Continuous flow methods for evaluating the ... - [ RSC ] Publishing

View Online

Continuous flow methods for evaluating the response of a copper ion selective electrode to total and free copper in seawater Ruth S. Eriksen,a Denis J. Mackey,*b Peter Alexander,a Roland De Marcoc and Xue Dong Wangd aSchool of Applied Science, University of Tasmania, PO Box 1214, Launceston, Tasmania 7250, Australia b CSIRO Division of Marine Research, PO Box 1538, Hobart, Tasmania 7001, Australia. E-mail: [email protected] cSchool of Applied Chemistry, Curtin University of Technology, GPO Box U 1987, Perth, WA 6845, Australia dSchool of Chemistry, La Trobe University, Bundoora, Victoria 3083, Australia

Downloaded on 26 August 2011 Published on 01 January 1999 on http://pubs.rsc.org | doi:10.1039/A903819A

Received 12th May 1999, Accepted 19th July 1999

This work describes the development of an instrument for measuring free and total copper in seawater by continuous flow analysis (CFA) with an Orion copper( II ) ion selective electrode (CuISE ). Sample analysis times are reduced considerably by using an extrapolation technique based on the fitting of an empirical mathematical expression to the electrode time–response curve enabling a prediction of the final equilibrium potential. CuISE measurements in seawater samples containing nanomolar levels of total copper can be very time consuming, and this predictive approach significantly reduces sample analysis time, and improves sample throughput. The time taken to measure pCu in seawater to a precision of ±0.1, using conventional potentiometry, varies considerably depending on the condition of the electrode membrane but can be reduced by a factor of 3–6 (typically from 60 to 10 min) by using the extrapolation technique in conjunction with CFA. Details are given of the protocols used for preconditioning the CuISE. The system can be used as a portable instrument for field measurements or for shipboard measurements of free copper in seawater. Extrapolated equilibrium potentials are within ±0.5 mV of true steady state values.

1 Introduction Trace metals such as copper are important elements because of their role as essential nutrients to aquatic organisms, and potential toxicants when present in excess of the trace amounts required for normal physiological functioning.1 The degree of toxicity is dependent upon the chemical form of the copper (e.g. ‘free’, inorganic, organic complexed) and the uptake pathways of the associated organisms. Sunda and Guillard2 demonstrated that the free copper ion concentration was the significant parameter in determining copper toxicity, hence this is the chemical form that is of most interest to environmental scientists. The jalpaite copper() ion selective electrode (CuISE ) has been shown to respond to the activity (concentration) of Cu2+ in seawater,3–5 and offers a simple and elegant way to make direct measurements of the free copper ion concentration. Ideally this measurement should be made in situ, to overcome the problems with sample preservation, sample aging and subsequent speciation changes that may occur when samples are brought back to the laboratory for ex situ analyses. The jalpaite CuISE is subject to interference from Hg2+ and Ag+ ions and, to a lesser extent, from Fe3+ ions. In seawater, the total concentrations of silver and mercury are lower than the total concentration of copper by about 2–3 orders of magnitude. These elements are also strongly complexed by the halide ions (and possibly organic ligands) in seawater and hence should cause no interference. Ferric ion is even less of a problem in natural seawater since complexation by hydroxide ion reduces the concentration of free Fe3+ to less than about 10−20 M. Even in UV-photooxidised seawater at pH 2, interference from Fe3+ should usually be negligible, in line with

the good agreement between ion-selective electrode (ISE) and graphite furnace atomic absorption spectrometry (GFAAS) measurements of total copper in seawater.4 Considerable work has been done on incorporating ISEs into portable field monitoring devices6–8 using techniques such as flow injection analysis (FIA) and continuous flow analysis (CFA). Flow methods are advantageous as they overcome most of the problems inherent in batch ISE measurements. For the Cu()ISE, these include sluggish response at low activities, cross contamination from one sample to another, contamination by membrane dissolution and/or silver chloride precipitation, and an observed alteration in response due to adsorption of natural organic matter present in seawater9,10 We found that the response of the CuISE was slow at the very low copper() activities that occur in seawater and that this was a significant factor in limiting sample throughput in an FIA system. CFA flow methods are inherently slower than FIA techniques, but offer distinct advantages over ‘dip’-type measurements, so we have investigated whether the throughput can be improved by using an extrapolation method. By acquiring the initial portion of the time–response curve in a CFA system, the final equilibrium potential may be accurately predicted without waiting for the establishment of a steady state signal. In this way, a sample measurement that may take an hour or more to reach a steady state signal under flow conditions, may be made within a matter of minutes. This approach has recently been used successfully with fluoride,11 cadmium, lead, iodide, chloride and nitrate ISEs12 and the results suggest that the technique should be equally applicable to Cu()ISE measurements at low copper() concentrations in natural waters. Studies of other ISEs have usually been J. Environ. Monit., 1999, 1, 483–487

483

View Online

limited to concentrations between 10−1 M and 10−5 M, the operating range over which many commercial ISEs give linear and relatively rapid response. For the technique to be applicable to marine and estuarine studies of copper speciation, the extrapolation technique must be robust in the nanomolar and sub-nanomolar range. A number of semi-theoretical relationships describing potential–time response transients of ISEs have been published elsewhere in the literature. In this paper, we investigate the extrapolation techniques described by Buffle and Parthasarathy13 and Mu¨ller,14 and develop an experimental protocol for the rapid measurement of free and/or total copper in seawater using a CuISE in a CFA system.

