Comparison of filtration techniques for size ... - ACS Publications

Filtration has become a universally applied first step in the preparation of water samples for trace metal analysis. By tradition this filtration is u...
3 downloads 0 Views 2MB Size
1350

Anal. Chem. 1982, 54. 1350-1355

500-1000 ppm 02. A high range method utilizing 8000 ppm O2in the carrier gas could measure 0-40 mg/L T N with up to 500 ppm CTOD competition. A general method using air as the carrier gas is suggested if it is not desired to simultaneously measure TOD.

ACKNOWLEDGMENT Shah is recognized for his skill in performing many of the TOD and T N analyses.

LITERATURE CITED Clifford, D. A. Proceedings, 23rd Industrial Waste Conference; Purdue University, Lafayette, IN, 1988, 772-785. Butzelaar, P. F.; Hoogeveen, L. P. J. Wafer S e n . (London) 1975, 79 (948), 50. Voorn, B.; Mltchell, P. Prog. Water Techno/. 1978, 6 (2, 3), 423-433. Dixon, J. K.; Longfleld, J. E. I n “Catalysis”; Emmett, P. H., Ed.; Van Nostrand-Reinhold: New York, 1980; Vol. 8, Chapter 4, pp 281-302. Musselwhlte, C. C. ”Technical Memorandum TM 107”; Water Research Center: Stevenage, U.K., 1970. Hopwocd, J. A.; Musselwhlte, C. C. “Enquiry Report ER428”; Water Research Center: Stevenage, U.K., 1976.

(7) “Notes on Water Research”; Water Research Center: Stevenage, U.K., Feb 1978; No. 14. ( 8 ) Clough. P. N.; Thrush, 6 . A. Trans. Faraday Soc. 1987, 63, 915-925. (9) Clyne, M. A. A.; Thrush, B. A.; Wayne, R. P. Trans. Faraday SOC. Ig84, 60, 359-370. (10) Ruka, R. J.; Weisstart, J., U S . Patent No. 3 400 054 (Reissue No. RE 28,7923, Sept 3, 1968. (11) APHA, AWWA, WPCF ”Standard Methods for the Examination of Water and Wastewater”, 14th ed.; American Public Health Association: Washlngton, DC, 1976; pp 423-427. (12) APHA, AWWA, WPCF “Standard Methods for the Examination of Water and Wastewater”, 14th ed.; American Public Health Associatlon: Washinaton, DC. 1976: DD 434-436. (13) APHA, AWW>, WPCF ‘Standard Methods for the Examlnation of Water and Wastewater”, 14th ed.; American Public Health Assoclation: Washington, DC, 1976; pp 437-440. (14) VanHall, C. E.; Safranko, J.; Stenger, V. A. Anal. Chern. W83, 35, 315-319. (15) Drushel, H. V. Anal. Chem. 1977, 49, 932-939.

RECEIVED for review August 26, 1981. Accepted March 29, 1982. The authors thank Ionics, Inc., and the University of Houston Department of Civil Engineering for financial support.

Comparison of Filtration Techniques for Size Distribution in Freshwaters Duncan P. H. Laxen” and

I. Michael Chandler’

Grant Institute of Geology, Universi@of Edinburgh, West Mains Road, Edinburgh, EH9 3JW, Scotland

Flltratlon of freshwaters through screen fitters (Nuciepore) and depth filters (Sartorius), of 8-0.015 pm nominal pore slre, is critlcaiiy evaluated. Screen filters (Nuciepore) provide more accurate slre fractionation, as long as the fltter Is not allowed to clog. No beneflt derives from stirring the sample during flltratlon. Field filtration is however necessary for the most reliable size fractionation. Criteria used to evaluate filtration properties are Iron and manganese fittrate concentrations and scannlng electron microscope examination of the filtered particulates.

Filtration has become a universally applied first step in the preparation of water samples for trace metal analysis. By tradition this filtration is usually performed with 0.45-pm membrane filters, the resultant fractions usually being termed particulate and dissolved. With the growing awareness of the need to define the speciation of trace metals it has become apparent that this distinction between particulate and dissolved is arbitrary and of uncertain value. Colloidal material can pass through 0.45-pm filters and recent speciation studies have found substantial fractions of colloidally associated trace metals in many waters (1-3). Stumm and Bilinski first suggested that trace metal species could usefully be classified according to their size distribution ( 4 ) . Laxen and Harrison (5) have subsequently applied this approach to the speciation of trace metals in freshwaters, using filtration through different pore size Nuclepore filters to differentiate the size fractions. A number of other workers have used filtration to provide information about the size distribution of particulate and colloidal metal species (&IO), Present address: D e p a r t m e n t of Chemistry, U n i v e r s i t y of Edinburgh.

