Published on 10 September 2024
This research summary is compiled from the recently published manuscripts “Quantification of reagent mixing in liquid flow cells for Liquid Phase-TEM” (Ultramicroscopy1) and “Toward sub-second solution exchange dynamics in flow reactors for liquid-phase transmission electron microscopy” (Nature Communications2). It describes the efforts of a team coordinated by Dr. Stefan Merkens to explore the capabilities of advanced flow control systems in the field of liquid–phase electron microscopy (LP-EM). The reported studies reflect a collaboration between the Electron Microscopy group at the Collaborative Research Center (CIC) nanoGUNE and the Colloidal Systems Chemistry group at the Center for Material Science (CFM) both in San Sebastian (Spain), as well as the Center for Sustainable Future Technologies@Polito of the Italian Institute for Technology in Turin (Italy).
The macroscopic properties of materials are defined by their structure, composition, and processes on the nanometer scale. Meticulous characterization with high spatial resolution is thus a prerequisite for increasing our understanding and optimizing technological applications. However, many relevant materials in areas including biochemistry, medicine, energy storage and conversion, and electro-catalysis function in liquid environments, which complicates their investigation with most standard imaging techniques, such as electron microscopy (EM).
Liquid-phase electron microscopy (LP-EM) is an emerging experimental technique that promises to overcome these limitations by allowing the monitoring of samples and processes in their native liquid environments with nanometer-scale resolution.
LP-EM relies on liquid cells (LCs), which are enclosures of ultra-thin (few tens to hundreds of nanometers) liquid layers between electron-beam transparent membranes, typically referred to as windows (compare Fig. 1). State-of-the-art setups enable for heating, cooling, electric biasing and exchange/mixing of liquids to be applied to the enclosed reaction solutions. LC setups can thus be perceived as chemical reactors that are inserted into the electron microscope to perform in situ experiments (lab-in-the-microscope approach).
During recent years, LP-EM has facilitated exciting, yet mostly qualitative, insights for a broad range of samples and processes in liquid environments. However, the main limitations that prevent the technique from becoming more quantitative are currently associated with:
Beyond the advantages of static LC designs for high-resolution imaging, many LP-EM systems are nowadays constructed as microfluidic devices. The diverse benefits associated with such flow setups include
Pumping systems (or flow controllers) are used to control the flow of reaction solution through the one or multiple inlets into the specialized sample holder and towards the viewing window where samples can be monitored. Considering the broad knowledge available in the field of microfluidics it is surprising that the associated benefits of flow setups are barely exploited. Current LP-EM flow experiments mainly suffer from a limited understanding and control of hydrodynamic properties as well as often primitive design of flow reactors, lack of established workflows and inaccessibility of quantitative insights due to unreliable data analysis.
It is important to realize that pressure plays a key role in flow experiments in general and in LP-EM in particular. Understanding and controlling its implications is hence detrimental to advance LP-EM. Foremost, the applied pressure difference between in- and outlet induces fluid flow through the channel configuration. The overall volumetric flow rate (in units of mm3/s) achieved at given pressure differences depends on the flow resistance of the flow device, while the local flow velocity (mm/s) varies significantly across the channel geometry. In addition, the internal pressure of the flow setup causes the ultra-thin viewing window to bulge outward. This effect is enhanced when inserting the setup into the high vacuum inside an electron microscope such that the achievable image resolution is drastically diminished.
For the aspects summarized above, there is currently a debate among LP-EM researchers whether pressure-driven flow controllers could be more suitable for LP-EM flow experiments compared to the widely applied syringe pump systems. Syringe pumps are widely used for flow control, but flow controller systems offer advantages in terms of reliability and precision. A pressure-driven flow controller enables fast settling times, high stability, and pulseless flow by utilizing piezoelectric technology instead of the mechanical pushing used in syringe pumps to control the flow. Pressure-driven flow control involves using gas input pressure within a sealed liquid tank to drive the liquid from the tank to a microfluidic device. In consequence, this approach is not limited by volume, is very precise, ultra-quick, and provides real–time adaptability.
