Comparative analysis of exchangeable aluminum in a tropical soil under long-term no-till cultivation

Aluminum toxicity is a key determinant of crop yield for acid soils. Techniques used to characterize a soil's exchangeable Al content often yield discordant results, thus impeding the implementation of corrective measures. We critically compared three such techniques - back-titration, ICP-AES, and Al-27 NMR - for a tropical Oxisol under long-term (8-years) no-till cultivation involving four different cropping-systems and three pH control regimens. ICP-AES was superior in terms of accuracy, level of quantification (LOQ), practicable pH range, speed, cost, and versatility. By comparison, back-titration consistently overestimated exchangeable Al, whereas Al-27-NMR, though accurate, had a higher LOQ and is inherently non-quantitative between pH 5.4 and 6.3 owing to extreme quadrupolar broadening of Al(OH)(n)(H2O)(6-n)((3-n)) + (n = 1-3) signals. The only Al species detected by NMR in any of the soil extracts was the toxic Al hexaaqua cation, Al(H2O)(6)(3+), despite an appreciable presence of dissolved organic matter. Application of either dolomitic lime or calcium-magnesium silicate (CMS) as a pH corrective significantly decreased exchangeable Al in the upper soil layers. None of the decrease, however, was attributable to insoluble aluminosilicates having formed as a consequence of CMS dissolution. Rotation crops of wheat (T. aestivum) and showy rattlebox (C. spectabilis) depleted exchangeable Al in the upper soil layers, whereas palisade grass (Brachiaria brizantha) had no discernable effect on Al availability. Plots which incorporated any one of the three rotation crops provided equally high yields of soybean in year-8 when pH (1 M KCl) exceeded 4.9 +/- 0.2 and Al saturation was less than 25 +/- 5% (Al-ICP = 22 +/- 21 mg kg(-1)), on average, in the 0-20 cm soil horizon.