1 Introduction

Low-cost, earth-abundant, highly efficient and stable electrocatalysts are crucial in the application of electrochemical techniques in modern industries, particularly for renewable energy conversion, fuel cell, energy storage and organic synthesis systems [1,2,3,4,5,6] in which, in general, they can be applied to either functionalize electrode surfaces or act directly as electrode surfaces to enhance reaction rates, efficiencies and selectivity. In recent decades, enormous improvements have been achieved in terms of electrocatalyst performances in a variety of applications such as water splitting, hydrogen fuel cells and CO2 reduction, mainly due to the detailed knowledge of electrocatalysts that has been achieved with the assistance of advanced characterization facilities and methods as well as theoretical calculations. As a result, the atomic and molecular-level understanding of electrocatalyst surfaces and corresponding electrochemical interfaces is indispensable for the rational optimization and design of electrocatalysts in which the knowledge of surface compositions, structures, defects and electronic structures as well as surface adsorbate configurations is key to identifying and increasing surface active sites and enhancing intrinsic activities [7, 8].

Scanning tunneling microscopy (STM) in ultra-high-vacuum (UHV) conditions has played a significant role in the field of surface sciences because it can allow for the study of surface structures at the atomic level to provide detailed information concerning the surface electronic structure of materials as well as the configuration, dynamics, dissociation and adsorption/desorption of surface adsorbates, all of which are vital in the application of materials such as noble metals, metal alloys and transition metal oxides in electrocatalytic or heterogeneous catalytic reactions [9, 10]. In recent years, researchers have also combined electrochemical cells (EC cell) with UHV-STM through the introduction of transfer systems to effectively avoid contamination and oxidation during sample exchange [11,12,13,14], thus allowing for the electrochemical measurement of clean and well-defined electrocatalyst surfaces without exposure to air and promoting the in-depth correlation between surface-related properties and electrocatalytic performances. And because STM can also be operated under atmospheric pressures and aqueous environments to study material surfaces at the atomic resolution in addition to UHV conditions, STM can also be applied in the in situ investigation of electrochemical reactions through the direct identification and monitoring of active sites during electrocatalytic reactions with the STM tip acting as a working electrode, which is known as electrochemical STM (EC-STM) [15,16,17,18].

Based on these characteristics, this review will present the recent developments in the use of ex situ and in situ STM to study the catalytic mechanisms of electrocatalysis and identify the surface active sites of electrocatalysts. Information concerning the use of STM for heterogeneous catalysts that are not relevant to electrochemical applications is not included in this review; however, it can be found in other review articles [7, 19,20,21]. Specifically, this review will present state-of-the-art studies that use ex situ UHV-STM and UHV-STM/EC cell combined systems in the investigation of electrocatalyst surfaces with the aim of identifying surface active sites and revealing corresponding electrochemical properties, with studies on different electrocatalysts being classified based on composite type. In addition, recent developments involving the use of EC-STM in solution environments mainly on metal surfaces are discussed with corresponding electrochemical interfaces and electrode processes being emphasized. Furthermore, this review will introduce emerging STM techniques, including high-speed EC-STM, scanning noise microscopy and tip-enhanced Raman spectroscopy.

2 UHV-STM and UHV-STM Combined with EC Cells

2.1 Noble Metal and Alloy-Based Electrocatalysts

Noble metals as well as corresponding alloys and oxides are traditional electrocatalysts in many electrochemical applications due to excellent catalytic performances with examples including Pt-based electrocatalysts for hydrogen evolution reactions (HER), oxygen reduction reactions (ORR) and hydrogen oxidation reactions (HOR) as well as Ir/Ru oxides for oxygen evolution reactions (OER). Here, surface and interface engineering approaches, including the control of surface structures and facets, alloying with other metals, construction of heterogeneous structures, etc., have been widely adopted not only to reduce noble metal loading, but also to enhance the intrinsic activity of surface active sites [22,23,24,25,26,27]. And in these attempts, both UHV-STM and EC-STM (Sect. 3) can play a vital role in the diagnosis of surface structures, morphologies and surface active sites as well as providing valuable assistance in the construction of structural models for theoretical simulations.

