Direct determination of metal concentrations in freshwater algae by total reflection X-ray fluorescence spectrometry† Katalin Barka�cs,a Anita Varga,b Kamilla Ga�l-Solymosc and Gyula Za�ray*b aDepartment of Chemical Technology and Environmental Chemistry, L . Eo�tvo�s University, P. O. Box 32, H-1518 Budapest 112, Hungary bDepartment of Petrology and Geochemistry, L . Eo�tvo�s University, P.O. Box 330, H-1445 Budapest, Hungary cResearch Group of Environmental and Macromolecular Chemistry, Hungarian Academy of Sciences, L.Eo�tvo�s University, P.O. Box 32, H-1518 Budapest 112, Hungary Received 11th October 1998, Accepted 15th February 1999 A direct TXRF method for the determination of metal impurities by applying three diVerent freshwater algae, viz., Chlorella keslerii, Synehococcus sp. and Cylindrospermopsis raciborskii, was developed. The applicability of the slurry sampling technique was confirmed by comparing the analytical results obtained with those obtained by total digestion of the algae.The concentration data showed an acceptable agreement for Cu, Fe, Mn and Zn. The calculated accumulation factors for the metals detected were highest for Chlorella keslerii. The problems caused by the increasing number of pollutants cants in biomass needs multi-elemental analytical techniques having also low detection limits. In practice, inductively from diVerent sources, and their eVects on the quality of the environment, also on aquatic life—from cellular to ecosystem coupled plasma atomic emission-,17 and mass spectrometry10 and also total reflection X-ray fluorescence spectrometry levels—are well known and have been widely investigated.1–4 Research published on the eVects of pollutants on water (TXRF)22–24 have been successfully used.These methods have generally been applied to the analysis of periphyton samples organisms has covered the testing of many diVerent species, as applicable organisms for alternative methods in aquatic previously digested with nitric acid,3,15–18 nitric acid–sulfuric or perchloric acid14 and also nitric acid–hydrogen peroxide biomonitoring and toxicology.4–9 These test organisms vary from aquatic animals: bacteria (bacterial bioluminescence mixtures.9,24,25 Pettersson and co-workers25–27 have developed a direct assays), Daphnia species (Daphnia magna), mussels, fish cells and fishes to plants; and among the most recent from hyacinths TXRF method for the investigation of periphyton communities.During their experiments, marine periphytons were to algae species.4,10–14 Much of the research work on toxic eVects has dealt with phototrophic algae and cyanobacteria, colonized on soda glass discs used as both benthic sample holders and TXRF carrier plates. They established that the as the primary producers in the hydrosphere.3,5,14 For monitoring pollutant inputs into the ecosystem and for evaluating analytical data of the wet digestion and direct methods for K, Ca, Mn, Fe, Ni, Cu, Zn, As, Rb and Sr were in good toxic eVects, the use of algae and algal communities, has some particular advantages.They easily meet the requirements of agreement. With all of the methods applied, the heterogeneity of the intensive toxicity screening by ensuring a considerable number of test animals. On the basis of high pollutant accumulation samples, and also the sample preparation procedure, including the eVective removal of the water medium, were found to be capacities compared with other species related to specific morphological and physiological properties, algae have been important.14,16,20 In this work, on the basis of the data obtained for sea-water found to be highly sensitive detectors of micropollutants, making them suitable biomonitors to quantify environmental periphytons, a direct TXRF method was developed for the analysis of freshwater algae. Three diVerent algae species were quality.15–17 The calculated accumulation factors (AF), i.e., the concentration ratio expressed as