Validation of Metrohm titrators

Application Bulletin 252/2 eValidation of Metrohm titrators (potentiometric) according to GLP/ISO9001Page 1/10Validation of Metrohm titrators(potentiometric)according to GLP/ISO9001GuidelinesOf interest to:General analytical chemistrySummaryGLP (Good Laboratory Practice) requirementsinclude the periodic check of analytical instrumentsfor reproducibility and accuracy usingstandard operating procedures (SOP).The user is advised to validate the Metrohmtitrators as a complete, integrated titration system,i.e. to perform a series of titrations usingstandard titrimetric substances (primary standards)and critically assess the results usingstatistical methods.Checking of the electronic and mechanical componentriesof measuring instruments can andshould be undertaken by qualified personnel ofthe manufacturing company as part of regularservicing. All Metrohm instruments are providedwith start-up test routines which check that therelevant assemblies are working perfectly whenthe instrument is switched on. If no error messageis displayed, it can be assumed that theinstrument is functioning faultlessly.As a guideline for the preparation of standardoperating procedures to check a titration systemcomprising a titrator, dispensing unit, measuringchain and possibly a sample changer, Metrohmsuggests the procedure described below. Thelimiting values specified must be considered asrecommendations. Specific limiting values mustbe defined in the particular standard operatingprocedure regarding in-house requirements tothe demanded accuracy of the measurementsystem.Application rangeThese test specifications are applicable to thefollowing Metrohm titrators:TitrandosTitrinosTitroprocessorsPotentiographsTest intervalsAnnual repetition of the testing of titrators appearsappropriate. If a dispensing unit is used incontinuous operation or if the work involves frequentuse of caustic, corrosive or precipitateformingtitration solutions which have a considerableadverse effect on the dispensing and/ormeasuring device, it may be advisable to decreasethe time between testing to, e.g. every 6or even 3 months.A special validation is advisable when one ormore components of the titration system arereplaced.Internal instrumental test routinesThe Metrohm titrators have an internal instrumentstart-up test and test routines. In the startuptest, the display elements are checked andthe contents of the program memories aretested by means of a checksum test. Properfunctioning of the data memory area is testedwith a write/read test.With the Titrino instrument series, the RS232interface is also subjected to an exhaustive test.If the titrators are regularly maintained, it is generallypossible to dispense with the specific validationof the instrument electronics.

Application Bulletin 252/2 eValidation of Metrohm titrators (potentiometric) according to GLP/ISO9001Page 2/10Maintenance/ServiceAn indispensable requirement to assure operationconforming to GLP for all instruments usedin the laboratory is careful maintenance andcleaning. Particular attention should also be paidto the accurate handling of such instruments.The instructions for use supplied with the instrumentshould be accessible to all workers inthe laboratory. We also recommend regularservicing of all relevant measuring instrumentsonce a year. Many Metrohm agencies offer favourablypriced servicing agreements for theirinstruments.MethodIf the daily work involves only a few, specifictitration methods, for the validation of the titratorit is advisable to select a combination of titrantand sample as similar as possible to those usedin one of the frequently employed methods andfor which a primary substance of specified, highpurity is available. In addition, it should be possibleto eliminate any error sources due to themethod.From the multiplicity of all possible combinationsof titrants and measuring chains, the pH titrationwith hydrochloric acid has been selected as anexample. The primary standard tris(hydroxymethyl)-aminomethane(TRIS) available from specialistdealers is titrated as a sample. The resultis calculated as a titer determination.Other possible combinations:TitrantPrimary standardCommentsHCl TRIS Trouble-freeNaOHAgNO 3Potassium hydrogenphthalateSodium chloride,NaClPotassiumhydrogen diioda-Carbonate-freeproceedingrequired *NaCl hygroscopicMind pH-valueSodium thiosulphateCe(IV) sulphate, As 2 O 3 Mind pH-valueCe 4+Potassium per-Disodium oxala-Mind pH-valuemanganate teTBAOH Benzoic acid Carbonate-freeproceedingrequired *Perchloric acid Potassium hydrogenMind tempera-phtha- tureNaNO 2 Sulphanilic acid MET modeEDTA CaCO 3 Buffering required* Sodium hydroxide readily absorbs carbon dioxidefrom the ambient atmosphere. Protect yourtitrant solution against the ingress of CO 2 byattaching a drying tube (6.1609.000) filled withCO 2 absorber or an absorption tube (6.1612.003) filled with NaOH.Other possibilities can be found in the relevantliterature, e.g. Metrohm Application Bulletin No.206.It is essential to ensure that only highly pure,dried primary substances (content min. 99.5%)are used. Primary standards should not be hygro-scopicand must be virtually completely insensitiveto CO 2 and air.If at all possible, aqueous solutions should beselected as titrants and these must also behighly stable to the influences of CO 2 , air andlight.Apparatus required• Titrator with dosing unit and stirrer (rod ormagnetic stirrer)• Exchange unit with anti-diffusion burette tip(6.1543.200)• Suitable measuring chain, e.g. combined pHglass electrode (6.0232.100)• Analytical balance, resolution min. 0.1 mg• 10 clean 100 mL titration vessels or beakers• Calibrated thermometer or temperature sensorChemicals required• Primary standard, e.g. TRIS, certified, declaredcontent min. 99.5 %, dried for 2 h at105°C and then allowed to cool off in a desiccator,where it is stored• Fresh titrant c = 0.1 mol/L, possibly c = 1mol/L (e.g. c(HCl) = 0.1 mol/L). Titer deviation

