Report

A bioinspired iron catalyst for nitrate and perchlorate reduction

See allHide authors and affiliations

Science  11 Nov 2016:
Vol. 354, Issue 6313, pp. 741-743
DOI: 10.1126/science.aah6886

Biological inspiration for reduction

Microorganisms have evolved sophisticated enzymatic machinery to reduce perchlorate and nitrate ions. Although the energetics of the pathways are different, the heme-containing active sites of the corresponding reductase enzymes are remarkably similar. Ford et al. constructed an inorganic catalyst to mediate these reactions based on these active sites, using a nonheme iron complex. A secondary coordination sphere near the iron center aligned the nitrate or perchlorate oxyanions and formed an iron-oxo complex. Regenerating the catalyst in the presence of protons and electrons released water—a potentially much more sustainable process than reduction strategies that require the use of harsh reagents.

Science, this issue p. 741

Abstract

Nitrate and perchlorate have considerable use in technology, synthetic materials, and agriculture; as a result, they have become pervasive water pollutants. Industrial strategies to chemically reduce these oxyanions often require the use of harsh conditions, but microorganisms can efficiently reduce them enzymatically. We developed an iron catalyst inspired by the active sites of nitrate reductase and (per)chlorate reductase enzymes. The catalyst features a secondary coordination sphere that aids in oxyanion deoxygenation. Upon reduction of the oxyanions, an iron(III)-oxo is formed, which in the presence of protons and electrons regenerates the catalyst and releases water.

The most efficient reduction of nitrogen- and chlorine-containing oxyanions is achieved by the microbial metalloenzymes (per)chlorate reductase and nitrate reductase during anaerobic respiration (1, 2). The active sites of the two metalloenzymes are similar and allow each enzyme to reduce both nitrogen- and chlorine-containing oxyanions (Fig. 1) (1, 2). Both reductases also feature extensive hydrogen-bonding networks, which facilitate the movement of protons and water about the active site and stabilize reactive intermediates (13). Disruption of this network in nitrate reductase via mutagenesis results in the complete loss of activity because of the low stability of high-valent Mo=O intermediates (3, 4). Furthermore, positively charged residues near the active site in (per)chlorate reductase aid in the binding of perchlorate to the Mo-center (3). The noncovalent interactions found within the metalloenzymes play an important role in facilitating reactivity (111). Incorporating these interactions into transition-metal complexes may aid in oxyanion reactivity because most transition-metal systems are not capable of these reductions (1217).

Fig. 1 Biological oxyanion reduction.

The active sites of the enzymes responsible for the reduction of chlorine- and nitrogen-containing oxyanions are depicted.

The challenge in reducing these oxyanions lies not only in their unfavorable reduction potentials but also in their low binding affinity to transition-metal centers (12). These inorganic oxyanions have long been touted for their weak complexation, poor nucleophilicity, and consequently their kinetic inertness toward oxidation and reduction (12). Therefore, harsh reaction conditions (such as low pH, high temperature, photolysis, and/or long reaction times) are required to facilitate oxyanion reduction in homogenous systems (1217).

Inspired by the active sites of nitrate and (per)chlorate reductase, we developed a nonheme platform that incorporates the following features: (i) an earth-abundant redox-active metal center (iron) and (ii) a secondary coordination sphere that facilitates deoxygenation of substrates and high-valent iron-oxo intermediates. Previously, we reported the synthesis and characterization of a family of late, first-row transition-metal complexes [N(afaCy)3MOTf]OTf (M = Mn, Fe, and Co), featuring the dative (azafulvene-amine) coordination mode of the ligand and the presence of the amino-derived secondary coordination sphere (1822). The ability of the ligand to undergo tautomerization may be a key feature during multi-electron reactions because it can facilitate proton and electron transfer between the substrate and the metal center. Furthermore, the secondary coordination sphere orients substrates binding to the metal center, as demonstrated in our nitrite reduction studies, in which a single hydrogen bond stabilized a key metal-nitrito intermediate (21, 22).

The addition of tetrabutylammonium nitrite ([NBu4][NO2]) to 2 equivalents (eq) of [N(afaCy)3FeOTf]OTf (FeII-OTf; OTf, trifluoromethanesulfonate) afforded the iron(III)-oxo complex [N(afaCy)3FeO]OTf (FeIII-O) and NO(g), which was trapped by FeII-OTf to furnish the iron(II)-nitrosyl species [N(afaCy)3FeNO]OTf2 (FeII-NO) (21). Given the facile one-electron reduction of nitrite by FeII-OTf, we sought to explore the reduction of nitrate. Hypothesizing that if the FeII-OTf could perform the two-electron reduction of nitrate to generate nitrite, the reduction of nitrite would then proceed as before to release NO(g), akin to the described reactivities of nitrate reductase and nitrite reductase, respectively. FeIII-O and FeII-NO formed when tetrabutylammonium nitrate, [NBu4][NO3], was added to 3 eq of FeII-OTf in the presence of triethylamine (NEt3) (fig. S3) (23). Two thirds of the isolated product consisted of FeIII-O, and one third was FeII-NO (67 and 28% isolated yield, respectively, based on the initial mass of FeII-OTf).

After the successful reduction of nitrate, we investigated the more challenging reduction of perchlorate (Fig. 2). Because of the low binding affinity of perchlorate to transition metals, our initial perchlorate reduction studies used iron(II)-perchlorate generated in situ via salt metathesis of FeCl2 and 2 eq of AgClO4. Ligand, H3N(piCy)3, was then added, resulting in an immediate color change from colorless to red-brown. Crystals suitable for x-ray diffraction (Fig. 2A) revealed an iron(III)-oxo with an outer-sphere perchlorate anion, [N(afaCy)3FeO]ClO4, [FeIIIO]-ClO4. The bond lengths of [FeIII-O]ClO4 were similar to those previously reported for FeIII-O, with an Fe–O distance of 1.8055(11) Å, compared with 1.8079(9) Å for FeIII-O (21). Furthermore, all three amino moieties are engaged in hydrogen-bonding to the oxygen atom.

Fig. 2 Solid-state molecular structures and stoichiometric chlorine oxyanion reduction.

Crystal structure of (A) [FeIII-O]ClO4, (B) ZnII- ClO4, and (C) FeII-Cl. Thermal ellipsoids are at the 50% probability level; solvent molecules and selected H-atoms have been omitted for clarity. (D) Stoichiometric chlorine oxyanion reduction.

In order to probe perchlorate binding to the metal center, the redox-inactive zinc perchlorate complex [N(afaCy)3ZnClO4]ClO4 (Zn-ClO4) was obtained from the addition of the ligand to Zn(ClO4)2. Solid-state structural characterization of the complex revealed a perchlorate anion bound to zinc with no hydrogen-bonding interactions from the amino moieties of the ligand (Fig. 2B). Instead, one arm engages in hydrogen-bonding to the outer-sphere perchlorate anion. Zn-ClO4 was robust and showed no perchlorate deoxygenation. These experiments confirmed that the secondary coordination sphere is not used in perchlorate coordination but rather facilitates the reduction of ClO4 through stabilization of the resulting metal-oxo center.

The labile ClO4 ligand was also transiently installed on iron from the reaction of tetrabutylammonium perchlorate ([NBu4][ClO4]) to 5 eq of FeII-OTf in the presence of NEt3. The 1H nuclear magnetic resonance (NMR) spectrum of the reaction revealed the formation of two products, one of which was identified as FeIII-O. We propose that the reaction proceeds via sequential oxygen atom transfer events from perchlorate to afford FeIII-O and the iron(II)-chloride compound [N(afaCy)3FeCl]OTf (FeII-Cl) in a 4:1 ratio. Independent synthesis of FeII-Cl confirmed this formulation. Crystallographic refinement revealed a trigonal bipyramidal geometry about the iron(II) center, with an axial chloride ligand and an outer-sphere triflate anion (Fig. 2C). Because of the low hydrogen-bonding affinity of chloride, the ligand does not hydrogen-bond to chloride and instead hydrogen-bonds to the outersphere triflate.

We propose that the reduction of perchlorate proceeds stepwise via sequential deoxygenations from perchlorate to chlorate, chlorite, and hypochlorite (Fig. 2D). To test this hypothesis, we investigated the reactivity of FeII-OTf toward the other chlorine oxyanions. In each case, we obtained mixtures of FeIII-O and FeII-Cl. In order to quantify the ratios of FeIII-O and FeII-Cl produced from the reduction of each chlorine oxyanion, we used the reaction FeIII-O with 1,2-diphenylhydrazine (DPH) (vide infra) to yield azobenzene. Because FeII-Cl did not react with DPH, the addition of DPH (in equimolar amounts based on FeII-OTf) to the chlorine oxyanion reduction mixture would yield a mixture of azobenzene and unreacted DPH whose ratio would match that of the oxygen and chlorine present in the starting oxyanion, assuming complete reduction. This method successfully demonstrated the ratios of oxygen and chlorine present in perchlorate, chlorate, and hypochlorite; the ratios of azobenzene to DPH quantified with 1H NMR spectroscopy were 4.36:1, 2.54:1, and 0.67:1, respectively (fig. S8).

We also investigated the reduction of FeIII-O to test the possibility of whether FeII-OTf could be regenerated with concomitant release of water; two protons and one electron were required for this transformation. For this purpose, DPH (a 2H+/2e source) (24) and Fc*OTf (as a sacrificial oxidant; Fc*OTf is decamethylferrocenium triflate) were added to FeIII-O, regenerating FeII-OTf (fig. S11) (23) in 74% isolated yield, concomitant with the formation of 0.89 eq of water, as assayed with Karl Fischer titration. Similarly, when FeIII-O was generated in situ from 0.5 eq of NaNO3 and reduced under identical conditions, 0.83 eq of water was formed.

Having shown that FeIII-O is cleanly regenerated with DPH/Fc*OTf, we examined the possibility that the oxyanion deoxygenation could be catalytic, beginning with nitrate. Using (TPP)Co (TPP, 5,10,15,20-tetraphenylporphyrin) to trap and subsequently quantify the NO(g) produced during the reaction, we tested the reactivity of 3 eq of NaNO3 [relative to (TPP)Co], 6 eq of DPH, and 3 eq of Fc*OTf as a control reaction for nitrate reduction. Quantifying by means of 1H NMR spectroscopy the amount of (TPP)CoNO formed revealed that NO(g) was produced under these conditions; 0.18 eq of (TPP)CoNO was detected after 27 hours, and 0.24 eq of (TPP)CoNO was detected after 42 hours [0.5 eq of (TPP)CoNO corresponds to TON of 1]. When FeII-OTf was added to the reaction mixture, over seven times the amount of NO(g) was trapped by (TPP)Co as compared with the control reaction; 1.46 eq of (TPP)CoNO was detected after 27 hours, and 1.74 eq of (TPP)CoNO was detected after 42 hours, resulting in a turnover number (TON) of 3.5 (Fig. 3).

Fig. 3 Catalytic nitrate and perchlorate reduction.

We hypothesize that the catalytic reduction of perchlorate would proceed similarly; however, because of the explosion hazards associated with perchlorate reagents, we limited the scale of the catalytic reaction to no more than 10 mg of [NBu4][ClO4]. FeII-OTf was mixed with [NBu4][ClO4] and 4 eq of DPH and stirred overnight. Analysis of the crude reaction mixture by means of 1H NMR spectroscopy revealed that 3.4 eq of the DPH was converted to azobenzene, providing an indirect measurement of the amount of perchlorate that was fully reduced. In the subsequent work-up, FeII-Cl was crystallized in 75% yield. This yield of FeII-Cl corresponds to a TON of 3 (Fig. 3). Although we were unable to determine whether the reaction proceeds by sequential two-electron reductions, from perchlorate to chlorate to chlorite to hypochlorite, the reaction proceeds through four steps because four oxygen atoms must be transferred from perchlorate to generate chloride. Moreover, the stoichiometric reaction of each chlorine oxyanion with FeII-OTf yielded the same two products, FeIII-O and FeII-Cl, in the expected ratios.

The uses of nitrogen- and chlorine-containing oxyanions are both extensive and varied, with notable applications in fertilizers, bleaching agents, propellants, and explosives (58). Because of their high solubility and mobility in water, they have become pervasive contaminants in many sources of drinking water (58). Remediation of polluted water by reduction of these oxyanions to benign products would thus have tremendous impact. The described catalytic deoxygenation of perchlorate and nitrate features mild reaction conditions. Our bioinspired iron catalyst is a first step toward a potentially more sustainable reduction strategy. Future improvements focusing on improved turnover numbers and a detailed mechanistic understanding will provide insights into catalyst design for future remediation efforts.

Supplementary Materials

www.sciencemag.org/content/354/6313/741/suppl/DC1

Materials and Methods

Figs. S1 to S14

Tables S1 to S8

References (2528)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: This work was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division under award DE-SC-0016026 and the University of Illinois at Urbana-Champaign. Y.P. was supported by the American Association for the Advancement of Science Milligan Mason Award for Women in the Chemical Sciences, awarded to A.R.F. We thank T. Betley, T. Rauchfuss, and K. Suslick for helpful discussions and S. Denmark and Y. Lu for use of their instruments. The crystallographic data CCDC-1510909-1510911 can be obtained free of charge from the Cambridge Crystallographic Data Centre (www.ccdc.cam.ac.uk/data_request/cif). Data are available in the supplementary materials.
View Abstract

Stay Connected to Science

Navigate This Article