This review offers a succinct overview of the development of a vacuum-compatible microfluidic reactor system for analysis at the liquid vacuum interface (SALVI), and its diverse applications in in situ, in vivo, and in operando imaging of liquid surfaces as well as the air-liquid (a-l), liquid-liquid (l-l), and solid-liquid (s-l) interfaces in the past decade. SALVI is one of the first microfluidics-based reactors that has enabled direct analysis of volatile liquids in vacuum surface tools such as scanning electron microscopy (SEM) and time-of-flight secondary ion mass spectrometry (ToF-SIMS). Its integration into ambient and vacuum spectroscopy and microscopy is illustrated. Several applications are highlighted including (1) imaging nanoparticles in liquid using in situ SEM; (2) mapping the evolving l-l interface using in situ x-ray absorption spectroscopy and ToF-SIMS; (3) following complex a-l interfacial oxidation reaction products using in situ ToF-SIMS; (4) capturing biological interfaces of cells and microbes via in vivo multimodal and correlative imaging; and (5) monitoring the dynamic solid electrode and liquid electrolyte interface using in operando molecular imaging. Finally, outlook and recommendations are presented. Besides showing the holistic information volume obtained by real-time multiplexed imaging, this review intends to convey the importance of tool development in revolutionizing surface and interface analysis using vacuum platforms previously limited to solid surfaces. Microfluidics is manifested to be not limited to ambient conditions in many examples in this review. Moreover, fundamental interfacial phenomena underpinning mass and charge transfer can now be pursued in real time via innovated chemical imaging and spectroscopy.

Microfluidics is the science and engineering of manipulating and controlling fluids, usually in the range of microliters (10−6) to picoliters (10−12), in networks of channels with dimensions from tens to hundreds of micrometers in a device. Since its inception in the early 1990s, microfluidics has seen exponential growth in fundamental research and various real-world applications. There are several known advantages of microfluidics: low cost, ease in fabrication and integration, small size, and low power consumption. Generally, microfluidic devices are employed in ambient conditions. Much of the published papers utilizing microfluidics are in separations [e.g., liquid chromatograph–mass spectrometry (LC-MS) and lab-on-chip capillary electrophoresis (LOC CE)], cell analysis for point-of-care applications, and flow sensors.1 Recently, on-chip synthesis has also surged as a new way for novel material discovery. Besides mimicking different types of flows in a well-defined channel or chamber with known geometry, droplet fluidics is another regime that attracts a lot of attention.2 It has been widely used to study emulsions, for example, to form different kinds of core-shell structured droplets. One prominent application is controlled drug delivery.3 

Due to its small size, microfluidic devices can be integrated to a large analytical platform or used solely on its own. One of the most mature examples is LOC CE for separation of ions.4 When combined with other platforms, one of the most successful examples is the microchip LC-MS.5,6 Integration with other modern analytical instruments is also possible. One such appealing application is nuclear magnetic resonance (NMR) spectroscopy or magnetic resonance imaging (MRI).7,8 However, only a handful of efforts have used microfluidics in vacuum compared to numerous ambient applications using optical microscopy and spectroscopy.9–11 Superficially, it is controversial to study liquids in vacuum. Vacuum itself poses significant scientific and physical challenges to keep liquid in its natural form due to the inevitable phase change of liquid to vapor. Therefore, placing a liquid of high volatility in vacuum is not a rational approach based on the known physical principle. Previous research of liquid in vacuum mostly is limited to low volatility ionic liquids.12 For example, massive, highly charged glycerol cluster impact was used to produce images of intact biomolecules with few-micrometer lateral resolution and few-minute acquisition times.13,14 Although some liquids are investigated, the advantages offered by vacuum instrumentation, such as surface sensitivity, high spatial chemical mapping, and high structural information, are not possible using conventionally known approaches. Therefore, most surface sensitive vacuum instruments are still largely limited to solid studies to date.

The taboo is overcome by recent development of microfluidic cells integrable to vacuum. The most wildly successful example is the popular liquid cell in transmission electron microscopy (TEM).15,16 Its smashing prosperity demonstrates that sealing liquids in the microchannel is feasible in vacuum systems. Besides liquid cell TEM, our group has led the parallel development of a transferrable microfluidic cell system for analysis for liquid vacuum interface (SALVI), integrable to both ambient and vacuum instrumentation around the same time.17,18 Although liquid analysis in ambient conditions is much easier compared to vacuum, the ability to transfer a liquid specimen between vacuum and ambient platforms is crucial to permit multimodal and correlative imaging and results in more comprehensive characterization of a system under study. We have demonstrated the feasibility of liquid analysis using vacuum-based techniques such as scanning electron microscopy (SEM), time-of-flight secondary ion mass spectrometry (ToF-SIMS), and ultraviolet vacuum single photon ionization mass spectrometry (VUV SPI-MS) in addition to a suite of ambient spectroscopy and microscopy techniques including MRI, optical microscopy [e.g., structured illumination microscopy (SIM), confocal laser scanning microscopy (CLSM), and x-ray absorption spectroscopy (XAS)].7,17–23 Enabling the ability to analyze and image complex systems, including liquids in situ, in vivo, and in operando, presents great opportunities in scientific research and new discoveries using vacuum instrumentation.

This paper offers an overview of how to develop and integrate microfluidic reactors to vacuum instruments and make it feasible for ambient measurement simultaneously. Several application examples using primarily vacuum-based techniques, such as SEM, ToF-SIMS, and VUV SPI-MS, are presented to show how unique scientific problems involving liquid as a major component in a complex system can be studied using microfluidics-enabled chemical imaging and spectroscopy approaches. The review consists of Secs. IIIV: introduction of how to design and fabricate microfluidics, applications in several interesting and emerging research areas using vacuum imaging and spectroscopy techniques, and finally, outlook and recommendations.

First, we present a brief synopsis of microfluidic liquid cells used in vacuum platforms such as ToF-SIMS and SEM primarily. Then, we give examples of how SAVLI microfluidic reactor has been integrated into vacuum and ambient analytical platforms in this section. There are three possible strategies to introduce liquids to surface characterization tools such as ToF-SIMS, SEM, or TEM: (1) the use of a small aperture over the liquid in a microchannel to reduce vapor load on a vacuum system; (2) the use of electron-transparent materials [e.g., silicon nitride (SiN), silicon oxide (SiO2), graphene] to permit liquid imaging through the thin membrane; and (3) the use of a thicker layer of liquid to allow analysis using tools with bigger analysis depth. The choice depends primarily on analysis volume and instrument configuration.

In situ SEM imaging of nanoparticle in liquid is attractive in many applications. Using SEM to study particles in liquid is not a new idea. An enclosed wet-SEM sample holder was developed for imaging wet biological specimens.24 An inverted SEM was first developed with a detachable, open-culture dish for imaging cells through a SiN window.25 SiN and SiO2 are favorable materials because of their electron transparency in electron microscopy (EM).26 These two types of cells are fully enclosed, and both used membranes to protect the wet sample from vacuum. The first microfluidic liquid cell demonstrated in liquid scanning transmission electron microscopy (STEM) consists of two SiN windows.16,26 This microfluidic liquid cell design is the most widely used for today's liquid STEM and liquid TEM.

Figure 1 depicts details of SALVI microfluidic cell developed for in situ liquid analysis in vacuum instruments such as ToF-SIMS and SEM. The SAVLI cell is compatible with multiple vacuum instruments including SEM, ToF-SIMS, and VUV SPI-MS. Unlike other fully sealed systems used in EM liquid cells, our flow cell uses the SiN membrane that is partially open to high vacuum yet maintaining the high vacuum while the micrometer-sized open liquid surface is being probed. The details of the design and performance validation of SALVI are described in two companion papers.17,18 The design involves many issues dealing with the challenges of liquid study in vacuum. Among them, vacuum compatibility and temperature drop due to water vaporization are the two biggest ones. Our basic idea is to have the liquid under study flow in microchannel and open micrometer-sized apertures on the SiN membrane at the side of the channel facing to the probing beam (i.e., x-ray, primary ion beam, and electron beam) while exposing the liquid directly to the vacuum. Surface tension is used to withhold the liquid against the pressure difference; therefore, liquid does not spill into the vacuum. The aperture is critical for surface sensitive techniques like ToF-SIMS. Because secondary ions are weak, they need to be extracted close to the surface. One option is to use thinner membrane windows. However, that does not solve the problem of not being able to extract ions from the liquid surface effectively compared to using no window at all at the extraction location as in our approach. The analysis volume in SEM is much higher, and the aperture is not critical as illustrated in our recent27 and many others’ works.28,29

FIG. 1.

Fabrication of the vacuum-compatible microfluidic interface. (a) A schematic showing the assembly of the microfluidic block. (b) An optical image showing the fabricated microchannel with the reduced width as the detection area. (c) A cross-sectional view of the PDMS block showing the aperture. (d) A photo of the electrochemical SALVI cell. (a)–(c) Adapted with permission from Yang et al., J. Vac. Sci. Technol. A 29, 061101 (2011). Copyright 2011, American Vacuum Society.