Application Bulletin 252/2 eValidation of Metrohm titrators (potentiometric) according to GLP/ISO9001Page 3/10When performing the titrations, ensure optimummixing of the sample solution. The setup illustratedbelow has proved its worth in practice.C02EP1Molar mass of primary standard (TRIS121.14 g/mol)Consumption of titrant in mLArrangement in titration beakerBEBEBurette tipElectrodeFor TitrandosTiter = R2 = C00*1000/CONC/Cl2/EP1mean value= R3 = SMN1abs. Std. Deviation = R4 = SSA1rel. Std. Deviation = R5 = SSR1The primary standard must be dried in a flat dish(e.g. 2 h at 105°C, depending on the type ofprimary standard) and allowed to cool off in adesiccator for at least 1 h. Standard substancesmust always be stored in a desiccator.With pH titrations (especially in the SET-mode),it is strongly recommended first to perform acalibration of the electrode to check the electrodeparameters. Fresh buffer solutions (specifiedvalue ± pH 0.02) must be used for this purpose.Calibration requirements:Slope > 0.97pH(as) 6.9...7.1(with comb. glass electrode and 3 M KCl aselectrolyte)In end-point titrations (SET) to a preset pHvalue, a calibration is essential. Further, it isadvisable to enter the working temperature forcompensation in the titrator or attach a Pt100 orPt1000 sensor to the titrator. The titrant solutionshould be in thermal equilibrium with the surroundings.C00 Sample size of primary standard in gCONC Concentration of the titrant in mol/LCl2 = ID2 = Molar mass of primary standard(TRIS 121.14 g/mol)EP1 Consumption of titrant in mLAll result variables and statistic functions areavailable as result templates.For Titroprocessor 796Titer = RT1 = SS*1000 /C/ID2/EP1SSCID2EP1Sample size of primary standard in gConcentration of the titrant in mol/LMolar mass of primary standardConsumption of titrant in mL2. Setting titration parametersThe settings of the titration parameters dependon the instrument and titration mode. The modewhich is used most frequently should be selected.If a start volume is used, the required factor hasto be calculated for each sample:For 796 Titroprocessor:Procedure1. Calculation formula of the titerFor TitrinosC00 * C01Titer = RS1= or C00 * C01/( C02 * EP1)C02 * EP1with 4 decimal placesfactor =1000 * C02C * ID2=1000 * 0.60.1* 121.14= 49.5 ≈ 501000 conversion in mLID2 molar mass of primary standard in g/molC concentration of titrant in mol/LC02 percentage of volume at EP (e.g 0.6)C00 Sample size of primary standard in gC01 Theoretical consumption of titrant for 1mol primary standard in mL (1000 or 10000 with 1 molar/ 0.1 molar titrant)

