METHODS

Ammonium concentrations and ¹⁵N, as well as the ¹⁵N of solid phases were determined through the sediment columns recovered from Sites 1227 and 1230. Bulk sediment samples used for isotopic analyses of solids were "squeeze cakes," kept frozen between the time of collection and analyses. Samples were oven-dried at 60°C for 24 hr and then finely ground. Total nitrogen (TN) and total carbon (TC) were measured on a Carlo Erba CHN-2500 elemental analyzer. Total organic carbon (TOC) was estimated from ¹³C of TC by assuming the ¹³C of carbonate is 0 and the ¹³C of organic matter is –21. TOC content obtained using this approach is in relatively good agreement with the values measured during Leg 112 (Meister et al., this volume).

Pore water was extracted by squeezing (D'Hondt, Jørgensen, Miller, et al., 2003). Pore water samples collected for isotopic analyses were acidified (to pH ~ 4 for Site 1227 samples and pH ~ 1 for Site 1230 samples) and immediately frozen on board the JOIDES Resolution. While acidifying vials for samples from Site 1227, the effect of alkalinity was not taken into account, which resulted in higher final pH values. Prior to isotopic analysis, samples were thawed and mixed well. Ammonium concentrations were determined at the University of Southern California (USC, USA), and the results were compared to the shipboard measurements in order to assure the adequate preservation of ammonium in the samples. The USC analyses and the shipboard analyses for Site 1227 agree within 2%. However, USC analyses of samples from Site 1230 are consistently lower than shipboard measurements by 10%. The amount of N collected on acidified filters during our ammonium extraction procedure (see below) and measured as a peak area on the mass spectrometer agrees within 2% with expected amount of N based on concentrations obtained at USC. Therefore, we conclude that the 10% difference between USC and shipboard analyses stems from discrepancies in the calibrations. The loss of ammonium during storage is unlikely because such loss would also have caused considerable scatter in ¹⁵N, which was not observed.

Ammonium was extracted on acidified glass fiber filters using a method of Sorensen and Jensen (1991), Sigman et al. (1997), and Holmes et al. (1998). In brief, diluted pore water was placed in polypropylene centrifuge tubes, and pH was adjusted to >10 by addition of finely ground MgO, precombusted at 450°C. This converted NH₄⁺ to uncharged NH₃. A "trap" made of a Teflon tape envelope containing a glass fiber filter grade C or glass fiber filter grade F wetted with 20 µL of 2-M H₂SO₄ was then added to each tube. The uncharged NH₃ diffused through the Teflon membrane was protonated inside the "trap" and collected on the filter.

Nitrogen isotopic ratios of sediments and of ammonium trapped on the filters were measured at USC on an Isoprime Micromass mass spectrometer, interfaced with a Carlo Erba CHN-2500 elemental analyzer in the continuous flow mode. Isotopic ratios were determined against a reference N₂ (ultra high purity grade, Gilmore) that has been calibrated against a set of National Institute of Standards and Technology (NIST) standards, IAEA N1 and IAEA N2, routinely run in the USC stable isotope laboratory. NIST standards were run daily as internal standards. The daily precision, based on the internal standards, was 0.2 or better. Procedural standards of known isotopic composition, similar in concentration and volume to samples, were prepared and carried through the whole extraction procedure with each set of samples. The sample standard deviation based on replicate analyses of both samples and standards was typically 0.50. The ¹⁵N of procedural standards of ammonium was usually ~0.5 lighter than expected, likely due to either a procedural artifact or perhaps a small reagent blank. The amount of nitrogen in standards was always close to that in samples, so ¹⁵N values of samples were corrected by adding 0.5 to the measured values. Typical size of the nitrogen blank was 0.2 to 0.3 µmol, while samples contained 6–7 µmol of N. Extraction efficiencies (yields) varied ranged between 97% and 103% for both sets of samples.