Stefan Merkens

Metabolic heterogeneity within the cell microenvironment is a key driver of cancer progression and resistance to therapy. However, current approaches lack the spatial and temporal resolution required to capture its dynamics in living systems. While recent advances in 3D cell culture models and metabolomic profiling have improved our understanding of the tumor niche, their integration with real-time optical sensing remains underdeveloped. Here, we present an integrated platform combining a 3D-printed microreactor culture chamber with Raman spectroscopy to enable non-invasive, spatially resolved metabolic monitoring of living cell cultures. Our microreactor platform generates controlled oxygen and nutrient cues while simultaneously acquiring label-free Raman spectra, revealing extracellular metabolic fingerprints linked to cell catabolism (e.g., glucose and lactate shifts) and acidification. Analysis across four cell lines uncovered temporal evolution as the dominant source of metabolic variance, while spatial heterogeneity along oxygen gradients is a secondary factor. In particular, diffusion-limited regions exhibited localized acidification and accumulation of stress biomarkers—such as the release of nucleotides—features that cannot be detected using conventional bulk assays. By providing a versatile platform for real-time mapping, this work enables the mechanistic dissection of cell adaptation to microenvironmental stress and supports the prediction of metabolic signatures underlying drug response and treatment outcomes.

Liquid-phase transmission electron microscopy is a burgeoning experimental technique for monitoring nanoscale dynamics in a liquid environment, increasingly employing microfluidic reactors to control the composition of the sample solution. Current challenges comprise fast mass transport dynamics inside the central nanochannel of the liquid cell, typically flow cells, and reliable fixation of the specimen in the limited imaging area. In this work, we present a liquid cell concept – the diffusion cell – that satisfies these seemingly contradictory requirements by providing additional on-chip bypasses to allow high convective transport around the nanochannel in which diffusive transport predominates. Diffusion cell prototypes are developed using numerical mass transport models and fabricated on the basis of existing two-chip setups. Important hydrodynamic parameters, i.e., the total flow resistance, the flow velocity in the imaging area, and the time constants of mixing, are improved by 2-3 orders of magnitude compared to existing setups. The solution replacement dynamics achieved within seconds already match the mixing timescales of many ex-situ scenarios, and further improvements are possible. Diffusion cells can be easily integrated into existing liquid-phase transmission electron microscopy workflows, provide correlation of results with ex-situ experiments, and can create additional research directions addressing fast nanoscale processes.

Electrochemical Liquid Phase Transmission Electron Microscopy (EC-LPTEM) is envisioned as an invaluable tool for the investigation of structural and morphological properties of functional materials in electrochemical systems for energy transition applications, especially in aqueous-based electrolytes. However, one major issue in accessing electrochemical conditions (e.g. overpotential, current density) relevant for energy applications (e.g. heterogeneous catalysis, batteries) is the saturation of the liquid cell by gaseous products resulting both from the reaction of interest and from water decomposition at the Working electrode (WE) and Counter electrode (CE). In this work, the performance of an optimized liquid cell geometry is investigated, with the aim of fine tuning the experimental set-up and optimizing the functional activity of the active materials under investigation. Ex situ and in situ comparative flow experiments are carried out in order to understand the role of the optimized liquid flow configuration for electrochemical experiments in aqueous electrolytes. The key role of the optimized flow geometry in reducing the formation of gas bubbles at both the WE and CE is highlighted, coupled with an improvement in the removal of gaseous products at the WE/CE. Finally, the validation of the optimized microfluidic set-up, through the electrodeposition of Zn nanostructures in aqueous electrolyte, is presented as a proof-of-concept experiment showing that the optimized geometry provides access to experimental conditions which have not been available with the standard liquid cell configuration.

Liquid Phase-Transmission Electron Microscopy research increasingly relies on liquid flow reactors to monitor nanoscale dynamics. Current challenges comprise fast mass transport dynamics inside the central nanochannel of the liquid cell, typically flow cells, and reliable fixation of the specimen in the limited imaging area. In this work, we present a novel liquid cell concept, the diffusion cell, that satisfies these seemingly contradictory requirements by providing additional on-chip bypasses to allow high convective transport around the nanochannel in which the diffusive transport predominates. Diffusion cell prototypes were developed using numerical mass transport models and fabricated on the base of existing two-chip MEMS-setups. Important hydrodynamic parameters such as the total flow resistance, the flow velocity in the imaging area and the time constants of mixing were improved by ~2-3 orders of magnitude compared to existing setups. Obtained solution replacement dynamics within seconds already matches the mixing timescales of many ex situ scenarios, with further improvements possible. Diffusion cells can be easily integrated into existing Liquid Phase Transmission Electron Microscopy workflows, provide correlation of results with ex-situ experiments, and can create entirely new research directions for fast nanoscale processes in liquids.

Applying a continuous flow to rinse radiolytic species from the irradiated volume is a widely proposed strategy to reduce beam-related artefacts in Liquid-Phase Transmission Electron Microscopy (LP-TEM). However, this has not been verified either experimentally or theoretically to date. Here we explore an extended numerical model implementing radiolytic chemistry, diffusion and liquid convection to study the peculiarities of beam-induced chemistry in the presence of a flowing liquid within a heterogenously irradiated nanoconfined channel corresponding to a LP-TEM flow cell. Intruigingly, the concentration of some principal chemical species, predominantly hydrogen radicals and hydrated electrons, is found to grow significantly rather than to decrease in respect to zero-flow when moderate flow conditions are applied. This counterintuitive behaviour is discussed in terms of reactants’ lifetimes, spatial separation of the reaction network and self-scavenging by secondary radiolytic species. In the presence of a flow the consumption of highly reactive species is suppressed due to removal of the self-scavengers, and as a result their concentration in the irradiated area increases. A proof of concept for the supply of scavengers by the flow is demonstrated. Unravelling the effect of flow on radiolysis spawns direct implications for LP-TEM flow experiments providing yet one more control parameter for adjusting the chemistry in the irradiated/imaging area, in particular for mitigation strategies by continuous supply of scavengers.

Hysteresis is an essential attribute of many solid-state devices and biological processes, yet it is often overlooked in colloidal and soft-matter dynamic systems. Herein, we show that gold nanoparticles can remain dispersed or aggregated at the same temperature depending on the trajectory of applied stimulus, featuring hysteretic behavior. Aided by real-time analytics and fine-tuning of experimental parameters, such as salt concentration, nanoparticle diameter, and surface potential, we disentangled the kinetic (rate-dependent) and thermodynamic (rate-independent) components of hysteresis in cyclic clustering of nanoparticles. The hysteresis originates from the difference in the aggregation and disassembly temperatures. Our findings enrich the repertoire of the experimental framework with potential for stimuli-sensitive nanotransducers, information storage, switchable catalysis, or autonomous chemical networks with feedback loops.

Hysteresis is an essential attribute of many solid-state devices and biological processes, yet it is often overlooked in colloidal and soft-matter dynamic systems. Herein we show that gold nanoparticles can remain dispersed or aggregated at the same temperature depending on the trajectory of applied stimulus, featuring hysteretic behaviour. Aided by real-time analytics and fine tuning of experimental parameters, such as salt concentration, nanoparticle diameter and surface potential, we disentangled kinetic (rate-dependent) and thermodynamic (rate-independent) component of hysteresis in cyclic clustering of nanoparticles. The hysteresis originates from the difference in the aggregation and disassembly temperatures. Our findings enrich the repertoire of experimental framework suitable for study life-like features on purely synthetic settings.