The Contamination of Commercial ¹⁵N₂ Gas Stocks with ¹⁵N–Labeled Nitrate and Ammonium and Consequences for Nitrogen Fixation Measurements

Abstract

We report on the contamination of commercial 15-nitrogen (¹⁵N) N₂ gas stocks with ¹⁵N-enriched ammonium, nitrate and/or nitrite, and nitrous oxide. ¹⁵N₂ gas is used to estimate N₂ fixation rates from incubations of environmental samples by monitoring the incorporation of isotopically labeled ¹⁵N₂ into organic matter. However, the microbial assimilation of bioavailable ¹⁵N-labeled N₂ gas contaminants, nitrate, nitrite, and ammonium, is liable to lead to the inflation or false detection of N₂ fixation rates. ¹⁵N₂ gas procured from three major suppliers was analyzed for the presence of these ¹⁵N-contaminants. Substantial concentrations of ¹⁵N-contaminants were detected in four Sigma-Aldrich ¹⁵N₂ lecture bottles from two discrete batch syntheses. Per mole of ¹⁵N₂ gas, 34 to 1900 µmoles of ¹⁵N-ammonium, 1.8 to 420 µmoles of ¹⁵N-nitrate/nitrite, and ≥21 µmoles of ¹⁵N-nitrous oxide were detected. One ¹⁵N₂ lecture bottle from Campro Scientific contained ≥11 µmoles of ¹⁵N-nitrous oxide per mole of ¹⁵N₂ gas, and no detected ¹⁵N-nitrate/nitrite at the given experimental ¹⁵N₂ tracer dilutions. Two Cambridge Isotopes lecture bottles from discrete batch syntheses contained ≥0.81 µmoles ¹⁵N-nitrous oxide per mole ¹⁵N₂, and trace concentrations of ¹⁵N-ammonium and ¹⁵N-nitrate/nitrite. ¹⁵N₂ gas equilibrated cultures of the green algae Dunaliella tertiolecta confirmed that the ¹⁵N-contaminants are assimilable. A finite-differencing model parameterized using oceanic field conditions typical of N₂ fixation assays suggests that the degree of detected ¹⁵N-ammonium contamination could yield inferred N₂ fixation rates ranging from undetectable, <0.01 nmoles N L⁻¹ d⁻¹, to 530 nmoles N L⁻¹ d⁻¹, contingent on experimental conditions. These rates are comparable to, or greater than, N₂ fixation rates commonly detected in field assays. These results indicate that past reports of N₂ fixation should be interpreted with caution, and demonstrate that the purity of commercial ¹⁵N₂ gas must be ensured prior to use in future N₂ fixation rate determinations.

Introduction

Nitrogen (N) is a major nutrient required universally by photosynthetic organisms. Its availability in the environment can directly affect the ecology and productivity of terrestrial and marine ecosystems, with important implications for the regional and global carbon cycles. The natural input of bioavailable N to the biosphere is dominated by nitrogen fixation, the biological reduction of dinitrogen (N₂) gas to ammonium (NH₄ ⁺). Two methods are commonly utilized to measure N₂ fixation rates in the field, the ¹⁵N₂ tracer assay [1] and the acetylene (C₂H₂) reduction assay [2,3]. The ¹⁵N₂ tracer assay was originally developed when artificially ¹⁵N-enriched substrate N₂ first became available [4]. This approach was then superseded by the acetylene reduction technique, as the cost and availability of high precision isotope ratio measurements proved restrictive [3]. The acetylene reduction assay, however, is associated with variations in the factor used to convert C₂H₂ reduction into N₂ equivalents, and with potentially biasing effects of C₂H₂ on the physiology of N₂ fixing organisms, among other issues [5]–[7]. Interest in the ¹⁵N₂ tracer assay later regained momentum, owing to the increased affordability of Isotope Ratio Mass Spectrometry (IRMS) instrumentation and to concurrent developments in ¹⁵N tracer techniques. Today, it is generally the preferred method to quantify N₂ fixation rates in both terrestrial and aquatic environments [1], owing to its high sensitivity, and ability to provide qualitative and quantitative constraints on the translocation and the fate of biologically fixed N [8]–[10].

A salient strength of the ¹⁵N₂ tracer assay is that ¹⁵N-enrichment detected in biomass can be ascribed to the biological reduction of N₂ exclusively, as no interfering processes can carry out the reduction of ¹⁵N₂ gas concurrently. This premise requires that the ¹⁵N₂ stock be devoid of any contaminant ¹⁵N-species that could be assimilated into biomass simultaneously. However, during recent research projects on N₂ fixation conducted independently in our laboratories at the University of Connecticut Avery Point and the University of Massachusetts Dartmouth, convergent observations indicated that some commercial ¹⁵N₂ stocks could be contaminated with ¹⁵N-enriched N-species other than N₂, including nitrate, nitrite and/or ammonium. These reactive forms of N would be readily assimilated by microorganisms, leading to significantly biased (i.e., overestimated) N₂ fixation rate measurements.These observations motivated the current study, with the goal of testing whether commercially available ¹⁵N₂ stocks contain ¹⁵N-contaminants at levels that would interfere with ¹⁵N₂ tracer N₂ fixation assays, particularly in the open ocean, and to assess if such contaminants are prevalent among ¹⁵N₂ stocks from different suppliers. We thus uncovered substantial contamination of one of three brands of commercial ¹⁵N₂ gas with bioavailable inorganic ¹⁵N-species. Our findings raise important concerns regarding the pervasiveness of reactive ¹⁵N contamination of the ¹⁵N₂ stocks, and the extent to which these contaminants may have affected the magnitude of the N₂ fixation rate estimates reported in the literature. We outline steps to contend with this issue to ensure the veracity of future N₂ fixation estimates.

Methods

Reagents

Four 33 mL lecture bottles of 98+ at% ¹⁵N-labeled N₂ gas were purchased from Sigma-Aldrich (produced by their subsidiary, Isotec Stable Isotopes; St. Louis, MO; Stock Keeping Unit 364584), three from lot # SZ1670V, synthesized in 2010, and one from lot # MBBB0968V, synthesized in 2014. Two 1L lecture bottles of 98+ at% ¹⁵N₂ were purchased from Cambridge Isotopes (Tewksbury, MA, part # NLM-363-1-LB) from respective lot #’s I1-11785A and I-16727. One 1L lecture bottle of 98+ at% ¹⁵N₂ was purchased from Campro Scientific (Berlin, Germany; catalogue # CS01-185_261) from lot # EB1169V. Ammonium and nitrate solutions were prepared with salts or with solutions obtained from different distributors: sodium nitrate (NaNO₃: CAS 7631-99-4), potassium nitrate (KNO₃: CAS 7757-79-1), and ammonium chloride (NH₄Cl: CAS 12125-02-9) from Fisher Scientific (Pittsburgh, PA); analytical-grade potassium nitrate (CAS 7757-79-1) from Fluka Analytical and a gravimetric solution of ammonium chloride (catalogue # AS-NH3N9-2Y) from SPEX CertiPrep (Metuchen, NJ).