2 Experimental

Downloaded on 26 August 2011 Published on 01 January 1999 on http://pubs.rsc.org | doi:10.1039/A903819A

2.1 Chemicals and instrumentation Analytical grade reagents (Ajax, Sydney, Australia) were used to prepare all solutions, and high purity water was obtained from a Milli-Q water purification system (Millipore, Milford, MA, USA). The CFA system employed in this study is based on a portable FIA field analyser described previously,8 except that the flowcell, injection valve and peristaltic pump were mounted on a Perspex stand. Where possible, non-metal components were used in the construction of the system to reduce the potential for contamination. The majority of the CFA work was carried out using a Hamilton 2-way valve (Model HV 86728, Reno, NV, USA) to switch between sample and standard solutions (Fig. 1). Later, a Rheodyne 6-way Type 50 Teflon rotary valve (Cotati, CA, USA) was used to switch between copper buffers, samples and pre-conditioning solutions. All studies were conducted using an Orion CuISE (Model 94–29, Boston, MA, USA) which has a pressed disk of jalpaite, Ag Cu S, as the membrane. A Perspex FIP-3 wall-jet flow1.5 0.5 cell (Chemflow Devices, Melbourne, Australia) was used in this study. A Ag/AgCl electrode was located in a separate carrier stream of 0.1 M KCl. This flowcell design minimizes streaming potentials and liquid junction potentials.15 Flow rates for both streams were optimized, with KCl and the carrier/sample stream pumped at flowrates of 0.5 ml min−1, and 2 ml min−1, respectively, using an Alitea peristaltic pump (Model C-XV, Stockholm, Sweden). Data was recorded via a laptop PC using a Picolog datalogger (Pico Technology, Cambridge, UK ). For the portable system, potentials were recorded on a battery powered A/D system connected to a Macintosh laptop computer. For laboratory measurements, the analogue signal from the system was also connected to Radiometer PHM85 Precision pH meter (Copenhagen, Denmark) and the output measured on a YEW Type 3056 Pen Recorder (Tokyo, Japan). Processing of data for extrapol-

Fig. 1 Schematic diagram of the CFA/CuISE system. Switching valve positions were as follows: positions 1–4, calibrating buffers; position 5, filtered seawater or sample, and; position 6, sacrificial buffer (pCu 14–15). A portable field kit was assembled by employing a simple injection system, a battery operated A/D converter, a peristaltic pump, and a Macintosh computer.

484

J. Environ. Monit., 1999, 1, 483–487

ation studies was undertaken off-line using Microsoft Excel and software written in MS-Visual BASIC for Windows. Data were acquired at 100 ms intervals, although a data acquisition rate of 1 Hz is adequate for testing of the extrapolation procedure. Total copper was measured on archived samples that had been stored at pH 2 and previously analysed by conventional dithiocarbamate extraction and GFAAS.16 Before ISE analysis, the archived samples were UV-photooxidized to destroy organic matter that might interfere with the electrode response.4 Metal ion buffers containing ethylenediamine (15 mM ), CuSO (1 mM ) and NaCl (0.6 M ) were prepared 4 in accordance with the method described by De Marco.3 The pCu (pCu=−log[Cu2+]) of the buffer solution could be varied by simply adjusting the pH in order to change the competition between copper ions and hydrogen ions for ethylenediamine. The pH electrode was calibrated using NBS (National Bureau of Standards) buffers. Equilibrium speciation calculations of pCu were performed using MINTEQA2 17 and the stability constants reported by De Marco.3 The calculated pCu, as a function of pH, was then fitted to the following empirical equation: pCu=125.39–86.958×pH+22.788×pH2−2.7572 ×pH3+0.16079×pH4−3.6683×10−3×pH5 For example, changing the pH of the buffer from 6.00 to 8.00 changed the calculated pCu from 8.31 to 14.86. For measurements of total copper in acidified UV-photooxidized seawater, we need to correct the measured value of pCu for inorganic complexes since the ISE responds to the concentration of free Cu2+. The correction factor a ( log a=0.8) was calculated using MINTEQA2. For method development, we used seawater samples collected without clean protocols and stored in polyethylene containers that had been acid-washed. The pCu values for these samples were not indicative of the true values in the original water masses. Contamination of other samples was avoided by using standard ‘clean’ procedures for collecting, storing and handling samples.16 Where possible, components of the CFA system were made from Teflon or low density polyethylene, and all plasticware was soaked in 10% HCl and rinsed copiously with Milli-Q water before use. All laboratory analyses were carried out in a Class-100 clean room. 2.2 Analysis procedure The CuISE was polished before each series of experiments using alumina powder (0.3 and 0.05 mm) on separate moistened soft felt pads (Buehler, Lake Bluff, IL, USA) for each grade of alumina (BAS Inc., West Lafayette, IN, USA). The ISE was rinsed extensively with deionised water after each polishing step and blotted dry. The polished ISE was then conditioned in the CFA system with filtered open-ocean seawater until a steady state signal was observed, usually within 3 h. A stability criterion of 0.1 mV min−1 was used to define the steady state. The system was calibrated by switching between a series of pCu buffers (9