Scientific publications

Elastocaloric technology is a new way to heat and cool spaces by using stretchy metals, called shape-memory alloys, instead of harmful refrigerant gases. When these metals are squeezed or stretched, they heat up; and when they relax, they cool down. This process is called the elastocaloric effect and it is more energy efficient than traditional cooling systems, making it a cleaner, greener alternative. Elastocaloric systems could cool homes, schools, and workplaces, and they could refrigerate food and medicine in areas with limited electricity. Researchers are also testing this technology for cooling and heating of electric vehicles, where it could help conserve battery life, and for heating buildings in colder climates. Despite its promise, elastocaloric technology faces challenges, such as improving the durability of materials and making the shape-memory alloys more affordable. With continued research, this technology could someday help to reduce greenhouse gas emissions, lower energy costs, and bring life-saving cooling to more people all over the world.

Structurally heterogeneous materials present major challenges for characterization due to their complex nanoscale order. Sodium poly(heptazine imide) (NaPHI), a layered carbon nitride photocatalyst, exemplifies this complexity, with its precise structure remaining unresolved. Here, we uncover new structural insights into NaPHI using energy-filtered four-dimensional scanning transmission electron microscopy combined with machine-learning-based diffraction image segmentation, supported by transmission electron microscopy, atomic force microscopy, X-ray diffraction, and Raman spectroscopy. At the mesoscale, NaPHI flakes display bent morphologies, while nanodiffraction patterns reveal features characteristic of stacking disorder. Based on these insights, we modeled a NaPHI-layered structure incorporating out-of-plane undulations (waves) with amplitudes of ∼0.5 Å and wavelengths of 2–3 nm. This model reproduces the observed line features in nanodiffraction patterns and agrees with powder X-ray diffraction data, thereby bridging local and bulk structural information. The introduced approach uses data-driven machine learning to identify statistically significant features, offering a robust framework for structural analysis of semi-crystalline materials.

Replica molding is a widely used technique for the fabrication of polymer microstructures. As structural dimensions decrease, anti-stick surface treatment of the mold becomes increasingly critical to ensure clean demolding and preserve structural integrity. We fabricated arrays of micropillars with 20 µm diameter and 60 µm height using medical-grade polydimethylsiloxane (PDMS), MDX4-4210, and observed a high fraction of collapsed pillars for the first molding after fluorosilanization of the mold to reduce sticking. To address this issue, we systematically investigated the surface treatment protocol for the molds, made from the PDMS Sylgard 184. We provide results from complementary measurement methods, to show that an additional vacuum step partially removes unbound fluorosilane, but does not improve pillar stability. In contrast, a method based on multiple replications, where the first replication effectively removes residual fluorosilane from the mold, significantly enhances structural stability. Mechanical testing further revealed that the presence of fluorosilane lowers the Young’s modulus of both PDMS materials, MDX4-4210 and Sylgard 184, suggesting interference with the curing process. Confocal Brillouin microscopy indicated an elongation of replicated pillars and revealed a softening close to the surfaces, as well as mechanical inhomogeneities in collapsed pillars. We discuss modifications to the molding protocol to improve the reproducibility and mechanical stability of the replicated microstructures, offering insights towards more reliable routes for the fabrication of residue-free, high-aspect ratio features with controlled surface chemistry.

There is a growing scientific interest in material unpleasantness, yet the role of distinct physical parameters in perceptual and affective haptic experiences with liquids remains to be fully understood. To address this, we investigated how perceptual qualities of liquids relate to measurable physical properties and unpleasantness during active touch. We prepared 15 custom liquid samples using everyday materials. Rheological measurements showed that samples varied between physical viscosity 1mPA s and 45 Pa s ⁠. Participants explored each sample using circular rubbing motions with their index fingers. A camera system tracked finger movements, and a force sensor revealed applied normal forces, pull-off force (PoF) and the coefficient of friction (CoF). We compared these physical properties with the perceptual dimensions from our earlier work: perceived viscosity and slipperiness. Perceived viscosity correlated strongly with both physical viscosity and PoF, but not with CoF. Conversely, perceived slipperiness was associated with CoF, but not PoF or physical viscosity, demonstrating distinct links between physics and perception of liquids. Interestingly, PoF but not CoF was significantly linked to unpleasantness, suggesting that PoF but not CoF is crucial for liquid unpleasantness. These findings advance our understanding of how distinct physical properties relate to perceptual and affective experiences of liquids.

Contact resonance atomic force microscopy (CR-AFM) has been used in many studies to characterize variations in the elastic and viscoelastic constants of materials along a heterogeneous surface. In almost all experimental work, the quantitative modulus of the surface is calculated in reference to a known reference material, rather than calculated directly from the dynamics models of the cantilever. We measured the cantilever displacement with very high sampling frequencies over the course of the experiment and captured its oscillations that result from thermal energy. Using short-term Fourier transformations, it was possible to fit the thermal resonance peak of the normal displacement to track the frequency and Q-factor of the cantilever during an experiment, using a similar process to that used to calibrate the normal bending stiffness of cantilevers. With this quantitative data, we have used the dynamic mechanics models relating the contact stiffness of the tip/cantilever pressing into a surface with the oscillation frequency of the cantilever and show that they did not accurately model the experiment. Several material combinations of tip and sample were examined; tip size and cantilever stiffness demonstrate that existing models cannot capture the physics of this problem. While concrete solutions to use analytical models to interpret CR-AFM data have not been found, a possible solution may include revisiting the analytical model to capture a potentially more complex system than the current model, improved matching the cantilever/sample stiffness to obtain a larger variation in contact stiffness with frequency, or investigating the use of higher-order modes that may achieve this improved match.

