Gry Christensen was a 15-year-old year 11 student when she took part in a “citizen science” project to understand how the different crystals in mussel shells form. But unlike most school experiments, the samples that she and her 1,000 fellow secondary school pupils prepared were then blasted by scientists in a particle accelerator using X-rays 10bn times brighter than the sun.
“It was a bit of an eye opener,” Christensen says of the study, called Project M, involving students from 110 schools. They prepared different samples of calcium carbonate (the main component of mussel shells) that scientists then examined at the UK’s national synchrotron (a type of circular particle accelerator), the Diamond Light Source in Oxfordshire. The aim was to help scientists better understand how to form different types of crystal structures from the same chemical. “I was more interested in chemistry afterwards,” says Christensen, who went on to study agricultural science at Gråsten Landbrugsskole in Denmark. “The chemistry really helped me to have an insight into the natural world.”
But while such an approach may be new, understanding how crystals form is an old problem with serious ramifications. Crystal structure can affect the strength of steel, and even the therapeutic activity of medicines developed to treat Aids and Parkinson’s disease.
Calcium carbonate is the main compound in rocks such as chalk, limestone and marble, that derive from organic materials including shells. It is responsible for the annoying limescale stains around taps, as well as having useful applications from antacid tablets to concrete blocks. “Calcium carbonate is all around us,” says Dr Claire Murray, a chemist who led Project M in 2017 along with colleague and fellow chemist at the Diamond Light Source, Dr Julia Parker. But one outstanding challenge is controlling its crystal forms.
A crystal is a solid in which components are arranged in a highly ordered and repeating pattern, and the shape of this pattern – the crystal structure – determines the material’s properties. A common example of the effect of crystal structure is carbon – useful for jotting down notes when the atoms lie in sheets of honeycomb-shaped lattices in pencil lead (graphite), but much harder, and far more expensive, when the atoms are arranged in the cubic crystal lattice that forms diamond.
In other materials, control over a substance’s possible crystal structures – or “polymorphs” – has been a matter of life and death. In the early 1980s, life expectancy following an Aids-related diagnosis was less than two years. Patient outcomes began to improve significantly by the mid-1990s, thanks to the development of antiretroviral treatments, including a drug called ritonavir. However, two years after its initial release in 1996, the drug was withdrawn from the market because of issues with the stability of its crystal structure.
The ritonavir capsules were originally dispensed with the active component in a highly concentrated solution. Unfortunately, these conditions prompted the active drug to change structure, becoming less soluble than the original and hence much less effective as a drug. Further drug development has since solved the problem. However, the Parkinson’s drug rotigotine faced similar issues with a less soluble crystal structure emerging in 2008, triggering a batch recall in Europe while in the US the drug was listed out of stock until 2012, when drug developers had found a reformulation.
“There are many recent examples, but they are not all public,” says Dr Marcus Neumann, CEO and scientific and technical director of Avant-garde Materials Simulation (AMS), a German company that develops software for crystal structure prediction. “Examples go public when they affect a drug that is already on the market. And fortunately, that does not happen very often.”
For over 20 years AMS has been finessing computer code that can predict what crystal structures can form for a given chemical compound, to help drug companies catch problematic polymorphs before a medicine is brought to market. In 2019, AMS showed its code could predict the appearance of a problematic form of rotigotine. Recent updates to the algorithm incorporate the effects of temperature and humidity, and also use comparisons with crystal structure data from drug companies AMS has worked with including AstraZeneca, Novartis, AbbVie (which now produces reformulated ritonavir), and UCB Pharma (which produces the reformulated rotigotine patches).
Nonetheless, identifying the experimental conditions required to produce a specific crystal remains challenging, since different structures can occur with little change in conditions, and one structure can change into another. You can think of it like oranges stacked in a box. You can lay out a square grid of oranges and balance each orange in the layer above directly on top of the orange below, and they will balance fine for a while. However, just a tap will cause oranges on top to nestle in the dip between oranges in the layer below – the more stable structure.
“There’s still a lot of need for experiments because a lot of factors are not 100% understood as far as how to achieve certain crystal structures,” says Dr Adam Raw, head of materials science R&D in the life science division at Merck. He highlights the “large number of factors that can come into play” when introducing additives to nudge the system towards a certain crystal structure, precisely the approach Project M investigated.
Calcium carbonate has three possible crystal structures: aragonite, vaterite and calcite. A mussel selectively grows which ones it needs – the more durable calcite for the outer shell, for example – and “not using harsh chemical conditions,” says Dr Julia Parker. “Just additives , organic molecules.” Parker and Murray wondered whether the right additive at the right concentration would help them control the growth of vaterite versus calcite.
At Diamond Light Source, the pair could quickly distinguish tiny changes in the crystal structure of hundreds of samples by examining the paths of the X-rays from the synchrotron as they scattered from each crystal’s lattice. (The synchrotron accelerates electrons, which emit X-rays as they change direction to move round it.) The bottleneck was preparing all the samples – to test factors including the additive used, concentration and mixing time – until the idea dawned to work with UK schools, exploiting the similarities in the labs and environmental conditions.
Christensen and fellow pupils at Didcot girls’ school, based near Diamond, were the first to trial the sample preparation kits and helped guide Parker and Murray towards the equipment and instructions needed in each kit. The data needed to characterise each sample was gathered in just a day at the synchrotron.
The results, published in January this year, help shed light on the conditions that significantly favour or deter vaterite formation, and provide insights into the ways these crystals form. “I think they’ve made progress in showing which factors are most likely to correspond to biomineralisation [living creatures making minerals] and formation of these calcium carbonate crystals in biological applications,” says Raw. “But of course there’s a lot more work to do.” However, results from the project were not just scientific: school participants enthused by chemistry later turned up at Diamond for internship interviews.
“With the project it was like you were doing something for the real world, not just an experiment at school,” says Christensen.