FIG. 1.

Fabrication of the vacuum-compatible microfluidic interface. (a) A schematic showing the assembly of the microfluidic block. (b) An optical image showing the fabricated microchannel with the reduced width as the detection area. (c) A cross-sectional view of the PDMS block showing the aperture. (d) A photo of the electrochemical SALVI cell. (a)–(c) Adapted with permission from Yang et al., J. Vac. Sci. Technol. A 29, 061101 (2011). Copyright 2011, American Vacuum Society.

Close modal

The dimension of the microchannel is flexible, ranging from 100 to 500 μm wide depending on the instrument and the spot size of the probe beam.30 The depth of the channel ranges from 50 to 300 μm. The deeper channel is developed for culturing microbes and avoiding biofouling, for example.19–21 Soft lithography is used to fabricate microfluidic devices. Polydimethyl siloxane (PDMS) is used due to its ease in prototyping. Although PDMS has some interference by its distinctive signatures in SIMS mass spectra, they can be used for mass calibration.31–35 Irreversible bonding is used to ensure no leakage in vacuum.17,18,36

SALVI was primarily designed to enable liquid analysis using vacuum instrument such as ToF-SIMS and SEM. In this section, we will describe three examples of how a microfluidic reactor is integrated with SEM, ToF-SIMS, and VUV SPI-MS. Due to its small footprint, SALVI integration does not need considerable change when adapting to an instrument most of the time. Figures 2(a) and 2(b) depict two types of cells used for in situ liquid SEM: one is coated with gold (Au) and the other graphite to reduce the charging effect in imaging and analysis.37 More than one device can be installed depending on the SEM stage configuration. In an experiment, more than one micrometer-sized holes can be milled using the gallium focused ion beam (FIB). Most recently, we have also shown that when using secondary electron (SE) and backscattered electron (BSE) imaging with thinner SiN membrane (e.g., 30 nm), milling apertures is not necessary for in situ SEM imaging of particles because of the SEM analysis depth.37 It is worth noting that the high accelerating voltage should be adjusted and optimized for the element being analyzed. An example of boehmite particles was given in a recent paper.37 

FIG. 2.

SALVI integration to vacuum instrument. Gold (Au) (a) and graphite coated (b) SALVI device installed on the SEM stage; (c) SALVI installed on the ToF-SIMS sample stage; and (d) SALVI integrated to the ion optics plate of the VUV SPI-MS. (a) and (b) adapted with permission from Yu et al., Surf. Interface Anal. 51, 1325 (2019). Copyright 2019, Wiley Publishing and (d) adapted with permission from Komorek et al., Rev. Sci. Instrum. 89, 115105 (2018). Copyright 2018, AIP Publishing LLC.

FIG. 2.

SALVI integration to vacuum instrument. Gold (Au) (a) and graphite coated (b) SALVI device installed on the SEM stage; (c) SALVI installed on the ToF-SIMS sample stage; and (d) SALVI integrated to the ion optics plate of the VUV SPI-MS. (a) and (b) adapted with permission from Yu et al., Surf. Interface Anal. 51, 1325 (2019). Copyright 2019, Wiley Publishing and (d) adapted with permission from Komorek et al., Rev. Sci. Instrum. 89, 115105 (2018). Copyright 2018, AIP Publishing LLC.

Close modal

One of the benefits of microfluidics is its transferability. The device used in one analytical platform can be moved to another without further modifications. Figure 2(c) depicts the SALVI installed on the ToF-SIMS stage.32,34,35 Unlike SEM, in situ liquid SIMS would require the micrometer-sized holes to allow weak secondary ions to be extracted and detected.17,18,36 The finely focused primary ion beam is another factor that drives the need of micrometer-sized apertures as detection windows to reach submicrometer lateral spatial resolution using imaging mass spectrometry like ToF-SIMS. Similar to in situ SEM, a series of apertures as detection windows on the SiN membrane can be used in an in situ ToF-SIMS analysis of liquids. This allows measurement reproducibility and functions as a moving detection window to further reduce the beam effect.17,18,36

Another vacuum instrument we have integrated is VUV SPI-MS at the advance light source (ALS).23Figure 2(d) depicts the integration scheme. In this application, the SALVI device was coated with gold and used as one of the electrodes in the ion optics module. Ions are extracted after the photon beam is interacted with evaporated gas above the apertures in VUV SPI-MS. Two holes of 4 μm in diameter were milled by SEM-FIB to facilitate liquid evaporation in vacuum prior to device assembly and installation. The coated device can sustain an applied voltage of 1 kV or more to function as an electrode to enable volatile liquid studies using VUV SPI-MS.

Compared to vacuum integration, ambient applications are easier. The vacuum integrable microfluidic device is applicable in ambient measurements of liquids because of tolerance of high vapor pressures in the latter. Although the integration of microfluidic cells to ambient techniques is not as challenging as that in vacuum, the advantage of transferability among different instrument platforms permits correlative and multimodal imaging and spectroscopy and promotes multiscale measurements of complex systems containing liquid, thus achieving mesoscale imaging and approaching more understanding of the complexity of real-world phenomena. Here, we illustrate several examples in SALVI integration into optical spectroscopy and microscopy, MRI, and hard XAS. Optical microscopy and spectroscopy are used frequently in biology and microbiology to map living cells. Figures 3(a) and 3(b) depict the schematic of SALVI integration to CLSM (Ref. 21) and SIM,31,38 in which bacterial biofilms are cultured for in vivo imaging. In optical imaging, apertures are not necessary. The same device can also be used in multimodal imaging and spectroscopy offered by MRI as shown in Fig. 3(c). In this case, a thinner version of the SALVI was developed without any metal parts to be compatible with the magnetic fields and glass tubes used in MRI.7In situ imaging of liquids using XAS is demonstrated in Fig. 3(d) besides in vivo imaging.22 No major changes are made in the assembly and key design parameters of the SALVI devices among these integration efforts. It is almost plug and go, demonstrating the unique feature of transferability of microfluidics. It is worth noting that a central factor to consider is the beam size to facilitate easy and successful integration to different instruments. For example, a wider channel of 500 μm was selected for XAS, because the XAS beamline had a spot size of 300 μm. The microchannel was adjusted in dimension for the systems of interest when used for biological samples.

FIG. 3.

SALVI integration to ambient instrument for in vivo and in situ imaging: (a) a schematic of the microfluidic cell with biofilms cultured within and imaged with CLSM (b) and SIM (c); (d) a schematic showing integration to MRI; and (e) a schematic showing integration to the hard XAS spectroscopy of liquids. (a)–(c) adapted with permission from Ding et al., Analyst 144, 2498 (2019). Copyright 2019, Royal Society of Chemistry and Hua et al., Analyst 139, 1609 (2014). Copyright 2019, Royal Society of Chemistry, (d) adapted with permission from Renslow et al., Analyst 142, 2363 (2017). Copyright 2017, Royal Society of Chemistry, and (e) adapted with permission from Zheng et al., J. Phys. Condens. Matter 30, 18LT01 (2018). Copyright 2018, IOP Science.

FIG. 3.

SALVI integration to ambient instrument for in vivo and in situ imaging: (a) a schematic of the microfluidic cell with biofilms cultured within and imaged with CLSM (b) and SIM (c); (d) a schematic showing integration to MRI; and (e) a schematic showing integration to the hard XAS spectroscopy of liquids. (a)–(c) adapted with permission from Ding et al., Analyst 144, 2498 (2019). Copyright 2019, Royal Society of Chemistry and Hua et al., Analyst 139, 1609 (2014). Copyright 2019, Royal Society of Chemistry, (d) adapted with permission from Renslow et al., Analyst 142, 2363 (2017). Copyright 2017, Royal Society of Chemistry, and (e) adapted with permission from Zheng et al., J. Phys. Condens. Matter 30, 18LT01 (2018). Copyright 2018, IOP Science.

Close modal

In situ chemical imaging has shown to be a powerful tool in many research areas in recent years. This section is broken down into several subsections to give illustrations on how in situ chemical imaging and spectroscopy can be achieved in nanoparticles in liquids, l-l, a-l, living biological, and s-l interfaces enabled by the SALVI microfluidic cells.