The alloying of noble metals is exemplified by the alloying of Pt with 3d transition metals such as Fe, Co, Ni and Ti that can result in significant improvements in ORR activity as compared with pure Pt catalysts [28,29,30,31]. These enhanced electrocatalytic activities can usually be explained by the modified surface electronic property and activity of active sites and are known as the d-band center theory. Here, corresponding clean single-crystal surfaces prepared in UHV environments are suitable for the study of these modified surface electronic properties and their correlation with electrocatalytic activity.

For example, in the study of noble metal alloy Pd3Fe(111) obtained after annealing in UHV at 1200 K and 1000 Kin which the samples were quickly transferred to an EC cell with the assistance of a small high-pressure antechamber/load lock, Yang et al. [32] reported that the sample annealed at 1200 K exhibited the best ORR performance as compared with Pd(111) and Pt(111) based on electrochemical ORR testing and was able to attribute this performance to the highly active dispersed Fe adatoms based on STM results (Fig. 1a, b). With the same setup, Yang et al. [33, 34] also conducted the nanofaceting of Re(11\({\bar{\text{2}}}\)1) as induced by carbon in UHV and used STM (Fig. 1c) to propose a surface with three-sided nanopyramids exposing (01\({\bar{\text{1}}}\)1), (10\({\bar{\text{1}}}\)1) and (11\({\bar{\text{2}}}\)0) facets (Fig. 1d). This surface was subsequently adopted as a template for monolayer Pt nanostructures to allow for higher activities toward HER than Pt(111) in acidic media. Furthermore, Wang et al. [34] used STM to study the nitrogen-induced nanofaceting of Re(11\({\bar{\text{2}}}\)1) consisting of two-sided ridges formed by (13\({\bar{\text{4}}}\)2) and (31\({\bar{\text{4}}}\)2) facets.

Fig. 1
figure 1

a STM image of a Pd3Fe(111) surface obtained after annealing at 1200 K showing Pd adatoms on the surface. b STM image of a Pd3Fe(111) surface obtained after annealing at 1000 K showing a flat alloy surface with bright spots being identified as Fe atoms. Inset is the corresponding low-energy electron diffraction (LEED) pattern. Reprinted with permission from Ref. [32]. Copyright © 2011 WILEY-VCH Verlag GmbH&Co. KGaA, Weinheim. c STM image of a nanofaceted C/Re(1\({\bar{\text{2}}}\)11) surface. d Hard spear model of a single pyramid for the nanofaceted C/Re(11\({\bar{\text{2}}}\)1) surface. Reprinted with permission from Ref. [33]. Copyright © 2012 American Chemical Society. e Atomic-resolution STM image of a Pt(111) surface. f Atomic-resolution STM image of a Ni/Pt(111) surface. Upper left insets in (e, f) are corresponding RHEED patterns. g Large-scale STM image of an as-prepared Ni/Pt(111) surface. h Large-scale STM image of a Ni/Pt(111) surface after 1000 potential cycles. Lower right insets in (e–h) are corresponding line profiles along the arrows. Reprinted with permission from Ref. [35]. © 2013 The Electrochemical Society

Full size image

In a further study, Todoroki et al. [35] deposited Ni onto the surface of Pt(111) and obtained eight times higher ORR activities as compared with clean Pt(111), but reported that activities decreased steeply during potential cycling between 0.6 and 1.0 V to a two-time higher activity after 1000 potential cycles. Here, these researchers combined UHV-STM with molecular beam epitaxy (MBE) and an EC cell in a stainless steel glove box to avoid the exposure of their sample to air before and after cyclic voltammetry (CV) tests to allow for the examination of corresponding surface structures and the linking of surface structure to electrochemical activity. As a result, these researchers were able to propose based on the obtained STM topologies before and after 1000 potential cycles (Fig. 1g, h) that the underlying Ni-induced specific surface strain and the electronic state of the Ni/Pt(111) alloyed surface were the contributors to the remarkable ORR activity. With the same setup, the ORR activity of bimetallic surfaces prepared under UHV, including Pt/Au(hkl) [36], Pt/Ni/Pt(111) [37], Pt/Ni/Pt(110) [38], Pt/Pd(111) [39] and Pt/Pt25Ni75(111) [40], has also been investigated.