Application Bulletin 252/2 eValidation of Metrohm titrators (potentiometric) according to GLP/ISO9001Page 4/10For Titrinofactor =1000 * C02C01* ID2=1000 * 0.60.1* 121.14= 49.5 ≈ 501000 conversion in mLID2 molar mass of primary standard in g/molC01 concentration of titrant in mol/LC02 percentage of volume at EP (e.g 0.6)For Titrando:The start volume is calculated at the beginningof the method and saved as result R1C00 * 1000 * CV02R 1=CV01* Cl21000 conversion in mLCl2 molar mass of primary standard in g/molC00 sample size in gCV01 concentration of titrant in mol/lCV02 percent of volume at EPThe following table lists the recommended, relevantparameters for the instruments and modesfor the titration of TRIS with c(HCl)=0.1 mol/L.ParameterDET pHmeas.pt.dens.min. incr.start Vfactorstop pHsignal driftMET pHV steptitr.ratestart Vfactorstop pHsignal driftSET pHEP1 at pHdynamicsmax. ratemin. ratestart Vfactorstop drift719, 794,798, 799Titrino410.0 μlrel.502.850 mV/min0.10 mlmax.rel.502.850 mV/min5.135 ml/min0.5 μl/minrel.5020 μl/minInstrument796 Titroprocessor410.0 μlrel.502.850 mV/min0.10 mlmax.rel.502.850 mV/min5.135 ml/min0.5 μl/minrel.5020 μl/min808,809Titrando410.0 μlR12.850 mV/min0.10 mlmax.R12.850 mV/min5.135 ml/min0.5 μl/minR120 μl/minshould be varied in random order and result in aconsumption of titrant of ca. 0.2 to 0.9 cylindervolume. Refilling of the cylinder should beavoided, except between samples.The recommended sample weight ranges forTRIS are given in the following table:Cylinder vol. Weight of TRIS c(HCl)=5 mL 120...550 mg 1 mol/L10 mL 250...1100 mg 1 mol/L10 ml 25…110 mg 0.1 mol/L20 mL 50...220 mg 0.1 mol/L50 mL 120...550 mg 0.1 mol/LAs low sample weights increase the weighingerror and hence the scatter of the results, it isadvisable to avoid these by using 1.0 molar titrantsolutions with 5 mL and possibly 10 mLcylinders.The weighed samples are dissolved in ca. 40mL distilled or deionised water and then immediatelytitrated. The preparation of stock solutionsand titration of an aliquot introduces a furthersource of error (pipetting error) and is thusnot recommended.Interpretation of the resultsThe relevant parameters for the validation ofmeasuring instruments are the reproducibility(precision) and the accuracy of the measurementresults. To assess these quantities, proceedas follows:The values obtained from the 10 determinations(titer of the titrant) are used for the calculation ofthe mean value x and the absolute standarddeviation s abs . These calculations can be performeddirectly with the built-in statistics functionof the instrument, if available, or by using apocket calculator or a PC (Personal Computer)with a suitable software package (e.g. spreadsheetprogram). As slightly different results canbe obtained in complex calculations with differentcomputing aids owing to the different calculationaccuracies, preference should always begiven to values calculated in the instrument itself.3. Method10 titrations are performed with the same instrumentsettings and different weights of theprimary standard (e.g. TRIS). The sample size

Application Bulletin 252/2 eValidation of Metrohm titrators (potentiometric) according to GLP/ISO9001Page 5/10Mean valuenx1+ x 2 + K + x n 1x = = ∑ xn ni=1Sum of the individual values=Number of individual valuesStandard deviationiTiter theo (at X°C) = 1.000 + 0.0002 * (20 − x)b. Calculation of the systematic deviation d relThe systematic deviation is calculated fromtiterd rel =− titertitermeantheotheo*100sabs==1n − 1n∑i=1xn∑i=12i( x − x)i⎛⎜⎝−n∑i=1n − 12⎞x ⎟i⎠n2Requirement: The systematic deviation should bemax. ± 0.5 %.Note: In sample titrations, reproducibility andlinearity (volume vs sample size) are important.There are normally no problems with the accuracyas long as all titrant solutions are subjectedto a regular titer determination and the titer andthe sample are determined with the same titrationsettings.1. Reproducibility, scatter (precision)The reproducibility of the measurement is expressedby the relative standard deviation.rel. standard deviations rel =s * 100 abs.standard deviation *100=xmean valueRequirement: The relative standard deviationshould be ≤ 0.3 %.(While the limiting value of 0.3 % for the rel.standard deviation is a limit conforming to practiseand can easily be met in the normal case,under optimum conditions rel. standard deviationsof 0.1 % and lower are obtainable withMetrohm titrators.)2. AccuracyThe accuracy of the results obtained dependson the content of the primary standard guaranteedby its producer (assumption: 100.00%).a. Calculation of the theoretical titer value as a functionof temperatureThe theoretical titer value of the titrant solutionat 20°C is 1.000 with a reduction in the titer of0.02 % per degree temperature rise (with aqueoussolutions, see warranty of the chemical producer).3. Systematic errorsa. Linear regression volume/sample sizeTo discover systematic errors, e.g. disturbinginfluences due to the method or solvent blankvalues, a linear regression of volume (in mL)against sample size (in g) can be calculated.This requires use of a powerful pocket calculatoror a statistics package or spreadsheet programon a personal computer. The sample size isplotted as the x-coordinate (independent variable)and the volume as the y-coordinate (dependentvariable). The linear regression draws aline through the experimental points whichminimises the sum of the squares of the individualdeviations. The regression line is describedby the formula: y = bx+ a, where a representsthe intercept on the y-axis and b is the slope ofthe line (see diagram below).Systematic errors of the titration method aremanifested in a significant deviation of the zeropoint coordinates of the y-axis (intercept), i.e.the regression line calculated from the valuepairs volume/sample size does not intercept they-axis exactly at the origin of the system of coordinates.volumeasyssample size