First, in situ SEM imaging of particles in liquids is given as an example. In situ SEM imaging has been an appealing application in biological cells, ionic liquids, and particles in liquid.26 Besides major instrument development such as the atmospheric inverted SEM (Ref. 39) and the environmental SEM or ESEM,40 most people take the approach of liquid cell in SEM and TEM that requires vacuum compatibility of the liquid cells on the order of 10−6 Torr or better.26 SPI and QuantomiX have been the first two companies that offer commercial wet cells for SEM. BSE imaging was possible in these commercial cells using SiN, SiO2, or polyimide as the electron-transparent membrane. Notably, the development of two-dimensional (2D) materials has a positive influence on in situ SEM. The most significant recent example is the usage of graphene as another promising window material to surpass issues in liquid imaging in vacuum.28,29 Recently, graphene was also illustrated to provide improved ion imaging of slowly dried neurons and skin cells using ToF-SIMS.41 

Unlike other commercial or research grade liquid cells, our SALVI cell is the only one that can be used with and without a micrometer-sized aperture on the detection window, namely, the SiN membrane [see Figs. 1(a) and 1(b)]. We first depicted the EDX spectral determination of Au nanoparticles in a tris buffer coupled with validation of in situ liquid SIMS spectra.36 Since then, we have made technical improvement for in situ liquid SEM. Figures 4(a)4(c) show the SE imaging, BSE imaging, and the energy dispersive x-ray (EDX) spectroscopy spectrum of a mixture consisting of boehmite (γ-AlOOH), zinc oxide (ZnO), and silver (Ag) nanoparticles in the microchannel with a micrometer-sized aperture on the 100-nm SiN membrane in the high vacuum mode.37 These results illustrate that both SE and BSE imaging can be achieved using the SALVI cell. Moreover, EDX spectra can be used to determine elemental composition of mixed particles in liquid. The SAVLI microfluidic cell can be used as other liquid cells, namely, imaging on the intact SiN membrane without opening a viewing window or aperture. This is illustrated in Figs. 4(e)4(g), where uranium oxide (UO2) and magnetite (Fe3O4) nanoparticles are observed in the BSE and SE mode, respectively, via a thinner 30 nm SiN membrane in the low vacuum mode using a SALVI device.42 

FIG. 4.

In situ SEM imaging of particles in liquids using SALVI. (a) SE imaging, (b) BSE imaging, and (c) an EDX spectrum of a mixture of boehmite, ZnO, and Ag nanoparticles in liquid in the high vacuum mode using devices with holes on 100 nm SiN window in the high vacuum mode. In situ imaging of an empty channel (d), a BSE image of a UO2 particle in liquid, and images of Fe3O4 particles in liquid in SE (e) and BSE modes, respectively, under an intact 30 nm SiN window in the low vacuum mode. (a)–(c) adapted with permission from Yu et al., Surf. Interface Anal. 51, 1325 (2019). Copyright 2019, Wiley Publishing, (d)–(g) adapted with permission from Yu et al., in Proceedings of the 17th International High-Level Radioactive Waste Management (IHLRWM) Conference (American Nuclear Society, Knoxville, TN, 2019), pp. 311–315. Copyright 2019, American Nuclear Society.

FIG. 4.

In situ SEM imaging of particles in liquids using SALVI. (a) SE imaging, (b) BSE imaging, and (c) an EDX spectrum of a mixture of boehmite, ZnO, and Ag nanoparticles in liquid in the high vacuum mode using devices with holes on 100 nm SiN window in the high vacuum mode. In situ imaging of an empty channel (d), a BSE image of a UO2 particle in liquid, and images of Fe3O4 particles in liquid in SE (e) and BSE modes, respectively, under an intact 30 nm SiN window in the low vacuum mode. (a)–(c) adapted with permission from Yu et al., Surf. Interface Anal. 51, 1325 (2019). Copyright 2019, Wiley Publishing, (d)–(g) adapted with permission from Yu et al., in Proceedings of the 17th International High-Level Radioactive Waste Management (IHLRWM) Conference (American Nuclear Society, Knoxville, TN, 2019), pp. 311–315. Copyright 2019, American Nuclear Society.

Close modal

The l-l interface is ubiquitous in nature and it has crucial impacts on fundamental science and engineering applications. Microfluidics are developed for liquid studies in the micrometer scale. One of the immediate applications of SALVI coupled with in situ imaging and spectroscopy is to map the evolving l-l interface. Here, we present two recent examples of switchable ionic liquids (SWILs). The first example is to use in situ ToF-SIMS to reveal the coexisting two liquid phases in a carbon dioxide (CO2) capturing SWIL showing mesoscopic structural changes containing disparate regions of ionic and nonionic solvents, even though stoichiometry would indicate a 100% ionic solvent at full CO2 loadings.43 The second example is in situ hard XAS including x-ray absorption fine structure (XAF) and x-ray absorption near edge structure to reveal the structure of nanocrystalline green rust synthesized in the SWIL.30 

Figure 5(a) depicts the reaction scheme consisting of DBU (1,8-diazabicyclo[5.4.0]undec-7-ene) and 1-hexanol, a two-component SWIL. The size of spherical [DBUH] + C6H12OCO2 ionic clusters is approximately 5–20 nm, and it is key to reactivity and solvation environment, for example, in forming nanoparticles. The ion pair of [DBUH] + C6H12OCO2 could facilitate charge stabilization or polar immediate formation as important steps in nucleation in liquid in the presence of CO2. In situ ToF-SIMS has been employed to determine the l-l interface for the first time, offering three-dimensional (3D) distribution of the SWIL's solvent structure as a result of ToF-SIMS dynamic depth profiling.43Figure 5(b) shows a schematic of the SIMS depth profiling in the microfluidic channel of 200 μm wide and 300 μm deep covered by a 100 nm thick SiN membrane. The Bi3+ primary ion beam mills holes with 2 μm in diameter on the SiN membrane and images the liquid surface directly at the liquid-vacuum (l-v) interface during depth profiling [Fig. 5(c)].

FIG. 5.

(a) The DBU and 1-hexanol IL chemistry; (b) the schematic of the SALVI and in situ ToF-SIMS imaging based on dynamic depth profiling; (c) depth profile data corresponding to steps in (b); (d) a representative m/z spectrum in the positive mode; and (e) reconstructed 3D images showing coexistence of the polar and nonpolar constituents of SWIL after CO2 capture. (a)–(e) adapted with permission from Yao et al., Phys. Chem. Chem. Phys. 19, 22627 (2017). Copyright 2017, Royal Society of Chemistry.

FIG. 5.

(a) The DBU and 1-hexanol IL chemistry; (b) the schematic of the SALVI and in situ ToF-SIMS imaging based on dynamic depth profiling; (c) depth profile data corresponding to steps in (b); (d) a representative m/z spectrum in the positive mode; and (e) reconstructed 3D images showing coexistence of the polar and nonpolar constituents of SWIL after CO2 capture. (a)–(e) adapted with permission from Yao et al., Phys. Chem. Chem. Phys. 19, 22627 (2017). Copyright 2017, Royal Society of Chemistry.

Close modal

Figure 5(d) gives a representative liquid SIMS spectrum, and Fig. 5(e) shows 3D reconstructed superimposed images of ion clusters (i.e., m/z+ 349, 367, 451) and the neutral molecule attached ions (m/z+ 349), indicating that the SWIL components are not evenly distributed in the 100% loaded SWIL. As a comparison, the m/z+ 152 DBU 3D image of the solvent components shows a homogeneous distribution as one of the original ionic liquid components before CO2 capture. Similarly, negative cluster ions (i.e., m/z 213, 247, 257) are not evenly dispersed whereas the solvent peak of hexanol m/z 99 is. The ionic liquid after CO2 uptake forms a new liquid phase showing different spatial distributions. At the vacuum-liquid interface, the temperature drop was estimated to be ∼12 K if starting the experiment at room temperature. This temperature change will not induce freezing and have a slight effect on the diffusion rate of ions. The concentration change was estimated to be approximately 1.2–2.2 times of the initial concentration.17 Both positive and negative 3D images give direct evidence that two liquid phases coexist. The ionic liquid samples are also verified using the VUV SPI-MS at the chemical dynamic's beamline (9.0.2) at the ALS.43 Needless to say that these approaches are not limited to a particular system of interest. We recently also show that a zwitterionic SWIL can be studied using in situ ToF-SIMS.44 

XAS is useful in studying solvation environments and it can offer the local structure of the primary solvation spheres of the SWIL during catalysis or the formation (nucleation and growth) of metallic nanoparticles. The archetypical DBH:1-hexanol:CO2 solvent is investigated using in situ XAS in a SALVI microreactor in a recent study.30 In this work, iron acetate [Fe(OAc)2] is used as a model surrogate to produce a nano green rust. Green rusts are highly reactive minerals containing octahedral iron hydroxide layers with positive charges and interlayers with anions and water molecules. Green rusts are air sensitive, and it is difficult to characterize in its native liquid state, because they undergo changes during isolation and analysis. The microfluidics reactor offers the opportunity to study the solvation of green rust following synthesis.45 The integration of SALVI to the advanced XAS beamline was demonstrated earlier [Fig. 3(e)].22In situ spectroscopy provides an rich understanding of the solvation environment and crystal growth of catalyst solvation in the SWIL compared to postmortem analysis. Figures 6(a)6(d) show that the electronic structures of previously unidentified chemical precursors and key reaction intermediates for synthesized nanocrystalline green rust as a function of the solvation environment by means of different ionic or nonionic states of the SWIL using the density functional theory.

