In Situ Liquid Phase TEM

• In Situ Liquid Phase TEM is a methodology that enables real-time, nanometer-to-atomic scale imaging of dynamic liquid processes by confining a thin liquid layer within specialized cell architectures.
• It employs diverse approaches including closed-cell holders, open-cell ETEM, and 2D heterostructure cells to achieve spatial resolutions from sub-nanometer to true atomic scale under controlled stimuli.
• Advanced imaging protocols using STEM, EELS, and EDS facilitate quantitative chemical mapping and operando tracking of phenomena such as colloidal self-assembly, nanocrystal growth, and electrochemical reactions.

In situ liquid phase transmission electron microscopy (LP-TEM) is a suite of methodologies enabling real-time, nanometer-to-atomic scale imaging and spectroscopy of dynamic processes occurring in liquids. By confining a thin layer of liquid within the electron-beam path—using either hermetically sealed microfluidic cells, windowless open-cell geometries in environmental TEM (ETEM), or, in some cases, native sample wetting—LP-TEM overcomes the vacuum incompatibility of liquids in conventional TEM. This facilitates operando investigations of phenomena such as colloidal self-assembly, nanocrystal growth, phase transitions, electrochemical reactions, and interfacial dynamics, under precisely controlled chemical and physical stimuli.

1. Architectures and Instrumentation for In Situ Liquid Phase TEM

Three principal approaches to LP-TEM have emerged:

• Closed-Cell Liquid Flow Holders: The canonical platform consists of two electron-transparent windows (Si₃N₄, graphene) separated by spacers defining a narrow liquid gap (typically 50–800 nm) (Grogan et al., 2010, Holtz et al., 2012). The cell is hermetically sealed (polymer, epoxy, o-rings) and integrates fluidic ports for solution exchange, as well as optional electrodes.
• Open-Cell Environmental TEM (ETEM): In open-cell ETEM, a standard TEM grid is exposed to a controlled water-vapor atmosphere. Hygroscopic salt particles (e.g., NaCl) nucleate and sustain stable liquid droplets on the grid, provided temperature and partial pressure of water vapor are tightly regulated (T = 1–10 °C, pH₂O ≤ 18 Torr) (Levin et al., 2023). No confining windows are used, offering unrestricted geometry.
• 2D Heterostructure Liquid Cells: Recent advances utilize stacks of 2D materials (graphene, hBN) to form ultrathin liquid enclosures (~70 nm total, including a monolayer MoS₂) with high mechanical stability and minimal scattering (Clark et al., 2022). This enables true atomic resolution.

Auxiliary functions—such as in situ electrical biasing (three-electrode configuration for electrochemistry (Holtz et al., 2013)), on-chip microheating/cooling (Dumitraschkewitz et al., 2021), or integrated illumination for photochemical studies—are implemented through MEMS structures or custom holder designs.

Key setup attributes:

Parameter	Closed-cell	Open-cell ETEM	2D heterostructure
Window material	Si₃N₄, graphene	None	Graphene, hBN
Liquid thickness	50–800 nm	1–6 μm droplets	~70 nm
Pressure control	<atm (sealed)	pH₂O ≲ 18 Torr	Hermetic (ambient)
Sample load	Wet or dry	Dry (hydrated in situ)	Dry stack

2. Imaging Modalities and Experimental Protocols

LP-TEM is compatible with both conventional TEM and scanning TEM (STEM) modes. Bright-field, dark-field, and high-angle annular dark-field (HAADF) detectors are routinely used. Energy-dispersive X-ray spectroscopy (EDS) and, crucially, electron energy loss spectroscopy (EELS) provide complementary chemical sensitivity (Holtz et al., 2012, Holtz et al., 2013).

