Home

References - International Atomic Energy Agency

1. Spectra for the calibration directory QXASdemo Cd 109 SPE INP Ti spe Cr spe Fe spe Co spe Ni spe Cu spe Zn spe Ge spe Zr spe Nb spe Mo spe Hf spe Ta spe W spe Au spe Pb spe KH2PO4 spe CaCO3 spe K2Cr207 spe MnO2 spe Input files directory QXASdemo Cd 109 SPE INP Ti inp Cr inp Fe inp Co inp Ni inp Cu inp Zn inp Ge inp Zr inp Nb inp Mo inp Hf inp Ta inp W inp Au inp Pb inp KH2PO04 inp K2 Cr207 inp CaCO3 inp Cr2 K207 inp MnO2 inp AXIL result files directory QXASdemo Cd 109 ASR Std Ti asr Cr asr Fe asr Co asr Ni asr Cu asr Zn asr Ge asr Zr asr Nb asr Mo asr Hf asr Ta asr W asr Au asr Pb asr KH2PO4 asr K2 Cr207 asr CaCO3 asr Cr2 K207 asr MnO2 asr Source file Cd 109 sou Calibration file Final cal Instrumental parameter file Cd 109 fpc 8 2 1 Elemental sensitivities Fifteen standards asr files for K calibration are available for the METHOD Elemental sensitivities KH2PO4 for potassium CaCO3 Ti Cr Cr2 K207 for chromium MnO2 Fe Co Ni Cu Zn Ge Zr Nb and Mo K2 Cr207 may not be used because of the enhancement effect Chromium excites potassium Final cal is the result of this calibration values for the angle of incidence of 52 5 and take off 77 1 were used The Mn calibration point is problematic this behaviour was already observed for the secondary target excitation 131 alibration file C 11_NOUNQXASDEMONCD 109 F inal cal
2. Input files E cal inp Ti inp Mo inp S7 1l inp Cr inp KBr inp Hf inp PbO2 inp Samp bl inp Instr bl inp 1 2 1 For anew spectrometer what settings of amplifier and ADC are appropriate To lead off with a completely new spectrometer both settings can be at preset values according to the needs of gamma spectrometry far away from the needs of XRF consequently one can maybe observe nothing meaningful in measured spectra or at the level of QXAS problems can occur The ADC spectrum size should be adjusted to either 1024 or 2048 channels for each spectrum because AXIL cannot handle spectra with more channels Otherwise during the spectrum conversion process one either has to have cut parts of the too long spectra or compression must be selected In many cases both options are undesirable With the coarse and fine gain of the amplifier one sets the amplification to 10 20 eV channel With such a selection the spectra will be displayed up to the scatter peak region For the sake of fitting it will be advisable to truncate the sample spectra only after the Kg scatter peak e g at approximately 25 keV for Rh secondary target excitation For practical purposes one will do the adjustment acquiring a spectrum of a sample or a single element standard like a pure metal Fe Cu etc in the sample position while any time the gain settings are changed a new spectrum acquisition is initiated 1 2 2 What is an energy calibration How t
3. The first three criteria as already used to evaluate the fit results of the standards apply also for samples total Chi square gt 3 residual within 3 individual Chi square values gt 3 After the first round the total Chi square is found with a value of below 3 but the residual as well as the fitted spectrum itself show a problem at Ca and Fe and the report provides bad 151 individual Chi square values for the two elements Therefore remove Ca and Fe and add instead X LINES ADD CA KA CA KB FE KA FE KB Alternative to this approach one can define the matrix of the sample but for unknown samples this is unrealistic because the composition is known only after the quantification step For severe problems with fluorescence peaks either the peak shape correction an increase of the background parameter a slightly better defined energy calibration or the splitting of the ROI into sub regions should be the last choice can improve the situation The background parameter value also is to be increased because the edges of the fitted ROI are not well described as can be seen from the residual change with the COMMAND BACKGR PARAM to a value of e g 10 The refit with the named improvements splitting of Ca and Fe inclusion of the peak shape correction and an increase of the background parameter will result in the fulfilment of the three criteria for this spectrum Two more criteria must be checked f
4. 10 POTTS P J TINDLE A G Evaluation of spectrum overlap corrections in EDXRF using the digital filter deconvolution procedure application to selected interferences encountered in the microprobe analysis of minerals X ray Spectrom 20 1991 119 11 KUMP P Some considerations on the definition of the limit of detection in X ray fluorescence spectrometry Spectrochim Acta Part B At Spectrosc 52 1997 405 12 CURRIE L A Detection and quantification limits origins and historical overview Anal Chim Acta 391 1999 127 157 13 14 15 16 17 18 19 20 21 22 23 24 25 26 158 MAACK BISGARD K LAURSEN J SCHMIDT NIELSEN B Energy dispersive XRF spectrometry using secondary radiation in a cartesian geometry X ray Spectrom 10 1981 17 KRAUSE M O Atomic radiative and radiationless yields for the K and L shells J Phys Chem Ref Data 8 1979 307 MCMASTER W H DELGRAND N K MALLET J H HUBBELL J H Compilation of X ray cross sections UCRL 50174 Lawrence Livermore Laboratory University of California USA 1968 RANI A NATH N CHATURVEDI S N Effect of Coster Kronig transitions on L3 sub shell XRF cross sections X ray Spectrom 18 1989 77 SCOFIELD J A Exchange corrections of K X ray emission rates Phys Rev A Gen Phys 9 1974 1041 SCOFIELD J A Hartree Fock values of L X ray emission rates Ph
5. TSample D i1_NOU QKASDEMO TRRFE MULTIE 3 ASR Measurement date 66 86 4 Live time 1666 sec Tube current 38 666 mA Reference Y Ka Int 483166 3465 Conc 385 000 ppb Analysed elements E counts compound 4212 518 U 2843 214 4876 214 1 493 ppb 23197 256 7235 169 16548 4 435 ppb I i I I I I i i Fig 9 14 Quantitative results obtained for an aqueous sample measured with TXRF In a strict sense the copper results are out of the calibration range calibration standards for this element e g with 5 and 10 ppb should be prepared and included into the calibration Due to the addition of the internal standard solution a dilution factor of 1 02 applies I e the reported results must be multiplied with this factor 143 CHAPTER 10 ELEMENTS OF QUALITY CONTROL References 27 50 25 51 52 For good laboratory praxis keeping a log book is essential Better are two log books one documenting the sample preparation date of preparation unique sample preparation number calibration standard sample description weight for pressing fluxing procedure etc The other documenting the spectrum acquisition date of measurement spectrum file name sample preparation number as shake hand between the two log books acquisition time live time LT dead time DT tube current when applicable Any essential changes replacement of protective foils changes in geometry other high voltage set
6. KANNGIEBER B LANGHOFF N WEDELL R WOLFF H Eds Practical X Ray Fluorescence Analysis Springer Verlag 2006 3 VAN ESPEN P Spectrum evaluation Handbook of X ray Spectrometry 2nd edn revised and expanded VAN GRIEKEN R E MARKOWICZ A A Eds Marcel Dekker 2002 4 HUBBELL J H VEIGELE W J BRIGGS E A BROWN R T CROMER D T HOWERTON R J Atomic form factors incoherent scattering functions and photon scattering cross sections J Phys Chem Ref Data 4 1975 471 5 VINCZE L JANSSENS K VEKEEMANS B ADAMS F Monte Carlo simulation of X ray fluorescence spectra Part4 Photon scattering at high X ray energies Spectrochim Acta Part B At Spectrosc 54 1999 1711 6 STATHAM P J Deconvolution and background subtraction by least squares fitting with prefiltering of spectra Anal Chem 49 1977 2149 7 STATHAM P J A comparative study of techniques for quantitative analysis of the X ray spectra obtained with a Si Li detector X ray Spectrom 5 1976 16 8 MCCARTHY J J SCHAMBER F Least squares fit with digital filter a status report HEINRICH K F J et al Eds Proc workshop on energy dispersive X ray spectrometry held at the National Bureau of Standards Gaithersburg MD 20760 April 23 25 1979 NBS special publication 604 1981 443 9 SCHAMBER F H Curve fitting techniques and their application to the analysis of ED spectra ibid
7. The relevant calibration file clb and the blank file asr names usually the same blank as already used for this calibration procedure both without extension are to be entered Independent of the selection for a report file name previously used the results can be saved in a file with the extension rpt It is possible to create a new report file or use the same report file as for the calibration measurements The calculation will start after the unknown sample asr file name was entered For a correct calculation run the number of iterations will briefly be displayed ending with a screen named Results of analysis With the calibration Soil clb it is possible to evaluate the IAEA reference standard SL 3 AXIL result file LakeSed3 asr because the calibration standards were selected in order to represent the sample elements except Rb and the scatter peak ratio compares well for Soil 7 and the sample 122 Results of analysis Z El Concentration Si 2 564E 6661 53 7E 8663 1 266E 6661 2 815E 8663 4 526E 8064 1 663E 6662 2 188E 6605 3 714E 6606 3 97 7E 6685 5 217E 0004 1 454E 0005 2 399E 0004 6 182E 8666 1 916E 6665 ETI Cwt Cwt Cwt Cwt Cwt Cwt ETI ETI ee co oe Cwt Cwt Cwt Cut f Std Dev 2 8E 6663 1 2E 6664 3 6E 6664 3 3E 0005 1 GE 6665 1 3E 6665 5 9E
8. Therefore in strict terms XRF falls into the category of an analytical process that can not be calibrated is unfortunately through its corrections A and H a function f c c C dark matrix 155 The results M repetition measurements are reported as cx and confidence intervals C in the form of c CI The concentration cx is calculated according to the formalism of CHAPTER 2 Fundamentals of XRF Theory and the confidence interval Cl is given by _ shy JM The standard deviation s has to be taken from a precision control chart ty is a t table value for a certain significance level usually 95 as a function of the degrees of freedom CI 10 8 XRF with thin samples the ET METHOD absorption correction is established independent from the sample measurement and TXRF are analytical processes that can be calibrated the unknown concentration can be calculated with the pre mentioned calibration according to T a Cc 10 9 b 10 9 re yes a ANA Eee 10 10 re 10 10 Also the confidence interval CI can be established from the calibration data Teor _ corr cal 2 eeso E 10 11 N M b Q The process standard deviation s is defined as pet ajor 2 s t 2 N 10 12 with ror bec a 10 13 and oDe Ee 10 14 156 REFERENCES 1 TSUJI K INJUK J VAN GRIEKEN R Eds X Ray Spectrometry Recent Technological Advances Wiley 2004 2 BECKHOFF B
9. asr files of standards that do not contain any useful fluorescence peaks but only the scatter peaks information cellulose HWC etc or to fit only the scatter region and neglect the fluorescence region Soil 7 etc The asr file must contain information of the form COH_SCAT and INC_SCAT In the later case the matrix composition must be specified FORM Menu for setting up options 7 There are known compositions 110 enu for Seting up Options Sample contains no matrix Elements exist as elements Matrix composition is not known Dilution material is not used Scatter peaks Cif used gt are from the same spectrum Secondary enhancement is corrected AE known compositions Normalization of concentrations is applied Report is not surely done Fig 7 6 The toggle field 7 there are known compositions will permit to define other elements than fluorescing ones later during the execution of the program Consequently one will have to define the matrix For e g cellulose with the chemical formula C6H 005 one can calculate with the utility Calculation of average atomic number the weight percentage of the elements nown Composition Ele Percent le Percent H 6 226600 Cc 44 445000 0 49 34000A W 666088 aalala 6 8 Ba 6 roan AG aa 8 a08008 LALA 9900008 6 666000 a 668000 6 660068 Fig 7 7 Definition of the composition of a cellulose pellet 7 2 1 Example Establishin
10. 74 5 895 1 8HHHH 74 1 6 484 143 172 12 9 S4E 1 31 6 491 1 8HHHH 31 3 2 532 951 178 44 9 65E 1 Fig 10 3 Full report for an AXIL fit of a Mn standard used for the control of the FWHM 10 1 3 Stability of the energy calibration For the Rh secondary target spectrometer a drift of the energy calibration was observed This behaviour is unpractical because each spectrum as can be seen by a closer inspection of the provided demonstration inp files needs an energy calibration for its own In case also a deterioration of the resolution had been found the combination of the two effects can be an indicator for some severe detector problem For the isolated problem either the amplifier or the MCA card is responsible but otherwise it is harmless because intensities must be constant independent of the gain In order to control the situation a Zr foil had been measured repeatedly The channel number of the maximum of the Zr Ka peak determined during the AXIL fit as the median of the Gaussian and accessible by the COMMAND REPORT FULL GO was plotted against the time since the first evaluation in days Sometimes measurements were taken within one day like at day zero 3 measurements and after four days 2 measurements This should demonstrate that the drift appears only during longer 146 periods of time Short term drifts would affect individual spectra and could not be accepted At day thirty five the amplifier gain was changed
11. GOLD LAYER THICKNESS CM 6666856 SI_DETECTOR DEAD LAYER CM o660166 DETECTOR SENSITIVE DEPTH CN 3668 NUMBER OF EVENTS 999999 NUMBER OF NET EVENTS 202600 COSECANT FOR PRIMARY X RAY 26896 COSECANT FOR SECONDARY X RAY 1 62579 EFFECTIVE SAMPLE RADIUS CNM 1 834 DETECTOR EFFECTIV RADIUS CN 316 SAMPLE ELEN ABS INT Wel FRAC SENS 2 JEI Ti 128 06 50 94 1 9688 6524 22 THICK STD CR Cr 165 26 193 95 1 9688 11457 24 THICK STD FE Fe 94 71 266 98 1 6608 19626 26 THICK STD co Co 33 31 261 40 1 6608 23215 2 THICK STD NI Ni 39 62 331 33 1 6606 29695 28 THICK STD cu Cu 82 46 411 87 1 6608 33963 29 THICK STD cN n 83 34 496 60 1 6608 41633 36 THICK STD GE Ge 76 368 689 40 1 6688 52945 32 THICK STD ZR r 36 56 1459 75 1 6688 126264 46 THICK STD NB Nb 89 73 1534 57 1 9688 137693 41 THICK STD MO Mo 91 35 1572 71 1 6606 144455 42 THICK STD VALUES FOR THE INDIVIDUAL COSECANTS ELEN CSCO csct1 SENS IND GEQ IND GEQEFF agli 1 269 1 626 527 5516 r 15E 05 Cr 1 268 1 026 1147E 05 S284 62396 05 Fe 1 266 1 026 1904E 05 5174 r254E 05 Co 1 265 1 626 2323E 6S 5191 P3848E 65 Ni 1 264 1 926 2971E 05 5198 364E 05 Cu 1 263 1 026 3398E 05 5139 621E 05 n 1 262 1 626 4164E 05 5187 536E 05 Ge 1 259 1 026 5291E 05 5284 9410E 05 er 1 252 1 026 1256E 86 5588 25026 05 Nb 1 252 1 026 1369E 06 5533 8207E 05 Mo 1 251 1 026 1436E 06 5519 8024E 05 AVERAGE GEOMETRICAL FACTOR G 5328 VARIATION COEFFICIENT OF G 4H 3
12. The accumulative spectrum containing all the lines of the named elements was totally counted for 519 s The same procedure was repeated but for the second round with equal counting times and current settings for all standards with the sample Soil 7 in front The settings for the live time for the individual target elements had been selected to have comparable counts in the peaks of the target plus sample in front spectrum the tube current was modified to adjust the dead time 5 1 3 Example Sample Soil 7 and target T1 Demonstration files directory QXASdemo ET Calibration file Soil7 cal Spectra Soil7 1 spe T1 spe S7 1_T1 spe directory QXASdemo ET SPE Input files S7 1 inp Tl inp S7 1_T1 inp directory QXASdemo ET INP AXIL result files Soil7 l asr Tl asr S7 1_Tl asr The sample was pressed as pellet with 2 5 cm diameter and consists of equal parts of the reference standard material IAEA Soil 7 and binder As binder HWC chemical formula C3gH76N202 was used The total mass of the intermediate thick pellet is 0 1055 g The sample spectrum Soil7 1 spe was acquired for 1000 s with a tube current of 40 mA For the Rh secondary target the incident and take off angle are 45 respectively The path length in air is 0 5 cm In the fit model S7 1 inp the matrix of the sample was defined 50 HWC 50 Soil 7 The characteristic lines of the following elements had been included X LINES ADD S
13. The calibration established by this METHOD is also used as input for the Emission transmission METHOD 49 lew calibration Elemental sensitivities Excitation source Angle of incidence ETAN 45 BAHA Detector take off angle deasa 45 000A Source type Tube voltage CKU gt 5a CEE Source name iiim ENE Average excitation energy KeU gt 20 6000 File name HAVRE Date lt mm dd yyyy gt 16 14 2666 lt Arrows gt move lt Enter gt Save_Change lt Esc gt Done Fi HELP replace on lt Ins gt Fig 3 1 Definition of a calibration file as used by the METHODs Elemental sensitivities and Emission Transmission 3 2 1 Example compound cal For the calibration compound cal using chemical compounds twenty four calibration standards asr files were included Add standards MgSO4 K2CO3 K KH2P04 CaCO3 Ti02 V205 Cr2 K207 MnO2 Fe203 CoO NiO CuO ZnO As203 SeO2 Br KBr SrCO3 Y203 ZrO2 Nb205 Ge Std Mo Std and Pb LStd were included because no compound standards for these elements were available S Std was included to assure the calibration point for sulphur Pb Lstd asr is the only representative of an L line emitting standard No effort was made for this calibration to extend the L lines calibration with more elements because all samples treated later contain only lead as sole emitter of L lines The standards P KH2P04 K2 Cr207 K KBr may not be used for this METHOD because phosphorus is enhanced by
14. assumed that the matrix of the sample SL 3 is comparable to Soil 7 as one consequence From the ratio of the scatter peaks for the intermediate thick pellet of the sample Soil 7 with binder added an average atomic number of 8 88 follows as another consequence When one wants to treat IAEA SL 3 lake sediment as unknown sample the scatter standard coming closest in respect to the scatter peak ratio is IAEA Soil 7 ignoring the fluorescence standards spectra K KH2PO4 and P KH2PO4 7 2 Full Fundamental Parameters METHOD Demonstration files directory QXASdemo FP Scatt Instrumental parameter files Soil7 fpc OrgaMatr fpce Spectra directory QX ASdemo Fp Scatt SPE LakeSed3 spe Input files directory QXASdemo Fp Scatt INP SL3 A inp SL3 B inp SL3 1 inp S7 2 1 inp AXIL result files LakeSed3 asr OragaMatr asr Soil7 2 asr 109 It is to be mentioned that the METHOD uses only the total scattering cross sections Ocoh Oinc instead of the differential scattering cross sections in this respect the BFP METHOD soon introduced is superior The fluorescence calibration of this METHOD for e g Soil fpc had been already established chapter 3 3 Calibration for the Full Fundamental Parameters METHOD This file carries the average and some individual instrumental constants information but for samples with dark matrix the scatter calibration must be established In principle it is possible to work with a fp
15. incorporating elements not even represented by a single calibration standard It is recommended to make as little use as absolutely necessary of this possibility The same arguments apply as already given for the calibration of METHOD Elemental sensitivities For reliable quantitative results each element contained in a sample should be represented by several calibration standards 9 3 Quantitative analysis of a water sample for the elements Mn Cu and Sr Demonstration files directory QXASdemo TXRF Spectrum MultiE 3 spe AXIL result file MultiE 3 asr Input file MultiE inp To an aqueous sample taken from a mixture of standards multi element solution X yttrium had been added The content of Y in this samples is 385 ppb Onto a reflector 30 ul of the yttrium spiked sample was pipetted and treated like the standards concerning the matrix removal The tube current could be again set to the maximum of 38 mA the spectrum was acquired for 1000 s 142 Spectrum MULTIE 3 SPE Iteration 10 ChiSquare 1 5 Dif Multi Element X 3 Display BEG beg chan END end chan MIN min cnts MAX max cnts ROI SPECTR RES IDUAL LIN LOG GO CANCEL gt gt DISPLAY RESIDUAL Fig 9 13 AXIL fit of the spectrum MultiE 3 spe measured in TXRF geometry The calibration included only the elements Mn Cu and Sr quantification is only possible for them For the sake of the AXIL fit other elements had to be included
16. intention to keep the system dead time below 20 for all standard measurements These 20 are far below the problematic range of non linear response and as another consequence the spectra do not suffer from spectral distortions as observed with spectra of standards collected with higher dead time The acquisition time live time LT was selected to adjust the total counts in the peaks of interest The rather high number of 50 000 counts for the peaks of relevance results in statistically well defined signals relative standard deviation lt 0 5 whereas the approximate equality gives all standards equal statistical weight Reference standard materials are usually not well suited as calibration standards because of their trace elements content which will introduce high statistical uncertainties As a recommendation if more than one element that could be used for calibration is contained in a standard two independent spectra should be collected with measuring times according to the above considerations All standards spectra had been fitted with input files as specified in the EXCEL standards xls file By use of the correct fit model among others for all spectra a standard deviation SD 2 N with Mn being the net peak area for the element n of interest could be found In CHAPTER 10 Elements of quality control a detailed guideline for calibration standard spectra treatment is given Note Information about all relevant details and
17. standard material for which the complete composition is known therefore also standards used for fluorescence calibration are incorporated All above used spectra as defined in table 7 1 and selected spectra spe and inp files for the definition of the scatter region Al Std Al Std 1 Si wafer Si 1 MgSO4 MgSO4 1 P KH2PO4 P KH2 1 K KH2PO4 K KH2P 1 K2CO3 K2CO3 1 Ti02 TiO2 1 K2Cr207 K2Cr20 1 S Std S Std 1 Ti Std Ti Std 1 K KBr K KBr 1 and Br KBr Br KBr 1 were utilized The scatter peak ratios follow a trend but some standards like MgSO4 pure aluminium silicon etc are problematic in this respect The more standards can be used to establish such a plot the better the predictions can be made concerning the correlation between average atomic number and scatter peak ratio of unknowns C38H76N203 binder HWC CgH 905 cellulose KHzPO binder 2 spectra Soil7 binder 3 intermediate thick MgSO binder pure aluminium o SL3 lake sigon l sediment Soil7 7 e sulphur KBr binder usta 2 spectra Tio pure titanium KeCO4 binder binder KCrO binder ratio incoherent to coherent scatter peak 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 average atomic number Fig 7 4 Ratio of incoherent to coherent scatter peak versus average atomic number of the substance used for scattering Aluminium and silicon do not follow the trend It can be
18. 1 2 1 Example Preliminary calibration file Cd_52_90 cal ee eee ence 125 8 1 3 Definition of the excitation source and geometry for the Monte Carlo STATA ALI ODD sas gente Garant E Meee cutee Sane enieneie nd gaan eset aaah 126 8 1 4 Results for the incidence and take off angle and verification eeeeeeeeeeeeteeeee 130 8 2 CALIBRATION vii cite tiieseese Seti alienate a nade nate ad 131 82l Blemental Sensitivities acecc ase etiy tole oie kee ei glee de de ee 131 8 2 2 Full fundamental parameters vic ciccsisissssaecassaseesasdsvseeaavenaceise soaeonavepaanasevsaceasnnecavs sobenas 132 CHAPTER 9 TOTAL REFLECTION XRAY FLUORESCENCE TXRF aaeeea 135 9 1 FUNDAMENTALS enaena eins ohh ae eee 135 9 2 CALIBRATION nenatis e eei le e See ae esa 138 9 3 QUANTITATIVE ANALYSIS OF A WATER SAMPLE FOR THE ELEMENTS MN CU AND SR jaissscsisisssveitesscrsctiscyanctasectasti aazeiessavertuscsenes 142 CHAPTER 10 ELEMENTS OF QUALITY CONTROL eee eecesecseeeneeeeeeseeeseenaes 144 10 1 PARAMETERS UNDER CONTROD ws scsi saris dus bende 144 10 1 1 Linearity as a function of the tube Current eee ceeeeeceeececeeeeeceeeeeceeeeecsteeeesaeees 144 10 1 2 Detector resolution as function OF TIC igs oui oSiattega is shad wasn doy sadness eaves 145 10 1 3 Stability of the energy Calibration c cis cg csascesdcctasssine case sedarasnssacevensdcnenseesencncesscecaanseads 146 10 2 HOW TO FIT STANDARDS SPECTRA ACCORDING TO QA 147 10 2 1 Be
19. 1000 GO Channel Number CANCEL gt gt DISPLAY RESIDUAL _ Fig 9 6 Spectrum of the calibration standard containing 100 ppb of Mn Cu Sr and Y as internal standard element with 500 ppb A calibration file Extra2 cal was created FORM New calibration Regression of count rate vs conc As Procedure for TXRF always Element as internal standard will be selected default the other options can be applicable for thin film samples analysis measured in standard XRF geometry Essential are the definitions for the internal standard element its fluorescence peak used Line which normally will be Kg and the concentration units if the lt gt character is needed it must be typed twice lt gt The instrument used for the measurements taken was an ATOMIKA Extra 2 spectrometer equipped with a Mo X ray tube and for excitation beam shaping a so called cut off reflector A tube voltage of 50 kV was applied for all the measurements In the pre last line the file name has to entered under which the calibration will be stored 139 ew calibration Regression of count rate vs conc Procedure Element as internal standard Element 3 Line K_a Conc Unit ppb Excitation source Tube excited XRF Source type Tube voltage lt KU gt 56 6606 Source name EXTRA II File name JESSEN es Date lt mm dd yyyy gt 1 17 2007 Fig 9 7 Definition of the calibration file TXRF cal With Add Standards the ten calibration standards
20. 16 Fig 8 9 The temporary file temp dat contains a wealth of information among others individual and average geometrical factor The average geometrical factor has a value of 5328 with a so called variation coefficient of 3 1 130 In the spectra of various samples thick samples Cabbage spe Lichen spe SL3 spe Soil7 spe etc intermediate thick sample SL7 thin spe the Compton peak of Ag Kg is found at a constant energy position namely at 20 45 0 02 keV By transformation of the Compton formula formula 2 18 one can deduce the scattering angle from the position of the Compton peak maximum in the spectra coss tesir LZ 8 3 0 v All energies are in keV With Ey 22 10 keV weighted average energy of Ag Ka and Ag Koy and Es 20 45 keV the scattering angle for this particular geometry is 150 0 The sum of the calculated effective incidence and take off angles 129 6 is not close to this experimentally determined scattering angle A slight deviation could be explained by a the inaccuracies introduced by taking all necessary measures with sliding callipers and b the fact that also the maximum of the Ag Kg Compton peak in the spectra exhibits slight variations depending on sample type and thickness but these deviations are much less than the observed difference between the sum of the calculated angles and the observed scattering angle 8 2 Calibration Demonstration files directory QXASdemo FP Scatt
21. 6606 4 6E 66067 8 5E 0007 2 GE 6606 6 3E 0007 1 4E 6606 4 1E 6667 1 6E 6666 U_Tot 8 92E 0002 4 04E 0002 3 15E 0002 3 06E 0002 1 51E 0002 1 21E 0002 5 8 E 8061 2 58E 6061 1 97E 6061 1 74E 6061 1 56E 6061 1 46E 6061 1 27E 6661 3 47E 6661 lakesed3 ASR gt No of iterations 16 F_Enh_Tot 1 164E 6666 1 136E 6066 1 645E 6666 1 642E 8668 1 627E 6666 1 626E 6666 1 628E 6666 1 631E 6666 1 636E 6606 1 628E 8666 1 029E 000A 1 629E 8666 1 636E 6068 1 629E 6666 F_Enh_Scat 1 619E 6606 1 626E 6606 1 626E 6606 1 626E 6606 1 621E 6666 1 622E 6666 1 624E 6606 1 028E 000A 1 629E 8606 1 629E 8606 1 636E 8686 1 625E 6660 Estimated mass per unit area of the sample 2 65591E 6061 g cm2 Total weight fraction of fluorescent elements 4 614E 80661 Average atomic number of the sample dark matrix Z 6 4 Save peak ratios CY Y ND in Next sample CY ND Fig 7 17 Quantitative results for the reference standard IAEA SL 3 are displayed with normalized concentration values where 1 0E 0 corresponds to 100 The concentration values are given as weight fraction 2 564E 0001 corresponds to 25 64 for e g Si In case the AXIL fit had not been carefully examined in order to eliminate elements with too low peak areas all results have to be checked so that the standard deviation Std Dev does not exceed the corresponding concentration value which would mean the respective result i
22. For these X ray lines the number of the characteristic photons produced is proportional to Ty O Pca 2 28 Tn Ox Pree 2 29 Ta Ox Pcp 2 30 Tn Ox Pcp 2 31 L L To Ta 5 TiO im fas em ce fis fiafas Pn 2 32 L L og ts 5 T O HE faa E iat fist files Pe 2 33 L TH T O fis ha pis 2 34 L L Ea Ta Tr O a fas gm Fa fis fioto3 Np 2 35 respectively In the absence of interfering effects when atoms of chemical element x emit characteristic radiation the intensity of a given characteristic line J is proportional to the mass fraction of the emitting atoms cx This holds for all characteristic lines seen in a XRF spectrum A set of equations can be defined Z c Lc I c etc I c etc up to Z c Unfortunately the peaks in the spectrum do not represent all chemical elements present in a sample In particular the Xray characteristic radiation originating from the sample organic matrix typically composed of elements like H C O N can not be detected It is due low fluorescence yields and strong absorption of the energetically low lt 1 keV characteristic radiation of these elements The not detectable part of the sample is called dark matrix If expressed in per cent the sum of concentrations of all elements present in any sample is 100 In terms of mass fractions 35 Cy pate C dark matrix z 1 2 36 I The direct proportionality between the intensity of X ray flu
23. The 41 spectra sufficient for our demonstration are only a small subset of all spectra collected during the scan In order to establish an input model describing not only a single spectrum or a limited number of spectra a sum spectrum Sum Spec spe was generated by adding channel by channel all the 41 individual spectra The fit of the sum spectrum with the model from SumSpec inp is shown in Figure 1 28 The elements Ca Fe Zn Sr and Pb were identified The adequate fitting of this spectrum is only possible with a linear background and highest order of the polynomial PARAM 30 and inclusion of sum peaks but this is not really the task of this demonstration 28 spectrum SUM SPEC SPE Iteration 3 ChiSquare 2 0 Dif 007 Display BEG beg chan END end chan MIN min cnts MAX max cnts ROI SPECTR RESIDUAL LIN LOG PoaHMon 1500 20668 cn Channel Number CANCE gt gt DISPLAY RESIDUAL Fig 1 28 An artificial spectrum Sum Spec spe obtained by channel wise summation of forty one individual spectra For the evaluation of the individual spectra also a linear background was selected but of order 0 to prevent unnecessary background overestimation The region of interest includes the escape peaks of Ca and stops after the Sr Ka peak The following peaks were included peaks in the model X LINES ADD CA KA CA KB FE ZN KA ZN KB PB LA PB LB PB LG SR KA All the fitting settings were stored in the in
24. The last used set directory is proposed as the default for the next run When the QXAS is run for the first time the C AXIL SPECT is proposed as the default set directory 1 1 1 Example Input model Targetl inp and spectrum Target1 sp of a demonstration sample QRAS 3 6 Quantitative K ray Analysis System Developed under the auspices of the International Atomic Energy Agency In cooperation with The University of Antwerp Belgium La Direccion Nacional de Tecnologia Nuclear Uruguay La Direccion General de Energia Nuclear Guatemala Ruder Boskovic Institute Yugoslavia Instituto de Asuntos Nucleares Colombia Current Directory C Last Working Dir C OK S NBS ASR Set directory i Fig 1 3 Definition of the default working directory for run of QXAS After defining the Set directory in the start screen the main QXAS screen containing the topmost program navigation menu is displayed To carry out the example select the option Spectrum fitting Axil X ray Analysis Package Spectrum format conversion r Q Quantitative analysis Utilities Fig 1 4 For access to AXIL Spectrum fitting must be selected Next select AXIL amp Voigt peak profiles option for high energy K lines Spectrum fitting Specify TRUM abe for ciate min noe ray library management Fig 1 5 Before loading a spectrum and a specific input model Load the input model line Select model file Targetl
25. They are found in the QXASdemo folder and its sub directories When executing the examples it is recommended to set the QXASdemo folder as the set directory Usually the C axil spect is the set directory The profit of distributing the example files among different folders is to limit their number in a given directory Another advantage when spe and asr files are put into different locations is that previous results of e g example asr files that are provided for demonstration purposes are not overwritten after attempts of the user to store his or her own results 1 1 Spectrum fitting AXIL Demonstration files directory QXASdemo Getstart Spectra Target1l spe Pb pure spe Input files Target1l inp Pb inp For historical reasons many users will associate the spectrum fitting part of QXAS with the name of AXIL which stands for Analysis of X ray spectra by Iterative Least square fitting The best way to start the QXAS software package is by using the QXAS icon on the WINDOWS desktop Fig 1 2 QXAS icon on the WINDOWS desktop The start screen of QXAS allows for setting the default working directory Set directory a default folder where input data are read from and result files are stored Still it is possible during programme execution to select other places to load files from however storing the majority of data is restricted to the working directory there are few exceptions
26. a calibration standard consisting of 4 g of KxCr2O7 and 0 9 g of HWC as binder The information gained with this utility cannot be transferred directly to the METHOD Elemental sensitivities taking notes will be necessary Alternatively open QXAS in a second window simultaneously and arrange the two windows such that both can be seen In the one run the METHOD Elemental sensitivities in the other the utility Calculation of average atomic number For the addition of a standard to the calibration knowledge about the tube current the measurement date not essential for tube excitation the live time usually transferred already from the asr file correctly is needed The Cr standard is infinitely thick in terms of XRF so the standard s sample mass was left at the default value 0 00000 all used example standards are infinitely thick The concentration values for all elements of compound and binder must be specified they are treated as one entity The most problematic fact in this context is the expected knowledge about the relative error in the concentration value of the element used as calibration point stddev The default value of 5 is definitely too high for well prepared standards It will overestimate this contribution to the total uncertainty The second contribution to the total uncertainty comes from the AXIL fit No effort was made to gain knowledge about the uncertainty in the concentration but 0 1 for a standard seems to be
27. and expanded VAN GRIEKEN R E MARKOWICZ A A Eds Mrcel Dekker 2002 WOBRAUSCHEK P Adjustment and working instructions for the total reflection attachment module IAEA TXRF module internal report Atominstitut der sterreichischen Universit ten Stadionallee 2 1020 Wien Austria 1989 DARGIE M MARKOWICZ A TAJANI A VALKOVIC V Optimized sample preparation procedures for the analysis of solid materials by total reflection XRF Fresenius J Anal Chem 357 1997 589 KLOCKENK MPER R VON BOHLEN A Elemental analysis of environmental samples by total reflection X ray fluorescence a review X Ray Spectrom 25 1996 156 GREAVES E D BERNASCONI G WOBRAUSCHEK P STRELI C Direct total reflection X ray fluorescence trace element analysis of organic matrix materials with a semiempirical standard Spectrochim Acta Part B At Spectrosc 52 1997 923 KLOCKENK MPER R Total reflection X ray fluorescence analysis Wiley 1997 KLOCKENK MPER R VON BOHLEN A Survey of sampling techniques for solids suitable for microanalysis by total reflection XRF J Anal At Spectrom 14 1999 571 DE BOER D K G Calculation of X ray fluorescence intensities from bulk and multilayer samples X Ray Spectrom 19 1990 145 INTERNATIONAL ATOMIC ENERGY AGENCY Sampling storage and sample preparation procedures for X ray fluorescence analysis of environmental materials IEAE TECDOC 950 I
28. are calculated for characteristic lines from standards correcting the self absorption in the standards It is assumed that the excitation radiation can be represented by a single X ray line with specified energy intensity weighted energy The sensitivity S of all elements n is calculated for which a concentration c 1s supplied to the current cal file and a peak area N can be read in by means of an asr file NA T LT c 8 1 LT is the acquisition time of the respective spectrum and A the absorption correction Ko lines and L lines are treated independently The decay correction T is defined as rse 221 8 2 1 2 where T is the half life of the used isotope t is the time interval between measurement and calibration both in matching units of time The report will contain individual results for each standard asr file and a summary of the calculated sensitivities lew calibration Elemental sensitivities Excitation source Radio source XRF Angle of incidence lt degrees gt z 52 0000 Detector take off angle degrees 94 8088 Source type z Cd 169 File name GAIN ONe SCS Ne ey ae Date lt mm dd yyyy gt 16 18 2066 Fig 8 1 Definition of the preliminary calibration file Cd_52_90 8 1 2 1 Example Preliminary calibration file Cd_52_90 cal As almost first input for the establishment of the preliminary calibration file an input for the Angle of incidence degrees is needed defined by the pr
29. brief descriptions of the above FORM with measures to be taken from reality 127 6 61 2 Fig 8 5 Sketch depicting the relevant dimensions in mm the used Cd 109 source excitation system Note Description of the geometry parameters as expected as input in the form Enter experimental setup parameters All measures had been taken with sliding callipers Figure 8 5 with dimensions in mm refers to the used letters and stated dimensions of Figure 8 4 and relates them to data input as expected by QXAS Warning It is easy to destroy a Be window of a detector Utmost care must be taken not to touch it Also the front surface of the sealed excitation source must not be scratched Hazard Sources as used for XRF will have activities around 1 GBq 10 100 mCi when they are new Minimize the duration of exposure when taking any measures Never touch a radioactive source with hand use forceps Respect the Vr law which means in words the greater the distance to a source the less dose rate received After the FORM had been filled in with relevant data or had been loaded utilizing the file Cd angle sen a source file has to be loaded Cd test sou The relevant data to be contained therein are the spectral lines Ag Ky Ag Kg and a y line at 88 keV with their respective wavelengths in Angstrom and their relative intensities describing a Cd 109 source Usually for a sou file also the geometry were of relevance defined
30. calibration standards and the samples the same approach must be used For the remainder of this book the model 1 will be used 7 1 1 Classification according to the scatter peak ratio Depending on the choice for representation of the scatter peaks the samples must be treated accordingly A classification of samples is possible through the ratio of the incoherent to coherent scatter peak area For every sample a scatter calibration must be achieved with a scatter calibration standard as similar as possible in the incoherent to coherent peak ratio Table 7 1 Incoherent and coherent peak area and their ratio for low Z matrix representatives and several sample spectra for Model 1 scatter peak representation me i Peak area Ratio Standard sample pmo Compton Peas incoherent aoe asr file for the Rh K coherent Rh description f Rh K to coherent scatter region K scatter scatter peak C3gH76N20 gt HWC HWC 1 3077129 475877 6 47 CoH 1905 Cellulos Cell 1 2033878 432407 4 70 IAEA 336 Lichen336 Lic336 1 2789663 590248 4 73 IAEA 359 Cabba359 Cab359 1 2925592 820512 3 57 IAEA Soil7 binder HWC Soil7 1 S7 1 1 243539 87794 2 77 IAEA SL3 LakeSed3 SL3 1 930573 501443 1 86 IAEA Soil7 Soil7 2 7 2 1 787234 492611 1 60 Instrument blank Instr Bl Ins Bl 1 11164 6187 One can further reduce the problem by plotting this ratio as a function of the average atomic number of all elements contained in a sample or standard Soil 7 is the only reference 108
31. edges The overall mass attenuation coefficient due to photoelectric effect is a sum of all the individual contributions Tn Th To pa T T m m ae 2 12 32 The photoelectric mass attenuation coefficient depends on the energy of the incoming photon Tn Tn E 2 13 The experimentally obtained values of the total photoelectric mass attenuation coefficients of elements tabulated as functions of photon energy can be found in literature At any given photon energy the contribution from a given absorption edge can be estimated with the use of the so called photoelectric absorption jump coefficient The absorption jump coefficient J is calculated as a ratio of the photoelectric mass attenuation coefficient at the energy slightly larger than the energy of the absorption edge e g for the K edge E Ex E to its value at the energy just below the edge E Ex E 8E lt lt Ex _ 1 EB T E EB r E h E 7 E 2 14 GE T E cht 7 cy E T E ty E K j t alm E 2 15 The value defined in 2 15 tells us what is the fractional contribution of a given photoelectric absorption edge to the overall value of the mass attenuation coefficient for photon energies only slightly larger than the energy of the edge It is assumed that the proportion given by formula 2 15 holds also for other photon energies significantly larger than the energy of the photoelectric
32. for K shell it is denoted by ax The relation between fluorescence and Auger yields depends on the atomic number Z For low Z elements up to and including Z 30 ay gt Wx 2 17 33 For the elements with atomic number Z gt 30 relation 2 13 is reverted which means that prevailing number of the K shell ionized atoms relaxes with the emission of X ray characteristic photon In case of the non s type atomic subshells e g Ly Lin Mn Mm Mrv My etc there is another significant process taking place during atom relaxation called Coster Kronig vacancy transition In this process the number of vacancies in a non s type atomic subshell is increased by radiationless and radiative transfer of vacancies from deeply bounded subshells of that shell The redistribution of vacancies increases the number of characteristic photons emitted form a given non s type subshell Without taking into account the Coster Kronig transitions the number of Xray characteristics photons emitted due to vacancies created by the photoelectric effect directly in the Lm subshell would be proportional to L Tr Oa 2 18 Considering Coster Kronig vacancy shifts from Ly to Lm from Ly to Lm radiationless and radiative and from L over Ly to Ly this number is increased to L L L Li g L Ta QO F Tn fi 3Q Tn fis T Tn fis Tn fiotr3 2 19 The fix factors are the Coster Kronig yields They define the fraction of vacancies that are shifted from i