FIG. 6.

Structure scheme of (a) DBU molecule and (b) C6H12OCO22; N* and O* indicate the bonding sites for Fe; DFT optimized structures of (c) (DBU)2-Fe and (d) Sol-(b), green, purple, red, gray, and white spheres represent Fe, N, O, C, and H atoms, respectively. Comparison of Fe k-edge extended x-ray absorption fine structure χ(R) spectra for (e) 1 and (DBU)2-Fe and (f) 2 and Sol-(b). Insets show the respective x(k) plots. (a)–(e) adapted with permission from Zheng et al., Chem. Commun. 55, 11239 (2019). Copyright 2019, Royal Society of Chemistry.

FIG. 6.

Structure scheme of (a) DBU molecule and (b) C6H12OCO22; N* and O* indicate the bonding sites for Fe; DFT optimized structures of (c) (DBU)2-Fe and (d) Sol-(b), green, purple, red, gray, and white spheres represent Fe, N, O, C, and H atoms, respectively. Comparison of Fe k-edge extended x-ray absorption fine structure χ(R) spectra for (e) 1 and (DBU)2-Fe and (f) 2 and Sol-(b). Insets show the respective x(k) plots. (a)–(e) adapted with permission from Zheng et al., Chem. Commun. 55, 11239 (2019). Copyright 2019, Royal Society of Chemistry.

Close modal

The best matched model is [(DBU)2-Fe], corresponding to two DBU molecules coordinated to Fe [Figs. 6(c) and 6(e)], consisting of two distorted tetrahedrons. One N atom and two C atoms from the DBU bind to the Fe with an average bond distance of Fe–N and Fe–C of 1.9 and 2.1 Å, respectively. Sol-(b) [Figs. 6(d) and 6(f)] gives the best explanation to the observation in the fully loaded ionic [DBUH] + C6H12OCO2 solution. This structure has a distorted planar coordination, and it has one Fe atom bridged to two hexylcarbonate anions with four O atoms and an average Fe–O distance of ∼2.0 Å. Our results suggest that a solvated FeOOH-like iron species in the intermediate synthesis step could be the precursor of green rust. The Fe–Fe bonds may not form until the removal of CO2 from the SWIL. This finding has offered the first in situ analysis of the chemical speciation of green rust at various stages of synthesis ionic liquids by XAS, shedding new insights into the mechanism of green rust formation.30 

The a-l interface presents great challenges in fundamental surface science and has applications in many fields. For instance, the a-l interfacial chemistry plays an indispensable role in aqueous secondary organic aerosol (aqSOA) formation and multiphase chemistry that changes the evolution of atmospheric composition.46–48 The usefulness of in situ liquid ToF-SIMS in studying the a-l interfacial change of a model aqSOA system consisting of glyoxal and hydrogen peroxide (H2O2) is recently illustrated.32 Furthermore, two recent systematic studies show that aqSOA formation has a time dependence on ultraviolet (UV) and dark aging conditions.34,35 More importantly, dark aging is another important yet much underrepresented source of aqSOA.

Figures 7(a)7(d) illustrate the temporal evolution of representative carboxylic acids, oligomers, cluster ions, and water clusters in the negative ion mode in the UV aging conditions. Similarly, Figs. 7(e)7(h) depict the temporal evolution of selective carboxylic acids, 2-hydroxy-2-hydroperoxyethanal (HHPE m/z 91 C2H3O4), oligomers, cluster ions, and water clusters in the negative ion mode during dark aging. Key oxidation products including carboxylic acids and various oligomers are observed with UV irradiation. In addition, in situ liquid ToF-SIMS can provide unique observations of cluster ions and water clusters. Interestingly, oxidation products are also observed without UV irradiation. Glyoxal still undergoes oxidation by H2O2 without UV irradiation. Although reactions may be slower in dark aging than photochemical aging, carboxylic acids, organic peroxides, and other products are still detected.

FIG. 7.

Temporal evolution of (a) carboxylic acids, (b) oligomer, (c) cluster ions, and (d) water clusters in UV aging; and temporal evolution of (e) oxidation products, (f) oligomers, (g) cluster ions, and (h) water clusters in dark aging. All are in the negative ion mode consisting of glyoxal and H2O2. FA, formic acid; GA, glyoxylic acid; and HHPE, 2-hydroxy-2-hydroperoxyethanal. The error bars represent the standard deviation of three measurements. (a)–(c) adapted with permission from Zhang et al., Environ. Sci. Technol. 53, 10236 (2019).Copyright 2019, American Chemical Society and (d) and (e) adapted with permission from Zhang et al., npj Climate. Atmos. Sci. 2, 28(2019). Copyright 2019, Nature Publishing Company.

FIG. 7.

Temporal evolution of (a) carboxylic acids, (b) oligomer, (c) cluster ions, and (d) water clusters in UV aging; and temporal evolution of (e) oxidation products, (f) oligomers, (g) cluster ions, and (h) water clusters in dark aging. All are in the negative ion mode consisting of glyoxal and H2O2. FA, formic acid; GA, glyoxylic acid; and HHPE, 2-hydroxy-2-hydroperoxyethanal. The error bars represent the standard deviation of three measurements. (a)–(c) adapted with permission from Zhang et al., Environ. Sci. Technol. 53, 10236 (2019).Copyright 2019, American Chemical Society and (d) and (e) adapted with permission from Zhang et al., npj Climate. Atmos. Sci. 2, 28(2019). Copyright 2019, Nature Publishing Company.

Close modal

For example, the oxidation products (e.g., glyoxylic acid m/z 73 C2HO3, formic acid m/z 45 CHO2) form in the first 2 h of dark reactions. The chemical 2-hydroxy-2-hydroperoxyethanal (HHPE m/z 91 C2H3O4) is found in dark aging within the first 2 h. HHPE has been reported in recent field and laboratory studies, and it is not easy to detect because of its instability in solution. The in situ imaging of hydroperoxides first demonstrates the sensitivity of ToF-SIMS. Moreover, it suggests the oxidizing capacity of the aqueous surface in dark conditions. Hydroxyhydroperoxides (HHPs) have high Henry's law constants and, therefore, high potentials to partition into the aqueous phase. It has a prime implication on the uptake of volatile organic compounds and trace gases in clouds, fogs, or wet aerosols.

Aerosol particles vary from solid, semisolid, and liquid phases in atmospheric conditions. The aerosol mixing state is an important parameter in evaluating aerosol evolution. Both cluster ions and water clusters at the a-l interface can undergo reactions with ions or particles in the atmosphere and result in atmospheric transformation of the aerosol, affecting the aerosol's ability to act as cloud condensation nuclei (CCN) and moderate radiative forcing. Figure 8 depicts the evolution of different key species (i.e., water clusters, oxidation products, oligomers, and cluster ions) in 3D reconstructed images under UV aging conditions. Compared to spectral results, 3D images give direct visualization of oxidation products in space and provide unprecedented insight into the a-l interfacial change due to photochemical reactions. It is worth noting that the formation of water clusters and cluster ions are important steps in particle nucleation and growth in the atmosphere, promoting the CCN number in ambient air. In situ ToF-SIMS could provide more insight into the hygroscopic behavior of aerosols in a photochemical event (Fig. 8). The increased relative intensity of cluster ions suggests their growth and demonstrates that the organic compounds, either in the form of organic and water or organic and organic association, can condense on nucleated water clusters, leading to CCN formation and ultimately affecting aerosol interfacial composition and radiative forcing.34,35

FIG. 8.

Time evolution of key species in reconstructed 3D images in the negative ion mode. (a) Carboxylic acids formic acid m/z 45 CHO2, succinic acid m/z 117 C4H5O4, tartaric acid m/z 149 C4H5O6, (b) oligomers m/z 161 C5H5O6, 185 C4H9O8, 209 C7H11O7, (c) oligomers, m/z 285 C8H13O11, 329 C9H13O13; 421 C14H13O15, (d) cluster ions m/z 269 C7H9O11, 617 C19H21O23, 635 C19H23O24, and (e) large water clusters m/z 467 (H2O)25OH, 521 (H2O)28OH, 611 (H2O)33OH. 3D images are normalized to total ion intensities excluding interference peaks. The darker color indicates higher intensity, and the brighter color indicates lower intensity. (a)–(c) adapted with permission from Zhang et al., Environ. Sci. Technol. 53, 10236 (2019). Copyright 2019, American Chemical Society.

FIG. 8.