Todoroki et al. [41] also examined the electrochemical CO2 reduction performance of epitaxially grown Ni and Pt on Cu(111) using STM and found that the growth of Pt monolayers on Cu(111) can increase total faradaic efficiencies to be comparable to that of Pt(111) and that the growth of Ni monolayers on Cu(111) can increase the selectivity for CH4 production above that of clean Ni(111), demonstrating that both the activity and selectivity of electrochemical CO2 reduction using Cu-based alloys can be modified through surface alloying by affecting the formation of adsorbed CO intermediates.

In another study, Todoroki et al. [42] applied UHV-STM to assist in the study of low-index Au(hkl) surfaces using online electrochemical mass spectrometry in which Au(110) exhibited superior electrochemical CO2 reduction to CO as compared with Au(100) and (111) surfaces. Using the same experimental facilities, these researchers also demonstrated that with a 0.1 monolayer of Co on Au(110), the selectivity and partial current density of electrochemical CO2 reduction to CO can be enhanced as compared with clean Au(110), whereas with a 0.1 monolayer of Sn on Au(110), the selectivity of HER to H2 can be increased with the selectivity of CO2 reduction to CO decreased [43].

2.2 Transition Metal Oxide-Based Electrocatalysts

Earth-abundant transition metal oxides (TMOs) are appealing alternatives to noble metal (Ir and Ru)-based electrocatalysts for electrochemical OER in alkaline media [44, 45]. Great efforts have also been made in the development of TMOs for other electrocatalytic applications, especially as bifunctional electrocatalysts for HER/OER and OER/ORR reactions [46,47,48,49,50]. To increase the electrocatalytic activity of these TMOs, strategies such as the engineering of defects, the construction of mixed oxides and the introduction of noble metal nanostructures have been adopted in the attempt to increase active site numbers and enhance activities. And as compared with the surface of noble metals, the characterization of TMO surfaces using STM is more difficult due to lower conductivities and limited access to corresponding single crystals.

Over the years, the surface of TMOs as well the adsorption and reaction of corresponding adsorbates such as molecules of water, oxygen and CO on TiO2 has been intensively studied [51, 52]. Because the intrinsic conductivity of most TMOs is poor, however, corresponding defects, heterogeneous centers containing noble metals or the addition of conductive supports is indispensable in their application as electrocatalysts. Here, STM can be combined with ex situ electrochemical measurements for performance validation. For example, Feng et al. [53] combined STM with ex situ electrochemical measurements to study the reduced pure single crystal of rutile TiO2(110) and found that corresponding electrocatalytic HER activities in alkaline media relied on Ti3+ ions to promote electron transfer and hydrogen desorption.

Instead of single-crystal TMOs, ultra-thin oxide islands or films are more accessible on metals produced through epitaxial growth and can be used as model catalysts in the study of catalytic mechanisms and interfacial (TMO/metal) effects in catalysis [54, 55]. As a result, a variety of metal oxides and TMOs on metal surfaces has been successfully fabricated and studied using STM [56,57,58,59,60,61,62,63], and among these, cobalt oxides have received broad attention due to excellent OER performances [64,65,66]. For example, Fester et al. [67] were able to find through the in situ dosing of water molecules in an UHV chamber that exposed Co atoms on cobalt oxide nanoparticle edges rather than oxygen vacancies on basal planes acted as OER active sites to dissociate water and involved the hydroxylation of cobalt oxides. Here, atom-resolved STM snapshots of single-layer CoO islands and bilayer CoO islands on Au(111) (Fig. 2a, b) revealed that new hydroxyl oxygen atoms with point-like features first formed at the edges and subsequently diffused to the basal plane during in situ water exposure in the UHV chamber and that at higher water dosages, an ordered √3 × √3 R30° structure of hydroxyls formed on the surface of CoO (Fig. 2c), which was also supported by density functional theory (DFT) calculations (Fig. 2d).

