Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Integrated luminometer for the determination of trace metals in seawater using fluorescence, phosphorescence and chemiluminescence detection

Integrated luminometer for the determination of trace metals in seawater using fluorescence,... Department of Environmental Sciences, Plymouth Environmental Research Centre, University of Plymouth, Plymouth PL4 8AA, UK 2 Department of Physical and Analytical Chemistry, University of Oviedo, c/Julian Claver|èa 8, E-33006 Oviedo, è Spain 3 Chimie Organique Physique and 4 Laboratoire de Traitement des Eaux et Pollution, Universite Libre de Bruxelles, 50 Av. F. D. Roosevelt, 1050 Brussels, è Belgium manifolds has been determined by analysis of certi® ed seawater reference materials. The ability to perform sequential determinations of diå erent analytes in the same sample using the three diå erent modes has been eå ectively demonstrated using Scheldt estuarine water samples. Cadmium, zinc, lead and cobalt were determined using FL (Cd, Zn), RTP (Pb) and CL (Co) and the results were validated against standard laboratory methods. The paper describes an integrated luminometer able to perform £uorescence (FL), room temperature phosphorescence (RTP) and chemiluminescence (CL) measurements on seawater samples. The technical details of the instrumentation are presented together with £ow injection (FI) manifolds for the determination of cadmium and zinc (by FL), lead (RTP) and cobalt (CL). The analytical ¢gures of merit are given for each manifold and results are presented for the determination of the four trace metals in seawater reference materials (NASS-5, SLEW-2) and Scheldt estuarine water samples. Experimental Reagents Analytical reagent-grad e chemicals were used throughout and high-purity water (Milli-Q, Millipore, Watford, UK) was used for the preparation of all solutions unless otherwise stated. For the determination of Cd and Zn using FL, anthylazamacrocycl e 1 (9-(1 0 ,4 0 ,7 0 ,10 0 , 13 0 ,16 0 -hexa-azacyclooctadecyl)methyl-anthracene ) was synthesized in the laboratory in Brussels according to literature procedures [6] (for further details, see below). Optimum reagent conditions included 0.5 mm anthylazamacrocycle 1 in NaOH (Micro Select, Fluka, Buchs, Switzerland), resulting in a reaction pH 9 for Zn and 13 for Cd determinations. Dowex 1X2-400 (Supelco, Bornem, Belgium) anion-exchang e resin was cleaned before use with 1 m HCl (37% Trace Select, Fluka) for 30 min, and then with Milli-Q water for 1 h. The resin was packed in a 10-ml column (Omni® t, Cambridge, UK). Between samples, the Dowex column was cleaned by pumping a solution of 1 m NaCl/0.01 m HCl over the resin (NaCl, > 99%, ACS reagent, Aldrich, St Louis, MO, USA). About 9.5 ml NaCl (5 m) and 312 ml HCl (concentrated) per 50-ml sample was added to samples and standards prior to injection onto the column. The column was eluted using a solution of 0.75 m NaCl/ 0.065 m HCl. For the determination of Pb(II) using RTP, 8-hydroxy7-quinoline sulphonic acid (Fluka) was used without further puri® cation. Aqueous solutions of the reagent were prepared by dissolving the appropriate amount of the solid reagent in Milli-Q water. Dowex 1X2-200 anion exchange resin (Sigma, Munich, Germany) was cleaned thoroughly before use with 2 m HCl to remove trace metal impurities, then with Milli-Q water and ® nally with ethanol to displace air from the pores of the resin and to remove residual monomers and solvents. The resin was placed into the RTP ¯ ow-cell (volume about 25 ml; see below for further details). Introduction It is important to determine trace metal concentrations in marine (and freshwater) environments in order to monitor compliance with legislation and ensure good water quality [1]. There are various techniques available, but the family of luminescence techniquesÐ ¯ uorescence (FL), room temperature phosphorescence (RTP) and chemiluminescence (CL) [2]Ð oå ers the potential for highly selective and sensitive determinations [3], providing that the appropriate reagents can be found and the methods optimized. The challenge of determining trace metals in seawater environments is compounded by the potential interference from matrix constituents, particularly the major ions. This paper reports on the development of a combined luminometer that can be readily switched between the various luminescence modes to provide optimized methods for the determination of a range of trace metals. Samples and reagents are presented to one of the photomultiplier-based detectors by means of automated ¯ owinjection manifolds controlled by the instrument software. Where necessary, matrix interferences are removed by means of in-line microcolumns containing chelating or ion-exchange resins [4, 5]. The accuracy of the various *To whom correspondence should be addressed. e-mail: pworsfold@ plymouth.ac.uk Journal of Automated Methods & Management in Chemistry ISSN 1463± 9246 print/ISSN 1464± 5068 online # 2002 Taylor & Francis Ltd http://www.tandf.co.uk/journals Reagents for the FI± CL determination of Co(II) included pyrogallol (1,2,3-trihydroxybenzene , ACS grade, Aldrich, Gillingham, UK), methanol (HPLC-grade, Rathburn, Walkerburn, UK), CTAB (cetyl-trimethylammonium bromide, Microselect 99%, Fluka), sodium hydroxide and hydrogen peroxide (both Analar, Merck, Poole, UK). The optimum reagent concentrations were 50 mm pyrogallol, 25 mm CTAB, 1 m hydrogen peroxide, 20% v/v methanol and pH 10.35 (0.15 m sodium hydroxide). A 0.1 m ammonium acetate buå er (Merck) was used for on-line pH buå ering (pH 5.1) of the sample before Co preconcentration on an 8-hydroxyquinolin e (8-HQ) microcolumn. The chelating microcolumn (Perspex casing) was packed with 8-HQ immobilized on a commercial hydrophilic vinyl co-polymer (Toyopearl TSK gel, HW 75F, 32± 63 micron ® ne, TosoHaas, Stuttgart, Germany) according to [5]. The 8-HQ column contained about 0.05 ml resin and had a lifetime of at least 6 months. A 0.05 m HCl (quartz-distilled) solution was used as the microcolumn elution solution. The 8-HQ column was cleaned daily by rinsing for 10 min with 0.01 m HCl. Standard solutions for calibration of the various luminescence techniques were prepared daily from appropriate stock solutions (Spectrosol, Merck). Standard solutions for the diå erent elements were prepared separately. Certi® ed reference materials were obtained from the National Research Council of Canada. Estuarine water samples were collected from the Scheldt estuary near Antwerp, Belgium, on an ebb tide (25 July 2000). Salinity was measured in situ using a homemade salinometer. Samples were ® ltered immediately after collection through an acid-cleaned (0.1 m HCl, 24 h) 0.45 mm polycarbonate membrane ® lter (Sartorius ) in a ® eld-laboratory on a pontoon in the estuary. Samples were then acidi® ed to pH 1.5 with 200 ml HNO3 (65% Suprapur, Merck Eurolab, Ontenay Sous Bois, France) per 100 ml sample. Instrumentation The luminometer was a modi® ed version of a commercial instrument (FL/FS 900 from Edinburgh Analytical Instruments) capable of monitoring ¯ uorescence (FL) and room temperature phosphorescenc e (RTP) decay curves after pulsed excitation, FL and RTP steady-stat e signals and transient chemiluminescence (CL) emission. A block diagram of the instrumentatio n is shown in ® gure 1. Emission was monitored at 908 to the excitation beam for FL and RTP. A ¯ ipping mirror was incorporated to select the appropriate light source in front of the excitation monochromator and a special cell compartment for CL measurements was added. Excitation wavelengths (200± 750 nm) were selected by a 300-mm focal length grating (G318 HOm25) blazed at 250 nm. Emission wavelengths (200± 900 nm) were selected by a second monochromator equipped with a grating (G318HOm5) blazed at 500 nm. A Hamamatsu R1527 PMT was used to measure light intensities in the photon counting mode. The excitation source used for steady-stat e FL was a continuous 450 W Xe-arc (Xe 900, 230± 2000 nm). The light source used to record time-resolved ¯ uorescence PG900 (GATE DELAY AND GATE WIDTH) uF900 (FREQUENCY 1100 Hz) nF900 NANOSECOND FLASHLAMP (1-100 kHz) LAMP SELECTING SWING MIRROR BOX M300 EXCITATION MONOCHROMATOR (1800 gr/mm) START PMT STARTER (TCSPC) PMT POWER SUPPLY BOX PMT POWER SUPPLY BOX BLUE SENSITIVE PMT (R1527) M300 EMISSION MONOCHROMATOR (300 gr/mm) Xe900 STEADY STATE XENON LAMP SAMPLE CHAMBER CL PMT (R6095P) CL PMT POWER SUPPLY BOX Figure 1. Block diagram of the integrated luminometer. emission was an nF 900-nanosecon d ¯ ashlamp ® lled with hydrogen. The source emitted approximately 108 photons per pulse in the wavelength range 110± 850 nm with a FWHM of 0.8 ns. Its power output was > 5 W at 100 Hz; its frequency could be adjusted between 1 and 100 Hz and its FWHM was in the range 1.5± 3.0 ms. The pulses were viewed by a 9661B Electron Tubes PMT by the time-correlated single photon counting (TCSPC) technique. A mF 900 ns ¯ ashlamp served as an excitation source for monitoring RTP spectra and decay pro® les. The delay and gate times were 0.05 ms and the excitation and emission slits were set at 10 and 20 nm, respectively. CL signals were measured with an R 6095 Hamamatsu PMT. A PC controlled the luminometer. Data acquisition and analysis were performed by the FS900 CDT Edinburgh Instruments Software for CL and steady-state FL and RTP emissions. Data acquisition and analysis in the TCSPC mode were performed with the FL 900 CDT and FLA 900 level 2 software packages. Fluorescence FI manifold for cadmium and zinc The FI manifold shown in ® gure 2 was used for the sequential determination of Cd and Zn in a sample. The analytical methodologies used the same reagents, eluents and elution procedures for the two metals, with the exception of the use of a reaction pH 9 for Zn analysis and pH 13 for Cd analysis. The manifolds for Cd and Zn were identical, with three channels, one for each metal and one for the reagent solution. Zinc and Cd were therefore determined separately with a changeover of reagent solution. The solutions were pumped and injected on the Dowex 1X2-400 microcolumns and through the ¯ uorescence cell via a three-way peristaltic pump (Gilson Minipuls 3, Villiers le Bel, France). A mixing ` T’ piece (Cole Parmer, Vernon Hills, IL, USA), PTFE tubing (0.8 mm i.d.) and peristaltic pump tubing (Elkay, Shrewsbury, MA, USA) were used to connect the reservoirs, the three-way valves and the ¯ ow-through cell (450 ml, Hellma). The two metal lines were alternately connected to the waste and to the ¯ uorescence ¯ ow-through cell in such a way that during the preconcentration step, the eluted fraction was discarded NaCl 1 M / HCl 0.01 M NaCl 0.75 M / HCl 0.065 M Zn(II) / NaCl 0.75 M / HCl 0.065 water NaCl 1M / HCl 0.01 M NaCl 0.75 M / HCl 0.065 M Cd(II) / NaCl 0.75 M / HCl 0.065 M water air luminophore / NaOH V1 V2 V3 V4 Dowex waste V9 V5 V6 V7 pump 2 V8 waste Dowex V10 waste waste V11 3 flow-through cell Figure 2. FL manifold for Cd and Zn. while during the wash step, the metals were passed through the ¯ uorescence cell. Flow rates for lines 1 and 2 (® gure 2) were 2.1 and 0.9 ml min¡1 , respectively. The resin was washed for 1 min using eluent 1. The samples or standards were injected for 5 min (line 2). The metals were eluted using eluent 2 (2 min), eluent 1 (3.5 min) and Milli-Q water, successively. The eluted metals was then mixed with the reagent, and after on-line complexation the metal chelate was passed through the ¯ ow-through cell where its ¯ uorescence was measured. The ¯ uorescent measurement took about 200 s. A complete analytical cycle was performed within 17 min using a 5-min preconcentration . Ten three-way PTFE solenoid switching valves V1± 10 were used to direct the diå erent solutions sequentially through the ¯ ow-through cell. The valves were controlled via the parallel port of the PC using a CoolDrive TM 24-V DC command interface. The sequences were fully automated using an in-house computer proNaCl NaCl / HCl échantillon H2O NaCl pompe NaCl / HCl échantillon H2O E F G ® H poubelle A B poubelle C D résine 1 poubelle J résine2 analyse étape gram written in Visual Basic. The graphic interface software (® gure 3), constructed in an Excel environment, had three distinctive components. The ® rst was a table showing the diå erent sequences of elution. It distinguished between the introduction step, the main step (which could be repeated several times) and the end step. Each step was assigned a diå erent code (1, 2 or 3) for the introduction step, the main step and the end step respectively. The second part was a diagram of the manifold showing the currently active part of the manifold, e.g. valves, peristaltic pump. Clicking on the RUN button started a new run. The third part was an information panel that gave the times for the start and ® nish of each step and the nature of the current operation, e.g. preconcentration, elution. The control device for the operation of the valves was interfaced with the PC (® gure 4, item 1) via the parallel port (Centronics). Adaptation of the levels and protection of the parallel port were done by means of an octal durée totale: 1 minute boucle 1 (sur 1) conditionnement / élution NaCl temps: 1 minute 38 secondes Run ligne 1 ligne 2 Etape 1 2 3 4 5 6 7 8 Durée 10 20 60 300 120 210 330 30 Proc. 1 1 2 2 2 2 2 3 Code1 85 86 80 81 88 80 82 82 Code2 21 22 16 17 24 16 18 18 Code3 128 128 128 128 128 128 128 128 Code4 0 0 0 0 0 0 0 0 20 ### 0 293 120 210 330 0 Figure 3. Graphic interface for the control of the integrated luminometer. V1 V2 Figure 4. Control device for operation of the valves. 