the concentration of a investigated, by applying slurry sampling or microwaveassisted vapour-phase acid digestion procedures.pollutant in dried algae relative to its ambient concentration in water, are on average within the interval 102–105.3,10,14–20 These high accumulation factors and also high tolerance to pollutants enable the algae to create a time integrated accumu- Experimental lation (the LD50 values, characterizing their tolerance, for Algae samples most of the metals are generally around several thousand mg g-1).3,15,21 Algal communities might provide Three diVerent freshwater algae species were investigated: more complex information on water quality than the direct Cylindrospermopsis raciborskii (filamentous green algae), water qualification carried out by chemical tests at defined Synehococcus sp.(unicellular, ellipsoidal, blue–green algae of times and locations. 5 mm length), and Chlorella keslerii (unicellular green algae, The various environmental stresses result in changes in the round or ellipsoidal of 2–10 mm length). These species were metabolism of the algal communities, which are reflected in chosen as test materials because of their diVerent morphologithe content of trace elements and also nutrients in algae. The cal and physiological characteristics and also their widespread determination of concentrations of micro-nutrients and toxi- occurrence in surface waters, such as the River Danube and Lake Balaton.The selection of a suitable sample preparation procedure, †Presented at the 8th Solid Sampling Spectrometry Colloquium, Budapest, Hungary, September 1–4, 1998. combined with metal content determinations carried out by J. Anal. At. Spectrom., 1999, 14, 577–581 577TXRF, required a large number of well defined, homogeneous samples. For this purpose, the algae species were grown in ‘Allen-nutrient solution’ under controlled conditions (sterility during aeration, artificial UV lighting, temperature).The concentrations of chemical compounds in this culture medium are shown in Table 1. In order to determine the metal impurities in these biological materials, the sampling of the tested algae species was carried out in the logarithmic-phase of algal growth. Homogeneous suspensions of algae samples were handled during a multi-step sample preparation procedure preceding their metal content determination by TXRF.Sample preparation The multi-step algae sample preparation method is demonstrated in Fig. 1. The first step was the separation of the suspended algae cells from the nutrient solution. This separation was carried out by using a laboratory centrifuge (2000 rpm, 5 min) or by filtration. Sartorius membrane filters of 0.2 mm pore size were applied and the filtration speed was enhanced by means of a vacuum pump.During this mechanical separation process, deionized water was applied as a washing medium according to the suggestions of other workers. 14,16,19,20,23 The eYciency of media removal was controlled by measuring the specific conductivity, and metal and nitrate ion concentrations of the filtrates. The calculation of metal concentrations in algae samples requires the exact determination of their dry mass. Therefore, the washed algae samples were freezedried,5 and the powder samples were divided into two portions. One portion of the samples was used for direct measurements by preparing slurries from known weights (2–10 mg) of dry algae samples suspended in 1 ml of deionized water by ultrasonic treatment (Branson – 3200 E1 ultrasonic cleaner working at 47 kHz for 10 min).The other portion of the samples was digested by applying a microwave-assisted vapour-phase acid digestion procedure (2–10 mg of dry algae were decomposed in the vapour of 9 ml of concentrated nitric acid and 1 ml of hydrogen peroxide) based on the CEM-2100 (Matthews, NC, USA) pressure controlled microwave digestion system.28 Fig. 1 Sample preparation procedure. The digestion program used for