Time evolution of key species in reconstructed 3D images in the negative ion mode. (a) Carboxylic acids formic acid m/z 45 CHO2, succinic acid m/z 117 C4H5O4, tartaric acid m/z 149 C4H5O6, (b) oligomers m/z 161 C5H5O6, 185 C4H9O8, 209 C7H11O7, (c) oligomers, m/z 285 C8H13O11, 329 C9H13O13; 421 C14H13O15, (d) cluster ions m/z 269 C7H9O11, 617 C19H21O23, 635 C19H23O24, and (e) large water clusters m/z 467 (H2O)25OH, 521 (H2O)28OH, 611 (H2O)33OH. 3D images are normalized to total ion intensities excluding interference peaks. The darker color indicates higher intensity, and the brighter color indicates lower intensity. (a)–(c) adapted with permission from Zhang et al., Environ. Sci. Technol. 53, 10236 (2019). Copyright 2019, American Chemical Society.

Close modal

An important research area we have enabled and expanded using SALVI as a microreactor to culture biological samples is systems biology, namely, in vivo bioimaging of microbes and single cells.7,19–21,31,38 Here, we illustrate a recent example of the toxicity study of living microbial biofilms using correlative SIM and ToF-SIMS. We show the ability to probe single cell membrane using SALVI and complemented with SIM. The feasibility of SALVI to study bacterial biofilms in the hydrated state is described earlier (Fig. 3).19,21 Two studies show promising results of the biofilm matrix using in vivo molecular imaging including MRI, ToF-SIMS, and CLSM.7,31,38

Bulk approaches [e.g., LC-MS, infrared (IR) spectroscopy, and NMR] are often used to study the bacteria cell and its associated extracellular polymeric substance (EPS) or the matrix. However, they could not determine whether the observed effect is from the bacteria or the EPS after the microbes are dried or extracted. In situ multimodal imaging, including ToF-SIMS, gives the spatial distribution of key biofilm components and potentially reveals how toxic chemicals affect the biofilm and its matrix.

Figures 9(a)9(d) show the SIM imaging of the wild type Shewanella biofilm and hyper-adhesive mutant hyper-adherent mutant CP2-1-S1 (CP) biofilm with and without exposure to toxic chemicals. The responses of the biofilm to the toxic Cr (VI) are compared using in vivo SIM and ToF-SIMS [Fig. 9(e)]. In situ molecular imaging reveals that more larger water clusters are present in the mutant CP biofilms possibly due to the formation of hydrophobic polymers. The CP biofilm has more polysaccharides, an indispensable component for biofilm formation. The thicker and more defensive CP biofilm has more active riboflavin, implying that more quorum sensing (QS) molecules are more responsive to the environmental perturbation by a toxic heavy metal ion Cr (VI). In comparison, important biological insights would be missed when only dried biofilms are analyzed. Our imaging results show that water clusters are indicative of the biofilm's cohesiveness in the CP mutant case. The thicker mutant biofilm has higher counts of fatty acids, riboflavin, and QS molecules corresponding to higher sustainability.38 

FIG. 9.

CLSM imaging of Shewanella oneidensis MR-1 wild type [(a) and (b)] and CP mutant [(c) and (d)] with and without chromium Cr (VI) contamination, respectively; and (e) reconstructed ToF-SIMS 3D images of MR-1 and CP mutant biofilms with and without Cr (VI) reduction, respectively. (a)–(c) adapted with permission from Ding et al., Analyst 144, 2498 (2019). Copyright 2019, Royal Society of Chemistry.

FIG. 9.

CLSM imaging of Shewanella oneidensis MR-1 wild type [(a) and (b)] and CP mutant [(c) and (d)] with and without chromium Cr (VI) contamination, respectively; and (e) reconstructed ToF-SIMS 3D images of MR-1 and CP mutant biofilms with and without Cr (VI) reduction, respectively. (a)–(c) adapted with permission from Ding et al., Analyst 144, 2498 (2019). Copyright 2019, Royal Society of Chemistry.

Close modal

Lipids could provide the structural support and affect the viscosity of the EPS, and they are major components of the biofilm's EPS, encompassing small molecules such as fatty acids and their derivatives. Fatty acids can be viewed as a crucial player in biofilms. Higher fatty acid signals in the CP biofilm suggest a higher lipid production and biofilm formation capability. This is verified with the SIM observation, that is, the mutant is denser. SIMS findings show more cyclic lipids in the mutant, suggesting that the CP biofilm is more robust because of the higher cyclic lipid content.38 Our multimodal and correlative approach using microfluidics has extended bioimaging to include vacuum analytical platforms and increase molecular level information of microbes and single cells significantly. More applications of microfluidics and label free bioimaging employing ToF-SIMS are anticipated in future research.

The s-l interface represents grand scientific challenge and engineering significance in surface and interface analysis using vacuum techniques. Bulk approaches are not suitable to probe the s-l interface. Of particular interest is the solid electrode-liquid electrolyte interface that is ubiquitous in batteries, energy storage materials, electrocatalysis, and corrosion. Many techniques are used to study the chemistry at the s-l interface with liquid and gas and sometimes under pressure. The commonly used analytical tools are scanning tunneling microscopy, atomic force microscopy, reflection-absorption infrared spectroscopy, Raman spectroscopy, second-harmonic generation, sum frequency generation, surface extended XAF, x-ray powder diffraction, voltammetry, laser-induced desorption, and scanning electrochemical microscopy induced desorption, to name a few.49 However, real-time chemical speciation at the s-l interface as potential is applied to the electrochemical cell is difficult to achieve. We have developed an electrochemical microreactor based on the SALVI principle (or the SALVI E-cell) and soft lithography incorporating three electrodes into a microfluidic device and enabled in operando chemical imaging of the s-l interface to elucidate the intermediate formation using ToF-SIMS.33,50 It is worth mentioning that in situ TEM has quickly become a tool to study the electrochemical interface using microfluidic liquid cells, giving another great example of microfluidics changing liquid analysis using vacuum technology. Both cases show successful applications of microfluidics to permit measurements previously deemed impossible. Compared to liquid cell TEM, our approach offers real-time molecular imaging. This is unobtainable using EM, because EM can only provide elemental information.

Figures 10(a)10(e) show the schematic of the SALVI E-cell. More than one aperture can be drilled with the primary ion beam (e.g., Bi+, Bi3+) in the positive or negative ion mode as seen in Fig. 10(g) during in operando analysis while maintaining a reasonably high vacuum (<5 × 10−7 mbar) in the main chamber. Three operation modes are possible as depicted in Fig. 10(h), including (1) hold at a potential to collect SIMS spectra; (2) hold at a potential to do depth profiling; and (3) sweep potentials to acquire depth profiles dynamically. A well-known electrochemical system containing gold working electrode (WE) as well as platinum reference electrode (RE) and counter electrode (CE) is used to illustrate feasibility. A dilute electrolyte of potassium iodide (KI) with a concentration in the millimolar is used as the liquid electrolyte. In operando ToF-SIMS is effectively a 2D analysis offering electrochemical signals from an electrochemical station and mass to charge ratios (m/z) from the ToF-SIMS at the dynamic s-l interface simultaneously [Fig. 10(h)]. Since the iodine electrochemistry is reversible, cyclic voltammograms are collected in this experiment.

FIG. 10.

(a) Cross section of the SALVI E-cell design; (b) overview of the RE and CE; (c) overview of the WE; (d) a finished E-cell; (e) the E-cell assembled on the ToF-SIMS stage; (f) the schematic of the detection SiN surface prior to dynamic depth profiling; (g) multiple holes drilled for positive and negative in operando ToF-SIMS; and (h) voltage and count profiles of the three operation modes. (a)–(e) adapted with permission from Liu et al., Lab Chip 14, 855 (2014). Copyright 2014, Royal Society of Chemistry.

FIG. 10.

(a) Cross section of the SALVI E-cell design; (b) overview of the RE and CE; (c) overview of the WE; (d) a finished E-cell; (e) the E-cell assembled on the ToF-SIMS stage; (f) the schematic of the detection SiN surface prior to dynamic depth profiling; (g) multiple holes drilled for positive and negative in operando ToF-SIMS; and (h) voltage and count profiles of the three operation modes. (a)–(e) adapted with permission from Liu et al., Lab Chip 14, 855 (2014). Copyright 2014, Royal Society of Chemistry.

