The 2025 Nobel Prize in Chemistry: Reflections from The University of Melbourne

The 2025 Nobel Prize was awarded in Chemistry to Susumu Kitagawa, Richard Robson and Omar Yaghi "for the development of metal-organic frameworks".^[1] Early work was undertaken in the School of Chemistry at the University of Melbourne by Richard Robson. When Robson began this work, these network materials were referred to as coordination polymers. Robson was aided in these investigations by his long-time colleague and trusted collaborator, Bernard Hoskins. This article describes the origins of Robson's early work and how single crystal X-ray diffraction was used to characterise these new materials.

The original publication in the area, co-authored by Bernard Hoskins and Richard Robson, was a 1989 communication that was followed by a full article in 1990.^{[2, 3]} The 1990 paper opens with:

"In this section we present some simple structural ideas that appeared to us to have far-reaching implications with regard to the deliberate design and construction of a new and potentially extensive class of solid materials with infinite framework structures resembling scaffolding. We also put forward sensible reasons for expecting these new materials to show unusual and useful properties. The synthetic and crystallographic results presented in later sections represent some of our early attempts at the deliberate construction of infinite 3D frameworks. The results obtained provide considerable encouragement that a wide range of materials with scaffolding-like structures should prove synthetically accessible."

The early inspiration for this work can be traced to the 1970s when Robson was assigned the task of producing models of inorganic solids such as diamond, wurtzite, sodium chloride and rutile for use in first year chemistry lectures at the University of Melbourne. In the construction of the models, wooden balls were linked by metal rods. In order to give the intended structure, holes needed to be drilled into the balls at calculated positions to allow the rods to connect the spheres together. During the assembly process, it occurred to Robson that the wooden balls, with their drilled holes at specific geometric positions, were geometrically predisposed to perform specific structural roles within a targeted structure. After construction of the models, Robson was intrigued by the large space between the balls in the ball-and-stick models, but of course, in the dense inorganic solids these spaces do not exist.

As a coordination chemist, Robson wondered whether it was possible to generate crystalline solids in which the metal rods represented molecular links instead of single bonds. His early efforts involved the synthesis of a tetrahedral organic building block in which four 4-cyanophenyl groups are bound to a tetrahedral carbon centre. The four cyano groups are thus directed to the vertices of a tetrahedron. When combined with a metal centre that prefers tetrahedral coordination, such as Cu(I), a network material is generated that possesses the same underlying connectivity as diamond (Fig. 1). Instead of a dense inorganic solid, the resulting coordination polymer possesses large intraframework voids filled with highly disordered (essentially liquid) solvent molecules and uncoordinated counterions. The solvent filled voids occupy ~2/3 of the crystal volume.

This early success prompted investigations using different solids as topological models. For example, using PtS as a model, an open network crystalline material was generated in which planar 4-connecting Cu(II)[tetra(4-pyridyl)porphyrin] units are linked by tetrahedrally coordinated Cu(I) centres.^[4] Similarly, networks were generated by linking the trigonal ligand tricyanomethanide with octahedral divalent metal centres, which correspond to the trigonal oxide ions and octahedral Ti(IV) centres, respectively, in rutile, TiO₂.^{[5, 6]}

Over the years, other network topologies were explored. Often the topological model was a common inorganic solid but other nets corresponding to more obscure crystalline solids were also used. In this regard, the work of A. F. Wells was instructive. In his book, Structural Inorganic Chemistry, Wells surveys the field of inorganic structures and systematically describes a vast number of inorganic solids with a strong emphasis on the connectivity of these materials.^[7] In 1977 and 1979 he published two further books in which he focuses on a mathematical description to describe the connectivity.^{[8, 9]} These books served not only to assist in the design of new network materials but also proved useful in describing the underlying connectivity in complex coordination polymers.

With Robson focusing on the design and synthesis of these unique crystalline materials, Hoskins was responsible for the structural characterisation using single crystal X-ray diffraction. Hoskins started in the Department of Inorganic Chemistry at the University of Melbourne in the mid-1960s following a post-doctoral appointment at Oxford University with Nobel Prize winner Dorothy Hodgkin. Upon his arrival, Hoskins established a vibrant X-ray crystallography laboratory at the University of Melbourne. He was a meticulous and insightful researcher with a deep understanding of small molecule crystallography. Over four decades he trained numerous research students in crystallography, many of whom established themselves as distinguished crystallographers.