1 ˆ PC parallel port; 2 ˆ 74HCT541; 3 ˆ 74HCT75; 4 ˆ valve drivers 225D5X24; 5 ˆ power supply 5 V DC; 6 ˆ power supply 24 V DC. buå er and line drivers with a three-state output integrated circuit (74HCT541) (® gure 4, part 2). Since the number of valves that needed to be controlled (10) was larger than the number of data lines (eight) of the parallel port, demultiplexing and storage of the valve positions was necessary. These two functions were ful® lled by means of three integrated 74HCT75 circuits (Dual 2-Bit Bistable Transparent Latch) (® gure 4, part 3). Two valve drivers (CoolDrive 225D5X24 Valve Drivers, NResearch, Inc.) were used to power the valves (® gure 4, part 4). To minimize interference problems, two independent power supplies were used, 24 V DC for the valves and 5 V DC for the integrated circuits. Room temperature phosphorescence FI manifold for lead The FI manifold for the determination of Pb(II) is shown in ® gure 5. A four-channel peristaltic pump (Gilson Minipuls 2) was used to generate the ¯ owing streams. Rotary valves (Omni® t 1106) were used for sample and reagent introduction (valves A and B) and for the elution of the retained species. An Omni® t 2401 mixing T-piece, PTFE tubing (0.8 mm i.d.) and ® ttings were used to connect the ¯ ow-through cell, the rotary valves and the carrier solution reservoirs. Samples (2 ml) were injected via valve A and chelating reagent solution (2 ml) through valve B into the ¯ ow system. Both solutions, at a ¯ ow rate of 1.4 ml min¡1 , were mixed in a T-piece and passed through the 0.5-m reaction coil to ensure chelate formation. After on-line reaction, the anionic Pb complex passed into the detector ¯ ow cell (Hellma Model 176.52-QS, 25 ml, Mullheim, Baden, Germany), placed « inside the sample compartment of the luminometer, and was retained on a packing of the Dowex 1X2-200. At the bottom of the ¯ ow cell a small piece of nylon net was placed to prevent particle displacement by the carrier stream. The Dowex resin was loaded with the aid of a syringe and the exit end of the ¯ ow cell was kept free. Then, the cell was connected to the ¯ ow system and 10 min was allowed for the particles to settle. To ensure that the compound retained by the packing solid material was in the light path, the resin level was maintained 1 mm below the cell window. The RTP intensity was then measured at an excitation wavelength of 395 nm and an emission wavelength of 595 nm. After measurement, 500 ml 6 m HCl was injected via valve C (to strip the chelate retained on the solid phase) before proceeding with the next sample injection. Chemiluminescence FI manifold for cobalt The FI manifold used for the CL determination of Co(II) is shown in ® gure 6. Two peristaltic pumps (Gilson Minipuls 3) were used to deliver the sample and buå er, Milli-Q water, eluent and reagents. All manifold tubing was PTFE (0.75 mm i.d., Fisher) except for the peristaltic pump tubing, which was Tygon (Elkay). The ¯ ow rates for the various sample and reagent ¯ ows are presented in ® gure 6. A six-port PTFE rotary injection valve (Rheodyne, model 5020) was used for sample introduction. The ¯ ow cell was a quartz glass spiral (1.1 mm i.d., 130 ml internal volume) positioned in front of a mirror in the sealed housing. The reagent ¯ ow created a background CL emission of 0.15 mV with a peak-to-pea k noise of 0.1 mV. The operating procedure for the FI± CL Co manifold was as follows: all PTFE ¯ ow lines, ® ttings and connectors of the FL manifold were initially cleaned with 0.5 m quartz-distilled HCl and UHP water for at least 2 h. For the determination of Co(II) in seawater, sample was buå ered in-line to pH 5.1 with ammonium acetate (0.1 m) and loaded onto the 8-HQ chelating resin column for 60 s at 1.2 ml min ¡1 . Milli-Q water was then passed through the column for 30 s to remove the major seawater cations and anions. The injection valve was switched to the elute position for 60 s and 0.05 m HCl was passed through the column in the reverse direction at 1.2 ml min¡1 to elute the Co(II). The eluent stream then Figure 5. RTP manifold for Pb. Figure 6. CL manifold for Co. merged with the reagent solutions and passed through a 5-m reaction coil (immersed in a heated oil bath at 808C) to the ¯ ow cell. The injection valve was then returned to the load position and washed with Milli-Q water for 30 s to remove residual HCl before starting the next load sequence. escence response was linear over the range 35± 2000 pm (r 2 ˆ 0:993) and the limit of detection was 35 pm (3 s, n ˆ 5). The relative standard deviation (SD) was 10% with a signal-to-noise ratio ˆ 30 for a 200-pm Cd(II) standard. Variation of the preconcentration time from 5 to 20 min for a 500-pm Cd solution also gave a linear response (r 2 ˆ 0:998), demonstrating that this was a good variable for altering the linear range to suit particular environmental conditions. The accuracy of the FI± FL technique for Cd was determined using marine-certi® ed water reference materials (SLEW-2, NASS-5). A small baseline drift was observed during the analysis and corrected by a linear adjustment of the baseline before and after the elution peak. The results for SLEW-2 and NASS-5 obtained after this linear correction are given in Results and discussion Fluorescence determination of cadmium and zinc The method used to determine Cd and Zn was based on the ¯ uorescence modi® cation resulting from the binding of these metals to a polyazacrown covalently linked to a luminophore via one methylene group. In this work, the anthylazamacrocycle 1 (® gure 7) was used [6, 7]. In aqueous alkaline conditions, the ¯ uorescence yield of 1 was very small. This was ascribed to an intramolecular quenching by electron transfer from the amines to the excited ¯ uorophore. When chelated to non-quenching metal ions such as Cd or Zn, the amine lone pairs become involved in bonding and cannot donate an electron to the excited state of the anthryl group. Consequently, chelation-enhanced ¯ uorescence and a bathochromi c shift of the spectra are observed. Elements including Cu(II), Hg(II) and Pb(II) interfere with the FL determination of Cd and Zn. The use of the Dowex 1X2-400 column not only resulted in a preconcentration of Cd and Zn on the column, but also allowed metals separation to prevent ¯ uorescence quenching by other trace metals. Cadmium. For Cd(II) standards and blanks prepared in Milli-Q water and preconcentrated for 5 min, the ¯ uor- NH NH N HN HN HN Figure 7. Structure of the anthylazamacrocycle 1 used for the determination of Cd and Zn. table 1 and show good agreement with certi® ed values [8]. Zinc. The ¯ uorescence response was linear over the range 4± 100 nm (r 2 ˆ 0:985) for a 5-min preconcentration and the limit of detection was 4 nm (3 s, n ˆ 5). The relative SD was 8% for a 50-nm Zn(II) standard. Variation of the preconcentration time from 30 s to 5 min gave a linear response (r 2 ˆ 0:980) for a 100-nm Zn(II) standard and showed that preconcentration time could be reduced when determining Zn in highly contaminated samples. The high detection limit for Zn compared with that obtained for Cd can be explained by a higher background signal due to the residual ¯ uorescence of the reagent at pH 9 and higher external contamination . The choice of CRMs to determine the accuracy of the method was therefore restricted to SLEW-2 and a coastal seawater sample from the Irish Sea. The data obtained for SLEW-2 and for the Irish Sea are shown in table 2 and they are in good agreement with certi® ed values. Cd in SLEW-2 and the Irish Sea was 50± 100 times less than Zn and therefore it did not interfere with the Zn determination. Room