Close modal

The s-l interface is dynamic in this measurement. The information depth of static ToF-SIMS is known to be a few nanometers from published emission depth measurements.17,51 Diffusion is an important factor for in situ ToF-SIMS, and the obtained SIMS spectra are expected to contain information from the electrode-electrolyte interface of a few nanometers thick and from the bulk electrolyte liquid of up to several micrometres.49 

Figures 11(a)11(c) give an example of dynamic depth profiling of the electrode-electrolyte interface while holding at different potentials. Figures 11(d)11(f) show corresponding reconstructed SIMS spectra from the shaded areas along the depth profiles. Distinct mass signatures in the SIMS spectra acquired at different potentials during depth profiling of the s-l interface give the direct evidence of the intermediate species formed. For example, key species (i.e., m/z 127 I, 143 IO, 159 IO2, 175 IO3, 254 I2, 324 AuI, 381 I3, 451 AuI2, and 521 Au2I) have different abundances at different potentials. At a low potential, adlayer species such as AuI2, Au2I, and AuI are more dominant, indicating their immediate formation at the Au WE. As the oxidation potential increases, the intensities of oxidation products such as I2, I3, IO, IO2, and IO3 grow higher, indicating that they are chain reaction products following the formation of the gold adlayers. Figures 11(g)11(j) show two sets of results in the potential sweeping mode. The potential and current temporal profiles are depicted in Figs. 11(g) and 11(h) from the simultaneous electrochemical station measurement. The dynamic s-l interfacial speciation changes are reflected in the SIMS spectral temporal profiles in Figs. 11(i) and 11(j). These real-time results offer the first molecular level reaction products and intermediates at different stages of the redox cycle as a result of electron transfer. Additionally, the spatial distribution of the s-l intermediate species and products can be reconstructed in 3D based on ToF-SIMS imaging. Furthermore, more intermediate species are discovered according to the in operando liquid ToF-SIMS measurements, expanding the fundamental knowledge of the seemingly well-known iodine electrochemistry.

FIG. 11.

(a)–(c) Dynamic probing the s-l interface at different voltages using in operando liquid ToF-SIMS. (d)–(f) Reconstructed SIMS spectra corresponding to the shaded sputtering time showing speciation differences at different potentials. Temporal profiles of potential (g), current (h), and intermediate species (i) and (j) during three consecutive sweeps. (a)–(j) adapted with permission from Yu et al., Chem. Commun. 52, 10952 (2016). Copyright 2016, Royal Society of Chemistry.

FIG. 11.

(a)–(c) Dynamic probing the s-l interface at different voltages using in operando liquid ToF-SIMS. (d)–(f) Reconstructed SIMS spectra corresponding to the shaded sputtering time showing speciation differences at different potentials. Temporal profiles of potential (g), current (h), and intermediate species (i) and (j) during three consecutive sweeps. (a)–(j) adapted with permission from Yu et al., Chem. Commun. 52, 10952 (2016). Copyright 2016, Royal Society of Chemistry.

Close modal

Our results show that it is imperative to study the dynamic s-l interface using advanced in operando chemical imaging techniques to provide molecular level understanding and enrich kinetics in electrochemistry at the solid electrode-liquid electrolyte interface. New tool development is indispensable for capturing the short-lived intermediate species and elucidating the reaction mechanism at the s-l interfaces. Although our recent work has focused on ToF-SIMS applications, the microfluidic cells are not limited to one instrument platform. The inherent transferability of microfluidics will undoubtedly empower traditionally vacuum-based methods and technology to study liquid surfaces and interfaces involving liquids in conditions simulating real-world phenomena.

This epitome gives an overview of how to design and integrate microfluidics to enable liquid analysis using vacuum instrumentation previously limited to solid analysis and expand the vacuum instrument and technology to interfacial studies involving liquids, that is, a-l, l-l, s-l, and l-v interfaces. Several examples are given in enabling vacuum imaging and spectroscopy including in situ SEM, ToF-SIMS, and VUV SPI-MS. Undoubtedly, vacuum-compatible microfluidics can be applied to ambient platforms as well. In fact, most known microfluidics are developed for ambient measurements as noted earlier. Although crossing the barrier from ambient conditions to vacuum to study liquids is antagonistic in nature, it is possible as demonstrated in many recent efforts from my group in this review. Specifically, several cases in optical microscopy (e.g., CLSM, SIM), NMR (or MRI), and XAS are illustrated to show the wealth of information that can be obtained using a transferrable microfluidic sample holder, SALVI. Transferability due to the device small footprint is a feature inherent to microfluidics and it should be a big selling point in using microfluidics in modern vacuum analysis of surfaces and interfaces. Novel material synthesis is equally cardinal in new instrument and capability development. Electron-transparent thin film, namely, SiN, is quite critical in the microfluidic device design and its compatibility with multiple analytical platforms in the case of SALVI. Novel 2D material development should boost the application of microfluidics and nanofluidics in the next generation of surface and interface analysis.

The initial success of our SALVI device in expanding vacuum analysis and technology marks the beginning of a new era of surface and interface spectroscopy and microscopy involving the condensed phase combining technical breakthrough in fundamental physics, chemistry, engineering, and advancement in material synthesis and microfabrication. The successful illustrations of how microfluidics can be employed to study new materials, material synthesis, evolution of reaction products, characterization of soft material interfaces, mass transfer, and electron transfer at the various interfaces involving liquids signify that vacuum instrumentation is not limited to solid surface and solid interfaces anymore. Advancement in novel material synthesis combined with microfluidics and possibly nanofluidics that are integrated with microelectromechanical systems (MEMS) would open a new passport to vacuum-based analysis in the future.

There are multiple aspects to consider for future development. First, we should expand microfluidics integration to more vacuum platforms. We have only shown integration with a few vacuum instrumental platforms; more can be achieved using similar design or with minimal change. One of the instruments of the most imminent interest is x-ray photoelectron spectroscopy (XPS) and the emerging ambient pressure x-ray photoelectron spectroscopy (APXPS). APXPS has seen rapid development in the past few decades and it has shown various studies of vapor-solid interfaces using differentially pumped electronic lens systems.52,53 However, studying liquid is still not as easy as solid materials. Several groups have shown development of near-ambient pressure x-ray photoelectron spectroscopy (NAPXPS) to study moderately volatile liquids.54,55 However, microfluidics has not been used as part of this instrument development. XPS, APXPS, and NAPXPS offer complimentary information to SEM and ToF-SIMS, permitting liquid analysis to vastly enhance quantitative measurements of surface and interface and provide valence information. Another appealing instrument is soft XAS and x-ray emission spectroscopy. These capabilities present exceptional opportunities to expand vacuum capabilities using microfluidics to study buried interfaces and could potentially offer a tremendous impact on the fundamental science of mass and charge transfer and will remarkably change how s-l interfaces can be studied using instruments that are limited to solid or only available to a few people who developed the novel reactors for liquid analysis currently.

Second, we should develop new capabilities to expand the range of conditions that we can model in microfluidic reactors. Currently, most of our experiments are conducted at room temperature. More features would significantly accelerate the whelm of conditions we can simulate using a versatile microreactor. For example, temperature control, more complex device patterns as well as generation, manipulation, and monitoring of complex flows would allow investigations of interfacial phenomena more like the real-world systems, such as single cell sorting, material phase change, membrane dynamics, catalytic reactions, corrosion, or electrode degradation. In addition, utilization of different fabrication approaches beside photolithography would broaden the types of vacuum-compatible reactors. One emerging approach for fast prototyping is 3D printing. The challenge in using a 3D printed device in vacuum lies in material choice and availability. For example, materials that are not off gassing in vacuum would be preferable than porous polymers for analysis of liquid in vacuum. The marriage of MEMS and microfluidics will mark a new generation of novel reactors suitable for in vacuo studies involving the condensed liquid phase similar to the general trend of the application of microfluidics in other industries.

Finally, a natural question to ponder is whether we have reached the limit of experimental systems we could explore using vacuum instrument and technology. This is a rather open-ended topic. While the physics and chemistry of liquids and interfaces involving liquids are not immediately germane to vacuum conditions, we firmly believe that the main physical constraints can be overcome by advancement in material science, development in engineering, and most importantly our curiosity and courage to triumph over obstacles presented by nature.

The author is grateful for the support from Pacific Northwest National Laboratory (PNNL) with funding from the Department of Energy (DOE) Nuclear Energy Spent Fuel Waste Science Technology (SFWST) program. The authors acknowledge the PNNL Nuclear Processing Science Initiative (NPSI) Laboratory Directed Research and Development (LDRD) project. PNNL is operated by Battelle under Contract No. DE-AC05-76RL01830.

