Instruction manuals

A flow-batch analyzer with piston propulsion applied to automatic preparation of calibration solutions for Mn determination in mineral waters by ET AAS

Description
The increasing development of miniaturized flow systems and the continuous monitoring of chemical processes require dramatically simplified and cheap flow schemes and instrumentation with large potential for miniaturization and consequent
Published
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
   Available online at www.sciencedirect.com Talanta 73 (2007) 906–912 A flow-batch analyzer with piston propulsion applied to automaticpreparation of calibration solutions for Mn determination inmineral waters by ET AAS Luciano F. Almeida a , Maria G.R. Vale b , Morgana B. Dessuy b , M´arcia M. Silva b ,Renato S. Lima c , 1 , Vagner B. Santos c , 1 , Paulo H.D. Diniz c , 1 , M´ario C.U. Ara´ujo c , ∗ a Universidade Federal de Pernambuco, CCEN, Departamento de Qu´ımica Fundamental, Brazil b Universidade Federal do Rio Grande do Sul, Instituto de Qu´ımica, Brazil c Universidade Federal da Para´ıba, CCEN, Departamento de Qu´ımica, P.O. Box 5093, 58051-970 Jo˜ ao Pessoa, PB, Brazil Received 1 February 2007; received in revised form 2 May 2007; accepted 11 May 2007Available online 21 May 2007 Abstract The increasing development of miniaturized flow systems and the continuous monitoring of chemical processes require dramatically simplifiedandcheapflowschemesandinstrumentationwithlargepotentialforminiaturizationandconsequentportability.Forthesepurposes,thedevelopmentofsystemsbasedonflowandbatchtechnologiesmaybeagoodalternative.Flow-batchanalyzers(FBA)havebeensuccessfullyappliedtoimplementanalytical procedures, such as: titrations, sample pre-treatment, analyte addition and screening analysis. In spite of its favourable characteristics,the previously proposed FBA uses peristaltic pumps to propel the fluids and this kind of propulsion presents high cost and large dimension, makingunfeasibleitsminiaturizationandportability.Toovercomethesedrawbacks,alowcost,robust,compactandnon-propelledbyperistalticpumpFBAis proposed. It makes use of a lab-made piston coupled to a mixing chamber and a step motor controlled by a microcomputer. The piston-propelledFBA (PFBA) was applied for automatic preparation of calibration solutions for manganese determination in mineral waters by electrothermalatomic-absorption spectrometry (ET AAS). Comparing the results obtained with two sets of calibration curves (five by manual and five by PFBApreparations), no significant statistical differences at a 95% confidence level were observed by applying the paired  t  -test. The standard deviationof manual and PFBA procedures were always smaller than 0.2 and 0.1  gL − 1 , respectively. By using PFBA it was possible to prepare about 80calibration solutions per hour.© 2007 Elsevier B.V. All rights reserved. Keywords:  Flow-batch analyzer; Automatic preparation of calibration solutions; ET AAS; Manganese; Mineral water analysis 1. Introduction The minimization of human interaction in analytical pro-cedures is an exhaustively persecuted target by the moderninstrumental analytical chemistry studies, mainly when a largenumber of samples are involved [1,2]. In general, the automated procedures are independent of errors caused by the operator andprovide high repeatability [3]. Several flow analyzers (FA) [4] have been developed in order to automate and to simplify ana-lytical procedures [5–16]. However, even with the successful ∗ Corresponding author. Tel.: +55 83 3216 7438; fax: +55 83 3216 7438.  E-mail address:  laqa@quimica.ufpb.br (M.C.U. Ara´ujo). 