Paper contributes to the question of what is a metal?
This is a semi-technical lay-perspective of our work published in the June 5th edition of Science, which made it’s way to the front cover. The full author contributor list is below:
Tillmann Buttersack*, Philip E. Mason* Ryan S. McMullen* ,H. Christian Schewe*, Tomas Martinek, Krystof Brezina, Martin Crhan, Axel Gomez Dennis Hein, Garlef Wartner, Robert Seidel, Hebatallah Ali, Stephan Thürmer, Ondrej Marsalek †, Bernd Winter †, Stephen E. Bradforth †, Pavel Jungwirth â€
*co first-authors; †corresponding authors
Think of magnesium in fireworks and gold in jewellery. Both metals are entrancing but for opposing reasons. With the help of a spark, one reacts with oxygen and gives out a bright glow in a classic school demonstration and the other remains stable under the same conditions indefinitely. If we were asked to divide things into metals and non-metals however, we could place both magnesium and gold in the correct category almost immediately, despite differences in both appearance and reactivity.
Buried deep within the bewildering properties of metals, there is a research question up for grabs: which properties are inherent to a metal and which are incidental? Our intuition might tell us that metals are ‘heavy’, or dense. At two extremes lithium is a featherweight, able to float on water, but osmium is a heavyweight champion: the densest element, at twice the density of lead—yet both are metals.
Lithium (left), sodium (right) stored under an argon atmosphere; both are light, silvery metals but react quickly in the presence of oxygen and water, which dulls their appearance.
Many other definitions of ‘metal-hood’ suffer similar contradictions, thwarting our plans at neat categorization. Cutting indium with a knife is easy, titanium pointless. As for melting points, tungsten melts at 3400 oC; caesium below body temperature. One of the 20th century experts on the physics of metals, Nobel Laureate Nevill Mott weighed in on this conundrum of contrasts in a 1996 letter:
I’ve thought a lot about ‘What is a metal?’ and I think one can only answer the question at T = 0. There a metal conducts, and a non-metal doesn’t. 1
In this understated summary, after a lifelong dedication to the topic, Mott declares that the only thing he was sure about is that only metals are able to conduct electricity at absolute zero (T = 0), the coldest possible temperature at which molecular motion would stop except for the limits imposed by quantum mechanics. Conduction it turns out, unlike density or hardness, is an important inherent property of metals.
Experimentalists often compare properties of one substance to another instead of directly relating phenomena to definitions. But can we do better and make a metal from scratch? In a mystical chemical transformation the answer is yes, with a trick first noted by Humphry Davy in 1809. 2
Solvated Electrons . . . what are they?
Like a chef starts with a base to which further ingredients are added to make soup, our unlikely process also starts with a liquid basis, albeit an unobvious one. Ammonia, a nitrogen atom attached to three hydrogen ‘legs’, is a gas at room temperature. Plunging the temperature below its boiling point at -330C, ammonia condenses to a clear, colourless liquid with a sharp odour. If not for its olfactory punch, liquid ammonia could easily be confused for water.
Liquid ammonia is a solvent similar to water in many ways with the exception of some hidden unexpected chemical talents. Here it is kept below its boiling point of minus 33 degrees Celsius in a cold thermal flask.
One curious difference between water and ammonia arises from the interaction with alkali metals. When these reactive metals (sodium, potassium all the way down to caesium) are dropped into water, a spectacular reaction occurs, which ranges from fizzing with lithium down to explosions with potassium, often accompanied by colourful flames. On the other hand, when alkali metals are added to liquid ammonia, something stranger happens.
For each alkali metal atom added, ammonia molecules can pull an electron from the alkali metal, which does not go on to react with ammonia as it does with water but becomes trapped in the gaps between ammonia molecules, becoming a solvated electron.
Liquid ammonia dissolves metals such as lithium by stripping an electron from the metal to liberate solvated electrons, as seen by the appearance of a blue colour to the solution (left) as well as positively charged lithium ions. Once all dissolved, the solution appears almost black (right).