Nomenclature
2D

two-dimensional

3D

three-dimensional

Ag

silver

aqSOA

aqueous secondary organic aerosol

Au

gold

a-l

air-liquid

AFM

atomic force microscopy

ALS

advance light source

APXPS

ambient pressure x-ray photoelectron spectroscopy

BSE

backscattered electron

CCN

cloud condensation nuclei

CE

counter electrode

CLSM

confocal laser scanning microscopy

CO2

carbon dioxide

CP

hyper-adherent mutant CP2-1-S1

Cr

chromium

CV

cyclic voltammogram

DBU

1,8-Diazabicyclo[5.4.0]undec-7-ene

DFT

density functional theory

DOE

Department of Energy

EDX

energy dispersive x-ray

EM

electron microscopy

EPS

extracellular polymeric substance

ESEM

environmental scanning electron microscopy

EXAFS

extended x-ray absorption fine structure

FA

formic acid

Fe3O4

magnetite, iron (II, III) oxide

Fe(OAc)2

iron acetate

FIB

focused ion beam

GA

glyoxylic acid

HHP

hydroxyhydroperoxide

HHPE

2-hydroxy-2-hydroperoxyethanal

H2O2

hydrogen peroxide

IHLRWM

International High-Level Radioactive Waste Management

IR

infrared

KI

potassium iodide

LC-MS

liquid chromatograph–mass spectrometry

LDRD

Laboratory Directed Research and Development

l-l

liquid-liquid

LOC CE

lab-on-chip capillary electrophoresis

l-v

liquid-vacuum

NMR

nuclear magnetic resonance

μm

micrometer

m/z

mass to charge ratio

mM

milimolar

MEMS

microelectromechanical system

MR1

Shewanella oneidensis MR-1

MRI

magnetic resonance imaging

NPSI

nuclear processing science initiative

PDMS

polydimethyl siloxane

PNNL

Pacific Northwest National Laboratory

QS

quorum sensing

γ-AlOOH

boehmite

RE

reference electrode

SALVI

system for analysis at the liquid vacuum interface

SE

secondary electron

SEM

scanning electron microscopy

SFG

sum frequency generation

SFWST

spent fuel waste science technology

SHG

second-harmonic generation

SIM

structured illumination microscopy

SiN

silicon nitride

SiO2

silicon oxide

s-l

solid-liquid

STEM

scanning transmission electron microscopy

STM

scanning tunneling microscopy

SWILs

switchable ionic liquids

TEM

transmission electron microscopy

ToF-SIMS

time-of-flight secondary ion mass spectrometry

UO2

uranium oxide

UV

ultraviolet

VOC

volatile organic compound

VUV SPI-MS

ultraviolet vacuum single photon ionization mass spectrometry

WE

working electrode

WT

wild type

XAF

x-ray absorption fine structure

XANES

x-ray absorption near edge structure

XAS

x-ray absorption spectroscopy

XES

x-ray emission spectroscopy

XPS

x-ray photoelectron spectroscopy

XRD

x-ray powder diffraction

ZnO

zinc oxide

1.
N.
Convery
and
N.
Gadegaard
,
Micro Nano Eng.
2
,
76
(
2019
).
2.
L.
Shang
,
Y.
Cheng
, and
Y.
Zhao
,
Chem. Rev.
117
,
7964
(
2017
).
3.
S.
Damiati
,
U. B.
Kompella
,
S. A.
Damiati
, and
R.
Kodzius
,
Genes
9
,
103
(
2018
).
4.
Y.
Temiz
,
R. D.
Lovchik
,
G. V.
Kaigala
, and
E.
Delamarche
,
Microelectron. Eng.
132
,
156
(
2015
).
5.
A.
Oedit
,
P.
Vulto
,
R.
Ramautar
,
P. W.
Lindenburg
, and
T.
Hankemeier
,
Curr. Opin. Biotechnol.
31
,
79
(
2015
).
6.
X.
Wang
,
L.
Yi
,
N.
Mukhitov
,
A. M.
Schrell
,
R.
Dhumpa
, and
M. G.
Roper
,
J. Chromatogr. A
1382
,
98
(
2015
).
7.
R. S.
Renslow
,
M. J.
Marshall
,
A. E.
Tucker
,
W. B.
Chrisler
, and
X. Y.
Yu
,
Analyst
142
,
2363
(
2017
).
8.
H.
Ryan
,
S.-H.
Song
,
A.
Zaß
,
J.
Korvink
, and
M.
Utz
,
Anal. Chem.
84
,
3696
(
2012
).
9.
I.
Rodríguez-Ruiz
,
T. N.
Ackermann
,
X.
Muñoz-Berbel
, and
A.
Llobera
,
Anal. Chem.
88
,
6630
(
2016
).
10.
V.
Bianco
,
B.
Mandracchia
,
V.
Marchesano
,
V.
Pagliarulo
,
F.
Olivieri
,
S.
Coppola
,
M.
Paturzo
, and
P.
Ferraro
,
Light Sci. Appl.
6
,
e17055
(
2017
).
11.
P.
Paiè
,
R.
Martínez Vázquez
,
R.
Osellame
,
F.
Bragheri
, and
A.
Bassi
,
Cytometry A
93
,
987
(
2018
).
12.
E. F.
Smith
,
F. J. M.
Rutten
,
I. J.
Villar-Garcia
,
D.
Briggs
, and
P.
Licence
,
Langmuir
22
,
9386
(
2006
).
13.
J.
Zhang
,
K.
Franzreb
, and
P.
Williams
,
Rapid Commun. Mass Spectrom.
28
,
2211
(
2014
).
14.
J.
Zhang
,
K.
Franzreb
,
S. A.
Aksyonov
, and
P.
Williams
,
Anal. Chem.
87
,
10779
(
2015
).
15.
C.
Zhu
 et al,
Nat. Commun.
9
,
421
(
2018
).
16.
H.-G.
Liao
and
H.
Zheng
,
Annu. Rev. Phys. Chem.
67
,
719
(
2016
).
17.
L.
Yang
,
X.-Y.
Yu
,
Z.
Zhu
,
T.
Thevuthasan
, and
J. P.
Cowin
,
J. Vac. Sci. Technol. A
29
,
061101
(
2011
).
18.
L.
Yang
,
X.-Y.
Yu
,
Z.
Zhu
,
M. J.
Iedema
, and
J. P.
Cowin
,
Lab Chip
11
,
2481
(
2011
).
19.
X.
Hua
,
M. J.
Marshall
,
Y.
Xiong
,
X.
Ma
,
Y.
Zhou
,
A. E.
Tucker
,
Z.
Zhu
,
S.
Liu
, and
X.-Y.
Yu
,
Biomicrofluidics
9
,
031101
(
2015
).
20.
X.
Hua
 et al,
Integr. Biol.
8
,
635
(
2016
).
21.
X.
Hua
 et al,
Analyst
139
,
1609
(
2014
).
22.
J.
Zheng
,
W.
Zhang
,
F.
Wang
, and
X.-Y.
Yu
,
J. Phys. Condens. Matter
30
,
18LT01
(
2018
).
23.
R.
Komorek
,
B.
Xu
,
J.
Yao
,
U.
Ablikim
,
T. P.
Troy
,
O.
Kostko
,
M.
Ahmed
, and
X. Y.
Yu
,
Rev. Sci. Instrum.
89
,
115105
(
2018
).
24.
S.
Thiberge
,
O.
Zik
, and
E.
Moses
,
Rev. Sci. Instrum.
75
,
2280
(
2004
).
25.
H.
Nishiyama
,
M.
Suga
,
T.
Ogura
,
Y.
Maruyama
,
M.
Koizumi
,
K.
Mio
,
S.
Kitamura
, and
C.
Sato
,
J. Struct. Biol.
169
,
438
(
2010
).
26.
X.-Y.
Yu
,
B.
Liu
, and
L.
Yang
,
Microfluid. Nanofluid.
15
,
725
(
2013
).
27.
X.-Y.
Yu
,
J.
Yao
,
Z.
Zhu
, and
E.
Buck
, in 17th International High-Level Radioactive Waste Management Conference (IHLRWM 2019) (
American Nuclear Society
,
Knoxville
,
TN
,
2019
), p.
311
.
28.
D. J.
Kelly
,
M.
Zhou
,
N.
Clark
,
M. J.
Hamer
,
E. A.
Lewis
,
A. M.
Rakowski
,
S. J.
Haigh
, and
R. V.
Gorbachev
,
Nano Lett.
18
,
1168
(
2018
).
29.
W. N.
Yang
,
Y. N.
Zhang
,
M.
Hilke
, and
W.
Reisner
,
Nanotechnology
26
,
315703
(
2015
).
30.
J.
Zheng
,
X.-Y.
Yu
,
M.-T.
Nguyen
,
D.
Lao
,
Y.
Zhu
,
F.
Wang
, and
D. J.
Heldebrant
,
Chem. Commun.
55
,
11239
(
2019
).
31.
Y.
Ding
,
Y.
Zhou
,
J.
Yao
,
C.
Szymanski
,
J.
Fredrickson
,
L.
Shi
,
B.
Cao
,
Z.
Zhu
, and
X.-Y.
Yu
,
Anal. Chem.
88
,
11244
(
2016
).
32.
X.
Sui
,
Y.
Zhou
,
F.
Zhang
,
J.
Chen
,
Z.
Zhu
, and
X.-Y.
Yu
,
Phys. Chem. Chem. Phys.
19
,
20357
(
2017
).
33.
J.
Yu
,
Y.
Zhou
,
X.
Hua
,
S.
Liu
,
Z.
Zhu
, and
X.-Y.
Yu
,
Chem. Commun.
52
,
10952
(
2016
).
34.
F.
Zhang
,
X.
Yu
,
X.
Sui
,
J.
Chen
,
Z.
Zhu
, and
X.-Y.
Yu
,
Environ. Sci. Technol.
53
,
10236
(
2019
).
35.
F.
Zhang
,
X.
Yu
,
J.
Chen
,
Z.
Zhu
, and
X.-Y.
Yu
,
npj Clim. Atmos. Sci.
2
,
28
(
2019
).
36.
L.
Yang
,
Z.
Zhu
,
X.-Y.
Yu
,
E.
Rodek
,
L.
Saraf
,
T.
Thevuthasan
, and
J. P.
Cowin
,
Surf. Interface Anal.
46
,
224
(
2014
).
37.
X.-Y.
Yu
,
B.
Arey
,
S.
Chatterjee
, and
J.
Chun
,
Surf. Interface Anal.
51
,
1325
(
2019
).
38.
Y.
Ding
,
Y.
Zhou
,
J.
Yao
,
Y.
Xiong
,
Z.
Zhu
, and
X.-Y.
Yu
,
Analyst
144
,
2498
(
2019
).
39.
K.
Hirano
 et al,
Ultramicroscopy
143
,
52
(
2014
).
40.
A. M.
Donald
,
Nat. Mater.
2
,
511
(
2003
).
41.
S. Y.
Lee
,
H.
Lim
,
D. W.
Moon
, and
J. Y.
Kim
,
Biointerphases
14
,
051001
(
2019
).
42.
X.-Y.
Yu
,
J.
Yao
,
Z.
Zhu
, and
E.
Buck
, in
International High-Level Radioactive Waste Management Conference
(American Nuclear Society,
Knoxville
,
TN
,
2019
).
43.
J.
Yao
 et al,
Phys. Chem. Chem. Phys.
19
,
22627
(
2017
).
44.
X.-Y.
Yu
,
J.
Yao
,
D. B.
Lao
,
D. J.
Heldebrant
,
Z.
Zhu
,
D.
Malhotra
,
M.-T.
Nguyen
,
V.-A.
Glezakou
, and
R.
Rousseau
,
J. Phys. Chem. Lett.
9
,
5765
(
2018
).
45.
D.
Lao
,
R. K.
Kukkadapu
,
L.
Kovarik
,
B. W.
Arey
,
D. J.
Heldebrant
, and
S. K.
Nune
,
Curr. Inorg. Chem.
6
,
92
(
2016
).
46.
Y.
Fu
,
Y.
Zhang
,
F.
Zhang
,
J.
Chen
,
Z.
Zhu
, and
X.-Y.
Yu
,
Atmos. Environ.
191
,
36
(
2018
).
47.
X.
Sui
,
Y.
Zhou
,
F.
Zhang
,
Y.
Zhang
,
J.
Chen
,
Z.
Zhu
, and
X.-Y.
Yu
,
Surf. Interface Anal.
50
,
927
(
2018
).
48.
K.
Artyushkova
,
D. R.
Mullins
,
L.
Gregoratti
, and
X.-Y.
Yu
,
Surf. Interface Anal.
50
,
911
(
2018
).
49.
X.-Y.
Yu
,
Curr. Opin. Electrochem.
6
,
53
(
2017
).
50.
B.
Liu
,
X.-Y.
Yu
,
Z.
Zhu
,
X.
Hua
,
L.
Yang
, and
Z.
Wang
,
Lab Chip
14
,
855
(
2014
).
51.
A.
Wucher
,
S.
Sun
,
C.
Szakal
, and
N.
Winograd
,
Anal. Chem.
76
,
7234
(
2004
).
52.
H.
Bluhm
 et al,
J. Electron Spectrosc. Relat. Phenom.
150
,
86
(
2006
).
53.
D. E.
Starr
,
Z.
Liu
,
M.
Hävecker
,
A.
Knop-Gericke
, and
H.
Bluhm
,
Chem. Soc. Rev.
42
,
5833
(
2013
).
54.
T.
Roychowdhury
,
S.
Bahr
,
P.
Dietrich
,
M.
Meyer
,
A.
Thißen
, and
M. R.
Linford
,
Surf. Sci. Spectra
26
,
014025
(
2019
).
55.
A.
Broderick
,
Y.
Khalifa
,
M. B.
Shiflett
, and
J. T.
Newberg
,
J. Phys. Chem. C
121
,
7337
(
2017
).