33. is described by Beer Lambert law T 1 exp z P0 2 40 where Jp and Z are the photon fluxes before and after the absorber respectively The product pt can be expressed in terms of mass m in g and area F in cm of the absorber m 2 41 ane 2 41 When the beam impinges on the layer surface under an angle of the angle between the beam direction and the surface the path length is increased I exp zee p exp ean 2 42 sin Q 36 Note that the mass attenuation coefficient of a sample that is composed of several elements i is calculated as a weighted sum of the individual mass attenuation coefficients of elements lim My c Hh 2 43 where cj is the mass fraction of element in the sample The coherent scattering contribution outweighs the incoherent for lower energies but for higher energies the incoherent scattering contribution will reach a constant value outweighing the decreasing coherent contribution and even become more important than the photo effect towards the end of the X ray region 10000 1000 4 1 Bg 99 energy keV Fig 2 1 The contributions to the attenuation coefficient are the absorption coefficient photo effect and coherent and incoherent scattering coefficients Iron was selected as absorber which has its K absorption edge at 7 1 keV The self absorption effects depend on the path length from the sample surface to the location of the
34. lines namely Pb Ly The As Kg peak also would miss its Kg it is the Pb Ly peak Identifying the element Kr is a common mistake for beginners because it is only contained in air in very low quantities nowhere else and is usually not seen in spectra it is the Pb Lg peak The proposed Zr is one of the L lines of Pb Sulphur would not match well in terms of energy for the close by peak the Pb M lines are responsible Unfortunately M lines are not displayed with the COMMAND KLM MARK Recommendation for unknown samples containing lead identified by approximately equally high Pb La and Pb Lg peaks Due to the problem of line interference one has to check carefully for arsenic overlap of Pb La and As Ka and for sulphur in the presence of higher quantities of lead proximity of Pb M and S K peaks Add the element lead to the model with the COMMAND X LINES ADD PB LA PB LB PB LG or alternatively PB L3 PB L2 PB L1 The usual approach would be to add lead as one element entry with PB only for minor amounts of this element in a sample For this spectrum the matrix had been defined which already improves the fit results still the three sub shells of the L radiation are taken into consideration separately 1 2 Frequent questions raised when working with AXIL Demonstration files directory QXASdemo Getstart Spectra E calib spe Ti Std spe Mo Std spe Soil7 1 spe Cr Std spe KBr spe Hf LStd spe PbO2 2 spe HWC spe Instr bl spe
35. must be located in a single folder It is advisable to generate a single sum spectrum by adding all spectra channel by channel in order to identify all possible peaks of chemical elements 27 A suited fit model has to be defined before with the help of the sum spectrum in order to finally evaluate all spectra in batch mode An ASCII file must be created outside QXAS that carries the minimum information needed for each spectrum to be fitted by AXIL After having fulfilled all the previous the batch fitting routine is initiated with the COMMAND BATCH followed by keying in the file name containing the instructions Please examine the example batch file FitAll bat In contrast to all other kinds of files to be loaded when working with QXAS when entering the batch command the filename extension for our batch file bat must not be omitted 1 3 1 Example Forty one spectra of a u XRF scan All files necessary to carry out batch processing of spectra are contained in the directory QXASdemo Batch Original This directory must be set as default directory for QXAS data processing It is done in the QXAS start screen with Set directory QXASdemo batch Original For this demonstration forty one spectra from an area scan over a sample bone cross section containing Ca Zn Sr and Pb were taken The sample was excited by monochromatic synchrotron radiation beam energy of 17 1 keV Counting time for each spectrum was 1 s
36. n the axis intercept a can be written as a c b 1 10 6 Usually for QXAS applications the calibration for each element of interest as described above is substituted by one or only few calibration measurements of either pure standards with 100 concentration metals or simple compounds of highest possible concentration pressed as pellets with the following qualities Metal standards are usually homogenous and have a good surface quality smoothness but also must be flat resulting in a good precession Compounds are easier available but have to fulfil certain criteria like stability homogeneity particularly when a binder must be used and should not be hygroscopic otherwise they cannot be stored and when needed re used The purity must be close too 100 otherwise the impurity concentrations should be well specified The self absorption enhancement which must be calculated in order to obtain the corrected intensities is more accurately described for simple standards For some compounds the enhancement effect must be taken into consideration for which some METHODs cannot be used When a compound is used which forms a crystalline structure diffraction peaks complicate the peak fitting The corrected intensity for the sample element x net peak area Nx je Ne le 10 7 i t A H sample i e of the concentrations of the other fluorescing elements the dark matrix and the element itself
37. only constituents In order to get an overview about the element specific concentration ranges of the sample and for really unknown samples one will establish an intermediate calibration with pure metals only according to the sample elements found in the spectrum Note Description of the calibration standards spectra and input files for the AXIL fit in directory QXASdemo NBS SPE amp INP AXIL result and std files in QXASdemo NBS and concentration values as expected from the preparation process of a synthetic bronze sample SynBronz can be found in the EXCEL nbs_standards x s file After this intermediate quantification one usually will have to get appropriate standards For the used NBS 1103 1107 1108 and 1115 the concentration ranges spanned by the standards are not ideal except for Cu The highest copper concentration pure metal standard is 100 the lowest is 59 27 NBS 1103 therefore the expectation value for the sample Cu 95 87 56 is well bracketed Zn has a much lower Pb a much higher expectation value and Sn is only slightly above 6 2 AXIL fit for NBS bronze alloy standards The peak evaluation of the pure metallic calibration standards is in analogy to the other METHODs calibration standards Problematic are the reference standards materials NBS As example the spectrum fitting of NBS1103 spe is detailed The spectrum was split into 3 different ROIs with 3 different input models The careful description
38. potassium enhancement correction factor 1 036 potassium is enhanced by chromium correction factor 1 41 and potassium by bromine correction factor 2 21 respectively The secondary excitation effect cannot be handled by Elemental sensitivities As example how a standard is included as calibration point the file Cr2 K207 asr was chosen The Cr peak in the spectrum K2Cr207 spe of the calibration standard which consists of 81 63 K2Cr207 and 18 37 HWC chemical formula C33H76N202 as binder is used to establish the Cr calibration point The weight percentage of all elements can be calculated with the utility Calculation of average atomic number in Utilities 50 Calculate Mean 2 iol xi 301 an Enter formula and weight of your compounds Note CaCO3 has to be entered as Ca1C 103 Fig 3 2 Definition of the composition of a calibration standard consisting of 4 g of K2Cr207 and 0 9 g of HWC as binder in order to obtain the weight for each chemical element The calculation of concentrations in the mixture of the compound with the binder is initiated with a blank line after the last entry of a compound and its weight with lt Enter gt 51 CI ox Compound 1 K2CR207 4 0010 Compound 2 C38H 6H202 0 9010 Hr Element Z Weight Atoms 1 H 1 2 374 42 5 2 C 6 14 146 21 2 3 H 7 0 868 1 1 4 0 8 32 062 36 1 5 K 19 21 694 10 0 6 CR 24 28 856 10 0 MeanZ 14 5455 Fig 3 3 Weight results for
39. problem with the format of the date also other MCAs spectra formats might be affected which only is of relevance for source excitation because of the calculations for the decay correction For repair of spe files see CHAPTER 4 Editing of data files When the date format is left unchanged the ET METHOD will crash 8 1 Establishment of numeric values for incidence and take off angle Demonstration files directory QX ASdemo Cd 109 Spectra for the preliminary calibration directory QXASdemo Cd 109 SPE INP Ti spe Cr spe Fe spe Co spe Ni spe Cu spe Zn spe Ge spe Zr spe Nb spe Mo spe Input files directory QXASdemo Cd 109 SPE INP Ti inp Cr inp Fe inp Co inp Ni inp Cu inp Zn inp Ge inp Zr inp Nb inp Mo inp AXIL result files directory QXASdemo Cd 109 ASR Std Ti asr Cr asr Fe asr Co asr Ni asr Cu asr Zn asr Ge asr Zr asr Nb asr Mo asr Source file Cd test sou Calibration file Cd_52_90 cal Spectra Cabbage spe Lichen spe SL3 spe Soil7 spe SL7 thin spe Input file Cabbage inp Lichen inp SL3 inp Soil7 inp SL7 thin inp The major problem with source excited XRF is the ill definition of the incident and take off angle in respect to the sample surface The wide spread of angles of X rays the divergence complicates the quantification In the ideal case one would treat this problem with the Monte Carlo method or with an additional integration over the solid angles Non
40. procedure 24 2 5 Dif 017 Show lt T gt lt 44 gt lt Pg Up gt lt Pg Dn gt Fitting Region channels 718 1332 lt Home gt lt End gt Line Ener KeV Peak area i i lt Esc gt Pb La 10 542 542018 GO CANCEL Fig 1 25 Fit results that do not fulfil certain quality criteria for a PbO2 sample The above spectrum should not be incorporated into any method calibration due to another reason it was acquired with the specimen in a wrong position 1 2 28 How to handle blank problems Two kinds of blank spectra can be measured the sample blank and the instrument blank both must be taken with the same acquisition time as for the samples and the highest current setting for tube excitation The sample blank measurement provide information about the impurities introduced during the sample preparation procedure as well as the contribution coming from the instrument itself modified by the presence of the sample The sample for the sample blank measurement should be of similar shape and kind as the samples to be analyzed e g for thick and intermediate thickness samples prepared in the form of pellets a blank sample should be also a pellet It should simulate the sample matrix without presence of any elements exhibiting fluorescence lines besides the ones originating from impurities Usually such blank sample is made of pure binder e g cellulose boric acid wax or flux if the sample
41. sample chamber the unit of the thickness um is um input 12 The next line Length of the air path takes into account the absorption of the characteristic radiation by air on the way from the sample to the detector input 0 5 Three different detectors predefined so that dead layer and gold layer are taken into account are for selection The Si PIN detector description is the same as for the Si Li detector no Au layer is used for the first one Type exactly Si Ge or Si PIN input Si The thickness of the detector crystal input 3 is important for the higher energy intrinsic efficiency whereas the beryllium window thickness in um is of influence to the lower energetic radiation input 25 118 Ti 2 of the source 1 lt T1 2 lt 10 9 days or O for X ray tube Excitation energy of the et radiation 3 lt Eo lt 166 gt keV1 Energy of coherently EPEE a T AR lt 3 lt ECoh lt 166 gt keU1 Incidence angle sample surface Bepi beam 16 lt alpha lt gt degrees1 Exit angle sample surface ieee eee radiation 16 lt beta lt 6 gt degrees1 Mylar foil thickness between eer surface detector lt lt Mylar lt 166 gt um Length of the air path pn ad detector lt lt Air path lt 5 gt cm Detector kind si or Ge or i PIND i Thickness of the detector crystal 6 61 lt thickness lt 16 gt mm 3 Thickness of the detector berillium window lt i lt Be window lt 26
42. selected for standards not for samples Instrumental constants for the Full fundamental parameters METHOD 1 3 1 2 4 1 14 aa 0 9 KCO 0 8 0 7 x 10e 7 0 6 4 0 5 4 K2C0 KBr pure standards compounds e K compounds av Instr Const 1 165 0 0215 Ti Pb av Instr Const 1 133 0 0543 Ca Nb All standards average instrumental constant Ca Pb 0 4 1 149 0 0435 x 10e 7 0 3 0 2 0 1 10 20 30 40 50 60 70 80 atomic number Z Fig 3 28 Instrumental constants for all available calibration standards Rh secondary target spectrometer 50 kV Red dots refer to pure element standards black circles to elements contained in compound Potassium standards blue dots scatter depending on the chemical compound For pure metals and compound standards different sub group average values are found slightly deviating from the total average instrumental constant For all cases the elements Al Si P S and K were not included into the averaging An effort was made to gain further insight for the elements sulphur and potassium The individual instrument constants are plotted against the concentration of the respective element in the standard as potential influence parameter also the compound is specified In contrast to the Elemental sensitivities METHOD potassium in KBr and K2Cr20 7 also phosphorus in KH 2POg which is not included in the graph can be treated and us
43. standard sample information Calculations of Geometry constants Analysis of unknown samp Fig 3 14 An instrument parameter file can be defined or modified with the selection of Set up instrumental parameters 3 3 1 1 Example Test fpc With Select instrument parameter file the demonstration file Test fpc adequately defined for the used setup Sample changer spectrometer at Seibersorf laboratories can be loaded directory QXASdemo FullFP 61 Set up Instrument Parameter Select instrument parameter file Create new instrument parameter file Set up fundamental libraries Fig 3 15 An already existing instrument can be loaded with the selection of Select instrument parameter file In the FORM Excitation Conditions from all options for Mode the choice of Secondary target excitation will enable to get predefined values for the spectral distribution of the used secondary target The spectrometer was run with an evacuable sample chamber but the air gap between the sample chamber exit window and the detector Be entrance window can be incorporated by the definition of Atmosphere Air instead of the more logic option Vacuum The term Collimator refers to a SPECTRACE spectrometer lucky the owners of such an instrument for all others there is no relevance for any selection therefore the calibration continues with the selection of No collimator xcitation onditions Mode Secondary target excitation Air Atmosp
44. the demonstration sample spectra of the previous chapters particularly for alloys or the inclusion of scatter peaks for dark matrix representation it is evident that no rigid rules can be formulated But it is possible to give some advices The first step for the AXIL fit always must be the energy calibration By loading an input model with established energy calibration e g of a calibration standard one must already be able to identify elements For the demonstration spectra this does not apply because the gain was changed in between calibration and measurement of unknowns Next remove all previous element entries with the COMMAND X LINES REMOVE ALL set the ROI to AUTO 10 3 1 Example intermediate thick pellet pressed from 50 Soil 7 and 50 HWC as binder Soil7 I spe Fine tuning of the energy calibration For this spectrum channel number 295 as the maximum position of the highest low energy peak was used By the coarse energy calibration this element must be identified as calcium In the high energy end usually the elastic scatter peak will be used as second calibration point channel number 1602 is the Rh scatter peak maximum position Identify all possible elements with KLM MARK For a really unknown sample there is neither the need to find all of them already in the beginning nor will any over interpretation of realistic elements do any harm as long as the below named criteria are used Usually on
45. use of the average calibration constant for fluorescence For the purpose of the scatter calibration it is possible for this METHOD and highly advisable to use asr files of standards that do not contain any useful fluorescence peaks but only the scatter peaks information cellulose HWC etc or to fit only the scatter region and neglect the fluorescence region Soil 7 etc Otherwise insignificant peaks can contribute to the fluorescence calibration The asr file must contain 117 information of the form COH_SCAT and INC_SCAT At least one so called blank sample file must exist Usually this will be a spectrum acquired without any standard sample in the usual position of the sample holder instrument blank and it should represent the scatter contribution from the spectrometer and the ambient air If necessary blank problems for certain elements can be treated here but the basic idea is the determination the scatter blank As result of a calibration run a clb file will be created only then unknowns can be analyzed 7 3 1 1 Example Creation of Soil clb For demonstration a calibration file had been created containing twenty one calibration standards asr files for the establishment of the fluorescence calibration constant Si wafer K2CO3 K KH2PO4 potassium K2 Cr207 potassium K KBr potassium CaCO3 Ti Std TiO MnO2 Fe Std Fe203 Zn Std ZnO Br KBr bromine SrCO3 Y203 Zr Std ZrO2 Nb Std Nb205 and Pb
46. value for the pre defined Cd 109 excitation source as used by the Full fundamental parameters method The radiation travels between source sample and the detector beryllium window through air therefore the FORM Excitation Detection Geometry will expect an average path length for the two distances Dist Source sample Dist Sample detector The first of the two path lengths is approximated with 1 3 cm the distance between sample and detector entrance window with 2 4 cm Excitation Detection Geometry Dist source sample Ccm 1 3 Incident angle lt degree gt Dist sample detector cm Emergent angle lt Cdegree gt Fig 8 13 Definition of incidence and take off angle in respect to the sample surface normal and the air path Hint An important fact to treat correctly the FP METHOD is the use of the angles defined between sample surface normal and incoming and outgoing radiation Therefore the expected input will be 90 52 5 37 5 degrees for the Incident angle and 90 77 1 12 9 degrees as Emergent angle For the secondary target spectrometer and other Cartesian geometry set ups the differentiation between the angles 90 45 45 for Full fundamental parameters and the other METHODs is only academic for source excitation systems it is important The file Cd 109 fpc was created with the use of twenty one standards asr files KH2PO4 for potassium K2 Cr207 for potassium CaCO3 Ti Cr Cr2 K207 for chromium Mn
47. were included in the calibration file The acquisition time was set to 1000 s and a tube current of 38 mA was used throughout all the calibration measurements Usually the tube current setting is used to fulfill certain dead time criteria imposed by the spectrometer As for none of the data collections the maximum permissible dead time limit was passed the maximum current permissible for this machine could be utilized throughout Data for standard D 11_NOU QKASDEMONTXRFEN 1G6PPB ASR Sample ID 166 Live Time lt sec gt 1600 006 Tube Current mA gt SSBB OOOO Date lt mm dd yyyy 12 16 2666 Fig 9 8 Usually the acquisition time and the date are read from the respective asr file used for calibration The concentration values have to be entered in the second FORM named Data for standard for the elements Mn Cu Sr and the internal standard Y Iron and zinc values included only for the sake of the fit will be left at their defaults 0 0 standard D 11_NOU QRASDEMO TRRF 1G6G6PPB ASR of analysed elements units ppb counts compound conc Mn 166 6666 Fe 6 660088 Cu 166 6666 Zn 6 666666 Sr 166 6666 CERCEI t T E 1123076 Fig 9 9 The concentration values as known from the standard s preparation must be entered Elements only included for the sake of the AXIL fit will be left with their concentration at the default 0 0 After the successful inclusion of all standards editing of any of th
48. willingly with a measurement before and after The instability continued in the course of time 1500 0 position channel number of T w 2 the Zr Ka peak maximum 1400 0 1300 0 channel number max of Zr Ka 1200 00 ooo ooo Fig 10 4 Drift of the energy calibration controlled by the peak maximum position of Zr K The observed energy drift represents a parameter that is out of control but has no direct impact on quantitative results 10 2 How to fit standards spectra according to QA 10 2 1 Before the fit with AXIL Optional Copy the converted spectra spe into a directory with a directory name with less than 8 characters for the examples as used in this book the directory name SPE stds had been found adequate Edit if not already done at an earlier stage the spe files so that the header SPEC_ID will contain the sample preparation number taken from the sample preparation log book Include also the spectrum file name from the spectrometer log book in case the name was changed like for the example standards Helpful but not mandatory will be some information about the standard itself Optional Create a model file inp already before the fit with AXIL in order to define the matrix Sample absorption Excitation conditions Detector characteristics Path length etc This information for most of the standards will be of no relevance except if line rati
49. 0 2000 00 00 00 MEAS TIM 25 1000 PEAKS it 33 d 10 532 51818 For Help press F1 Fig 4 1 WordPad representation of the file As2O3 asr Text in red refers to the changes made to the original file Example For spectra acquired with an ORTEC MCA a problem was encountered that is revealed for source excitation when the so called decay correction is to be calculated The date format in the spectrum header is differently defined as expected by QXAS For the example Cr spe directory QXASdemo Cd 109 SPE INP measured with Cd 109 excitation source the modification should be SSPEC_ID SSPEC_ID Cr Std 12 7 06 cd0002 spe Cr Std 12 7 06 cd0002 spe SSPEC_REM SSPEC_REM DET 1 DET 1 DETDESC NAALXRF GUEST1A MCB 9 DETDESC NAALXRF GUEST1A MCB 9 AP Maestro Version 6 01 AP Maestro Version 6 01 SDATE_MEA SDATE_MEA 077 1272006 16 33 50 07 12 2006 16 33 50 SMEAS_TIM SMEAS_TIM 100 102 100 102 SDATA SDATA 0 2047 0 2047 4 2 Editing of calibration files cal The step Polynomial fit of sensitivities is not recommended prohibited according to QA requirements Nevertheless a calibration file cal directory QXASdemo ElSens can been processed with Polynomial fit of sensitivities and the relevant calibration data for Rb 7 01E 004 4 36E 003 as an example must be noted 75 e a 5 17E 004 5 81E 004 6 36E 004 7 01E 004 7 69E 004 8 83E 004 7 Fees 3 22E 863 3 62E 003 3 96E 00
50. 0 8 5 0 7 5 0 6 0 5 4 0 4 4 intrinsic efficiency 0 3 4 0 2 4 0 1 5 0 5 10 15 20 25 30 35 40 45 50 energy keV Fig 2 5 Intrinsic calculated efficiency as to be expected for a Si Li detector The discontinuity at 1 8 keV is caused by the Si dead layer 2 3 Elastic and inelastic Scattering The spectral background results from a variety of processes For photon excitation the main contribution is the incoherently and coherently scattered primary radiation by sample sample holder ambient air etc and therefore depends on the shape of the usually poorly described excitation spectrum and on the later to be determined sample composition To the low 44 energy side of a dominant peak the low energy tailing often increases the scatter background significantly The most straightforward method to obtain the net peak area under a peak of interest consists of interpolating the background under the peak by a suited function as also done in AXIL In another frequently used approach the discrete de convolution of a spectrum with a so called top hat filter suppresses the lower frequency component i e the slowly varying background A severe distortion of the peaks is introduced But applying this filter to both the unknown spectrum and well defined experimentally obtained reference spectra will result in the net peak areas of interest A disadvantage of this method is that reference and unknown spectra should be acqu
51. 0 ppb to 200 ppb in steps of 20 ppb and yttrium as internal standard element with fixed concentration of 500 ppb for all standards had been prepared The concentration ranges were selected in order to find the finally obtained results for unknown samples within the spanned calibration range As this cannot be known before usually one would prepare only one calibration solution and try to get preliminary results for the unknown sample The combination of the elements Mn Cu Sr had been selected with the intention to avoid peak overlaps that would introduce through the peak de convolution an additional source of uncertainty The element Y had been chosen as internal standard because the inspection of the spectrum of the later used sample does not contain this element In case a sample of interest contains elements with overlapping peaks other sets of calibration standard solutions covering the other elements needed to be prepared Single element mother solutions of Mn Cu Sr and Y with 1000 ppm concentration as used for AAS or ICP were firstly diluted to 10 ppm solutions in order to use pipette volumes for the final calibration standards high enough to avoid unnecessary big relative errors as found for volumes of less than 10ul Plastic vials with a volume of 10 ml served as containers In order to obtain 10 ml as final volume with a concentration of 10 ppm from a 1000 ppm mother solution 100 ul 0 1 ml were pipetted into the vial and 9 9 ml of t
52. 04 3 621E 002 ame 39 1 8 661E 004 4 251E 002 38 1 40 1 9 217E 004 890E 002 39 1 41 055E 005 5 384E 002 40 1 152E 005 5 617E 002 41 1 874E 004 1 691E 002 42 1 82 2 346E 003 120E 003 805E 003 640E 003 617E 003 164E 004 217E 004 749E 004 061E 004 599E 004 946E 004 561E 004 1 744E 002 414E 004 2 205E 002 745E 004 2 141E 002 211E 004 3 336E 002 554E 004 5 785E 001 010E 004 4 360E 003 864E 004 3 621E 002 661E 004 4 251E 002 217E 004 4 890E 002 055E 005 5 384E 002 152E 005 5 617E 002 874E 004 1 691E 002 922E 002 286E 002 130E 001 760E 001 403E 001 676E 001 200E 001 758E 001 265E 002 196E 002 506E 002 POOP PWOHOWNNEFERR EP ODNWUOWN WwW UD NFPrPoO aI OTOP BEWNYHONNFPRFPRFOGHWNE Rubidium results for unknown samples might have a higher uncertainty but all other elements represented by measured calibration points are well defined 4 3 Removing element entries from asr files For the sake of an appropriate fit sometimes elements have to be included but for final results as obtained by quantitative METHODs should be suppressed Within a defined ROI all peaks must be defined in order to enable an optimized fit One example to suppress an element from a report is the use of energy values instead of an element symbol e g X LINES ADD 2 7 76 2 8 2 9 instead of RH L Alternatively all necessary elements are included for the fit but the asr file can
53. 1108 asr directory QXASdemo NBS the reference standard material NBS 1108 can serve as example The equivalent FORMs as for the calibration for the METHOD already used have to be passed For Select Calculation Mode this time the default Sample concentrations is appropriate All elements in the sample Mn Fe Ni Cu Zn Sn Pb are represented by calibration standards included for the average instrument constant no low Z elements the default for the Type of Instrument Constant Average can be used The FORM Information on sample will have to be edited for the tube current value of 10 mA and the Sample type must be toggled to Thick sample otherwise one would have to enter the area related mass The date is of no relevance for tube excitation and all other fields of this FORM can be left unchanged The next FORM Menu for setting up options will need attention The default of 1 Sample contains may not be left at its default otherwise the presence of a dark matrix will be expected and together with the data available to the program it will crash So toggle to no matrix The default of 6 Secondary enhancement is not appropriate it must be changed to is corrected It will be seen from the results that it is 70 important to correct for the secondary excitation As rule of thumb the secondary excitation correction can be selected for every sample there is no known disadvantage The default Normalization of concentrations is applied i e the s
54. 14 2 Br 7 3 0 10 Rb 51 47 56 Sr 108 103 114 Y 21 15 27 Zr 185 180 201 Nb 12 7 0 17 Pb 60 55 71 Th 8 2 6 5 8 7 U 2 6 2 2 3 3 low Z Na 0 24 Mg 1 13 Al 4 70 S 0 12 dark matrix H 0 89 C 7 18 O 47 36 The warning message No elements analysed in this sample is together with the correct definition of the matrix no problem Fig 7 8 Warning message displayed for asr files handled with this method not containing any fluorescence peak information When the so far not seen FORM Information on scatter peaks appears in the regular case it will display the relevant information taken from the asr file It is possible to edit this FORM but this is not advisable Only for testing purposes e g when the effect of slight variations in the scatter peak areas should be examined it is recommended All other problems should be solved either by changes in the definition of the fpc file or different choice for the sample description etc In case the values for Coherent and Incoherent scatter peak area are displayed as 0 0 the selected asr file was not set up for the scatter 112 peaks For the example of Soil 7 the displayed information matches the values given in table 7 1 Information on Scatter Peaks Measure time Sec gt LONA 660 Tube current m 46 66666 Characteristic line Rh Ka coh Coherent scatter peak area 492611 6 Std 819 66 Characteristic line Rh Ka inc Incoherent scatter peak ar
55. 2P04 CaCO3 Ti Std Ti02 MnO2 Fe Std Fe203 Cu Std CuO Zn Std ZnO As203 SrCO3 Y203 Zr Std ZrO2 Nb Std Nb2O5 and Pb LStd All elements except Rb identified in the demonstration sample can be represented by elements contained in at least one calibration standard The Rb sensitivity value of 7 15 0 241 10 had been obtained Polynomial fit of sensitivities from the optimum fit with a linear polynome of order 4 weighted option After the polynomial fit the calibration was refitted with order 0 no fit in order to use the original calibration points data Finally the Rb value was pasted into this calibration file 85 Current calibration D 11_NOU QKASDEMO ET SOIL CAL Select calibration file Change Measuring Parameters Select Files for Analysis Perform Calculation of Concentrations Define File to Save Analysis Results alibration Data Filename SOIL CAL Created 69 28 2006 Excit SecTarget R kU 56 Hr of St 15 Elements K lines Si Nb L Lines Pb Pb Arrows move selector box lt CR gt select item lt ESC gt exit Fig 5 3 After the selection of the calibration file and definition of measuring parameters three relevant asr files must be loaded After having selected the correct files and entered the sample mass all proposed target elements V Co Cu Se Sr and Mo will be selected Current calibration D 11_NOU QKASDEMO ET SOIL CAL Select calibration file Ch
56. 3 4 36E 003 4 79E 003 5 50E 003 i ee 5 21E 66 5 64E 804 7 86E 004 8 66E 804 TE TUB 3 34E 002 2 66E 002 3 62E 002 4 25E 002 l I I I I I I 50E 004 5 91E 863 9 22E 804 4 89E 882 Fig 4 2 Interpolation of Rb obtained with Polynomial fit of sensitivities Next the calibration has to be re run with order 0 for the polynome Order of polynomial 0 no fit 0 Consequently the calibration points are not fitted at all and the original calibration points are not biased Finally outside of the QXAS environment the relevant data for Rb is pasted into the now unfitted calibration file and the total number of available sensitivities CAL_SENS is increased from 22 to 23 SEXP_COND SEXP_COND 0 0 09 28 2006 09 28 2006 0 0 00 SecTarget Rh SecTarget Rh 45 0 45 0 45 0 45 0 20 600 20 600 50 00 0 00 50 00 0 00 SSTAND SSTAND 24 24 1 4142 1 4142 1 4142 1 4142 etc etc SCAL_SENS SCAL_SENS 22 23 16 1 1 346E 003 5 922E 002 16 J 19 1 2 120E 003 3 286E 002 19 1 20 805E 003 2 130E 001 20 1 22 640E 003 3 760E 001 22 1 23 617E 003 5 403E 001 23 1 24 164E 004 676E 001 24 25 1 1 217E 004 200E 001 25 1 26 1 1 749E 004 8 758E 001 26 J 27 1 2 061E 004 1 265E 002 27 28 1 2 599E 004 1 196E 002 28 29 1 2 946E 004 1 506E 002 29 i 30 E 004 1 744E 002 30 1 32 414E 004 2 205E 002 32 33 745E 004 2 141E 002 333 1 34 211E 004 3 336E 002 34 1 35 1 5 554E 004 5 785E 001 35 1 38 1 7 864E 0
57. 3 From this comparison one can learn that the targets should be numbered even front and back side may behave differently and care must be taken that the same target is used for the target and target plus sample measurements 10 6 Calibration according to QA norms For each element under consideration in principle for each element found in an unknown sample an adequate number of very well defined standards with a range of concentration values bracketing the sample concentration must be prepared Let us assume a linear calibration function can be defined for each element although it were possible to define the following equations also for a non linear function A linear calibration function for the element x established with N standards with concentrations c where n is the running index for the various concentration values can be written as I b c a 10 1 In order to work with a linear calibration function the corrected intensity has to be calculated which is built up from the net peak area N of standard element St x and is corrected in terms of self absorption A enhancement effect E tube current i alternative decay correction T and live time t St x ae jer Ni G 10 2 154 The slope b measure of sensitivity is defined as Sher 8 02 T b 10 3 peng ery Using the mean of the measured intensities r 10 A N and the mean of the standard concentrations Fely 10 5 N
58. 3 0 For very intense peaks peak area equal or larger than 10 counts the corresponding Chi square value can be larger than 3 All the net Peak areas should be positive and greater than 3 times the standard deviation otherwise remove the respective peak from the fitting model Conclusion _As a basic rule one initiates the first fit with coupled lines and no entry E g start with V for vanadium not with V KA V KB The escape peaks should be added for intense peaks already from the beginning The starting value for the order of the background model for samples PARAME the parameter for the linear exponential orthogonal polynomial background should be small say between 0 exponential 1 and 5 1 1 2 Example Input model Pb inp and spectrum Pb pure spe for a pure lead calibration standard As an example for a quantification calibration standard pure lead was selected The input model file Pb inp does not only among others carry information about the energy calibration etc but also the absorption correction is defined Under Specify spectrum analysis parameters the Sample absorption had been defined Specify spectrum analysis parameters Background parameters Calibration parameters Fitting control parameters Sample absorption Fig 1 15 Access to more parameters of an input model file Set sample absorption Sample thickness g cm 2 gt 106 988 Sample composition element amount element amoun
59. 3 2 85 2 85 Zn ug g 104 101 113 106 103 As ug g 13 4 12 5 14 2 10 9 10 5 Rb ug g 51 47 56 52 31 Sr ug g 108 103 114 120 117 Y ug g 21 15 27 23 23 Zr ug g 185 180 201 219 215 Nb ug g 12 7 0 17 11 11 Pb ug g 60 55 71 62 60 slope 2 543 2 877 regression c 0 997 0 991 std deviation 0 076 0 142 5 1 5 Example Lichen 336 Demonstration files directory QXASdemo ET Calibration file Lichen cal AXIL result files samples LichenO asr Lichenl asr Lichen1B Lichen 2 asr Lichen2B asr Lichen3 asr Lichen4 asr Lichen5 asr AXIL result files targets TargetF asr TargetA asr TargetlB asr TargetB asr TargetC asr TargetD asr TargetE asr AXIL result files target plus sample Li0O_T_F asr Lil_T_A asr LilB_T1B asr Li2_T_B asr Li2B_T1B asr Li3_T_C asr Li4_T_D asr Li5_T_E asr Demonstration spectra directory QXASdemo ET SPE 90 Sample spectra Lichen0 spe Lichenl spe Lichen1B spe Lichen2 spe Lichen2B spe Lichen3 spe Lichen4 spe Lichen5 spe Target spectra TargetF spe TargetA spe TargetlB spe TargetB spe TargetC spe TargetD spe TargetE spe Target plus sample spectra Li0_T_F spe Lil_T_A spe LilB_T1B spe Li2_T_B spe Li2B_T1B spe Li3_T_C spe Li4_T_D spe Li5_T_E spe Input files Lichen inp Lichen_T inp Lich_T1B inp Li4 amp 5_T inp directory QXASdemo ET INP The reference material IAEA 336 Lichen had been selected to demonstrate the behaviour of this METHOD with the sam
60. 3 inp directory QXASdemo FP scatt INP the file HWC 3 asr directory QXASdemo FP scatt has to be manipulated in order to obtain the standardized asr file SSPEC_ID HWC binder 50 40 1000 vc2401 spe SDATE_MEA 00 00 2000 00 00 00 SMEAS_TIM 1000 1000 SCOH_SCAT 0 1615 1670 491253 491253 SINC_SCAT 0 1250 1610 3557881 3557881 SSPEC_ID HWC binder 50 40 1000 vc2401 spe SDATE_MEA 00 00 2000 00 00 00 SMEAS_TIM 1000 1000 SCOH_SCAT dl 45 1 20 167 491253 INC_SCAT 1 45 1 19 401 3557881 80 CHAPTER 5 EMISSION TRANSMISSION METHOD References 54 27 28 The EMISSION TRANSMISSION ET METHOD is restricted to intermediate thick samples in terms of XRF in practical terms they have to be less than 1 mm thin This technique corrects adequately the absorption caused by those elements which are not recognized by characteristic lines in the spectrum the dark matrix A carefully selected irradiator the so called target its characteristic lines should not interfere with the ones of the sample is positioned at the backside of the unknown sample and is excited by the excitation source In a second measurement the sample is removed and the irradiator is kept at the same position Finally the unknown sample is excited in a third measurement without the irradiator As a consequence three asr files are needed in order to analyze unknown samples with an unspecified matrix The intensity ratios of the charact
61. 30 140 150 160 170 180 scattering angle in degree Fig 2 6 Differential scattering cross sections for a thin scatterer of carbon Z 6 and an incident energy of 17 4 keV as a function of the scattering angle Prior to the computation of an element s concentration for the calculation of the absorption as well as the enhancement correction the concentration of all elements in a sample should be known This vicious circle can be overcome by a suitable iterative procedure Still there are the elements forming the dark matrix also their composition needs to be known The fundamental parameters METHODs use the incoherent coherent scattering peak ratio for establishing the average atomic number of a sample with dark matrix in order to calculate the absorption correction and when not defined by an input also the sample thickness The count rates Neon Ninc of these two peaks are respectively given by dO oni Noon Goon ECE AcE NE Xic 5 E OZ 2 65 and dO pc N nc Gne aE Anl E I E Ve Fo E84 2 66 In analogy to the formalism for fluorescence respectively for elastic and inelastic scattering is G the instrument constant is the detector efficiency A is the absorption correction is the primary intensity of the line with energy E which is scattered The summation over i includes do f all elements with concentration c The differential cross section jo U Z is a function of the energy
62. 5 18 75 20 5 20 75 21 22 5 ZR KB The scatter region has to be split off because the background is described for the first spectrum part as linear polynomial of order 10 for the second part also as linear of order 15 Any such high parameter would subtract the background from the scatter peaks in an uncontrollable manner The first saved result was expanded with the second coming result with the COMMANDs SAVE_RES and ADD For the scatter peaks it is advantageous to use UPDATE instead of ADD when the spectrum had been fitted before for the scatter peak region Otherwise a second block with scatter peak information will be stored only the first block will be used by AXIL In doubt inspect the final file LakeSed3 asr The definition of the sample is achieved with Specify standard sample information and selection of the default Sample concentration for Select calculation mode The Type of Instrument Constant is to be selected as Individual and or average which will correct for the lighter elements the systematic deviation from the average instrumental fluorescence constant The sample thickness one usually will leave for unknowns at the default intermediate thick but not specify the actual value because the program will calculate the thickness itself For true intermediate thick samples as used also for the ET METHOD the thickness is better specified explicitly by the known value In the FORM Menu for setting up options the only alteration s
63. 6 gt um 25 Fig 7 13 In the form Enter calibration data the excitation source the geometry the detector and the air path must be specified After the successful completion of the FORM the so called blank sample file name input of an asr file name but the extension ASR must be omitted has to be entered Instr Bl It contains the fit results of the scatter peaks originating from scattering of the excitation radiation by the empty sample holder instrument blank and if applicable the ambient air Enter the x ray tube current refers to this blank measurement The question Save results in the file report will enable to save the individual calibration results to a report file name rpt which later also can be used to store final results of sample analysis in readable ASCII format but the further calculations are not affected by this decision The actual calibration file will have the extension clb and is created at the end of the calibration procedure 7 3 1 2 Example pellet of KxCO3 plus HWC as binder K2CO3 asr as fluorescence calibration standard The next FORMs refer to the first standard No of sample components lt 1 50 gt and expresses the need of the program for an input of the total number of chemical compounds constituting the first standard the input must be an integer between 1 and 50 For the K2CO3 calibration standard pressed as pellet with added binder the input will be 2 The area related mass aerial de
64. AEA Vienna 1997 FUNK W DAMANN V DONNEVERT G Quality assurance in analytical chemistry VHC Verlagsgesellschaft 1995 52 53 54 MACDOUGALL D CRUMMETT W B et al Guidelines for data acquisition and data quality evaluation in environmental chemistry Anal Chem 52 1980 2242 MARQUARDT D W An algorithm for least square estimation of nonlinear parameters SIAM J Appl Math 11 1963 431 MARKOWICZ A A VAN GRIEKEN R E Quantification in XRF analysis of intermediate thickness samples Handbook of X ray Spectrometry 2nd edn revised and expanded VAN GRIEKEN R E MARKOWICZ A A Eds Marcel Dekker 2002 161