• Spatial resolution: Closed-cell geometries (liquid thickness ≲ 500 nm) routinely achieve ≲ 1 nm resolution in STEM (Grogan et al., 2010), with true atomic resolution (~0.1–0.2 nm) attainable in ultrathin 2D heterostructure cells (Clark et al., 2022). Open-cell ETEM achieves ≲ 1 nm in the best-case regions but is limited by thicker pathlengths (Levin et al., 2023).
• Temporal resolution: Frame rates ~0.5–2 Hz are typical for video capture (limited by readout and SNR), with sub-0.5 s/frame routine for dynamic events (Grogan et al., 2011).
• Beam dose management: Dose rates are minimized to suppress radiolysis, typically ≲10³–10⁴ e⁻ nm⁻² s⁻¹, using low accelerating voltages (20–300 kV) and short dwell times. For sensitive soft matter, lower dose rates (∼0.25–2.7 e⁻ nm⁻² s⁻¹) are favored (Kaczmarczyk et al., 29 Dec 2025).

Example closed-cell protocol: Assemble liquid cell, fill with aqueous nanoparticle suspension, seal; mount in TEM, evacuate; apply desired electrical or thermal stimuli, capture images or spectra over time.

Open-cell protocol: Deposit salt and specimen dry on standard grid, mount in cooled holder, insert into ETEM, raise water-vapor pressure to induce droplet formation, then image (Levin et al., 2023).

3. Quantitative Capabilities and Chemical Mapping

LP-TEM allows the extraction of quantitative kinetic, structural, and chemical information from dynamic, solvated systems:

• Particle dynamics: Brownian motion, aggregation rates, and trajectory statistics are captured at nanometer resolution. Mean-squared displacement (MSD) analysis yields diffusion coefficients (e.g., for Pt adatoms on MoS₂: D_liquid ≳ 0.25 nm²·s⁻¹ (Clark et al., 2022)).
• Colloidal self-assembly: In situ visualization of diffusion-limited aggregation (DLA), with real-time fractal dimension analysis matching bulk scattering data (Grogan et al., 2010).
• Electrochemical processes: LP-TEM liquid-flow holders with integrated electrodes enable operando monitoring of ion transfer in batteries. Valence EELS fingerprinting discriminates lithiated/delithiated phases (E_PEAK ≈ 5 eV for FePO₄; E_PEAK ≈ 6.2 eV for solvated LiSO₄) and maps spatial concentration of charge carriers (Holtz et al., 2013).
• Phase transitions in soft matter: Beam-induced and thermally driven mesophase transitions in nanoconfined liquid crystals can be temporally resolved and linked quantitatively to thermal modeling (ΔT ≈ 8–15 K in 30–40 s at D ≈ 1 e⁻ nm⁻² s⁻¹ for 8CB LC) (Kaczmarczyk et al., 29 Dec 2025).

EELS in Liquids

• Thickness measurement: Log-ratio method $t = \lambda \ln(I_t/I_0)$ relates the inelastic mean free path (λ) to zero-loss and total counts.
• Optical gap and plasmon peaks: Valence-loss (< 50 eV) spectra remain interpretable up to t/λ ≈ 7.
• Chemical mapping: For LiFePO₄: EFTEM at 5 eV images FePO₄ domains, at 6.2 eV images solvated Li⁺ (Holtz et al., 2013).

4. Beam–Sample Interactions and Artifacts

Beam–liquid interactions are central to LP-TEM experimental design:

• Radiolysis: Electron impact on water generates hydrated electrons, H·, ·OH, leading to byproducts (H₂, H₂O₂) and bubble formation. Bubble nucleation rate $J \sim \exp(-\Delta G^*/k_BT)$ depends on H₂ supersaturation (Grogan et al., 2012).
• Thermal effects: For soft matter (e.g., liquid crystals), beam heating, not radiolysis, typically dominates phase transition kinetics; the temperature rise is quantitatively modeled by $$\frac{\partial T}{\partial t} = \alpha \frac{\partial^2 T}{\partial x^2} + Q(x,t)/(\rho C_p)$$ (Kaczmarczyk et al., 29 Dec 2025).
• Spatial artifacts: Closed cell windows (Si₃N₄, graphene) can bulge, affecting sample thickness and resolution (Holtz et al., 2012). Open cell avoids window-induced artifacts.
• Dose mitigation: Pulsed/blanked beam, radical scavengers, and ultrathin liquid layers (< 200 nm) suppress bubble nucleation and minimize artifacts (Grogan et al., 2012).

