A wrinkle that challenges a long-held assumption

Graphene is flat. But that doesn’t mean other 2D materials—particularly those containing more than one element—necessarily share the same structure. In the past, researchers have often simplified these materials by modelling them as perfectly flat sheets. New research from the ARC Centre of Excellence for Carbon Science and Innovation (ARC COE-CSI) suggests they’ve been working from the wrong assumption.

Graphitic carbon nitride (g-C₃N₄) is a material that is attracting considerable interest for its potential to produce hydrogen and break down environmental pollutants. It has often been treated as if it consisted of perfectly flat atomic sheets, and this flat structure has subsequently become the starting point for numerous studies attempting to predict how the material will behave and how its properties might be modified through the introduction of defects.

There’s just one problem. It’s not flat. It’s wrinkled. And that matters.

“The chemistry and structure of the material determine its properties,” says COE-CSI Affiliated member, UNSW Canberra PhD student Gbemi Abass. “If the structure is wrong, the predicted properties will definitely be wrong.”

Why nitrogen changes the picture

Graphene consists entirely of carbon atoms and forms an extraordinarily flat sheet. Graphitic carbon nitride has a related layered structure, but introduces another player: nitrogen. These nitrogen atoms change the way the carbon atoms around them interact.

“The structural challenge has been a fundamental issue in graphitic carbon nitride research for a long time,” says Abass. “When the structure is viewed from above, it looks perfectly flat. But when it is viewed from the side, it isn’t.”

When the researchers allowed their computer models to move away from the conventional flat arrangement and find a lower-energy—and therefore more stable—structure, the sheets buckled. Rather than appearing like a perfectly flat sheet of paper, the team found they were more like one with a gentle ripple running through it.

“Previously people had talked about buckling theoretically, but there was a lack of experimental evidence to prove that structure,” says ARC COE-CSI PhD researcher Adnan Ahmad from ANU. “We introduced the buckled structures, and then we validated them with experiment.”

To do so, the research team used X-ray scattering data from real samples. The experimental researchers effectively asked whether the atomic structure predicted by the computers was consistent with the material they could actually observe.

By testing different ways the material could buckle and comparing their computational predictions with experimental evidence, they produced compelling evidence that the long-held assumption of a flat structure was incorrect.

A stack of slightly wonky books

The discovery didn’t stop there. Graphitic carbon nitride doesn’t necessarily exist as one isolated sheet. Its layers can sit on top of one another.

Ahmad uses a stack of books as a simple analogy. The conventional model effectively assumes these books are flat and arranged neatly on top of each other. But the research team found the layers weren’t just buckled; they could also be shifted sideways to make the overall structure more stable.

It’s an important distinction because changing the structure also changes predictions about how the material will behave.

Why the wrinkle matters

Graphitic carbon nitride is of particular interest as a photocatalyst—a material that uses light to drive a chemical reaction. Researchers are investigating its potential to harness sunlight to produce useful products such as hydrogen.

One way researchers seek to tailor graphitic carbon nitride for different applications is by introducing other elements, known as dopants, to change its properties. The Centre team tested phosphorus and nickel and found that their predicted effects changed depending on whether graphitic carbon nitride was modelled as a flat sheet or in its more realistic buckled and stacked structure.

The problem is obvious. Assuming that the material has a flat structure is simply not a realistic starting point for any comparison.

Getting the structure right should allow researchers to make more accurate predictions about how modifications will affect the material before moving into the laboratory.

That more accurate structural picture is already informing further research within ARC COE-CSI. In related work, researchers are investigating how paired phosphorus and sulphur atoms can be introduced into graphitic carbon nitride to improve its catalytic performance. The wrinkled structural model is helping them understand why these paired dopants can behave differently—and potentially more effectively—than isolated ones.

When experiment and computation work both ways

The finding is the result of a collaborative approach between two PhD students approaching the same problem from different directions: Ahmad through experiment and Abass through computation.

For more than two years, Ahmad and Abass worked across that experimental-computational divide, continually exchanging results and ideas. A calculation could suggest something for the experimental researchers to investigate. An experimental result could send the computational researchers back to reconsider their model.

For Ahmad, who spent around a decade working in the textile industry before moving into advanced materials research, the experience has changed the way he approaches science. It connected him with researchers across other COE-CSI nodes, and gave him the opportunity to work with researchers specialising in the engineering and application of catalytic materials.

“Previously my focus was on the application,” he says. “Now I think first we need to understand the structure, rather than just focusing on the applications.”

This was also his first experience of catalysis research. “I received strong support and learned a lot from working closely with Professor Rose Amal and Professor Liming Dai’s teams,” he says.

The collaborative approach mirrors the longstanding way in which Professor Yun Liu, an experimental materials scientist at ANU, and computational chemist A/Professor Terry Frankcombe at UNSW Canberra, have worked together for more than fifteen years.

“We need computational and experimental researchers working together, understanding each other and communicating with each other,” says Professor Liu. “In that case, our research outcome will be more convincing.”

She says this is because frequently the relationship can be one-directional. An experimental researcher might obtain an interesting result and then approach a computational researcher to explain it. Alternatively, a computational prediction might later be handed to an experimental team to test.

The race to develop materials with better performance can make it tempting to move quickly towards applications. But without understanding the material itself, those efforts can be built on shaky foundations.

“Fundamental research is very critical,” says Professor Liu. “It may take a long time to understand it in the first step, but that provides the solid foundation to move forward.”

And sometimes, that first step can reveal the odd wrinkle.

Read more about this publication via https://doi.org/10.1039/D6MA00065G