65. AS instrumental constant amp is the detector efficiency see later and fo fy correct for absorption of radiation between excitation source sample and sample detector air protective foils The summation over n takes into account that the excitation radiation spectrum may not only be composed of a single energy but also contains a continuum Qx is the product of the so called fundamental parameters see 2 28 2 35 A is the absorption correction factor and the enhancement correction term H a rather lengthy expression taking into account the secondary excitation It corrects for the increase of intensity of the characteristic peak of element x through secondary excitation by characteristic radiation originating from the same other X ray lines of that element and or other elements 1 H TEE ee 2 48 m The term Y is rather complex especially for intermediate thick samples For a thick sample for the thickness classification in terms of XRF see the following paragraph it can be expressed as 5 Ee nts Hy Eo aud in aE 2 49 L Ey Lt E sin o Un E L E sin y Note The METHOD Full fundamental parameters uses a different notation for the incident and take off angles They are defined in respect to the sample surface normal opposite to all other METHODs of QXAS where they are defined with respect to the sample surface itself 2 1 3 Classification of samples according to their thickness In X
66. CO3 K2 Cr207 for potassium and K KBr for potassium For Al 2 70 10 Si 5 65 10 and P 4 32 10 the constants could also have been generated by the program directly but for sulphur two standards are available S Std 7 78 10 and MgsS04 6 70 10 8 for potassium even four standards K KH2PO4 9 67 10 K2CO3 8 26 10 K2 Cr207 9 37 10 and K KBr 6 999 10 For sulphur the average constant 7 24 10 for potassium the average constant 8 575 10 will be used The entries to be incorporated have to follow strictly the syntax rules for the fpc files 77 SINDIV_INSTRU_CONSTS SEXCIT_MODE SINDIV_INSTRU_CONSTS 2 70E 08 1 2 2 70E 08 65E 08 SEXCIT_DATA 4 65E 08 32E 08 47 4 00 3 000e 002 LS 325 08 24E 08 50 00 50 6 24E 08 57E 08 45 45 00 45 00 74 19 sD IE 08 SEXCIT_SPECT SEXCIT_MODE 2753E 010 20 216 L 6932E 009 20 074 EXCIT_DATA 4509E 009 22 724 47 4 00 3 000e 002 2978E 009 22 699 50 00 50 1308E 008 23 173 45 45 00 45 00 74 7610E 007 2 916 SEXCIT_SPECT 0322E 007 2 892 2753E 010 20 8506E 006 3 364 6932E 009 20 074 4509E 009 22 724 etc L 2578E 009 22 699 1308E 008 23 173 7610E 007 2 916 0322E 007 2 892 8506E 006 3 364 etc 4 5 File format of asr files as expected by the NBS METHOD The original AXIL result file of the copper calibration file Cu Std asr directory QXASdemo ASR Stds as expected by other quantitative METHODs is not adequate for the NBS alpha coeff
67. E of the line that is scattered and the atomic number Z of the scattering atom Separating the contribution of the high atomic number fluorescent elements from the dark matrix elements the scattering coefficients of the latter can be calculated from 46 dO ohi SEZ Sc Cont N ooh Na eh 2 67 iwz dQ ME Gon ECE Ang E rignz coh and dO N Sua ye inci _ inc c O nei 2 68 iwz dQ NE Gue ElEne An E nignz where the concentrations in the summation over the high Z elements are assumed to be known from the fluorescence intensities All values on the right hand side can be accessed either by measurement or theory As a result the dark matrix can be represented by two elements e g Z and Z 2 where Z is calculated in such a way that lowZ d aO onz42 gt S oh AO onz 2 69 d S lowZ 2 d i O nez 2 inc O ine Z and the concentrations of these two representative elements are calculated by solving the system of the two equations in two unknowns cz and cz42 do do Si ey e o e 2 70 and do do a o e 271 These two concentrations can now be used for the calculations of the absorption correction factors and enhancement effect in order to obtain correct results for concentrations of the fluorescing elements 47 CHAPTER 3 CALIBRATION ISSUES Reference material 25 26 3 1 Calibration standards instrumental requirements Demonstration files the example calibration standards spect
68. HOD Backscatter FP 9537 12 06 2815 3 7 39 8 522 METHOD Certified values 8740 11 11 2610 5 6 38 8 0 47 C I 95 7910 9570 10 72 11 5 2300 2920 4 8 6 4 36 9 40 7 0 45 0 49 115 The comparison of the two sets of results as obtained by the Full fundamental parameters and the Backscatter fundamental parameters METHOD exhibits for both a problem with the element strontium for this sample Bromine is underestimated by both by approximately a factor of two but its low intensity only demonstrates that working close to the detection limit is generally problematic another fit model could have resulted in e g an overestimation For potassium calcium and titanium there is no other statement than the Full FP METHOD tends to the lower confidence interval limit and Backscatter fundamental parameters to the upper limit Rubidium is better described by Backscatter fundamental parameters 7 2 3 Artificial scattering calibration standards As can be seen from Figure 7 4 it might be difficult to impossible to find a suited scattering calibration standard E g the intermediate thick pellet of Soil 7 prepared with binder has no near neighbours except MgSOx The composition of this sample is well known it consists of half binder HWC half standard reference material with the elements content above 0 1 of H 1 9 C 48 7 N 1 4 O 25 3 Mg 0 6 Al 2 3 Si 9 1 K 0 6 Ca 8 1 Ti 0 15 and Fe 1 3 The res
69. I K CA KA CA KB TI MN FE KA FE KB CU ZN RB SR Y ZR KA NB KA PB AS 83 Note Description of sample SOIL 7 used to acquire Soil7 1 spe and the input model details to generate Soil7 1 asr can be found in the EXCEL soil7_description xls file Due to the dominance of Ca and Fe both element entries were split into Ka and Kg and the escape peaks and peak shape corrections were included Furthermore for the sake of Si the Rh L scatter region had been defined X LINES ADD SUM 2 7 2 8 2 9 As will be seen from the later obtained results silicon needed not to be included therefore also not the Rh L scatter region but for the time being one is interested is the behaviour of this METHOD to low Z elements There are blank problems with the element copper it resulted in a count rate of 0 1 counts second definitely not 10 times higher than the Cu instrument blank value Not included were elements like Cr because the Fe Kg escape peak will potentially mislead its quantification Ni because of the proximity to the strong Fe Kg signal Kr because the peak at 12 6 keV is rather Pb Lg The AXIL fit result was saved under Soil7 1 asr Spectrum SOIL 1 SPE Iteration 3 ChiSquare 1 4 Dif 3104 Soil binder 0 1055g dilution factor 1 99 40mA vuc215 Display a BEG He Wy EE A beg chan ia ERII ad MEE FETT END ANY E ted Ah 1 LLIA yf by end chan it i MIN min cnts MAX max cnts ROI SPECTR RESIDUAL LIN LOG R e s i d u
70. LStd One standard for the scatter constants pellet of IAEA Soil 7 only evaluated for the scatter peaks Soil7 2 asr The first FORM of this METHOD Select will lead to the calibration 1 Perform calibration by typing 1 Any other character will not be accepted and 2 brings to the later to be passed quantification of unknown samples The FORM Enter calibration data and all other FORMS of this program will have to be filled in correctly e g numbers must be entered that are between certain limits any lt signs will rather mean that the actual limits are included into the range and are also applicable When capital small letters are expected as usually specified one line above the expected input the operator will have to obey otherwise the program will respond with an error message and the input will have to be repeated correctly For tube excitation the first line of the FORM will have to be passed with 0 The excitation energy input 20 6 keV can be different from the coherent scatter peak energy because the weighted average for K excitation is higher than the coherent scatter peak energy Energy of the coherently scattered radiation of the K line Ee sct Eka input 20 167 keV Incident and Exit angle input for both 45 refer to the sample surface beam geometry lines 4 5 A Mylar foil between detector and sample can be used to protect the detector window and or represent the exit window of an evacuable
71. MAND KLM MARK It will start by default with the markers at the energies of the iron lines for iron the Fe Kg Fe Kg and the rarely visible L lines are proposed With arrow keys left right and up down jumps of Z 10 potential peaks can be checked for their presence qualitative analysis The lines for the elements to be include can be defined with the COMMANDs X LINES then ADD and type in V CO CU SE SR MO And terminate the COMMAND sequence with either GO or lt enter gt In between each entry of an element a space must be left Immediately after each entry for an element new markers will be superimposed on the spectrum for each line of the element The information gained with the COMMAND KLM MARK can not be transferred other than manually to the COMMAND X LINES The background can be defined e g with BACKGRND LINEAR PARAM and an input of 5 after PARAM Which means a linear polynomial as function of the channel number energy of order 5 is used to approximate the background within the region of interest Finally the fit of the region of interest in order to strip the background and resolve peak overlaps will be initiated with the COMMANDs FIT N_ITER 10 The adoption of the number 10 is fairly arbitrary A number of 10 20 will usually be sufficient in order to enable the iteration procedure with enough loops After the AXIL fit some information about the quality of the fit can be gained by first looking at the valu
72. MMAND REPORT then DUMP and GO An ASCII file dmp is created channel energy measured fitted fitted number spectrum spectrum background T sl al wl BWWWWNHNNHEEE ON RPoORRPY AT ONDDNOUOCADHAO A N H ol eD A A A OrRN Now Ds 5 De 5 5a De 5y DS 5 5 5 Os 5 B 5 Dy Sx 6 6 E O O O O O O O 0 2 O O OGO OGO OO OO OO a PES Eo o E a ES a S O S a POODOVDVDDVDDVDVDVDAMOOWUUOUMOMMO OO WO OO O TO OOGO CE G eO A MO MOO OO OO amp O OO 22 The data in the file are arranged in columns There are five columns and as many lines as the spectrum has channels The first column contains the channel number The second column the energy corresponding to that channel columns 3 4 and 5 contain the measured spectrum the fitted spectrum and the fitted background respectively EXCEL can read this file type and one can prepare detailed graph of the spectrum selecting the data from the second and or the third column A spectrum graph prepared in EXCEL is shown in Figure 1 20 As mentioned direct printing by using the COMMANDs like PRINT or PLOT may fail However in many QXAS routines there is also an option available to save the results to a file depending on the METHOD as ARP PRN RPT file Such file can be then retrieved by a word processor program for eventual editing and printing 1 2 25 What is the setup ax file What can be done when it gets corrupted and how to recognize it The set
73. NBS 1108 In the displayed results contained in the column Enhancement the correction factor for the respective element can be found This can be interpreted for e g Mn that its concentration would be overestimated by factor of 2 6 when the correction is not applied Except Zn and Pb they only enhance the other elements all elements are influenced by the others in respect to the secondary excitation substantially Mn Fe Sn or at least not negligible Ni Cu 71 Table 3 3 Comparison with certified values in for both experimentally obtained results average and individual instrumental constants the concentration values are normalized Mn Fe Ni Cu Zn Sn Pb measured average instrumental constant 0 030 0 067 0 043 65 51 33 80 0 47 0 075 measured individual instrumental constants 0 032 0 066 0 042 65 66 33 65 0 47 0 075 0 025 0 05 0 033 64 95 34 42 0 39 0 063 0 001 0 001 0 001 0 05 0 05 0 01 0 001 certified value 3 332 Definition and use of individual instrumental constants In case one prefers individual instrumental constants also for elements otherwise well covered by the average instrumental constant the individual instrumental constants for the elements Mn Fe Ni Cu Zn Sn and Pb must be defined Use the instrument parameter file NBSalloy fpc or to work without predefinitions test fpc but
74. O2 Fe Co Ni Cu Zn Ge Zr Nb Mo and L line standards Hf Ta W Au and Pb all files from directory QXASdemo Cd 109 For the establishment of the average instrumental constant for fluorescence 38 38 0 849 the two potassium standards in analogy to the secondary target excitation the one for Mn this compound is problematic as was already found out and the L line standards the intra element effect K lines of an element excite also its L lines which is not considered by the FP METHOD were not included 133 Instrumental constants for the Full FP METHOD Cd 109 excitation 45 aaua 40 fe 35 Pb 30 1 MnO 25 i average instrumental constant Ca Mo excl Mn 38 38 0 849 20 J 0 1 T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T T 10 20 30 40 50 60 70 80 atomic number Z Fig 8 14 Instrumental constants for all available calibration standards Cd 109 source excitation Red dots refer to standards used to establish the average instrumental constant for fluorescence The elements K Mn and the L line emitting standards were not included into the averaging For potassium an individual instrumental constant of 36 05 was defined which is the average value from standards KH2PO4 asr and K2 Cr207 asr For Pb a value of 36 81 was defined 134 CHAPTER 9 TOTAL REFLECTION X RAY FLUORESCENCE TXRF References 41 42 43 44 45 46 47 48 49 9 1 F
75. O3 with 5 binder HWC as C3gH76N202 added Target A has a similar composition with slightly different concentrations but Sr is missing All concentration values had been selected to have comparable heights for the target element peaks with a representative sample intermediate thick Soil 7 pellet in front After careful mixing of all components the six pellets were pressed with 10 tons weight 82 Table 5 1 Concentration values for elements in synthetic targets B F and acquistion time and tube current for the accumulative target Targets B F 3114 3118 V V205 Co CoO Cu CuO Se SeO2 Sr SrCO3 Mo M003 concentration 42 58 7 47 5 31 1 35 0 28 0 32 pure metals Ti Cr Co Cu Ge Mo live time s tube 400 40 50 40 30 10 25 5 10 5 4 5 current mA 5 1 2 Example Accumulative counting of pure targets As alternative approach to the preparation of a suited compound target the target and target plus sample measurements can be replaced by pure metal targets accumulative counting For the example spectrum TMetal spe as first target element the Ti calibration standard was measured for 400 s with 40 mA tube current The Ti was replaced after the first measurement by Cr but its spectrum was counted onto the Ti spectrum for extra 50 s with 40 mA After Cr was replaced by Co the spectrum was acquired again for 30 s with 10 mA This was continued in the same manner for Cu 25 s 5 mA Ge 10 s 5 mA and Mo 4 s 5 mA
76. O3 Y203 Zr Std ZrO2 and Pb LStd All elements except Rb identified in the demonstration sample can be represented by elements contained in at least one calibration standard The Rb sensitivity value of 6 51 0 246 10 had been obtained Polynomial fit of sensitivities from the fit with a linear polynomial of order 5 weighted option After the polynomial fit the calibration was refitted with order O no fit in order to use the original calibration points data Finally the Rb value was pasted into this calibration file Note Description of the samples prepared from the reference standard material IAEA Lichen 336 with different weights and the input model details to generate the asr files can be found in the EXCEL lichen_description x s file All lichen spectra had been fitted with the same model file with the inclusion of the elements X LINES ADD K CA KA CA KB TI MN FE KA FE KB ZN BR RB SR Y KA ZR KA PB Chlorine although of interest bears the inherent risk of under or overestimation due to the proximity to the Rh L scatter region that only could be described poorly Molybdenum only observed for Lichen0 spe and copper will originate from the sample chamber blank problems Copper is included in order to make the blank problems for this element obvious Mo for spectrum Lichen0O spe because this element is contained in the target and it needs to be included for intensity correction Tungsten lines as observed originate fr
77. Prepare alpha coefficients file Quanitative analysis of unknown samples Fig 6 3 The asr files needed for calibration must be converted to std files With the knowledge about the concentration values of the calibration standards at hand the following FORM is easily passed There is no possibility to add any confidence interval information therefore the standard deviation values of the edited asr files also were ignored previously 98 El amp line Conncen formula NBS1163 16mA vc2107 spe Live Time 1666 6 Number of peaks 7 Fig 6 4 The concentration values are attached to the NBS1103 asr file information for the creation of the NBS1103 std file Due to the need for consistency elements concentration values might be questioned which are not present at all in certain standards Leave these concentration values at their default 0 When for an element neither a small peak in the asr file was added as recommended nor its concentration value is different from 0 expect a warning message later which still would not prevent to evaluate samples Warning Mn missing or int amp conc zero Press any key to continue Fig 6 5 Warning message displayed when an element entry neither was specified in the asr file with a small intensity value nor a concentration value was entered during the generation process of the std file For each calibration standard intended to be used later one ha
78. QUANTITATIVE X RAY ANALYSIS SYSTEM USER S MANUAL AND GUIDE TO X RAY FLUORESCENCE TECHNIQUE FOREWORD This guide covers trimmed and re arranged version 3 6 of the Quantitative X ray Analysis System QXAS software package that includes the most frequently used methods of quantitative analysis QXAS is a comprehensive quantitative analysis package that has been developed by the IAEA through research and technical contracts Additional development has also been carried out in the IAEA Laboratories in Seibersdorf where QXAS was extensively tested New in this version of the manual are the descriptions of the Voigt profile peak fitting the backscatter fundamental parameters and emission transmission methods of chemical composition analysis an expanded chapter on the X ray fluorescence physics and completely revised and increased number of practical examples of utilization of the QXAS software package The analytical data accompanying this manual were collected in the IAEA Seibersdorf Laboratories in the years 2006 2007 Any queries comments or suggestions regarding this guide and the QXAS software package may be directed to XRF Group IAEA Laboratories Seibersdorf A 2444 Seibersdorf Austria E mail official mail iaea org Fax 43 1 2600 28222 EDITORIAL NOTE The IAEA is not responsible or liable for the accuracy of analytical results produced using QXAS Such accuracy depends among other factors on sample preparation issues cal
79. RF analysis samples are classified according to their thickness or mass loads For any particular sample the intensity of X ray fluorescence peak given by formula 2 47 depends on the sample thickness or the sample mass load m F Formula 2 47 can be simplified for two extreme cases 1 thin samples and 2 thick samples For the so called thin sample its mass load is considered infinitely thin m F lt lt 1 2 50 If 2 50 holds m m exp a l a 2 51 Pf a z E 2 51 It can be proved that if 2 50 holds the enhancement correction term can also be neglected H Cys lt lt 1 2 52 C dark matrix Taking into account 2 51 and 2 52 the formula for the intensity of characteristic peak emitted by a thin sample takes the following form 39 1 Gef DU EMhO d e Z Se 2 53 For a thin sample the intensity of X ray characteristic peak is proportional to the element mass load c m F expressed in g cm The intensity does not depend on the mass loads of other elements present in the sample The proportionality constant is the sensitivity factor Sy firstly introduced in the formula 2 44 For a given X ray peak and given measuring conditions the sensitivity factor does not depend on the analyzed sample Its value can be established by measuring thin standard samples with known mass loads of elements The definition of a thin sample given in 2 50 is not very practical or specific A more specif
80. ROI SPECTR RES IDUAL LIN LOG R e s 1 d u a 1 15608 2000 GO Channel Number CANCEL gt gt DISPLAY RESIDUAL _ Fig 1 21 AXIL fit results for the spectrum KBr spe which shows a diffraction peak at 9 75 keV Other artefacts e Fluorescence lines of other origin than the sample itself A typical example are the Zr K peaks identified in the hafnium standard Hf LStd spe with Hf inp Zirconium is a contamination of the Hf standard in a strict sense they originate from the sample Tungsten L lines can originate from the mill or dye used for the sample preparation of pressed pellets 17 Au L lines can originate from the gold contact of a Si Li detector Of importance are elements found in the sample blank and the instrument blank e g Fe Cu Zn Pb Spectrum HF LSTD SPE 166 26mA vcz2092 spe contamination rk a gt atnti y atn i T atnt16 atn 16 Cc o u m t S gt C h a m n e 1 GO CANCEL 1500 26606 Channel Number Fig 1 22 Spectrum of a hafnium standard Hf LStd spe showing zirconium as contamination f Sometimes M lines can be identified while L lines of an element of interest are used for quantification Typically Pb L is accompanied by M lines which have energies in range of 2 3 2 4 keV This can lead to misinterpretations because of the proximity to sulphur K For the example Hf LStd spe the Hf M lines are found around channels 150 160 1 6 1 7 keV Unfortunatel
81. X a2 X X 88 497 1 463 0 943 9 097 Y a0 al X a2 X X 88 454 1 497 0 942 9 106 expected values 87 6 1 8 1 4 9 2 Intuitively one would prefer to use different calibration functions for each element But have in mind that the as expected concentration values are not certified Of course it is not permitted to combine element wise results from different calibration functions It also not permitted to construct average values and standard deviations from the four calibration functions This could be achieved by measurements of several at least two individually prepared samples and their evaluation by QXAS 104 CHAPTER 7 UTILIZATION OF THE SCATTER PEAKS References 26 27 32 33 34 35 28 36 7 1 Fitting of scatter peaks example Rh secondary target excitation Demonstration files Spectra directory QXASdemo Fp Scatt SPE HWC spe Cellulos spe Liche336 Cabba359 spe Soil7 1 spe LakeSed3 spe Soil7 2 spe Instr Bl spe Spectra directory QXASdemo SPE Stds Al Std Si wafer MgSO4 P KH2P04 K KH2P04 K2CO3 Ti02 K2Cr207 S Std Ti Std K KBr Br KBr Input files for the Rh K scatter region directory QXASdemo FP Scatt INP HWC l inp HWC 3inp Cell l inp Lic336 1 Cab359 1 inp S7 1 1l inp SL3 1 inp S7 2 l inp Ins Bl 1 inp Si 1l inp Al l inp MgSO4 1 inp P KH2P 1 inp K KH2P 1 inp K2CO3 1 inp Ti02 1 inp K2Cr2O0 1 inp S Std 1 inp Ti Std 1 inp K KBr 1 inp Br KBr 1 inp AXIL re
82. XRF Only monochromatic excitation can be defined quasi mono energetic excitation One draw back of the METHOD neither varying tube current nor decay corrections are treated For tube excitation standards and samples must be measured with constant current alternatively asr files have to be edited In the case of source excitation the decay correction must be calculated by the user and the asr files have to be corrected edited manually Samples are classified as Element system Oxide system Fused disk system The term element system refers to samples where the analytes are present as unbound chemical elements all of them produce a detectable peak in the spectrum Examples are stainless steel or brass alloy where low Z components or impurities can be neglected For oxide sample systems the analytes are present as oxides with known stochiometry which is usually the case for minerals and geological materials In fused disk system samples had been diluted with a suitable chemical flux and then melted together at high temperatures 1200 C producing a glass bead The sample classification must apply for all standards and samples E g a MnO calibration standard cannot be incorporated for an Element sample e g alloy calibration although maybe needed for the quantification of samples Select the type of specimen composition Element sample e g alloy Oxide system e g ore Fused disc sample lt lithyum compou
83. a 1 15008 2000 GO Channel Number CANCEL gt gt DISPLAY RESIDUAL _ Fig 5 1 AXIL fit of Soil7 1 spe The spectra of the target T1 spe and target plus sample S7 1_T1 spe were measured with a tube current of 40 mA The spectrum for the sample Soil7 1 spe was acquired for 1000 s live time whereas the target and target plus sample spectra were acquired for 200 s respectively Note Description of the target used in combination with the Soil 7 sample and the input model details to generate Tl asr and S7 1_Tl asr can be found in the EXCEL target1_description xls file The target was fitted with split Ka and Kg lines and peak shape corrections input model files Tl inp S7 1_T1 inp 84 X LINES ADD V KA V KB CO KA CO KB CU KA CU KB SE KA SE KB SUM For Sr and Mo no peak shape correction exists they were included with X LINES ADD SR KA SR KB Mo KA The fit results were saved under Tl asr and S7 1_T1 asr Spectrum TARGET1 SPE Iteration 4 ChiSquare Targeti i V Co Cu Se Sr Mo vuc2176 spe Axil LOAD STOP DISPLAY ROI CALIB X LINES KLM MARK FIT REPORT SAVE_RES PLOT BATCH BACKGRND SCAT_ROI Cc o u n t 4 C h a n n e 1 1500 2000 Channel Number G0 CANCEL Fig 5 2 AXIL fit of T1 spe A calibration file Soil7 cal suited for the sample was created with the METHOD Elemental sensitivities Twenty one calibration standards asr files were included Si wafer K2CO3 K KH
84. a certain class of samples For many users this is a bit unrealistic one will usually try to find a compromise First of all one has to get a chemical compound containing a fluorescing element with rather low atomic number which is not present in the sample The lower its characteristic line energy the better with the constraint that the line is not completely absorbed when passing through the sample Usually potassium and calcium are present in samples sulphur and chlorine lines will usually suffer already too much from absorption Potential candidates are Sc Ti V or Cr The concentration of this compound must dominate the target composition All other target elements should cover uniformly the sample spectrum in terms of energy preferably with decreasing concentration with increasing atomic number They should not be present in the sample at least not dominant like e g Fe in reference standard material Soil 7 In principle the METHOD accounts for these coincidences the sample contribution is subtracted from the sample plus target spectrum but a higher uncertainty for such a target element can be related to this fact 5 1 1 Example Targets B F The approximate composition of five targets B C D E F with respective spectra TargetB spe TargetF spe target T1 is identical to TargetB but different spectra were recorded is V 42 V205 Co 7 CoO Cu 5 CuO Se 1 SeO2 Sr 0 3 SrCO3 Mo 0 3 Mo
85. absorption edge E gt Ex The contribution of the edge to the photoelectric mass attenuation coefficient of photons at the energy E can be calculated by using the tabulated overall value of the photoelectric mass attenuation coefficient at that energy 7 and the absorption jump coefficient of the edge cX E t oh 2 16 Jk The emission of X ray fluorescence photon is not the only process leading to atom relaxation there are at least few other processes competing with each other In an assembly of ionized atoms each atom with a vacancy in a given electronic shell only a fraction will de excite with the emission of X ray characteristic photon This fraction is assigned a number called fluorescence yield of a given shell e g for K shell it is denoted by x One of the competing paths of relaxation is a radiationless process called Auger effect In this process the vacancy e g in K shell is filled in by an electron from less tightly bound electronic shells e g Li shell the atom energy is reduced but the energy excess Ex Ern is not emitted in the form of characteristic photon instead in a radiationless process it is transferred to another electron from e g Lm shell binding energy of which is less that Ex Ern This electron is expelled from the atom Its kinetic energy equals to Ex Ern Erm The fraction of atoms relaxing through the emission of Auger electrons is assigned a number called Auger yield of that shell e g
86. ality assurance every time an asr file is created or altered by saving the AXIL fit results SAVE_RES the results are also appended to the spectrum file This is sometimes problematic because even after an asr file is erased and created newly it can happen that the old fit result information is partly transferred to the newly created asr file Otherwise this feature is harmless but also not helpful In the worst case one will erase the block that represents the fit result from the spe file SSPEC_ID Soil7 2g 40mA vc2139 spe SDATE_MEA 00 00 2000 00 00 00 SMEAS_TIM 1000 1000 SCOH_SCAT 45 I 471788 SINC_SCAT 45 1 3 811436 SCOH_SCAT 492611 787234 21 1 2 23 Although a spectrum or file is definitely stored in a directory its file name will not be displayed in a SCROLL BOX SCROLL BOXes within QXAS can handle only 100 file names of a kind Although present in a directory any files exceeding this limit will not be selectable This is a reason to move files into sub directories 1 2 24 How can I get spectra into a format that can be used for decent looking printouts It often happens that the PLOT and PRINT COMMANDS of AXIL do not function as expected One way to overcome this problem is to use the WINDOWS build in lt Prt Scr gt key often in combination with lt Alt gt or lt Shift gt which will copy the QXAS window to the clipboard memory When this is not
87. ally it will not be necessary to include escape peaks scatter peaks or contamination elements e g there is no need to include the Sr contamination into the fit model for CaCO3 spe Examples For Fe Std spe and Fe203 spe define the elements with X LINES ADD FE KA FE KB For Pb LStd spe X LINES ADD PB L3 PB L2 PB L1 As an example for exceptions Rh L scatter peaks which are included by their approximate energies for e g CaCO3 spe X LINES ADD CA KA CA KB 2 7 2 8 2 9 Avoid stand alone peaks because the four calibration parameters two for the energy calibration two for the resolution will have too much freedom during the fit As example 148 although elements with higher atomic number could be fitted by their Kg peak only one still works with Ka and Kg As alternative approach for e g S Std spe where the two K lines are not resolved the fitting of these four parameters was restricted with D_FANO 0 00000 Use the automatic region of interest ROI AUTO Exceptions are low Z standards In order to enable a better defined background shape for e g S Std spe use the COMMAND ROI with Begin channel 80 End channel 350 As background model COMMAND BACKGER select either LINEAR or EXPON free of choice and begin with the lowest value PARAM which is 0 or 1 respectively Not recommended are the BREMSstrahlung background model meant for charged particle excitation the FILTER background did not work at al
88. alues are added to the model The background under the scatter peaks is subtracted in such a way that it fits the neighbouring continuum background BACKGRND LINEAR PARAM2 0 The proposed model has a ROI from the middle of the spectrum to the very end in order to minimize the subtraction of the background channels 1024 to 2048 The lines of interest are defined by X LINES ADD RH INC RH KA COH The tailing X LINES ADD 17 17 25 17 5 17 75 18 18 25 18 5 18 75 and the Kg scatter region are described by X LINES ADD 21 75 INC RH KB COH 20 5 20 75 21 22 5 The Kg Compton peak would not be positioned correctly therefore it was added as an energy value Sometimes Zr Kg is also added when the peak is interfering with the ROI All added lines were included with the goal to stabilize the fit of the incoherent and coherent Ka peaks Spectrum SOIL 2 SPE Iteration 8 ChiSquare 87 2 Dif Soil 2g 40mA vc2139 spe Axil LOAD STOP DISPLAY ROI CALIB X L INES KLM MARK FIT REPORT SAVE_RES PLOT BATCH BACKGRND SCAT_ROI Cc o u n t s Ta C h a n n e 1 1500 2000 Channel Number GO CANCEL Fig 7 1 AXIL fit of the Rh K scatter region according to model 1 Looking at the residual or the Chi square value can be frustrating but is of no relevance for the results In the resulting asr file only the incoherent and coherent Ka peak information will be stored All other peak areas can be inspected with REPORT FULL GO or e
89. aneously a set of markers corresponding to the X ray lines of that element is shown in spectrum display If the calibration was established correctly the markers should match the X ray peak positions One can first check the matching for the peaks used for calibrating the spectrum In the example spectrum E calib spe Ti K peaks Mo K peaks and Rh Kg as well as Rh L peaks should match with the corresponding set of markers The Rh Kg peaks are out of the range for E cal spe All Rh lines originate from the Rh secondary target They are detected after being scattered by the sample The ill defined hump around channel 1790 is the Rh Kg Compton peak incoherently or inelastically scattered radiation 1 2 7 How to fill the spectrum with characteristic lines correctly The usual procedure after establishing a correct energy calibration will be a careful check with KLM MARK The default markers will start to propose Fe The first action will be to Jump with arrow down to the lowest line in terms of energy The first element as inherently defined that could be identified is sodium Na 13 1 2 7 1 Example IAEA reference standard spectrum Soil7 1 spe Before loading the spectrum Soil7 1 spe load the input file S7 1l inp The first element present in the sample is silicon Then from left to right the elements will be identified The elements with low atomic number Na Cl will have unresolved Kg and Kg peaks consequently they have on
90. ange Measuring Parameters Select Files for Analysis Perform Calculation of Concentrations Define File to Save Analysis Results rSelect Peaks for Abs Calc alibration Data A 23 U 195635 Filename SOIL CAL Created 69 28 2006 paetarser se 15 Elements K lines Si Nb L Lines Pb Pb Fig 5 4 Relevant target elements must be selected The absorption correction calculation does not depend on the selected calibration file All target elements are accepted although the parameter for V Z 23 Lg Lg A 1 094 is already very close to the threshold limit of 1 097 for rejection The slope 2 543 is not close to the expected value of 2 8 the regression coefficient regress 0 997 and the standard deviation stdev 0 076 for the fit are good The edge correction Enter element to start iter was applied from Fe on 86 Lg E Lg g a Fact 4 950 1 094 3 285 6 925 0 267 1 814 6 041 0 022 1 527 11 208 0 874 1 210 14 142 1 475 1 113 17 443 2 155 1 065 slope 2 543 intercept 5 212 regress 0 997 stdev 0 076 Accept ysn y Enter Element to start iter Fe Fig 5 5 The absorption correction factors are calculated at the energies of the accepted target elements Criteria to judge the calculation are the slope the regression coefficient and the standard deviation A discontinuity in the absorption behaviour of the sample caused by an element present in higher quantity can be taken into ac
91. ank spectra have a comparable shape one can assume the sample substrate does not contribute to the blank problems The comparison of the three asr files provides consequences Have in mind the previously defined rule that peak areas for results to be used must exceed 10 times the respective blank values Si Ca and Zn peak areas in comparison with the two blank spectra can be accepted Fe will have about 5 blank contribution Ni Cu Sr and Pb intensities approximately will be half generated by the sample but the other half is caused by the instrument therefore their results must not be reported The Mo peak does not originate at all from the sample 59 Table 3 1 Net peak area for elements identified in spectra of a standard reference filter material a blank foil and the instrument blank Elements in red colour can be used for quantification Si Ca Fe Ni Cu Zn Sr Mo IPb AirFilt asr 160 6 153 5 1910 14 574 3 1524 4 274 6 254 5 387 8 48 2 Instr Bl asr 22 3 0 112 4 37 3 94 4 15 2 200 5 381 7 18 2 Blamk AF asr 8 11 l 34 108 12 28 8 64 10 34 8 200 15 401 24 21 6 A calibration file suited for this sample was generated Airfilt cal It carries the calibration points as needed for the four elements of interest defined by six calibration standards Si wafer asr CaCO3 asr Fe Std asr Fe203 asr Zn Std asr and ZnO asr no extensi
92. ation trace elements should be omitted At least three entries have to be marked For a too high absorption for a target element which is more likely for elements with lower atomic number it will not be used for the calculation Ie when the ratio 7 590 equivalent to Indn 7 78 7 gt 1 097 For the other elements a linear regression of the form In In a versus In attenuation coefficient a energy E is displayed Fact is the sample self absorption correction factor Both logE and Lg E stand for the natural logarithm When the regression coefficient regress is smaller than 0 99 usually one or more elements will be removed from the list A higher standard deviation stdev value is usually only an indicator of what can be already seen in the plot the target element points scatter around the fitted line The expectation value for the slope is around 2 8 Rather cryptic Enter Element to start iter the chemical symbol of an element can be entered for which an edge correction should be considered The edge correction refers to the sample and makes sense for elements with concentration values expected to be greater than 1 such a correction will be applied only after a first run without any edge correction It applies to the selected element and to all others in the sample with lower atomic number 5 1 How to compose the target In the ideal case one would mix a target according to the sample spectrum of
93. atom of interest as well as on the path length from the fluorescing atom through the sample in the direction of the detector The attenuation of the incoming radiation is characterized by the corresponding path length and the weighted sample mass attenuation coefficient for exciting photon energy Eo The attenuation of the characteristic radiation in the direction of the detector is characterised by the corresponding path length and the weighted sample attenuation coefficient for the energy of characteristic radiation e g Exo or Era 37 excitation to detector absorption by all other atoms along the path Fig 2 2 Both the exciting radiation and the fluorescence radiation of an atom element of interest suffer from sample self absorption For a monochromatic excitation theory provides a summation integration over all possible locations for flat and homogenous samples with the thickness f as parameter I Cy i Sy A t Cy C C Jarkmatrix 2 44 where Sx is the sensitivity factor for the element x The term sensitivity is used in the context of e g the Elemental sensitivities and Emission Transmission METHODs A is the absorption correction factor which is a function of the concentrations of all elements cx Ci Cdark matrix present in a sample I expf a2 _lexptra p t _ ME 2 45 A t Cys Cis C darkmatriz a a using the expression a which depends on the incidence and take off y angles and the weigh
94. ause no scatter information is found in the respective asr files answer must be n otherwise a further spectrum bust be loaded After the successful completion of the first standard the question Next sample Y N will permit the definition one standard after the 66 99 other input y in the same way as for the first standard 7 3 1 3 Example pellet of IAEA Soil 7 Soil7 2 asr as scatter calibration standard The input needed to describe the scatter standard is for No of sample components lt 1 50 gt 12 From the certificate the essential elements concentration values gt 0 1 are represented by Na 0 24 Mg 1 13 Al 4 70 Si 18 0 S 0 12 K 1 21 Ca 16 3 Ti 0 3 Fe 2 57 the dark matrix is described by H 0 89 C 7 18 and O 47 36 in accordance with reference 27 Tube current input 40 file name input Soil7 2 The successful termination of the calibration is achieved with n 120 Sample no 2 No of sample components lt 1 5 gt 12 Enter the sample mass per unit arealg cm2 166 Compound element amp i weight fraction wt i CaCoO3 Si 18 Compound element amp i weight fraction wt 4 i CaCO3 K 1 21 Compound e lement Ca 16 3 Compound e lement Ti 3 Compound e lement Fe 2 57 Compound e lement Na 24 Compound e lement Mg 1 13 weight fraction wt Z i CaCO3 weight fraction wt i CaCO03 weight fraction wt i CaCO03 w
95. be edited E g the air filter spectrum AirFilt spe was fitted with all elements as identified by peaks but due to the problem with the blank spectrum several elements entries were removed from the original AirFilt asr file directory QXASdemo EISens in order to suppress the mistaken use of results for elements that should not be quoted S SPEC_ID Ref Std SRM2783 267 40mA vc2406 spe SDATE_MEA 00 00 2000 00 00 00 SMEAS_TIM 1000 1000 SPEAKS 9 14 20 26 28 29 30 38 42 82 NwWDOOUDPWHUD Consequently the edited version AirF cor asr will only enable to get quantitative results for the elements Si Ca Fe and Zn Edit also the total number of fluorescence net peak areas PEAKS from 9 to 4 SSPEC_ID Ref Std SRM2783 267 40mA vc2406 spe SDATE_MEA 00 00 2000 00 00 00 SMEAS_ TIM 1000 1000 SPEAKS 4 4 Individual instrumental constants in instrument parameter files fpc Individual instrumental constants are used for the Full Fundamental Parameters METHOD at least for low Z elements The METHOD inherent approach to incorporate these constants will only allow for one standard per element A way to overcome this shortcoming is to edit instrument parameter files e g Soil fpc or OrgaMatr fpc The instrumental constants for the standards of relevance for this problem are obtained from the asr files Al Std Si wafer P KH2P04 for phosphorus K KH2PO04 for potassium S Std MgSO4 for sulphur K2
96. c file from the previous program No further changes in the fpc file are permissible when the constants are entered otherwise this would affect the calibration 113 7 2 2 Analysis of unknown samples containing dark matrix As demonstration sample the reference standard material IAEA SL 3 lake sediment is adequate because its scatter peak ratio comes closest to the Soil 7 The sample had been prepared as pressed pellet mass 2 0 g no binder added the spectrum LakeSed3 spe had been acquired for 1000 s with a tube current of 40 mA Note Description of reference standard IAEA Lake Sediment SL 3 and its spectrum acquisition conditions along with the names of the three input files necessary for an adequate AXIL fit resulting in LakeSed3 asr can be found in the EXCEL s13_dm xls file The spectrum had been fitted in three parts with input models SL3 A inp for the lower energy part of the spectrum for the element peaks included with X LINES ADD SI K CA KA CA KB TI MN FE KA FE KB 2 7 2 8 2 9 The higher energy part of the spectrum fitted with the input model SL3 B inp X LINES ADD ZN BR RB SR Y ZR KA NB KA PB had been split off because some of the individual Chi square values were too high with a single ROI for the fluorescence peaks Finally is the Rh K scatter region SL3 1 inp approximated with the previously introduced model 1 defined by X LINES ADD RH INC RH KA COH RH KB COH 21 75 INC 17 17 25 17 5 17 75 18 18 25 18