temperature phosphorescence determination of lead The RTP method for the determination of Pb in seawater was based on the formation in a FI manifold (® gure 5) of a Pb chelate with 8-hydroxy-7-quinolinesulphoni c acid [9], which was then on-line immobilized on the anionic exchange resin Dowex 1X2-200. A ¯ ow rate of 1.4 ml min¡1 , sample injection loop of 2 ml and a reaction coil of 500 ml ensured the complete formation of the metal chelate. Once the metal chelate was retained onto the Dowex resin and the RTP measurements were taken, the chelate was removed from the resin by 500 ml 6 m HCl. The maximum sampling frequency that can be achieved with the proposed optosensing method for Pb analysis was 12 samples h¡1 . RTP measurements of the metal chelate were performed in the absence of oxygen (solutions contained 3 £ 10¡3 m sodium sulphite). A delay time of 0.05 ms was used in all RTP measurements to Table 1. Cadmium concentrations in the CRMs SLEW-2 and NASS-5. CRM SLEW-2 NASS-5 FI± FL (pm) 190 § 34 213 § 18 Certi® ed value (pm) 170 § 20 200 § 30 ensure that any background ¯ uorescence had ceased before RTP measurement. When using shorter delay times (i.e. 0.01± 0.04 ms), residual ¯ uorescence from the complex and the solid support itself was observed. The excitation spectrum of the complex showed intense absorption bands at 395 nm, and in the absence of oxygen, the RTP emission spectrum showed a maximum at 595 nm. The limit of detection was 0.5 nm, the relative SD was § 3% (n ˆ 4) for a 0.5-mm Pb(II) standard solution. Moreover, a good regression coeæ cient for the FI-RTP calibration graph was obtained (r ˆ 0:9989) over the range 0.5± 10 000 nm. A detailed study of potential interferences on the RTP determination of Pb(II) was performed. The presence of other heavy metals such as Zn, Cd and Co at concentrations several times higher than observed in seawater did not aå ect Pb(II) determination. Therefore, it was possible to determine Pb(II) using RTP in seawater in the presence of Cd, Zn, Pb and Co. Chemiluminescence determination of cobalt This FI-CL method uses the oxidation of pyrogallol in the presence of methanol and cetyltrimethylammoniu m bromide, with a basic solution of hydrogen peroxide as the oxidant [9, 10]. A detection limit of 5 pm (3 s of the baseline noise) was achieved using the FI-CL manifold shown in ® gure 6, with a linear range (r 2 consistently > 0:999) of 0.5± 850 pm. This shows that the method was suitable for the determination of Co(II) in coastal and open ocean waters. One sequence of triplicate injections of sample and three standard additions with a 60-s preconcentration (sample load) took 40 min. In interference studies [10], the only element showing any detectable interference with the Co FI-CL method was Ag(I), at concentrations 50 times the typical Ag concentration in open ocean seawater. Because the Ag concentration in seawater is typically 0.5± 35 pm, it is therefore unlikely that it would interfere in the analysis. The accuracy of the method was determined by analysing three marine CRMs and an Irish Seawater sample and the results (table 3) show good agreement with the certi® ed values (for the CRMs) and with a cathodic stripping voltammetry (CSV) method (for the Irish seawater). Scheldt study The Scheldt is a perturbed estuarine system situated on the border between Belgium and The Netherlands. The River Scheldt originates in France and receives high Table 3. Cobalt concentrations in the CRMs NASS-4, CASS-3 and SLEW-2 and an Irish Sea sample. Sample NASS-4 CASS-3 SLEW-2 Irish seawater FI± CL (nm) Certi® ed value (nm) CSV (nm) 0:16 § 0:01 0:60 § 0:09 0:93 § 0:13 0:35 § 0:02 0:15 § 0:02 0:68 § 0:11 0:87 § 0:21 ± ± ± ± 0:34 § 0:01 Numbers are the mean §2 s (n ˆ 4). Table 2. Zinc concentrations in the CRM SLEW-2 and an Irish Seawater sample. CRM SLEW-2 SLEW-2 ‡ 50 nm Zn(II) Irish seawater FI± FL (pm) 15:0 § 2:5 65 § 7 30 § 6 Certi® ed value (pm) 16:8 § 2:1 67 27± 29 Numbers are the mean § 2 s. 46 Numbers are the mean § 2 s (n ˆ 4). 1.2 1.0 Cd (nM) 0.8 0.6 0.4 0.2 0.0 10:00 Cd by FL Salinity 5 4 Salinity 3 2 1 0 12:00 14:00 16:00 18:00 Sampling Time (GMT+1) Table 4. Dissolved Pb(II) in Scheldt samples determined using FIºRTP and ICPºMS. Sample time (GMT ‡ 1 h) 13:53 14:53 15:53 FI-RTP (mg l¡1 ) 1.2 1.3 n.d. ICP± MS (mg l¡1 ) 0.9 1.2 n.d. Figure 8. Time series for dissolved Cd in the Scheldt with FL analysis. 300 250 Zn (nM) 200 150 100 50 0 10:00 Zn by FL Salinity 5 4 3 2 1 0 12:00 14:00 16:00 18:00 Sampling Time (GMT+1) Salinity by passing 500 ml of the water samples on-line through a microcolumn packed with 8-hydroxyquinolin e (8-HQ) immobilized on a commercial hydrophilic vinyl co-polymer (Toyopearl TSK gel). After preconcentration, the Pb was eluted from the column by injecting 2 ml HCl (0.5 m), and this solution was subsequently injected in the FI manifold. The data for dissolved Pb are shown in table 4. Dissolved Pb and Co were also determined using ICPMS and CSV respectively and the results showed good agreement with the data from the luminometer. Conclusions The three diå erent luminescence techniques (FL, RTP, CL) can be combined in a single instrument and ¯ ow injection manifolds can be used to pretreat and deliver samples to the relevant detector. Microcolumns are eå ective for removing interfering seawater matrix ions. The sensitivity is suitable for the determination of a range of trace metals (Cd, Zn, Pb, Co) in estuarine and coastal (and open ocean) waters with good accuracy. Supporting technical data can be found at the MEMOSEA website [3]. Figure 9. Time series of dissolved zinc in the Scheldt with FL analysis. 3.5 3.0 Co (nM) 2.5 2.0 1.5 1.0 0.5 0.0 10:00 Co by CL Co by C SV Salinity 5 4 3 2 1 0 12:00 14:00 16:00 18:00 Sampling Time (GMT+ 1) Salinity Acknowledgements The authors thank the EU for funding under the MAST Programme (Grant No. MAS3-CT97-0143, MEMOSEA). Figure 10. Time series of dissolved Co in the Scheldt with CL and comparative CSV analyses. levels of waste water inputs. Samples were collected during an outgoing tide in the summer of 2000 as described above, and dissolved trace metal concentrations were determined following sample ® ltration at the University of Brussels, using the integrated luminometer. The time series data for dissolved Cd, Zn and Co are shown in ® gures 8± 10 respectively. The Cd-salinity relationship was linear (r 2 ˆ 0:95) and was assigned to the desorption of Cd from suspended solid matters due to the increasing concentration of chloride at high tide during the mixing of freshwater with seawater with the formation of dissolved Cd chloride complexes. Owing to the low concentrations of dissolved Pb present in the water samples, a preconcentration step before the analysis was necessary. The preconcentration was carried out http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Automated Methods and Management in Chemistry Hindawi Publishing Corporation

Integrated luminometer for the determination of trace metals in seawater using fluorescence, phosphorescence and chemiluminescence detection

Loading next page...