While syringe pumps only affect the inlet pressure (with the outlet usually kept at ambient conditions), pressure-driven flow controllers promise additional control over the outlet, and thus the internal pressure inside the LC (compare Fig. 2). Operating at negative outlet pressures is expected to reduce bulging of the viewing window and therefore promises improved image quality. However, the pressurizing gas (e.g. N2 or Ar) that saturates the reaction solution in pressure-driven systems could interfere chemically with the experimental observations and/or may accumulate into gas bubbles inside the flow channel.
How sophisticated flow control systems can impact LP-EM experiments and what hydrodynamics properties are involved? This research summary addresses procedures to unravel the hydrodynamic properties of LP-EM flow reactors and explores the impact of sophisticated flow control systems to perform quantitative LP-EM experiments. It will demonstrate and explore
Two different experimental configurations were employed to control flow and supply reaction solution in LP-EM flow setups:
In both scenarios, the flow systems were set up by connecting the liquid in the inlet reservoirs (syringes and falcon tubes, respectively) with PEEK tubings to the LP-EM sample holder, and further to the outlet reservoir which was kept at either ambient (1) or controlled negative pressure (2).
Desired volumetric flow rates, Q, were applied either directly through the syringe pump (1), or by applying a pressure gradient, Δp, between the in- and outlet (2).
Flow sensors (MFS FS2 flow sensor) were included into the flow setup to feed volumetric flow rates back into the pressure controller unit. The photograph in Fig. 3 exemplarily illustrates a LP-EM flow system operated with an OB1 MK3 Flow controller.
An experimental workflow was established to calibrate the hydrodynamic properties of LP-EM flow systems:
First, we compared syringe and pressure–driven pumping systems to control fluid flow in LP-EM setups. It was confirmed that in addition to standard syringe pumps, LP-EM flow systems can be operated with pressure-driven control systems as also explored elsewhere.3 It was further confirmed that pressure-driven pumping can provide superior flow control compared to syringe pump systems (Fig. 4; compare Introduction).
A method to quantify the overall flow resistance of LP-EM flow systems was developed. It relies on the experimental setup depicted in Fig. 3. In brief, two flow sensors measured the volumetric flow rate for a given pressure difference, Δp, applied with the pressure controller.
Fig. 5 illustrates the Δp–Q-curves of two exemplary flow setups constructed from LP-EM sample holders. The overall flow resistance (Rtotal is calculated as the slope of the depicted data; the flow resistance of the sample holder () is obtained upon minor mathematical exercise:
where denote the resistances of the in- and outlet tubings connecting the solution reservoirs with the sample holder, respectively. The slope of the data in Fig. 2 reflects Rtotal of a setup with relatively low (red, 30 mbar/(µL/min)) and high (blue, 60 mbar/(µL/min)) flow resistance, respectively.
Fig. 5: Δp–Q-curves for different LP-EM flow setups. The slope of the plot corresponds to the overall resistance of the entire flow setups. The blue1 & red (unpublished) curves represent setups with high/low flow resistance, respectively; examples of such setups are depicted in Fig. 7a.
Double-inlet flow setups – independent of the flow control system – can be applied to mix two solutions. In certain setups, this can be achieved by running both inlets simultaneously. However, in flow setups with fixed viewing area (as is the case for LP-EM), one solution must be replaced by another to replicate mixing dynamics of batch chemistry scenarios. Experimentally this is achieved by switching the flow between the two inlets. We demonstrated that the mixing timescale in such scenarios can be quantified by tracking the concentration of a contrast agent over time. For in situ LP-EM measurements, phosphotungstic acid (PTA, 40 mM) is a convenient choice.1 Fig. 6 illustrates the calibration procedure in more detail. Notably, the experimentally determined mixing timescale was found to be drastically larger than those known for classical bench-top mixing experiments, e.g, involving magnetic stirring, for which mixing usually occurs within a few seconds.