97. c file without defined coherent and incoherent instrumental constants because in such a case a default value of 47 the fluorescence constant will be used The program will not necessarily fail but an adequate scatter calibration is usually superior verage Instrumental Constant for fluorecence 1 149HE 7 for coherent scattering 4 6666 for incoherent scattering 6 6666 Fig 7 5 For the quantification of samples the instrumental constant for fluorescence must have a defined value The range of values accepted by the METHOD is unfortunately relatively narrow Not only the ratio of incoherent to coherent peak but also their absolute values can sometimes lead to a program crash for samples with dark matrix Usually one will employ several standards for fluorescence calibration representing the elements identified in the unknown sample s and one or more standard s for scatter calibration being close in its ratio incoherent to coherent scatter peak area to the unknown sample s Elements not represented by standards as well as elements for which no individual fluorescence instrumental constant had been defined will be calculated by use of the average calibration constant for fluorescence The calculation of scatter constants is selected from the TOGGLE FIELD Select Calculation mode with the selection of Instru Constants for fluorescence amp scatter For the purpose of the scatter calibration it is possible for this METHOD to use
98. cannot be matched with any KLM MARK markets funn naaa sete alte ac oocciaterns aeons ocmangal a SRE a E EN EEY 16 Ar and or Kr are identified in a spectrum sssesesssssseesseessesseresseeessresseesseesseresseee 18 How to define scatter peaks for the fit model eee eeeeeceereeceeeeeceeeeeeseeeeeeeeeees 18 What are the criteria for including or excluding a weak peak eeeeceeeseeeeeteeee 19 How to remove elements or lines from the list 0 cece eeeeeeeseecseeceeeeeeeeeeaeeeaeees 20 Why will one need an input model file inp for treating a spectrum with AXIL 20 What are the essentials of a model input inp file How is it created 0 20 Will the parameters stored in an inp file influence any of the quantitative METHODS Eo ete ete ie E en Bacal Rd ad 20 What is the FANO factor scnssiiinesninusisn eines si 21 Why does the ASR file information get into the respective SPE file 0 21 Although a spectrum or file is definitely stored in a directory its file name will not be displayed in a SCROLL BO Nice 5 aa cas peda aae ees tect ee an ndan Gove eed svesss waseaetees 22 How can I get spectra into a format that can be used for decent looking printouts onnios a a banca E E A a E ala sd iacauuacen eae E aE 22 What is the setup ax file What can be done when it gets corrupted and how to BOC ORTIZ IA ys an E aera Saati andes alae alae eaaahed dalle aed 23 1 2 26 Under what measuring conditions th
99. cify spectrum analysis parameters or Specify experimental parameters which make such parameters accessible for editing 1 2 19 What are the essentials of a model input inp file How is it created The model input file contains parameters for the energy calibration definition of the region of the fit the X ray lines defined with the COMMAND X LINES the background model the sample standard composition the detector resolution calibration the allowed variability ranges for the energy and peak resolution calibrations and several other data The file should be created saved immediately after a successful fit is accomplished For this purpose use the COMMAND STOP and select Save model action For creation of a new file use the COMMAND line In new file By selecting the option In current file one will update the file associated with the current model If this step is not performed all the alternations to the current model done in the AXIL session will be lost Choosing the option In existing file allows the operator to select other existing input model file and overwrite it content with the current model settings 1 2 20 Will the parameters stored in an inp file influence any of the quantitative METHODs No the parameters e g the detector specification the air path between the sample and the detector the average excitation energy etc are only used for peak ratio correction calculations They are not used by any of the quantitati
100. cixRFdown php The QXAS software package can be installed under the operating system WINDOWS but it is actually executed in the command prompt as a DOS program This fact carries the advantages of the WINDOWS inherent possibility to switch between programs without termination because of their simultaneous execution For the purpose of switching from QXAS DOS environment to WINDOWS environment and back one uses the combination of the lt ALT gt and lt TAB gt key By this action active windows are brought in front one after the other without termination Particularly to toggle between any of the QXAS programs and e g the WINDOWS Explorer will facilitate searches for files or enable to edit quickly files It is not recommended to open QXAS in two command prompt windows simultaneously and run them in parallel It is also possible to run QXAS software under other operating systems including 64 bit systems in a DOS emulator It is recommended to run QXAS in full screen mode as default For this purpose with the right mouse button the QXAS icon on the desktop should be selected From the drop down menu select Properties then in the option Screen enable the Full screen General Program Font Memory Screen Mise Compatibility Usage Full screen O Window Fig 1 1 Setting of the full screen mode for QXAS permanently Throughout the guide example spectra and files can be used in order to follow hands on
101. condary target and direct X ray tube excitation spectrometers will have linear influence on the spectrum emitted by the sample Any increase or decrease of the tube current will increase or decrease the fluorescence and scatter intensities strictly linear proportional of course with statistical deviations Given that all circuitry of the electronic chain from the preamplifier output to the MCA is correct the detector side of the spectrometer might not react accordingly because the signal processing is count rate dependent Between the two extremes of a very low count rate and the saturation of the detection system a non linear behaviour can exist the measured intensity will not follow the increase of the tube current with the same rate A dead time dependency can be specified In order to avoid dead time dependent misinterpretations one has to establish a series of measurements with the tube current as parameter In the praxis a calibration standard producing already a high count rate for the lowest current value will be measured repeatedly with increasing tube current as parameter from the lowest to the highest setting In this manner the iron calibration standard peak area was recorded between 5 and 50 mA in 5 mA steps 144 140000 120000 Linearity for the Fe K peak as a function of 48 9 the tube current 100000 80000 counts 60000 40000 6 3 R 0 9998 20000 It 0 5 10 15 20 25 30 35 40 45
102. cording to the scatter peak ratio e seeseseseeeeeeesreeresresserererrerseesee 108 7 2 FULL FUNDAMENTAL PARAMETERS MET THOD eeeeeeeeeeeesee 109 7 2 1 Example Establishing the instrumental constants for scattering for Soil7 fpc 111 7 2 2 Analysis of unknown samples containing dark matrix eile eeeeeereeeneeeeeeeees 114 7 2 3 Artificial scattering calibration standards 00 eee ceeeeeeeesneeceseceseeeeeeesaeessaeenseesees 116 7 3 METHOD FUNDAMENTAL PARAMETERS MONOCHROMATIC EXCITATION SCATTER PEAK So rea iir EA ase ea R EE 116 Tkl Calibration ea rae EE EE E acts a e AE E ovale 3s oes 117 do Lele Example Creation of SOU lbe cet etait ta eased Seek aa lates 118 7 3 1 2 Example pellet of KxCO3 plus HWC as binder K2CO3 asr as fluorescence calibration Stand ar xcs d0ccccssseasdceaiepaes bua roedhsaeedaapelaceselahcaees 119 7 3 1 3 Example pellet of IAEA Soil7 Soil7 2 asr as scatter calibration standardni erson a r aes aden a E eased TEE A ENE ESES 120 7 3 2 Analysis of unknown sample ssessenesesssessseeeesseessesseessesssseessseesseesseesseesseeessseesseese 122 CHAPTER 8 SOURCE EXCITATION sesssesesssessssessssesssessessressesssesrtssessssssessessesseeso 124 8 1 ESTABLISHMENT OF NUMERIC VALUES FOR INCIDENCE AND TAKE OFF ANGILE rested ena e a e a I 124 Sol Calibration standards i noes ee A Oe ee Eaa aS 124 8 1 2 Preliminary calibration with the METHOD Elemental Sensitivities 125 8
103. count The results for Si K and Ca have to be ignored because the METHOD can not calculate an adequate F factor for them The target consists of the fluorescing elements V Mo therefore silicon potassium and calcium are out of range for an appropriate absorption correction ANALYS IS REPORT Date 11 67 66 Sample 3164 S0il tbinder 1 055g dilution factor 1 99 46 mA vc215 Method Emission Transmission Mass Cg cm2 gt 0 0215 Element Energy counts Concentration Error F 1397 25 62 wx 1 99 45 7293 2384 1 19 wx 66736 13 71 wz 2470 1766 3 ppm 1869 1 ppm 105429 43 wz 165 ppm 1152 5 ppm 186 5 ppm 1 Js 3 4 5 6 8 8 O eRe NS NO A oO OO a 28 8 oe 8 Fig 5 6 Quantitative results for an intermediate thick pellet of the reference standard material Soil 7 mixed with HWC as binder A dilution factor of 1 99 applies Sample elements not within the range spanned by the target elements must be disregarded 87 5 1 4 Example Sample Soil 7 and target Tmetal Demonstration files directory QXASdemo ET Calibration file Soil7 cal Spectra Tmetal spe S7 1_Tme spe directory QX ASdemo ET SPE Input files Tmetal inp S7 1_Tme inp directory QXASdemo INP AXIL result files Soil7 1 asr Tmetal asr S7 1_Tme asr As alternative the accumulative spectra with the artificial accumulative target can be used for the establishment of the absorption correction The spectra of the ta
104. current mA Fig 10 1 Check of the linearity of the Fe Kg peak as a function of the tube current up to a dead time level of 48 9 Two facts in this respect are to be considered Small peaks can behave differently than the intense peaks causing the dead time Sometimes an energy dependence can be observed but not all peaks of a spectrum ranging from the lowest to highest energy must have the same behaviour For the used system the iron Kg net peak area followed the tube settings up to the maximum equivalent to almost 50 dead time When a non linear response is observed for a spectrometer all measurements must be taken within the linearity approximation range For source excitation such an easy going approach does not exist although by insertion of absorbers with varying thickness a comparative concept can be followed For such a study the transmission factors for the absorbers must be precisely defined 10 1 2 Detector resolution as function of time In principle the resolution of a detector has no influence on the quantification due to the principle of constant peak area both a detector with poor resolution and lower peak height as well as a good detector with a slimmer resolution and a more elevated peak maximum will result under otherwise identical conditions in the same peak area Still the FWHM is a major criterion for the quality of a detector The better the resolution the better overlapping peaks can be apportioned The detec
105. d 5 First element Last element Mo CL meeeemeLinear Weight of fit Yes Fig 3 6 A polynomial interpolation of sensitivities for elements not represented by calibration standards is possible For the elements between S and Mo the calibration compound cal was extended with Order of polynomial 5 as Type of fit Linear and Weight of fit Yes With this selection a Mean Diff of 6 2259 as to be found in the later displayed results was achieved Any other combination of fit parameters has worse results The weighted fit option could be used because all standards were treated properly for their uncertainties in the AXIL fit standard deviation greater than the square root of the peak area and realistic uncertainties of the 54 concentration values relative error 0 1 were entered In case this can not be assured Weight of fit No must be selected One will usually have to go through this Polynomial fit of sensitivities several times testing all possible combinations of parameters in order to get knowledge about the optimum Of major interest are the fit results of neighbouring calibration points of missing elements needed for unknown samples Calibration file E QXASDEMONELSENS COMPOUND CAL Created on 09 28 2006 Calibration date 11 08 2006 71 5e 003 u 1 0e 003 5 0e 002 40 Atomic number lt Esc gt continue lt F9 gt LOG Plot lt F10 gt Printer Plot Fig 3 7 Polynomial representation of sen
106. d take off angle in respect to the sample surface and S Diam is the diameter of the sample in cm For tube excitation there is the possibility to measure the unknown sample s spectrum with a different higher current Cur S than the target plus sample and target only spectra Cur T With Select Files for Analysis three SELECTOR BOXes open consecutively and an asr file has to be choose in this order sample asr file from first box Select Sample file ASR target asr file from second box Select Target file ASR sample plus target asr file from third box Select Sample Target file ASR There is no consistency check An input for the sample mass in g is needed Define File to Save Analysis Results enables to create an already previously created file for saving the results arp or to retrieve a new one 81 Warning Such a file must be created selected before the final calculation step otherwise the program will crash QXAS during the calculations for the next sample With an applicable calibration file loaded the measuring parameters correctly defined three appropriate asr files loaded and the mandatory definition of the result file the absorption correction calculation can be initiated with Perform calculation of Concentrations A list of proposed target elements together with their net peak areas pops up The user has to select appropriate elements for the absorption correction calcul
107. d to good results from the theoretical point of view it is not recommended to apply this METHOD to polychromatic excitation A handicap of the program is the missing lt ESCape gt command throughout the various lines that have to be entered with appropriate information Any single mistake in the input will force the user to terminate the program with lt BREAK gt some computers lt Ctrl gt plus lt BREAK gt lt Fn gt plus lt BREAK gt No re runs to add or remove standards from the calibration procedure are possible In such cases the calibration has to be re done 7 3 1 Calibration Generally it is recommended to start QXAS with the directory information of the directory where all calibration standards are contained in also the blank file and scatter calibration standards should be found in this directory for the demonstration QXASdemo ASR Stds As minimum one spectrum asr file together with the complete knowledge about its composition will serve for the calibration The more the standards cover the element range of unknown samples the better because individual calibration constants will be used Usually one will employ several standards for fluorescence calibration representing the elements identified in the unknown sample s and one or more standard s for scatter calibration being close in its ratio of incoherent to coherent scatter peak area to the unknown sample s Elements not represented by standards will be calculated by
108. dary target K Lines Ag anode 50kV SCONTINIUM 2 Ts pO 1 2 0 SCHARLIN 5 C1516 53 85 S60752 26 15 50555 00282 54610 5 30 253493 2 56 Example For the excitation with the Cd 109 source emitting silver K radiation treated as Ka and Kg and a y line of 88 keV the source description file Cd 109 sou directory QXASdemo Cd 109 is adequate but only for the geometry of the used spectrometer It contains the angles as found by the Monte Carlo approach implemented in the utility Incident and take off angles Geometry constants sen file 79 SIDENT Cd 109 annular source incident take off angle 75 0 77 0 degrees Cd 109 0 SGEOM 75 0 77 0 SCONTINIUM 2 1g is 1 2 0 SCHARLIN 3 5608 82 770 2961 17 230 1408 3 74 4 7 Editing of asr files to bring the scatter peak information obtained with the COMMAND SCAT_ROI to a format expected by quantitative METHODs Three different approaches for obtaining the scatter peak information are proposed CHAPTER 7 others exist The use of the AXIL COMMAND SCAT_ROI will not fit the scatter peaks but sum up the channel contents within the ROI Unfortunately saving of the so obtained peak areas is not in the format as expected by the quantitative METHODs When this approach is utilized it is advisable to edit the so obtained asr file For the example of the spectrum HWC spe directory QXASdemo FP scatt SPE treated with the input model HWC
109. data files as needed for calibration of the Rh secondary target spectrometer can be found in the EXCEL standards xls file 3 2 Calibration for the METHOD Elemental Sensitivities Demonstration files directory QXASdemo E ISens 48 Calibration files Compound cal Metals cal AirFilt cal Spectra AirFilt spe Blank Af spe Instr Bl spe Input files AirFilt inp AXIL result files AirFilt asr AirF cor asr Blank Af asr Instr Bl asr For the calibration of the Elemental sensitivities METHOD sensitivities are calculated for characteristic lines from standards correcting the self absorption in the standards It is assumed that the excitation radiation can be represented by a single X ray line the intensity weighted energy Average excitation energy E average Prai Exa t Par Eraz t Pep Exp Prpa Erp t Pros Erps 3 1 For a Rh secondary target the weighted average of its Ka and Kg lines yields Eaverage 20 6 keV For direct X ray tube excitation as primary radiation sometimes so called effective energies are defined Such a procedure might lead to good results from the theoretical point of view it is not recommended to apply this METHOD to polychromatic excitation The calibration for this METHOD is limited to 25 standards in total Preferably such elements should be covered by standards which are found in unknown samples The sensitivity S of all elements n is calculated for which a concentration cn is
110. defined for equation 6 1 which are used for the calculation of the coefficients for the later to be selected kind of polynomial CALCULATED RELATIVE INTENSITIES OF STANDARDS Fe Ni si Zn Ph Sn NBS1167 001308 56594 39517 00153 i 5 515 5 5 NBS1163 00393 00176 af K 35211 03188 z 5 5 5 5 5 NBS1168 OOE 00044 61205 Be ire y 00054 a 00042 NBS1115 00265 00079 86276 13169 OOOHA 0 51515 5 CU NBS 3 OOHOA 6 7 66666 OOAHA 86666 FE NBS 866866 5 86668 86686 66666 NI NBS 86668 gt 66606 86606 86606 PB NBS 86668 5 86606 86606 86606 SN NBS E 86668 gt 86606 86608 5 86666 ZN NBS 86666 OOOHO 1 66660 866686 86666 Fig 6 11 Relative intensity values for all calibration standards are displayed in the course of the evaluation of unknowns The SCROLL BOX CALCULATED PURE INTENSITIES FROM STANDARDS carries valuable information because a column for each element should contain in the ideal case only values slightly deviating from the average the standard deviation has to be calculated by the user This is the case for Cu relative standard deviation 3 and Zn relative standard deviation 2 no wonder because these elements are contained in high quantities in all NBS standards For Fe and Ni there is a systematic bias to the pure element standards This can be anticipated because they are only contained as minor elements in the NBS standards Fo
111. determination using the energy dispersive X ray fluorescence technique and the emission transmission method X ray spectrom 32 2003 317 ALVAREZ M MAZO GRAY V Determination of trace elements in organic specimens by energy dispersive X ray fluorescence using a fundamental parameters method X ray spectrom 20 1991 67 TAO G Y PELLA P A ROUSSEAU R M NBSGSC A FORTRAN program for quantitative X ray Analysis Using X Ray Tube Excitation NBS Technical Note No 1213 NIST Washington DC USA 1985 LACHANCE G R CLAISSE F Quantitative X ray Fluorescence Analysis Wiley 1995 PELLA P A FENG L Y SMALL J A An analytical algorithm for calculation of spectral distributions of X ray tubes for quantitative XRF analysis X ray Spectrom 14 1985 125 WEGRZYNEK D MARKOWICZ A CHINEA CANO E Application of the backscatter fundamental parameter method for in situ element determination using a portable energy dispersive X ray fluorescence spectrometer X ray spectrometry 32 2003 119 RACHETTI A WEGSCHEIDER W Background intensities and their utilization in quantitative analysis by monochromatically excited EDXRF Adv X Ray Anal 30 1987 143 RACHETTI A WEGSCHEIDER W A fundamental parameters approach including scattered radiation for mono energetically excited samples in energy dispersive X ray fluorescence spectrometry Anal Chim Acta 188 1986 37 NIELSON K K Matrix corrections fo
112. down and will read as COSECANT FOR PRIMARY X RAY 1 26096 129 COSECANT FOR SECONDARY X RAY 1 02579 8 1 4 Results for the incidence and take off angle and verification The cosecans of an angle is defined as sin of the angle Therefore we obtain as average effective angle of incidence 52 5 and for the take off angle 77 1 both of them in respect to the sample surface How can one know after all this lengthy procedure that the results are meaningful There is not a complete proof but some kind of verification is possible For consistency of the calibration typed in values for the geometry and detector etc the result file to be saved as sen and also found as temporary file temp dat when the COMMAND lt F10 gt Print is used after the evaluation run until the next run will provide a wealth of parameters like the AVERAGE GEOMETRICAL FACTOR and its VARIATION COEFFICIENT which can be compared with the individual values like the relative efficiencies etc LIST OF DATA AND RESULTS What angles DATE 11720706 START TIME 18 21 57 END TINE 18 23 15 SOURCE INSIDE RADIUS CM 656 SOURCE OUTSIDE RADIUS CN 1 266 SOURCE SAMPLE DISTANCE CN 1 250 SAMPLE DETECTOR DISTANCE CN 2 368 DETECTOR BelJ DISTANCE CN 566 DETECTOR RADIUS CCM 316 COLLIMATOR RADIUS CCM 476 COLLIMATOR HEIGHT CCM 766 COLL TOP DETEC DISTANCE CNM 1 668 SOURCE WINDOW THICKNESS CNM 00G Be WINDOW THICKNESS CCM 96256
113. e The energy calibration is needed for the proper identification of elements qualitative analysis whereas the term calibration quantification calibration relates intensities to concentrations quantitative analysis For an appropriate energy calibration some conditions should be fulfilled At least two peaks have to be identified in the spectrum A peak is identified if it is recognized as corresponding to particular X ray line or a group of lines The peaks utilized should have good statistical significance sufficient number of counts in the peak so that the maxima peak position can be established with good accuracy 10 The selected peaks should be more or less evenly ditributed through out the spectrum If only two peaks are used one should be located in the lower part of the spectrum and the other in the upper spectrum region Such selection of peaks ensures the energy calibration accuracy Consequently it is not recommended to use the Kg and the Kg peaks of one chemical element as the only two calibration points Note It is possible but in most of the cases hardly necessary to use more than two X ray peaks for the energy calibration The values of the energy calibration parameters are subsequently optimized during the fit One has to note that after performing the energy calibration procedure the spectrum abscissa is still displayed as Channel Number The validation of the energy calibration can be performed
114. e Lm shell The atom is still ionized its energy 3 E E E 2 4 The energy Erm is the binding energy of an electron in the Lm electronic shell The transition K Ly can occur spontaneously because the energy of the ionized atom is reduced in such a process E lt E 2 5 Consequently the atom gives out the energy There are several ways the energy can be given out One of the possible paths is by emission of X ray photon which is essentially called X ray fluorescence XRF For K lt Ly transition the emitted X ray line is denoted as Koy in so called Siegbahn notation or K L in IUPAC International Union of Pure and Applied Chemistry notation The emitted X ray photon has a characteristic energy E Ka 31 Ex E Ey Ex E 2 6 For the vacancy in the K shell the other most probable transitions include K lt Ly K Mi K lt Mn They are associated with emission of KO KB and KB3 X ray characteristic lines respectively New vacancies are created in this process They are successively filled in by electrons from outer electronic shells which results in cascade emission of L and M series X ray fluorescence lines The probability of a photoelectric effect is characterized by so called cross section In atomic scale the cross section Tis expressed in barns or in barns per atom 1 barn 10 cm 2 7 In a macroscopic scale more practical unit is in use the cross section is called photoe
115. e example standards and samples had been collected sesioen ees aeee niian n EE aneia AREER area ee een 23 1 2 27 A spectrum can not be fitted properly what to dO sssssssssesssesssssesssressessseessesesseee 24 1 2 28 How to handle blank problems seesesseseesesessesessessresresserssesressressesrssresseseresresseseess 25 1 3 BATCH MODE nai e a r a aA A OR E ate aes 27 1 3 1 Example Forty one spectra of a U XRF SCan eeeseseesseseresesrresressrsrresresrreeresressesere 28 CHAPTER 2 FUNDAMENTALS OF XRF THEORY 0 0c ceceeeeeeeeeeeeeceecneeeneeeeeeeees 31 2A FLUORESCENCE RADIATION W0 cc ccescosscssctsorssescnsesscoeesacesctsosscosesesaeens 31 Zale Sampleself absorptionsnes iia ean vatican ees 36 2 1 2 Ant r element Sets 2 22 2 cgstes ntessieestigeienieiaali eed uteden E KEAS i E 38 2 1 3 Classification of samples according to their thickness 00 0 0 eeceeseeeseeeseeeeeeeeeeeneeees 39 2 2 SKED DETECTOR Aen a a Re a as 42 2 3 ELASTIC AND INELASTIC SCATTERING seeeeeeeeeseseseesrsssersersrserse 44 CHAPTER 3 CALIBRATION ISSUES wisissivcessivecdisscpueceasvadestvpanteseasarennseectiononetoaeaens 48 3 1 CALIBRATION STANDARDS INSTRUMENTAL REQUIREMENTS erem a a a T N a 48 3 2 CALIBRATION FOR THE METHOD ELEMENTAL SENSITIVERE Si iihi a a ae a a e aoe aiaa aaae eae Paeetenaahoonaees 48 3 2 1 Example COnmpouna Cal 52 4521 e n ena a E N RR a E T E 50 32 2 RAMPS metals cal Kareen a E E a E E a E 57 3 2 3 App
116. e indicated range Results COMMANDs REPORT GO e Individual Chi square values less than 3 0 e For standards The standard deviation must be greater than the square root of the peak area e For samples The peak area must be positive and greater than 3 times the standard deviation Full report COMMANDs REPORT FULL GO e For samples The peak area must be greater than 3 times the background under the peak 1 2 12 What is the difference between X LINES ADD e g FE and FE KA FE KB What are the secrets behind and amp The command to include e g iron K series peaks into a fitting model is X LINES ADD FE The K series peaks entered this way have fixed tabulated ratio of Kg Kg If due to the absorption effects the observed line ratios do not agree with the tabulated ones a better fit is obtained if the ratio is not fixed during the fit This can be achieved by entering the iron peaks with the COMMANDs X LINES ADD FE KA FE KB before issuing this new command please eventually remove the iron peaks already entered in the model COMMANDs X LINES REMOVE FFE If the sample composition is know as for the standard samples a good fit can also be obtained using fixed line ratios providing that the sample composition has been defined at earlier stage For intense peaks the appearance of the escape peaks can be treated with the inclusion of suffix For Si Li detectors the escape peak energy is the parent pea
117. e of it is performed by any of the QXAS METHODs Rather average effective angles will have to be defined For this purpose use the utility Incident and take off angles Geometry constants sen file This utility is rather user unfriendly will need a lot of input but this is the only way to calculate these angles by QXAS Unfortunately during the establishment of values for the angles a calibration file will have to be loaded Such a cal file must be generated before with the METHOD Elemental sensitivities All what was mentioned to be essential for this METHOD has to be applied with slight modifications for a source excitation system 8 1 1 Calibration standards All example standards spectra for Cd 109 source excitation had been collected with the intention to have approximately 10 000 counts in the peaks of relevance for calibration The measuring time live time LT was selected to adjust the total counts in the peaks of interest The number of 10 000 counts for the lines of relevance should result in statistically well defined peaks relative standard deviation 1 A Cd 109 source emits silver radiation Ag Ka Ag Kg and a y line at 88 keV 124 Note Description of the calibration standards available for Cd 109 source excitation can be found in the EXCEL cd109_standards xls file 8 1 2 Preliminary calibration with the METHOD Elemental Sensitivities For the calibration of the Elemental sensitivities METHOD sensitivities
118. e of the ChiSquare 3 0 displayed only after the first fit In the ideal case this number should be 1 0 or close to it it is defined as reduced Chi square and is an overall information for the region of interest As next information source inspect the residual with the COMMANDs DISPLAY RESIDUAL This residual offers information about each spectrum channel within the ROI For a well defined fit model all points should be between 3 and 3 indicated by the two red bars Of course with certain probability random fluctuations of greater magnitude can occur spectrum T1 5PE Iteration 4 ChiSquare Target 1 U Co CursSe Sr 7Mo TrA Ad i pe Display F t BEG beg chan END end chan MIN min cnts MAX max cnts ROI SPECTR RESIDUAL LIN LOG R E s i d u a 1 1500 2666 GO Channel Number CANCE gt gt DISPLAY RESIDUAL Fig 1 12 Spectrum Targetl spe displayed after the first AXIL fit run The Chi square value of 3 0 as well as the displayed residual suggest that there is room for improvements remove the previous entries for the elements with the COMMANDs XLINES REMOVE ALL and add instead X LINES ADD V KA V KB CO KA CO KB CU KA CU KB X LINES ADD SE KA SE KB SR KA SR KB MO KA SUM This way of adding X ray lines forces that the K and Kg peaks are treated independently as if they originated from different elements It means that the ratio between Ky and Kg peaks of a given element is not fixed during the fi
119. e optimized parameters are improperly evaluated for some of the spectra The four parameters namely ZERO eV GAIN eV ch NOISE eV and FANO unitless parameter concern the energy the first two and resolution calibration They can vary within their respective limits D_ZERO eV D_GAIN eV ch D_NOISE eV D_FANO This variability is necessary to compensate for slight drifts in the energy calibration which may occur due to electronic or temperature related effects In certain situations especially when processing a bunch of spectra with poor counting statistics some of the parameters e g GAIN and FANO should be kept constant It helps stabilizing the fit and keeps the results within physical limits 30 CHAPTER 2 FUNDAMENTALS OF XRF THEORY Reference material 13 14 15 4 16 17 18 19 20 21 22 23 24 2 1 Fluorescence radiation Photo absorption of X rays with energy Eo is usually associated with an inner shell ionization of an atom A photoelectron is hereby emitted from the atom and carries the energy Ep Ep Here Eg is the binding energy ionization energy of the electron The energy of the ionized atom is increased by Eg Configuration of electrons in the ionized atom is not the optimum one in terms of energy Therefore in a very short time of the order of 10 s 10 s the arrangement of electrons in the electronic shells of the ionized atom starts to change until it achieves a minimum T
120. e respective peak area is followed by the information about its standard deviation st dev For the sake of statistical significance an element with a peak area lt 3 times the standard deviation or a negative area value should be excluded in the full report COMMANDs REPORT FULL GO underneath each individual peak area the under laying background backgr for this peak is displayed When the area of a relevant element peak is lt 3 times the square root of this background the element should be excluded These criteria are related to the definition of the detection limit and the fact that the peak area must be a positive number The definition assumes the presence of a peak if the peak area N is greater than three standard deviations of the background underlaying the peak BG N gt 3 VBG 1 2 From AXIL full report the background counts can be read for Co and Ni rel int peak area st dev chi sq fwhm eV backgr tot abs For the element cobalt the criterion is fulfilled 920 gt 3 510 2 for nickel 43 22 lt 3 525 a therefore Ni peaks should be removed from the fitting model After the model is modified the fitting should be repeated and the results examined again until no other insignificant peaks are present 1 2 17 How to remove elements or lines from the list To remove all X ray peaks apply COMMANDs X LINES REMOVE ALL To remove individual peaks or group of peaks replace the keyword ALL with t
121. e the detector at low temperatures contact to liquid nitrogen or a Peltier element The statistical spread is affected by the average energy required to produce an electron hole pair Usually the FWHM as one of the measures for the detector quality is given at the energy of Mn Ka 5 89 keV and typically ranges between 130 and 180 eV The preamplifier stage integrates each detector charge signal proportional to the energy of the incident photon to generate a voltage step proportional to the charge This is then amplified and shaped in a series of integrating and differentiating stages Owing to the finite pulse shaping time in the range of us the system will not accept any other incoming signals in the meanwhile dead time but extend its measuring time instead The dead time is defined as RT LT dead time RT 100 2 62 where RT real time refers to the time interval that could be measured with a stop watch during the acquisition of a spectrum LT live time is the time during which the system is ready to process incoming pulses Real time is always greater than live time In the next step the amplitude of the shaped and amplified pulse is digitized converted to a number in analog to digital ADC converter The content of a memory location corresponding proportional to that number called channel is increased by one The process is repeated for every valid pulse As a result a histogram of the puls amplitudes is
122. e will agree upon the elements Si K Ca Ti Mn Fe Zn Rb Sr Y Zr and Pb Chlorine is not a good choice because may it be present or not the Rh L scatter region will prevent good results due to its proximity to Cl K Adding of the identified peaks to the fit model with the COMMANDs X LINES ADD The intensive peaks in this spectrum calcium and iron also need an input for their escape peaks Possible sum peaks are to be included X LINES ADD SI K CA TI MN FE ZN RB SR Y PB SUM Do not split the elements into e g KA and KB for the first fit run of a sample Silicon is separated from the other elements by the Rh L scatter peaks therefore the energy values 2 7 2 8 2 9 will have to be included Arsenic is always problematic in the presence of lead therefore it will be included and carefully checked by its standard deviation and background counts Niobium might be present therefore it will be included The elements zirconium and niobium will be included with the Kg peaks only because Zr Kg and Nb Kg 150 carry no relevant information but suffer already from the high background caused by Rh K scatter For the used spectrometer copper is known as a problematic element in respect to the blank spectra The obtained peak area must be compared to the blank results as well as for the elements iron and strontium X LINES ADD 2 7 2 8 2 9 AS ZR KA NB KA CU The region of interest for fitting must include all elements of interest For thi
123. ea 787234 0 Std 1613 6 Atomic difference of Low Zs Fig 7 9 The form Information on Scatter peaks can be edited but scarce use of this option should be made After leaving the last FORM the calculation will be initiated with Calculation of Geometry constants Analysis of unknown samp le 11 16 2666 11 44 69 Sample identity Soil 2g 46 mA vc2139 spe Spectrum fitting data E QKASDEMO FP SCATT SOIL 2 ASR Instrument parameter data E QRASDEMO FP SCATTN SOIL FPC Instrumental identity Secondary Target The secondary target Rh The tube anode g Tube voltage 580 KU Tube current 46 666 mA Measuring time 1600 Sec Collimator No Collimator Filter used No Filter Atmosphere Air Report of Calculated Instrumental Constants Sample thickness infinitely thick Instru constant for coherent scatter 1 7883E 6 Instru constant for incoherent scatter 1 4103E 6 Concentration Instr constant Absorption Enhancement ll Fig 7 10 After a successful run the instrumental constants for coherent and incoherent scattering can be obtained The two instrumental constants for scattering must be manually transferred to the relevant fpc file Instrument Parameter File E QRASDEMO FP SCATT SOIL FPC verage Instrumental Constant for fluorecence 1 1490E 7 for coherent scattering 1 7883E 6 for incoherent scattering MRESEEST Fig 7 11 Also the instrumental constants for scattering are not transferred automatically to the fp
124. ed elastic or Rayleigh scatter peak either channel 1844 or 1845 originating from the Rh secondary target could have been used Not recommended are the incoherent Compton peak around 19 6 keV which for this spectrum is obscured by the stronger Mo Kg peak nor the Ti Kg peak channel 453 The Compton peak has neither a Gaussian peak shape nor has its maximum a tabulated energy value because its position is a function of the geometry scattering angle In difficult cases when only one element of a sample is clearly identified also the Kg peak can be used as second energy calibration point The obtained preliminary calibration can be used to identify other peaks in the spectrum The newly gained knowledge is then applied to establish a more precise energy calibration 12 1 2 3 In the spectrum E calib spe the maximum of Ti K is well defined in contrast to the ill defined Mo Kg A closer inspection of channels 1591 1600 of the above discussed spectrum will reveal the fact that there is no pronounced maximum of the Mo Kg peak within this region all channels have similar counts gt 400 The one better solution to this problem is to acquire Mo for longer counting times The more praxis oriented solution is not to adopt the channel with the highest counts value as peak maximum but rather the mean channel number which of course leaves some room for the experience of the user 1 2 4 With arrow up and down I can jump from one peak to t
125. ed for calibration No clear trend can be found In the ideal case all individual values would coincide with the total average instrumental constant 69 1 3 eae Instrumental constants for S in sulphur 1 14 nd K in jum com n KH PO and potassium compounds 0 9 K2Cr207 0 8 KBr kco pure 0 7 e sulphur 0 6 MgSO 0 5 0 4 4 0 3 0 2 0 14 x 10e 7 concentration in Fig 3 29 Plot of the average instrumental constants versus element concentration for selected calibration standards containing sulphur red dots and potassium blue dots 3 3 3 Application bronze alloy sample With the so far achieved calibration establishment of the average instrumental constant for fluorescence samples with either well specified matrix or no matrix at all can be treated The scatter calibration as needed for samples containing a dark matrix will be discussed later CHAPTER 7 Utilization of the scatter peaks To test the calibration an alloy reference standard material is well suited There is no dark matrix contained but the enhancement effect must be taken into account For the details of the rather complicated definition of the input model used to fit bronze alloys properly the reader is referred to CHAPTER 6 Use of NBS Alpha coefficients 3 3 3 1 Use of the average instrumental constant By use of the predefined instrumental parameter file NBSalloy fpc and the AXIL fit result file NBS
126. ediate thickness samples There are also methods which can handle all kind of sample types e g the fundamental parameter methods Usually the methods which deal with narrower rage of sample types are more accurate 41 2 2 Si Li detector In energy dispersive XRF the acquisition of X ray spectra is performed using detectors that directly measure the energy of photons The resulting signal is proportional to the energy of the incident photon the detector is capable of detecting simultaneously photons of different energy A lithium drifted silicon crystal of a Si Li detector consists of a p i n structure referring to the p type contact dead layer on the entry side the intrinsic active volume and the lithium diffused n type contact Typically it is less than 10 mm in diameter and about 3 5 mm thick When a reverse bias in the range of 500 1000 V is applied to the device the drifted region acts as an insulator with an electric field throughout its volume A photon reaching the active volume generates photo electrons as well as Compton and Auger electrons which lose energy by producing ionization in the form of electron hole pairs The free charges are swept away by the applied bias and collected within a time of typically 25 100 ns Since the average energy to create an electron hole pair is well defined 3 76 eV at 77 Kelvin the total number of charges is directly proportional to the energy of the incident photon The frequency of s
127. eeees 79 4 7 EDITING OF ASR FILES TO BRING THE SCATTER PEAK INFORMATION OBTAINED WITH THE COMMAND SCAT_ROI TO A FORMAT EXPECTED BY QUANTITATIVE METHODG 80 CHAPTER 5 EMISSION TRANSMISSION METHOD 1 00 eeeeeeeseceecneeeeeeneeeeees 81 5 1 HOW TO COMPOSE THE TARGE D2 segcccycscecyeedoseacd coosesseetncda cous etasees 82 5 1 1 Example Targets B F nasienne tea joie anid E RE ES 82 5 1 2 Example Accumulative counting of pure targets ssseseseseseeeesseessressresseesseeesseee 83 5 1 3 Example Sample Soil 7 and target T1 eeeeeseeseeeeeseesesresseserssresseseresressrssresressessrs 83 5 1 4 Example Sample Soil 7 and target Tmetal eee ceecceceeneeceeeeceeneeceseeeceeeeeenteeeenaes 88 Sloe Exampl Lich n 356 5 sassctes ieri oner n a a a a e eei anes 90 CHAPTER 6 USE OF NBS ALPHA COEFFICIENTS 00 0 ceccscesecesecesecesecneeeneeeeeeeeees 94 6 1 EXAMPLE CALIBRATION WITH NBS STANDARDS AND PURE METALS UNKNOWN SAMPLE SYNTHETIC BRONZE cesses 95 6 2 AXIL FIT FOR NBS BRONZE ALLOY STANDARDS nsss 96 6 3 CREATION OF CALIBRATION STANDARD FILES eseese 97 6 4 CREATION OF THE ALPHA COEFFICIENT FILE seese 99 6 5 QUANTITATIVE ANALYSIS OF A SYNTHETIC BRONZE SAMPLE Gney bean insect ea E E E AA O S 101 CHAPTER 7 UTILIZATION OF THE SCATTER PEAKS eeeeeseeseeeseresseeressee 105 7 1 FITTING OF SCATTER PEAKS EXAMPLE RH SECONDARY TARGET EXCITATION csini gnntieei n kasaan 105 7 1 1 Classification ac
128. eight fraction wt i CaC03 weight fraction wt i CaCo03 weight fraction wt i CaCo3 weight fraction wt i CaCoO3 Pr a a a a weight fraction wt i CaCo03 Fig 7 15 For the description of the composition the reference standard Soil 7 used for scatter calibration twelve chemical elements are considered to contribute significantly The last but one SCROLL BOX for the calibration procedure summarizes the AVERAGE CALIBRATION CONSTANTS for the coherent and incoherent scatter peaks and for fluorescence Finally all relevant data will be stored FORM Enter the name for the calibration file A meaningful name but no extension it is by default clb is needed It is advisable to match it with the report file name rpt in case this option to create and save a report had been selected previously The report Soil rpt contains a wealth of information Among others the spectrometer specification as entered previously data for all standards and the calculated individual and average calibration constants 121 oherent scat constan RETE TIE TILI CONSTANTS FOR THE STAND Incoherent scat kah 7 2 ASR Limp s mA Cimps mAl x mi nittaaerneartolemenertm nc LUO mD CO OUIDAOUMOW IDC mm HQAhOLWOWOW wa SOON ONON D T mrono woop a ph pei pet pet pat pat E AVERAGE CALBIRATION CONSTANTS AND ESTIMATED SCATTER BACKGROUND Coherent scat constant 2 227150E 0003 Cimp s mAl Coh sca