Preparation of nitrate and ammonium solutions equilibrated with ¹⁵N₂ gas

In order to determine whether the ¹⁵N₂ gas stocks contained ¹⁵N-labeled ammonia (NH₃) or nitrate and/or nitrite (NO_x) contaminants, aqueous solutions of natural abundance (unlabeled) ammonium and nitrate salts were equilibrated overnight with an air headspace supplemented with an injection of ¹⁵N₂ gas. After equilibration, the ¹⁵N/¹⁴N ratio of ammonium and the ¹⁵N/¹⁴N and ¹⁸O/¹⁶O ratios of nitrate/nitrite in solution were measured, as well as the ¹⁵N/¹⁴N ratio of N₂ gas in the headspace, as described below. The isotope ratios of nitrate and ammonium were compared to those in control solutions, which were not supplemented with ¹⁵N₂ gas. Experiments with the Campro Scientific ¹⁵N₂ stock were verified for ¹⁵N-nitrate/nitrite contaminants only (and not for ¹⁵N-ammonium).

Initial experiments consisted of 40 mL or 100 mL solutions of 10, 50, 100, 200, or 300 µmol L⁻¹ nitrate and 5 µmol L⁻¹ ammonium chloride in 60 mL or 120 mL serum vials that were sealed with Thermo Scientific gas-impermeant stoppers (part # C4020-30) or with Bellco Glass septum stoppers (catalogue # 2048-11800). The 20 mL of air headspace in each of the treatment vials was supplemented with 0.1 mL of ¹⁵N₂ gas from respective bottles from each of the three suppliers (three lecture bottles from Sigma-Aldrich lot # SZ1670V and one bottle from lot # MBBB0968V, two bottles from Cambridge Isotopes lot # I1-11785A and lot # I-16727, and one bottle from Campro Scientific lot # EB1169V). The solutions were equilibrated overnight on a shaker, after which the ¹⁵N/¹⁴N and ¹⁸O/¹⁶O isotope ratios of nitrate were analyzed as described below. The ¹⁵N/¹⁴N isotope ratio of ammonium was also analyzed (described below) in experimental solutions treated with the Sigma-Aldrich and Cambridge Isotopes stocks, but not the Campro Scientific stock.

The experimental sensitivity to ¹⁵N-contaminants was increased in subsequent experiments involving ¹⁵N₂ stocks that did not show clear evidence of contamination in the experiments described above (see Results ) by increasing the volume of ¹⁵N₂ gas injections and decreasing solution volumes. Experiments were initiated in which 2 mL ¹⁵N₂ gas was equilibrated overnight in 20 mL serum vials containing 10 mL solutions of 10 µmol L⁻¹ sodium nitrate, after which the ¹⁵N/¹⁴N and ¹⁸O/¹⁶O ratios of nitrate were measured as described below. Similarly, 10 mL solutions of 5 µmol L⁻¹ ammonium chloride were dispensed in 20 mL serum vials and equilibrated overnight with 2 mL ¹⁵N₂ gas, after which the ¹⁵N/¹⁴N isotope ratios of ammonium were analyzed (described below).

The measured ¹⁸O/¹⁶O ratios of nitrate/nitrite in solutions equilibrated with ¹⁵N₂ gas from some stocks suggested the presence of ⁴⁶N₂O contamination. As our analyte for isotope ratio analysis is N₂O, and ⁴⁶N₂O can be explained by both ¹⁵N¹⁵N¹⁶O and ¹⁴N¹⁴N¹⁸O, N₂O that is doubly labeled with ¹⁵N is falsely detected as δ¹⁸O_NO3 enrichment. The presence of ⁴⁶N₂O contamination in ¹⁵N₂ gas was verified directly for one of the Sigma-Aldrich stocks (Lot # SZ1670V) by adding 0.0125, 0.020, or 0.025 mL of ¹⁵N₂ stock to 20 mL serum vials containing 10 nmoles of reference N₂O in helium. The N and O isotopic composition of the N₂O was analyzed as described below, and compared to unamended N₂O injections.

Dunaliella tertiolecta cultures

The marine green alga Dunaliella tertiolecta was cultured in growth media equilibrated with ¹⁵N₂ gas in order to ascertain the susceptibility of ¹⁵N-labeled gas contaminants to assimilation by non-N₂-fixing organisms. Culture medium was prepared from filtered Long Island Sound sea water supplemented with 50 µmol L⁻¹ NaNO₃, 36.3 µmol L⁻¹ NaH₂PO₄*H₂O, and 107 µmol L⁻¹ Na₂SiO₃*9H₂O, as well as f/2 trace metals and f/2 vitamins [11], added from filter sterilized stock solutions. Medium (200 mL) was dispensed in 250 mL stoppered glass bottles. Experimental treatment bottles were equilibrated overnight with 0.2 mL ¹⁵N₂ gas from either a Cambridge Isotopes (lot #I-16727) or Sigma-Aldrich (lot # SZ1670V) lecture bottle. Following inoculation, cultures were left loosely capped and placed on a windowsill with exposure to natural light. Nitrate concentrations were monitored daily. Upon the complete depletion of nitrate, 8 days after inoculation, the cultures were harvested on pre-combusted GF/F filters. Filters were dried at 60°C for 18 h pending N isotopic analysis of the particulate nitrogen (described below).

Nitrate and ammonium concentrations

Nitrate concentrations in the experimental solutions were verified via reduction to nitric oxide in hot vanadium (III) solution followed by detection with a chemiluminescence NO_x analyzer (model T200 Teledyne Advanced Pollution Instrumentation) [12]. Ammonium concentrations were measured by derivatization with orthophthaldialdehyde (OPA) and fluorometric detection on an AJN Scientific f-2500 Fluorescence Spectrophotometer [13].

