Towards Unveiling the Mystery of Electron-Liquid Interaction in Liquid-Phase TEM: Implications for Practical Application

Liquid-phase TEM (LP-TEM) represents a game-changing in situ characterization method that enables real-time observations of nanostructure evolution in liquid media down to the atomic scale. However, a comprehensive understanding of electron-liquid interactions is still incomplete, due to the overall infancy of LP-TEM. To enable LP-TEM becoming a quantitative method, artifacts such as electron-beam induced radiolysis, charging, and heating must be clarified. While beam-induced heating was predicted to only have a minor impact in LP-TEM [1], in situ temperature measurements revealed significant sample heating (see Figure 1) when applying the bubble method for increasing the resolution capability [2]. To probe this effect, we quantified temperature changes in situ by measuring the lattice expansion of gold nanostructures using charge-compensated parallel-beam electron diffraction [3, 4], a method allowing for detection of temperature changes down to sub-Kelvin precision [4]. Although not fully clarified, these results already imply that the most infamous phenomenon in LP-TEM, namely radiolysis, can be heavily influenced under certain conditions. Yet, implementing a comprehensive radiolysis simulation model without additional dependencies, is already a sophisticated task comprising a large set of coupled chemical reactions for pure water [5]. To overcome this obstacle, we introduced an advanced software solution for automated radiation chemistry simulations (AuRaCh), which simplifies setting up comprehensive reaction sets [6]. This was demonstrated for classification and quantification of the redox chemistry in irradiated HAuCl4 containing solutions. Our findings are in good agreement with experiments using graphene-supported microwell liquid cells [7, 8], as well as phenomena reported in literature, such as gold nucleation [9], volatile gaseous species [10], or concentration-dependent etching [11, 12]. On top, it was shown that radiation chemistry simulations can be used to investigate the influence of electron irradiation on the Gibbs free energy landscape by quantifying dissolution kinetics [6]. This directly allows not only to thermodynamically access electron beam-induced radiolysis, but also to map the Gibbs free energy landscape with respect to parameters that cause little distortions on the observation of interest. While electron beam-induced radiolysis heavily influences the redox chemistry of a solution, it also causes the simultaneous generation H+, and OH-, strongly affecting the acid-base chemistry in LP-TEM. We show that pH is in general not suited for describing the acidity in irradiated solutions and, thus, introduce alternative measures that account for the concentrations of both, H+ and OH-, which we denote as radiolytic acidity π* and radiolytic ion product Kw* [13]. By applying this concept, we shown that the acidity in irradiated water is not shifted towards acidic conditions as assumed so far [5]. Instead, irradiation drives the system towards neutral conditions, yet with strongly increased H+ and OH- concentrations (see Figure 2). In addition, by adding a-priory pH-irrelevant solvents, π* can be heavily influenced. This is evaluated by suitable kinetic models for aqueous solutions of LiCl, LiBr, LiNO3 using AuRaCh [13]. Moreover, the simulated radiation-dependent acidity change is validated in situ using acidity-dependent anisotropic dissolution of Goethite mineral needles in LP-(S)TEM. The findings are in good agreement with both, kinetic simulations of aqueous solutions containing FeIII, and comprehensive ex situ studies on goethite dissolution. Finally, we discuss that the general concept is not only relevant to LP-TEM but to any in situ method utilizing ionizing radiation such as high-intensity X-rays [15]. a) principle of parallel-beam electron diffraction for in situ temperature measurements in gaseous environment: lattice expansion of gold is detectable by contracting diffraction rings. b) dose-rate dependency of beam-induced heating in LP-TEM utilizing the bubble method compared to literature [1,14] and own models. a) acid-base correlation in non-irradiated solutions governed by autoprotolysis of water and b) governed by radiolysis products in irradiated solutions causing an increase of both, H+ and OH- concentrations simultaneously.