129. em still is possible 140 Fig 9 10 The calibration file Extra2 cal contains calibration data for three elements originating from ten standards The calibration is achieved with Perform calibration For one element after the other a regression calibration will be performed Due to the fact that the uncertainty of calibration solutions was not even attempted to be known for the demonstration the Type of fit should be specified as Unweighted Straight line fit It is possible but not recommended to include the point with zero concentration which would correspond to a sample reflector blank measurement forced to zero intensity Fig 9 11 For the regression applied to the data points representing the calibration element Mn the origin zero concentration and zero intensity should not be forced in For the element Mn comparable also for Cu and Sr a fairly good linear correlation between the concentration and the intensity ratio Mn Kg to Y Ka is observed 141 Fit of Int Mn Ka Int Y Ka vus Conc of Mn in ppb Mean difference 1 7918E 000 z Ge OG1 Ora aoe Get OG1 Get OG1 0e 000 150 Conc Mn ppb Fig 9 12 Regression line for the element Mn After having passed this routine as many times as calibration elements are found in this case after Mn for Cu and Sr a report will be displayed The option Extend calibration for not measured elements will permit to interpolate and extrapolate the calibration
130. ement correction term Hy is given in 2 48 and 2 49 It can be noticed that the formula 2 58 does not depend on the sample mass load m F It implies that for a thick sample the intensity of X ray peak does not increase with increasing sample mass the sample has reached a saturated thickness However the intensity of any X ray peak depends in a non liner way on concentrations of all elements present in the sample To arrive at practical definition of a thick sample it has been agreed that assuming monochromatic excitation for a given X ray peak the sample is considered to be a thick sample if the error resulting from applying equation 2 58 instead of 2 47 is less than 1 It is equivalent to the following condition m m 4 615 a gt lnl0l amp gt 2 59 F F H Eo u E sin Q sin y The samples with the mass loads greater than that defined in 2 55 and less than that given in 2 59 are called intermediate thickness samples For an intermediate thickness sample the intensity of an X ray peak is a non linear function of the sample mass load It also depends in a non linear way on the concentrations off all elements present in the sample There are quantitative analysis methods which deal with only one type of samples e g methods assuming direct proportionality between the X ray peak area and the element mass load work only with thin samples the Emission Transmission METHOD works only with interm
131. erate individual entries in the fpc file Having created such individual constants for samples the option Individual and or average will make use of these constants for elements were they are available The finally obtained individual instrumental constants Instr constant for the seven standards are stored automatically into test fpc there is no need to save the results The only difference in the treatment of the sample file NBS1108 asr as compared to calculations using the average instrumental constant is the selection of Individual and or average in the FORM Type of Instrument constant There are no significant differences observed between the results obtained with either average or individual instrumental constants for the standard reference material NBS 1108 73 CHAPTER 4 EDITING OF DATA FILES The data files in QXAS are block structured Each file is divided in blocks that are identified by a header name with as the first character and as last The order of the blocks is random but within each block the data is well defined Each program in the package searches modifies and writes only the blocks known to that program leaving all other blocks unchanged The most important block structured ASCII file is the set up file setup ax in which all the information on the current status of the package is given This file contains the menu tree of all the installed programs in the package the general information for all the
132. ergent angle for secondary target 45 00 Fig 3 18 The secondary target is specified by its chemical symbol and incident and take off angle According to the available detector data sheet the Be entrance window of the used Si Li detector has a thickness of 25 um and a crystal thickness of 3 mm One should not expect any information about the Au contact layer and Si dead layer from the data sheet usual values are 0 02 0 05 um for the contact layer let us wish is really made of gold because it could be other materials too and 0 02 0 2 um for the dead layer These values importantly influence the light element analysis as they are so poorly known this always serves as explanation for poor results of these elements etector Characteristics Detector type Be window Cmicron gt 6666 u layer lt micron gt 0 02000 Dead layer micron 6 16060 Active depth mm 3 66600 Fig 3 19 A Si Li detector is specified in respect to its intrinsic efficiency by its Be entrance window thickness the thickness of the contact layer the dead layer thickness and the crystal thickness The Excitation Detection Geometry FORM does not need data for the distances in agreement with the reality but rather defines the air path previous choice Atmosphere Air between secondary target and sample the radiation travels only in vacuum therefore an input of 0 is adequate and sample and detector gap between detector and sa
133. eristic lines of the target are used to calculate the energy dependent matrix absorption and by interpolation absorption correction factors for selected elements in the sample The target can either be a pressed pellet mixed from simple compounds or a set of pure element standards metals can be used instead The sample element range is defined by the first and the last target element for which the absorption correction factors are calculated When the intensity ratio for a target element becomes too high it is not considered for the absorption calculation the sample element range will become narrower As basic assumption of this METHOD the excitation source must be monoenergetic To fulfil this requirement at least partially either a secondary target set up or isotope source excitation has to be used The ET METHOD is applicable with good precision for fluorescing elements in the minor and trace element range The calibration for this METHOD is to be established with the METHOD Elemental sensitivities previously and is loaded into this METHOD with Select calibration file Certain measuring parameters must be defined by activating Change Measuring Parameters If applicable for source excitation the half life of the excitation source is defined with an input for HalfL in days which will correct for the decay of the excitation source in between the calibration and the measurement of unknown samples Incid A and Emerg A refer to incident an
134. esting lines of Sn L Fe Ni Cu Zn and Pb Lg but one can recognize the undesirable rise of the background under the Pb La peak 96 Spectrum NBS1103 SPE Iteration 7 ChiSquare NBS1103 10mA vc2107 spe F it Pole pik BEG beg chan END end chan MIN min cnts MAX max cnts ROI SPECTR RES IDUAL LIN LOG Display 15008 2000 GO Channel Number CANCEL gt gt DISPLAY RESIDUAL _ Fig 6 2 In principle it is possible to find a fit model to describe the entire ROI of NBS1103 spe but as can be seen from the background rise under Pb L the minor and trace elements might be badly treated In analogy to NBS 1103 for the other reference standards NBS 1107 and NBS 1108 it also contains Mn similar models were defined In the spectrum NBS1115 spe neither Sn nor Pb were identified consequently only one input model is sufficient The spectrum of the synthetic bronze sample Synbronz spe was again fitted in 3 parts In SynBronz inp Fe and Ni were not included but the Pb L line with its energy of 9 185 keV 6 3 Creation of Calibration standard files In order to generate the std files the asr files have to be modified Each element to be analyzed in unknown samples or present in any of the standards has to be presented as a row entry in the standards asr files irrespective whether the element s peak is present in a spectrum or not One approach is to use a single inp file for the standards and for the sa
135. fferences between the three Source selection Amersham source Isotope Products source Fig 8 2 From three brands of sources the NEN source is selected From the SELECTOR BOX Isotope selection choose the item Cd 109 source from Spectral lines selection the item K alpha lines the calculations by use of L alpha line emitters is possible too but the calibration file Cd_52_90 cal was only defined in terms of K lines standards because there are more standards available for the first selection from Data input selection the item Enter data from keyboard This is now deep enough down in QXAS 126 Data input selection Enter data from old file Fig 8 3 The manual input of certain dimensions describing the source sample detector geometry will be activated A FORM Enter experimental setup parameters distances in centimetres has to be filled in with relevant data describing the geometry of the spectrometer For lazy users a pre defined FORM of course only applicable in its details for the used spectrometer can be loaded with Enter data from old file and the sensitivity file Cd angle sen Fig 8 4 Several dimensions in cm describing the geometry of the spectrometer must be known From the sketch in scale with the real situation see Figure 8 5 all relevant dimensions in mm can be taken The used letters from A to L and X can serve together with Figure 8 5 and its explanations to correlate the
136. ficients alp file Prepare alpha coefficients file is the definition of the element set as found in one of the std files As each of the standard std files must contain the same element list any of them can be used for the generation of the alp file Then a source description sou file describing the excitation source must be loaded essential are the incidence and take off angle in respect to the sample surface and discrete wavelength values accompanied by intensity ratios decribing the excitation source Finally the previously created std files and the appropriate alp file have to be recalled in order to evaluate the samples asr files Quantitative analysis of unknown samples Depending on the number of standards only one or maximum four polynomials concerning the relationship between concentration and intensity of the calibration standards are for disposition All example AXIL result files the standard description files std source files sou and alpha coefficient files are contained in the directory QXASdemo NBS 6 1 Example Calibration with NBS standards and pure metals unknown sample synthetic bronze As set of calibration standards bronze reference standards NBS 1103 1107 1108 1115 and pure metals Fe Ni Cu Zn Pb Sn is used to quantify a synthetic bronze sample As can be seen in its spectrum SynBronz spe and known from its preparation the elements Cu Zn Pb Sn are present and should be its
137. fore the fit With A Xk cst crscsensarsetotaagaedatecenan ihe tates acast da actAudatansstutahtesiganat betes 147 10 2 2 Spectrum fit With A XID cenene R E ia 148 10 3 RECOMMENDATIONS FOR THE FIT OF SAMPLE SPECTRA 150 10 3 1 Example intermediate thick pellet pressed from 50 Soil7 and 50 HWC as binder SOU LSS icc 2 he asda aaah ce ide ue le end de el eae 150 10 4 ELEMENTAL SENSITIVITIES METHOD 000 cece eeceeseeeeeeeeeesecneeeneeees 152 10 5 EMISSION TRANSMISSION METHOD eee eeeeeseeseeeeeeeeeeecesecnneeneeees 153 10 6 CALIBRATION ACCORDING TO QA NORMS CHAPTER 1 GETTING STARTED References material 1 2 3 5 6 7 8 9 10 11 12 53 This Manual has been prepared as a guide a hand on reference for the users of QXAS Quantitative X ray Analysis System 3 6 software package It should be of help for users with basic experience in X ray fluorescence analysis XRF It can also be used by scientists who have not worked before with the QXAS at all More experienced users should find here answers for many specific questions Some fundamental features of XRF are discussed as far as the use of QXAS is concerned some examples are worked out in detail and a guideline for quality assurance QA matters is given All references are made to the version 3 6 of the software package QXAS The manual and the latest version of QXAS are available on http www iaea org OurW ork ST NA NAAL pci ins xrf p
138. g the instrumental constants for scattering for Soil7 fpc For the use of Soil 7 as scatter calibration standard the concentration values of the Soil 7 certificate are needed The spectrum Soil7 2 spe had been fitted only for the scatter region according to model 1 with the input model S7 2 1 inp For the definition of the matrix selection of 7 There are known compositions in the FORM Known compositions the essential elements concentration values gt 0 1 are represented by Na 0 24 Mg 1 13 Al 4 70 Si 18 0 S 0 12 K 1 21 Ca 16 3 Ti 0 3 Fe 2 57 the dark matrix is described by H 0 89 C 7 18 and O 47 36 in accordance with reference 27 This information is stored as Soil7 reb file The specimen was prepared as a thick pellet total weight 2 0 g the measurement was taken with 40 mA Table 7 2 Concentration values for selected elements taken from the certificate for the IAEA reference standard material Soil 7 the values for the elements forming the dark matrix are in accordance with reference 27 IAEA Soil 7 selected elements concentration ug g C I 95 Si 18 16 9 20 1 K 1 21 1 13 1 27 Ca 16 3 15 7 17 4 Ti 3000 2600 3700 111 IAEA Soil 7 selected elements concentration ug g C I 95 V 66 59 73 Cr 60 49 74 Mn 631 604 650 Fe 2 57 2 52 2 63 Co 8 9 8 4 10 1 Ni 26 21 37 Cu 11 9 0 13 0 Zn 104 101 113 Ga 10 9 0 13 AS 13 4 12 5
139. he for the authors available standards is not sufficient to make any detailed study in this respect As a possible criterion in order to take a decision for the order of the polynomial the self reference results for the standards can be used Le the certified standards used for calibration treated as samples must lead to acceptable results At this stage reference 29 should be quoted The authors use the nomenclature l Straight line for Y a0 al X 2 Quadratic line for Y a0 al X a2 X X 3 Straight line constrained to zero intercept for Y al X 4 Quadratic line constrained to zero intercept for Y al X a2 X X In many cases calibration curve 4 seems to particularly compensate for inaccuracies in fundamental parameters used in the calculation of theoretical alpha coefficients especially over a wide range of analyte composition and better results have been observed However when the concentrations of the unknowns are out of the range of the standards it is suggested that the other calibration curves listed above be used for consistency in the results The extrapolation provided by the quadratic line 2 can sometimes give large errors When only one multi element standard is available calibration curve 3 is the only option At least two multi element standards are required for calibration curve 1 or 4 while three are required for calibration curve 2 As problematic in this respect wa
140. he incidence angle to differentiate between a residue s fluorescence signal of e g Sc on the surface and a bulk signal like Si the substrate which is important e g during the search for contaminations A thin layer Ni on top of a reflector also shows a distinct angular behaviour TXREF is a very sensitive technique for trace element determination therefore contaminations can be a major issue Strict cleaning procedures for the sample reflectors must be followed and all chemicals used must be of spectroscopy grade Furthermore it is not sufficient to examine the instrument and sample blank once in while Before each deposition of a standard or sample onto a reflector the clean reflector must undergo a blank measurement 137 In the following the concentration units of ppm parts per million and ppb parts per billion will be used They are defined for liquids as ug ml and ng ml respectively 9 2 Calibration Demonstration files directory QXASdemo TXRF AXIL result files 20ppb asr 40ppb asr 60ppb asr 80ppb asr 100ppb asr 120ppb asr 140ppb asr 160ppb asr 180ppb asr 200ppb asr Calibration file Extra2 cal Input file TXRF inp Spectra directory QX ASdemo TXRF SPE 20ppb spe 40ppb spe 60ppb spe 80ppb spe 100ppb spe 120ppb spe 140ppb spe 160ppb spe 180ppb spe 200ppb spe Ten calibration solutions containing the elements manganese copper and strontium with concentrations ranging from 2
141. he next COMMAND CALIB is this position correctly used for the energy calibration In most cases the cursor will indicate the correct position calculated by AXIL as a centroid of the peak in some cases it will not find a peak at all or it will place the cursor in the proximity of the maximum In order to identify the last case it is advisable to check several channels left and right of a proposed peak maximum for a possibly better solution Unfortunately experience is needed sometimes 1 2 5 Is it important to find the exact energy calibration for each spectrum to be fitted The answer is mixed No because AXIL does not treat the energy calibration as fixed throughout the optimization process iterations but varies the values of ZERO and GAIN So to some extend natural fluctuations caused by the statistical nature of the number of counts small amplifier drifts etc are corrected during the fit Yes because the better the energy calibration is described the smaller usually will be the Chi square values total and individual and the residual deviations criteria to judge the goodness of a fit Sometimes an energy re calibration will help to overcome serious fitting problems 1 2 6 How can I check the quality of the energy calibration The usual way is to select the COMMAND KLM_MARK which displays in the comment line directly under the AXIL blue spectrum frame the atomic number and the element s chemical symbol Simult
142. he symbol of the peak or a group of peaks e g FE KA or FE Exceptions are the peaks specified directly by using peak energy They become GR01 GR02 etc see X LINES SHOW To remove particular peak e g GRO1 from the list use COMMAND REMOVE GR01 1 2 18 Why will one need an input model file inp for treating a spectrum with AXIL There are good reasons to create such files however the AXIL fit itself can be performed without going through an intermediate step of creating an input model file One motivation for its creation is for repeated fits of similar spectra Such model file modified as required can be also used as a base for fitting different types of spectra The alternations being made can be saved under a different name for further use Another motivation to create and store input model file for every sample or group of samples comes from quality assurance In a properly defined and implemented quality assurance system all analytical operations should be documented It helps tracking back mistakes and formulating appropriate corrective actions To store a newly created or modified input model file one have to exit the AXIL fitting screen with the COMMAND STOP and select Save model then select appropriate action from the list An input model file will be created If necessary other parameters which can not be changed inside the AXIL window can be altered by selecting the option Specify parameters for spectrum analysis followed by Spe
143. here Collimator No collimator Fig 3 16 Various modes of excitation permit to select a pre defined secondary target arrangement For the excitation of the secondary target a Ag anode X ray tube was used Because it is a diffraction type of tube the anode take off angle will be around 4 its Be exit window has according to the description of the manufacturer a thickness of 300 um The tube high voltage was adjusted to 50 kV for a once established calibration this value never may be changed Although the scattered continuum originating from the anode will not significantly contribute to the excitation the Number of continuum intervals taking into account this effect had been defined as 50 For direct tube excitation one would work with a value of 200 or slightly less for the case of computation problems Parameters for K ray Tube Tube anode a Take off angle lt degree gt 4 6600 Be window thickness mm 6 3668 Operating voltage CkU gt 56 666 Number of continuum intervals 56 Fig 3 17 The X ray tube used for exciting a secondary target can be specified The used secondary target Rh is irradiated with an incident angle of 45 the direction of the emergent Rh characteristic radiation K and L lines is also 45 both in respect to the target surface and defined by the collimation system 62 Parameters for Secondary Target Secondary target material Incident angle for secondary target 45 00 Em
144. hiSquare HWC binder 5074071000 vc2401 spe KLM mark atn 1 e atn 1 T atn 10 4 atn 10 KPRIISVTFONVaAS Song GO CANCEL 1500 2000 Channel Number Fig 1 26 Sample blank spectrum obtained with a sample pressed from pure binder HWC The instrument blank Instr Bl spe Instr Bl inp is obtained by leaving the sample position empty Scattering when a sample is in the measuring position is mainly caused by the sample itself When no sample is present the scattering intensity is low and the exciting radiation can hit parts of the sample chamber otherwise shielded by the sample Consequently other peaks can be seen above the spectral background This blank measurement is representative for thin layer samples like air filters For instrument blank problems a background subtraction might be the only solution in order to get any results although this will drastically increase the uncertainty in reported concentration values for affected elements For the used spectrometer Fe with a count rate of 0 1 counts second Cu 0 1 counts second Sr 0 2 counts second and Mo 0 4 counts second were identified in the instrument blank 26 Spectrum INSTR BL SPE Tray removed instrument blank 3 vuc2403 spe KLM mark atn 1 e atn 1 T atn 10 4 atn 10 GO CANCEL cS o u n t S 7 Cc h a n n e 1 1500 2000 Channel Number Fig 1 27 Instrument blank obtained from a measurement without any sample in the usual measur
145. his process is called relaxation One of the relaxation paths starts by filling in the vacancy in the ionized shell by another electron of the same atom Not all electrons can take part in such transition For a given electron there is a certain probability that such transition will occur For the vacancy in the K shell the most probable electronic transition to fill in the vacancy is by the electron of the Lm shell Such transition is denoted as K lt Ly To describe this process in terms of energy we have to note that the binding energy of an electron in an atom is expressed by a negative number In this convention the energy associated with a fully ionized atom an atom totally stripped out of electrons is equal to zero On the opposite end the energy associated with an atom possessing all of its electrons is equal to a negative number the smallest energy which can be associated with that atom Therefore in terms of energy removing one electron from an atom e g by photoelectric effect is equivalent to subtracting a small negative number from the overall also negative atom energy E E 2 1 E E Eo 2 2 E gt E 2 3 E is the energy of a neutral atom negative value Eg E2 is the energy of the ionized atom lacking one K shell electron and the Ex is the binding energy of the expelled K shell electron If the vacancy in K shell is filled in by an electron from the Ly shell it creates a new vacancy this time in th
146. hould be to switch to Secondary enhancement corrected because only for ancient PCs one had to think of the speed for the computation Otherwise select this option automatically for samples The final iterations are initiated with Calculation of Geometry constants Analysis of unknown samples Hint for problems during the final calculations of instrument constants or sample concentrations When after a short flash the calculation will stop and bring back to the previous COMMAND Calculations of Geometry constants Analysis of unknown samp 114 the conventional memory did not suffice Terminate QXAS completely and re start it go again to the lat COMMAND Calculations of Geometry constants Analysis of unknown samp the temporary file fundp tmp still carries all data and initiate the calculations again 11 12 2666 18 12 26 Sample identity Lake Sediment IAEA SL 3 2g 46mA vc2144 spe Spectrum fitting data E QKASDEMO FP SCATT LAKESED3 ASR Instrument parameter data E QKASDEMO FP SCATT SOIL FPC Instrumental identity Secondary Target fiverage instrumental constant 1 1490E 08 7 Instrumental constant for coherent scatter 1 7883E 06 Instrumental constant for incoherent scatter 1 4163E 066 The secondary target Rh The tube anode Ag Tube voltage 56 KU Tube current 46 666 mA Measuring time 16606 Sec Collimator No Collimator Filter used No Filter Atmosphere Air Report of Calculated Concentrations Sample thickness infinite
147. i Std T102 V205 Cr Std Cr2 K207 for chromium MnO2 Fe Std Fe203 Co Std CoO Ni Std NiO Cu Std CuO Zn Std ZnO Ge Std As203 SeO2 Br KBr for bromine SrCO3 Y203 Zr Std ZrO2 Nb Std Nb205 Mo Std Cd LStd Sn LStd Hf LStd Ta LStd W LStd Au LStd and Pb LStd For a detailed description see the EXCEL standards xls file Conclusion In Figure 3 28 all individual values are displayed Several features can be noticed Between the elements Ca and Pb the points scatter around an average value 1 149 0 0435 107 this value is also used as average instrumental constant One can differentiate between the elements contained in pure form in the calibration standards namely elements Ti Pb with their sub group average of 1 165 0 0215 107 and the elements contained in compounds which automatically also have binder added namely elements Ca Nb with a sub group average of 1 133 0 0543 10 There is a systematic difference between the two sub groups and not unexpected the standard deviation for elements present in compounds is higher As a consequence one would obtain slightly different results by using either only compounds or metals for calibration From the spread of the points one can conclude the higher the total number of standards used for calibration the better defined will be the average value and potential troublemakers could be identified and excluded For the elements below Ca namely Al Si P S a
148. ibration procedures experimental aspects experience etc many of which are in the hands of the user Although great care has been taken to maintain the accuracy of information contained in this publication neither the IAEA nor its Member States assume any responsibility for consequences which may arise from its use The mention of names of specific companies or products whether or not indicated as registered does not imply any intention to infringe proprietary rights nor should it be construed as an endorsement or recommendation on the part of the IAEA CONTENTS CHAPTER gt GETTING STARTED 25 oon ae a elon cain E AA A a 1 1 1 SPECTRUM FITTING AXIL raa ia i a a E 2 1 1 1 Example Input model Target1 inp and spectrum Target1 sp of a demonstration Samples cA nten g aa teste E A Stele ete S ogee tina 2 1 1 2 Example Input model Pb inp and spectrum Pb pure spe for a pure lead c libration standardisere an a e a a a e mu a a secon ees 8 1 2 FREQUENT QUESTIONS RAISED WHEN WORKING WITH AXIL 9 1 2 1 For a new spectrometer what settings of amplifier and ADC are appropriate 9 1 2 2 What is an energy calibration How to get an energy calibration for a spectrum measured with my own spectrometer scree vadyianse diets saaseesorcaan dey Harm vngstageaase cv ahansencuas 10 1 2 2 1 Example spectrum E calib Spe ic sscisssccesiscasectavsassadsssasdeatasessaac nncdevssaseaanns 11 1 2 3 In the spectrum E calib spe the max
149. ic definition is used in laboratory practice namely assuming monochromatic excitation and neglecting in 2 47 the enhancement correction term it has been agreed that for a given X ray peak the sample can be considered a thin sample if the error resulting from applying equation 2 53 instead of 2 47 is less than 5 This definition is equivalent to the following condition x A _ l expt a m F 1 m F a m F 1 05 2 54 Graphical solution of the condition 2 54 is presented in Figure 2 3 A miF gt 1 05 min samp A mIF 1 exp a mMF a mF AJ mIF a MIF pn semcie lt 0 1 a miF Fig 2 3 Graphical solution of the condition 2 54 leading to a definition of a thin sample critical mass load The solution of 2 54 tells us that for a given X ray peak the sample can be considered a thin sample if 40 m m 0 1 lt 01 amp 7 10 mie 2 55 i F F Un Eo 4 Em E sin Q siny The opposite extreme case is the thick sample approximation in which the mass load of the sample is assumed to be infinitely thick m IF gt gt 1 2 56 Condition 2 56 implies that m exp a 0 2 57 Pf 2 57 Taking into account 2 57 the formula 2 47 for the intensity of characteristic peak emitted by a thick sample is transformed to 1 l Gc Ef L E oQ u E u E 1 H es Cis T 2 58 n m pini y sin Q sin y The enhanc
150. icients METHOD SSPEC_ID Cu 100 5mA vc2099 spe SDATE_MEA 00 00 2000 00 00 00 SMEAS_TIM 40 1000 SPEAKS 1 29 1 50520 Certain other element entries as imposed by the elements found in the unknown sample and also in the other calibration standards must be represented by rows Following the syntax of asr files their atomic number an entry of either 1 or 2 for either Ka or Lo peak representation its characteristic energy the net peak area its standard deviation established during the AXIL fit and the individual Chi square value must be contained Of relevance is the net peak area As another complication the copper calibration standard had been measured with 5 mA tube current but the sample with 10 mA The NBS METHOD can neither handle variable tube current nor decay corrections for the case of source excitation therefore the net peak area is multiplied by a factor of 2 in order to correct to the expectation value at the double current Edit also the total number of fluorescence net peak areas SPEAKS from 1 to 7 Instead of a value of 0 0 for the net peak area for the elements that had to be included for the sake of the program the lowest possible value 0 1 was entered in order to suppress warning messages No attention was paid to the standard deviations and Chi squares none of them is used All these alterations to the original file had been saved as Cu NBS asr directory QXASdemo NBS now ready for creation of a co
151. ies Y and the normalized concentrations X This choice is valid for all sample elements The functions 2 Y a X 6 1 0 with Y being defined as the ratio of the intensity for a calibration standard element to the pure element intensity normalized or relative intensity and X being the normalized concentration i e 100 gt 1 As it is the case with the preliminary calibration for one point per element the pure element standards only a linear relationship makes sense For the two higher order polynomials program internal constants will be used to establish the curves coefficients an Fig 6 10 An intermediate calibration with only pure elements as standards leads to preliminary results After the preliminary results are obtained as a consequence one usually will be in desperate need of suited final calibration standards for a more reliable quantification of the unknown sample s For the demonstration example the named NBS reference standards were available 101 With the set of NBS1103 NBS1107 NBS1108 and NBS1115 std files the pure element standards std files Fe NBS Ni NBS Cu NBS Zn NBS Sn NBS and Pb NBS and the a coefficients file NBS amp pure alp the final quantification of the synthetic bronze sample by its AXIL result file SynBronz asr is possible The sample asr file needs not be edited The SCROLL BOX CALCULATED RELATIVE INTENSITIES OF STANDARDS displays the relative intensities as
152. imary radiation and the sample surface the take off angle between sample surface and detector axis Detector take off angle degrees One has to breath deeply but then one can relax any angle not too far away from reality of course will do these angles are needed for the absorption correction calculation of the standards By a first guess for the take off angle a value of 90 had been chosen parallel to the detector axis direction For the incidence angle the basis angle of a 125 cone defined by the central ring of the annular source as basis and as peak the centre of the sample front surface is good enough From the naturalistic sketch in Figure 8 5 an angle of 52 can be depicted The two angles 52 for the incidence and 90 for the take off angle had been chosen firstly they can be justified by simple arguments and secondly in order to demonstrate even though the take off angle value is relatively far away from the later to be calculated value that the choice is not too critical In the preliminary calibration file Cd_52_90 cal eleven asr files of pure metallic standards had been included Ti Cr Fe Co Ni Cu Zn Ge Zr Nb and Mo No effort was made for this calibration to establish also a L lines calibration because it serves only as a slave for the calculation of incident and take off angles For the addition of a standard to the calibration knowledge about the measurement date essential for source exc
153. imum of TiKg is well defined in contrast to the ill defined MOK gsc es conve decbecadesnccnsdictonsdes etre sie earel sak ENE ti havens 13 1 2 4 With arrow up and down I can jump from one peak to the next COMMAND CALIB is this position correctly used for the energy calibration ee 13 1 2 5 Is it important to find the exact energy calibration for each spectrum to be 1 2 6 1 2 7 1 2 8 1 2 9 1 2 10 1 2 11 1 2 12 1 2 13 1 2 14 1 2 15 1 2 16 1 2 17 1 2 18 1 2 19 1 2 20 1 2 21 1 2 22 1 2 23 1 2 24 1 2 25 fitted 13 How can I check the quality of the energy calibration eseeseeeseeeserrsereerrrreesere 13 How to fill the spectrum with characteristic lines correctly oo eles eeeeeeeseeeeeeee 13 1 2 7 1 Example IAEA reference standard spectrum Soil7 1 spe 0 0 0 eeeeeeeeeee 14 What background model should be Used 0 eee eee ceeceesseecneeceseeeeeeeeseecaecneesseeeenees 14 How to select the ROI region of interest properly eee eeeeeceeeeeceeeeeeeeeeeeeteeeees 14 What are the minimum requirements in order to initiate a fit of a spectrum 15 What are the criteria to be satisfied with a fit result oo eeeeeeneeeneecneeeeeeeeenees 15 What is the difference between X LINES ADD e g FE and FE KA FE KB What are the secrets behind and 807 Loic ccsessesccccccccccssssssssceccecessusnensseeees 15 What are the peaks in a spectrum that
154. in this FORM in another run verage Instrumental Constant for fluorecence 4 8668 for coherent scattering 6 6666 for incoherent scattering 6 6606 Fig 3 21 The average instrumental constant values will be zero for the first time the parameters are defined These values are to be obtained by suited calibration standards and have to be entered later 3 3 2 Description of calibration standards for this METHOD In principle a single standard can be sufficient to calibrate for this METHOD but with the exception of exotic problems one will try to use as many standards as possible to establish the instrumental constants In the ideal case all calibration standards will result in the same instrumental constant different to the concept of e g the Elemental sensitivities METHOD With the selection of Specify standard sample information from the SELECTOR BOX Current Instrument Parameter File and Select samples for quantitative analysis all standards have to be processed It is possible but not necessary to do this in one run Not very fortunate because misleading the same text Select samples for quantitative analysis is used for both calibration and the later analysis of unknown samples 3 3 2 1 Calibration standard example K KBr asr From the directory QXASdemo ASR Stds the AXIL result file K KBr asr was selected to demonstrate the calculation of the potassium instrumental constant in order to build up finally from all calibrati
155. ing position Intermediate thick samples will exhibit an intermediate behaviour both instrument blank and sample blank must be considered for each element affected Conclusion when blank peaks cannot be eliminated and a precise estimate of the result uncertainty is required the intensity of a peak in the analyzed sample should exceed tenfold the intensity of that peak observed in the blank Otherwise the uncertainty of the result for that element should be carefully estimated taking into account the contribution of the blank 1 3 Batch mode Demonstration files files in the sub directories Original MovePeak and ZeroCont of directory QXASdemo Batch Spectra Sum Spec spe 01 spe 41 spe Input files SumSpec inp CaZnSrPb inp MovePeak inp Batch file FitAll bat The BATCH COMMAND is very useful for processing a number of similar spectra to be treated with the same input model in an unattended manner Large groups of spectra are obtained during e g scanning measurements analysis with the use of X ray micro beam XANES etc or during bulk analysis of large batches of samples The number of generated spectra may reach several thousands for a run or batch It would be very impractical if not impossible to process such large number of spectra in the interactive mode When processing the spectra in batch mode several requirements must be fulfilled All spectra to be processed already converted to QXAS format
156. inp created for the purpose of this demonstration When no particular model file is loaded certain parameters like background model energy calibration coefficients etc are set to default values They require further adjustments AXIL Spectrum analysis using Voigt function Model file Analyse Spectra Select model Save model Parameters of current model Atmosphere is AIR elements using groups of lines Background model Smooth Filter with 256 iteration Energy calibration zero 9 06 eU gain 20 08 ev ch Resolution calibr noise 120 6 eV fano factor 114 Region of interest set AUTO m e x 7 Fig 1 6 Default input model parameters To retrieve the input model use the SCROLL BOX to select the directory QXASdemo GetStart and load the Target1 inp file Select model file INP gt E QKAS DEMO saN FI GURESN ASR STDSN ETN INP STDSN SPE STDSN Fig 1 7 SCROLL BOX for loading the input model file Targetl inp The input model file Target1 inp contains the correct energy calibration an appropriate region of interest ROD which brackets all the peaks of interest to be fitted in our example spectrum the peak positions and several other pre defined parameters AXIL Spectrum analysis using Voigt function Model file E QKASDEMO GETSTARTNTARGETI INP Select model Save model Parameters of current model Atmosphere is AIR elements using groups of lines Background model LINEAR of
157. ired under preferably identical conditions The effects of elastic coherent Rayleigh and inelastic incoherent Compton scattering are usually much weaker than photo absorption Elastic scattering for a single electron can be deduced from classical electrodynamics but due to its coherent nature the form factor has to be introduced to account for all electrons of an atom The differential cross section which describes this kind of interaction depends on the energy of the primary radiation the atomic number of the scattering atom and the scattering angle because of its inherent anisotropy It shows a pronounced maximum for forward scattering Inelastic scattering is relatively independent of the atomic number of the scattering atom The observed energy shift AE Eo Eg of the Compton peak depends on the scattering angle v B E EE ET 511 Eg 2 64 For this expression all energy values used are in keV Note that the scattering angle is the sum of the incidence and take off angle V g g With increasing atomic number scattering effects get weaker and the ratio of coherent to incoherent scattering intensity will change favouring elastic scattering 45 0 14 Differential scattering cross sections for Mo Ka radiation 0 12 scattered by carbon elastic S inelastic cm2 g sr 2 o 0 04 0 02 0 T T T T T T T T T T T T T T T T T 1 Oo 10 20 30 40 50 60 70 80 90 100 110 120 1
158. is element are questionable 56 Calibration file E QKRASDEMO ELSENS COMPOUND CAL Created on 69 28 2606 Fit optimization performed on 11 68 2606 Tube excitation SecTarget Rh Operating at 56 6 KU Polynomial fit of Ka sensitivities Sens mu AB AL Z AZ Z 2 Weighted fit Mean z diffence 6 2259E 600 2 5829E 664 5 9240E 002 7842E 003 1907E 002 3974E 882 3159E 000 1537E 661 546G6E 661 97645E 661 5 761E 863 2288E 663 7455E 665 I I i I i i C a a jeh peh pah OD pa 3 4 5 6 4 FTSummary of fitted Ka sensitivities Arrows lt PgUp gt lt PgDn gt lt Home gt lt End gt move lt ESC gt exit lt Fi gt Print Fig 3 10 Report displayed as scroll box containing the mean difference between all measured calibration points and fitted sensitivities as criterion to judge the polynomial representation 3 2 2 Example metals cal For the calibration metals cal twenty one asr files of pure metallic standards had been included Al Std Si wafer S Std Ti Std Cr Std Fe Std Co Std Ni Std Cu Std Zn Std Ge Std Zr Std Nb Std Mo Std Cd LStd Sn LStd Hf LStd Ta LStd W LStd Au LStd and Pb LStd For the elements with K emission lines between Al and Mo fourteen calibration points the calibration metals cal was extended Optimize Calibration with Least Square Fit with Order of polynomial 5 as Type of fit Linear and Weight of fit Yes A
159. itation and the live time usually transferred already from the asr file correctly are needed The standards can be treated as infinitely thick in terms of XRF so the standard s sample mass is left at the default value 0 00000 The relative error in the concentration value of the standards stddev with a default value of 5 it is of no relevance for this intermediate calculation was set to 0 1 After the correct inclusion of all standards the calibration for all elements found in the standards was finalised with Perform calibration The option Optimize Calibration with Least Square Fit was not used only the raw data is needed 8 1 3 Definition of the excitation source and geometry for the Monte Carlo simulation To be aware as pre requisites for the calculations are needed Two properly defined files namely a cal file Cd_52_90 cal and a sou file Cd test sou a wealth of information about the geometry of the system and the detector characteristics to be taken from the detector data sheet hopefully supplied by the manufacturer The detector characteristics are not critical when medium to high energy lines are used for calibration Select from the SELECTOR BOX Excitation selection the item Annular source excitation next select from Source selection the item NEN source The used Cd 109 source is of unknown provenience but the two other options were not brought to working despite the fact that there are only slight di