 
/lp/hindawi-publishing-corporation/integrated-luminometer-for-the-determination-of-trace-metals-in-MkJPstpBHS
Publisher
Hindawi Publishing Corporation
Copyright
Copyright © 2002 Hindawi Publishing Corporation.
ISSN
1463-9246
eISSN
1464-5068
Publisher site
See Article on Publisher Site

Abstract

Department of Environmental Sciences, Plymouth Environmental Research Centre, University of Plymouth, Plymouth PL4 8AA, UK 2 Department of Physical and Analytical Chemistry, University of Oviedo, c/Julian Claver|èa 8, E-33006 Oviedo, è Spain 3 Chimie Organique Physique and 4 Laboratoire de Traitement des Eaux et Pollution, Universite Libre de Bruxelles, 50 Av. F. D. Roosevelt, 1050 Brussels, è Belgium manifolds has been determined by analysis of certi® ed seawater reference materials. The ability to perform sequential determinations of diå erent analytes in the same sample using the three diå erent modes has been eå ectively demonstrated using Scheldt estuarine water samples. Cadmium, zinc, lead and cobalt were determined using FL (Cd, Zn), RTP (Pb) and CL (Co) and the results were validated against standard laboratory methods. The paper describes an integrated luminometer able to perform £uorescence (FL), room temperature phosphorescence (RTP) and chemiluminescence (CL) measurements on seawater samples. The technical details of the instrumentation are presented together with £ow injection (FI) manifolds for the determination of cadmium and zinc (by FL), lead (RTP) and cobalt (CL). The analytical ¢gures of merit are given for each manifold and results are presented for the determination of the four trace metals in seawater reference materials (NASS-5, SLEW-2) and Scheldt estuarine water samples. Experimental Reagents Analytical reagent-grad e chemicals were used throughout and high-purity water (Milli-Q, Millipore, Watford, UK) was used for the preparation of all solutions unless otherwise stated. For the determination of Cd and Zn using FL, anthylazamacrocycl e 1 (9-(1 0 ,4 0 ,7 0 ,10 0 , 13 0 ,16 0 -hexa-azacyclooctadecyl)methyl-anthracene ) was synthesized in the laboratory in Brussels according to literature procedures [6] (for further details, see below). Optimum reagent conditions included 0.5 mm anthylazamacrocycle 1 in NaOH (Micro Select, Fluka, Buchs, Switzerland), resulting in a reaction pH 9 for Zn and 13 for Cd determinations. Dowex 1X2-400 (Supelco, Bornem, Belgium) anion-exchang e resin was cleaned before use with 1 m HCl (37% Trace Select, Fluka) for 30 min, and then with Milli-Q water for 1 h. The resin was packed in a 10-ml column (Omni® t, Cambridge, UK). Between samples, the Dowex column was cleaned by pumping a solution of 1 m NaCl/0.01 m HCl over the resin (NaCl, > 99%, ACS reagent, Aldrich, St Louis, MO, USA). About 9.5 ml NaCl (5 m) and 312 ml HCl (concentrated) per 50-ml sample was added to samples and standards prior to injection onto the column. The column was eluted using a solution of 0.75 m NaCl/ 0.065 m HCl. For the determination of Pb(II) using RTP, 8-hydroxy7-quinoline sulphonic acid (Fluka) was used without further puri® cation. Aqueous solutions of the reagent were prepared by dissolving the appropriate amount of the solid reagent in Milli-Q water. Dowex 1X2-200 anion exchange resin (Sigma, Munich, Germany) was cleaned thoroughly before use with 2 m HCl to remove trace metal impurities, then with Milli-Q water and ® nally with ethanol to displace air from the pores of the resin and to remove residual monomers and solvents. The resin was placed into the RTP ¯ ow-cell (volume about 25 ml; see below for further details). Introduction It is important to determine trace metal concentrations in marine (and freshwater) environments in order to monitor compliance with legislation and ensure good water quality [1]. There are various techniques available, but the family of luminescence techniquesÐ ¯ uorescence (FL), room temperature phosphorescence (RTP) and chemiluminescence (CL) [2]Ð oå ers the potential for highly selective and sensitive determinations [3], providing that the appropriate reagents can be found and the methods optimized. The challenge of determining trace metals in seawater environments is compounded by the potential interference from matrix constituents, particularly the major ions. This paper reports on the development of a combined luminometer that can be readily switched between the various luminescence modes to provide optimized methods for the determination of a range of trace metals. Samples and reagents are presented to one of the photomultiplier-based detectors by means of automated ¯ owinjection manifolds controlled by the instrument software. Where necessary, matrix interferences are removed by means of in-line microcolumns containing chelating or ion-exchange resins [4, 5]. The accuracy of the various *To whom correspondence should be addressed. e-mail: pworsfold@ plymouth.ac.uk Journal of Automated Methods & Management in Chemistry ISSN 1463± 9246 print/ISSN 1464± 5068 online # 2002 Taylor & Francis Ltd http://www.tandf.co.uk/journals Reagents for the FI± CL determination of Co(II) included pyrogallol (1,2,3-trihydroxybenzene , ACS grade, Aldrich, Gillingham, UK), methanol (HPLC-grade, Rathburn, Walkerburn, UK), CTAB (cetyl-trimethylammonium bromide, Microselect 99%, Fluka), sodium hydroxide and hydrogen peroxide (both Analar, Merck, Poole, UK). The optimum reagent concentrations were 50 mm pyrogallol, 25 mm CTAB, 1 m hydrogen peroxide, 20% v/v methanol and pH 10.35 (0.15 m sodium hydroxide). A 0.1 m ammonium acetate buå er (Merck) was used for on-line pH buå ering (pH 5.1) of the sample before Co preconcentration on an 8-hydroxyquinolin e (8-HQ) microcolumn. The chelating microcolumn (Perspex casing) was packed with 8-HQ immobilized on a commercial hydrophilic vinyl co-polymer (Toyopearl TSK gel, HW 75F, 32± 63 micron ® ne, TosoHaas, Stuttgart, Germany) according to [5]. The 8-HQ column contained about 0.05 ml resin and had a lifetime of at least 6 months. A 0.05 m HCl (quartz-distilled) solution was used as the microcolumn elution solution. The 8-HQ column was cleaned daily by rinsing for 10 min with 0.01 m HCl. Standard solutions for calibration of the various luminescence techniques were prepared daily from appropriate stock solutions (Spectrosol, Merck). Standard solutions for the diå erent elements were prepared separately. Certi® ed reference materials were obtained from the National Research Council of Canada. Estuarine water samples were collected from the Scheldt estuary near Antwerp, Belgium, on an ebb tide (25 July 2000). Salinity was measured in situ using a homemade salinometer. Samples were ® ltered immediately after collection through an acid-cleaned (0.1 m HCl, 24 h) 0.45 mm polycarbonate membrane ® lter (Sartorius ) in a ® eld-laboratory on a pontoon in the estuary. Samples were then acidi® ed to pH 1.5 with 200 ml HNO3 (65% Suprapur, Merck Eurolab, Ontenay Sous Bois, France) per 100 ml sample. Instrumentation The luminometer was a modi® ed version of a commercial instrument (FL/FS 900 from Edinburgh