although they used Sartorious or Millipore filters, which are structurally very different to the Nuclepore filters used by Laxen and Harrison (5). Nuclepore filters are thin (5-10 pm) polycarbonate membranes with individual channels of uniform diameter. The pores vary from the rated size to -20%. Millipore filters (mixture of cellulose acetate and cellulose nitrate) and Sartorious filters (cellulose nitrate) are thicker (-100-150 ym), have a spongelike structure, and provide a tortuous pathway for the liquid. Particles may be trapped within the depth of the filter, as well as on the surface. Hence for the purpose of this paper Nucleopore filters will be called screen filters, while Sartorious filters (and similar types from other manufacturers) will be termed depth filters. It should be recognized, however, that these latter filters are not true “depth” filters, as all particles greater than the rated filter pore size are all trapped on the surface. Glass fiber filters are examples of true depth filters. Despite the widespread use of filtration, and the different filters and filtration techniques used, there have been few comparative studies of filtration methods for natural waters (11-13). The study reported here was therefore designed to provide a more systematic insight into the filtration of freshwaters, more specifically to examine the following points: (1) to compare depth filters (Sartorius) and screen filters (Nuclepore) of different nominal pore size with a view to determining their respective suitability for size distribution studies; (2) to compare filtration using stirred and unstirred samples, the former being designed to prevent or minimize accumulation of material on the filter; (3) to compare size distributions based on immediate filtration in the field with filtration performed several hours later in the laboratory; (4) to compare filtration through glass fiber filters with membrane filters. Glass fiber filters are necessary for the determination of particulate organic carbon, the values of which can be

0003-2700/82/0354-1350$01.25/00 1982 Amerlcan Chemical Society

ANALYTICAL CHEMISTRY, VOL. 54, NO. 8, JULY 1982

1351

Table I. Samples Used in This Study date

location

depth, m

pII

alkalinity, mequiv L"

ng mL-'

MnT n g mL-'

4/8/81 5/8/81 6/8/81 7/8/81 11/8/81 12/8/81 13/8/81 17/8/81 19/8/81

Esthwaite Water Esthwaite Water Black Beck Esthwaite Water Windermere Rostherne Mere Esthwaite Water Black Beck Esthwaite Water

8 4

6.7 7.3 6.8 6.7 7.1 7.4 6.7 7.0 6.6

0.44 0.47 0.91 0.57 0.24 1.92 0.58 0.97 0.52

96.9 123 547 840 15.0 71.0 947 415 292

247 77.1 125 1450 27.7 55.9 1550 67.4 1250

1

11.5 10 11.5 11 1 10

FeT

s4

9'

Total metal concentrations. compared with particulate constituent values determined with membrane filters (14). The principal analytical feature used to distinguish the filtration properties of the different filters and filtration techniques was the iron concentration of freshwater filtrates. Use has also been made of the filtrate manganese concentrations, as well as scanning electron microscope examination of the filters. Several different types of freshwater were used in these studies.

EXPERIMENTAL SECTION Samples. The samples used in this study were obtained from three freshwater lakes: Esthwaite Water and Windermere in Cumbria, and Rostherne Mere in Cheshire, as well as from a slow moving stream, Black Beck, feeding Esthwaite Water (Table I). Water was pumped from the required depth through polythene tubing by means of a persitaltic pump. In addition to filtration in the field a 5-L sample was collected in a well-rinsed acid-washed (15) polythene container for return to the laboratory. Filtration. The following filters were used: Nuclepore 47 mm diameter filters (Nuclepore Corp., Pleasanton, CA, supplied by Sterilin Ltd, Teddington, Middlesex) of nominal pore size 8, 3, 1, 0.6, 0.4, 0.2, 0.1, 0.05, 0.03, and 0.015 pm; Sartorius 47 mm diameter filters (Sartorius-Membranfilter GmbH, Gottigen, West Germany, Suppled by V. A. Howe, London) of nominal pore size 8, 3, 1.2, 0.45, 0.1, 0.05, and 0.01 pm. (Lot numbers of the filters used are available from the authors); glass fiber 25 mm diameter filters (Whatman Lab Sales Ltd, Maidstone, Kent) type GF/A, GF/C, and GF/F. The Nuclepore and Sartorius filters were prepared by soaking individual filters, housed in Millipore plastic petri slides, in 0.5% Aristar HNO, for a minimum of 24 h and then in Milli Q pure water for two 24-h periods. The filters were stored in pure water until required and then transferred to the filter holders using Millipore stainless steel tweezers. Filtration blanks were negligable,