Natural abundance-level measurement of the nitrogen isotopic composition of oceanic nitrate: an adaptation of the ammonia diffusion method

We have adapted the ''ammonia diffusion'' method of nitrate extraction for natural-abundance level nitrogen isotopic measurement of oceanic nitrate. The method involves: (1) sample concentration (by boiling or evaporation); (2) conversion of nitrate to ammonia using Devarda's alloy; and (3) the gas-phase diffusion of ammonia onto an acidified glass fiber disk which is sandwiched between two porous Teflon membranes. We have investigated the conditions necessary to effect complete ammonia recovery from natural seawater samples and the use of Devarda's alloy under these conditions. In addition, we have characterized the blanks in this method and designed a protocol to minimize them. Here, we report our protocol for nitrate extraction from seawater and provide an explanation of the protocol based on our method development work. To demonstrate the performance of the method, we present nitrate nitrogen isotopic data from nitrate standard additions to Sargasso Sea surface water and from several Southern Ocean depth profiles. The nitrate extraction method gives highly reproducible, complete recovery of nitrate and a standard deviation for isotopic analysis of < 0.2 parts per thousand down to 5 mu M nitrate (or lower). Replicate extractions of a nitrate standard added to Sargasso Sea surface water demonstrate agreement between the isotopic composition of the added and recovered N, with the extraction blank causing less than or equal to 0.3 parts per thousand discrepancy for 5 mu M nitrate. The blanks inherent in the extraction procedure are from Devarda's alloy and seawater dissolved organic nitrogen (''DON''). The N blank of the Devarda's alloy reagent depends on brand and lot number. The Devarda's alloy which we are currently using results in a blank of similar to 0.4 nmol N per 100 mi of seawater (effectively 0.4 mu M). An isotopic correction is made for this blank. For standard incubation conditions, stored Woods Hole seawater(with similar to 10 mu M DON) gives a similar to 0.6 mu M DON blank, while stored Sargasso Sea(with similar to 6 mu m DON) surface water gives a DON blank of 0.3-0.5 mu M The DON blank appears to cause the less than or equal to 0.3 parts per thousand difference between the measured and actual isotopic composition of nitrate added to Sargasso Sea surface water at the 5 mu M nitrate level. We discuss several ways to lower the DON blank for samples in which the DON concentration is high relative to the nitrate concentration. The nitrogen isotopic data from several Southern Ocean profiles, in conjunction with the other results presented in this paper, demonstrate the consistency of the data produced by the ammonia diffusion method. The ammonia diffusion-based protocol is more reliable and allows for better precision than the nitrate reduction/ammonia distillation method (Cline and Kaplan, 1975) in our hands. While the samples have an incubation time of 4 days or longer, we find that the diffusion method allows for higher throughput than the distillation method because samples can be run conveniently in large batches.