1 Tel.: +55 83 3216 7438; fax: +55 83 3216 7438. applicationofFAforautomationandsimplificationofanalyticalmethodologies, their flexibility and versatility are still limited.The FA manifolds require significant changes in their physicalassemblies when it is necessary to analyze samples with a largevariation of analyte concentration and/or physical–chemicalproperties.Automated micro batch (AMBA) and flow-batch analyzers(FBA)proposedbySweilehandDasgupta[17,18]andHonorato et al. [19], respectively, are systems more flexible and versatile (multi-task characteristic). In these analyzers, it is possible towork in any analyte concentration range as well as to imple-ment different analytical processes. It may be accomplished just by changing the operational parameters in their controlsoftware, without significant alterations on the physical config-urations of the analyzer. AMBA and FBA combine favourable 0039-9140/$ – see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.talanta.2007.05.009   L.F. Almeida et al. / Talanta 73 (2007) 906–912  907 characteristics of both flow and batch analyzers (BA). As in FA,thetransportationofreagents,samplesorothersolutionsarecar-ried out in a flow mode, and, as in BA, the sample processing iscarriedoutintoamixingchamber(MC).InAMBA,aninjectingloop is used on the sampling stage (as in FA), while in FBA thesample amounts are added into the MC by controlling the ONswitching time of one solenoid valve.As most of the FA, FBA and AMBA also present goodprecision and accuracy, high sample throughput and low con-tamination,consumption,manipulationofreagentsandsamples,cost per analysis and waste liberation for the environment, etc.Theyhavebeenusedtoimplementseveralanalyticalproceduressuchas:titrations[19,20],analyteaddition[21,22],internalstan- dard [23], screening analysis [24], exploitation of concentration gradients [25], on line matching of pH [26] and salinity [27], liquid–liquidextraction[17],distillationofvolatileanalyte[17], sample digestion [28] and kinetic approach [18]. In general, FBA and AMBA present the following character-istics: the use of solenoid valves and MC; highly precise fluidaliquots can be delivered by microcomputer controlling the ONvalves switching times; high sensitivity because the physicalandchemicalequilibriainherenttotheanalyticalprocessesmaybe attained and the dispersion and/or dilution of the samplesmay be negligible; the analytical signal measurements can beperformed in flow cells or directly inside MC and the multicom-mutation [29,30] may be used in order to manipulate the fluidsin a simultaneous and/or in an intermittent way.In spite of their favourable characteristics, the previouslyproposed FBA and AMBA make use of a peristaltic pumpand/or pneumatically pressurized reservoirs to propel the flu-ids. These kinds of propulsion result in manifolds with largedimensions, making their miniaturization and portability unfea-sible.Toovercomethesedrawbacks,alowcost,robust,compactflow-batchanalyzerbasedonpistonpropulsionisproposed.Thisnewapproachtoaccomplishautomaticanalysiswasdesignedtoadd a new study prospect in the present flow-batch technique. Itwas named as the piston propelled flow-batch analyzer (PFBA)becauseitusesastepmotor-controlledpistoncoupledtotheMC.To illustrate the feasibility of PFBA, it was applied to preparecalibration solutions for determination of manganese in min-eral waters by electrothermal atomic absorption spectrometry(ET AAS). It is worth noting that the most critical and time-consuming steps of an analytical methodology, during whicherrors may be introduced, are the sample pre-treatment and thepreparation of calibration solutions. The proposed system is agood alternative to perform these tasks. 