A direct instrument comparison for measurement of total dissolved nitrogen in seawater

Abstract

A direct comparison has been made for the measurement of total dissolved nitrogen (TDN) in seawater using high temperature combustion (HTC) instruments. A side-by-side comparison in one laboratory (Lewes exercise) was made with two hybrid total organic carbon (TOC) instruments linked to chemiluminescence detectors (CLD), a combined TOC total nitrogen (TN) instrument, and two stand-alone TN instruments. Later, a series of field samples was run in home laboratories of the participants (Homelab exercise) using four hybrid TOC–CLD systems, the combined TOC/TN instrument, and three stand-alone TN instruments. In the Lewes and Homelab exercises, samples were also run by the more established persulfate oxidation method. There was fairly good agreement (group precision) among all the HTC methods; average %CV of ±12.0% for five nitrogen compounds and average %CV of ±7.7% for five natural samples in the Lewes exercise and %CV of ±10.7% for 14 field samples in the Homelab exercise. In both cases, the persulfate method values were similar to those of the HTC instruments. On a finer scale of evaluation, the stand-alone TN instruments, operating at higher combustion temperatures, gave considerably more scatter and more divergent values than the set of HTC instruments that used a singular combustion environment with a lower temperature. The five hybrid systems, with Shimadzu TOC instruments combined with five different CLD systems and employing individual integration programs, gave a group precision of ±5.8% for the field samples. These systems also provided simultaneous dissolved organic carbon (DOC) concentrations, and the DOC group precision for the 14 samples was ±5.9%. Our conclusion is that the five Shimadzu instrument systems provide more accurate TDN measurements than the stand-alone instruments. However, when subtracting inorganic nitrogen (DIN) to calculate dissolved organic nitrogen (DON), there is not good agreement, especially for deep ocean samples. The persulfate method gave slightly higher TDN concentrations for the field samples than the Shimadzu HTC averages, indicating a possibility of incomplete conversion of some nitrogen compounds by the HTC instruments and/or peak integration problems. The inequity with persulfate oxidation and the scatter with the HTC methods require further investigation. A new broad community DON comparison is currently underway which includes TDN and DIN analyses. In addition, several of us are informally collaborating to improve and evaluate routine analysis with HTC instrumentation.