Nitrate N and O isotope ratio analyses

Nitrate/nitrite nitrogen (¹⁵N/¹⁴N) and oxygen (¹⁸O/¹⁶O) isotope ratios were measured using the denitrifier method [14,15]. Nitrate (and nitrite) in experimental samples was converted stoichiometrically to nitrous oxide (N₂O) by a denitrifying bacterial strain (Pseudomonas chlororaphis f. sp. aureofaciens, ATCC 13985) that lacks nitrous oxide reductase. The N and O isotopic composition of N₂O was then measured on a Delta V Advantage Isotope Ratio Mass Spectrometer (IRMS) interfaced with a modified Gas Bench II gas chromatograph (Thermo Fisher) purge and trap system. The isotope ratio measurements are reported in the conventional delta (δ) notation in per mille (‰) units, defined for N and O by the following equations:

The ¹⁵N/¹⁴N reference is N₂ in air, and the ¹⁸O/¹⁶O reference is Vienna Standard Mean Ocean water (V-SMOW). Individual analyses on the GC-IRMS were referenced to injections of N₂O from a pure N₂O gas cylinder, and then standardized through comparison to the international nitrate standards USGS-34 (δ¹⁵N of −1.8‰ vs. air; δ¹⁸O of −27.9‰ vs. V-SMOW), USGS-32 (δ ¹⁵N of +180‰ vs. air; δ¹⁸O of +25.7‰ vs. V-SMOW), and IAEA-NO-3 (δ¹⁵N of +4.7‰ vs. air; δ¹⁸O of +25.6‰ vs. V-SMOW) [16]–[18], using standard bracketing techniques. Nitrate samples from experiments with Campro Scientific ¹⁵N₂ were standardized with USGS-32 and IAEA-NO-3, and an additional internal lab nitrate standard (UBN-1; δ¹⁵N of 14.15‰ vs. air; δ¹⁸O of +25.7‰ vs. V-SMOW). Precision for analytical replicates was ≤0.2‰ for δ¹⁵N_NO3 and ≤0.2‰ for δ¹⁸O_NO3 for isotope ratio amplitudes encompassed by the standards. Above 200‰, the precision decreased in proportion to δ¹⁵N_NO3 amplitude, with standard deviations of 7.5‰ for δ¹⁵N_NO3 at or above 1000‰. Similarly, precision decreased with increasing δ¹⁸O_NO3, with standard deviations of 1.3‰ to 6.1‰ for δ¹⁸O_NO3 values ≥80‰. The poor precision of the higher range measurements is likely due to the variable contribution of a trace NO_X contaminant in denitrifier preparations with δ¹⁵N and δ¹⁸O values that are in the range of natural abundance samples [19].

Nitrous oxide N and O isotope ratio analyses

N₂O isotope ratios were measured directly on the GC-IRMS, and referenced against the N₂O tank, which was standardized indirectly by comparison to the δ¹⁵N and δ¹⁸O of nitrate standards.

Ammonium N isotope ratio analyses

The ammonium δ¹⁵N_NH4 was measured using the hypobromite-azide method [20]. Ammonium in basic solution was converted to N₂O via oxidation to nitrite (NO₂ ⁻) with hypobromite, followed by reduction of nitrite to N₂O with sodium azide in acetic acid. The δ¹⁵N of the N₂O analyte was measured on the GC-IRMS, as outlined above. Measurements were calibrated using solutions made from the international standard ammonium salts, IAEA-N1 and IAEA-N2, with assigned δ¹⁵N values of +0.4‰, +20.3‰ vs. air, respectively [16,17,21,22]. Our standard error for analytical replicates was ≤0.6‰ at relatively low ¹⁵N-abundances, but increased substantially for δ¹⁵N_NH4 from 100‰ to 9000‰, varying from 2.9‰ to as high as 59.7‰. As with the nitrate analyses, the low precision of higher range measurements likely stems from the variable contribution of a trace ammonium or nitrite contaminant with a natural abundance δ¹⁵N value, inadvertently introduced during the analyses.

Headspace N₂ isotope ratio analyses

To measure the δ¹⁵N of N₂ gas in the headspace of experimental samples, 75 µL of headspace was injected into 12 mL Exetainer vials previously flushed with helium, then analyzed on a Gas Bench II GC-IRMS (Delta V Advantage Plus) operated in continuous flow mode. N₂ and (O₂+ Ar) were separated on a 5-Å mole-sieve capillary gas chromatography column. The analyses were standardized with parallel analyses of ambient N₂ gas in air. These direct N₂ gas measurements were carried out for experiments conducted using two of three lecture bottles from Sigma-Aldrich lot # SZ1670V, and for experiments conducted using the lecture bottle from Cambridge Isotopes lot # I1–11785A. The ¹⁵N₂ concentration in the headspace of other experiments was estimated from the tracer injection volume rather than from direct measurements.

Particulate nitrogen isotope ratio analyses

The δ¹⁵N of particulate nitrogen (PN) was analyzed using a Costech Instruments elemental combustion system (model 4010) coupled to a Thermo Scientific Delta V Advantage IRMS. Analyses were standardized using L-glutamic acid reference materials, USGS-40 (δ¹⁵N of −4.52‰ vs. air) and USGS-41 (δ¹⁵N of +47.57‰ vs. air) [23].

Results

Nitrate solutions equilibrated with any of three ¹⁵N₂ gas stocks from Sigma-Aldrich lot # SZ1670V (referred to hereafter as ‘Sigma A1, A2 and A3’) showed a substantial increase in the δ¹⁵N of nitrate (and possibly nitrite) compared to control solutions in the lower sensitivity nitrate dilutions (Fig. 1a). Respective ¹⁵N enrichments evidenced by the δ¹⁵N_NO3+NO2 were inversely proportional to the concentration of nitrate in the solutions, from nearly 1000‰ at 10 µmol L⁻¹ nitrate to 30‰ at 290 µmol L⁻¹ nitrate, compared to a δ¹⁵N_NO3 of 23.5±0.5‰ in the corresponding potassium nitrate control solutions. The ¹⁵N enrichments imparted on the nitrate solutions were comparable among the three lecture bottles from this lot (# SZ1670V). The δ¹⁵N_NO3+NO2 resulting from equilibration with a single Sigma-Aldrich gas stock from lot # MBBB0968V (‘Sigma B’) was relatively modest, but still significant, averaging 28.4±0.3‰ at 10 µmol L⁻¹ nitrate, compared to 23.5±0.5‰ in the corresponding control solutions (Fig. 1a). When tested at the more sensitive experimental dilution, equilibrations of nitrate solutions with the Sigma B stock resulted in a δ¹⁵N_NO3+NO2 of 200.2±70.9‰, compared to a δ¹⁵N_NO3 of 1.3±0.1‰ in control solutions (Fig. 2a). These measurements thus indicate that N₂ gas stocks sourced from lot # SZ1670V contained 410±80 µmoles of ¹⁵N-nitrate and/or nitrite per mole of ¹⁵N₂, whereas the bottle from lot # MBBB0968V contributed 1.8±0.6 µmoles of ¹⁵N-nitrate and/or nitrite per mole of ¹⁵N₂ (Table 1). The ¹⁵N-nitrate additions were determined by a mass balance calculation:

where ¹⁵NO₃ ⁻ + NO₂ ⁻ _added is the moles of ¹⁵N-labeled nitrate and/or nitrite added by the ¹⁵N₂ gas injection, δ¹⁵N_{NO3+NO2,added} is presumed to be equivalent to the δ¹⁵N of ¹⁵N₂ tracer gas (266,540‰), NO₃ ⁻ _initial and NO₃ ⁻ _final refer to the moles of nitrate in solution before and after ¹⁵N₂ equilibration, and δ¹⁵N_NO3,_initial and δ¹⁵N_NO3+NO2,_final refer to the δ¹⁵N_NO3 of nitrate solutions before and after ¹⁵N₂ equilibration. The quantity of ¹⁵N₂ gas added to experimental treatments was measured explicitly in Sigma A1 and A2 bottles, and was calculated from the ¹⁵N₂ injection volumes for experiments treated with Sigma A3 and B stocks.

Figure 1. (a) δ¹⁵N_NO3+NO2 (log scale) of nitrate solutions (10–300 µmol L⁻¹) following equilibration with 0.1 mL ¹⁵N₂ gas from lecture bottles procured from three distributors.

Solutions were 40 mL for Sigma-Aldrich and Campro Scientific equilibrations, and 100 mL for Cambridge Isotopes equilibrations. The solid line corresponds to the δ¹⁵N_NO3 of the control solutions for Sigma-Aldrich and Cambridge Isotopes experiments (δ¹⁵N_NO3 = 23.5±0.5‰); the dashed line corresponds to controls for Campro Scientific experiments (δ¹⁵N_NO3 = 14.15±0.1‰). Paired symbols identify replicate experimental treatments. (b) Corresponding apparent δ¹⁸O_NO3+NO2 of the experimental nitrate solutions. The solid line corresponds to the δ¹⁸O_NO3 of control solutions for the Sigma-Aldrich and Cambridge Isotope experiments (δ¹⁸O_NO3 = 18.9±0.3‰); the dashed line corresponds to controls for Campro Scientific experiments (25.4±0.3‰).

Figure 2. (a) δ¹⁵N_NO3+NO2 (log scale) of higher sensitivity equilibrations of 10 µmol L⁻¹ nitrate solutions (10 mL) with 2 mL of ¹⁵N₂ gas from a Cambridge Isotopes or a Sigma-Aldrich bottle.

(b) Corresponding apparent δ¹⁸O_NO3 (log scale) of higher sensitivity equilibrations of the two stocks. n = the number of experimental replicates.

Table 1. The quantity of ¹⁵N-labeled contaminants detected relative to ¹⁵N₂ additions.

µmoles 15NX per mole 15N2
	15NO3 −/NO2 −	15NH4 +	46N2O
Sigma A1lot # SZ1670V	420±110	34±11	≥21±3
Sigma A2lot # SZ1670V	420±40	520±30	109±5§
Sigma A3 †lot # SZ1670V	350±80	N/A	≥63±15
Sigma B †lot # MBBB0968V	1.8±0.6	1900±560	≥49±17
Cambridge Alot # I1-11785A	n.d.*	0.052±0.020	n.d.*
Cambridge B †lot # I-16727	0.024±0.006	0.014±0.004	≥0.81±0.24
Campro Scientific†lot # EB1169V	n.d.*	N/A	≥11±3

The µmoles of ¹⁵N contaminants (NO₃ ⁻+NO₂ ⁻, NH₄ ⁺, and N₂O) detected per mole of ¹⁵N₂ gas from lecture bottles provided by different suppliers. N/A = not available; n.d. = not detected.

*Not explicitly tested in high sensitivity ¹⁵N₂ dilutions.

^†Moles of ¹⁵N₂ estimated from the injection volume rather than direct measurements.

^{§ 46}N2O measured directly.

In contrast to the Sigma-Aldrich stocks, ¹⁵N_NO3+NO2 contaminants were significantly lower, or possibly absent, in Cambridge Isotopes and Campro Scientific ¹⁵N₂ stocks. The δ¹⁵N_NO3+NO2 values of solutions treated with Cambridge Isotopes ¹⁵N₂ gas (lots # I1–11785A and I-16727, hereafter referred to as Cambridge A and Cambridge B, respectively) and with Campro Scientific ¹⁵N₂ gas (lot # EB1169V) were indistinguishable from those of control solutions at all experimental nitrate concentrations in the lower sensitivity tests (Fig. 1a). In the more sensitive experimental treatments, however, solutions treated with Cambridge B ¹⁵N₂ gas (lot # I-16727) had a δ¹⁵N_NO3+NO2 of 4.8±0.8‰, compared to a δ¹⁵N_NO3 of 1.3±0.1‰ in control solutions (Fig. 2a). This stock thus contributed trace contaminants on the order of 0.024±0.006 µmoles of ¹⁵N-nitrate and/or nitrite per mole of ¹⁵N₂ (Table 1). Nitrate isotope ratios in the Cambridge A (lot # I1–11785A) and Campro Scientific ¹⁵N₂ gas stocks were not tested at these lower experimental dilutions.