Solvated electrons begin to pair up
Solvated electrons (represented by large green bubbles) exist in the gaps between solvent molecules. Image courtesy of Ondřej Maršálek
Single electrons come in two types: spin-up and spin-down. In atoms and molecules, two electrons occupying the same level, or orbital, will closely pair up their spins to reduce their energy. This doesn’t typically happen outside of atoms or molecules because one cannot build up a high enough concentration of ‘free’ electrons for this to happen: unpaired electrons, more commonly found in free radicals, usually rapidly undergo reaction with something else. Solvated electron reactivity towards ammonia is low however, so these electrons then go on to pair up within a solvent cavity.
The number of solvated electrons in liquid ammonia can be built up to sufficient amounts that spin-pairing can be seen. Image courtesy of Ondřej Maršálek
For those who like brightly coloured compounds as seen in sci-fi films, solutions of solvated electrons and dielectrons come close to the ticket, being extremely blue-coloured, even at dilute concentrations. This result should seem surprising, since adding the silvery alkali metals to a colourless liquid gives a blue solution, independent of the metal added, and yet no reaction in the traditional sense is occurring.
And now . . . making the metal
Dilute blue solutions of solvated electrons bear little relevance to metals so far. After all, solvated electrons are tethered to a particular solvent cavity or are in other words ‘localized’. Metals on the other hand consist of a ‘sea’ of delocalized electrons, spread out and not tied to a particular atom or molecule. If we add sugar to water, a point will be reached where no more can dissolve and the additional solid remains as a deposit left in the bottom of the glass. This is also true with the alkali metals, but before that happens, the colour changes from a deep blue to a beautiful bronze.
Adding more alkali metal to the solvated electron solution is all it takes to generate a new liquid metal with a bronze lustre, which provides some clues as to its metallic nature.
Based on visual appearance alone, we would probably be inspired to include this new substance along with gold and magnesium in our list of metals and non-metals.
Our molecular picture shows solvated electrons coalescing into pairs of electrons as more alkali metal is added to an already dilute solution. But when the concentration is increased so much to induce a transition to a metal, we can then imagine electron pairs further associating to form a network of electrons well-spread out in solution.
Metallic Ammonia Probed with New Method
Within this rich history of metal-ammonia solutions comes our experiment. Nested within atoms or molecules, electrons are tethered to positively charged nuclei. Light in the UV or X-ray region can be used to punch electrons from their tethered positions and out of the molecule, to see how tightly bound they are to the constituent nuclei of the molecule. This technique is known as photoelectron spectroscopy. For molecules with many electrons, some of those electrons, the so-called core electrons, will be closer to the nuclei and harder to remove and some will be further out, the so-called valence electrons. By using high energy X-ray light, all of these electrons from the core to the valence band can be removed to collect information about the energy levels of the electrons in the molecule.
In our case, this is an important diagnostic tool to discern the difference between electrons from ammonia and those from the dissolved metal; the latter electrons responsible for electrical conduction are not tethered to particular nuclei and are hence the weakest bound electrons. These excess electrons should hit our detector with the most residual energy from X-ray light, as little energy is required to remove them from their original locations.
A specific method of collecting photoelectron spectra from a liquid microjet in vacuum was used to collect a spectrum of pure liquid ammonia previously. 3,4 Armed with a liquid ammonia spectrum, how would this look when alkali metals are added?
Results from Liquid Microjets in Vacuum
Solvated electrons were not measured in this particular experiment, due to signal to noise constraints, so the most dilute spectrum measured is the spin-paired electron species: a concentration region well below the limit for ammonia to transition to a metal. The spectrum to the bottom left shows a representative metal-ammonia spectrum at a dilute yet spin-paired electron concentration (bottom), which shows up as a single, featureless peak. As predicted, these electrons are much less bound than the electrons which are part of the ammonia molecules themselves; the electrons belonging to ammonia are not shown in this energy region.