Fig. 2
figure 2

a–c In situ STM snapshots and saturation level of hydroxylation. a, b Image sequences from atom-resolved STM video recorded at the initial stage of water exposure on single- (a) and double- (b) bilayer CoO nanoislands. Insets are corresponding high-magnification STM images. c STM image of a heavily hydroxylated single-bilayer island. d DFT-calculated hydroxylation structure of the CoO basal plane as a function of H chemical potential and pressure at room temperature. Reprinted with permission from Ref. [67]. Copyright © 2017, Springer Nature. e STM image of Co–O bilayer nanoislands on Au(111). f Structural model corresponding to Co–O bilayer nanoislands. g STM image of a 0.3 monolayer cobalt oxide/Au(111) surface after HER. i Cathodic scan of the sample in (g) in the HER potential region as compared with clean Au(111). h STM image of a Co–O/Au(111) surface after OER with corresponding CV curves in (j). Reprinted with permission from Ref. [68].© 2018 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim

Full size image

Fester et al. [68] further used these atomically defined CoO nanoislands on Au(111) (Fig. 2e, f) to study interfacial effects in electrocatalytic reactions through a CV measurement and STM combined system in which a preparation chamber was used as a transfer system to avoid the exposure of samples to air. As a result, these researchers were able to find that Co–O/Au(111) exhibited greatly enhanced HER and OER activities in alkaline media as compared with pure Au(111) (Fig. 2g, i). In addition, STM images of Co–O/Au(111) after HER CV tests (Fig. 2h) indicated the significant agglomeration of Co–O islands due to reordering, whereas after OER CV tests, the morphology of Co–O islands showed only slight changes, confirming the OER activity enhancement of Au-supported cobalt oxide. These results also demonstrated that optimal enhancements occurred at below one monolayer of oxide due to the active stabilizing effect of Au(111), which will diminish if oxide films are too thick.

2.3 MoS2-Based Electrocatalysts

Nanoparticulate MoS2 electrocatalysts have also exhibited high activities in terms of electrochemical HER in acidic media that are competitive with state-of-the-art Pt-based electrocatalysts. Here, Hinnemann et al. [69] identified the active sites of nanoparticulate MoS2 electrocatalysts as the edge of MoS2 nanoparticles through the evaluation of calculated adsorbed H free energies and proposed a model electrocatalyst system composed of MoS2 nanoparticles and graphite supports that was successfully fabricated and characterized using UHV-STM (Fig. 3). This model was subsequently validated through numerous experimental results [70, 71], and in particular, Jaramilloa et al. [70] fabricated a series of MoS2 nanoparticles on Au(111) with different ratios of basal plane sites to edges sites and used UHV-STM for in-depth characterizations. These researchers also transferred the well-defined samples of MoS2 on Au(111) into EC cells without exposure to air for electrochemical characterization, thus allowing for the easy correlation of the lower overpotential of the sample obtained at 440 °C to the larger number of active sites due to smaller island sizes and more edge sites and demonstrating that the edge sites were the active sites of MoS2 in electrochemical HER.

Fig. 3
figure 3

Reprinted with permission from Ref. [69]. Copyright © 2005 American Chemical Society

Left: Polarization curve for HER on Pt, Daihope C support and MoS2 as cathodes. Right: STM image of MoS2 nanoparticles on graphite with enlargement in the inset.

Full size image

Based on this fundamental understanding, great efforts have been made in corresponding material synthesis to increase the ratio of edge sites to basal sites for higher activities [72,73,74] in which molybdenum sulfide clusters including [Mo3S4]4+, [Mo3S13]2− and [Mo2S12]2− with more S atoms arranged in a structural motif similar to the edges of MoS2 have demonstrated outstanding HER activities with lower overpotentials and excellent turnover frequencies (TOF) [75,76,77]. In terms of the use of STM, Kibsgaard et al. [76] successfully applied STM in the study of [Mo3S13]2− to acquire the atomic-resolution structure of single clusters on a highly orientated pyrolytic graphite (HOPG) substrate and reported that STM can play a vital role in the identification of intrinsic structures, which can further be applied in the calculation of TOF by providing the number of active sites.