Small molecule crystallography in the late 1980s and 90s was particularly challenging and often required the input of a specialist crystallographer in regard to the handling of crystals, measurement of data and structure determination. When Hoskins first began work on Robson's crystals, the instrument in use was an Enraf-Nonius CAD4 diffractometer. Whilst this was a beautifully engineered instrument, its point detector allowed the intensity of only one reflection to be measured at a time. As a consequence, data collections typically took days and sometimes weeks to complete.

The new network materials presented additional challenges. Crystals commonly contained large regions of solvent, which could be lost upon removal from the mother liquor, leading to a deterioration in the crystal quality. Cooling of crystals immersed in a protective oil was seldom performed because the cryogenic equipment at the time was extremely temperamental. As a result, the crystallographer relied upon mounting a crystal in a Lindemann glass capillary tube along with a small quantity of the solution from which it was obtained, to protect the crystal from solvent loss. Whilst good quality diffraction was commonly evident at low 2θ angles for the open framework structures, there was often a paucity of higher angle data.

The highly disordered solvent could seldom be modelled, with Fourier difference maps indicating broad smears of electron density rather than clearly defined peaks. This was at a time before Anthony Spek's invaluable SQUEEZE tool was in widespread use.^[10] As a result of the disordered solvent, elevated R values were a common occurrence for these framework materials and referees regularly queried the high R values. Analysis of diffraction data often revealed high symmetry, and it was not uncommon to find solutions in multiple space groups that shared the same Laue symmetry and systematic absences. Manual determination of the correct space group was sometimes challenging but a 1995 paper of Richard E. Marsh provided instructive advice that regularly guided space group selection.^[11]

Refinements of structures, particularly those with large data sets and high numbers of parameters, were very time-consuming given the limitations in computer power. The representation of structures using ORTEP could be a slow, laborious process requiring the specification of atoms and symmetry operations in unforgiving fixed format files. Nevertheless, the program yielded high quality black and white images. One of the advantages of using ORTEP was that the user gained a deep appreciation of the underlying symmetry of the network material.

In the last 35 years there have been enormous advancements across all areas of X-ray crystallography with the improvement in diffractometer hardware probably the single most important development. When Robson and Hoskins first started working on network systems, the queue for the diffractometer as well as the time to measure a dataset were the rate determining steps in obtaining results. Remarkable advancements in computer hardware and software have streamlined the process enormously but it is interesting to note that the group still utilises various SHELX programs.^[12]

It is both satisfying and astonishing to see how the field of coordination polymers and metal-organic frameworks has developed. There appears to be no limit to the type of framework materials that can be synthesised. We are restricted only by our imaginations in generating materials that can be tailored for specific purposes. Despite all these changes and advancements, it appears that single crystal X-ray diffraction will continue to have an important role in driving this field forward, deep into the 21st century. Sadly, Bernard Hoskins passed away in 2002 and never lived to see the maturation of the field and its global impact.

References

2. Hoskins, B. F. & Robson, R. (1989). J. Am. Chem. Soc. 111, 5962–5964.

3. Hoskins, B. F. & Robson, R. (1990). J. Am. Chem. Soc. 112, 1546–1554.

4. Abrahams, B. F., Hoskins, B. F., Michail, D. M. & Robson, R. (1994). Nature 369, 727–729.

5. Batten, S. R., Hoskins, B. F. & Robson, R. (1991). J. Chem. Soc. Chem. Commun. pp. 445-447.

6. Batten, S. R., Hoskins, B. F., Moubaraki, B., Murray, K. S. & Robson, R. (1999). J. Chem. Soc. Dalton Trans. pp. 2977–2986.

7. Wells, A. F. (1984). Structural inorganic chemistry, 5th ed. Oxford University Press.

8. Wells, A. F. (1977). Three-dimensional nets and polyhedra. New York: Wiley.

9. Wells, A. F. (1979). Further studies of three-dimensional nets, ACA Monograph No. 8. New York/Pittsburgh: American Crystallographic Association.

10. Spek, A. L. (2015). Acta Cryst. C71, 9–18.

11. Marsh, R. E. (1995). Acta Cryst. B51, 897–907.

12. Sheldrick, G. M. (2008). Acta Cryst. A64, 112–122.