Fig. 6: Image contrast variation method. a) The applied solution is alternated sharply between pure water and an electronically dense contrast agent. b), c) The induced intensity (b) changes of the transmitted electron beam can be converted into solute concentration. The characteristic time constants (delay time ∆t & decay time constant ) are obtained through exponential fitting. The red vs. the blue curve represents a setup with relatively slow (& more gradual) vs. fast solution replacement, respectively; examples of such setups are discussed in reference 1.
From the previous paragraph it became apparent that further optimization of the flow reactor is required to replicate real-world processes inside the LC. Here, we will focus on optimizing the LC design which is hence independent of the pumping systems. To facilitate the prototyping of such devices, a numerical model of the mass transport in the realistic channel geometry was implemented. In the summarized research articles, we excessively report on the details and how results from the two hydrodynamic calibration steps (see above) were crucial to validate the model.1,2
Supported by the numerical simulations, the reactor design was optimized. The geometrical adjustments are schematically illustrated in Fig. 7a (Readers interested in the working principle, the prototyping and/or fabrication procedure are referred to the original manuscripts.1,2).
Even more important is the drastic improvement of a series of hydrodynamic properties by (at least) 2 orders of magnitude including a reduced flow resistance, a reduced local flow velocity in the viewing window, and accelerated time constants of solution exchange, which are summarized in Fig. 7b-d. Reportedly, these improvements pave the way to develop LP-EM in a more quantitative technique for the characterization of nanoscale dynamics in a liquid environment.
Image contrast can be substantially improved by reducing the bulging of the membrane when working at reduced pressures as demonstrated in stationary measurements.4 We recently evaluated the benefit of conducting flow experiments in the negative pressure regime for nanoscale imaging. Fig. 8a & b depict electron microscopy images of a gold nanoparticle dimer obtained under continuous flow conditions with ambient and negative (-0.6 bar) outlet pressures applied, respectively. In the negative pressure regime, more features are apparent in the transmitted intensity profiles (Fig. 8c) indicating improved image resolution.
LP-EM was introduced as a burgeoning characterization technique for monitoring samples and processes in native liquid sample environments which has taken advantage from recent advances in flow reactor fabrication. Even though diverse stimuli to control the experimental conditions were recently unlocked, knowledge established in the field of microfluidics has been barely transferred to the field of LP-EM, which has slowed down its development.
General methods to calibrate and quantify the hydrodynamic properties of LP-EM flow devices were introduced which provide a much-need workflow for flow characterization studies. Based on the acquired knowledge, improved reactor designs were developed, and their properties and benefits were characterized. Particular attention was paid to flow control systems. The compatibility of pressure-driven pumping systems with LP-EM flow setups was demonstrated, and their advantages and disadvantages compared to syringe pump setups were evaluated.
The reported results indicate potential benefits of pressure-driven pumping systems for:
Stefan Merkens is a postdoctoral researcher in the Electron Microscopy group at the Collaborative Research Center (CIC) nanoGUNE BRTA in San Sebastian, Spain. He obtained his PhD degree in Physics of Nanostructures and Advanced Materials from the University of the Basque Country (UPV/EHU) in 2023. He has acquired interdisciplinary knowledge in fields comprising microfluidics, in situ sample characterization techniques, microfabrication and nanotechnology. His main research interest is directed towards the development of quantitative methods for the study of nanoscale dynamics in native liquid environment.
How can we help you?
Name*
Email*
Message
I hereby agree that Elveflow uses my personal data Newsletter subscription
We will answer within 24 hours
Get a quote
Newsletter subscription
By filling in your info you accept that we use your data.
Collaborations
Need customer support?
Serial Number of your product
Support Type AdviceHardware SupportSoftware Support
Subject*
Message I hereby agree that Elveflow uses my personal data Newsletter subscription