160. iteria must be fulfilled otherwise one or more of the parameters of the input model have to be varied The order of the background PARAME must be increased usually the first choice a different background model has to be chosen lines could have to be added the ROI could have to be set manually the self absorption etc must be defined Each of such changes will necessitate a re fit and re inspection of all the above named criteria The last two criteria must be fulfilled at any rate for lines of elements which the standard is composed of After all the criteria were fulfilled the results have to be saved as asr files SAVE_RES 149 For QA requirements although not necessary to proceed with any of the quantitative METHODs it is mandatory also to save the input model This is not executed automatically nor is there any reminder to this action After each successful fit of a calibration standard spectrum the fit program AXIL has to be terminated with the STOP in order to save also the inp file Save model With any further lt Escape gt after the STOP any changes to the model are lost even when an inp file had been created before loading the spectrum All asr files are stored with the same name as the spe file into the same directory It is advisable to move the asr files in a different directory 10 3 Recommendations for the fit of sample spectra Remembering the complicated definitions of the input models for
161. k s energy minus 1 74 keV E g an intense iron signal should be defined either by X LINES ADD FE or FE KA FE KB The asterisk suffix added to X ray line entry e g FE KA FE KB corrects for the deviation of the peak shape from an ideal Gaussian profile This peak shape correction is counted as background under the peaks not as part of the parent peak As a consequence applying this peak shape correction will slightly diminish the net peak area as compared to a fit without this correction added The peak shape correction includes also correction for the presence of the escape peaks When the amp suffix is appended to K series or L series peaks they will be fitted with the Voigt peak profile instead of the Gaussian one It may be of use in the case of high energy K series peaks such as Pb Kg g It also improves fitting of high intensity L series lines of heavy elements The suffix amp can be mixed with the other two 15 The three symbols amp can not stand alone but have to be attached to an element name e g FE or a peak FE KA in the COMMANDs X LINES ADD Verification with the COMMANDs X LINES SHOW The lines included together with any of these symbols will be accompanied by the respective remark Escape peak Peak shape corr Voigt peak 1 2 13 What are the peaks in a spectrum that cannot be matched with any KLM MARK markers There is a list of peaks that can sometimes be found i
162. l background models The exponential background is preferably used if a strong curvature of background is observed or the background level varies significantly in the ROI Linear background model will usually do better for slowly varying background the magnitude of which is more or less of the same order in the fitted ROI The orthogonal polynomials overestimate the background and small peaks might be lost their peak area is either underestimated as compared to the recommended models or even negative The Bremsstrahlung background model was originally meant for charged particle excitation It makes use of some parameters as defined in the input model At the time of writing this text for the current version of QXAS the smooth filter background model did not provide meaningful results 1 2 9 How to select the ROI region of interest properly The region of interest ROI should be selected such that all peaks of interest are included Its definition comes after the definition of the peaks Many times the automatic region of interest is appropriate COMMAND ROI AUTOMATIC it is already the default It will start with sufficient room left from the lowest energy peak present in the fit model previously defined with the COMMAND X LINES and will stop with sufficient room right of the highest energy peak There are good reasons to select the ROI manually Sometimes the low energy part of the spectrum is difficult to describe so either an exclusio
163. l with the version of QXAS available although in principle it was an alternative to the above named background models for former versions ORTPOL has one more parameter R with default 1 5 to be optimised usually between 0 5 0 75 1 0 1 25 1 5 1 75 2 0 It showed a strange behaviour for some example spectra overestimation of background with the consequence of negative intensities for small peaks which could be identified positively with other background models Initiate the first fit with a sufficiently high number of iterations sufficient means that the number of iterations executed in a fit run will be found less than the starting value FIT N_ITERe entered by the user The following criteria are to be checked in descending order Total Chi square ChiSquare less than 3 0 Residual COMMANDs DISPLAY RESIDUAL between 3 When there is no trend it is tolerable that a few points can be ouside of the indicated range The residual should please the eye therefore it is a soft criterion because the experience of the user will be of influence Results COMMANDs REPORT GO e Individual Chi square values less than 3 0 e The standard deviation st dev must be greater than the square root of the peak area this criterion only can be checked by use of a calculator Not necessarily the peak area will change much due to this criterion for a refit with changed parameter s Each of the cr
164. lculations of Geometry constants Analysis of unknown samp will initiate the calculation of the element specific geometry instrumental constant for potassium The instrumental constant for potassium in KBr is obtained as 6 998 10 Note the enhancement correction factor of 2 21 which means the K intensity was corrected with this factor to account for the excitation of potassium by bromine 67 61 14 2687 13 63 57 Sample identity For K 3122 KBrthbinder 5mA VUC2452 spe Spectrum fitting data D 11_NOU QKASDEMO ASR STDS K KBR ASR Instrument parameter data D 1i1_NOU QRASDEMO FP SCATTNTEST FPC Instrumental identity Secondary Target The secondary target Rh The tube anode Ag Tube voltage 50 KU Tube current 5 666 mf Measuring time Sec Collimator No Collimator Filter used No Filter Atmosphere ir Report of Calculated Instrumental Constants Sample thickness infinitely thick Concentration Instr constant Absorption Enhancement ll 26 83 z 6 99 75E 68 1 1648E 63 2 2136 II Fig 3 27 Instrumental constant for potassium of the KBr calibration standard In this manner all available calibration standards there is no limit to their number can be treated For each of the forty four asr files an instrumental constant for fluorescence can be obtained Al Std Si wafer P KH2PO4 for phosphorus K KH2P04 for potassium S Std MgSO4 for sulphur K2CO3 K2 Cr207 for potassium K KBr for potassium CaCO3 T
165. lectric mass attenuation coefficient Tn It is expressed in centimetres square per gram of a substance cm g It is possible to convert between the atomic cross section expressed in barns per atom and the photoelectric mass attenuation coefficient expressed in centimetres square per gram One can assume an infinitely thin layer of a substance composed of atoms of single chemical element where each atom K shell can be ionized by incoming photon with the same cross section denoted as 7 barns atom The assumption about the thickness of the layer is necessary to neglect shadowing or any other interfering effect which could eventually diminish the cross section of some atoms in the layer The total atomic cross section for the whole layer expressed in barns is equal to nt 2 8 where n is the number of atoms in that layer The photoelectric mass attenuation coefficient expressed in em g is obtained by multiplication of the total atomic cross section by the factor 10 and division by the total mass m of all the atoms in the layer K 4 nt T 107 2 9 m The number of atoms in the layer can be expressed in terms of the layer mass m atomic weight of the element A and the Avogadro s constant Na n N 2 10 By combining 2 9 and 2 10 qA fem g7 10 Magt barns atom 2 11 The probability of photoelectric effect has contributions from various electronic shells also called photoelectric absorption
166. lication air filter sample ssneeeeneeesesessensseesseeesseeesseesseessesssetessseessresseesseesseeesseee 59 3 3 CALIBRATION FOR THE FULL FUNDAMENTAL PARAMETERS MELHOD nna N e a A A a AN E e A Raes 61 3 3 1 Set up of the instrumental parameter file 6 tess caisiapsessaeaciscat ap nnstesiunpasmdaneteas 61 Del si go eao E EC EEE E EE A E E E 61 3 3 2 Description of calibration standards for this METHOD sesseessseessseesserssesssesesseee 64 3 3 2 1 Calibration standard example K KBt aSt ssneseesseeeeseeesseesseessereseeeesseessreso 64 3 3 3 Application bronze alloy sample ccs s 4 225 fo csacuets ots ted ta scacss ceeds bea ita Poem tate 70 3 3 3 1 Use of the average instrumental Constant 00 0 0 ee eeeeeseceeeeeeeeeeneecnseenseeeees 70 3 3 3 2 Definition and use of individual instrumental constants eee eeeeeeeeee 12 CHAPTER 4 EDITING OF DATA FILES o oo eceeseceeecneeeaceeeceseeeeceaecneseaceteeeseees 74 4 1 EDITING OF SPECTRUM FILES SPE jes ca0 tye casei paestetbend sami iotetaas 74 4 2 EDITING OF CALIBRATION FILES CAL 0 0 eee eeeeceseecesecneeeeeeeeeeeees 75 4 3 REMOVING ELEMENT ENTRIES FROM ASR FILES oioi 76 4 4 INDIVIDUAL INSTRUMENTAL CONSTANTS IN INSTRUMENT PARAMETER PILES EPC orenen einna e Berar i T11 4 5 FILE FORMAT OF ASR FILES AS EXPECTED BY THE NBS METHOD viren a E E E TAA E A R T A 78 4 6 SOURCE DESCRIPTION FILES SOU ce eeceecsseceeeeneeeeeeeeeeeenaeen
167. lose to this value As representatives the elements nitrogen its atomic number represents the major constituents of the binder and as sample contribution calcium or silicon are proposed Their stochiometry was varied until the average atomic number was found The first compound N 4Ca is equivalent to N 85 59 and Ca 14 41 It results in the instrumental constants for coherent scattering 1 8328 10 and incoherent scattering 1 6765 10 The second compound NsSi is equivalent to N 71 375 and Si 28 625 It results in the instrumental constants for coherent scattering 2 0813 10 and incoherent scattering 1 6494 10 Other compositions are possible with the constraints of the given average atomic number and a realistic description of the sample composition 7 3 METHOD Fundamental Parameters monochromatic excitation scatter peaks Demonstration files directory QXASdemo FP Scatt Calibration file Soil clb 116 AXIL result files for calibration directory QXASdemo ASR Stds Soil7 2 asr Instr Bl asr Si wafer asr K2CO3 asr K KH2PO4 asr K2 Cr207 asr K KBr asr CaCO3 asr Ti Std asr TiO asr MnO2 asr Fe Std asr Fe203 asr Zn Std asr ZnO asr Br KBr asr SrCO3 asr Y2O3 asr Zr Std asr ZrO2 asr Nb Std asr Nb2OS asr Pb LStd asr AXIL result files sample Instr Bl asr LakeSed3 asr Fundamental Parameters for monochromatic excitation by use of the scatter peaks also
168. ly one peak per element From Ar on upwards also a Kg peak accompanies the Kg for any proposed element intensity ratio Kg Kg is roughly 100 15 From the element Fe upwards Z gt 26 the L line markers are displayed shifted towards lower energies as compared to K lines markers of the same element In case of samples highly abundant in e g Rb and Sr the L series peaks of theses elements can be seen in the low energy region With increasing atomic number the markers pass the region of spectrum containing the scatter peaks then the K lines markers move out of the spectrum and only L lines markers remain Note the differences in their overall number and proportions as compared to the K lines The Lg Lg peak ratio is roughly 1 1 M lines and other structures that may be observed in X ray spectra e g satellite lines diffraction peaks are left to the interpretation of experienced user Write down the identified elements peaks and then feed the information into the fit model with the COMMANDs X LINES option ADD 1 2 8 What background model should be used The background should be described by a smoothly changing function which follows the fluctuations in background only regions Particularly the edges of the fit ROI should not deviate from the trend of background Under the peaks especially the intense ones the background usually will rise due to the imperfect description of the peaks by Gaussian profile Recommended are either linear or exponentia
169. ly thick Iterations for scatter peaks 11 Pre set convergence 10 z Last convergence 084 z Ele line Constituent Concen lt elem gt Absorption Enhancement ll SA Ea EE ea a SR a a a a a A i Ka 14 69 102 z 6 5944F 64 1 80276 K Ka 7851 78 74 107 ppm 3 0667E 03 1 1333 Ca Ka 9 32 M a 9555E 03 1 0291 Ti Ka i 2188 21 447 ppm 1572E 03 1 6251 Mn Ka 315 46 867 ppm 4663E 83 1 0022 Fe Ka 8153 45 633 ppm 058 7E 62 1 66668 Zn Ka 16 682 131 ppm 2184E 82 1 6666 Br Ka 2 58 431 ppm 0122 E 002 1 6668 Rb Ka 29 11 548 ppm 5573E 82 1 6666 Sr Ka 394 37 371 ppm 4070E 02 1 6666 Y Ka 16 66 356 ppm 2996E 82 1 6666 Zr Ka 181 18 922 ppm 2212E 2 1 6666 Nb Ka 4 70 325 ppm 0119E i 1 6666 Pbh La Ph 19 66 1 455 ppm 2298E 82 1 6666 Total percent of fluorescent elements 25 93 z Total matrix is estimated as Na 19 35 Al 88 65 Dark matrix is estimated as C 54 15 0 49 857 I Iterations for concentration 26 I a A We AO Con TON Cook Oo Fig 7 12 Quantitative results for the reference standard SL 3 obtained with the Full Fundamental Parameters method Table 7 3 Selected element concentration values for IAEA SL 3 standard reference material established with Full fundamental parameters and Backscatter fundamental parameters METHOD in comparison with the certified values K ug g Ca Ti ug g Br ug g Rb ug g Sr ug g Full fundamental 7852 9 32 2188 2 6 29 1 394 parameters MET
170. minent for the case when a hole is created in the K shell of the silicon and the 43 accompanying emission of its characteristic radiation leaves the crystal Consequently an escape peak at an energy 1 74 keV less than the parent peak energy can be observed Escape peaks can be included into the AXIL fit model with suffix In order to include the tailing and the escape peaks simultaneously within AXIL the peak shape correction suffix can be used The intrinsic detector efficiency amp F for a Si Li detector has a characteristic dependence on the energy and is governed at the low energy side mainly by the thickness of the beryllium detector entrance window on the cryostat vacuum enclosure the contact layer and the silicon dead layer also responsible for the discontinuity at the absorption edge energy of 1 84 keV and at the high energy side by the thickness of the crystal itself Between 5 and 20 keV the efficiency is close to 1 for many conventional Si Li detectors The dependence of the detection efficiency on the energy and other detector parameters is given by the following formula E exp u E ty My lE tau My lE tal l exp us E ts p 2 63 where fge tau fg and ts are the area related masses in g cm of the Be window the gold contact layer the silicon dead layer and the detector crystal itself respectively Usually these values are specified poorly in the detector data sheet 0 9 5
171. models are proposed For all models the background subtraction is tried to be kept to the minimum because also the background is mainly caused by scattering As demonstration sample a thick pellet of Soil 7 spectrum Soil7 2 spe serves Model 1 The example spectrum Soil7 2 spe has to be fitted with the model file S7 2 1 inp A region of interest is selected such that all K scatter peaks incoherent Kg coherent Kg incoherent Kg and coherent Kg are well contained A Gaussian function is used to describe the coherent scatter peak COMMANDs X LINES ADD RH KA COH The incoherent scatter peak is added to the fit model with X LINES ADD RH INC The incoherent scatter peak maximum energy will be calculated by AXIL according to the angles defined in the input file default values are 45 incident and 45 take off angle which are acceptable for secondary target excitation but not for source excitation The incoherent scatter peak is much broader due to the fact that scattering occurs over a range of angles around the nominal 105 scattering angle divergence of beams and the Doppler broadening On the low energy side of the incoherent scatter peak a tail is formed due to multiple scattering in the sample This peak profile cannot be accurately described by a single Gaussian function Instead a number of Gaussian peaks are used to compose the incoherent scatter peak profile With a distance in terms of energy of 0 25 keV between each other energy v
172. mple chamber of 0 5 cm The incident and take off emergent angle is defined by the collimation system with an average value of 45 for both Excitation Detection Geometry Dist source sample cem Incident angle lt degree gt Dist sample detector cm Emergent angle degree gt Fig 3 20 The sample geometry is specified by the incident and take off angle Both are to be defined in respect to the surface normal The distances between the excitation source the sample and the detector entrance window need not reflect the real dimensions but are rather the distances radiation has to travel in air and therefore imposes a specific correction No filter between secondary target and sample was used Filter between source and sample Such a filter is often used for direct tube excitation in order to shape the primary spectral distribution The last FORM Average Instrumental Constant will display only zero entries for the file Test fpc simply because no calibration had been performed For any attempt to quantify unknown samples at least the line for fluorescence must be defined with the input of a spectrometer relevant number So far only the spectrometer was described by use of e g Test fpc one will be able to use calibration standards asr files to establish individual 63 instrumental constants By selection and averaging as outlined later the average instrument constant can be calculated which value will be entered
173. mples containing all possible elements The more correct solution is to edit the original asr files and save them under a different file name For the alloy example seven element entries need to be found Mn Fe Ni Cu Zn Sn and Pb E g the original Ni Std asr file SSPEC_ID Ni 100 10mA vc2073 spe SDATE_MEA 00 00 2000 00 00 00 SMEAS_TIM 25 1000 SPEAKS 1 28 i 53543 97 was edited and saved as Ni NBS asr SSPEC_ID pure Ni 10mA vc2073 spe SDATE_MEA 00 00 2000 00 00 00 SMEAS_TIM 25 1000 SPEAKS 7 25 26 28 29 30 50 82 The lines in red were added modified The lowest possible peak area accepted is 0 1 According to this also all NBS reference standards files except NBS1108 asr which contains all elements in measurable quantities were modified E g Mn has to be added into all files because although for the analysis of the synthetic bronze sample of no interest NBS 1108 as only standard contains this element SSPEC_ID NBS1103 10mA vc2107 spe SDATE_MEA 00 00 2000 00 00 00 SMEAS_TIM 1000 1000 SPEAKS 7 25 O41 26 6603 28 5056 29 f 1396881 30 1086374 50 3 3393 82 3 26401 After the delicate task to edit asr files the creation of calibration standard std files is initiated with the selection of Prepare standard concentrations file nd Par monochromatic excit NBS alpha coefficients Prepare standard concentrations file
174. n of unneeded lines e g escape peaks or an extension of the background region to facilitate the fit may improve the results In the high energy part of a spectrum the scatter peaks may sometimes cause a drastic increase in the background counts An exclusion of the Kg peak of the last and sometimes pre last element or the exclusion of the Lg and L lines of elements represented by L lines can facilitate the fit for the essential peaks Definitely one should not start the ROI at channel zero the first few channels of a spectrum are usually empty or contain trash nor should one mark the entire spectrum 14 1 2 10 What are the minimum requirements in order to initiate a fit of a spectrum Under X LINES at least one line has to be defined Change to another background model than the default smooth filter for the last version of QXAS When no ROI was selected manually the automatic ROI will be the default Select the COMMAND FIT then set the number of iterations e g N_ITER 20 and press lt enter gt The fitting should be completed after a few iterations 1 2 11 What are the criteria to be satisfied with a fit result The following criteria are to be checked in descending order as listed below Visual inspection by the experienced user Total Chi square ChiSquare less than 3 0 Residual COMMANDs DISPLAY RESIDUAL between 3 When there is no trend it is tolerable that a few points can be ouside of th
175. n spectra which are only indirectly or not at all caused by a fluorescing atom of a sample see example data Figure 1 20 Cr Std spe with Cr inp a Escape peaks for Si Li detector a parent peak e g Fe Ky with energy of 6 40 keV will be accompanied by a small escape peak with an energy of the parent peak minus 1 74 keV escape peak energy for Fe Kg at 4 66 keV The amplitude of an escape peak is usually much less than 1 of the parent For small parent peaks the escape peaks are swallowed by the background Escape peaks are included in the fit model with or b Sum peaks the sum peaks are observed at the double energy of the parent peak For intense lines there is a significant probability that two photons of the same energy are absorbed in the detector crystal within short time interval They are not collected as two individual signals but rather as a single pulse with double energy Amplitude of sum peak is a small fraction of the parent peak All possible sum peaks can be included in the fit model COMMANDs X LINES ADD with SUM keyword Their intensity can be influenced to some extent with a change of the Pile up resolution time default value 2 microseconds in the FORM Set detector characteristics c Scatter peaks the inclusion of the K line scatter region in the fit model is discussed in the next paragraph the utilization of this so gained extra information about samples is given in CHAPTER 7 Utilizati
176. nal standard element and the known concentration of the internal standard the sample element concentration values are determined detector E fluorescence radiation incidemt beam reflected beam reflector sampio Fig 9 1 Total reflection XRF geometry where a small sample residue of a liquid sample rests on top of a suited reflector The angle of incidence in respect to the reflector surface and the angle of the reflected beam are identical There exists also a refracted beam that penetrates into the sample but due to the low value of the transmission coefficient the resulting scattering by the reflector and the silicon fluorescence radiation are well under control Scattering will predominantly be caused by the residual therefore usually higher sample volumes will be counter productive for background reduction 135 ce vacuum air l At l Q7 Z N O 9 reflector Fig 9 2 Incident reflected and transmitted beam with respective intensities Io Ig and Ir Due to the special geometry not only the primary beam will excite the sample but also parts of this beam that impinge onto the reflector in front of the sample This results in double excitation detector a a sample reflector Fig 9 3 Double excitation of the sample by primary 1 and reflected beam 2 For standard TXRF a calibration standard will be a solution consisting of several elements with properly chosen concentration values in orde
177. named Backscatter Fundamental Parameter BFP is a very versatile METHOD for quantification and is suited for completely unknown samples thickness and dark matrix are calculated Source and X ray tube excitation can be covered The fundamental parameters approach calculates calibration constants from standards asr files and takes several theoretical values into account The fundamental parameters various absorption corrections the enhancement effect and the detector efficiency The sample self absorption by the dark matrix is corrected by the use of the scatter peaks As important feature of the program are the differential scattering cross sections used for the calculations The range of elements that can be analyzed by this METHOD is not limited or correlated to any range defined by the elements of the standards but when elements of standards represent unknown sample elements the individual calibration constant is used One constraint of the METHOD is the fact that only monochromatic excitation can be handled Even for more than one line of a source or secondary target the intensity weighted average energy must be calculated Pra Ekai t Pra Ee Kart Prei Exp Prp Erpa Prp Erg For a Rh secondary average target the weighted average of its Kg and Kg lines yields Eaverage 20 6 keV For direct X ray tube continuous excitation as primary radiation sometimes so called effective energies are defined Such a procedure might lea
178. nd K there is a clear tendency to lower instrument constant values with decreasing atomic number This is an indicator that against all efforts the spectrometer could not be described sufficiently for the low Z elements For this purpose individual instrument constants can be established for each element where a calibration standard is available In principle one can do this also for all elements which can be described by an average constant appropriately The two demonstration files Soil fpc and OrgaMatr fpc make use of the individual constants for the low Z elements For the remainder from Ca on the average instrumental constant established by metals and compounds together is used The two instrumental parameter files have individual constants for fluorescence for the low Z elements Al 2 70 10 Si 5 65 10 P 4 32 10 S 7 24 10 and K 8 575 10 68 The incorporation of individual instrument constants is preferably achieved by editing of the fpc file with a suited text editor CHAPTER 4 Editing of data files because for more than one entry per element e g for potassium there are four entries only the last is recognized by the program Care must be taken that individual constants are incorporated within one run all relevant asr files must be selected from the SELECTOR BOX because all old entries for individual instrument constants of a previous run are overwritten by a new run when Individual and or average is
179. nds flux Fig 6 1 Classification of samples as alloys oxides or samples prepared by the fusion technique In case the above named criteria can be fulfilled and after the proper definition of certain files the METHOD is easy to use and the results compare well with the other more complex fundamental parameters approaches 94 The analysis of samples should be performed in four stages Proper AXIL fit and adaptation of asr files Creation of standard files std Definition the a coefficients file alp Analysis of unknown samples Each of the elements of a sample must be represented by an element contained in at least one standard preferably more standards The standards asr files must be edited such that they become compatible with the asr files of the samples and the other standards Also pure element standards must contain intensity information at least a negligible small peak about the set of the other elements E g as in the later worked out example the sample does not contain manganese but one standard contains this element then also the other standards asr files must contain an entry with the otherwise unnecessary Mn information The standards file format expected by the program is obtained through the conversion of the asr to std file format with input of the concentration values Prepare standard concentration file The first step in order to generate the correct alpha coef
180. ned by the correct set of elements With the selection of the source description file Rh K sou which describes the excitation source the Rh secondary target and the sample geometry the O coefficient file NBS amp pure alp was created element set Mn Fe Ni Cu Zn Sn and Pb For the preliminary quantification pure Ele arp will do Cu Zn Sn and Pb for its creation any of the std files _Cu NBS _Zn NBS _Sn NBS _Pb NBS and the source description file Rh K sou are to be loaded 100 6 5 Quantitative analysis of a synthetic bronze sample Let us assume we do not know the composition of the sample but from the inspection of the spectrum SynBronz spe it is evident that the elements Cu Zn Sn and Pb are to be taken into account Select concentration calculation mode Calculate using standards Calculate using intensity calib coefs Calculate using sensitivity calibration Fig 6 8 The first stage approach to quantify unknowns is to use calibration standards With the preliminary oa coefficient file pure Ele arp and the preliminary calibration standards _Cu NBS std _Zn NBS std _Sn NBS std and _Pb NBS std intermediate quantitative results can be obtained From the SELECTOR BOX Select the function only the first option should be considered Select the function X a al k aix a2 K K a al K a2 k K Fig 6 9 Four polynomials are available to describe the correlation between the relative intensit
181. not Soil fpc or OrgaMatr fpc otherwise their individual constants as needed later were overwritten Due to the targeted creation of individual instrumental constants it is not necessary to have the average instrumental constant defined for this file As outline in the previous only one standard can be used per element for more than on standard per element the relevant fpc must be edited manually for incorporating the averaged value entry as detailed in CHAPTER 4 Editing of data files The pure element calibration standards asr files Fe Std Ni Std Cu Std Zn Std Sn Std and Pb LStd and MnO2 asr compound mixed with binder were selected to contribute All of them must be processed in one run Select unknown or standard files ASR C QRAS DEMON AS R STDS TEMP ZN STD ASR MNO2 ASR FE STD ASR PB LSTD ASR SN LSTD ASR CU STD ASR Fig 3 31 Selection of calibration standards in order to establish individual instrumental constants that will be written to the fpc file With the exception of the FORM Type of Instrument constant where one will toggle from the default to Individual and or average all steps for establishing the individual instrumental constants are in analogy with Calibration standard example K KBr asr 72 Type of Instrument Constant Instrument constant type BERET Uy er 7 Fig 3 32 After having selected as Instrument constant type Individual and or average calibration standards will gen
182. not correct the enhancement effect Therefore they had not been included in the calibration compound cal 53 4000 KBr 3500 K and S in compounds asummed error of 5 in the s concentrations 3000 o KCr 0 x 5 2 2500 KH PO lt 2000 a 2 KCO 1500 MgSO Fa 2 S 1000 4 pure B sulphur 500 0 TEE CORE ae Ee ee Ne ee Le OR eee ON ee Mee ON ee Oe ee eh eee a eae ee ee ON ekg a ee NS Re Og 0 10 20 30 40 50 60 70 80 90 100 concentration Fig 3 5 Plot of the sensitivity versus element concentration for selected calibration standards containing sulphur red dots and potassium blue dots The compounds KBr and K2Cr207 are only displayed for comparison but may not be used for this method because of the secondary excitation of the elements of interest For the case when some elements in unknown sample can not be covered by a calibration point one will proceed As provided by QXAS an extension of the element range can be achieved with Optimize Calibration with Least Square Fit The use of this option is very problematic because otherwise well defined calibration points can have a systematic bias afterwards At least with this procedure an interpolated data point for e g the element Rb can be established as needed later Polynomial fit of Ka sensitivities 23 sensitivities found elements S K K Ca Ti UV Cr Mn Fe Co Ni Cu Zn Ge As Se Br Sr Y Zr Nb Mo Order of polynomial no fit gt
183. nsity of the first standard must be entered after the line Enter the sample mass per unit area for thick standards any high number will do e g 100 In the next line for the K2CO part of the calibration standard with a concentration of 81 63 input K2CO3 81 63 Consecutively enter C38H76N202 18 37 for the binder The number of times a line with the need of an input for an element compound and its concentration will show up corresponds to the previously entered integer No of sample components lt 1 50 gt 119 Sample no 1 No of sample components lt 1 5 gt 2 aa the sample mass per unit arealg cm2 Compound element amp its weight fraction wt in the sample e g CaCO3 25 5 K2C03 81 63 K2C03 is composed of wt x 56 582 8 689 34 729 Compound element amp its weight fraction wt in the sample e g CaCO3 25 5 C38H76N202 18 37 Fig 7 14 The calibration standard KCO must be described concerning the compound of interest and the binder used for pelletizing The next FORM Enter sample filename will need the asr file name for the calibration standard matching the defined elements compounds as specified above input K2CO3 There is no immediate check for consistency The consecutive FORM refers to the tube current as used for the measurement of this standard input 40 The question Scatter peak intensity from other ASR file will be asked for fluorescence calibration standard files bec
184. ntal parameters the absorption correction the inter element effect and the detector efficiency The sample self absorption by the dark matrix can be corrected by the use of the scatter peaks Standards are used to determine the geometry factor average instrumental constant In theory the range of elements that can be analyzed by this METHOD is not limited or correlated to any range defined by the elements of the standards In principle with only one standard the METHOD can be calibrated As first step in order to run the METHOD correctly the fundamental parameter file fpc has to be defined Hint For problems during the final calculations of instrument constants or sample concentrations When after a short flash the calculation will stop and bring back to the previous COMMAND Calculations of Geometry constants Analysis of unknown samp the conventional memory did not suffice Terminate QXAS completely and re start it go again to the last COMMAND Calculations of Geometry constants Analysis of unknown samp the temporary file fundp tmp still carries all data and initiate the calculations again 3 3 1 Set up of the instrumental parameter file fpc The excitation conditions geometry detector characteristics and other influence parameters are to be defined or the pre defined files Test fpc OrgaMatr fpc Soil fpc can be inspected with Set up instrumental parameters Full Fundamental Parameters Set up instrumental parameters Specify
185. ny other combination of fit parameters has worse results The weighted fit option could be used because all standards were treated properly for their uncertainties in the AXIL fit standard deviation greater than the square root of the peak area and realistic uncertainties of the concentration values relative error 0 1 were entered With this procedure an interpolated data point for e g Rb with 7 36 0 0884 10 could be established 57 alibration file C _T SEPTNETNMMETALS CAL reated on 09 29 2006 Calibration date 09 29 2006 46 Atomic number lt Esc gt continue lt F9 gt LOG Plot lt F10 gt Printer Plot Fig 3 11 Polynomial representation of sensitivities fitted as a function of the atomic number All used calibration standards are pure elements and are K line emitters For the L lines between Cd and Pb seven calibration points a logarithmic polynomial of order 3 weighted option yes was found to suffice scroll down in the SCROLL BOX to find the L results Mean Diff coming after the K results alibration file C _T SEPTNETNMMETALS CAL reated on 09 29 2006 Calibration date 10 03 2006 80 Atomic number lt Esc gt continue lt F9 gt LOG Plot lt F10 gt Printer Plot Fig 3 12 Polynomial representation of sensitivities fitted as a function of the atomic number All used calibration standards are pure elements and are L line emitters Conclusion By intuition one could deduce from the compari
186. o get an energy calibration for a spectrum measured with my own spectrometer There exist spectra converted from certain multichannel analyzer s MCA spectrum formats that contain already a correct energy calibration intrinsically Usually this will not be the case When an appropriate model inp file can be loaded which is the case for all example spectra also no energy calibration step is necessary The energy calibration is defined as the correlation between the channel number i of a spectrum obtained from the MCA with the contents for each channel in counts per channel usually referred to as intensity and the energy in eV or keV E ZERO GAIN i 1 1 In praxis it means to know the energies corresponding to at least 2 positions channel numbers in the spectrum and transfer this knowledge to AXIL Each chemical element measurable by XRF shows characteristic lines seen as peaks in X ray spectrum Using a simple standard or sample at least two correlation pairs i E must be available where 7 denotes the channel number corresponding to the position of a recognized X ray peak Er denotes the tabulated energy of that X ray peak As a rule of thumb well defined Ka and or Ly peaks should be used to energy calibrate the spectrum To help the identification of peaks present in the example X ray spectra used through out this guide the spectra files carry self describing names related to the sample composition Not
187. obtained It is called a multichannel spectrum The number of counts in every channel of the spectrum is proportional to a number of photons which deposited in the detector the energy vide the energy calibration equation 1 1 covered by this channel In a more advanced approach the output signal of the detector preamplifier is digitized directly with a minimum analog shaping the main shaping and pulse filtering in performed in digital domain by a digital signal processing DSP circuitry For high count rates there is an increasing probability that two photons of a very intense line are absorbed in the detector crystal within such a short time interval the sweeping time needed to collect the free charges that their charges are not collected as two individual signals with a certain energy but rather as a single one with twice the energy sum peak All possible combinations of sum events can be included into the AXIL fit model with the SUM keyword COMMANDs X LINES ADD SUM In an ideal detector the charge is collected completely and the response of the system would be a single peak containing only counts of a certain energy and no other counts elsewhere in the spectrum In practice some artefacts are observed incomplete charge collection resulting in a tail to the low energy side of a principal peak low energy tail depends on the design of the crystal The escape of photons or secondary particles from the detector surface is most pro
188. of the strong Cu and Zn peaks has too much impact on the Sn L and Pb Lz lines evaluation With a high value of the background model the entire region could be described but in this case background oscillations will cause unrealistically differences for the weak peaks as a function of the value PARAM For the model file NBS1103 inp the peaks for the Cu Zn region can be described with X LINES ADD FE NI CU KA CU KB ZN KA ZN KB 9 185 Unfortunately Ni and Fe hardly can be split off because of the Cu and Zn escape peaks overlaps Because the Pb L intensity is high the otherwise small Pb L peak with an energy of 9 185 keV is taken into account by its energy value The model file N1103 Sn inp describes the Sn L region with X LINES ADD SN L and a linear background of order 1 Param The file N1103 Pb inp describes the narrow region around Pb La ROI begin 910 end 1020 sandwiched by the Zn Kg and Pb Lg peaks X LINES ADD PB LA There is no need to include the other Pb L lines nor does one have to care for the other lines originating from the L3 shell transition defined by PB LA which are outside of the ROI Note Description of reference standard NBS 1103 and its spectrum acquisition conditions along with names of the three input files necessary for an adequate AXIL fit resulting in NBS1103 asr can be found in the EXCEL nbs_srm xls file As can be seen from Figure 6 2 it is possible to fit the entire region for the inter
189. of the two target approaches can handle their absorption correction Ti by target T1 Zn Rb Y Nb and Pb results fall within the 95 confidence interval as given by the certificate That the titanium result for target Tmetal falls out of the range is due to the rejection of Ti as target element therefore the working range starts only from sample element chromium Mn Fe Sr and Zr are overestimated Rubidium is a positive surprise because it was only calibrated by interpolation Arsenic although having the interference with Pb La is well specified The trend for Fe and Sr is due to the blank problem with the two elements the used standard for Mn MnOz is not well suited for calibration as already mentioned earlier Copper was not even taken into account because of its blank problem Only slight differences can be found for the two target approaches 89 Table 5 2 Certified and measured concentration values for IAEA Soil 7 established by two different target approaches target T1 target Tmetal Concentration values within the certified C I are coloured red The respective slope regression coefficient and standard deviation for the absorption correction calculations are to the bottom lines certified values concentration 95 C I Target T1 Target Tmetal Si 18 16 9 20 1 50 79 K 1 21 1 13 1 27 2 37 2 97 Ca 16 3 15 7 17 4 27 28 33 1 Ti ug g 3000 2600 3700 3313 3959 Mn ug g 631 604 650 828 850 Fe 2 57 2 52 2 6
190. ollowing samples should be rejected because The slope for Lichen0O asr is 4 675 as to be compared to the expected value of 2 8 The regression coefficients for LichenO asr Lichenl asr and Lichen1B asr are 0 976 0 958 0 966 respectively recommended better than 0 99 The standard deviations for LichenO asr Lichenl asr LichenlB asr are 0 306 0 320 0 279 respectively which is a factor of roughly 3 5 higher than for the other samples When working with several targets like for the comparison of the lichen reference standard samples one must respect that inhomogeneities may be found even when the mother mixture was thoroughly homogenized and the pellets were prepared identical 153 element residual for targets B C D E and F Gres Target B Gres Target C E res Target D E res Target E E res Target F fluorescing constituents of targets Fig 10 7 Residual for the fluorescing elements contained in five targets used for the ET method In the ideal case all results were found within the range 3 For target E a systematic deviation is observed For other targets inhomogeneities for selected elements can be identified For the five targets B F the residual was constructed from the asr file results peak area for the respective elements by calculation of the element average and individuals subtraction divided by the standard deviation The fluctuations should not exceed
191. om the dye X LINES ADD CU W SUM 91 Note Description of the targets used in combination with the lichen samples and the input model details to generate the target and target plus sample asr files can be found in the EXCEL lichentargets_desc x s file The target spectra were fitted with split Ka and Kg lines and peak shape correction X LINES ADD V KA V KB CO KA CO KB CU KA CU KB SE KA SE KB For Sr and Mo there exists no peak shape correction so they were included with X LINES ADD SR KA SR KB Mo KA SUM Lg E Lg Lg A Fact 4 950 0 717 2 266 6 925 0 277 1 423 8 041 0 737 1 267 11 208 1 784 1 100 14 142 2 244 1 051 17 443 2 835 1 028 slope 2 821 intercept 5 175 regress 0 997 _ __ stdev 0 087 1 80 2 00 2 26 2 46 2 60 2 80 Log E Accept ysn y Enter Element to start iter Fig 5 10 Absorption correction calculation results obtained in connection with the lichen sample spectrum Lichen2B spe The values for the elements which are certified agree fairly well for K Mn although the Mn calibration point is questionable Fe Zn Br Rb Sr Pb But the element potassium must not be included for a report therefore no average value because it is out of the range spanned by the targets This is because the first target element V is already too high in terms of energy For the 0 4 and 0 5 g samples Lichen4 spe and Lichen5 spe V was rejected from the abs