decomposing the algae samples consisted of the following steps: (1) 1 min at 25% thoroughly cleaned and siliconized (silicone swer (pressure, 138 kPa); (2) 4 min at 66% power (pressure, Feinbiochemica, Heidelberg, Germany), and dried in a clean- 689 kPa); and (3) 30 min at 33% power (pressure, 689 kPa) box at 80 °C for 30 min on a ceramic-coated hot-plate.The (100% power=950±50 W). amount of the Ga was 100 ng in the 25 ml droplets. The surface After digestion, the liquid samples were made up to 1 ml densities of the algae samples on the carrier plates were with deionized water. The pre-treated samples (slurries, calculated, assuming that the droplets had a homogeneous digested sample solutions) were mixed with a Ga standard distribution regarding their dry residues and covered a surface solution (as internal standard).After homogenization, a 25 ml of 20 mm2. The data varied between 2 and 10 mg mm-2 dry aliquot was dropped onto a quartz-glass carrier, previously algae. The surface densities of the digested samples shown in Fig. 2 are not the true surface densities; they were calculated Table 1 Concentrations of chemical compounds in the ‘Allen’ for comparative evaluations, assuming that the original, nonnutrient solution digested dry algae content of the samples was still present.Component c/mg l-1 Instrumentation NaNO3 1500 The analysis of the algae samples, nutrient solutions, filtrates, K2HPO4 39 deionized water, reagents (nitric acid and hydrogen peroxide) MgSO4·7H2O 75 Na2CO3 20 and digested blank solutions was carried out with an EXTRA CaCl2 27 IIA total reflection X-ray fluorescence spectrometer (produced Na2SiO3·9H2O 58by ATOMIKA Instruments, Oberschleissheim, Germany).EDTA 1.0 The specific analytical parameters were as follows: Mo Citric acid 6.0 micro-focus X-ray tube (50 kV, 38 mA); high energy cut-oV Iron(III) citrate 6.0 filter (quartz glass mirror); attenuation filter (200 mm Mo, Boric acid 2.9 MnCl2 1.8 240 mm Al ); Si(Li) detector (80 mm2 area); integration time ZnSO4·7H2O 0.2 500 s; energy of applied analytical lines: K (Ka) 3.312 keV; Na2MoO4 0.4 Ca (Ka) 3.690 keV; Mn (Ka) 5.894 keV; Fe (Ka) 6.398 keV; Co(NO3)2·6H2O 0.08 Cu (Ka) 8.040 keV; Zn (Ka) 8.630 keV.CuSO4·5H2O 0.05 In order to check the influence of the inhomogeneous 578 J. Anal. At. Spectrom., 1999, 14, 577–581be a more beneficial tool in handling very small sample amounts. Filtration was carried out in several (4–5) steps, so that the residue of the nutrient solution was first removed, then the algae were resuspended in deionized water on the surface of the filter and filtration was repeated. The metal concentrations of the subsequent filtrates were measured by TXRF.As Fig. 3 demonstrates, for Synehococcus sp., the element concentrations of the filtrates decreased stepwise during the washing procedure. The fourth filtrate was virtually a metal-free solution. However, the specific conductivity and metal concentrations (mainly Ca) increased in the next washing step, possibly because the use of deionized water caused the algae cells to rupture on osmotic shock. There is a risk, therefore, of metal content loss with increasing time or water volumes applied during separation. Similar results were Fig. 2 EVect of surface densities of the digested and suspended observed for each algae sample; no variations were experienced Synehococcus sp. samples on the Fe concentrations measured by in the eYciency of the washing procedure owing to the TXRF. Surface density data are expressed in mg dry algae per mm2 diVerences in algae species. In summary, the application of and calculated also for the digested samples, assuming that the original four washing steps, and a total volume of deionized water no dry algae mass was present. greater than twice the original volume of the algae medium, resulted in the eVective separation of interstitial water and hence the eVective removal of the disturbing metals originating from the sample