Application Bulletin 252/2 eValidation of Metrohm titrators (potentiometric) according to GLP/ISO9001Page 6/10a sys as a measure of the systematic error is calculatedfrom the mean values of the x values,the mean values of the y values and the regressioncoefficient b (slope).The calculation formulae:b =n∑i=1( x − x)( y − y)ni∑i=1( x − x)ii2=n∑i=1x y −ni∑i=1xi2in∑xni=1*ii= 1 i=1⎛⎜−⎝∑nn∑⎞x ⎟i⎠n2yiof the titer on the magnitude of the volume or thesample size. This can also be an indication ofsystematic disturbing influences due to themethod.The slope b T/Vol (regression coefficient b, calculationformula, see p. 9) from the equation of thelinear function y = bx+ ashould here be 0.000in the ideal case, i.e. the line should be horizontalthrough y=1.000.titera sys = y-intercept =x − b * ywith systematic errorAssessmentvolumeIf a sys > ± 10 μL for 1 mL-burettes or a sys > ± 50μL for 5, 10, 20 and 50 ml burettes, it must beassumed that a systematic error is present. Acheck on the titration method and other possibledisturbing influences due to the system is thenimperative.titercorrected with a sysIf no optimisation of the validation method ispossible, the individual values of the consumptionin mL must be corrected by the value of a sys(volume–a sys in mL) to ensure that the systematicerror associated with the method is not incorporatedin the assessment of the titrator. Therelevant characteristic data for the reproducibilityand the accuracy of the titration results mustthen be recalculated with the corrected consumptionvalues.If they are necessary, these time-consuming calculationsshould be performed only with a computeror powerful calculator. However, it must benoted that slightly different results can also beobtained here on different computing systemsowing to the different calculation accuracy.b. Linear regression titer/volumeA further possible method to discover systematicerrors involves plotting the regression line (scatterdiagram) of the value pairs titer/volume. It isadvisable to plot such a diagram as it also providesa good visual impression of the scatter ofthe results.AssessmentvolumeIf b T/Vol is greater than ± 0.0010 ml -1 , a systematicerror due to the method must also be assumedhere. A correction of the consumptionvalues by a sys (volume in mL – a sys in mL) and asubsequent recalculation of the titer shows adramatic improvement when the regression line(titer against volume) is replotted.ConclusionIf systematic errors are found, an attempt mustbe made to optimise the titration method andadapt the standard operating procedure (SOP)accordingly. If no optimisation is possible or aspecified method must be used unchanged, therelevant characteristic data must be calculatedwith corrected consumption values (volume inmL – a sys in mL).A significant positive or negative slope of theregression line indicates a fictitious dependence

