ALUMINUM SPECIATION - VARIATIONS CAUSED BY THE CHOICE OF ANALYTICAL METHOD AND BY SAMPLE STORAGE

We have compared a new FIA method (M1) for aluminium speciation in nat ural waters and a manual one based on one of Driscoll's proposals (M2) . In synthetic solutions, aluminium's fluoro complexes were measured a s 'labile monomeric Al' (Al-i) by M2, but not as 'quickly reacting Al' (A(qr),) by M1. Aluminium's complexes with Nordic Reference Humic Aci d were measured neither as Al-i nor as A(qr), and the same result was obtained for Al's citrate complexes. After excluding aluminium's fluor o complexes from Ali, the results of the two methods agreed well for s oil leaching samples and fairly well for natural water samples. Detect ion limits: 1O mu gL(-1) L(-1) (ca. 0.4 mu M) for both methods. Sample throughputs: M1: 66 injections h(-1); M2: 5 samples h(-1). Repeatabil ities (RSD) on natural water samples: M1: 0.6-5.8% and M2 0.7-4.6%. Sa mple storage studies on soil solutions (FIA method) indicated that sto rage effects were 'sample type specific'. A sample with a low level of Al-qr, and a high DOC lever (P2A) was more sensitive to storage than one with a high [Al](qr)], and a low [DOC] (P2B()). The decrease in [ Al](qr) was statistically significant after 10 h (P2A) and 24 h (P2B( )), respectively. After 3 days' storage, [Al],, had decreased by some 40% in P2A, while the decrease in P2B() was less than 1O% after 8 day s. The results of this part of the study also emphasize the importance of careful method standardization in sample storage studies.