2. Experimental 2.1. Reagents and solutions The 1000mgL − 1 Mn stock solution was prepared from aTitrisol (Merck, Germany) ampoule in 0.5% (v/v) bidistilledHNO 3  (Merck, Germany). Calibration solutions were preparedby dilution of Mn stock solution in 0.5% (v/v) HNO 3 . A 0.5%(v/v) bidistilled HNO 3  was used as the blank solution. Table 1Graphite furnace heating program of the Shimadzu AA6800 for determina-tion of Mn in mineral water samples. The optimized pyrolisis and atomizationtemperatures were 2100 and 700 ◦ C respectivelyStep Temperature ( ◦ C) Time (s) Gas flow rate(mLmin − 1 )Ramp HoldDrying 1 110 10 15 500Drying 2 250 7 15 500Pyrolysis 1 700 45 10 500Pyrolysis 2 700 0 10 500Atomization a 2100 0 5 0Clean-out 2300 0 3 500 a The signal reading is performed in this step. For the measurements of minimum times for switching ONvalves, the dye solution was composed of 0.1molL − 1 aceticacid/sodium acetate buffer, a 1000mgL − 1 Fe(II) in medium of 0.5molL − 1 HCl prepared from a Titrisol (Merck, Germany)ampoule and the chromogenic reagent was a 0.25% (w/v) 1,10-phenantroline solution prepared in medium of 0.05molL − 1 HCl.All solutions were prepared with chemicals of analyticalgrade and freshly distilled-deionized water from a MilliporeMilli-Q device. The mineral water samples were purchased inlocalsupermarketsandanalyzedwithoutanyprevioustreatment. 2.2. Instrumental parameters and operation of thespectrometer  A Shimadzu AA6800 furnished with a longitudinally heatedgraphite tube atomizer was used for all atomic absorptionmeasurements and pyrolytic-coated graphite tubes were used.Samples were delivered to the furnace using a Shimadzu ASC-6100autosamplerandstoredinacidwashedpolypropylenecupspriortoinjection.ThelampusedwasaMnhollowcathodelamp(Hamamatsu Photonics) operated at 10mA; the wavelength at279.5nmwasusedwithaslit-widthof0.2nm.Theinertgasusedwasargon(99.999%).DilutionswerecarriedoutwithcalibratedEppendorfpneumaticpipettes.Thegraphitefurnaceheatingpro-gramisgiveninTable1.Thevolumestakenofallsolutionswere always 20  L. 2.3. The piston propelled flow-batch analyzer (PFBA) TheproposedanalyzerisshowninFig.1a.FourColeParmer three-way solenoid valves were used: V B  and V CS  to directthe blank and work calibration solution or sample towards MC,respectively; V D/S  to seal the MC outlet during the preparationofcalibrationsolutionsortodischargethesesolutionsordilutedsamples of MC and V A/W  to direct these solutions into the cupof the spectrometer auto sampler or towards waste.A Pentium 550MHz microcomputer was used to controlPFBA, the spectrometer and to acquire and treat the analyticaldata. Software written in Labview ® 5.1 with a friendly inter-face and easy operation was developed to manage PFBA. Anelectronic actuator (EA) was used to increase the power of the  908  L.F. Almeida et al. / Talanta 73 (2007) 906–912 Fig. 1. (a) Schematic diagram of PFBA at starting configuration: MC=mixing chamber (the dimensions of MC are described elsewhere [16,17]); MS=magnetic stirrer;EA=electronicactuator;PS=propulsionsystemandPC=microcomputer,V B ,V CS ,V D/S  andV A/W  =blank,workcalibrationsolution,MCdischarge/sealingand auto sampler/waste three-way solenoid valves, respectively. Note that the valves V B , V CS  and V D/S  present one of its outlets completely sealed in order tominimize any internal pressure variation. (b) A transversal view of MC-piston assembly at starting configuration: a=step motor; b=motor axis; c=connection pin;d=screw; e=motor support columns; f=piston head; g=guide bar of the piston; h=connection between the MC and the motor set; i=support of the motor set; j=piston body; k=rubber o-ring; l=system support columns; m=solenoid valves connections; o=MC and p=magnetic bar. The piston course edges are indicatedby top and bottom marks. The air volume into MC with the piston at bottom position is 210mm 3 , taking account the magnetic bar volume (about 40mm 3 ). Thedimensions are expressed in millimetres. signal sent by the parallel port of the microcomputer in order tocontrolthesolenoidvalvesandthestepmotor.The378and37Ahexadecimal addresses were used to control the ON switchingtimes of the solenoid valves and to send the activation pulses tothe step motor, respectively.A transversal view of PFBA is shown in Fig. 1b. The lab- madepropulsionsystem(PS)comprisesabrasspistonconnectedmechanicallytoadot-matrixprinterstepmotorthroughascrew.This