Application Bulletin 252/2 eValidation of Metrohm titrators (potentiometric) according to GLP/ISO9001Page 7/10Possible error sources• Primary standard• Balance /weighings• Titration vessel• Solvent• Titrant• Dispensingunit• Measurement• Titrator• Temperatureunsuitable, impure, moist,inhomogeneous, no guaranteedprimary standardqualitybalance too inaccurate,draughts, temperature influences,contaminatedbalance, temperaturegradient titration vessel/balance,carelessweighing, sample weighttoo lowcontaminatedimpure (blank value), poorsolubilising powercontains CO 2 , impuretubing connections nottight, contaminated cylinder(visible corrosionmarks), leaky piston (liquidfilm or crystals belowthe piston), filling rate toohigh, leaking burette tip,air in tubing system,three-way stopcock leakingcontaminated electrode,blocked diaphragm, loosecontact at connector,faulty cable, poor mixingof sample solution, unfavourablearrangement ofburette tip and electrode,excessive response timeof electrodeunsuitable titration mode,wrong measurement parameters,titration rate toofast or too slowtemperature fluctuations,especially perceptiblewith titrants in organicsolventsRecommendations for troubleshootingWith rel. standard deviation too high (poor reproducibility))• Finely grind fresh the primary standard substancein a mortar, dry and allow to cool off ina desiccator (possibly grind again).• Ensure complete dissolution of the weighedsample in the solvent.• Optimise arrangement of burette tip, electrodeand stirrer.• Regenerate or change electrode.• Optimise titration parameters (see MetrohmApplication Bulletins).• Remove Exchange Unit, clean and possiblychange tubing as well as piston and/or cylinder.• Weigh out sample only after temperatureequilibrium established between balance andtitration vessel.• Possibly increase sample weight.With rel. systematic deviation too high (accuracyunsatisfactory)• Use pure solvent (without blank value), boilout water if necessary.• Dry standard primary substance.• Ensure complete dissolution of the weighedsample in the solvent.• Use fresh titrant (possibly use different productionbatch).• Visual inspection of Exchange unit and replacementif need be.• Check electrode and titration parameters,regenerate or replace electrode.• Check balance.

Application Bulletin 252/2 eValidation of Metrohm titrators (potentiometric) according to GLP/ISO9001Page 8/10Procedure with values not conforming tospecificationsAll non-conforming values must be commentedon in the validation record and the subsequentprocedure noted.If excessive deviations are found, the differentpoints under the sections “Possible error sources”and “Recommendations for troubleshooting”must be carefully checked and the disturbinginfluences eliminated. It is essential to repeatthe validation. If unsatisfactory results are stillobtained when the test series is repeated, thevalidation must be performed again by a differentperson on a different day.If doubt exists regarding the precision of thedosing unit, this can be checked separately (seeMetrohm Application Bulletin No. 238).LiteratureFurther information on titrations and titer determinationscan be found in the following publications:• Metrohm Application Bulletin No. 206, Titerdeterminations in potentiometry• Metrohm Application Bulletin No. 238, Check ofDosimat according to GLP/ISO• Instrumental Titration Techniques, F. Oehmeand W. Richter, Hüthig Verlag, Heidelberg, 1987• Practical Aspects of Modern Titration, W. Richterand U. Tinner, Monographs Metrohm AG, 1988• Electrodes in potentiometry, U. Tinner, MonographsMetrohm AG, 1989On the following pages you will find an example of avalidation record and a diagram of the linear regressionmentioned above.

Application Bulletin 252/2 eValidation of Metrohm titrators (potentiometric) according to GLP/ISO9001Page 9/10Validation Record Company : Metrohm Ltd.Titrators Division : Ti.-MarketingTemperature in °C : 25 Instrument : 809 TitrandoTitrant : HCl Electrode : 6.1231.000Conc. in mol/L : 1.0 Slope : 99.0Lot / Date of manufact. : 22.01.2002 pH (as) : 6.961Primary standard : TRIS Exchange Unit : 800Molar mass in g/mol 121.14 Burette size : 10 mlTitration parameters : Mode : DET pHStop pH off signal drift 50 mv/minStop EP 1 measr. Pt. Density 4Vol. after EP 2 ml min. increment 10 μlEP criterion 5Smpl size : Volume : Titer : No. Remark :0.5883 g 4.858 ml 0.9996 10.3439 g 2.843 ml 0.9987 20.7883 g 6.511 ml 0.9994 30.9979 g 8.243 ml 0.9994 41.0701 g 8.837 ml 0.9997 50.8896 g 7.347 ml 0.9995 60.4897 g 4.048 ml 0.9987 7 Mean = 0.9993 s abs = 0.00040.3691 g 3.031 ml 0.9987 8 Titer ref = 0.9990 s rel = 0.04 %0.2604 g 2.150 ml 0.9997 9 d rel = 0.03 % a sys = -0.0044 ml0.6814 g 5.628 ml 0.9994 10 b T/Vol = 0.0001Result :

Application Bulletin 252/2 eValidation of Metrohm titrators (potentiometric) according to GLP/ISO9001Page 10/10Date : 23.01.2002 Signature : Haider Vis.:1.1000Validation of TitratorsRegression Titer vs Volume y = 9E-05x + 0.99881.0500Titer1.00000.95000.90000.000 ml 1.000 ml 2.000 ml 3.000 ml 4.000 ml 5.000 ml 6.000 ml 7.000 ml 8.000 ml 9.000 ml 10.000mlVolume