medium.In order to check the eYciency of the digestion procedure, chemical oxygen demand (COD) tests were applied.29 (COD of the organic materials expressed in milligrams of O2, determined according to international standards by a rapid, microscale Merck-test. The method is analogous to ISO 6060.) The algae samples were tested both in slurries (without digestion), and also after the digestion procedure.The data measured by the COD method showed that, after digestion, the samples generally contained less than 5% of the organic materials originally present. In order to select the appropriate concentration of algae slurries, the influence of the total algae mass on the carrier plate on the determined element concentrations was measured.It should be emphasized that the amount of Ga added in the Fig. 3 Element concentrations of the filtrates measured by TXRF as droplets was in all cases 100 ng. Typical data of these experi- a function of the washing steps; amount of Synehococcus sp. algae on ments are demonstrated in Fig. 2, which summarizes the iron the filter: 13 mg (dry mass); washing medium: deionized water, 4 ml per washing step.concentrations of the digested samples and of the directly measured Synehococcus sp. slurries as a function of the algae surface densities. It can be seen that the measured concen- distribution of the dried slurry droplets, the TXRF measuretration values show the lowest deviation at a surface density ments were carried out at three diVerent orientations (0°, 120°, of 5 mg mm-2. The RSD values of these measurements also 240°) of the carrier plates.For the investigation of dried slurry demonstrate that, at low surface densities, an increased uncer- droplets, a scanning electron microscope (Amray 1830 I/T6) tainty of the measurements was observed. However, a surface equipped with an EDAX PV9800 ED spectrometer (20 kV density higher than 5 mg mm-2 created less homogeneous acceleration voltage, 2 nA beam current) was used. samples having a relatively high thickness. Hence, a surface density of 5 mg mm-2 is appropriate for both sample types.Results and discussion Since during drying of the slurry droplets thin layers having a relatively rough surface are formed, the analytical data are During the algae sample preparation process—as an important expected to be influenced by the surface properties. In order and determining step in the correct measurement of the metal to study this phenomenon, digested and non-digested Chlorella content of the cells—the removal of interstitial water was keslerii samples (surface density 5 mg mm-2) were investigated tested first.For a proper separation of algae cells and their in three diVerent positions of the carrier plates. The data in medium, the filtration procedure proved to be more eVective Table 2 suggest the existence of slight diVerences measured as and less time consuming than the use of a centrifuge. Filtration required smaller volumes of washing water and was found to a function of the carrier position and hence a relatively uniform Table 2 Mean metal concentration values in Chlorella keslerii determined by TXRF at three diVerent orientations of the carrier plates (0°, 120°, 240°) containing the same sample and relative standard deviations calculated from the data obtained at diVerent positions Sample type Mn Fe Cu Zn K Ca Slurry— Mean value/mg g-1 4510 15 400 118 646 4170 35 600 RSD (%) 7.1 6.6 5.7 7.3 4.7 6.2 Digested— Mean value/mg g-1 4530 13 600 125 728 3270 29 400 RSD (%) 2.7 0.3 3.6 1.5 4.4 4.4 J. Anal.At. Spectrom., 1999, 14, 577–581 579Fig. 5 Calculated accumulation factors (AF) of metals for the three diVerent algae species. concentrations were measured in the slurries of Chlorella keslerii and Cylindrospermopsis raciborskii. Similar Cu concentrations were determined in both sample types for all the algae species, whereas lower Zn concentrations were found in each Fig. 4 Back-scattered electron image of the