system is coupled to the MC so that the MC-PS assemblyworks similarly to an automatic syringe. A rubber o-ring placedat the extremity of the piston is used for internal sealing of MC.The full-step motor operation was chosen in order to providean enough torque to displace the piston. In this operation mode,two coils are simultaneously activated, generating a strongermagnetic field than the yielded by one coil activation.The transmission lines were constructed with 0.8mm i.d.PTFE tubing. The 2.0mL laboratory-made MC was also con-structed in Teflon ® . 2.4. Procedure The preparation of each calibration solution is always initial-izedwithallthevalvesswitchedOFFandthepistonplacedonthebottom position (Fig. 1b). After that, 48 pulses (see Section 3.1) areappliedtothestepmotorwithallthevalvesswitchedOFFinorder to displace the piston towards the top position, yielding aslightlowerpressureinsidetheMCregardingtotheatmosphericpressure. Afterwards, the V CS  valve and the piston are simulta-neously switched ON during a previously defined time interval t  CS , promoting the aspiration of an aliquot of work calibrationsolution (CS) towards MC. It occurs due to the internal pressurevariation inside MC caused by the piston displacement. In thesequence,thesameprocedureisappliedtotheV B valve,whichisswitchedONduringatimeinterval t  B  andanaliquotofblank(B)isaspiratedtowardstheMC.Soonafterwards,thevalveV D/S  andthe piston are simultaneously switched ON during 6s in orderto displace the piston to the top position. This step is accom-plished in order to insert enough air inside the MC to transportthe prepared mixture into the cup of the spectrometer auto sam-pler and to empty completely the MC and analytical path. Thistransport is always carried out by ON switching the V D/S  valveandthepistonduring16swithreversalpistondisplacement(topto bottom).All calibration solutions are prepared by using the sameprocedure, however, with  t  CS  increasing (1–9s),  t  B  decreasing(9–1s) and  t  CS  + t  B  =10s.If a dilution of the sample (S) is needed in order to obtain ananalytical signal into the linear range of calibration curve, it canbe carried out by replacing CS from the sample and by accom-plishing the same procedure of calibration solution preparationwith the appropriated time intervals for  t  S  and  t  B  regarding tothe required dilution.Among the preparations of calibration solutions or sampledilutions, a cleaning step may be carried out by using onlythe blank solution and the same procedure described above.However, it was unnecessary to perform this step because nocarryover was observed, owing to the hydrophobic property of PTFEusedintheMC,valvesandtransmissionlinesofthePFBAmanifold.The magnetic stirrer was always activated during the inser-tion of the fluids aliquots into MC in order to assure a goodhomogenization of the solutions.   L.F. Almeida et al. / Talanta 73 (2007) 906–912  909 2.5. Theoretical IntheautomaticpreparationofcalibrationcurveswithPFBA,the analytical response may be directly related to the switchingON time,  t  CS , of the V CS  valve as demonstrated below.In PFBA, since  v = Qt   (where  Q  is the channel flow-rate),the valve timing courses,  t  , define the volumes,  v CS  and  v B ,added into MC. So,  t  CS ( i ) and  t  B ( i ) can be used instead of   v B ( i )and  v CS ( i ). Moreover, if significant statistical differences do notoccur among the flow-rates of the channels ( Q CS  = Q B  = Q ), thefollowing expression is valid: R CS ( i )  = KC 0CS   t  CS ( i ) t  CS ( i ) + t  B ( i )   ( i = 1 , 2 , 3 ,...,n ) (1)where  R CS ( i ) is the  i th analytical responses for the  i th concen-tration of a prepared work calibration solution,  C 0CS  and  n  is thenumber of points of the calibration curve.Inordertosimplifythemathematicalmodel,thetotaltimeof switching ON valves  t  Total  = t  CS ( i )+ t  B ( i ) in a given calibrationset is maintained constant, and Eq. (1) assumes the form below: R CS ( i ) =  KC 0CS t  Total  t  CS ( i ) (2)that can be rewritten as: R CS ( i ) = K  t  CS ( i ) (3)where  K   is the angular coefficient of the curve  R CS ( i ) versus t  CS ( i ).Therefore, it is demonstrated that a linear relation betweenthe analytical responses,  R CS ( i ), and the switching ON timeof the calibration solution valve,  t  CS ( i ), is obtained from Eq.