5. Applications and Case Studies

LP-TEM offers access to a wide spectrum of dynamic in-liquid phenomena across diverse domains:

• Colloid and nanomaterial physics: Real-time imaging of gold particle aggregation and nanorod capillarity-induced alignment (Grogan et al., 2010, Grogan et al., 2011).
• Battery science: Nanoscale mapping of lithiation/delithiation and ion distributions in operando electrode/electrolyte systems (Holtz et al., 2013).
• Catalysis: Single-atom dynamics on 2D supports in liquid environments, providing direct evidence for solvation-modified diffusion and binding (Clark et al., 2022).
• Materials phase transitions: Observation of melting/solidification and spheroidization in alloy systems using MEMS-based heating (Dumitraschkewitz et al., 2021). Dynamic beam-induced phase transitions and defect dynamics within nanoconfined liquid crystals (Kaczmarczyk et al., 29 Dec 2025).
• Biological and soft matter systems: Studies of protein aggregation, gelation, and membrane interactions at nanometer resolution under true solution conditions (prospective).

6. Comparative Analysis of LP-TEM Geometries

Attribute	Open-Cell ETEM (Levin et al., 2023)	Closed-Cell (1010.32861212.1501)	2D Heterostructure (Clark et al., 2022)
Sample loading	Dry, standard grid	Wet or dry, device assembly	2D transfer, dry stack
Liquid thickness	μm-scale (droplets)	100–800 nm (fixed)	~70 nm (total, incl. windows)
Imaging artifacts	No windows	Window bulging, finite t	Minimal, atomic-scale
Spatial freedom	Unconstrained	Fixed cell geometry	2D/planar confinement
Max resolution	≲1 nm (best-case)	≲1 nm (optimal), ≳10 nm (t ≳ 1 μm)	Atomic (sub-Å) for single atoms
Environmental control	T: 1–10 °C, pH₂O limited	T: wide (−196 °C to >100 °C), P: multi-atm	Ambient, no P control
Major advantages	Flexible geometry	Robust, high throughput	True atomic imaging in liquid
Major limitations	Temp. stability, droplet size	Windows, liquid thickness	Fabrication complexity

7. Limitations, Challenges, and Future Developments

Challenges in LP-TEM are centered on:

• Beam-induced chemistry: Even at low doses, radiolysis and thermal effects can dominate observed behavior in sensitive materials or induce non-equilibrium states.
• Thickness/resolution trade-off: Achieving atomic to nanometer resolution requires total sample thicknesses ≲ 100 nm. Advanced windowless or graphene-window geometries minimize scattering but are not yet widely available.
• Sample environment flexibility: Open-cell ETEM affords unique geometry for external stimuli (illumination, magnetic fields), but requires precise humidity/temperature regulation (Levin et al., 2023).
• Quantitative chemical mapping: EELS-based chemical fingerprinting in thick liquid requires signal optimization (selecting sub-gap and sub-plasmon energy-loss windows). Multiple scattering and window artifacts limit conventional EELS for t/λ > 3 (Holtz et al., 2012).
• Soft matter and biological systems: Precise control of dose and temperature is critical to capture reversible, intrinsic phase dynamics without beam-induced artifacts (e.g., for nanoconfined LCs (Kaczmarczyk et al., 29 Dec 2025)).
• Future directions: Integration of in situ illumination, improved thermal management, advanced direct-electron detection, microfluidic mixing, and atomic layer encapsulation (graphene, hBN) are priorities to expand the scope and accuracy of LP-TEM for complex, multidimensional in-liquid phenomena.

Research continues to improve liquid-cell architectures, spectroscopic sensitivity, and methods for artifact mitigation, with the aim of enabling robust, artifact-free operando imaging and spectroscopy of dynamic processes in real liquids at atomic spatial and sub-second temporal resolution.