Analytical Instruments) capable of monitoring ¯ uorescence (FL) and room temperature phosphorescenc e (RTP) decay curves after pulsed excitation, FL and RTP steady-stat e signals and transient chemiluminescence (CL) emission. A block diagram of the instrumentatio n is shown in ® gure 1. Emission was monitored at 908 to the excitation beam for FL and RTP. A ¯ ipping mirror was incorporated to select the appropriate light source in front of the excitation monochromator and a special cell compartment for CL measurements was added. Excitation wavelengths (200± 750 nm) were selected by a 300-mm focal length grating (G318 HOm25) blazed at 250 nm. Emission wavelengths (200± 900 nm) were selected by a second monochromator equipped with a grating (G318HOm5) blazed at 500 nm. A Hamamatsu R1527 PMT was used to measure light intensities in the photon counting mode. The excitation source used for steady-stat e FL was a continuous 450 W Xe-arc (Xe 900, 230± 2000 nm). The light source used to record time-resolved ¯ uorescence PG900 (GATE DELAY AND GATE WIDTH) uF900 (FREQUENCY 1100 Hz) nF900 NANOSECOND FLASHLAMP (1-100 kHz) LAMP SELECTING SWING MIRROR BOX M300 EXCITATION MONOCHROMATOR (1800 gr/mm) START PMT STARTER (TCSPC) PMT POWER SUPPLY BOX PMT POWER SUPPLY BOX BLUE SENSITIVE PMT (R1527) M300 EMISSION MONOCHROMATOR (300 gr/mm) Xe900 STEADY STATE XENON LAMP SAMPLE CHAMBER CL PMT (R6095P) CL PMT POWER SUPPLY BOX Figure 1. Block diagram of the integrated luminometer. emission was an nF 900-nanosecon d ¯ ashlamp ® lled with hydrogen. The source emitted approximately 108 photons per pulse in the wavelength range 110± 850 nm with a FWHM of 0.8 ns. Its power output was > 5 W at 100 Hz; its frequency could be adjusted between 1 and 100 Hz and its FWHM was in the range 1.5± 3.0 ms. The pulses were viewed by a 9661B Electron Tubes PMT by the time-correlated single photon counting (TCSPC) technique. A mF 900 ns ¯ ashlamp served as an excitation source for monitoring RTP spectra and decay pro® les. The delay and gate times were 0.05 ms and the excitation and emission slits were set at 10 and 20 nm, respectively. CL signals were measured with an R 6095 Hamamatsu PMT. A PC controlled the luminometer. Data acquisition and analysis were performed by the FS900 CDT Edinburgh Instruments Software for CL and steady-state FL and RTP emissions. Data acquisition and analysis in the TCSPC mode were performed with the FL 900 CDT and FLA 900 level 2 software packages. Fluorescence FI manifold for cadmium and zinc The FI manifold shown in ® gure 2 was used for the sequential determination of Cd and Zn in a sample. The analytical methodologies used the same reagents, eluents and elution procedures for the two metals, with the exception of the use of a reaction pH 9 for Zn analysis and pH 13 for Cd analysis. The manifolds for Cd and Zn were identical, with three channels, one for each metal and one for the reagent solution. Zinc and Cd were therefore determined separately with a changeover of reagent solution. The solutions were pumped and injected on the Dowex 1X2-400 microcolumns and through the ¯ uorescence cell via a three-way peristaltic pump (Gilson Minipuls 3, Villiers le Bel, France). A mixing ` T’ piece (Cole Parmer, Vernon Hills, IL, USA), PTFE tubing (0.8 mm i.d.) and peristaltic pump tubing (Elkay, Shrewsbury, MA, USA) were used to connect the reservoirs, the three-way valves and the ¯ ow-through cell (450 ml, Hellma). The two metal lines were alternately connected to the waste and to the ¯ uorescence ¯ ow-through cell in such a way that during the preconcentration step, the eluted fraction was discarded NaCl 1 M / HCl 0.01 M NaCl 0.75 M / HCl 0.065 M Zn(II) / NaCl 0.75 M / HCl 0.065 water NaCl 1M / HCl 0.01 M NaCl 0.75 M / HCl 0.065 M Cd(II) / NaCl 0.75 M / HCl 0.065 M water air luminophore / NaOH V1 V2 V3 V4 Dowex waste V9 V5 V6 V7 pump 2 V8 waste Dowex V10 waste waste V11 3 flow-through cell Figure 2. FL manifold for Cd and Zn. while during the wash step, the metals were passed through the ¯ uorescence cell. Flow rates for lines 1 and 2 (® gure 2) were 2.1 and 0.9 ml min¡1 , respectively. The resin was washed for 1 min using eluent 1. The samples or standards were injected for 5 min (line 2). The metals were eluted using eluent 2 (2 min), eluent 1 (3.5 min) and Milli-Q water, successively. The eluted metals was then mixed with the reagent, and after on-line complexation the metal chelate was passed through the ¯ ow-through cell where its ¯ uorescence was measured. The ¯ uorescent measurement took about 200 s. A complete analytical cycle was performed within 17 min using a 5-min preconcentration . Ten three-way PTFE solenoid switching valves V1± 10 were used to direct the diå erent solutions sequentially through the ¯ ow-through cell. The valves were controlled via the parallel port of the PC using a CoolDrive TM 24-V DC command interface. The sequences were fully automated using an in-house computer proNaCl NaCl / HCl échantillon H2O NaCl pompe NaCl / HCl échantillon H2O E F G ® H poubelle A B poubelle C D résine 1 poubelle J résine2 analyse étape gram written in Visual Basic. The graphic interface software (® gure 3), constructed in an Excel environment, had three distinctive components. The ® rst was a table showing the diå erent sequences of elution. It distinguished between the introduction step, the main step (which could be repeated several times) and the end step. Each step was assigned a diå erent code (1, 2 or 3) for the introduction step, the main step and the end step respectively. The second part was a diagram of the manifold showing the currently active part of the manifold, e.g. valves, peristaltic pump. Clicking on the RUN button started a new run. The third part was an information panel that gave the times for the start and ® nish of each step and the nature of the current operation, e.g. preconcentration, elution. The control device for the operation of the valves was interfaced with the PC (® gure 4, item 1) via the parallel port (Centronics). Adaptation of the levels and protection of the parallel port were done by means of an octal durée totale: 1 minute boucle 1 (sur 1) conditionnement / élution NaCl temps: 1 minute 38 secondes Run ligne 1 ligne 2 Etape 1 2 3 4 5 6 7 8 Durée 10 20 60 300 120 210 330 30 Proc. 1 1 2 2 2 2 2 3 Code1 85 86 80 81 88 80 82 82 Code2 21 22 16 17 24 16 18 18 Code3 128 128 128 128 128 128 128 128 Code4 0 0 0 0 0 0 0 0 20 ### 0 293 120 210 330 0 Figure 3. Graphic interface for the control of the integrated luminometer. V1 V2 Figure 4. Control device for operation of the valves. 