Inorganic chemistry

Ammonia transformed

Ammonia is produced industrially by combining nitrogen and hydrogen gas, catalysed over a solid iron surface. How about a catalytic reaction that could take place in solution? The first steps have now been taken.

The synthesis of ammonia from its constituent elements, nitrogen and hydrogen, ranks as one of the most important discoveries in industrial catalysis1. It merited the award of two separate Nobel prizes — first to Fritz Haber in 1918 and then to Carl Bosch in 1931, recognizing the discovery of the process and its implementation, respectively. For more than 70 years, millions of tons of ammonia have been produced annually through the Haber–Bosch process; fertilizer made from this ammonia is estimated to be responsible for sustaining roughly 40% of the world's population and is the source for 40–60% of the nitrogen in the human body. Now a remarkable chemical transformation has been discovered (page 527 of this issue2) that is likely to have important implications for the production of ammonia.

As originally devised, the Haber–Bosch process involves the elements in their gaseous state interacting at high temperature and pressure over an activated iron surface. The iron is a heterogeneous catalyst, meaning that it is in a different phase from the reactants. The process continues to be refined, particularly now with the introduction of heterogeneous ruthenium catalysts3. A desirable improvement would be a process that operates at much lower temperatures and pressures. This could be achieved through homogeneous catalysis, involving a soluble metal complex reacting with nitrogen and hydrogen to form ammonia in a solvent. But as yet there is no known homogeneous catalyst that can effect this simple process. However, the work now reported by Pool et al.2 suggests some elementary steps that could eventually be incorporated into a homogeneous catalytic reaction.

Pool et al. have traced the activation of molecular nitrogen by an organometallic complex. The complex, shown in Fig.1a, consists of a zirconium ion to which are attached two substituted cyclopentadienyl rings (C5Me4H). When molecular nitrogen, or dinitrogen, is introduced, a side-on N2 fragment is generated, bridging between two zirconium centres. This in itself is not surprising, but what happens next to the complex, in pentane solution and in the presence of hydrogen, is unprecedented: a new complex forms, in which hydrogen atoms are added to the dinitrogen bridge. Although predicted4 a few years ago, this is the first observation of such a transformation. Why has it taken so long?

Figure 1: A new route to ammonia.

a, Pool et al.2 have traced these steps to produce ammonia from bis(tetramethylcyclopentadienyl)–zirconium dichloride (symbolized as [Zr]Cl2). b, In contrast, if a slightly different ligand is used — the pentamethyl version of the complex, ([Zr′]Cl2) — the reaction follows a different path, and no ammonia is produced.

The answer is probably that molecular nitrogen is not that good a ligand. It is so chemically inert that even binding it to metal complexes in solution, just as Pool et al.2 have done here, was a decades-long challenge for inorganic chemists5. When it can be coerced to coordinate to a metal, in many cases the interaction is so weak that other (better) ligands can easily displace the dinitrogen unit. A good example6 is shown in Fig. 1b: the reduction of the pentamethylated cyclopentadienyl derivative (η5-C5Me5)2ZrCl2 with sodium amalgam (Na/Hg) in the presence of dinitrogen. The result is a complex with dinitrogen units bound end-on and bridging end-on between two zirconium centres. When exposed to hydrogen, all three dinitrogen units dissociate; hydrogen is added to each zirconium centre to generate (η5-C5Me5)2ZrH2. This is pretty typical for dinitrogen ligands5.

What is remarkable is how small, seemingly insignificant changes in the supporting ligands can drastically affect the outcome of such transformations. As part of an extensive study of the effects of substituents, Pool et al.2 examined the tetramethylated version of the same cyclopentadienyl derivative — that is, one methyl group on each carbon ring is replaced by a hydrogen atom. Here, reduction of the complex with sodium amalgam in nitrogen produces a different complex, incorporating a side-on dinitrogen bridge unit, as discussed above. So, by replacing one methyl group with hydrogen, the coordination geometry of the dinitrogen unit has changed.

Now the reaction with hydrogen proceeds rapidly, without dissociation of the dinitrogen unit, to form a diazenido complex (Fig. 1a). This is worthy of comment but does have some precedent in the literature, both experimentally and theoretically. What happens next, upon thermolysis and further reaction with hydrogen, are brand new transformations: when the diazenido derivative is heated, hydrogen is lost and the central bridging N2H2 moiety is rearranged; further reaction in excess hydrogen results in the release of ammonia, NH3.

The sequence of transformations is not catalytic — this would require regeneration of the dinitrogen complex by the reaction of N2 with (η5-C5Me4H)2ZrH2, ready to repeat the process. But it is relevant, for two main reasons, to devising a homogeneous catalytic process for the production of ammonia. First, although both hydrogen and dinitrogen have been shown to bind at metal centres in solution individually and even at the same time, they have only recently been seen to interact with the formation of one N–H bond7. Pool and colleagues' work2 suggests that these two simple molecules can react to generate a variety of NHx ligated species. Second, and more interesting, is the remarkable effect of substituents on the ancillary ligands. The fact that changing from a pentamethylated ligand to a tetramethylated ligand can completely change how the dinitrogen unit coordinates with the metal complex, and then how it reacts with hydrogen, is truly surprising. No one could have predicted such a profound change in outcome with such a small change in ancillary ligands.

In truth, it is unlikely that any homogeneous catalytic process will ever compete on an industrial scale with the heterogeneous Haber–Bosch reaction and its modern variants. But now that ammonia has been produced from its elements in solution, one can only begin to imagine what other kinds of transformation might be possible for molecular nitrogen.


  1. 1

    Tamaru, K. in Catalytic Ammonia Synthesis (ed. Jennings, J. R.) 1–18 (Plenum, New York, 1991).

    Google Scholar 

  2. 2

    Pool, J. A., Lobkovsky, E. & Chirik, P. J. Nature 427, 527–530 (2004).

    ADS  CAS  Article  Google Scholar 

  3. 3

    Schlögl, R. Angew. Chem. Int. Edn 42, 2004–2008 (2003).

    Article  Google Scholar 

  4. 4

    Basch, H., Musaev, D. G. & Morokuma, K. Organometallics 19, 3393–3403 (2000).

    CAS  Article  Google Scholar 

  5. 5

    Shaver, M. P. & Fryzuk, M. D. Adv. Synth. Catal. 345, 1061–1076 (2003).

    CAS  Article  Google Scholar 

  6. 6

    Manriquez, J. M. & Bercaw, J. E. J. Am. Chem. Soc. 96, 6229 (1974).

    CAS  Article  Google Scholar 

  7. 7

    Fryzuk, M. D., Love, J. B., Rettig, S. J. & Young, V. G. Science 275, 1445–1447 (1997).

    CAS  Article  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Fryzuk, M. Ammonia transformed. Nature 427, 498–499 (2004).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.