(3) f or the constructed calibration curve with PFBA. However,the calibration curves used in analysis by ET AAS are usuallyconstructed by plotting the integrated absorbance data versusanalyte mass (nanograms, in general). The  t  CS ( i ) values may betransformed in analyte mass values,  m CS ( i ), of each preparedcalibration solution with PFBA, by using the expression below: m CS ( i ) =  v AS C 0CS t  Total  t  CS ( i ) (4)where  v AS  is the inserted volume into the graphite tube by theauto sampler of the spectrometer.SubstitutingEq.(4)in(3),thefollowingequationisobtained: R CS ( i ) = K   1 v AS  m CS ( i ) (5)making  θ   = K (1 /v AS ) Eq. (5) can be rewritten as: R CS ( i ) = θm CS ( i ) (6)Finally, if a sample dilution is carried out with PFBA, theanalyte concentration in the sample flask,  C 0S , can be calculatedby Eq. (6) below: C 0S  =  t  CS + t  B t  CS  R S θ  (7)where  t  CS  and  t  B  are the switching ON time V CS  and V B  valves,respectively, during the sample measurement,  R S , and  θ   is theangular coefficient of Eq. (6) estimated by linear least-squares regression fitting. 3. Results and discussion 3.1. The effect of the air elasticity inside the MC  The amounts of blank and work calibration (or sample) solu-tions inserted into the MC depend on the ON switching time of the V B  or V CS  valves when the piston is displaced towards thetop position (Fig. 1b). However, the air elasticity inside the MC impairstheinstantaneoustransmissionofmovementtothefluidswhen the piston starts its displacement. In order to compensatethis effect, the piston must be initially displaced to a suitableposition by applying an adequate velocity of piston displace-ment towards the top position. To find this position and velocity,it was investigated the number and the frequency of pulses to beapplied to the step motor. It was verified that the application of 48 pulses with a frequency of 50Hz was enough to compensatethe air elasticity inside the MC. 3.2. The flow-rate of channels The flow-rates of the blank and work calibration solutionchannels were estimated by aspirating water into the MC dur-ing interval times from 1 to 5s and by weighting the sampledwateraliquots.ThewaterintotheMChasalwaysbeenaspiratedafter the air elasticity compensation (48 pulses at 50Hz). Byusing the well known density values of the water at monitoredtemperature (21–23 ◦ C), the volumes were calculated and theflow-rates for different pulse frequencies (50–200Hz) appliedto the step motor were estimated from the volume versus timecurves (Fig. 2).No significant statistical difference was observed at a 95%confidence level between the flow-rates of the blank and work calibration solution channels for each different pulse frequen-cies applied to the step motor (Table 2). Thus, the condition Q B  = Q CS  = Q , proposed in the PFBA mathematical modellingis valid. 