Functional groups regulate ion concentration and pH in nanopores

 

To understand the chemical reactions occurring inside the nanopores of nanostructured materials—whether synthetic or natural, such as those found in membranes or ion channels in biological systems—it is essential to determine the ion concentration within them. For this purpose, nanopores are functionalized with specific chemical groups.

Until now, it had not been possible to determine how functional groups influence ion concentration inside nanopores.

In this study, a group of researchers from the United States reported the development of a core–shell-type plasmonic nanosensor, consisting of a gold nanorod coated with mesoporous silica functionalized with phenyl and methyl groups. This nanosensor can measure the local concentration of protons, anions (such as phosphates, nitrates, sulfates, and arsenates), as well as cations (such as mercury, lead, and copper) in functionalized nanopores. The measurements were performed using Surface-Enhanced Raman Spectroscopy (SERS), applied in situ.

The obtained values were compared with those of bulk silica. Moreover, results indicated that ion concentrations differ in pristine and hydrophobic nanopores compared with those functionalized with phenyl and methyl radicals. In the latter, an increase in anion concentration and a concurrent decrease in cation concentration were reported. Additionally, the pH within the nanopores was found to depend on the composition of the solution. In some cases, the pH inside the nanopores decreased by as much as 2.5 units compared to the bulk value.

These findings provide insight into ion–nanopore chemical interactions and enable precise and selective control of contaminants, with direct applications in water chemistry for membrane-based desalination processes, CO₂ storage, and catalysis in porous materials.

More information at: ACS Applied Materials and Interfaces

Ab initio structural solutions from nanocrystal powder diffraction using diffusion models

 

Over the past century, the development of materials science has increasingly relied on the precise determination of atomic arrangements—that is, the crystal structure and its properties. To this end, X-ray diffraction (XRD) is commonly applied, with the sine qua non being the availability of a single crystal or monocrystal. However, this is not always feasible, especially with atomic clusters of nanometric size (smaller than 1000 Å), known as the nanostructure problem. In such cases, powder X-ray diffraction (PXRD) patterns are degraded due to peak broadening, intensity loss, and Bragg peak overlap.

Researchers from the United States and Germany have proposed a procedure that uses a generative machine learning model* based on diffusion processes, trained on 45,229 known structures. The model, called PXRDnet, conditioned solely on the compound's chemical formula, can solve simulated nanocrystals up to 10 Å in 200 materials with various symmetries and complexities, including all seven crystal systems.

PXRDnet correctly identifies structural candidates in 4 out of 5 cases, with an average error of just 7% in the Rietveld refinement factor R. Moreover, it is capable of resolving structures from noisy experimentally obtained diffraction patterns.

The authors argue that this data-driven, theoretically bootstrapped approach opens new avenues for determining previously unsolved nanomaterial structures. However, the model has limitations: it requires prior knowledge of the chemical formula and is restricted to structures with fewer than 20 atoms per unit cell.

*The term “generative” refers to a class of statistical models as opposed to discriminative models. Generative models can generate new data instances, while discriminative models distinguish between different types of data instances.

The work was published by Nature Materials