Dr. Yu's research focuses on in situ mesoscale chemical imaging of soft materials in atmospheric, biology, energy, and material sciences using microfluidics as well as their impacts on the environment and human health. She was interested in chemistry as a teenager growing up with parents who were scientists at the Chinese Academy of Sciences before retirement. Dr. Yu acquired her undergraduate training in materials sciences and engineering and went to graduate school at the University of Michigan, Ann Arbor. She started with inorganic chemistry under the guidance of Professor Stephen Lee. Shortly after, she decided to pursue physical chemistry with Professor John Barker. She obtained her doctoral degree with a dissertation focused on studying kinetics of complicated photochemical chain reactions in the aqueous phase in 2001.

After graduation, she immediately took a postdoctoral position in the Atmospheric Sciences Division at the Brookhaven National Laboratory (BNL). This position offered her the first experiences working at different field campaigns and learning the gears of developing new analytical tools for real-time observations using both ground and airborne platforms. It also helped her transition from a laboratory kineticist to a researcher with broader experiences in both the laboratory and field settings.

Because of her experiences and technical expertise of in situ characterization of aerosols using novel analytical approaches, Dr. Yu got a postdoctoral fellowship with Professor Jeff Collett and Professor Sonia Kreidenweis at the Colorado State University (CSU) right after the BNL appointment. She conducted many field campaigns to study aerosol properties while at CSU. She also learned microfluidic fabrication in Professor Chuck Henry's group and led a couple of projects developing microfluidic devices for in situ aerosol measurements. Due to her outstanding performance, she was promoted to a senior staff scientist (equivalent of associate research professor) in 2004.

An opportunity came up at the Pacific Northwest National Laboratory (PNNL) for a senior scientist in Atmospheric Chemistry requiring extensive field and laboratory experiences and she joined the Atmospheric Sciences and Global Change Division in 2006. She has recently transferred to the Energy Processes and Materials Division in 2019 to continue her career as a senior scientist at PNNL. Since joining PNNL, her work portfolio has expanded. For instance, she mastered full scale and scale model engineering approaches to study effluent mixing with radioactivity in stacks. This is of importance in nuclear energy production and facility qualifications to meet EPA air quality requirements. She continued some work on field observations of changes of atmospheric composition to understand their impacts on global climate change via aerosol-cloud interactions, particle evolution during transport, and transformation due to multiphase chemistry. She not only participated in multiple large-scale campaigns but also led several field studies. Dr. Yu also has become the technical leader of applied toxicology for the DOE emergency management program. She has been the chair of the DOE chemical exposure working group and led the development of protective action criteria (PAC) and chemical mixture methodology (CMM) for consequence assessment of toxic health effects. She uses her expertise in applied toxicology in mixture exposure assessment for multiple sponsors.

A crucial transition after becoming a staff scientist at PNNL was that she was able to go back to her root in physical chemistry and chemical engineering by working with colleagues in the DOE Basic Energy Science (BES) Condensed Phase and Interface Science Program. She led an important breakthrough and invented a microfluidic interface that enabled studying liquid under high vacuum conditions using SEM and ToF-SIMS since 2007. The versatile reactor was named SALVI or system for analysis at the liquid vacuum interface. SALVI has generated multiple issued US patents and attracted commercialization interests. Dr. Yu was the lead principal investigator to win the prestigious R&D 100 award for SALVI as a new analytical instrumentation in 2014 and the Federal Laboratory Consortium Technology transfer award for their outstanding effort in commercialization in 2015.

Moreover, this invention enables direct chemical imaging of liquid surfaces and interfaces involving volatile liquids by vacuum sensitive techniques and opens a new avenue to obtain molecular level understanding of interfacial phenomena at the liquid-vacuum, air-liquid, liquid-liquid, and solid-liquid interfaces. Dr. Yu's group and her collaborators have published over 50 papers reporting new findings and applications using in situ chemical imaging using microfluidics in respectful journals such as Lab on a Chip, Biomicrofluidics, Microfluidic and Nanofluidic, Analyst, Analytical Chemistry, Journal of Physical Chemistry Letter, Chemical Communication, PCCP, Integrative Biology, Environmental Science and Technology, and Journal of Vacuum Science and Technology A. The peer reviews generally rank the papers high, commenting the technique as an exciting, significant, and innovative technical breakthrough. Dr. Yu's group is continuing their effort in enabling more analytical capabilities using microfluidics and expanding imaging and spectroscopic studies in material interfaces as exemplified in this review. It is anticipated that more papers and hopefully inventions associated with in situ, in vivo, and in operando imaging and spectroscopy using vacuum science and technology will come from her group in the future.