In treatments using several ¹⁵N₂ gas stocks, δ¹⁸O_NO3 was found to be elevated relative to control solutions. The denitrifier method, employed for δ¹⁸O_NO3 measurements, involves the bacterial reduction of NO₃ ⁻ and NO₂ ⁻ to N₂O, and the subsequent analysis of N₂O using an IRMS. However, the elevated δ¹⁸O_NO3 values detected within experimental treatments are expressly not explained by the formation of ¹⁴N¹⁴N¹⁸O during bacterial reduction of ¹⁵N-enriched nitrate, which could only account for a negligible portion of the observed δ¹⁸O_NO3+NO2 increase. Instead, the values are best explained by the presence of doubly-labeled ¹⁵N^–N₂O (i.e., ⁴⁶N₂O) in the ¹⁵N₂ gas stocks. The apparent δ¹⁸O_NO3+NO2 of nitrate solutions equilibrated with all of the Sigma-Aldrich stocks, the Campro Scientific stock, and the Cambridge B ¹⁵N₂ stock proved to be greater than that of control solutions in the low sensitivity treatments (Fig. 1b). At 10 µmol L⁻¹ nitrate, the apparent δ¹⁸O_NO3+NO2 of treated solutions was 188.5±83.8 among the Sigma A1-A3 stocks (lot # SZ1670V), 169.8±17.9‰ for the Sigma B stock (lot # MBBB0968V), and 20.1±0.2 for the Cambridge B stock, compared to 18.9±0.3‰ in corresponding control solutions. The apparent δ¹⁸O_NO3+NO2 of the Campro Scientific stock at 10 µmol L⁻¹ nitrate was 70.4±1.4‰, compared to 25.7±0.1‰ in corresponding control solutions. The apparent δ¹⁸O_NO3+NO2 of the samples decreased coherently with increasing nitrate concentrations for respective stocks. The apparent δ¹⁸O_NO3+NO2 values of solutions equilibrated with the Cambridge A stock, at 19.1±0.2‰, were not distinguishable from the control solutions. In the more sensitive equilibrations, nitrate solutions equilibrated with Sigma B ¹⁵N₂ gas had a δ¹⁸O_NO3+NO2 of 13,129±1186‰ compared to 23.9±0.2‰ in control solutions, whereas the δ¹⁸O_NO3+NO2 of the Cambridge B stock was 216.7±78.4‰ (Fig. 2b). Given that the apparent δ¹⁸O_NO3 enrichments are explained by the presence of ⁴⁶N₂O, the inverse relationship between δ¹⁸O values and nitrate concentration stems from the fact that the detected ⁴⁶N₂O derives from ⁴⁶N₂O dissolved in the nitrate solutions, and solutions containing higher nitrate concentrations require lower sample volume injections when using the denitrifier method for IRMS analysis. The observed excess ⁴⁶N₂O levels indicate ¹⁵N¹⁵N¹⁶O contaminants (µmole ⁴⁶N₂O per mole of ¹⁵N₂) on the order of 41±21 among the Sigma A1-A3 bottles, 49±17 in Sigma B, 11±3 in Campro Scientific, and 0.81±0.24 in Cambridge B (Table 1). The Cambridge A bottle was not tested in higher sensitivity dilutions that could have revealed the presence of some N₂O therein. The presence of ⁴⁶N₂O contaminant was verified directly for the Sigma A2 lecture bottle from analyses of N₂O amended with injections of ¹⁵N₂ gas. Among four experimental samples, 109±5 µmoles of ⁴⁶N₂O were detected per mole of Sigma A2 ¹⁵N₂ added, more than double the ⁴⁶N₂O that was detected in samples analyzed by the denitrifier method (39±8 µmole ⁴⁶N₂O per mole of ¹⁵N₂). This discrepancy likely resulted because samples analyzed by the denitrifier method were uncapped immediately prior to their injection into P. aureofaciens denitrifier cultures, allowing N₂O to escape to the atmosphere. As contaminant N₂O was not the target analyte of the denitrifier measurements, precautions were not taken to prevent N₂O gas loss at this step. The ⁴⁶N₂O concentrations derived from solution equilibrations of respective ¹⁵N₂ stocks thus constitute lower limits (Table 1).

Solutions equilibrated with Sigma-Aldrich ¹⁵N₂ gas showed substantial ¹⁵N-enrichments of ammonium compared to control solutions (Fig. 3a): Equilibration with ¹⁵N₂ from the Sigma A1 lecture bottle (lot # SZ1670V) resulted in a δ¹⁵N_NH4 of 99±39‰, compared to 0.6±0.5‰ for the control solution (NH₄Cl, SPEX CertiPrep); equilibration with ¹⁵N₂ from the Sigma A2 bottle (lot # SZ1670V) yielded a δ¹⁵N_NH4 of 940±60‰, compared to 7.6±0.3‰ for the control solution (NH₄Cl salt, Fisher Scientific); equilibration with ¹⁵N₂ from the Sigma B bottle (lot # MBBB0968V) resulted in a δ¹⁵N_NH4 of 7030±2100‰, compared to 9.0±0.06‰ for the corresponding control solutions (NH₄Cl salt, Fisher Scientific). Mass balance calculations based on these isotope ratio values thus evidence the presence of 34±11, 518±26, and 1890±560 µmoles of ¹⁵N-ammonium per mole of ¹⁵N₂ injected from Sigma A1, A2, and B ¹⁵N₂ bottles, respectively (Table 1). Unlike ¹⁵N-labelled nitrate/nitrite contaminants, the ¹⁵N-ammonium contaminants appeared to be variable among bottles of lot # SZ1670V (Fig. 3a). The Sigma A3 lecture bottle was not tested for ¹⁵N-ammonium.

Figure 3. (a) δ¹⁵N_NH4 (log scale) of 5 µmol L⁻¹ ammonium solutions after equilibration with 0.1 mL ¹⁵N₂ gas from respective Sigma-Aldrich and Cambridge Isotopes lecture bottles vs. control solutions.

Sigma-Aldrich treatments utilized 40 mL ammonium solutions, whereas Cambridge Isotopes treatments utilized 100 mL ammonium solutions. (b) δ¹⁵N_NH4 of higher sensitivity equilibrations of 5 µmol L⁻¹ ammonium solutions (10 mL) with 2.0 mL ¹⁵N₂ gas from Cambridge Isotopes lecture bottles vs. control solutions. n = the number of experimental replicates.

In contrast to Sigma-Aldrich stocks, ammonium solutions equilibrated with ¹⁵N₂ from the Cambridge A bottle had a δ¹⁵N_NH4 of 8.3±1.0‰, comparable to that of the corresponding control solution of 7.6±0.3‰ (NH₄Cl salt, Fisher Scientific) in the lower sensitivity experiments (Fig. 3a). In the more sensitive dilutions, however, ¹⁵N-ammonium contaminants were detected in both of the Cambridge A and B stocks (Fig. 3b). Solutions equilibrated with Cambridge A had a δ¹⁵N_NH4 of 24.0±5.9‰, compared to 8.7±0.1‰ for the control solutions (NH₄Cl salt, Fisher Scientific) and solutions equilibrated with Cambridge B had a δ¹⁵N_NH4 of 13.1±1.1‰, compared to 9.0±0.1‰ for the control solutions ((NH₄Cl salt, Fisher Scientific). The enrichment relative to control solutions invariably originates from a ¹⁵N-ammonium contaminant, and cannot be attributed to a trace N₂O (¹⁵N¹⁴N¹⁶O) contaminant, because the samples were purged when conducting ¹⁵N-ammonium analyses, following the oxidation of ammonium to nitrite with hypobromite. These more sensitive treatments thus reveal the presence of minuscule ¹⁵N-ammonium concentrations in the Cambridge Isotopes stocks, on the order 0.052±0.020 and 0.014±0.004 µmoles of ¹⁵N-ammonium per mole of ¹⁵N₂ gas in lots # I1–11785A and I-16727, respectively (Table 1).

