Iron interference in the quantification of nitrate in soil extracts and its effect on hypothesized abiotic immobilization of nitrateAbstractIntroductionIron in nitrate/nitrite standard solutionsIron in soil extractsWhy wasn’t NO3− immobilization observed by Colman et al. (2007)?Status of the ferrous wheel hypothesis and nitrate quantificationConclusionsAcknowledgmentsReferencesORIGINAL PAPERIron interference in the quantification of nitrate in soilextracts and its effect on hypothesized abioticimmobilization of nitrateEric A. Davidson Æ D. Bryan Dail Æ Jon ChoroverReceived: 16 May 2008 / Accepted: 6 August 2008 / Published online: 30 August 2008Ó Springer Science+Business Media B.V. 2008Abstract Human alteration of the nitrogen cyclehas stimulated research on nitrogen cycling in manyaquatic and terrestrial ecosystems, where analyses ofnitrate (NO3-) by standard laboratory methods arecommon. A recent study by Colman et al. (Biogeo-chemistry 84:161–169, 2007) identified a potentialanalytical interference of soluble iron (Fe) with NO3-quantification by standard flow-injection analysis ofsoil extracts, and suggested that this interference mayhave led Dail et al. (Biogeochemistry 54:131–146,2001) to make an erroneous assessment of abioticnitrate immobilization in prior15N pool dilutionstudies of Harvard Forest soils. In this paper, wereproduce the Fe interference problem systematicallyand show that it is likely related to dissolved,complexed-Fe interfering with the colorimetric anal-ysis of NO2-. We also show how standard additionsof NO3-and NO2-to soil extracts at native dissolvedFe concentrations reveal when the Fe interferenceproblem occurs, and permit the assessment of itssignificance for past, present, and future analyses. Wedemonstrate low soluble Fe concentrations and goodrecovery of standard additions of NO3-and NO2-inextracts of sterilized Harvard Forest soils. Hence, wemaintain that rapid NO3-immobilization occurred insterilized samples of the Harvard Forest O horizon inthe study by Dail et al. (2001). Furthermore, addi-tional evidence is accumulating in the literature forrapid disappearance of NO3-added to soils, suggest-ing that our observations were not the result of anisolated analytical artifact. The conditions for NO3-reduction are likely to be highly dependent onmicrosite properties, both in situ and in the labora-tory. The so-called ‘‘ferrous wheel hypothesis’’(Davidson et al., Glob Chang Biol 9:228–236,2003) remains an unproven, viable explanation forpublished observations.Keywords Iron  Nitrate  Nitrite Nitrogen  Soil extracts  Abiotic immobilization Ferrous wheel hypothesisIntroductionNitrogen (N) is among the most commonly measuredelements in biogeochemical studies owing to itsimportance as an essential nutrient for plant, animal,and microbial metabolism. The nitrate ion (NO3-)isfrequently measured in studies of atmosphericE. A. Davidson (&)The Woods Hole Research Center, 149 Woods HoleRoad, Falmouth, MA 02540-1644, USAe-mail: [email protected]. B. DailDepartment of Plant, Soil, and Environmental Sciences,University of Maine, Orono, ME 04469, USAJ. ChoroverDepartment of Soil, Water and Environmental Science,University of Arizona, Tucson, AZ 85721-0038, USA123Biogeochemistry (2008) 90:65–73DOI 10.1007/s10533-008-9231-6deposition, agricultural runoff, and wastewater fromhuman sewage systems. Increasing human alterationof the nitrogen cycle (Galloway et al. 2004;UNEPand WHRC 2007) has stimulated research on nitro-gen cycling in many aquatic and terrestrialecosystems, and many studies include analysis ofNO3-by standard water chemistry methods. Com-mon methods for quantifying NO3-concentrationsinclude ion specific electrodes, ion chromatography,and colorimetric analysis of nitrite (NO2-) followingNO3-reduction (Mulvaney 1996). The latter methodis widely used, partly because of commerciallyavailable flow injection analyzers (FIA) that auto-mate this analysis, allowing large sample numbers tobe analyzed quickly with reasonably good accuracy,precision, and sensitivity.In the process of carrying out an experimentdesigned to measure abiotic immobilization of NO3-,Colman et al. (2007) discovered a potential analyticalinterference of NO3-quantification caused by dis-solved iron (Fe) in extracts of thrice-autoclaved soils,leading to an underestimate of NO3-concentration.These authors demonstrated that NO3-concentra-tions were underestimated when they used thestandard ammonium-chloride–ethylenediaminetetra-acetic acid (NH4Cl–EDTA) buffer in FIA; that theunderestimation of NO3-was correlated with solubleFe concentrations; and that this interference did notappear when they substituted an imidazole buffer.The authors concluded that previously publishedreports of evidence for abiotic immobilization ofNO3-into a dissolved organic-N (DON) fraction(Dail et al. 2001; Davidson et al. 2003) were due tothis analytical artifact, because underestimation ofNO3-concentrations would cause an underestimationof recovery of experimentally added15NO3-labeland an over estimation of DON (estimated by thedifference between total dissolved N [TDN] anddissolved inorganic-N [DIN]).Interference of soluble Fe in NO3-quantificationis a potentially important issue not only for researchpertaining to abiotic immobilization of NO3-, butalso potentially for thousands of other studies thathave reported NO3-concentrations in soil extracts,rainwater, streamwater, lakes, groundwater, and sea-water. We have recently encountered severalgraduate students who have read the Colman et al.(2007) paper and who were worried about the validityof their NO3-concentration data from a variety ofdifferent experiments and studies. Hence it is veryimportant to accurately characterize this solubleFe interference of NO3-quantification whenemploying the widely used NH4Cl–EDTA buffermethodology.Ironically, the standard methodology found in theSSSA Methods for Soil Analysis (Bundy and Mei-singer 1994; Mulvaney 1996), which follows standardmethods published by the U.S. Environmental Protec-tion Agency (US EPA 1983) and by the AmericanPublic Health Association (APHA 1989), recommendsusing the NH4Cl–EDTA buffer expressly to remove Feinterference for NO3-quantification. The reduction ofNO3-to NO2-is typically accomplished in a copper-coated granulized cadmium (Cd) ‘‘reduction column’’,and several potential interferences of this step havebeen identified, including Fe,