1 ˆ PC parallel port; 2 ˆ 74HCT541; 3 ˆ 74HCT75; 4 ˆ valve drivers 225D5X24; 5 ˆ power supply 5 V DC; 6 ˆ power supply 24 V DC. buå er and line drivers with a three-state output integrated circuit (74HCT541) (® gure 4, part 2). Since the number of valves that needed to be controlled (10) was larger than the number of data lines (eight) of the parallel port, demultiplexing and storage of the valve positions was necessary. These two functions were ful® lled by means of three integrated 74HCT75 circuits (Dual 2-Bit Bistable Transparent Latch) (® gure 4, part 3). Two valve drivers (CoolDrive 225D5X24 Valve Drivers, NResearch, Inc.) were used to power the valves (® gure 4, part 4). To minimize interference problems, two independent power supplies were used, 24 V DC for the valves and 5 V DC for the integrated circuits. Room temperature phosphorescence FI manifold for lead The FI manifold for the determination of Pb(II) is shown in ® gure 5. A four-channel peristaltic pump (Gilson Minipuls 2) was used to generate the ¯ owing streams. Rotary valves (Omni® t 1106) were used for sample and reagent introduction (valves A and B) and for the elution of the retained species. An Omni® t 2401 mixing T-piece, PTFE tubing (0.8 mm i.d.) and ® ttings were used to connect the ¯ ow-through cell, the rotary valves and the carrier solution reservoirs. Samples (2 ml) were injected via valve A and chelating reagent solution (2 ml) through valve B into the ¯ ow system. Both solutions, at a ¯ ow rate of 1.4 ml min¡1 , were mixed in a T-piece and passed through the 0.5-m reaction coil to ensure chelate formation. After on-line reaction, the anionic Pb complex passed into the detector ¯ ow cell (Hellma Model 176.52-QS, 25 ml, Mullheim, Baden, Germany), placed « inside the sample compartment of the luminometer, and was retained on a packing of the Dowex 1X2-200. At the bottom of the ¯ ow cell a small piece of nylon net was placed to prevent particle displacement by the carrier stream. The Dowex resin was loaded with the aid of a syringe and the exit end of the ¯ ow cell was kept free. Then, the cell was connected to the ¯ ow system and 10 min was allowed for the particles to settle. To ensure that the compound retained by the packing solid material was in the light path, the resin level was maintained 1 mm below the cell window. The RTP intensity was then measured at an excitation wavelength of 395 nm and an emission wavelength of 595 nm. After measurement, 500 ml 6 m HCl was injected via valve C (to strip the chelate retained on the solid phase) before proceeding with the next sample injection. Chemiluminescence FI manifold for cobalt The FI manifold used for the CL determination of Co(II) is shown in ® gure 6. Two peristaltic pumps (Gilson Minipuls 3) were used to deliver the sample and buå er, Milli-Q water, eluent and reagents. All manifold tubing was PTFE (0.75 mm i.d., Fisher) except for the peristaltic pump tubing, which was Tygon (Elkay). The ¯ ow rates for the various sample and reagent ¯ ows are presented in ® gure 6. A six-port PTFE rotary injection valve (Rheodyne, model 5020) was used for sample introduction. The ¯ ow cell was a quartz glass spiral (1.1 mm i.d., 130 ml internal volume) positioned in front of a mirror in the sealed housing. The reagent ¯ ow created a background CL emission of 0.15 mV with a peak-to-pea k noise of 0.1 mV. The operating procedure for the FI± CL Co manifold was as follows: all PTFE ¯ ow lines, ® ttings and connectors of the FL manifold were initially cleaned with 0.5 m quartz-distilled HCl and UHP water for at least 2 h. For the determination of Co(II) in seawater, sample was buå ered in-line to pH 5.1 with ammonium acetate (0.1 m) and loaded onto the 8-HQ chelating resin column for 60 s at 1.2 ml min ¡1 . Milli-Q water was then passed through the column for 30 s to remove the major seawater cations and anions. The injection valve was switched to the elute position for 60 s and 0.05 m HCl was passed through the column in the reverse direction at 1.2 ml min¡1 to elute the Co(II). The eluent stream then Figure 5. RTP manifold for Pb. Figure 6. CL manifold for Co. merged with the reagent solutions and passed through a 5-m reaction coil (immersed in a heated oil bath at 808C) to the ¯ ow cell. The injection valve was then returned to the load position and washed with Milli-Q water for 30 s to remove residual HCl before starting the next load sequence. escence response was linear over the range 35± 2000 pm (r 2 ˆ 0:993) and the limit of detection was 35 pm (3 s, n ˆ 5). The relative standard deviation (SD) was 10% with a signal-to-noise ratio ˆ 30 for a 200-pm Cd(II) standard. Variation of the preconcentration time from 5 to 20 min for a 500-pm Cd solution also gave a linear response (r 2 ˆ 0:998), demonstrating that this was a good variable for altering the linear range to suit particular environmental conditions. The accuracy of the FI± FL technique for Cd was determined using marine-certi® ed water reference materials (SLEW-2, NASS-5). A small baseline drift was observed during the analysis and corrected by a linear adjustment of the baseline before and after the elution peak. The results for SLEW-2 and NASS-5 obtained after this linear correction are given in Results and discussion Fluorescence determination of cadmium and zinc The method used to determine Cd and Zn was based on the ¯ uorescence modi® cation resulting from the binding of these metals to a polyazacrown covalently linked to a luminophore via one methylene group. In this work, the anthylazamacrocycle 1 (® gure 7) was used [6, 7]. In aqueous alkaline conditions, the ¯ uorescence yield of 1 was very small. This was ascribed to an intramolecular quenching by electron transfer from the amines to the excited ¯ uorophore. When chelated to non-quenching metal ions such as Cd or Zn, the amine lone pairs become involved in bonding and cannot donate an electron to the excited state of the anthryl group. Consequently, chelation-enhanced ¯ uorescence and a bathochromi c shift of the spectra are observed. Elements including Cu(II), Hg(II) and Pb(II) interfere with the FL determination of Cd and Zn. The use of the Dowex 1X2-400 column not only resulted in a preconcentration of Cd and Zn on the column, but also allowed metals separation to prevent ¯ uorescence quenching by other trace metals. Cadmium. For Cd(II) standards and blanks prepared in Milli-Q water and preconcentrated for 5 min, the ¯ uor- NH NH N HN HN HN Figure 7. Structure of the anthylazamacrocycle 1 used for the determination of Cd and Zn. table 1 and show good agreement with certi® ed values [8]. Zinc. The ¯ uorescence response was linear over the range 4± 100 nm (r 2 ˆ 0:985) for a 5-min preconcentration and the limit of detection was 4 nm (3 s, n ˆ 5). The relative SD was 8% for a 50-nm Zn(II) standard. Variation of the preconcentration time from 30 s to 5 min gave a linear response (r 2 ˆ 0:980) for a 100-nm Zn(II) standard and showed that preconcentration time could be reduced when determining Zn in highly contaminated samples. The high detection limit for Zn compared with that obtained for Cd can be explained by a higher background signal due to the residual ¯ uorescence of the reagent at pH 9 and higher external contamination . The choice of CRMs to determine the accuracy of the method was therefore restricted to SLEW-2 and a coastal seawater sample from the Irish Sea. The data obtained for SLEW-2 and for the Irish Sea are shown in table 2 and they are in good agreement with certi® ed values. Cd in SLEW-2 and the