3.3. The choice of the pulse frequency applied to the stepmotor  To elect the pulse frequency to be applied to the step motorin fluids sampling stage (  f  S ), both the accuracy and the ana-lytical velocity inherent to each studied frequency were takeninto account. At 50Hz or less, PFBA provides a good precision,although significant losses in analytical velocity may occur. Atlarger  f  S  values (above 100Hz), PFBA provides higher analyt-ical velocities. However, the sampling is more susceptible toerrors due to the uncertainty in the valves timing control. Evenwith the rigorous controlling furnished by the control software,slight timing variations (0.05–0.1s) may occur, propagating,thus,significanterrorsincalibrationsolutionspreparationand/orsample dilutions. The pulse frequency of 100Hz has been cho-  910  L.F. Almeida et al. / Talanta 73 (2007) 906–912 Fig. 2. The volume vs. time curves for different pulse frequencies applied to the step motor used to estimate the flow-rates (Table 2) of the blank (a) and work  calibration solution (b) channels.Table 2Estimated equations from the curves of volume vs. time (Fig. 2). Volume and time are expressed in mL and seconds, respectively Pulses frequency (Hz) ChannelsBlank Work calibration solutionEquations  R 2a Equations  R 2a 50  v = 0 . 0417 t  − 0 . 0199 1.0000  v = 0 . 0417 t  − 0 . 02 0.9999100  v = 0 . 0846 t  − 0 . 0173 0.9992  v = 0 . 0847 t  − 0 . 0179 0.9998150  v = 0 . 1217 t  − 0 . 0357 0.9998  v = 0 . 1217 t  − 0 . 0355 0.9993200  v = 0 . 1578 t  − 0 . 0463 0.9984  v = 0 . 1575 t  − 0 . 0459 0.9990 a Correlation coefficient. sen as the better compromise between precision and analyticalvelocity. 3.4. PFBA dilution capability The available internal volume of the MC ( v ma ) used in thiswork is about 1.583mm 3 and it is limited by the piston dis-placement course (about 31.5mm). A portion corresponding to2/3  v ma  was adopted as the maximum work volume ( v mw ) tobe inserted into the MC of the blank plus the work calibrationsolution in order to guarantee an enough amount of air to expelthe fluids and empty the MC and the analytical path completely.As described in Section 2.5, the ON switching valves times can be used instead of volumes, thus, the volume  v mw  is corre-sponding to the work total time of piston displacement course, t  Total . Since the pulse frequency of 100Hz has been elected forPFBA operation, a  t  Total  =16s was determined.Otherparameterthatmustbetakenintoaccounttodeterminethe PFBA dilution capability is the minimum valve switchingON time ( t  min ) necessary to delivery a fluid volume into theMC without significant impairing of the analytical precision.Thereby,  t  min  =0.2s was found. By using  t  min  and  t  Total  val-ues above, the PFBA dilution capability can be determined by t  Total  /  t  min  =16/0.2=80. 3.5. Analysis by manual and PFBA procedures For comparison purposes, five calibration curves into0.2–0.5  gL − 1 Mn range were constructed using thecalibration standard solutions prepared by manual andPFBA procedures. The same working calibration solution(5.0  gL − 1 Mn) was utilized in both procedures. Theabsorbance measurements were always carried out in tripli-cate. The parameters of the constructed curves are shown inTable 3.The angular coefficients of the curves built from the applica-tion of both procedures were very similar. It can be attested bya simple comparison between the mean values of this parameterfor the manual (8.566 ± 0.4038) and PBFA (8.369 ± 0.3371)procedures. No significant statistical differences at a 95% con-fidence level were observed by applying the paired  t  -test. The Table 3Angular( b )andcorrelation(  R 2 )coefficientsandcharacteristicmass( m 0 ),limitsofdetection(LOD=3 s blank   /  b )andquantification(LOQ=10 s blank   /  b )obtainedbyPFBA and manual procedures. LOD and LOQ values were calculated based onthe blank standard deviation,  s blank   ( n =10)Method a b  (ng − 1 )  R 2 LOD (ng) LOQ (ng)  m 0  (pg)M-1 8.798 0.9982 0.0014 0.0045 0.51M-2 7.907 0.9954 0.0015 0.0050 0.51M-3 8.750 0.9986 0.0014 0.0045 0.51M-4 8.463 0.9974 0.0014 0.0047 0.50M-5 8.912 0.9982 0.0013 0.0044 0.50P-1 8.566 0.9954 0.0014 0.0046 0.51P-2 8.514 0.9979 0.0014 0.0047 0.52P-3 7.768 0.9968 0.0015 0.0051 0.50P-4 8.495 0.9849 0.0014 0.0047 0.51P-5 8.502 1.0000 0.0014 0.0047 0.44 a M and P represent the manual and PFBA preparations, respectively.
Search
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks