Nanoscale Fe2O3-Based Catalysts for Selective Hydrogenation of Nitroarenes to Anilines

See allHide authors and affiliations

Science  29 Nov 2013:
Vol. 342, Issue 6162, pp. 1073-1076
DOI: 10.1126/science.1242005

Lighter Hydrogenation Catalysts

Enzymes have evolved to use abundant metals such as iron, cobalt, and nickel for redox catalysis. However, synthetic catalysis has generally relied on the rarer, heavier relatives of these elements: ruthenium, rhodium, iridium, palladium, and platinum (see the Perspective by Bullock). Friedfeld et al. (p. 1076) used high-throughput screening to show that the right cobalt precursor can be activated for asymmetric hydrogenation catalysis by using the traditional ligands developed for the precious metals. Zuo et al. (p. 1080) focused on iron, demonstrating a highly effective asymmetric transfer hydrogenation catalyst that uses a ligand rationally designed after careful mechanistic study. Jagadeesh et al. (p. 1073) prepared supported iron catalysts that selectively reduce nitro substituents on aromatic rings to amines, thereby facilitating the preparation of a wide range of aniline derivatives.


Production of anilines—key intermediates for the fine chemical, agrochemical, and pharmaceutical industries—relies on precious metal catalysts that selectively hydrogenate aryl nitro groups in the presence of other easily reducible functionalities. Herein, we report convenient and stable iron oxide (Fe2O3)–based catalysts as a more earth-abundant alternative for this transformation. Pyrolysis of iron-phenanthroline complexes on carbon furnishes a unique structure in which the active Fe2O3 particles are surrounded by a nitrogen-doped carbon layer. Highly selective hydrogenation of numerous structurally diverse nitroarenes (more than 80 examples) proceeded in good to excellent yield under industrially viable conditions.

Beginning in the 1950s, the development of organometallic catalysts proceeded to revolutionize organic synthesis at scales ranging from the laboratory bench to the industrial manufacture of fine and bulk chemicals. This success was mainly due to the use of noble metal complexes—for example, palladium, rhodium, ruthenium, and iridium (1). However, the high price and limited availability of these precious metals (2) have spurred interest in catalysis with more earth-abundant alternatives, especially iron (38). The durability and copious supply of iron salts coupled with their environmentally benign nature and low toxicity make them ideal catalysts. Recently, structurally well-defined molecular iron complexes have been applied successfully in contexts where previously noble metals were required (919). However, most of these homogeneous complexes are rather sensitive and/or incorporate sophisticated (and thus synthetically demanding) ligand systems. In contrast, heterogeneous iron oxides are extremely stable and can be easily recycled. Important applications of heterogeneous iron catalysts include the production of olefins through the Fischer-Tropsch process (20, 21) and the hydrogenation of CO (22). Unfortunately, these known catalysts work under drastic conditions (>275°C) and are therefore of limited use for the refinement of more complex substrates. In this context, we report special iron oxide–based catalysts that allow for a general and highly selective hydrogenation of nitroarenes under comparably mild conditions. The resulting anilines constitute key building blocks for the synthesis of fine (agrochemicals, dyes, pigments, and pharmaceuticals) as well as bulk chemicals (polymers) (23, 24). In particular, anilines used in life science applications are often structurally complex molecules decorated with diverse functional groups. Thus, achieving high chemoselectivity is of crucial importance for any catalyst development. Among the different known reductions of nitroarenes, catalytic hydrogenation represents the most benign and cost-efficient route (2531). Until today, such hydrogenations have been mainly carried out using noble metal catalysts. With respect to chemoselectivity, important advancements have been reported by Corma and co-workers using gold catalysts (30) and recently by our group applying Co3O4-based materials (31).

To improve the hydrogenation activity of bulk iron oxides, we explored the synthesis of nanoscale iron oxides supported on nitrogen-doped surfaces. For this purpose, iron nitrogen complexes were prepared by mixing iron(II) acetate in ethanol at room temperature with 1,10-phenanthroline (phen) (for other nitrogen ligands, see the supplementary materials). The resulting metal complex was deposited onto carbon (Vulcan XC72R), titanium dioxide (P25), or aluminum oxide at 60°C followed by pyrolysis at higher temperatures under inert gas atmosphere (for details, see the supplementary materials). Hereafter, the carbon-supported catalysts are labeled as Fe-phen/C-x, where x denotes the pyrolysis temperature.

All the prepared materials were tested for their hydrogenation activity toward the industrially important substrate nitrobenzene in a water-tetrahydrofuran (THF) solvent mixture. Parameters such as pyrolysis temperature, pyrolysis time, type of support, nitrogen ligands, and Fe:ligand molar ratios were systematically investigated (table S1). Neither the homogeneous iron complexes nor supported Fe(OAc)2 without any ligand were active (table S1, entries 1 to 6). Similarly, the material containing supported Fe(OAc)2-phenanthroline without pyrolysis was also not active (table S1, entry 7). However, after pyrolysis the resulting catalysts showed different activity (table S1, entries 8 to 12). On increasing the pyrolysis temperature from 200°C to 800°C, the activity of the resulting catalyst steadily increased. The material pyrolyzed at 800°C (Fe-phen/C-800) showed maximum activity: The benchmark substrate nitrobenzene was hydrogenated to give aniline in excellent yield (98%) (table S1, entry 11). Upon further raising the pyrolysis temperature to 1000°C, the activity of the resulting catalyst decreased (table S1, entry 12). Pyrolysis of the phenanthroline-based iron complexes on TiO2 and Al2O3 supports also gave active catalyst materials but produced aniline only in lower yields of 49 to 78% (table S1, entries 13 and 14).

To demonstrate the general applicability of the catalyst, we explored the hydrogenation of more than 80 diverse nitroarenes and consistently obtained the corresponding anilines in excellent yields and selectivities (Figs. 1 and 2 and fig. S2). In addition to amino and halogen substituents (the latter present in key agrochemical intermediates), sensitive functional groups such as aldehydes, ketones, nitriles, esters, amides, and olefins were well tolerated without being reduced to any substantial extent. In addition to benchmark substrates, marketed nitro-substituted drugs—for example, Nimodipine, Nilutamide, Nimesulide, Flutamide, and Niclosamide—were hydrogenated to the respective amines with high selectivity (Fig. 2). Moreover, amino-benzonitriles, which constitute common building blocks for the chemical industry, could be accessed in up to 88% yield (Fig. 2 and fig. S2).

Fig. 1 Hydrogenation of nitroarenes to anilines catalyzed by Fe-phen/C-800 (Fe2O3-N/C).

Reaction conditions: 0.5 mmol nitroarene, 42 mg catalyst [4.5 mole percent (mol %) Fe], 120°C, 1:1 water-THF (4 mL), 50 bar H2. Conversion and yields were determined by gas chromatography (GC) using n-hexadecane (100 μL) standard. Yields in parentheses refer to isolated yields. In all cases, complete conversion of nitroarene was observed.

Fig. 2 Fe-phen/C-800 (Fe2O3-N/C)–catalyzed hydrogenation of functionalized nitroarenes and nitro-heterocyclic compounds.

Reaction conditions: 0.5 mmol nitroarene, 42 mg catalyst (4.5 mol % Fe), 120°C, 1:1 water-THF (4 mL), 50 bar H2. Conversion and yields were determined by GC using n-hexadecane (100 μL) standard. Yields in parentheses refer to isolated yields. In all cases, complete conversion of nitroarene was observed. *, in 0.5:4 water-THF solvent (4.5 mL). †, in 4 mL t-amylalcohol at 130°C. ‡, at 105°C with 60 mg catalyst.

We then tested a variety of heterocyclic nitro compounds (Fig. 2 and fig. S2). Amino-substituted N-heterocycles in particular are important intermediates in the pharmaceutical and agrochemical industries. Again, all the catalytic hydrogenations proceeded smoothly, and we obtained the corresponding heteroaromatic amines in good to excellent yields. To demonstrate the utility of this method, we performed several gram-scale reactions (>20 mmol) for selected substrates (see fig. S4). In all cases, similar yields to the standard 0.5 mmol–scale experiments were obtained.

Next, we investigated the stability and recyclability of the catalyst, which are crucial performance metrics for cost-effective industrial processes. The iron catalyst was recycled and conveniently reused seven times without any reactivation (fig. S3). After the fifth recycling, some gradual decrease in the activity is observed. Nevertheless, full conversion was obtained after prolonged reaction time.

To elucidate reasons for the markedly different catalytic activity of the Fe-phen/C-x catalysts compared with known iron oxides, the effects of pyrolysis temperature and ligand on the structure of the catalysts were investigated in detail by transmission electron microscopy (TEM), x-ray photoelectron spectroscopy (XPS), electron paramagnetic resonance (EPR), and Mössbauer spectroscopy. The TEM image of the inactive Fe-phen/C-400 shows essentially no iron oxide particles; only very small features (<1 nm) of weak contrast within cloudy agglomerates are observed (Fig. 3A). The energy-dispersive x-ray (EDX) analysis (fig. S5C) indicates a rather homogeneous distribution of iron within these agglomerates. However, the active catalyst Fe-phen/C-800 is characterized by Fe2O3 particles of different size (Fig. 3B and fig. S5D). Their oxidic nature is confirmed by Mössbauer, XPS, and EPR results presented in the supplementary materials. Most of the oxide particles have a size between 20 and 80 nm but occur together with smaller particles of 2- to 5-nm size [Fig. 3B, high-angle annular dark field (HAADF) image].

Fig. 3 HAADF and ABF TEM images of Fe-phen-/C-x catalysts.

(A) Fe-phen/C-400. (B) Fe-phen/C-800. (C) Fe-phen/C-1000.

The iron oxide particles of the active catalyst are surrounded by a shell of 3 to 5 nitrogen-doped graphene layers [marked by arrows in the annular bright field (ABF) image, Fig. 3B]. When the pyrolysis temperature is raised to 1000°C, small particles (d ≤ 5 nm) collapse into well-faceted larger ones of 50- to 60-nm size, which are shielded by a 10- to 40-nm-thick graphite layer. This is evident from the TEM images of sample Fe-phen/C-1000 in Fig. 3C. Supported iron-nitrogen complexes used in fuel cell applications were reported to have different structures (3234).

The growth of particles within the Fe-phen/C series is also supported by EPR and Mössbauer results, which confirm additionally that the Fe particles might have a γ-Fe2O3–like structure (see fig. S8). Because no conclusive information on the location of nitrogen in the samples could be obtained by TEM/EDX analysis, due to superposition of C and N signals, XPS studies have been performed (Fig. 4). In nonpyrolyzed Fe-phen/C, two N1s peaks occur at 398.7 eV and 400.1 eV, which are assigned to pyridinic nitrogen and Fe-N centers, respectively (35). The former peak might reflect N within the free ligand, whereas the latter arises from ligand N atoms coordinating to Fe. In Fe-phen/C-400 and, to a more pronounced extent, in the most active catalyst, Fe-phen/C-800, these peaks are shifted to lower binding energies (398.3 eV and 399.3 eV). A similar shift, and even a splitting, of the N1s peak was observed for 2H-tetraphenylporphyrin monolayers in contact with deposited Fe atoms on top. It was assigned to a so-called distorted FeN4 center (36), in which the Fe-N distances are slightly different, thus giving rise to the peak splitting. It is probable that similar FeNx centers are formed with progressing pyrolysis on the surface of the emerging iron oxide particles. The fact that no splitting of the peak at 399.3 eV is observed may be due to a distribution of Fe-N distances that just contributes to the line width. A further N1s peak is observed in sample Fe-phen/C-800 at 401.0 eV, which gains intensity after pyrolysis at 1000°C and which can be assigned to nitrogen in a graphite-like structure (35).This suggests that N is incorporated into the support lattice. In the less active catalyst Fe-phen/C-1000, the characteristic peak of FeNx centers around 399.3 eV disappears, leaving behind only the signals of pyridinic (398.6 eV) and graphite-enclosed nitrogen (401.0 eV). Simultaneously, a loss of nitrogen was observed with rising pyrolysis temperature in both the near-surface region and the bulk, and the surface Fe/N ratio decreased from around 7 in catalysts Fe-phen/C-400 and Fe-phen/C-800 to 3.6 in sample Fe-phen/C-1000 (table S3). These results indicate a decomposition of the FeNx centers during pyrolysis at 1000°C, accompanied by partial sublimation and incorporation of nitrogen into the support structure.

Fig. 4 XP spectra of the N1s electrons for the Fe-phen/C catalyst as prepared and after different pyrolysis temperatures.

In contrast to the most active sample, Fe-phen/C-800, the poorly active Fe-bipyridyl/C-800 catalyst shows only one N1s peak typical for pyridinic nitrogen (398.6 eV). No hints for Fe-N centers were found (fig. S9). Moreover, TEM indicates much larger Fe2O3 particles for this material that consist of several crystallites and that are not surrounded by a graphene layer (fig. S5A). This is also true when the catalyst is prepared in the same way, but without any ligand, in which well-faceted big particles of 100- to 800-nm size are formed [fig. S5B, sample Fe(OAc)2/C-800]. For sample Fe-phen/Al2O3-800, which is almost as active as sample Fe-phen/C-800 (table S1), XPS points to a very similar nature of the N species, including the presence of FeNx (fig. S10), whereas TEM/EDX indicates a rather homogeneous distribution of Fe on the support with hardly any pronounced particle formation (fig. S6 B). In contrast, very large needle-like particles are formed in catalyst Fe-phen/TiO2-800, which are covered by a few graphene layers (fig. S6 A) containing FeNx species as evidenced by XPS (fig. S11).

Comparison of characterization and catalytic results suggests that it is these particular FeNx centers formed in a narrow range of pyrolysis temperatures around 800°C that govern the unique catalytic activity. The size of the iron oxide particles seems to play a minor role as long as their growth does not reduce the exposed catalytically active surface, as might be the case in catalyst Fe-phen/TiO2-800.

Supplementary Materials

Materials and Methods

Figs. S1 to S11

Tables S1 to S3

References (3742)

References and Notes

  1. Acknowledgments: We gratefully acknowledge the support from the Federal Ministry of Education and Research (BMBF) and the State of Mecklenburg-Vorpommern. We thank the analytical department of the Leibniz-Institute for Catalysis, Rostock for the nuclear magnetic resonance measurements. The Leibniz Institute for Catalysis has filed a patent on the catalysts reported herein. Author contributions: M.B. and R.V.J. planned the project; R.V.J. developed and prepared the catalysts; R.V.J designed and performed all catalytic hydrogenation experiments; R.V.J., M.B., and A.B. wrote the paper; A.-E.S., H.J., and M.B were involved in the development of this class of catalysts; M.-M.P. performed TEM measurements and analysis; J. Radnik did the XPS measurements; J. Rabeah and A.B. conducted EPR measurements; and H.H. and V.S. performed Mössbauer measurements.
View Abstract

Stay Connected to Science

Navigate This Article