Introduction

Ammonia (NH3) was likely a prerequisite for the formation of amino acids and peptides on the Hadean Earth leading up to the origin of life. During the Hadean Eon, Fe2+ - bearing iron sulfide mineral surfaces likely existed in and around aqueous environments and may have served as a substrate for NH3 formation. This scenario provides a possible bridge between the iron-sulfur clusters observed within many present day enzymes (such as active sites present in nitrogenase and hydrogenase (Bertini et al. 2007)) and the mineral surfaces from which they may have evolved (Peters and Williams 2012; McGlynn et al. 2009).

It has been hypothesized previously that the levels of ferrous iron (Fe2+), sulfate (SO4 2−), nitrite (NO2 −) and nitrate (NO3 −) in the early ocean were likely significant and relevant in the context of prebiotic chemistry (Summers and Chang 1993; Walker and Brimblecombe 1985; Zohner and Broda 1979). Previous investigations have suggested that the reduction of NO2 − and NO3 − to NH3/NH4 + by aqueous Fe2+ may have been a key reaction for the fixation of nitrogen in prebiotic chemical environments (Summers and Chang 1993; Blöchl et al. 1992). Subsequent work also demonstrated that the ferrous iron-bearing sulfide FeS could facilitate the transformation of NO2 − to NH4 + at pH values between 4.7 and 6.9 (Summers 2005). Additional work in this area has shown that NH3/NH4 + can be generated from aqueous NO2 − (and NO3 −) in the presence of dissolved ferrous iron (Summers and Chang 1993) and a variety of Fe-containing surfaces, including Fe0 (s) (Smirnov et al. 2008), pyrite (Singireddy et al. 2012), pyrrhotite, basalt, and Fe3O4 (Brandes et al. 1998). Despite the myriad of previous work on this system there has been little molecular-level insight offered by prior studies that explains the reaction mechanism or mechanisms associated with the reduction of NO2 − and NO3 −. Furthermore, previous work examined this chemistry under acidic to mildly basic solution conditions, even though more alkaline environments may have existed during the Hadean (Russell et al. 1988, 1989, 1993; Macleod et al. 1994; Singireddy et al. 2012; Summers et al. 2012).

In this contribution, research is presented that investigated NO2 − and NO3 − reduction to NH3 under reducing alkaline conditions, simulating the FeS-containing hydrothermal vent environments which Russell has proposed to have existed on early Earth (Russell et al. 1988, 1989; 1993; Macleod et al. 1994). Specifically, Russell proposed that an FeS “protomembrane” existed in hydrothermal vent environments that acted as a cell membrane-like partition between the alkaline crust environment and the acidic Hadean ocean (Russell and Hall 1997). Our research study investigated chemistry in an alkaline FeS chemical environment and examined the ability of this alkaline environment to support NH3 production via the reduction of NO2 − and NO3 − . Specifically, this contribution focuses on both the kinetics of the reaction and the identity of surface-bound intermediates associated with the reduction of NO2 − and NO3 − to NH3 (detected in the conjugate acid form, NH4 +) at reaction temperatures of 22, 70, and 120 °C. In contrast to earlier work in this area, freshly precipitated FeS was selected as the reactive mineral surface rather than commercial FeS (Summers 2005). Commercial FeS is typically a pyrrhotite structure and by contrast, freshly precipitated FeS is a well defined, and well characterized nano-sized material with the mackinawite structure (Michel et al. 2005) similar in size and composition to what is thought to have been present in the Hadean around vents associated with the alteration of primitive crust (Russell et al. 1993). Results from attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR) spectra are presented and provide molecular-level insight into mechanistic details associated with the FeS surface-based reduction of NO2 − and NO3 − . Finally, interpretation of these results and their significance in the context of mineral based organic synthesis and the origin of life on Earth are presented.

Experimental

Nano-FeS and Reagent Preparation

Nano-sized iron sulfide (FeS) used in this study was prepared by the aqueous reaction of ferrous iron (Fe2+) with sulfide (S2−) (Rickard 1968; Michel et al. 2005). In brief, 1 M solutions of FeSO4 and Na2S.9H2O (ACS reagent grade, used without further purification) were prepared at room temperature with freshly drawn Ar-sparged, O2-free deionized (DI) water (18.2 MΩ) inside an anaerobic glovebox (90 – 95 % Ar/5 – 10 % H2). The H2 in the glove box was only present to work along with a commercially available catalyst in the glovebox to remove trace amounts of O2 (as H2O product). The gaseous H2 present in the glovebox was not thought to participate in the chemistry evaluated in this work due to its inherently low solubility in water (~1.5 mg H2/kg H2O) and since all the solutions were sparged with ultrahigh purity argon. The Fe2+ containing solution was then injected into the sulfide solution, resulting in the rapid precipitation of FeS. The resulting black precipitate was centrifuged, the supernatant discarded and replaced with 1 mM Na2S.9H2O solution. This “washing” process was repeated three times and was designed to remove residual SO4 2−and Na+ along with any other dissolved impurities. Prior research using high energy total X-ray scattering has shown that the resulting material is the mackinawite phase with particle sizes of roughly 3–5 nm (Michel et al. 2005). The particle loading was determined using a Pyrex® Guy-Lussac bottle and confirmed by desiccating an aliquot of known volume/weight of the slurry under anaerobic conditions. The FeS slurry was stored in a humidity-controlled anaerobic glovebox with constant O2 monitoring (Coy Laboratory Products) and it was used as necessary in batch experiments. The slurry was stored in a protective 1 mM Na2S solution in order to both prevent any oxidation of the FeS due to long-term storage and to ensure any Fe2+ (aq) will be instantaneously converted back to FeS thus eliminating a possible competing NO2 − and NO3 − reduction pathway by ferrous iron. Stock NO2 − and NO3 − solutions were prepared by dissolving reagent-grade NaNO2 or KNO3 in Ar-sparged, O2-free DI water (18.2 MΩ). The DI water used for making solutions was obtained from an EasyPure water treatment system and it was sparged with high purity argon gas for 2 h to remove any dissolved gases. All sample handling was conducted only under anaerobic conditions inside the glove box.

For the ATR-FTIR experiments, freshly precipitated FeS was generated within a glove bag, which was equipped with an O2 monitor (< 0.1 ppm). The FeS was generated using the same methodology as described above, after which 2 mL of the FeS slurry was combined with 2 mL of 10 mM NO2 − or NO3 − stock solutions yielding a 5 mM solution, in NO2 − or NO3 −, and FeS particle loadings of 7.5 g/L. These slurries were prepared and then used immediately for all the particular experiments conducted in this study.

Batch Experiments

NO2 −/FeS or NO3 −/FeS slurries were individually loaded into 6 mL PEEK and Teflon® -lined stainless steel reactors inside an anaerobic glovebox. Loaded and sealed reactors were then placed in a temperature controlled oven on an orbital shaker and agitated at ~150 rpm for the duration of an experiment. Experiments were conducted at 22, 70, and 120 °C. Individual reactors were withdrawn at desired time intervals, filtered using a 10,000 and 3,500 Dalton filter and either immediately analyzed or frozen until analysis. Background levels and NH4 + production were determined in a series of blank experiments at 70 and 120 °C to assess any reactivity of the reaction vessel and/or thermal degradation, as well as release of NH3 from the FeS used in the experiment. These blank experiments are important for the interpretation of the results given that ammonia is a common contaminant in reagents and starting materials (Smirnov et al. 2008; Schoonen and Xu 2001). The initial solutions had a pH of 10.3 and the final pH values of each experiment are shown in Table 1.

Table 1 pH values are shown for the ammonia producing experiments in this investigation
Full size table

Batch Reaction Measurements

Both NO3 − and NO2 − were analyzed on a Dionex DX500 ion chromatograph with a 4 mm IonPac® AS4A-SC anion column and a 1.7/1.8 mM NaHCO3/Na2CO3 eluent. Concentrations were calculated from a linear 4-point calibration curve with R2 values greater than 0.99. NH4 + was analyzed using an adapted fluorescence method (Amornthammarong and Zhang 2008) on a Barnstead Quantech FM109515 fluorometer with 360 and 420 nm narrow band excitation and emission filters, respectively. Samples and standards with individual volumes of 175 μL were pipetted into polypropylene vials containing 3 mL of working reagent and incubated in the dark for 24–48 h before analysis. Samples were acidified to pH 6 prior to analysis to convert NH3 to NH4 + for analysis. The concentrations of NH4 + were calculated from a linear 9-point calibration curve with R2 values greater than 0.99. Error bars in Figs. 1, 2, 3 are calculated by taking three readings for each point, averaging, calculating standard deviation and then the % error. The three readings were always very close to each other, thus the standard deviations were very small. Due to experimental complications with FeS, it was not possible to reliably analyze for sulfate concentrations.

Fig. 1
figure 1

NH4 + yields resulting from NO2 − and NO3 − reduction on FeS as a function of time at 22, 70, and 120 °C. Top: results of batch reaction measurements of NH4 + generated by FeS and 5 mM NO2 − at 22 °C (▲), 70 °C (○), and 120 °C (●) as a function of reaction time. The 120 °C blank is shown for comparison (■). Bottom: results of batch reaction measurements of NH4 + generated by FeS and 5 mM NO3 − at 22 °C (■), 70 °C (▲), and 120 °C (●) as a function of reaction time. The 120 °C blank is shown for comparison (○)

Full size image
Fig. 2
figure 2

NO2 − (●) and NO3 − (○) concentrations obtained from geochemical batch reaction measurements as a function of time conducted at 22 °C (top), 70 °C (middle), and 120 °C (bottom)

Full size image
Fig. 3
figure 3

Results of batch reaction measurements starting with NO3 − at 22 °C (■), 70 °C (▲), and 120 °C (●) as a function of reaction time

Full size image

ATR-FTIR Experiments

Two different ATR accessories were utilized, depending on the reaction temperature used for a particular experiment. The 22 and 70 °C experiments were conducted in a multibounce ATR accessory (Pike Technologies) with a Ge element, whereas the 120 °C experiments utilized a diamond single bounce ATR accessory (Specac) which was pressurized with 50 bar of ultra-high-purity argon to maintain the presence of liquid H2O which was required for experimental considerations. There was no added pressure to the 22 °C and 70 °C experiments. The accessories were individually mounted on a Nicolet Magna 750 FTIR spectrometer equipped with a liquid N2 cooled mercury cadmium telluride A (MCTA) detector. In each case data was collected with a resolution of 4 cm−1 (2 min/spectrum). During an experiment a spectrum was taken as soon as the reaction cell reached the desired experimental temperature (time 0) and this spectrum was used as a reference spectrum for all subsequent spectra.

Results

The results section is divided into two sections which describe results obtained from geochemical batch reactions and in situ ATR-FTIR experiments, respectively, on FeS in 5 mM NO2 − or NO3 − slurries at reaction temperatures of 22, 70, and 120 °C.

Batch reaction results

Results from aqueous geochemical batch reaction studies are shown in Figs. 1, 2, 3. NH4 + production resulting from FeS reacting with 5 mM NO2 − and NO3 − at reaction temperatures of 22, 70, and 120 °C is shown in Fig. 1 (top and bottom, respectively). The concentration of NH4 + generated by the 120 °C NO2 −/FeS system after 5 h of reaction time is more than 200 μmol/kg, while the 70 °C NO2 −/FeS and 22 °C NO2 −/FeS experiments generated ~14 and ~3 μmol/kg of NH4 +, respectively. The latter amount associated with the 22 °C reaction temperature is equivalent to background levels generated in blank experiments where no NO2 − is present. After 30 h nearly 1,200 μmol/kg of NH4 + was generated by NO2 −/FeS at a reaction temperature of 120 °C and lower yields were generated in the 70 °C NO2 −/FeS (85 μmol/kg) and 22 °C NO2 −/FeS (~5 μmol/kg) systems.

Compared to the data presented in the top panel of Fig. 1, data presented in the bottom panel shows that the NH4 + yields with NO3 − as the reactant are significantly less than yields in experiments with NO2 − as the reactant. For example, at a temperature of 22 and 70 °C the reaction of NO3 − with FeS yielded ~3 and ~5 μmol/kg of NH4 + after 30 h of reaction time, respectively. The same reaction at 120 °C yielded ~30 μmol/kg of NH4 + after 30 h (compared to 1,200 μmol/kg when NO2 − is the reactant).

Figure 2 exhibits the aqueous NO2 − concentration as a function of time resulting from the reaction of FeS with 5 mM NO2 − at 22 °C (top), 70 °C (middle), and 120 °C (bottom). These data are associated with the same experimental runs as the NH4 + measurements that were presented in Fig. 1. Based on Fig. 2, the 120 °C NO2 −/FeS experiment shows less NO2 − reactant remaining (1.2 mmol/kg) than in the 70 °C NO2 −/FeS experiment (3.5 mmol/kg) after 30 h. The 22 °C NO2 −/FeS experiment shows little change in NO2 − concentration over the course of the experiment. It is mentioned that there is no evidence for the presence of NO3 − in solution in the NO2 − reactant experiments. Overall, the trend of NO2 − consumption increasing with temperature is consistent with the yield of NH4 + product increasing with reaction temperature.

Figure 3 shows the NO3 − concentration as a function of time for the 22, 70, and 120 °C NO3 −/FeS experiments. There is no discernible change in the NO3 − concentration for any of the reactions. This experimental observation is not unexpected considering that the NH4 + yield from the reaction of NO3 − in the presence of FeS was limited at any of the reaction temperatures investigated.

ATR-FTIR Results

Vibrational mode assignments in this investigation were made both in view of prior work on other iron-sulfur surfaces (Hadjiivanov 2000; Singireddy et al. 2012) and on the basis of NH4 + yields observed in geochemical batch reaction measurements discussed in the previous section. These assignments are summarized and compared to prior work in Table 2.

Table 2 Vibrational assignments of NOx species associated with NO2 − and NO3 − reduction on FeS are summarized and compared to those in prior work on pyrite (FeS2)
Full size table

FeS Exposed to 5 mM NO2 −

ATR-FTIR spectra associated with freshly precipitated FeS that was exposed to 5 mM NO2 − at a temperature of 22 °C for reaction times of 0.08, 0.5, 1, 2 and 3 h are shown in Fig. 4 (b–f). A 22 °C control spectrum (a) in which FeS was exposed to a NO2 −/NO3 − free solution after 3 h is included for comparison. The two modes centered at 1747 and 1777 cm-1 are assigned to the υas(NO)2 and υs(NO)2 modes, respectively, of the dinitrosyl group adsorbed on FeS (most prominent in the 2 and 3 h spectra).

Fig. 4
figure 4

ATR-FTIR spectra of FeS exposed to 5 mM NO2 − at 22 °C as a function of time: 0.08 h (b), 0.5 h (c), 1 h (d), 2 h (e), and 3 h (f). Also included is a blank experiment (a) conducted under the same conditions after 3 h, but without added NO2 −. Note that spectra are offset for clarity

Full size image

Figure 5 shows ATR-FTIR spectra associated with FeS that was exposed to 5 mM NO2 − at 70 °C for 0.08, 0.17, 0.5, 2, and 3 h (b–f). A 70 °C control spectrum (a) in which FeS was exposed to a NO2 −/NO3 − free solution after 3 h is included for comparison, as is a spectrum (g) obtained after 3 h of a 5 mM NO3 − exposure. In the NO2 − data (b–f), the N-O stretching region exhibits three distinct NO stretching modes at 1777, 1747, and 1715 cm−1 that are assigned to the υs(NO)2 and υas(NO)2 of the dinitrosyl group on FeS and the υ(NO) stretch of a nitrosyl group on FeS, respectively. All three surface-bound N-O modes show a gain in intensity over the course of the 3 h run, although there are changes in the relative intensities of these modes. Specifically, spectra observed after 0.08 (b) and 0.17 h (c) show an increasing intensity of the band at 1715 cm−1. At longer times, however, the bands at 1777 and 1747 cm−1 show a greater growth in spectral intensity over the band at 1715 cm−1. Additional modes are observed at 1488 and 1270 cm−1 that are attributed to adsorbed NO3 − species. These modes show little growth over the 3 h run. Another weak mode is observed at 1120 cm−1 and is attributed to adsorbed sulfate on the FeS surface. This mode shows small increases in intensity over the course of the experiment.

Fig. 5
figure 5

ATR-FTIR spectra of FeS exposed to 5 mM NO2 − at 70 °C as a function of time: 0.08 h (b), 0.17 h (c), 0.5 h (d), 2 h (e), and 3 h (f). Spectra associated with NO3 − under the same conditions after 3 h (g) and a blank experiment conducted under the same conditions but without NO2 − and NO3 − (a) are shown for comparison. Note that spectra are offset for clarity

Full size image

Figure 6 exhibits ATR-FTIR spectra of FeS exposed to 5 mM NO2 − after 0.5, 1 and 3 h of reaction time at 120 °C (b–d). Also included is a 120 °C control spectrum (a) in which FeS was exposed to a NO2 −/NO3 − free solution after 3 h and a spectrum (e) obtained after 3 h of a 5 mM NO3 − exposure. In the NO2 − data, a broadened υ(NO) mode appears at 1806 cm−1 after a reaction time of 0.5 h and shows an increasing intensity at least up to a reaction time of 3 h. Consistent with the appearance of the υ(NO) are modes positioned at 1225, 1155, and 1105 cm-1 that are assigned to the S-O vibrations associated with sulfate adsorbed on FeS. Two pairs of modes are observed at 1407/1270 cm−1 and 1510/1309 cm−1 which are attributed to the υas(NO2)/υs(NO2) modes of surface bound NO2 − and the υ(NO3) modes of NO3 − species on the FeS surface, respectively.

Fig. 6
figure 6

ATR-FTIR spectra of FeS exposed to 5 mM NO2 − at 120 °C as a function of time: 0.5 h (b), 1 h (c), and 3 h (d). Also included is a blank experiment (a) conducted under the same conditions after 3 h, but without added NO2 − or NO3 − and a spectrum of FeS exposed to 5 mM NO3 − at 120 °C after 3 h (e). Note that spectra are offset for clarity

Full size image

FeS Exposed to 5 mM NO3 −

An ATR-FTIR spectrum (g) of FeS exposed to 5 mM NO3 − after a reaction time of 3 h at 70 °C is shown in Fig. 5. A vibrational mode having weak intensity at 1747 cm−1 is observed throughout the 3 h experimental run and is assigned to the υ(NO)as mode of adsorbed dinitrosyl species on FeS. Similar modes having much greater intensity were experimentally observed when NO2 − was the reactant, suggesting that the activation of NO3 − to form surface bound species is an energetically more difficult process.

Figure 6 shows the ATR-FTIR spectrum (e) associated with FeS that was exposed to 5 mM NO3 − for 3 h of reaction time at 120 °C. The υ(NO) mode of N-O bound to the FeS surface is observed at 1758 cm−1. This mode shows only a minor increase in intensity after a reaction time of 3 h. A pair of modes associated with adsorbed NO2 − are observed at 1407/1270 cm−1 and a second pair of modes attributed to NO3 − coordinated to the FeS surface are observed at 1510/1309 cm−1. Additional modes are also observed at 1225, 1155, and 1105 cm−1 which are attributed to S-O vibrations of adsorbed sulfate.

Discussion

There are three experimental observations presented above that warrant further discussion. First, the reduction of NO2 − on FeS in alkaline conditions will be discussed, which will highlight the temperature dependence of the reaction and the resulting implications for the mechanism of NH4 + production. Second, the results show that the activation of NO3 − and the production of NH4 + only occur at a temperature of 120 °C, while NH4 + production from NO2 − occurs at a reaction temperature at least as low as 70 °C. ATR-FTIR results will be discussed that lend insight into the reasons for this difference in reactivity. Finally, results obtained from this investigation will be discussed in view of relevant prior research.

Temperature Dependent NH4 + Production from NO2 − and FeS

Results from ATR-FTIR lend insight into the experimental observations associated with the batch reactions. With regard to NH4 + production from the reaction of NO2 − and FeS, little if any product formation occurs at a reaction temperature of 22 °C, while a significant amount of NH4 + production occurs at reaction temperatures of 70 and 120 °C. Analysis of the ATR-FTIR spectra shows both similarities and differences in surface speciation of the FeS at the three reaction temperatures that are likely relevant to NH4 + production. Comparison of the 22 and 70 °C spectra suggest that both surfaces are populated by dinitrosyl species, which we associate with the 1747 cm−1 (υas (NO)2) and 1777 cm−1 (υs(NO)2) vibrational modes. A further comparison shows that based on the intensity of these modes, the population of surface bound dinitrosyl is greater at the higher temperature, consistent with the surface mediated activation of NO2 − adsorption and NH4 + formation occurring more readily at the reaction temperature of 70 °C. Furthermore, there is a rather remarkable increase in surface bound NO modes on FeS compared to the 22 °C NO2 −/FeS experiment. At the same time, significant NH4 + production is experimentally observed (Fig. 1, top) in parallel with a concurrent loss of NO2 − from solution (Fig. 2, middle). In addition to the same dinitrosyl modes observed in the 22 °C experiments at 1747 cm−1 (υas(NO)2) and 1777 cm−1 (υs(NO)2), a third NO vibrational mode, the υ(NO) of (mono)nitrosyl, is observed at 1715 cm−1 and becomes apparent within the first 0.08 h of reaction. The mode at 1777 cm−1 is present after 2 h of reaction time. We believe that a possible reason for the absence of the υs(NO)2 in the first two hours is a change in the NO-Fe-NO angle on the FeS surface with increasing reaction time. As this particular angle of the dinitrosyl complex approaches 180°, the integral peak area of the υs(NO)2 is greatly decreased relative to the υas(NO)2, based on infrared surface selection rule arguments (Hadjiivanov 2000; Braterman 1975). It is noted that after 2 h, the relative intensity of these bands are similar in the 22 °C and 70 °C NO2 − experiments, which we suspect is due to the NO-Fe-NO angle becoming more acute as reaction time proceeds. The change in the NO-Fe-NO angle likely results to accommodate the increasing coverage of NO adsorbate as the reaction proceeds. The absence of modes attributable to adsorbed NO2 − and the overall low intensity of the dinitrosyl species on the FeS surface at 22 °C is consistent with the batch experiment results that show little NH4 + production at this temperature (Fig. 4, and Fig. 1, top).

At a reaction temperature of 70 °C there is a rather remarkable increase in surface bound NO modes on FeS compared to the 22 °C NO2 −/FeS experiment. At the same time, significant NH4 + production is experimentally observed (Fig. 1, top) in parallel with a concurrent loss of NO2 − from solution (Fig. 2, middle). In addition to the same dinitrosyl modes that were observed in the 22 °C experiments at 1747 cm−1 (υas(NO)2) and 1777 cm−1 (υs(NO)2), a third NO vibrational mode, the υ(NO) of (mono)nitrosyl, is observed at 1715 cm−1 and this mode becomes apparent within the first 0.08 h of reaction.

A reaction temperature of 120 °C for the NO2 −/FeS system exhibits the greatest NH4 + production of any experiment in this set (Fig. 1, top). Increasing the temperature to 120 °C increases the NH4 + yield greatly over the 70 °C case (1.2 vs. 0.04 mM after 30 h of reaction time). Based on the change in rate with temperature, the activation energy for NH4 + production on FeS from NO2 − is estimated to be ~15 kcal/mol. This activation energy barrier is significantly lower than the activation barrier associated with the homogeneous reduction of NO2 − by ferrous iron. For example, an estimation of the activation barrier associated with the reduction of NO2 − by aqueous Fe2+ (aq) that was investigated previously by Summers et al. (Summers and Chang 1993) yields a value of ~38 kcal/mol (based on measured rates between 0 and 40 °C using an Arrhenius-type treatment). Hence, the FeS surface provides a lower reaction barrier pathway for the production of NH4 + than does the homogeneous reaction.

In addition to the more significant NH4 + yields observed in the 120 °C case, the ATR-FTIR spectra show changes from the 70 °C case as well. Perhaps the most important aspect of the spectra is the single υ(NO) mode at 1806 cm−1 that we attribute to NO bound to an Fe3+ site. The position of this NO vibrational mode suggests that the back-bonding mechanism typically associated with the binding of NO on other electron-rich iron sulfides is no longer the primary bonding mode. Instead, the electron-deficient Fe3+ site likely coordinates to the 5σ antibonding orbital electrons of NO. This bonding configuration results in a single mononitrosyl species, with an υ(NO) mode at 1806 cm−1 which is blue-shifted relative to the υ(NO) in the 70 °C NO2 −/FeS experiment (1715 cm−1). Based on our results, we propose that NH3 formation can be represented by the composite Eqs. 1 and 2.

$$ \mathrm{FeS}+5{\mathrm{H}}_2\mathrm{O}+9{\mathrm{NO}}_2{}^{-}\to {\mathrm{Fe}}^{3+}+{\mathrm{SO}}_4{}^{2-}+9\mathrm{NO}+10\kern0.5em {\mathrm{OH}}^{-} $$
(1)
$$ 16{\ \mathrm{H}}_2\mathrm{O}+9\kern0.5em \mathrm{NO}+5\mathrm{FeS}\to 9{\mathrm{NH}}_3+5\kern0.5em {\mathrm{Fe}}^{3+}+5\kern0.5em {\mathrm{SO}}_4{}^{2-}+5{\mathrm{OH}}^{-} $$
(2)

Additional pairs of modes are observed at 1510/1309 cm−1 and 1407/1270 cm−1 and are attributed to surface bound NO3 − and NO2 − adsorbates. The assignment of modes in this region is dubious at best, considering that there is significant overlap of modes associated with NO2 − and NO3 − in multiple binding configurations (e.g., monodentate and bidentate geometries) on multiple binding sites (Hadjiivanov 2000).

Modes observed at 1510/1309 cm−1 are assigned to υ(NO3) of adsorbed NO3 − (nitrato). We surmise that the bonding geometry is either monodentate or bidentate based on the splitting of the υ3 mode of aqueous NO2 − upon adsorption on the FeS surface. The latter pair of modes observed at 1407/1270 cm−1 is assigned to the nitro bonding configuration of NO2 − . Based on the observed positions, it is possible that these modes correspond to monodentate nitrito species as well, however, based on prior investigations it was determined that the nitro configuration was likely the preferred configuration for NH4 + formation (Singireddy et al. 2012).

We speculate that the nitro species actively participate in NH4 + formation, which proceeds through a surface bound nitrosyl intermediate. The increased rate of NO2 − reduction observed throughout the reaction at 120 °C and the presence of the 1407/1270 cm−1 pair suggests that chemistry is occurring that is unique to a reaction temperature of 120 °C. Specifically, a temperature dependent adsorption of NO2 −on the FeS surface is observed at 120 °C given that there is little evidence for the presence of modes assigned to nitro species (i.e., at 1407/1270 cm−1) present in the 70 °C spectra, or the 22 °C spectra, which adds support to this temperature dependent activation of NO2 −.

Similarly, the υ(NO) mode intensity displays a dependence on temperature. ATR-FTIR spectra show little mode intensity ascribed to N-O species at 22 °C. Increasing the reaction temperature to 70 °C, however, results in a dramatic increase in intensity, which suggests that there is also a temperature dependent conversion of NO2 − to nitrosyl. We suspect that this surface bound NO is a relevant surface intermediate in the production of NH4 +. Though it is difficult to ascertain the binding configurations of both the NO2 − and NO3 − adsorbate observed in the 120 °C experiments, there is evidence of a disproportionation reaction

$$ {2\mathrm{NO}}_2{}^{-}\to \mathrm{NO}+{\mathrm{NO}}_3{}^{-}+{\mathrm{e}}^{-} $$
(3)

since modes experimentally observed at 1510/1309 and 1806 cm−1 are consistent with adsorbed NO3 − and NO, respectively, on the FeS surface. Unfortunately, it is not entirely clear how much of this disproportionation reaction results in the surface-bound NO species. It is mentioned that N2 production is a likely reaction channel, based on prior investigations on pyrite (Singireddy et al. 2012). N2 production is, however, not detectable using ATR-FTIR spectroscopy.

Equations (1) and (2) suggest that sulfate is generated as a product in all three reaction schemes. We suspect that the majority of sulfate product partitions to solution, given that we experimentally observe relatively low intensity vibrational modes on FeS at 1225, 1155, and 1105 cm−1, that can be attributed to surface bound sulfate. These assignments are consistent with sulfate assignments on pyrite in prior publications (Singireddy et al. 2012; Usher et al. 2005, 2004).

It is also noted that Fe3+ appears in Eqs. (1) and (2) which is the result of oxidation by NO2 − and appears to show a significant dependence on temperature. The position of the NO stretches at 70 °C and 22 °C are assigned to NO bound to Fe2+ sites, and at 120 °C the dominant modes are associated with Fe3+-NO. This temperature dependent shift is either due to the increased availability of Fe3+ sites, a possibility that would be consistent with the higher rate of ammonia production, or the Fe3+ binding site is thermally activated at 120 °C (or a combination of both). It is also noted that there is a decrease in spectral weight near 1750 cm−1 with reaction time in the 120 °C NO2 − experiment (see Fig. 6). This spectral region is associated with the Fe2+-NO stretch and suggests that at 120 °C, Fe2+-NO complexes initially form, but as the reaction proceeds Fe2+ sites are oxidized to Fe3+ and spectral weight appears that is associated with Fe3+-NO. This effect is not observed at 22 °C or 70 °C, but it is not readily obvious if this is because of a lack of Fe3+ sites available due to the lower temperatures (and lower ammonia yield), or due to the presence of a thermal activation barrier.

We mention that reactions (1) and (2) result in hydroxide ions being released into solution. However, the results of pH measurements taken before and after the reaction show little change in the final pH values (Table 1). We might speculate that the lack of observed pH increase is due to secondary reactions that include the production of iron oxides and oxyhydroxides which may result in the offset of the expected pH increase. Thus Fe3+ and OH− in reactions (1) and (2) represent ammonia forming steps and are not thought to be the final states of these particular species. Solid state analysis is not easily accomplished with FeS materials due to its rapid oxidation when dried or exposed to air, thus unfortunately we cannot offer direct solid state experimental observation to support this claim.

Activation of NO3 − at 120 °C

Prior studies have shown that the conversion of NO3 − to NH4 + is kinetically hindered on pyrite, FeS, and in the presence of Fe2+ (aq) at temperatures near 25 °C (Summers and Chang 1993; Summers 2005; Singireddy et al. 2012). Pyrite was shown in prior work, however, to facilitate NH4 + from NO3 − at a reaction temperature of 120 °C (Singireddy et al. 2012). The prior investigation of the NO3 −/FeS2 system agrees well with the 120 °C NO3 −/FeS system in this study that shows a small, but significant amount of NH4 + production. Perhaps the most distinct feature of the NO3 −/FeS experiment is that at 120 °C, the υ(NO) mode becomes observable on FeS. We interpret this experimental observation to indicate that reduction of NO3 − has occurred at this elevated temperature, resulting in the partial reduction of NO3 − to surface bound NO on FeS (Fig. 6). Only a very small amount of spectral intensity attributed to the υ(NO) mode is observed when FeS is exposed to NO3 − at a reaction temperature of 70 °C which suggests that activation of NO3 − has not occurred to any significant degree. We propose that the pathway to this NO species likely proceeds through surface bound NO3 − and NO2 − intermediate, based on modes at 1510/1309 and 1407/1270 cm−1 that we attribute to surface bound NO3 − (nitrato) and NO2 − (nitro), respectively, in the 120 °C NO3 −/FeS system. Calculations published in an earlier investigation (Singireddy et al. 2012) gives some insight into why NO3 − is more difficult to reduce than NO2 − and would presumably need higher temperatures to activate. Interestingly, calculations suggest that while NO2 − binds to iron via its nitrogen atom (i.e. nitro), NO3 − adopts a bonding configuration that has its oxygen atoms binding to the iron component of FeS. The result of these binding configurations is that the N-O stretching frequency associated with NO2 − is red-shifted upon adsorption while the N-O stretching frequency associated with NO3 − is blue-shifted upon adsorption. In short, the N-O bonds of NO3 − become more strongly bound to FeS compared to the aqueous species while the N-O bonds of NO2 − become weaker on FeS when compared to the aqueous species. These bond strength arguments and the absence of binding via the nitrogen atom of NO3 − to FeS are consistent with the experimental observation that NO3 − is more difficult to reduce than NO2 −.

Additional vibrational modes associated with adsorbed sulfate are observed at 1225, 1155, and 1105 cm−1 when FeS is exposed to NO3 − at 120 °C. Prior vibrational spectroscopy based work on pyrite surfaces under oxidizing conditions show the evolution of similar modes. In the pyrite circumstance it is thought that electron withdrawal from the sulfur component is followed by the nucleophilic attack of water to form sulfate (Usher et al. 2004). In our circumstance we suspect that at least a fraction of the sulfate product is due to the transfer of electrons from the sulfur component of FeS. These electrons likely go on to reduce a fraction of the adsorbed NO3 − to NO2 − and then to NO species. Such a scenario would be consistent with the following composite reactions:

$$ 2\mathrm{FeS}+9{\mathrm{NO}}_3{}^{-}+{\mathrm{H}}_2\mathrm{O}\to 2{\mathrm{Fe}}^{3+}+9{\mathrm{NO}}_2{}^{-}+2{\mathrm{SO}}_4{}^{2-}+2{\mathrm{OH}}^{-} $$
(4)
$$ \mathrm{FeS}+5{\mathrm{H}}_2\mathrm{O}+9{\mathrm{NO}}_2{}^{-}\to {\mathrm{Fe}}^{3+}+{\mathrm{SO}}_4{}^{2-}+9\mathrm{NO}+10\kern0.5em {\mathrm{OH}}^{-} $$
(5)

Significance of Results in View of Prior Work

NH4 + production via NO2 − and NO3 − reduction on pyrite (FeS2) has been investigated previously in our laboratory using a methodology that was similar to that used in the present study (Singireddy et al. 2012). NO2 − and NO3 − reduction has also been studied previously on FeS (Blöchl et al. 1992; Summers 2005) and pyrrhotite (Brandes et al. 1998). These prior investigations of NH3 formation have been carried out on the FeS at circumneutral or acidic pH conditions. In the context of prebiotic chemistry, alkaline conditions on the Hadean may have been more relevant. As mentioned before, Russell and coworkers, for example, postulated that hydrothermal vents served as natural incubators for prebiotic chemistry (Russell et al. 1988, 1989, 1993). Within these vent systems, a precipitated FeS phase served as a barrier between the acidic Hadean Ocean with the alkaline fluids which were rising from within the Earth. Therefore, the investigation presented is most relevant to NH3 formation within the alkaline boundary of the hydrothermal vent system framework presented in previous work (Russell et al. 1988, 1989, 1993).

It has been proposed that over the course of about 106 years that μM concentrations of nitrogen oxides could have built up within the Hadean Oceans (Ducluzeau et al. 2009). Research presented here shows that if an environment existed as suggested by Russell during the Hadean, the basic conditions proposed would have supported NH3 formation, with the caveat that high enough temperatures were present within the vents systems to drive the reactions on FeS.

Building on the reaction chemistry of NO2 − and NO3 − on FeS reported on in this contribution, there are a number of other processes that must be considered which may have important implications for reactions that produce ammonia and other raw materials needed for life. Reactions resulting in Fe3+ production must also be considered, as well as H2S/S2− availability, since these latter species influence both the pH, Fe3+ concentration, and perhaps even the reaction mechanism(s) involved in ammonia production. In order to create a more complete model, future work should involve investigating these parameters and how each affects ammonia production. Additionally, non-equilibrium reactions (i.e. continuous introduction of reactants and removal of waste product and unused reactant materials) must also be investigated to create a more comprehensive picture of ammonia production in realistic prebiotic scenarios.