192. on necessary all elements of interest are represented by calibration points Before the AXIL result file AirFilt asr was used for quantification with Elemental sensitivities the element entries of the lines containg Ni Cu Sr Mo and Pb peak areas had been erased and the edited file was saved as AirF cor asr for editing of files see CHAPTER 4 Editing of data files For the use with the above defined calibration file AirFilt cal this procedure were not necessary but interested readers might also use other calibration files which would enable to generate results for elements that should not be reported The tube current for the measurement was 40 mA as Sample type thin sample and as concentration units mass per unit area are applicable Sample C QXRASDEMO ELSENS AI RF COR ASR Measurement date 16 27 2606 Live time 1600 sec Tube current 46 666 mf Method is Elemental Sensitivities Thin sample Analysed elements El counts Si 166 Ca 153 Fe 1916 Zn 274 compound 6 Si 5 ug cm 2 Ka Ka Ka Ka ug cem 2 14 ug cm 2 6 ug cm 2 I I I I Fig 3 13 Quantitative results for the reference standard material SRM 2783 air filter obtained with the method Elemental sensitivities The reference certificate lists the concentration values in ng filter By use of the average area 9 96 cm as specified there one can convert the QXAS data in order to compare the results The standard deviation in the concentration value
193. on of the scatter peaks For the Rh secondary target excitation also the Rh L lines are scattered They easily can be confused with Cl K lines In this particular case it is recommended that energy values are added to the fit model with X LINES ADD 2 7 2 8 2 9 This unusual way to add peaks will suppress Rh L results in asr files and therefore also suppress unwanted concentration values in output reports for the element Rh usually not present in a sample 16 Cr Std spe 10000 4 eas 5 41keV ch 506 Pure Cr metal n Ag X ray tube Rh sec target 50kV 40mA 16s 1000 Cr KB 1 S 9 K V ch 556 Elastic scattering Rh Ko j 20 17keV ch 1886 I il 100 4 fi Sum Cr Ka Cr Ka Compton scatter Rh Ka M 1 Escape Cr Ka 10 82keV ch 1010 19 40keV ch 1815 3 67keV ch 340 i h 1 Ma sum cr Koucr KB i 4 Pa i Hh h ih 11 36keV ch 1063 yr I p Wi NAN et wy io li M M e Ie f if iy Hi i ill 0 0 10 0 15 0 20 0 energy keV Fig 1 20 Spectrum plot as obtained by use of the DMP file and further processing with EXCEL Cr Std spe d Diffraction peaks are observed for crystalline samples In the example shown in Figure 1 21 in the calibration standard spectrum KBr spe a peak at 9 75 keV was found input model to be loaded KBr inp Spectrum KBR SPE Iteration 11 ChiSquare pr Display GON BEG beg chan END end chan MIN min cnts MAX max cnts
194. on standards the average instrumental constant for fluorescence and how to define also individual instrumental constants Within the SELECTOR BOX Select Calculation Mode toggle by use of the lt space bar gt to Instru Constants for fluorescence otherwise with the default option Sample concentrations in the following the attempt will be made to treat the standard as unknown sample resulting in a crash of the program because of the so far not established and therefore missing average instrument constant electe Calculation Mode Instru constants for fluorecence Fig 3 22 In order to obtain the needed instrumental constants for fluorescence as mode Instru constants for fluorescence must be selected The consecutive SELECTOR BOX will offer a choice Average and or Individual that should not be taken for a first round calibration but will be needed for the second calibration round As can not be known a priori potassium will need to be described by an individual instrumental constant Instrument constant type Average For the beginning all calibration standards should only contribute to the average After all standards results are available one will decide to exclude some of them or enable individual element specific instrumental constants Potassium is already an example for the light elements that will be included individually 64 Type of Instrument Constant Instrument constant type iC TST Fig 3 23 In a firs
195. opriately with this METHOD None of these problems applies to air filter samples aerosols deposited as thin film on a suited support The major problem with thin samples is the usually low count rate for the peaks of interest The fit with AXIL will not be problematic but utmost attention must be paid to the instrument blank and the blank filter material i e the substrate where the aerosols are collected onto The following elements were identified in the spectrum of AirFilt spe standard reference material SRM 2783 batch 267 Si Ca Fe Ni Cu Zn Sr Mo and Pb The scattered Rh L lines have to be included for the sake of Si and are preferably described by energy values the elements had been included with X LINES ADD SI CA FE NI CU ZN SR MO KA PB 2 7 2 8 2 9 The input model AirFilt inp describes the background as linear of order 8 because of the strong curvature caused by Rh K scatter region extending to Mo Sulphur must not be included because of a potential interference with the Rh L scatter peak The same input model is used to fit also the blank filter SRM 2783 blank spectrum Blank Af spe and the instrument blank spectrum Instr Bl spe see Figure 1 27 In the instrument blank spectrum the elements Fe Cu Sr and Mo can be identified A molybdenum foil had been used for shielding purposes Fe and Cu are of unidentified origin and Sr very likely will originate from a crumb lost from a previously measured sample Both bl
196. or Most of the results fall within the certified C I or are close to it red color Conclusion There is an optimum sample thickness in respect to the absorption correction Unfortunately it will be different from sample type to sample type When a sample matrix shows too high absorption and the pellets can not be pressed thin enough it will not be possible to apply this METHOD at all or only to the elements bracketed by target elements accepted for absorption correction 93 CHAPTER 6 USE OF NBS ALPHA COEFFICIENTS References 29 30 31 The NBS alpha coefficients METHOD is restricted to thick samples containing only elements of relatively high atomic number Z metallic or oxides There is a detectable fluorescence signal from every constituent except oxygen samples with dark matrix can not be treated The fundamental parameters approach corrects for the self absorption and the enhancement effect The program was developed by the National Bureau of Standards now NIST therefore the name for this METHOD NBS and the Geological Survey of Canada for wavelength dispersive spectrometers The algorithm is called COLA COmprehensive Lachance Algorithm in which the influence coefficients are calculated only once despite the concentration range of unknown analytes to be considered The alpha coefficient file neither depends on the concentration values of the analytes nor on intensities It was adapted by the IAEA for energy dispersive
197. or small peaks The last criterion as used for the standards does not apply for samples Instead use for small peaks the two criteria A negative peak area or a peak area less than three times the standard deviation should result in the exclusion of the element The peak area must exceed three times the square root of the background otherwise exclude the element Applying these two criteria arsenic and niobium survived pplying Blank problems Copper was out of doubt identified but its count rate is only 0 2 counts second which is far less than 10 times the blank value Fe and Sr peak areas might lead to systematically elevated results but are well above the threshhold After having saved the fit results it is mandatory to save also the fit model 10 4 Elemental Sensitivities METHOD Under the aspect of QA the METHOD Elemental Sensitivities must not be applied when there is an input needed for the dark matrix of any sample Compound by difference except when it can be specified very precisely e g the dark matrix was specified by another METHOD element by element Chemical elements of calibration standards which are excited by the inter element effect must not be included into the calibration e g not potassium of KBr not phosphorus of KH2PO not potassium of K Cr207 The calibration for this METHOD is limited to 25 standards in total All elements which are found in unknown samples must be rep
198. order Energy calibration zero 26 21 eU gain 12 53 ev ch Resolution calibr noise 120 0 eV fano factor 6 114 Region of interest 218 1472 Arrows move selector box lt CR gt select item lt ESC gt exit Fig 1 8 A brief overview is available for certain parameters contained in the temporary valid input model file Target1 inp Next select Analyse Spectra and consecutively use the COMMANDs LOAD and DIR_SEL Axil LOAD STOP BATCH GO CANCEL Fig 1 9 COMMAND LOAD to retrieve an already converted measured spectrum Load SPEC filename DIR_SEL GO CANCEL Fig 1 10 Two options are available to load a spectrum by direct input of the known file name from the correct directory or by browsing through the various directories and files to load the file intended for this demonstration from the directory QXASdemo GetStart with the selection of the spectrum Targetl spe from the SCROLL BOX Select spectrum to fit file spe gt E QRAS DEMO GETSTART ao TARGET1 SPE PB PURE SPE E CALIB SPE Fig 1 11 SCROLL BOX for loading the spectrum T1 spe Because the input model file Targetl inp was loaded before the spectrum file the energy calibration is already correctly defined The sample which served to collect the spectrum Targetl spe contains the elements V Co Cu Se Sr Mo and elements that are not visible by XRF H C N O In order to verify the predictions select the COM
199. orescence radiation and the mass fraction of the chemical element holds only for infinitely thin layers of atoms For samples of finite thickness the intensity is affected by self absorption A and the secondary excitation effect H In general case the intensity is a non linear function of the mass fractions or concentrations of all elements present in the sample L I le C 3 C garkmatrix 2 37 The sample self absorption is treated as correction factor and is with a few exceptions of utmost importance The secondary excitation also named enhancement or inter element effect can be also treated mathematically but it is of importance for certain samples only 2 1 1 Sample self absorption The mass attenuation coefficient of an element Jn summarizes all possible photon interactions with atoms of that element and it is a function of the incident photon energy E Hn Hp E 2 38 For the photon energies below the pair production edge lt 1022 keV there are tree types of processes leading to photon absorption photoelectric effect coherent and incoherent scattering The mass attenuation coefficient is expressed accordingly Mn Tn Ooo tO 2 39 coh inc Where O p and o are the contributions due to coherent elastic or Rayleigh and incoherent inelastic or Compton scattering expressed in cm g The attenuation of a parallel beam of photons by a layer of atoms of thickness t cm and density p g cm
200. orption correction calculation Consequently one should not make use of the results for the elements below the next target element Co i e Mn Fe their entries were not used for the average value calculation The samples are already too thick For the lichen pellet with 0 05 g Lichen0 spe Mo and Sr of the target elements had been rejected because their characteristic line intensities where higher for sample plus target than for the target only Samples can also be too thin for this METHOD It is evident from the results that copper must be excluded from any report because of its contamination problem with the used setup Note Comparison of selected concentration values as obtained through the measurements of several samples prepared from IAEA 336 lichen reference standard material and quantification with the Emission Transmission method with the certificate can be found in the EXCEL lichen_results xls file Individual concentration values in blue color are acceptable 92 Potassium not contained within the range spanned by the target elements and copper blank problems values have to be disregarded completely Bad slope regression and standard deviation values are held in green color Due to the rejection of target elements during the absorption correction calculation certain concentration values are also disregarded Disregarded individual values are not used for the establishment of the average value and standard deviation red col
201. os that are kept fixed during the fit are affected Examples are the inclusion of escape peaks and L line ratios within one group 147 Set sample absorption Sample thickness g cm 2 gt Sample composition element amount element amount Ph 1 98808 86606 8688 s5 Fig 10 5 Definition of the sample standard matrix in order to improve the pre defined line ratios 10 2 2 Spectrum fit with AXIL Energy calibration Usually best with the Ka La peak of the major constituent of the standard and the coherent Ky scatter peak where not visible use the Kg Lg peak of the element of which the first peak already was used for energy calibration instead Example for CaCO3 spe with the COMMAND CALIB the Ca K peak maximum was found at channel number 340 and was identified with lt Fl gt CA lt Space bar gt channel number 1846 was identified with lt Fl gt RH lt Enter gt as the Rh Kg elastic scatter peak Add only the peaks COMMAND X LINES ADD that will result in the information needed K or La Add also all visible lines within the ROI which will be defined in the next step E g most of the times the Kg or Lg and Ly peak although not carrier of relevant information will overlap with the Kg La peak Hint The automatic region of interest can be displayed only after a fit with the COMMAND DISPLAY ROI GO Split the elements into Kg and Kg for K lines and L3 L2 L1 for L lines also instead LA LB LG Usu
202. ould not be displayed at all 1 3 8 For a standard it is not permissible to define a dark matrix 1 therefore it is irrelevant if it is known or not 3 and the sum of all concentrations is normalized 8 for standards without exception For 2 Elements exist as the option compounds is rather a nice feature for samples when concentrations are displayed for both the elements and the compounds simultaneously For samples prepared by use of a binder one would love to be able to use an entry for 4 but any input in order to correct for the dilution factor will crash the program 5 does not apply to any of the standards used for this calibration the scatter peak calibration will be discussed in CHAPTER 8 Utilization of the scatter peaks The default for 6 Secondary excitation enhancement should not be altered particularly not for this calibration standard because bromine contributes to the potassium signal substantially by secondary excitation 7 is of utmost importance because by toggling to There are known compositions all other chemical elements of the compound and the binder can and must be defined Whether one wants to save the final report or not 9 can still be decided when the results are displayed 65 Sample contains no matrix Elements exist as compounds Matrix composition is not known Dilution material is not used Scatter peaks Cif used gt are from the same spectrum Secondary enhancement is corrected iT kno
203. peaks Note In case one tries the background subtraction COMMANDs SCAT_ROI for this scatter peak modelling approach before at least one element peak free of choice has to be fitted otherwise AXIL will crash As example the spectrum HWC spe can be treated with the input model HWC 3 inp The incoherent Rh Kg region is set within channels 1250 to 1610 The coherent region is set between channels 1650 to 1670 No background subtraction was attempted The resulting file was renamed to HWC 3 asr For the further use with the quantitative METHODs the asr file must be edited and the scatter peak information brought to standard see CHAPTER 4 Editing of data files 107 Spectrum HWC SPE HWC binder 50 40 1000 vc2401 spe Axil LOAD STOP DISPLAY ROI CALIB X LINES KLM MARK FIT REPORT SAVE_RES PLOT BATCH BACKGRND SCAT_ROI Cc o u n t S Z cC h a n n e 1 1566 2000 Channel Number G0 CANCEL Fig 7 3 Summation of the channel contents for the Compton and the elastic Rh Ka scatter peaks model 3 Conclusion One can find other approaches to describe the scatter peaks within AXIL none of the known ones is completely satisfying because either the background subtraction can not be controlled sufficiently and or the deconvolution of elastic and inelastic peaks is not correct Each of the models will result in different peak areas for the scatter peaks For what so ever model one has preferences for scatter
204. ple mass as variation parameter Its overall absorption is much lower as compared to e g Soil 7 For Soil 7 the dynamic range in terms of thickness is very limited simply because intermediate thick pellets in respect to XRF are already very thin when it comes to the realization of pressing pellets for too thick pellets the METHOD will not work any longer Of the lichen material eight pellets had been pressed without addition of a binder The weight of the pellets spans a range from 0 05 0 5 g corresponding to 0 01 0 1 g cm diameter of the specimens 2 5 cm the weight for pressing was set to 5 tons The lightest pellets had been already so thin that they could be used only once For each sample absorption specification a specific target had been used for the measurement with and without sample in front Although the mixture for all targets was the same with exception of target A which does not contain Sr slight inhomogeneities can be observed even front and back side of a target might show slight differences All measurements were taken with 40 mA tube current The samples were measured for 1000 s the targets and target plus sample for 200s respectively The air path between sample and detector was 0 5 cm A calibration file Lichen cal was created with the METHOD Elemental sensitivities Eighteen calibration standards asr files were included K2CO3 K KH2PO4 CaCO3 Ti Std Ti02 MnO2 Fe Std Fe203 Cu Std CuO Zn Std ZnO Br K SrC
205. put model CaZnSrPb inp This model should be used for batch fitting the 41 spectra The sequence of commands to perform the batch fitting has been stored in the file FitAll bat Its structure is shown below load spec 01 spe fit n 10 save load spec 02 spe fit n 10 save etc load spec 40 spe fit n 10 save load spec 41 spe fit n 10 save For each file there are 3 commands executed 1 to load the spectrum 2 to perform the fitting and 3 to store the results of the fit Before starting the batch fitting a proper input model file should be loaded in this case CaZnSrPb inp Next with the COMMAND 29 BATCH one enters the full name of the file containing the batch fitting commands fitall bat without the apostrophes The batch mode is invoked spectrum after spectrum is fitted After fitting all the 41 spectra the batch mode is finished The last spectrum 41 spe is displayed in interactive mode The results of fitting of all the spectra are stored in corresponding asr files as Ol asr 41 asr There can be cases when the situation gets more complicated During a long lasting scan measuring points might be analysed that do not contain the usual composition of elements Instead peaks of other elements might appear imagine the case when the sample holder is hit If peaks of such elements are not included in the fitting model it might occur that due to mismatch of the fitting model and the data some of th
206. quality of the absorption correction Sample element range equivalent to the energy range of characteristic lines spanned by the first and the last element of the accepted target elements Sufficient target elements for acceptance by the absorption correction calculation routine the minimum is 3 to uniformly define the named range Slope expected value around 2 8 Regression coefficient regress better than 0 99 Standard deviation stdev low Example A The reference standard IAEA Soil 7 was analyzed by this METHOD with the use of a target composed of V Co Cu Se Sr Mo Consequently the results for elements named in the analysis report for Si K Ca must be disregarded the results for the element Ti can still be used because of its proximity to the established value of the target element vanadium Example B The reference standard I AEA 336 Lichen was studied with the total weight of the intermediate thick pellets as parameter For the lightest sample with the weight of 0 0497 g the target lines of Sr and Mo were rejected consequently the element range for this sample is narrowed to the characteristic energies between 4 5 12 keV excluding elements like K Rb and Sr from the analysis report For the thickest samples with 0 4039 g and 0 5007 g the target element V was rejected so the characteristic energies of the sample must range within 6 9 17 5 keV excluding K Mn and Fe from the analysis report The f
207. r energy dispersive X ray fluorescence analysis of environmental samples with coherent incoherent scattered X rays Anal Chem 49 1977 641 ARAUJO M F VAN ESPEN P VAN GRIEKEN R Determination of sample thickness via scattered radiation in XRF with filtered continuum radiation X ray spectrum 19 1990 29 TANG S M KUMP P YAP C T BILAL M G Calculation of relative X ray fluorescence intensity for annular source geometry by the Monte Carlo method X ray spectrometry 15 1986 289 BERNASCONI G BAMFORD S A DOSAN B HASELBERGER N MARKOWICZ A MAHMOUD A VALKOVIC V Applicability of annular source excited systems in quantitative X ray fluorescence analysis X ray spectrometry 23 1994 65 159 39 40 41 42 43 44 45 46 47 48 49 50 51 160 AVALDI L BUI C MILAZZO M The problem of irradiation and detection angles in quantitative XRF analysis X ray spectrometry 14 1985 159 LABRECQUE J J ROSALES P A MEJIAS G A study of the source sample detector geometries for an annular Am 241 source and a planar pure Ge detector for flat disc samples J Radioanal Nucl Chem 120 1988 13 KREGSAMER P Fundamentals of total reflection XRF Specrochim Acta Part B At Spectrosc 46 1991 1333 KREGSAMER P STRELI C WOBRAUSCHEK P Total reflection X ray fluorescence Handbook of X ray Spectrometry 2 edn revised
208. r to cover the elements as well as their concentration range as expected in the sample and an extra element the internal standard It must not interfere with the spectral lines of the calibration standard nor must it be contained in the samples The concentrations of such a set of elements will vary from one standard to the other whereas the concentration of the internal standard element usually will be kept fixed for all of them and is preferably close to the amount added to the unknown samples 136 counts channel Mo K scatter 0 5 10 15 20 energy keV Fig 9 4 A typical spectrum of a calibration standard for TXRF analysis All specified elements are present with equal amounts It should be mentioned that not only residues of liquid samples can be investigated by this technique but also the reflecting surface itself for contaminations A practical application is the search for traces on and in the plane and polished surface of silicon wafers as manufactured for the semiconductor industry By the controlled change of the incidence angle for total reflection a fluorescence signal originating from a residue can be distinguished from a thin layer and from homogenously distributed atoms embedded in the bulk material Shallow implantations in bulk material can also be investigated for their depth profile e Sc me Si bui Ni layer ee T ef e intensity a u angle phi phi Fig 9 5 It is possible by variation of t
209. ra are stored in the directory QXASdemo SPE Stds their asr files are found in the directory QXASdemo ASR Stds The user can create asr files having tried different parameters than recommended without danger of overwriting example files The spectra are accompanied by inp files already optimised to fulfil all AXIL fit criteria All inp files are found in the directory QXASdemo INP Stds A detailed description of all standards related data can be found in the EXCEL standards xls file Due to the fact that input parameters of a previous fit are used as starting values for the next fit the user might encounter slight differences when trying to repeat a fit with the demonstration files even when using the recommended input files All example standards spectra for Rh secondary target excitation had been collected with the intention to have approximately 50 000 counts in the peaks of relevance for calibration All spectra had been collected with 50 kV high voltage applied to the Ag X ray tube exciting the secondary target As an example of a non linear influence measuring parameter the tube voltage value may never be altered like many other parameters as the secondary target vacuum conditions filters geometry etc for a specific calibration and consequent unknown samples evaluations with this calibration The tube current was varied between the generator X ray tube limits of 5 40 mA it has in the ideal case linear influence with the
210. reasonable 52 Data for standard C _I SEPT ASR STDS CR2 K207 ASR Concentrations for ana ed elements oncentrations of other elements conc zstddev Compound conc 28 86000 6 1608 K a 21 76668 H Mu 2 370000 G u 14 14000 N 0 A 8666068 Fig 3 4 Definition of the composition of the calibration standard K2Cr207 with added binder After the correct inclusion of all standards the calibration will continue with the calculation of the sensitivities for all elements found in the standards Perform calibration When all elements identified in unknown samples can be covered with a calibration point standard containing the respective element this will be the end of the calibration procedure For the calibration points sulphur and potassium a plot of the sensitivities as function of concentration with the respective compound as parameter reveals some problematic behaviour see Figure 3 5 Although an unrealistically high relative standard deviation for the concentrations of 5 was assumed one can easily see that neither sulphur and the compound MgSOu nor the compounds KH2PO and K2COs respectively have matching sensitivities In the ideal case the sensitivities of standards for the same element calibration point would have the same values independent of their compound and the concentration The sensitivities of potassium in the compounds KBr and KCrO are also displayed in this graph although this METHOD can
211. reated on 11 21 2006 Calibration date 11 21 2006 46 Atomic number Fig 8 10 Calibration for K line emitters for Cd 109 excitation with the method Elemental sensitivities Five standards are available for L calibration Hf Ta W Au and Pb Calibration file C 11_NOUNQXASDEMONCD 109 F inal cal Created on 11 21 2006 Calibration date 11 21 2006 mE DUELI 80 Atomic number Fig 8 11 Calibration for L line emitters for Cd 109 excitation with the method Elemental sensitivities 8 2 2 Full fundamental parameters With the Full fundamental parameters METHOD a calibration established by all available standards can be performed The program is full of surprises for the source excitation with Cd 109 The source type is selected with Radioisotope excitation from the FORM Excitation conditions in the line Mode In the consecutive FORM Parameters for Radio 132 isotope the energy of the emitted Ag Kg line is defined wrongly the energy of 26 10000 Cd Kg must be edited to the correct value of 24 9870 For each editing run of the fpe file it will adopt the wrong default each time the user has to enter the correct value Instrument Parameter File E QRASDEMO CD 169 CD 169 FPC Parameters for Radio isotope Isotope name Cd 169 Half life days gt 462 6600 Energy CkeU gt Probability 22 10300 82 76998 24 98700 17 23000 88 63416 3 746686 4 666066 4 666068 Fig 8 12 Correction to the wrong energy
212. resented by standards For the case when some elements in unknown sample can not be covered by a calibration point one will proceed with Optimize Calibration with Least Square Fit The use of this option is very problematic because otherwise well defined calibration points can have a systematic bias afterwards With 152 this procedure interpolated missing data points can be established After having noted the desired sensitivities the calibration has to be re run with order 0 for the polynomial Order of polynomial O no fit 0 The calibration points are not fitted at all and the original calibration points are not biased Outside of the QXAS environment the relevant data is pasted into the calibration file The interpolated elements are worse defined than the surrounding elements but the rest of the elements do not suffer from a systematic error 10 5 Emission Transmission METHOD The calibration established by the Elemental Sensitivities METHOD is used as input for the ET METHOD so the same arguments concerning the calibration also apply in this case For the ET METHOD the dark matrix is well treated but the range of reported elements must be restricted to the range that is spanned by the elements with useful lines of the target In case some of them are rejected during the calculation of the F factors i e the sample is either too thin or too thick reported elements of such samples must fall within this range As criteria for the
213. rget Tmetal spe and target plus sample S7 1_Tme spe were measured with varying tube current 5 10 and 40 mA depending on the metal standard acquired for 519 s in total respectively A tube current value of 40 mA is assumed to be representative although not used for all measurements The target plus sample spectrum S7 1_Tme spe is fitted with the model file S7 1_Tme inp The target spectrum Tmetal spe two model files for split ROIs T metal inp and T metal2 inp is a typical example where the fit even with highest order of the linear background splitting of lines into KA and KB peak shape correction included and two ROIs will give a bad individual Chi square value for Ti In principle pure elements calibration standards spectra input models can be well defined but for the accumulative spectrum no individual treatment is possible Again Soil7 cal is used for calibration For the sample representation the previous Soil7 1 asr is used Current calibration C QRASDEMONET SOIL CAL Select calibration file Change Measuring Parameters Select Files for Analysis Perform Calculation of Concentrations Define File to Save Analysis Results Calibration Data Filename SOIL CAL 69 28 2606 SecTarget R 56 st 13 Elements K lin oh ee L Lines Pb Arrows move selector box select item lt ESC gt exit Fig 5 7 After the selection of the calibration file and definition of measuring parameters three relevan
214. ri distilled water were added All volumetric manipulations were cross checked with a balance 0 1ml astm 9 1 Li Pe GarO Dm en As example for the preparation of the final calibration standard solutions the 100 ppb standard can serve For 10 ml final volume 100 ul from each of the 10 ppm single element solutions of Mn Cu and Sr were pipetted into a new vial Yttrium was added with a volume of 500 ul Tri distilled water was added to the mixture with a volume of 9 2 ml For any of the three elements Mn Cu Sr the concentration is 0 1ml Dobim D x BROT PEE 0 1401 40 1 40 5 ml 9 2 138 For the internal standard Y the concentration is 500 ppb Onto the reflectors 20 ul of the mixed calibration solutions had been pipetted and dried on a hotplate inside a clean bench until complete removal of the water matrix before the respective measurements For the AXIL fit an input model TXRF inp was used for all calibration standards spectra with the inclusion of X LINES ADD MN KA MN KB CU KA CU KB SR KA Y KA The splitting of lines for manganese and copper was necessary for the sake of fitting of the spectra of the higher concentration standards Further included were elements occasionally showing up as contaminants with X LINES ADD FE ZN SUM Spectrum 100PPB SPE Iteration 4 ChiSquare Display BEG beg chan END end chan MIN min cnts MAX max cnts ROI SPECTR RES IDUAL LIN LOG R e s 1 d u a 1 800
215. rrect standard description file Cu NBS std In analogy all other asr files as needed for the calibration had been edited with the exception of NBS1108 asr which represents the reference standard material NBS 1108 and has all seven elements present in the spectrum 78 SSPEC_ID pure Cu 5mA vc2099 spe Intx2 SDATE_MEA 00 00 2000 00 00 00 SMEAS_TIM 40 SPEAKS 7 26 28 29 Sp 82 50 25 4 6 Source description files sou During the use of the METHOD NBS alpha coefficients and for the utility Incident and take off angles source description files sou have to be loaded Such files have to be created and filled with relevant data Most important are the entries for the correct incident and take off angle in respect to the sample surface and the assumed spectral distribution of the primary radiation exciting the sample As a speciality of this file type instead of energy values for the description of the primary spectrum of e g the secondary target or the radioactive source corresponding wavelength values in Angstrom must be used The conversion is rather simple 12 396 E keV A Angstrom 4 1 Example For the Rh secondary target with K energy values of 20 216 keV Ka 20 074 keV Ka 22 724 keV Kgi 23 173 keV Kp2 and 22 699 keV Kgs which were converted with the above formula and using the respective relative intensities the file Rh K sou was created directory QXASdemo NBS SIDENT Rh secon
216. rtunately these elements are not found in the sample For the element Pb relative standard deviation 18 there is a higher scatter around the average value Similar for the element Sn relative standard deviation 13 which unfortunately also shows a systematic bias to the pure element standard This fact can be explained by the small peaks found in the respective spectra due to the weak excitation as compared to the other elements Changes in the fit model can have a drastic impact Therefore Sn had been isolated with an input model of its own for the AXIL fits of the NBS standards Mn had only to be included for the sake of NBS 1108 no statement can be made As a conclusion for the everyday use of this METHOD after the analysis of the information gained from this SCROLL BOX one will maybe decide to exclude certain standards 102 NBS1167 NBS1163 NBS 1168 NBS1115 CU NBS FE NBS NI NBS PB NBS SN NBS ZN NBS AVERAGE Fig 6 12 Pure intensity values for all standards are calculated In the ideal case the average value is a good representative for all calibration standards For copper and zinc this holds true Due to the information gained by this scroll box certain standards may be removed from the list The weak part of the program is the almost uncontrollable impact imposed by the selection of the polynomial as defined in equation 6 1 The number and quality in terms of the concentration ranges spanned by the elements of t
217. s simple spectrum even the automatic ROI will do for all fluorescence peaks of interest In cases when a complicated fit situation is identified the spectrum has to be fitted in parts Examples are alloy spectra or the scatter peak fit for the sake of the dark matrix representation as needed by the fundamental parameters METHODs When another region is fitted after the first region had been fitted and saved the attempt will result in a warning message ASR exists The usual continuation of the saving will be to type in A for combining the results The background model must be simple for a first run i e the parameter describing the model must be low 0 5 The linear background model is usually the first choice The exponential background will be better suited for steep increases of the underlying background Select for the example spectrum e g BACKGR LINEAR and PARAM 5 because the background within the region of interest does not change drastically Spectrum SOIL 1 SPE Iteration 2 ChiSquare 2 2 Dif 007 3104 So0il7 binder 0 1055g dilution factor 1 99 40m vuc215 Display F t BEG beg chan END end chan MIN min cnts MAX max cnts ROI SPECTR RES IDUAL LIN LOG R e s 1 d u a 1 1000 1200 i GO Channel Number CANCEL gt gt DISPLAY LIN MAX 1500 ROI RESIDUAL _ Fig 10 6 AXIL fit results obtained by a rather simple definition of input parameters adequate for a first round
218. s are prepared as fused beads The impurities e g Fe Cu Zn Pb W found in blank spectra may originate from different sources E g tungsten may come from the dyes used for pressing It may also come from grinding in tungsten carbide mill Eventually its presence may be due to spattering of X ray tube cathode resulting in contamination with tungsten L series lines of the excitation spectrum The peaks of other elements may originate from the instrument chamber itself Sometimes it is possible to reduce the magnitude of the peaks by coating part of the chamber with pure aluminium or indium foil and or by adaptation of the sample chamber internal collimators The so called blank subtraction i e the subtraction of disturbing peaks either at the level of the MCA or by editing spectra or asr files is generally not recommended If performed the blank subtraction should be done carefully as the magnitudes of blank peaks may be not constant e g it may vary with the sample matrix In general the impurity peaks should be eliminated or they magnitude reduced to a degree where it is not significant An example blank spectrum HWC spe model file Samp Bl inp is shown in Figure 1 26 The blank sample was a pellet made of pure binder HWC wax chemical formula C3gH76N202 The element copper was identified net count rate of 1 3 counts second iron 1 3 counts second and tungsten L series lines 25 Spectrum HWC SPE Iteration 3 C
219. s as provided by QXAS may not be quoted In order to establish the standard deviation one has to prepare and measure samples repeatedly From these results the average and the standard deviation must be taken Table 3 2 Comparison of results obtained with the method Elemental sensitivities with certified values in mg for the reference standard material SRM 2793 air filter Si Ca Fe Zn measured 137 10 0 27 2 1 9 Certified value 58 600 1 758 13 200 1 716 26 500 1 590 1 790 0 125 60 Obviously silicon can not be described appropriately as it is the case some many times for light elements The other elements results compare well 3 3 Calibration for the Full Fundamental Parameters METHOD Demonstration files directory QXASdemo FP scatt Instrumental parameter files Test fpc Soil fpc OrgaMatr fpc NBSalloy fpc AXIL result files NBS1108 asr directory QXASdemo NBS Binder reb Full Fundamental Parameters is the most versatile METHOD for quantification in the QXAS package and is suited even for completely unknown samples Several modes of excitation with electromagnetic radiation in the range of X rays can be covered and many parameters can be selected to match the assumptions needed for the calculations with the experiment The fundamental parameters approach takes into account the spectral distribution of the excitation source the fundame
220. s below the detection limit U_Tot stands for the total absorption correction for the line of interest F_Enh_Tot for the inter element plus enhancement by scattering correction and F_Enh_Scat for the enhancement by scattering For the two last two columns values close to 1 will correspond to small correction less important enhancement by other elements or scattering For the experienced user The line Save peak ratios Y N aims to the fact that improvements in the AXIL fit will have a positive effect on the quantification When the question is answered with y files with the extension RAT will be created which can be used for the model file inp of AXIL manual input by copy and paste only This will affect the intra line peak ratios of K and L lines for a given element different sample self absorption for different energies With these improved fit results the quantification is to be repeated For samples measured under the same conditions and adequately represented by the fluorescence and scatter calibration the calculation can be started for another sample Next sample Y N with y 123 CHAPTER 8 SOURCE EXCITATION References 37 38 39 40 Note for users of ORTEC MCAs and software The MCA spectra can be saved in ASCII format and spe files are already generated at the level of storing the spectra after the measurement the spectrum conversion with QXAS is not necessary Unfortunately there is a
221. s found All elements will be described with the same calibration curve as selected by the user Sometimes calibration curves can be selected that according to the above recommendations should not be selected As consequence unrealistic results can be obtained because for the higher order polynomials coefficients a program internal constants will be used to establish the curves E g for the preliminary calibration all four calibration curves are available Select any of the quadratic line calibration curves and the preliminary results either not normalized to 100 or normalized can be like this 103 b 858 x Total 100 000 7 Fig 6 13 a and b Intentional creation of bad results obtained with the preliminary calibration and a Selection of polynomials not adequate for a quantification with only pure element standards As there could not be taken a clear decision what calibration function should be used all four calibration functions were tried The option Do you want to normalize conc entration was willingly accepted under the very likely fulfilled expectation that no other elements constitute the sample Table 6 1 Normalized concentration values as obtained by the use of four different calibration functions The as expected values refer to the preparation of the sample Concentration in calibration function Cu Zn Sn Pb Y al X 89 702 1 486 1 163 7 650 Y a0 al X 89 649 1 539 1 106 7 706 Y al
222. s of channel number the corresponding provisional energy not necessarily correct for this spectrum and the number of counts accumulated in that channel are displayed By default the starting point is the centre of a spectrum The navigation with arrow up and down or mouse clicks into the spectrum will lead to the approximate maximum The lines of interest have to be carefully examined for their real peak maximum Once at position 414 press the lt F1 gt key or select with the mouse the option F1 E_CAL which must be followed by the input IT for the weighted average energy of Ti Ka or the energy value 4 509 Very important after this first energy point the correct key to be used is the lt Space bar gt but not lt Enter gt in order to define also the second point for the energy calibration because lt Enter gt will be the last command to terminate the energy calibration The use of lt Enter gt during the calibration procedure will accomplish the calibration as a consequence the term ZERO will not be the true intercept as needed for a correct energy calibration but its value would be 0 The second peak maximum position should be found at channel number 1595 followed by lt F1 gt again The correct input will be MO or equivalently accepted the corresponding energy value 17 443 for the weighted average energy of Mo Ka As alternative for the Mo Kg peak either Mo Kg maximum channel 1791 or the coherent also nam
223. s to go through the loop of Prepare standard concentration file For the intermediate quantification the files _Cu NBS asr _Zn NBS asr _Sn NBS asr and _Pb NBS asr were created and edited such that four lines are found for element entries Cu Zn Sn and Pb These files were used to generate equivalent std files 6 4 Creation of the alpha coefficient file When the standards asr files had been edited according to the above guideline there is no need to make a fuss about the alpha coefficient file Any single file from the set of the std files NBS1103 NBS1107 NBS1108 NBS1115 Fe NBS Ni NBS Cu NBS Zn NBS Sn NBS Pb NBS is to be loaded There is no need to go through this procedure for more than once 99 Fig 6 6 The most convenient way to generate an alpha coefficient file is to load an appropriate std file The question Use standards can be answered with Yes comfortably There is no need to care for what so ever displayed Weight frac normalized concentration values even when they are asterisks nor about Intensity Remember the o coefficient file is concentration and intensity independent Element Line Weight frac Intensity El 0 6666000 6600006 2539 96808 Seoeooaoo Add an element Fig 6 7 Once the asr file that had been used to generate the loaded std file had been defined according to the syntax as expected by the NBS method the alpha coefficient file will be defi
224. se programs and the name of the files in use by the package Some file names are passed from one run to the next so they are will be used as default Programs can use the same data files without having to know all the details Modifications to one program have little or no influence on other programs Examples when it is necessary or profitable to edit data files are many Note the rigid syntax of the files has to be maintained at any rate with only a few exceptions like the number of space characters separating number values or the text of spectrum headers block SPEC_ID 4 1 Editing of spectrum files spe Many times it will be helpful and by QA requirements it is strongly recommend to modify the spe file header that should carry information in order to uniquely identify a spectrum when working with AXIL This information is further passed on to the result file asr as used for all quantitative METHODs Example in As203 spe directory QXASdemo SPE Sts the otherwise empty line following the header identifier SPEC_ID was edited such that the identification number from the sample preparation log book a brief description of the standard itself the tube current and the spectrum name from the spectrometer log book will be displayed when the file is loaded into AXIL 74 B AS203 ASR WordPad File Edit View Insert Format Help Ce 4 amp 4 amp fo SPEC_ID 3121 As203 binder Sm ve2416 spe DATE_MEA 00 0
225. sitivities fitted as a function of the atomic number When probing what kind of polynomial Linear or Logarithmic and what Order of the polynomial the maximum order is 5 is adequate and how to deal with the option weighted or not weighted several possibilities can be excluded when an obviously bad fit is obtained 55 alibration file C _T SEPTNETNMMETALS CAL reated on 09 29 2006 Calibration date 09 29 2006 40 Atomic number KEsc gt continue lt F9 gt LOG Plot lt F10 gt Printer Plot Fig 3 8 An obviously bad selection of parameters defining the polynomial fit Also a fit for the L line calibration is attempted by QXAS but due to the fact that only Pb LStd asr was added as L line emitting standard a warning massage will be displayed Press any key to continue Fig 3 9 Warning message for the case when less than tree calibration standards are available for an attempted polynomial fit Still it will be possible to quantify sample elements represented by standards In this particular case only Pb could be quantified but no other L line emitting elements The ultimate criterion to judge a fit is the Mean difference value found in the report displayed at the end of a loop The optimum value of 6 2259 was obtained Mean difference 6 2259E 000 for the above named selection rubidium sensitivity 7 01 0 436 10 The compound MnO is problematic the Mn calibration point will be used but results for th
226. son of compound cal with metals cal that the later calibration is superior This conclusion in the best case holds true for the better precision One should rather establish calibrations by inclusion of as many 58 necessary calibration points as possible where the necessity of calibration points is imposed by unknown samples The step Polynomial fit of sensitivities is not recommended prohibited according to QA requirements Nevertheless both calibration files compound cal and metals cal had been processed with Polynomial fit of sensitivities and the relevant calibration data for Rb was noted The element Rb was selected as example because the examples for the ET METHOD will need such an extension Both calibrations were re run with an order of 0 for the polynomial Order of polynomial 0 no fit 0 They were not fitted at all and the original calibration points are not biased Outside of the QXAS environment the relevant data for Rb was pasted into the now unfitted calibration files for further explanations see CHAPTER 4 Editing of data files Rb results for unknown samples might have a higher uncertainty but all other elements represented by measured calibration points are well defined 3 2 3 Application air filter sample A variety of samples e g organic matter geological samples both usually containing dark matrix or alloys for which the secondary excitation needs to be corrected cannot be treated appr
227. sufficient screen shots print screen will be forced with a combination of commands First use lt Alt gt and lt Enter gt which runs the DOS screen as reduced size foreground window then use lt Alt gt or lt Shift gt and lt Prt Scr gt which brings this window contents to the clipboard memory Finally paste with the usual lt Control gt V into any desktop publishing document from there print as for any other document On some combination of PC hardware this method may not work while the AXIL graphical mode is active A copy of the AXIL screen may be sufficient in many cases but for reports and publications a higher resolution graphics is usually required The spectral data can be exported to an ASCII file Such file can be read in by many data visualization programmes e g to MS EXCEL There are two ways to export the spectral data It is possible to convert the original MCA spectrum file to ASCII format directly In the MENU Spectrum format conversion choose the sub MENU Select format of source data and select the MCA format appropriate for your spectrum In the sub MENU Select format of target data choose ASCII The resulting file asc has as many lines as the spectrum has channels listed in a single column The major disadvantage of this way of conversion is lack of energy calibration data in the resulting file The second method of exporting spectral data becomes available after fitting the spectra After the fit select CO
228. sult files for the Rh K scatter region directory QXASdemo FP Scatt HWC l asr HWC 3 asr Cellulos asr Liche336 asr Cabba359 asr Soil7 1l asr LakeSed3 asr Soil7 2 asr Instr Bl asr Si l asr Al Std 1 asr MgSO4 1 asr P KH2P 1 asr K KH2P l asr K2CO3 1 asr TiO2 1 asr K2Cr2O 1 asr S Std 1 asr Ti Std 1l asr K KBr 1 asr Br KBr 1 asr Prior to the computation of an element s concentration for the calculation of the absorption as well as the enhancement correction the concentration of all elements in a sample should be known This vicious circle can be overcome by a suitable iterative procedure Still there are the elements forming the dark matrix also their composition needs to be known The fundamental parameters METHODs Full Fundamental Parameters and Fundamental Parameters monochromatic excitation scatter peaks use the incoherent coherent scatter peak ratio for establishing the average atomic number of the dark matrix of a sample in order to calculate the absorption correction and when not defined by an input also the sample thickness The two scatter peaks of the characteristic Ko lines e g Rh originating from the excitation source are usually engaged rather than the L lines or the continuum spectral background To generate useful results representative for the scatter by the sample is an art of its own Consequently there exist several approaches how to define the scatter peaks for the AXIL fit model Three different
229. supplied to the current cal file and a peak area N can be read in by means of an asr file gos 6 2 LT i c LT is the acquisition time of the respective spectrum i the tube current and A the absorption correction Kg lines and La lines are treated independently If more than one sensitivity is found for a calibration point one element the average value will be used When more than three calibration points elements are found for a Ka or L calibration non measured elements can be interpolated even extrapolated though not recommended In the ideal case no interpolations or extrapolations with Optimize with Least Square Fit will be necessary Following the guidelines of CHAPTER 10 Elements of quality control the use of such treated cal file is even prohibited The report contains all individual results for each standard asr file and a summary of the calculated sensitivities It can be saved as a text file extension arp which is not necessary for further calculations because the relevant data is stored in the current calibration file cal This file is used to determine the concentrations of elements in unknown samples again correcting for the sample self absorption The composition of the sample s dark matrix must be known Also intermediate thick and thin specimens can be evaluated their area related mass aerial density is need as input An on line HELP can be activated throughout the whole METHOD with the lt F1 gt key
230. t asr files must be loaded The proposed target element Pb will not be selected because it had been included for the sake of the spectrum fit with AXIL lead is a contamination for this measurement 88 rSelect Peaks for Abs Calc Pea Mi 368662 g Cr 163234 Co 46652 ETI 26301 Ge 15683 Fig 5 8 Relevant target elements must be selected During the absorption correction procedure the target element Ti is rejected Cr is then the first target element The iron and elements with lower atomic number edge correction Enter element to start iter was applied with the input of Fe Lg E Lg Lg A Fact 5 412 0 865 2 673 6 925 0 237 1 799 8 041 0 026 1 489 9 875 0 614 1 256 17 443 2 486 1 047 slope 2 877 intercept 5 842 regress 0 991 stdev 0 142 A O oe a oa wo 1 80 2 00 2 20 2 40 2 66 2 86 Log E Accept y n y Enter Element to start iter Fe Fig 5 9 The absorption correction factors are calculated at the energies of the accepted target elements Criteria to judge the calculation are the slope the regression coefficient and the standard deviation A discontinuity in the absorption behaviour of the sample caused by an element present in higher quantity can be taken into account In order to compare the results with the certified values for the two kinds of targets one has take into account the dilution factor of 1 99 No wonder the results for Si K Ca are out of competition because non
231. t Ph 1 696000 4 80000 4 66000 4 86000 6 66000 4 66000 6 66000 4 60000 4 88000 Fig 1 16 Definition of the sample standard matrix in order to improve the pre defined line ratios The goal of the definition of the sample absorption is to correct the tabulated line ratios for the energy dependent self absorption Such correction improves fitting spectra containing intense L series X ray peaks To complete the correct definition select under Specify experimental parameters with the COMMAND Excitation conditions and type in for the Primary excitation energy keV 20 6000 for the Detector characteristics Beryllium window micron enter 25 000000 and as Path length cm 0 500000 which are the parameter values specific for the used spectrometer Specify experimental parameters Excitation mode Excitation conditions Detector characteristics Filter absorption Funny filter absorption Path length Fig 1 17 Definition of more parameters affecting the pre defined line ratios After loading the input file load the spectrum Pb pure spe With the KLM MARK COMMAND all possible elements can be checked for their presence in the standard It is known that the sample consists of lead but one will find peaks matching the elements Ga As Kr Zr S None of these is present in the calibration standard The Ga Ka peak is not accompanied by a Kg peak in the spectrum therefore it is rather one of the so many lead L
232. t The additional for Sr Ky Sr Kg and Mo Ka includes the escape peaks the includes not only the escape peaks for an entry but also experimentally determined peak shape correction deviation from the usually assumed Gaussian profile of the peak The entry SUM includes any possible sum peaks due to pile up of the X ray detector pulses The order of the background polynomial is increased to 20 using the COMMANDs BACKGRND PARAME 20 Spectrum TARGET1 SPE Iteration 4 ChiSquare Targeti i V Co Cu Se Sr Mo vucZ2176 spe Axil LOAD STOP DISPLAY ROI CALIB X L INES KLM MARK FIT REPORT SAVE_RES PLOT BATCH BACKGRND SCAT_ROI GO CANCEL Fig 1 13 Spectrum Target1 spe displayed after the optimized AXIL fit After the re fit the Chi square value is reduced to 1 5 The overall magnitude of the residuals is also reduced Before storing the results one should inspect the numerical values of the calculated peak areas by using COMMANDs REPORT only available after fitting the spectrum and GO Spectrum T1 SPE Show lt t gt lt b gt lt Pg Up gt lt Pg Dn gt Fitting Region channels 232 1436 lt Home gt lt End gt Ener KeV Peak area A lt Esc gt 950 195783 925 44091 041 55094 208 47715 142 20134 443 52410 GO CANCEL I Fig 1 14 AXIL report of the fit results of spectrum T1 spe The individual Chi square values should be less than
233. t round for all fluorescence calibration standards instrumental constants should be calculated in way that they will not written as individual constants to the fpc file directly If necessary this should be done only after careful evaluation of all results The FORM Information on sample should rather be named information about standard but bear with us We profit again from the fact that the spectrum header had been edited such that all relevant information is displayed without search in log books The acquisition time is read in automatically but the tube current has to be edited manually the date as well of no relevance for tube excitation As Sample type toggle to Thick sample as for all other calibration standards otherwise the Sample mass needed to be specified The rest leave untouched Information on Sample D ii_NOU QKAS DEMO ASR STDS K KBR ASR Sample ID For K 3122 KBrtbinder 5mA UC2452 spe Live time sec 16660 66 Tube current mA gt 5 686008 Date Cmm dd r I 2000 SP Thick sample EEE Sample mass Cmg cm 2 gt 6 666066 Supporting material No support Thickne Cmg cm 2 gt 6 800000 Backing material No backing Thickne Cmg cm 2 gt 6 660800 Convergent value for calculation Fig 3 24 Definition of details concerning the potassium calibration standard K KBr asr like the X ray tube current and the sample type The consecutive FORM Menu for Setting up Options provides several options which sh
234. te a file Edit a new file with extension REB defining the composition of the used binder or one can toggle use the lt space bar gt to Input an old file and profit from the already pre defined file Binder reb For the eager readers who want to define such a file themselves One will use the utility Calculation of average atomic number and calculate for the used binder HWC C3gH76N202 by input for Formula of C38H76N202 and any arbitrary Weight the concentration values for 100 of this binder 66 m Result Compound 1 C38H76H202 1 0000 Hr Element Weight Atoms 12 918 380 0 76 962 190 0 4 724 10 0 5 396 10 0 Fig 3 27 Weight results for the chemical elements contained in the used binder The pressed pellet consists of 81 64 of KBr 4 024 g and 18 36 binder 0 905 g therefore multiply all HWC concentration values with 0 1836 it results in the concentrations to be used as input in order to generate a file like Binder reb Known Composition Percent Ele Percent 5 666600 4 OOOD 666066 666066 666666 666066 666666 666066 666668 666606 DUCES 666666 Ele H dla 9000608 4 5 A 4090000 90006 A bkn Fig 3 26 Definition of the binder composition This was the last step in order to define K KBr asr as a calibration standard for this METHOD One could have included many more suited asr files in one run and define them one after the other Ca
235. ted mass attenuation coefficients of the sample constituted of the element x the other fluorescing elements i and the dark matrix for exciting Eo and fluorescence Ex radiation acc 2 C C garkmatrix eS E F Pn E 2 46 s n Q s n WA 2 1 2 Inter element effects Sometimes fluorescent photons have sufficient energy to excite fluorescent radiation of other atoms in the specimen This effect is called secondary excitation also called inter element or enhancement effect and can be a major contribution to the observed photons This effect is strong for elements differing by 2 4 in atomic numbers Stainless steel Cr Fe Ni is the usual example for demonstrating secondary excitation iron excites chromium Fe gt Cr nickel excites chromium Ni gt Cr and iron Ni gt Fe The actual contributions from such effects are often around 5 10 in a few cases they are much higher For calibration standards like KBr and K2Cr2O7 this effect prohibits the use of the element potassium for the METHOD Elemental sensitivities analogue for KH2PO and the element phosphorus The intensity of a given characteristic peak of element x excited in a flat homogenous sample by a polychromatic primary beam of photons as utilized by the full fundamental parameters METHODs can be written as 38 I z GCE fy x tI E foQ A le C d C darkmatrix Ki A c C C garkmatrix 2 47 G is the geometrical factor called in QX
236. terest The concept to have equal counts for each peak used in method calibration ensures that the same statistical weight is given to each calibration point 1 2 27 A spectrum can not be fitted properly what to do X ray spectra should be fitted according to the criteria specified in paragraph 1 2 11 When a spectrum like PbO2 2 spe input file PbO2 inp can not fulfil a good fit criteria an explanation for the failure should be found Spectrum PBO2 2 SPE Iteration 5 ChiSquare 2 5 Dif vc2230 spe PbO2 3137 10mA reverse side 21 7 2006 Run 1 j Display BEG beg chan END end chan MIN min cnts MAX max cnts ROI SPECTR RES IDUAL LIN LOG R e s i d u a 1 1500 2000 GO Channel Number CANCEL gt gt DISPLAY RESIDUAL _ Fig 1 24 A spectrum of a PbO2 sample that cannot not be fitted by AXIL according to certain quality criteria In the example shown in Fig 1 24 the lead X ray lines were input by using PB LA PB LB PB LG the matrix of the standard was defined PbO plus binder the background was defined as linear polynomial of order 30 Still the fit results were not satisfactory despite acceptable value of the overall Chi square The residuals showed misfit around the most intense peaks and the individual Chi square value of the PB La was unacceptably huge When spectra like this are identified it usually indicates problems with linearity of the energy calibration or not properly coded fitting
237. th to k th subshell Factors f 2 fi 3 and 23 refer to radiationless transitions and fi to the radiative one For any given subshell the sum of yields fluorescent yield Auger yield Coster Kronig yields of all processes leading to its de excitation filling in or shifting out the vacancy in that particular subshell is equal to unity e g for the Ly subshell QO a fiz fis fis 1 2 20 and for the Ly Li and K subshell respectively a fy 1 2 21 a 1 2 22 Og a 1 2 23 Formula 2 19 can be written in a different form Ly L L T Tn L Ta On K a fra T En ia fis fiadas Tn OF iy 2 24 m Lin Tn The factor T 1 is called the hole transfer factor of i th subshell It is equal to unity for s type atomic subshells e g K Lr Mj subshell In the case of photoelectric excitation of Ly Lr and K atomic subshells the probability of atom relaxation through X ray fluorescence process is proportional to q TO fe he 2 25 TI T TI m 34 g O 2 26 T Ok 2 27 respectively Probability of emission of X ray characteristic line e g Koy is given by so called partial or fractional radiative rate pkai Partial radiative rate is a ratio of the radiative rate of that characteristic X ray line to the total radiative rate for a vacancy in the particular shell The most intense X ray characteristic lines include Ko Koz KBi KB3 K series Lo Loe LB1 and LB L series
238. ting etc and problems instabilities repair of equipment etc must noted For an established calibration during the measurements there have to be kept constant geometry anode material kind of secondary target high voltage for tube excitation vacuum conditions if applicable thickness and kind of protective foils The sample position must be reproducible and the irradiated area must be in the centre and smaller than the sample area itself For tube excitation the detector must be electrically insulated from the rest of the spectrometer because grounding problems will result in unexpected spectrum behaviour Reference Standard materials are to be avoided as calibration standards exception NBS METHOD because of their trace elements content which will introduce high statistical uncertainties 10 1 Parameters under control Certain parameters have to be carefully examined in order to establish a working range e g the influence of the applied tube current on the peak area behaviour some have to be under permanent regular surveillance e g sample blank and instrument blank others have to be checked but can vary within certain constraints because they have negligible influence on the results like the resolution of the detector FWHM and the stability of the energy calibration The experienced user will simply have an eye on the later category 10 1 1 Linearity as a function of the tube current The tube current for se
239. tion limits proportional to the square root of the background under a peak get lower A value for the FWHM close to specification of the detector should be reached The manufacturer had specified his detector under certain conditions usually at a low count rate fora Mn Kg peak originating from an Fe 55 source A ground loop to e g the high voltage generator of the X ray tube a higher count rate than specified detector aging and a non perfect energy calibration can lead to an increased FWHM value result Of importance in this context is a rather constant value over the time 145 Detector resolution 190 o gt x 180 ee a ge l 30th Ma 29th June 3rd July e 12th July oth August 31st Augusta z 7th July e 24th Augus g 3rd July 7th September S MO SSS 19th June 33 nA RS ARRS Se a AMU A R SS RARR T Ww 160 pS SS a 150 T T T T T T T T T T T T T 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 number of measurement Fig 10 2 The FWHM of Mn Kg controlled on a regular basis over a period of three months For the detector as part the secondary target spectrometer over a period of three months the FWHM of Mn K for the MnO calibration standard was recorded repeatedly An average of 174 37 1 61 eV was determined by use of the AXIL COMMAND REPORT FULL GO PEAK DATA Line EtKeU gt rel int peak area st deu chi sq chant fuhm lt el backgr tot abs
240. tter background 6 666600E 0006 Cimp s mA Incoherent scat constant 2 426575E 0003 imp7s mA Inc scatter background 6 666060E 0006 imp7s7 mA Fluorescence constant 4 619863E 0003 Cimp s mAl Fig 7 16 The calibration report is available as ASCII file 7 3 2 Analysis of unknown sample It is possible to work with sample asr files without scatter peak information contrary to the calibration procedure where at least one file must contain relevant scatter peak information This makes sense for samples without dark matrix Usually the unknown samples asr files should contain the fitted scatter ROIs However when no scatter peak information is found all concentration values will be normalized to 1 e g for alloys this will be a good option Warning The program will crash for elements included into the asr files with absorption edges above the excitation energy If this has to be done for fitting purposes the asr file has to be edited manually to remove such elements Although in the simplified secondary target excitation approach it were impossible to excite e g Zr with a Mo secondary target in reality it is possible because the secondary target scatters also continuum radiation from the x ray tube The first screen of this METHOD Select will lead to the analysis of unknown samples 2 Analyze unknown sample s by typing 2 Any other character will not be accepted and 1 opens the calibration procedure
241. uch events is proportional to the photon s intensity Si dead layer 3 5mm negative high vollage Fig 2 4 Cross section of a Si Li detector crystal and typical dimensions The response function of a Si Li detector to monochromatic radiation e g a characteristic line with characteristic energy E is nearly Gaussian The natural line width is usually neglected except for the Voigt peak shape model The number of counts Nj in X ray spectrum channel j in the vicinity of an isolated Gaussian peak with its top positioned at channel i is described by GAIN E Ef Ney a 2 60 J2n S E a Pa aie The factor A is the peak area which is equivalent to the total number of counts under the peak The and are the energies corresponding to channels j and i respectively The relation 42 between the channel number and the energy is linear It is given by the energy calibration equation 1 1 The GAIN in usually given in eV channel or keV channel is one of the two parameters of the energy calibration The relationship between the peak width S and the previously defined FWHM equation 1 3 is FWHM E V8 In2 S E 2 355 S E 2 61 The energy resolution FWHM of a semiconductor detector depends on the electronic noise and statistical fluctuations The electronic noise contribution is determined by the input amplifier stage and the detector leakage current therefore it is essential to operat
242. uld be included with fixed line ratios which for the scatter peak region fit is not recommended Setting the incoherent peak as described above puts only one line marker 18 corresponding to the calculated position of the Compton Kqi peak Another significant feature of the incoherent peak is that it is included with a resolution that is free i e independent of the resolution optimization gained from the other peaks Usually it will be advisable to fit the fluorescence peaks of interest in a first run with a different background model and within the relevant ROI then the scatter peaks As alternative to the fitting of the scatter peaks the COMMAND SCAT_ROI will sum all channel contents within a region of interest for the experienced user Hint The format of the scatter peak information obtained through the last approach is different when saved to the asr file than expected by some quantitative METHODs Editing of the asr file might be necessary when unexpected program crashes occur 1 2 16 What are the criteria for including or excluding a weak peak After setting up clearly identifiable peaks one can add to the model other peaks the origin of which is not yet fully confirmed The relevance and the need of their presence in the final fitting model need further checks Two criteria are helpful to decide whenever to remove a given peak from the fitting model in the report available after the fit COMMANDs REPORT GO th
243. ulting average atomic number is 8 88 It is possible to construct artificial scattering standards An instrumental parameter file OrgaMatr fpc identical with Soil fpc with the exception of the instrumental constants for scattering was created In order to obtain the scatter calibration the scatter peak information from the sample AXIL result file Soil7 l asr was copied and a pasted into a new file OrgaMatr asr containing this scatter peak information and supplementary information as needed to fulfil the asr file syntax This OrgaMatr fpc file was loaded by the METHOD Full fundamental parameters and the scatter calibration standard OragMatr asr was evaluated In the TOGGLE FIELD Select Calculation mode Instru Constants for fluorescence amp scatter was selected With an input for the tube current of 40 mA the sample type left at the default as intermediate thick use was made of the known aerial density area related mass of 21 5 mg cm There are no fluorescence peaks found in this file therefore the matrix composition must be specified FORM Menu for setting up options 7 There are known compositions Let us assume the matrix composition were unknown then one had to utilize the graph of Figure 7 4 The scatter peak ratio for the intermediate thick Soil 7 pellet is 2 77 from this an average atomic number of 8 95 can be deduced Consequently an artificial sample composition must be constructed that will result in an average number c
244. um of all later displayed concentrations will be summed up to 100 can be switched off for a first run of samples in order to identify otherwise disguised potential problems For alloys there is a long lasting dispute whether the normalization is allowed or not Depending on which party the reader belongs to the selection should be made After having gone through all this the final computations are initiated with Calculations of Geometry constants Analysis of unknown samp le 16 28 2666 Sample identity NBS1168 16mA vc2169 spe Spectrum fitting data C QRASDEMO NBS NBS1168 ASR Instrument parameter data C QRASDEMO FULL FP ORGAMATR FPC Instrumental identity Secondary Target fiverage instrumental constant 1 152 E 87 The secondary target Rh The tube anode Ag Tube voltage 50 KU Tube current 16 666 mf Measuring time 1000 Sec Collimator No Collimator Filter used No Filter Atmosphere Air Report of Calculated Concentrations Sample thickness infinitely thick Iterations for concentration 3 Pre set convergence 10 z Last convergence 62 z Ele line Constituent Concen elem gt Absorption Enhancement 296 44 12 191 4 2419E 83 671 26 13 070 5 552E 3 432 72 12 189 6 8227E 03 65 51 022 z 7 7724E 03 33 86 414 8 7263E 3 4743 3 532 381 1 1883E 83 749 41 72 165 3 6494E 3 Fig 3 30 Quantitative results obtained with the Full fundamental parameters method for the reference standard material
245. undamentals For standard Total Reflection X ray Fluorescence TXRF spectrometry a small droplet of a liquid sample volume 1 100 wl is deposited in the centre of the plane polished surface of a suited substrate usually a disk shaped quartz reflector with a diameter of 30mm Whenever possible the liquid matrix is removed by evaporation through heating or pumping Due to the geometrical arrangement the primary radiation of an X ray tube will impinge on the surface of the substrate under a rather shallow angle in the range of 1 6 mrad equivalent to 0 1 and will be totally reflected on this surface This physical phenomenon can be explained by the refraction index slightly smaller than 1 for X rays It can be shown that only a small amount of the primary photons can penetrate into the reflector On the other hand is the small sample excited by the primary radiation irregularly shaped therefore no total reflection effect and the interesting fluorescence radiation is collected very efficiently by the detector that is positioned much closer to the sample 5 mm than in conventional energy dispersive XRF As a result superior detection limits can be achieved A special equipment is needed and practically only samples in liquid form that must be spiked with an internal standard element can be analyzed Users of a TXRF equipment quantify their samples by the addition of a suited internal standard From the intensity ratios element of interest inter
246. up ax file is a dynamic file in editable ASCII format that accepts certain information from QXAS programs The current information relevant to the individual programs can be read by the programs and if necessary updated It contains the menu tree of some of the installed programs in the package names of the current data files information about libraries to be used etc The setup file is a block structured file i e it is divided into blocks that are identified by a header name with as starting character and as last character The actual position of a block of information within the file is of no importance When the setup ax file gets corrupted which can happen particularly after program crashes typically the SELECTOR BOX Axil X ray Analysis Package will display one line less than the usual four Axil X ray Analysis Package kad ae format conversion J s Utilities Fig 1 23 Possible consequences from a corrupted setup ax file In such a case delete the faulty setup ax copy the spare setup ax file which is found in the directory axil backup and paste it into axil folder Do not move the original backup file with drag and drop it shoul be stored in the axil backup folder for future use 1 2 26 Under what measuring conditions the example standards and samples had been collected Tube excitation Ag anode X ray tube in combination with Rh secondary target was
247. used to excite X ray fluorescence spectra The tube was operated at 50 kV constant potential The tube operating potential was kept constant for all the measurements otherwise the sensitivity calibration would be affected in a non linear manner All measurements were under vacuum conditions pressure in the sample chamber less than 5x10 mbar For the calibration standards in order to minimize pile up effects sum peaks and to keep the peak shape distortions low the standards spectra were collected at less than 20 dead time To achieve that the tube current was varied between 5 mA and 40 mA The counting time live time varied strongly among the collected spectra For standards it was selected in such a way to collect around 50 000 counts in the peak of interest With 50 000 counts a relative standard 23 deviation due to counting statistics of less than 0 5 can be expected The other samples were acquired for 1000 s live time and 40 mA tube current with a few exceptions when the dead time would have exceeded 40 e g for alloys For organic and geological samples the most intense were the scatter peaks By the acceptance of higher dead times as compared to the standards the low intensity fluorescence peaks were measured with relatively good counting statistics The Cd 109 annular radioisotope source excitation for standards the counting time was preset to a value that enabled collecting approximately 10 000 counts in the peak of in
248. ve METHOD routines 20 1 2 21 What is the FANO factor The response function of a Si Li detector is nearly Gaussian due to the fact that Poisson statistics applies the natural line width is usually neglected except for the Voigt peak shape model The FWHM Full Width at Half Maximum of a Gaussian peak is a function of the characteristic line energy E in eV and is described by FWHM AE 8In2 F E 1 3 where A Fie is the electronic noise contribution named NOISE in AXIL with a default value of 120 eV F stands for the Fano factor named FANO factor in AXIL with a default value of 0 114 The average energy required to produce an electron hole pair in the detector volume is with a value of 3 76 eV The usual assumption is that fluorescence lines are governed by Poisson statistics which describes independent events An example of Poisson statistics is the decay of a radioactive isotope each decay is completely independent from all past and future decays In semiconductor devices the details of the energy loss process are such that the individual events of the electron hole pair production are not strictly independent from each other therefore a deviation from Poisson behaviour is observed This deviation is taken into account by the inclusion of the Fano factor For true Poisson statistics the Fano factor were 1 1 2 22 Why does the ASR file information get into the respective SPE file Due to motivation driven by qu
249. ven stored as out file with REPORT FULL SAVE In order not to overwrite the asr file when other scatter peak models are tested the resulting file can be renamed after saving When problems are observed like a shift of the energy calibration or unexpected peak broadening during the iterations locking of input parameters D_GAIN D_FANO etc can be useful Model 2 In a rather puristic approach with the COMMANDs 106 X LINES ADD RH KA COH RH INC only the scatter peaks as needed for the quantification are fitted nothing else background linear order 0 In the linearly displayed AXIL fitted spectrum Soil7 2 spe one can see that the Compton peak is not well described This is a reason why one will prefer the rather exotic approach of model 1 or model 3 Spectrum SOIL 2 SPE Iteration 3 ChiSquare 516 6 Dif 087 Soil 2g 40mA vc2139 spe Axil LOAD STOP DISPLAY ROI CALIB X L INES KLM MARK FIT REPORT SAVE_RES PLOT BATCH BACKGRND SCAT_ROI c o u n t S Z C h a n n e 1 1500 2000 Channel Number GO CANCEL Fig 7 2 AXIL fit of the Rh Kg scatter region according to model 2 Model 3 The scatter peaks are not fitted at all Instead within a region of interest the respective counts per channel are summed up The background is usually not subtracted at all This is achieved with the so far not used COMMAND SCAT_ROI and consequently INCOHER and COHERENT by setting appropriate ROIs for the scatter
250. with the KLM MARK COMMAND option as explained in section 1 2 6 Spectrum TI STD SPE Ti Ka Axil i LOAD ee STOP Ti Kb DISPLAY ROI CALIB X LINES KLM MARK FIT PLOT BATCH BACKGRND SCAT ROI cC o u m t S gt Le h a m nm e 1 GO 1500 2000 CANCEL Channel Number Fig 1 18 Spectrum of a pure Ti calibration standard Ti Std spe 1 2 2 1 Example spectrum E calib spe The spectrum E calib spe had been obtained by firstly measuring pure Ti for 60 s live time similar to Ti Std spe Then the Ti foil was replaced by pure Mo similar to Mo Std spe the preset measuring time was extended to 63 s and the spectrum acquisition continued for 3 more seconds The resulting spectrum carries two well defined peaks Ti Ka and Mo Kg with their respective maxima at channel number 414 559 counts and 1595 438 counts 11 Spectrum E CALIB SPE hannel 18024 energy 11 188 KeU 8 counts Calib TM Mo Ka gt move move T jump 4 jump F1 E_CAL F2 R_CAL F3 ID_Ka F4 ID_Kb FS ID_La YA OE num CeV GAIN num CeV NO ISE 500 1000 1500 2000 num Cey Channel Number FANO F gt Gain 16 9 eV ch Noise 133 eV Fano 111 num RIP a SON tee OO GO CANCEL Fig 1 19 Spectrum obtained by additive acquisition of a pure Ti standard and a pure Mo standard E calib spe The energy calibration is performed with the COMMAND CALIB Underneath the spectrum name the actual position of the cursor in term
251. with the two numbers in the line after GEOM As the two numbers correspond to exactly the incidence and take off angle being under investigation the two values were set to zero 128 SIDENT Cd 109 annular source incident take off angle unknown Cd 109 0 SGEOM 0 0 SCONTINIUM 2 tepis 120 SCHARLIN 3 5608 82 770 4961 17 230 1408 3 74 One will pass through the SELECTOR BOX Current calibration by loading the previously generated preliminary calibration file Cd_52_90 cal without caring about the stupid ITEM Define sample current Current calibration C QRASDEMO CD 1869 CD_52 96 CAL Select calibration file Define Sample Current Fig 8 6 The file Cd_52_90 cal is used for the Monte Carlo calculation Finally exhausted one arrives at a FORM Enter number of Monte Carlo iterations This number will influence the time need for the execution of the Monte Carlo simulation For the sake of precision with nowadays computer processors the highest possible number of 999999 should be typed in Fig 8 7 The Monte Carlo calculation is initiated with the input for the number of events Having everything defined properly the calculation run will start and a flashing screen will show the progress Fig 8 8 During the calculation the screen will flash and display the progress in The SCROLL BOX LIST OF DATA AND RESULTS will provide the desired angles A bit cryptic they can be reached by scrolling
252. wn compositions Normalization of concentrations is applied Report is not surely done Fig 3 24 In the form Menu for setting up options the toggle field 7 must be changed to known compositions for standards containing chemical elements in form of a binder In the FORM Analyzed elements of as constituent the compound must be typed in such that the element under investigation is the leading element Different results will be obtained for KBr than for BrK this has to do with the calculation for secondary excitation correction As it complicated to define compound standards correctly for this METHOD for which more than one element should be used for calibration it is advised to generate different asr files for each element used for calibration This approach is well in accordance with the solution for problems sometimes encountered for fitting more than one standard element within one ROI Do not edit the net peak area For composition the input for the KBr concentration in the mixture of this compound with the binder can be calculated with the use of the EXCEL standards xls file finalyzed Elements of D 11_NOU QXASDEMO ASR STDS K KBR ASR constituent counts composition z KE ai 54562 2 81 63999 Fig 3 25 Definition of the major compound forming the calibration standard KBr and its concentration The binder will have to be defined later Due to the definition of 7 There are known compositions one will have to crea
253. y M lines will not be displayed as markers when KLM MARK is used for identification 1 2 14 Ar and or Kr are identified in a spectrum Argon Ar is usually identified in spectra collected with spectrometers having no vacuum chamber This chemical element is present in the ambient air If the measurements were performed in vacuum an argon signal either points to a misinterpretation or the vacuum chamber is not tight Krypton Kr is present in air at too low concentration to be observed in spectra It is usually the Pb Lg peak which may be mistakenly identified as Kr Kg by a novice spectrometer operator 1 2 15 How to define scatter peaks for the fit model Scatter peaks can be included in a fit model e g Rh Ka incoherent and coherent scatter peaks with the COMMANDs X LINES ADD RH INC RH KA COH In order to find a correct position for the relevant marker at the incoherent scatter peak maximum before adding RH INC the Angle of incidence degrees and Detector take off angle degrees have to be defined according to the spectrometer sample incidence and take off angle respectively The sum of incident and take off angle is the scattering angle For the secondary target measurements the default values of 45 for both angles will result in the correct scattering angle of 90 For source excitation or other geometry the defaults must be adapted For setting the coherent scatter peak also the line must specified otherwise both Ky and Kg wo
254. ys Rev A Gen Phys 10 1974 1507 THINH T P LEROUX J New basic empirical expression for computing tables of X ray mass attenuation coefficients X ray Spectrom 8 1979 85 SZALOKI I Empirical equations for atomic form factor and incoherent scattering functions X ray Spectrom 25 1996 21 DE BOER D K G Fundamental parameters for XRF Spectrochim Acta Part B At Spectrosc 44 1989 1171 TIKKANEN T HAMALAINEN K HUOTARI S Electrode thickness measurement of a Si Li detector for the SIXA array Nucl Instrum Methods Phys Res Sect A 403 1998 425 PROCOP M Estimation of absorbing layer thicknesses for a Si Li detector X ray spectrom 28 1999 33 TANG S M KUMP P YAP C T BILAL M G An XRF method for the determination of the efficiency of Si Li detectors in an extended source geometry by using thick specimens Nucl Instrum Methods Phys Res Sect A 241 1985 503 MOODY J R GREENBERG R R PRATT K W RAINS T C Recommended inorganic chemicals for calibration Anal Chem 60 1988 1203A HE F A generalized approach to quantitative EDXRF analysis using Fundamental parameters Ph D Thesis Department of Chemistry University of Antwerp UIA B 2610 Wilrijk Belgium 1991 27 28 29 30 31 32 33 34 35 36 37 38 WEGRZYNEK D MARKOWICZ A CHINEA CANO E BAMFORD S Evaluation of the uncertainty of element

References - International Atomic Energy Agency

Contents

Download Pdf Manuals

Related Search

Related Contents