A representative photoelectron spectrum of dilute metal-ammonia solution (bottom) and a metallic ammonia spectrum (top). The structure of the single peak changes as concentration increases and further bands emerge which are indicative of metallic behavior. The two energy scales (vacuum and Fermi) refer to two different methods of referencing how bound the electrons are to the environment from which they came.
As the concentration is increased up to a metallic concentration (top), the single peak as seen in the dilute spectrum changes shape and develops an ‘edge’. This sharp so-called Fermi edge is characteristic of a photoelectron spectrum of metals. The ‘sharpness’ of this edge is well defined by the theoretically derived Fermi-Dirac distribution, which describes how electrons are apportioned into the energy levels of a system at a given temperature. What this process also showed us was that this transition was gradual, with features of a metallic system showing up even at intermediate concentrations.
At the higher concentrations, to the left of the highest intensity peak (at around 2 eV below the Fermi edge), there are side peaks which relate to the golden colour of metallic ammonia. Absorption, which gives rise to colour, is different from a photoelectron experiment in that it involves absorption of light to promote an electron from one energy level to another (rather than ejecting an electron out of a molecule). Wavelengths which are not absorbed across all possible transitions give rise to the colour of the object. In metals, there are transitions involving collective oscillations of delocalized electrons, where the electrons move in unison.
Like oscillations on an instrument string are excited when plucked, absorption of light causes similar disturbances of ‘free’ electron collective motion in metals. Called plasmon resonances, transitions between different plasmon modes determines the bronze colour of metallic ammonia. In our experiment, some outgoing photoelectrons lose some of their kinetic energy in causing excitations between plasmon modes. So, the side bands in the above spectrum are electrons that would have been ejected from the peak near the Fermi edge, but just so happened to lose some of their energy in exciting plasmon resonances along the way.
To the Future
Of the myriad uses for metals, from aluminium in life-carrying aircraft to iridium in LEDs, there does not seem to be an obvious practical use for metallic ammonia. But, as hinted earlier, the chemical complexity of these solutions is immense.
Go the other way into the dilute regime of electrons in ammonia and you are now in an area of practical relevance to organic chemists. Solvated electrons have long been used in a reaction known as the Birch reduction, where the electron can attack organic molecules to make new molecules of greater interest and utility—this reaction was used in the manufacture of the first oral contraceptives in the 1950s 5 and is still used in selective synthesis today. With a better understanding of these electrons species at all concentrations, it’s entirely possible that new syntheses will open to the organic chemist in the future.
References
1. Edwards, P. P., Lodge, M. T. J., Hensel, F. & Redmer, R. ‘… a metal conducts and a non-metal doesn’t’. Philosophical Transactions Royal Soc Math Phys Eng Sci 368, 941–965 (2010).
2. Edwards, P. P. The Electronic Properties of Metal Solutions in Liquid Ammonia and Related Solvents. Adv Inorg Chem Rad 25, 135–185 (1982).
3. Buttersack, T. et al. Valence and Core-Level X-ray Photoelectron Spectroscopy of a Liquid Ammonia Microjet. J Am Chem Soc 141, 1838–1841 (2019).
4. Buttersack, T. et al. Deeply cooled and temperature controlled microjets: Liquid ammonia solutions released into vacuum for analysis by photoelectron spectroscopy. Rev Sci Instrum 91, 043101 (2020).
5. Birch, A. J. Steroid hormones and the Luftwaffe. A venture into fundamental strategic research and some of its consequences: The Birch reduction becomes a birth reduction. Steroids 57, 363–377 (1992).
6.) Buttersack, T. et al. Photoelectron spectra of alkali metal–ammonia microjets: From blue electrolyte to bronze metal. Science, Vol. 368, Issue 6495, pp. 1086-1091 DOI: 10.1126/science.aaz7607