slurry sample of case when slurries of algae were applied.The RSD values of Synehococcus sp. (a) and X-ray image of Ga Ka (b) showing the the determined metal concentrations were higher for direct distribution of the internal standard element at the border of the measurements than those for the digested samples for almost sample droplet; total amount of Ga added: 1000 ng; average surface all the metals. The lowest deviations in the mean concen- density of the algae sample in the dried droplet: 5 mg mm-2; image magnification: ×15 (a), ×113 (b).trations determined by the slurry and digestion techniques were found for Synehococcus sp. The algae species grown under the same conditions can be thin layer structure. Comparing the data obtained with and compared regarding their metal accumulation capacities without digestion, slightly lower RSD values, as a function of (accumulation factors). The calculated accumulation factors the carrier position, for almost all the metal concentrations of for the diVerent elements (see Fig. 5) emphasize the diVerences the digested samples were found. However, all the RSD values depending on the metal qualities regarding a particular species, for the non-digested algae slurries were also less than 10%. and also the variations in a given element accumulation ratio The investigation of the structure of droplets of algae slurries for the diVerent species. Among the algae species tested—on having diVerent concentrations and the distribution of the Ga the basis of the highest accumulation factors determined— internal standard in the dry thin layer was carried out with a Chlorella keslerii appears to be the best species for scanning electron microscope.In this case the amount of Ga biomonitoring. added was higher (1000 ng) than during TXRF measurements, to ensure the detection of Ga by the EDAX system. Independently of the slurry concentration, Ga shows a similar Conclusion distribution profile in the droplets of all the samples. A homogeneous distribution of the Ga internal standard did not A solid sampling algae preparation technique has been developed and modified as a multi-step sample preparation pro- occur on the total surface of the droplets; instead, the Ga was concentrated in the outer parts of the droplets, as the X-ray cedure. The main steps included the eVective removal of the nutrient solution, and hence the removal of the disturbing image of Ga Ka shows in Fig. 4. The photograph, of a dried Synehococcus sp. sample slurry having a 5 mg mm-2 surface metal content, by filtration and washing with deionized water. Freezedrying of the algae after washing ensured the homogen- density, demonstrates the characteristic Ga enrichment ‘ring’ found for the algae slurry droplets. eity of the samples. Slurries were prepared from the dried algae and deionized water by using ultrasonic treatment. In The element concentrations of the dried algae formed from freshly cultivated algae communities were measured by TXRF order to compare the analytical results, acid-digested algae samples were also analyzed.and are summarized in Table 3. The average concentrations of the elements (mean values calculated from three repeated The deviations of the metal concentration values determined by TXRF investigation of slurry and digested algae samples measurements) determined using both slurries and digested solutions of the same algae samples and the RSD values are lowest for Synehococcus sp.Since Chlorella keslerii shows only slightly higher deviations and its accumulation factors calculated from these measurements are also compared in Table 3. The data generally show higher mean values for Mn are considerably higher than those of the other algae, this species seems to be the most promising for biomonitoring. measured in slurries than in digested samples. Also, higher Fe Table 3 Element concentrations of dry algae mass measured for the diVerent species and sample types by TXRF (surface density: 5 mg mm-2) Sample Mn/ RSD Fe/ RSD Cu/ RSD Zn/ RSD K/ RSD Ca/ RSD Algae type mg g-1 (%) mg g-1 (%) mg g-1 (%) mg g-1 (%) mg g-1 (%) mg g-1 (%) Cylindrospermopsis Slurry 243 2.9 1155 7.4 72.8 1.7 36.3 8.0 6240 6.6 2175 4.5 raciborskii Digested 208 2.4 1090 7.4 72.8 0.6 41.4 1.3 6100 5.6 1440 3.9 Chlorella Slurry 4300 9.9 14 940 12.4 113 9.1 633 10.4 118 22.3 1085 10.3 keslerii Digested 3865 9.4 12 490 3.1 114 2.9 680 1.9 116 46.6 642 70.3 Synehococcus Slurry 77.8 5.8 1210 10.4 15.4 17.3 54.7 12.6 1470 2.9 614 7.1 sp.Digested 75.1 6.2 1285 6.1 15.0 12.3 60.5 7.5 1385 9.7 479 6.5 580 J. Anal. At. Spectrom., 1999, 14, 577–58113 J. M. C. Geuns, A. J. F. Cuypers, T. Michiels, J. V. Colpaert, A. Acknowledgements Van Larere, K. A. O. Van der Broeck and C. H. A. Vandecasteele, Sci. Total Environ., 1997, 203, 183. The authors are grateful to Dr. T. K. Kiss, Dr.A. Kozma, 14 M. D. K. Abo-Radey, Arch. Hydrobiol., 1980, 89, 387. Dr. E� . A� cs and Mr. I. Grigorszki for their valuable instructions 15 M. G. Kelly and B. A. Whitton, Interspecific diVerences in Zn, Cd and for the algal cultures. This research program was sup- and Pb accumulation by freshwater algae and biophytes, Durham, ported by grants T 014861 and T 023726 from the National UK, 1987. Scientific Research Foundation (Hungary). 16 B. A. Whitton, I. G. Burrows and M. G. Kelly, J.Appl. Phycol., 1989, 1, 293. 17 C. A. Mahan, V. Majidi and J. A. Holcombe, Anal. Chem., 1994, 66, 624. References 18 B. M. McHardy and J. J. George, presented at the Series of 1 S. Hall, J. Chamberlain and E. Godwin-Saad, Water Environ. Symposia Biologica Hungarica, Hungarian Academy of Sciences, Res., 1995, 67, 713. Budapest, Hungary, 1985, vol. 29, p. 37. 2 Gy. Lakatos, E� . A� cs, K. Buczko� and Cs. Cserha�ti, presented at 19 P. T. S. Wong, Y. K. Chau and J. L. Yaromich, Can.J. Fish. the 10th International Svedala Symposium on Ecological Design, Aquat. Sci., 1987, 44, 1257. Budapest, Hungary, 1992. 20 P. L. Foster, Nature (London), 1977, 269, 322. 21 J. W. Rijstenbil, A. Sandee, J. Van Drie and J. A. Wijnholds, 3 P. L. Foster, Freshwater Biol., 1982, 12, 17. FEMSMicrobiol. Rev., 1994, 14, 387. 4 E. Nusch, Z. Umveltschem. O� kotox., 1993, 5, 155. 22 K. Gu� nther, A. von Bohlen, G. Paprott and R. Klockenka�mper, 5 P. J. Say and B. A. Whitton, Hydrobiologia, 1980, 76, 255.Fresenius’ J. Anal. Chem., 1992, 342, 444. 6 L. S. Clesceri, A. E. Greenberg and R. R. Trussell, Standard 23 T. Bistriczki and M. Munwar, Can. J. Fish. Aquat. Sci., 1983, Methods for Examining Water and Wastewater, American Public 39, 506. Health Association, Port City Press, Baltimore, Maryland, USA, 24 J. Bohman, H. Blanck, P. Standzenieks, R. P. Pettersson and N. T. 17th Edition, 1989. Hong, X-ray Spectrom., 1993, 22, 260. 7 S. A. Badr and H. F. Abou-Waly, Bull. Environ. Contam. Toxicol., 25 R. P. Pettersson, Doctoral Thesis, Chalmers University of 1997, 59, 298. Technology, Go�teborg, Sweden, 1997. 8 M. A. El-Dib, H. F. Abou-Waly and A. M. H. El-Naby, Bull. 26 R. P. Pettersson and M. Olson, Proceedings of the EDXRS-96 Environ. Contam. Toxicol., 1997, 59, 438. Conference, Lisbon, Portugal, 1996, p. 32. 9 S. C. Gardner, C. E. Grue, J. M. Grassley, L. A. Lenz, J. M. 27 R. P. Pettersson, Proceedings of the 6th Conference on TXRF and Lindenauer and M. E. Seeley, Bull. Environ. Contam. Toxicol., Related Methods, Dortmund, Germany, 1996, p. 69. 1997, 59, 492. 28 Zs. Cze�ge�ny, B. Berente, M. O� va�ri, M. Garcý�a Tapia and Gy. 10 I. Bruns, K. Friese, B. Markert and G. J. Krauss, Sci. Total Za�ray, Microchem. J., 1998, 59, 100. Environ., 1997, 204, 161. 29 Validation of Merck Spectroquant COD Cell Test, Quality 11 Th. Braunbeck, Eco-Informa ’97, 1997, 12, 446. Assurance Group,Water Research Institute, Budapest, 1998. 12 R. Altenburger, U. Ensenbach, K. Jung, M. Knops, P. Popp, H. Segner, H. Weiss and G. Schu�u� rmann, Eco-Informa ’97, 1997, 12, 452. Paper 8/08806C J. Anal. At. Sp