Methods for the precise temporal control of cell surface receptor activation are indispensable for the investigation of signaling processes in mammalian cells. Optogenetics enables such precise control, but its application in primary cells is limited by the imperative for genetic manipulation of target cells. We here describe a method that overcomes this obstacle and enables the precise activation of the T cell receptor in nongenetically engineered human T cells by light. Our optogenetic receptor activation system OptoREACT employs a TCR-specific scFv fused to PIF6 that interacts with tetramerized PhyB in a light-dependent manner and thereby clusters and activates the T cell receptor in response to red light. OptoREACT not only omits genetic manipulation of the target cell but, because of its modular nature, is likely applicable to a broad range of oligomerization-activated cell surface receptors.

The construction and assembly of information-processing biomaterials are limited by the need for laborious assembly of various circuits. A new framework to assemble protein-based elements encoding complex Boolean operations enables user-defined release of biomolecules from these materials.

The most common characteristic observed in numerous diseases like rheumatoid arthritis or psoriasis is chronic inflammation. Endotoxemia is an important factor in these conditions as it is triggered by prolonged exposure to lipopolysaccharide (LPS), leading to inflammation and immune dysregulation. Therapeutic peptides are promising options to treat these chronic diseases with inflammatory characteristics. However, the applicability of therapeutic peptides is limited due to their poor stability in the body, which is typically overcome by cost-intensive modifications. Living therapeutics are emerging as a more cost-effective strategy to tackle this limitation by engineering microbes to produce and deliver the peptides right where they are needed. We developed an in-vitro endotoxemia (and psoriatic) model to test living therapeutics secreting anti-inflammatory peptides: KCF-18, I6P7, α-MSH (secreted from a genetically modified lactic acid-free strain of Lactiplantibacillus plantarum (TF103)) on murine macrophages, characterized the dose-response effects of these peptides and performed multi-array cytokine analysis. The model revealed that this living therapeutic approach enhanced the effects of the peptides, requiring lower amounts to achieve anti-inflammatory effects. Notably, α-MSH secreted by TF103 L. plantarum achieved significant pathway suppression, comparable to or exceeding that of synthetic controls, without inducing cytotoxicity. This points to potential synergistic effects between the peptides and the intrinsic anti-inflammatory properties of lactic acid bacteria. We will expand the applicability potential of these anti-inflammatory living therapeutic materials in an in vitro model of psoriasis.

The demand for lithium production has seen a significant rise, with the growing electric vehicle and stationary battery markets requiring further development of sustainable and scalable extraction methods. Direct lithium extraction technologies have been developed to address potential shortages, with adsorption emerging as a key method due to its efficiency and low environmental impact. Given that Al(OH)3 is already utilized as an adsorbent in various industrial applications, the practical importance of Al-based alternative systems for lithium ion extraction is increasing, yet lithium ion recovery requires harsh chemicals. In this study, we report a novel lithium extraction method combining chemical adsorption and electrochemical release using a synthesized aluminum layered double hydroxide (Al-LDH) material, developed under mild reaction conditions. The performance of the Al-LDH electrode was evaluated against a commercial Al(OH)3 adsorbent. Comprehensive characterization using techniques such as X-ray diffraction, Fourier-transform infrared spectroscopy, and scanning electron microscopy revealed detailed insights into the crystalline structure, particle size distribution, and surface morphology of the materials. The Al-LDH electrode exhibited a lithium ion adsorption capacity, achieving an average chemical uptake of lithium ions of 57.6 mg/g. In contrast, lithium-ion uptake capacity for Al(OH)3 was 1.0 mg/g over 15 cycles. Notably, this method operates under pH-neutral conditions, eliminating the need for harsh acidic or basic eluents. As a result, it prevents structural degradation and minimizes secondary pollution for potential future applications of lithium-ion recovery. The material’s layered structure selectively allowed lithium ion intake while blocking sodium ions, demonstrating its high selectivity and utility in lithium ion recovery processes. The integration of pH-neutral regeneration and high selectivity shows that Al-LDH electrodes as viable candidates for next-generation, green lithium extraction technologies.

Self-assembly is a fundamental property of living matter that drives the three-dimensional organization of cell collectives such as tissues and organs. Here, the co-assembly of synthetic and natural cells is leveraged to create hybrid living 3D cancer cultures. We screen a range of synthetic cell models for their ability to form augmented tumoroids with artificial but controllable micro-environments, and show that the balance of inter- and extracellular adhesion and synthetic cell surface tension are key material properties driving integrated co-assembly. We demonstrate that synthetic cells based on droplet-supported lipid bilayers can establish artificial tumor immune microenvironments (ART-TIMEs), mimicking immunogenic signals within tumoroids and eliminating the need to integrate complex living immune cells. Using the ART-TIME approach, we identify a AhR-ARNT-mediated co-signaling mechanism between PD-1 and CD2 as a driver in immune evasion of pancreatic ductal adenocarcinoma. Our study advances the field of hybrid organoid engineering, offers opportunities for the construction and modelling of artificial tumour environments, and marks a step towards the design of functional living/non-living cytomimetic materials.