Dataset: Fe reduction isotope effects

2.2 Stability of the Fe-ligand complexes
In order to ensure that our experiments are only capturing the isotope effect of Fe(III) reduction to Fe(II) due to our added reductants, we first assessed whether there was any spontaneous reduction of Fe(III)EDTA, spontaneous oxidation of Fe(II)FZ3, or isotope exchange between these complexes. This was tested by separately preparing concentrated solutions of Fe(II)-FZ3 from natural iron, and Fe(III)double-spike-EDTA from an Fe spike consisting of roughly equal portions 57Fe and 58Fe. The stocks were then diluted into surface seawater collected near Bermuda to final concentrations of 2·10-4 M Fe(III) double-spike, 2·10-4 M Fe(II), 10-3 M FZ, and 10-3 M EDTA. Eight such solutions were prepared, to be subsampled over the next 27 h.
At each timepoint, the Fe(II) was first separated from the experimental solution onto the C18 column and eluted with methanol. Then, Fe(III)-EDTA which remained in the experimental solution was reduced to Fe(II)-FZ3 by the addition of 1.5 mL of 1.4 M hydroxylamine hydrochloride, and separated by adsorption as Fe(II)-FZ3 onto C18 resin. Both the Fe(II) and Fe(III) fractions were then purified for isotope analysis as described below.

2.3 Chemical reduction experiments
10 mL experimental solutions were prepared by adding reagents to a 15 mL polypropylene centrifuge tube in the following order: 1) 100 μL Fe(III), 2) 100 μL EDTA, mixed and allowed to equilibrate for > 2 min to allow EDTA to bind the Fe(III), 3) 1 mL buffer, 4) 100 uL ferrozine, 5) 8.6 mL filtered trace-metal clean seawater collected from the surface ocean near Bermuda, 6) 100 μL reductant added in order to initiate the experiment. Experiments at 60°C and 100°C were performed by immersing samples in a heated water bath before adding reductant, while samples performed at 2°C were immersed in an ice bath.
The progress of reaction was monitored by visually examining tubes for the appearance of purple-colored Fe(II)FZ3. For comparison, a standard was made containing 10 μM Fe(II)-FZ3, equivalent to reducing 5% of the initial Fe(III)-EDTA, so that when the color in the experimental tube matched that of the standard, experiments could be terminated at a reaction extent of about 3-7% completion. The exact extent of reaction was then determined by comparing sample adsorption at 562 nm to iron standards prepared using the ferrozine method (Viollier et al., 2000).

2.4 Photochemical reduction experiments
Photochemical reduction experiments were performed as above, except that no chemical reductant was added. Instead samples were either amended either with glucaric acid or not, and the experiment was initiated by bringing samples outside and exposing them to natural sunlight on a nearly-cloudless summer mid-day in South Carolina.

2.5 Isotope analysis
Methanol samples containing Fe(II)FZ3 were evaporated, then reacted overnight on a hot-plate with 500 μL 0.1N HNO3 and 500 μL concentrated H2O2 to destroy the FZ. Samples were then evaporated again and observed to make sure that no purple color (Fe(II)FZ3) remained. If necessary, samples were reacted again with HNO3 and H2O2 until no purple color remained. Samples were analyzed for Fe stable isotope ratios (δ56Fe) at the University of South Carolina according to published methods (Conway et al., 2013) as modified for analysis of marine particles (Revels et al., 2015). Briefly, samples were amended with a ~1:1 57Fe:58Fe double spike, purified by anion exchange chromatography, and analyzed on a Neptune multi-collector ICP-MS. Sample δ56Fe was analyzed compared to IRMM-014 and the δ56Fe of the FeCl3 used to prepare stock solutions was also analyzed, so that all data are reported here as Δ56Fe, calculated as δ56FeFe(II)-δ56FeFe(III).