The control solutions in the ¹⁵N-ammonium experiments prepared from a single Fisher Scientific NH₄Cl salt stock revealed progressively heavier mean δ¹⁵N_NH4 values among experiments, at 7.6±0.3‰, 8.7±0.1‰, or 9.0±0.06‰. The solutions with a δ¹⁵N_NH4 estimated at 7.6±0.3‰ and 8.7±0.1‰ were made fresh from the salts for each experiment, such that the ¹⁵N-enrichment of ammonium cannot be attributed to the progressive degassing of ammonia in solution during storage. In turn, isotopic standards for NH₄ ⁺ (IAEA-N1 and N2) were stored in acidic solution, and thus were not subject to progressive degassing. Moreover, degassing of the isotopic standards would manifest as progressively lower δ¹⁵N values measured for control solutions. Inter-batch variability intrinsic to the hypobromite-azide method [20] is plausible, as this technique is relatively recent, such that subtle sensitivities may not yet be apparent.

D. tertiolecta cultures grown in medium equilibrated with ¹⁵N₂ gas from the Sigma A3 bottle expectedly showed substantial ¹⁵N enrichment of particulate nitrogen (δ¹⁵N_PN), averaging 44.3±6.1‰ among triplicate treatment cultures compared to 1.9±0.1‰ in control cultures (Fig. 4). Conversely, the δ¹⁵N_PN of cultures equilibrated with Cambridge B ¹⁵N₂ gas was 2.1±0.2‰, and thus not detectably different from that of control cultures, at 1.9±0.1‰ (Fig. 4).

Figure 4. The δ¹⁵N of particulate nitrogen (δ¹⁵N_PN) of D. tertiolecta harvested in stationary phase following growth in media containing sodium nitrate (and no ammonium) and equilibrated with ¹⁵N₂ gas from Sigma Aldrich or Cambridge Isotopes.

n = the number of experimental replicates.

Discussion

This study reveals that some commercial ¹⁵N₂ gas stocks contain contaminant ¹⁵N-labeled bioavailable nitrogen species, including nitrate/nitrite, ammonium and nitrous oxide. Substantial levels of ¹⁵N-labeled nitrate/nitrite, ammonium, and nitrous oxide were detected in Sigma-Aldrich stocks from lot # SZ1670V. Another Sigma-Aldrich stock from a different lot (# MBBB0968V) contained considerably less, but still significant, ¹⁵N-nitrate/nitrite contaminants, similar nitrous oxide concentrations, and a greater concentration of ¹⁵N-ammonium. Cambridge Isotopes stocks, in turn, contained relatively low concentrations of ¹⁵N-nitrate/nitrite, ¹⁵N-ammonium and ¹⁵N-nitrous oxide. A ¹⁵N₂ stock from Campro Scientific contained no detected ¹⁵N-nitrate/nitrite contaminants in low sensitivity experiments, but measurable ¹⁵N-nitrous oxide. Indeed, a certificate of analysis provided by Campro Scientific attests that stocks may contain up to 15 ppm N₂O. ¹⁵N-ammonium was not analyzed in Campro Scientific gas, nor was the stock tested at more sensitive ¹⁵N₂ dilutions, which could have revealed trace ¹⁵N-nitrate or ammonium contaminants in the stock. In any case, ¹⁵N-contamination with nitrous oxide is of no obvious consequence for biological ¹⁵N₂ applications, such as N₂ fixation rate measurements. However, the presence of ¹⁵N-nitrate, nitrite and ammonium has serious implications for measurements of N₂ fixation, as these contaminants could lead to the detection of false positives or inflated rates.

The propensity of the detected ¹⁵N-labeled contaminants to be assimilated into biomass was verified directly from cultures of D. tertiolecta, which acquired elevated δ¹⁵N of PN in media equilibrated with a Sigma-Aldrich ¹⁵N₂ stock. Expectedly, media equilibrated with ¹⁵N₂ from a Cambridge Isotopes stock did not cause detectable ¹⁵N-enrichment of biomass. At the given experimental conditions, however, the complete assimilation of the contaminant ¹⁵N-nitrate/nitrite from the Sigma-Aldrich stock should have yielded greater δ¹⁵N_PN values than those observed, of at least 56.3±5.4‰ (vs. control δ¹⁵N_PN values of ∼1.9‰), notwithstanding the additional contribution of any ¹⁵N-ammonium contaminant (¹⁵N-ammonium was not measured explicitly in the Sigma A3 stock). This discrepancy is difficult to reconcile. We tentatively posit that ¹⁵N-nitrite comprises a substantial fraction of the trace ¹⁵N-nitrate/nitrite contaminant, and that D. tertiolecta may not be able to transport nitrite at nanomolar to sub-nanomolar concentrations. Indeed, such trace nitrite concentrations are likely below the thresholds achievable by micro-algal nitrite transport systems [24].

The contaminants in the ¹⁵N₂ stocks ostensibly derive from the method of ¹⁵N₂ gas production. ¹⁵N₂ gas is generally produced by the catalytic oxidation of ¹⁵N ammonia (¹⁵NH₃) gas with cupric oxide [25]. The bulk of the oxidation product is N₂ gas, although more oxidized N species are also produced in lesser quantities, specifically N₂O and NO [26]. Thus, potential contaminants in a ¹⁵N₂ gas stock would expectedly consist of unreacted ammonia gas, N₂O, and nitric oxide (NO). In contact with any oxygen and water vapor, NO would inadvertently be oxidized to nitric and nitrous acid [27], which would, in turn, dissociate to nitrate and nitrite upon dissolution in water, respectively. Purification of the ¹⁵N₂ gas from unreacted ammonia and from the generated nitrogen oxides involves sequential acid and alkaline scrubbing, respectively [28] and/or cryo-trapping of ammonia and NOx gases. Upon personal communication, Cambridge Isotopes and Campro Scientific did not provide details on their method of ¹⁵N₂ production, whereas Sigma-Aldrich reported that the company’s subsidiary, Isotec, produces ¹⁵N₂ gas by the catalytic oxidation of ¹⁵N-ammonia gas with cupric oxide, followed by sequential rounds of cryo-trapping and alkaline scrubbing of the N₂ gas to increase purity.

In order to gauge the extent to which the observed ¹⁵N-ammonium contamination of Sigma-Aldrich and Cambridge Isotopes ¹⁵N₂ gas could skew estimates of N₂ fixation in incubations with ¹⁵N₂ gas, we modeled a field incubation experiment in which microorganisms assimilate the ¹⁵N-ammonium contaminant rather than reduce ¹⁵N₂ gas. A simple finite-differencing model of a ‘typical’ oceanic N₂ fixation assay was devised, in which 0.1 mL of ¹⁵N₂ gas was equilibrated in two different water sample volumes, 0.25 L or 4.5 L, then incubated for 24 hours. Prescribed biomass and growth rates were characteristic of those at the oligotrophic surface ocean, namely, a particulate N stock of 0.2 µmol L⁻¹ assimilating ammonium at a specific growth rate coefficient (µ) of 0.1 d⁻¹ to 0.3 d⁻¹, with a recycling rate (the rate at which particulate N is returned to the ammonium pool) equivalent to the respective growth rate. The prescribed δ¹⁵N of the initial particulate N was 0‰ [29], and the δ¹⁵N of ambient ammonium was −2‰ [30]. Incremental initial concentrations of ambient ammonium were prescribed, from 1 nmol L⁻¹ to 1 µmol L⁻¹. Ammonium concentrations in surface oligotrophic waters are typically very low (≤10 nmol L⁻¹), however, ammonium is a pervasive contaminant that could easily be introduced during sample preparation, as well as leached from incubation vial septa. We note that the ¹⁵N-ammonium introduced by the ¹⁵N₂ gas, while substantial in terms of the ¹⁵N/¹⁴N ratio of ammonium, is on the order of ∼20 nanomolar at most (under the modeled conditions), and thus has minimal effect on ambient ammonium concentrations. ¹⁴N-ammonium contamination is expected to be negligible, given the method of ¹⁵N₂ gas synthesis. Finally, ¹⁵N-ammonium assimilation was simulated for the broad range of ¹⁵N-ammonium contaminant concentrations observed among Sigma-Aldrich and Cambridge Isotopes lecture bottles. N₂ fixation rates inferred from the simulated δ¹⁵N increase of particulate N were computed based on the formulation of Montoya (1996):

[PN]_Δ is the change in particulate nitrogen concentration, [PN]_f is the final particulate nitrogen concentration, A _PNf is the final ¹⁵N enrichment of particulate nitrogen, A _PN0 is the initial ¹⁵N enrichment of particulate nitrogen, A _N2 is the ¹⁵N enrichment of the N₂ available for fixation, V is the specific rate of N₂ uptake, and ρ is the volumetric rate of N₂ fixation.

The model-derived ‘N₂ fixation rates’ resulting from Sigma-Aldrich ¹⁵N-ammonium contaminant levels ranged from undetectable, <0.01 nmol N L⁻¹ d⁻¹, to as high as 530 nmol N L⁻¹ d⁻¹ under the modeled conditions (Table 2). Rates were clearly sensitive to the concentration of ¹⁵N-contaminant, the ambient ammonium concentration, the incubation volume, and the specific growth rate. At the lower level of ¹⁵N-ammonium contaminant observed in the Sigma-Aldrich stocks, N₂ fixation rates were comparable to rates observed in situ for nearly all parameter permutations, from<0.01 to 9 nmoles N L⁻¹ d⁻¹. N₂ fixation rates reported for marine environments cover a broad range, from 0.01 nmoles L⁻¹ d⁻¹ to tens of nmoles N L⁻¹ d⁻¹ [31]–[33]. Rates simulated with the highest observed level of contaminant, in smaller incubation volumes at given ¹⁵N₂ additions, and/or with low ambient ammonium concentrations, tended to surpass rates observed in situ by 10 to 100 fold. The N₂ fixation rates modeled using the minute contaminant level detected in a Cambridge Isotopes stock ranged from undetectable to 0.02 nmoles N L⁻¹ d⁻¹ (Table 2), approximating the lower limit of some N₂ fixation rates reported in the literature [31],[34]–[38]. These simulated rates can be deemed conservative since the model does not account for any assimilation of contaminant ¹⁵N-nitrate/nitrite, and the 0.1 mL ¹⁵N₂ injection volume used in the model is on the lower end of ¹⁵N₂ injection volumes typically used in open ocean N₂ fixation rate measurements.

Table 2. Inferred N₂ fixation rates (nmoles N L⁻¹ d⁻¹) resulting from ¹⁵N-labeled contaminants.

Ambient [NH4+]	Cambridge Isotopes			Sigma-Aldrich			Sigma-Aldrich
(µmol L−1)	Cambridge A			Sigma A1			Sigma B
	Lot # I1–11785A			Lot # SZ1670V			Lot # MBBB0968V
	0.52 µmol 15NH4 +/mol 15N2			25 µmol 15NH4 +/mol 15N2			1900 µmol 15NH4 +/mol 15N2
	µ = 0.1	µ = 0.2	µ = 0.3	µ = 0.1	µ = 0.2	µ = 0.3	µ = 0.1	µ = 0.2	µ = 0.3	Incubation Volume (L)
0.001	0.019	0.019	0.019	9.0	9.0	9.0	310	460	530	0.25
0.01	0.014	0.016	0.016	7.5	8.5	8.6	250	400	470
0.1	n.d.	n.d.	n.d	1.6	2.7	3.6	90	170	220
1	n.d.	n.d.	n.d.	0.17	0.32	0.45	13	24	34
0.001	n.d.	n.d.	n.d.	0.50	0.50	0.50	38	38	38	4.5
0.01	n.d.	n.d.	n.d.	0.42	0.47	0.48	30	35	36
0.1	n.d.	n.d.	n.d.	0.08	0.15	0.19	6.5	11	15
1	n.d.	n.d.	n.d.	n.d.	0.012	0.017	0.7	1.4	1.9

N₂ fixation rates that would be inferred from 24-h field N₂ fixation assays conducted with ¹⁵N₂ stocks containing the respective concentrations ¹⁵N-ammonium contaminants detected in Sigma-Aldrich and Cambridge Isotopes ¹⁵N₂ gas. In the simulations, microbial plankton assimilate ¹⁵N-ammonium rather than fix ¹⁵N₂. Incubations are simulated in volumes of 0.25 L or 4.5 L equilibrated with 0.1 mL of ¹⁵N₂ gas, with 2.0×10⁻⁷ µmol L⁻¹ of plankton nitrogen (with a δ¹⁵N = 0‰) assimilating at a range of specific growth rates, µ (d⁻¹), countered by equivalent recycling rates, at incremental concentrations of ambient ammonium (δ¹⁵N_NH4 = −2‰). Inferred rates of <0.01 nmoles N L⁻¹ d⁻¹ are considered undetectable (n.d.).

Based on the simulations above, the likelihood of N₂ fixation rates being inflated when using contaminated ¹⁵N₂ gas stocks is high. It is surprising, then, that contamination of the ¹⁵N₂ stocks has not been reported previously. While growth solely upon N from N₂ fixation would eliminate the effect of ¹⁵N-labeled bioavailable contaminants, it is expected that nitrate and ammonium assimilation would be rapid relative to N₂ fixation due to the prohibitive energetic cost of N₂ fixation [39]. A review of pertinent literature reveals that soil scientists were once aware of the possible contamination of ¹⁵N₂ with bioavailable N, and took steps to mitigate it [28,40]. However, to the best of our knowledge, there is no mention of potential contamination of ¹⁵N₂ stocks in the marine literature, or in more recent terrestrial literature. The fact that this issue has gone unnoticed could mean that major contamination of ¹⁵N₂ gas stocks, such as that observed here in Sigma-Aldrich stocks, could be limited to the current lots. Supporting the notion that contamination is rare is the observation of undetectable N₂ fixation rates at the surface ocean, where phytoplankton readily assimilate ammonium [38] – even in investigations utilizing the Sigma-Aldrich (Isotec) ¹⁵N₂ gas [31]. However, a representative at Isotec stated that their procedures for synthesis and purification of ¹⁵N₂ gas have not changed in past decades, which suggests that ¹⁵N-contaminants may have been pervasive in previous lots. Failure to detect interferences from ¹⁵N-contaminants in previous studies may then stem from incubation conditions conspiring to yield expected rates of apparent N₂ fixation in spite of the presence of ¹⁵N contaminants (Table 2). Interference of ¹⁵N contaminants on N₂ fixation rate measurements may then be relatively minor in systems where bioavailable N assimilation rates are low and/or where ambient nitrate and ammonium concentrations are relatively elevated (≥100 nmol L⁻¹; Table 2), as ambient assimilable N effectively diminishes ¹⁵N enrichment resulting from ¹⁵N-labeled contaminants.

It is difficult, if not impossible, to discern whether N₂ fixation rate estimates in previous studies may have been confounded due to the assimilation of ¹⁵N contaminants in ¹⁵N₂ gas stocks. Given that ¹⁵N₂ stocks from only one of the three suppliers tested here contained contaminants to an extent that would interfere with any but the lowest reported N₂ fixation measurements, there is a strong likelihood that published estimates performed with ¹⁵N₂ from the other two suppliers have not been significantly inflated by labeled contaminants. In fact, many estimates may be lower than reality due to the incomplete equilibration of ¹⁵N₂ gas with the incubation medium, a pervasive problem with aqueous ¹⁵N₂ fixation assays that was diagnosed only recently [41]–[43]. Nevertheless, it is advisable at this point to analyze commercial ¹⁵N₂ stocks prior to their use to ensure their relative purity. In doing so, particular attention must be paid to the lower limit of detection for N₂ fixation rates. In recent years, workers have reported estimates of very low rates (≤0.1 nmol L⁻¹ d⁻¹) in environments where N₂ fixation is otherwise unexpected, which include oxygen-deplete regions of the water column at Pacific margins [34,35,38], as well as in the Beaufort Gyre of the Arctic Ocean [36]. Such minimal rates are questionable, considering that the relatively clean Cambridge Isotopes ¹⁵N₂ gas was found to contain enough ¹⁵N-ammonium to infer N₂ fixation rates of up to 0.02 nmoles N L⁻¹ d⁻¹. Campro Scientific and other commercially available ¹⁵N₂ gas stocks could similarly contain minute, but significant, concentrations of ¹⁵N-nitrate or ammonium. Therefore, it behooves investigators to not only verify the purity of their commercial ¹⁵N₂ prior to use, but also to generate constraints as to the lower limit of detection, allowing for the possibility that a trace-level ¹⁵N-contaminant could interfere with the detection of diminutive N₂ fixation rates.

Steps toward mitigation

The catalytic synthesis of ¹⁵N₂ gas from ¹⁵N-ammonia gas invariably entails the incidence of ¹⁵N-ammonium and ¹⁵N^–NOx contaminants, the removal of which is dependent on the stringency of scrubbing procedures to which a given batch is subjected. The consistency of ¹⁵N-nitrate/nitrite measurements among bottles from an individual lot from Sigma-Aldrich (Sigma A1-A3), in contrast to the lower ¹⁵N-nitrate/nitrite detected in a subsequent lot (Sigma B), supports the premise that the levels of ¹⁵N-contaminants are associated with discrete batch syntheses, identified by lot numbers, rather than with individual lecture bottles. The variability in ¹⁵N-ammonium among lecture bottles of the same lot suggests that ammonia gas does not disperse homogeneously in compressed N₂ gas. In any case, large-scale batch syntheses of ¹⁵N₂ occur relatively infrequently, on the order of every 2 years at Isotec (subsidiary of Sigma-Aldrich). We currently have a verbal agreement with Isotec to perform nitrate and ammonium isotopic analyses of ¹⁵N₂ batches, toward providing a certificate of analysis ensuring adequate purity for N₂ fixation assays. In the meantime, we advise that workers procure low-contaminant stocks from lots that we tested here. The very high purity of the batches from these suppliers suggests stringent and efficacious purification protocols, such that batches synthesized by these groups in the future are likely to be equally pure, notwithstanding the potential for human error during synthesis or subsequent purification.

Regardless of ‘expected’ purity, we recommend that workers explicitly test new batches availed by respective suppliers for ¹⁵N-nitrate and ammonium prior to using them in N₂ fixation assays, and actively disseminate the results to targeted web-based forums. To test a given batch, ¹⁵N₂ gas can be equilibrated with nitrate and ammonium solutions following protocols akin to the low and high sensitivity equilibrations herein. A number of laboratories perform commercial nitrate isotope analyses routinely at a modest cost per sample. Ammonium isotope analyses are substantially more involved, but are also performed routinely by a number of laboratories.

We further recommend that pertinent publications include not only the brand of ¹⁵N₂ stock, but also the associated lot number, and references to reported contaminants. With continued testing, our understanding of the prevalence of commercial ¹⁵N₂ contamination will grow and shed light on this problem, which may have plagued N₂ fixation estimates in the past.

Data Availability

The authors confirm that all data underlying the findings are fully available without restriction. Data files are available from Figshare using the link http://dx.doi.org/10.6084/m9.figshare.1170194.

Funding Statement

This work was supported by the following sources of funding: National Science Foundation Ocean Sciences, 1233897, JG; National Science Foundation Ocean Sciences, 1130495, PHM, MAA (http://www.nsf.gov/div/index.jsp?div=OCE); University of Connecticut Research Funding Large Grant, 461510, JG (http://research.uconn.edu); and Swiss National Science Foundation R’Equip 121258, MFL (http://www.snf.ch/en/Pages/default.aspx). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.