Irish Sea was 50± 100 times less than Zn and therefore it did not interfere with the Zn determination. Room temperature phosphorescence determination of lead The RTP method for the determination of Pb in seawater was based on the formation in a FI manifold (® gure 5) of a Pb chelate with 8-hydroxy-7-quinolinesulphoni c acid [9], which was then on-line immobilized on the anionic exchange resin Dowex 1X2-200. A ¯ ow rate of 1.4 ml min¡1 , sample injection loop of 2 ml and a reaction coil of 500 ml ensured the complete formation of the metal chelate. Once the metal chelate was retained onto the Dowex resin and the RTP measurements were taken, the chelate was removed from the resin by 500 ml 6 m HCl. The maximum sampling frequency that can be achieved with the proposed optosensing method for Pb analysis was 12 samples h¡1 . RTP measurements of the metal chelate were performed in the absence of oxygen (solutions contained 3 £ 10¡3 m sodium sulphite). A delay time of 0.05 ms was used in all RTP measurements to Table 1. Cadmium concentrations in the CRMs SLEW-2 and NASS-5. CRM SLEW-2 NASS-5 FI± FL (pm) 190 § 34 213 § 18 Certi® ed value (pm) 170 § 20 200 § 30 ensure that any background ¯ uorescence had ceased before RTP measurement. When using shorter delay times (i.e. 0.01± 0.04 ms), residual ¯ uorescence from the complex and the solid support itself was observed. The excitation spectrum of the complex showed intense absorption bands at 395 nm, and in the absence of oxygen, the RTP emission spectrum showed a maximum at 595 nm. The limit of detection was 0.5 nm, the relative SD was § 3% (n ˆ 4) for a 0.5-mm Pb(II) standard solution. Moreover, a good regression coeæ cient for the FI-RTP calibration graph was obtained (r ˆ 0:9989) over the range 0.5± 10 000 nm. A detailed study of potential interferences on the RTP determination of Pb(II) was performed. The presence of other heavy metals such as Zn, Cd and Co at concentrations several times higher than observed in seawater did not aå ect Pb(II) determination. Therefore, it was possible to determine Pb(II) using RTP in seawater in the presence of Cd, Zn, Pb and Co. Chemiluminescence determination of cobalt This FI-CL method uses the oxidation of pyrogallol in the presence of methanol and cetyltrimethylammoniu m bromide, with a basic solution of hydrogen peroxide as the oxidant [9, 10]. A detection limit of 5 pm (3 s of the baseline noise) was achieved using the FI-CL manifold shown in ® gure 6, with a linear range (r 2 consistently > 0:999) of 0.5± 850 pm. This shows that the method was suitable for the determination of Co(II) in coastal and open ocean waters. One sequence of triplicate injections of sample and three standard additions with a 60-s preconcentration (sample load) took 40 min. In interference studies [10], the only element showing any detectable interference with the Co FI-CL method was Ag(I), at concentrations 50 times the typical Ag concentration in open ocean seawater. Because the Ag concentration in seawater is typically 0.5± 35 pm, it is therefore unlikely that it would interfere in the analysis. The accuracy of the method was determined by analysing three marine CRMs and an Irish Seawater sample and the results (table 3) show good agreement with the certi® ed values (for the CRMs) and with a cathodic stripping voltammetry (CSV) method (for the Irish seawater). Scheldt study The Scheldt is a perturbed estuarine system situated on the border between Belgium and The Netherlands. The River Scheldt originates in France and receives high Table 3. Cobalt concentrations in the CRMs NASS-4, CASS-3 and SLEW-2 and an Irish Sea sample. Sample NASS-4 CASS-3 SLEW-2 Irish seawater FI± CL (nm) Certi® ed value (nm) CSV (nm) 0:16 § 0:01 0:60 § 0:09 0:93 § 0:13 0:35 § 0:02 0:15 § 0:02 0:68 § 0:11 0:87 § 0:21 ± ± ± ± 0:34 § 0:01 Numbers are the mean §2 s (n ˆ 4). Table 2. Zinc concentrations in the CRM SLEW-2 and an Irish Seawater sample. CRM SLEW-2 SLEW-2 ‡ 50 nm Zn(II) Irish seawater FI± FL (pm) 15:0 § 2:5 65 § 7 30 § 6 Certi® ed value (pm) 16:8 § 2:1 67 27± 29 Numbers are the mean § 2 s. 46 Numbers are the mean § 2 s (n ˆ 4). 1.2 1.0 Cd (nM) 0.8 0.6 0.4 0.2 0.0 10:00 Cd by FL Salinity 5 4 Salinity 3 2 1 0 12:00 14:00 16:00 18:00 Sampling Time (GMT+1) Table 4. Dissolved Pb(II) in Scheldt samples determined using FIºRTP and ICPºMS. Sample time (GMT ‡ 1 h) 13:53 14:53 15:53 FI-RTP (mg l¡1 ) 1.2 1.3 n.d. ICP± MS (mg l¡1 ) 0.9 1.2 n.d. Figure 8. Time series for dissolved Cd in the Scheldt with FL analysis. 300 250 Zn (nM) 200 150 100 50 0 10:00 Zn by FL Salinity 5 4 3 2 1 0 12:00 14:00 16:00 18:00 Sampling Time (GMT+1) Salinity by passing 500 ml of the water samples on-line through a microcolumn packed with 8-hydroxyquinolin e (8-HQ) immobilized on a commercial hydrophilic vinyl co-polymer (Toyopearl TSK gel). After preconcentration, the Pb was eluted from the column by injecting 2 ml HCl (0.5 m), and this solution was subsequently injected in the FI manifold. The data for dissolved Pb are shown in table 4. Dissolved Pb and Co were also determined using ICPMS and CSV respectively and the results showed good agreement with the data from the luminometer. Conclusions The three diå erent luminescence techniques (FL, RTP, CL) can be combined in a single instrument and ¯ ow injection manifolds can be used to pretreat and deliver samples to the relevant detector. Microcolumns are eå ective for removing interfering seawater matrix ions. The sensitivity is suitable for the determination of a range of trace metals (Cd, Zn, Pb, Co) in estuarine and coastal (and open ocean) waters with good accuracy. Supporting technical data can be found at the MEMOSEA website [3]. Figure 9. Time series of dissolved zinc in the Scheldt with FL analysis. 3.5 3.0 Co (nM) 2.5 2.0 1.5 1.0 0.5 0.0 10:00 Co by CL Co by C SV Salinity 5 4 3 2 1 0 12:00 14:00 16:00 18:00 Sampling Time (GMT+ 1) Salinity Acknowledgements The authors thank the EU for funding under the MAST Programme (Grant No. MAS3-CT97-0143, MEMOSEA). Figure 10. Time series of dissolved Co in the Scheldt with CL and comparative CSV analyses. levels of waste water inputs. Samples were collected during an outgoing tide in the summer of 2000 as described above, and dissolved trace metal concentrations were determined following sample ® ltration at the University of Brussels, using the integrated luminometer. The time series data for dissolved Cd, Zn and Co are shown in ® gures 8± 10 respectively. The Cd-salinity relationship was linear (r 2 ˆ 0:95) and was assigned to the desorption of Cd from suspended solid matters due to the increasing concentration of chloride at high tide during the mixing of freshwater with seawater with the formation of dissolved Cd chloride complexes. Owing to the low concentrations of dissolved Pb present in the water samples, a preconcentration step before the analysis was necessary. The preconcentration was carried out

Journal

Journal of Automated Methods and Management in ChemistryHindawi Publishing Corporation

Published: Sep 14, 2014

There are no references for this article.