Experimental determination of magnesium isotope fractionation during higher plant growth

Two higher plant species (rye grass and clover) were cultivated under laboratory conditions on two substrates (solution, phlogopite) in order to constrain the corresponding Mg isotope fractionations during plant growth and Mg uptake. We show that bulk plants are systematically enriched in heavy isotopes relative to their nutrient source. The Δ26Mgplant-source range from 0.72‰ to 0.26‰ for rye grass and from 1.05‰ to 0.41‰ for clover. Plants grown on phlogopite display Mg isotope signatures (relative to the Mg source) ∼0.3‰ lower than hydroponic plants. For a given substrate, rye grass display lower δ26Mg (by ∼0.3‰) relative to clover. Magnesium desorbed from rye grass roots display a δ26Mg greater than the nutrient solution. Adsorption experiments on dead and living rye grass roots also indicate a significant enrichment in heavy isotopes of the Mg adsorbed on the root surface. Our results indicate that the key processes responsible for heavy isotope enrichment in plants are located at the root level. Both species also exhibit an enrichment in light isotopes from roots to shoots (Δ26Mgleaf-root = −0.65‰ and −0.34‰ for rye grass and clover grown on phlogopite respectively, and Δ26Mgleaf-root of −0.06‰ and −0.22‰ for the same species grown hydroponically). This heavy isotope depletion in leaves can be explained by biological processes that affect leaves and roots differently: (1) organo–Mg complex (including chlorophyll) formation, and (2) Mg transport within plant. For both species, a positive correlation between δ26Mg and K/Mg was observed among the various organs. This correlation is consistent with the link between K and Mg internal cycles, as well as with formation of organo–magnesium compounds associated with enrichment in heavy isotopes. Considering our results together with the published range for δ26Mg of natural plants and rivers, we estimate that a significant change in continental vegetation would induce a change of the mean river δ26Mg that is comparable to analytical uncertainties.