fisher Scientific Inc. and its subsidiaries. Specifications, termsand pricing are subject to change. Not all products are available in all countries. Please consult your local sales representative for details.

Africa-Other +27 11 570 1840Australia +61 2 8844 9500Austria +43 1 333 50 34 0Belgium +32 53 73 42 41Canada +1 800 530 8447China +86 10 8419 3588

Denmark +45 70 23 62 60Europe-Other +43 1 333 50 34 0Finland /Norway/Sweden

+46 8 556 468 00France +33 1 60 92 48 00Germany +49 6103 408 1014

India +91 22 6742 9434Italy +39 02 950 591Japan +81 45 453 9100Latin America +1 608 276 5659Middle East +43 1 333 50 34 0Netherlands +31 76 579 55 55

South Africa +27 11 570 1840Spain +34 914 845 965Switzerland +41 61 716 77 00UK +44 1442 233555USA +1 800 532 4752www.thermo.com

PS40690_E 01/10C

Thermo Electron Manufacturing Ltd(Cambridge) is ISO Certified.

iCE 3300

iCE 3400

iCE 3500