The surfaces of MoS2 and many other transition metal dichalcogenides (TMDs) have also been widely studied using UHV-STM due to interesting properties in terms of both catalysis and physics. And although most of them, prepared either in UHV through MBE or through the in situ cleavage of single crystals, have yet been directly studied using electrochemical measurements, existing knowledge involving surface defects [78,79,80], dopants [81], structural phases and phase boundaries [82, 83] as well as hybridizations and heterostructures [84,85,86] is of importance in the revealing of corresponding surface active sites and the elucidation of electrocatalytic mechanisms.

2.4 Metal-Supported Organic Molecule-Based Electrocatalysts

In addition to solid crystal materials, many organic molecules are also promising candidates in numerous catalytic applications due to easy accessibility and chemical and thermal stability along with structural tunability at the molecular level [87,88,89,90,91,92]. As a result, metal-supported organic molecules including metallo-phthalocyanines (MPcs), metallo-porphyrins (MPps) and metal–organic coordination networks (MOCNs) have been widely studied as model electrocatalysts using UHV-STM to reveal structures as well as molecule–molecule and molecule–substrate interactions [93,94,95,96,97,98,99,100,101].

For example, Sedona et al. [99] were able to use STM to acquire the high-resolution image of Fe phthalocyanines (FePcs) on Ag(110) in which three different configurations were identified, with R1-LD and R2-LD in the sub-monolayer and O-HD in the approximate single monolayer (Fig. 4a–c). With these high-resolution images, these researchers were able to examine the corresponding activities of each structure toward oxygen reduction through the in situ dosing of oxygen in a UHV chamber and characterize the surfaces before and after oxygen dosing. Here, these researchers found that the Fe-centered bright features on the mixed R1/R2-LD surface at a positive sample bias progressively disappeared during oxygen dosing, but recovered after annealing at 370 K for 60 min (Fig. 4d–g), whereas the O-HD FePc structure on Ag(110) showed almost no activity toward oxygen reduction (Fig. 4h, i), clearly demonstrating the different oxygen reduction activities of the different surface structures of FePcs.

Fig. 4
figure 4

Reprinted with permission from Ref. [99]. Copyright © 2012, Springer Nature

a–c STM images of monolayer FePc on Ag(110): a R1-LD phase, b R2-LD phase and c O-HD phase with respective unit cells indicated by corresponding blue lines. d STM image of a R1/R2-LD FePc/Ag(110). e STM image of a R1/R2-LD FePc/Ag(110) after dosing with 1700 L oxygen at room temperature. f STM image of the annealed surface (370 K, 60 min) of an oxygen dosed R1/R2-LD FePc/Ag(110). g High-resolution STM images of FePc molecules in the R2-LD phase: top (as deposited), middle (after oxygen dosage) and bottom (healed surface after annealing). h STM image of an O-HD FePc/Ag(110). i STM image of an O-HD FePc/Ag(110) after dosing with 1700 L oxygen at room temperature.

Full size image

In another study, Grumelli et al. [100] rationally prepared a two-dimensional (2D) MOCN on Au(111) for electrochemical ORR and used an UHV-STM and EC cell measurement combined system connected by a transfer system to study the resulting samples in which two molecular networks of benzene–1,3,5-tricarboxylic acid–Fe (TMA-Fe) and 5,50-bis(4-pyridyl)(2,20-bipirimidine) (PBP-Fe) were observed (Fig. 5a–d). And as combined with electrochemical testing (Fig. 5g, h), these researchers were able to demonstrate that the metal centers of the samples determined the final product as well as the electrocatalytic mechanism in which the Fe centered TMA-Fe and PBP-Fe exhibited a (2 + 2)e− mechanism, whereas a TMA-Mn network can directly reduce O2to H2O through a 4e−pathway. The nearly unchanged STM images of the sample surfaces before and after electrochemical testing also revealed the excellent reversibility and stability of these metal–organic networks (Fig. 5e, f). In a subsequent study, these researchers also studied heterobimetallic catalysts involving M1-TPyP-M2 (M1, M2 = Fe or Co, TPyP = 5,10,15,20-tetrakis(4-pyridyl)porphyrin) and reported almost 2 orders of magnitude increases in OER activity as compared with homometallic catalysts [101].

Fig. 5
figure 5

Reprinted with permission from Ref. [100]. Copyright © 2013, Springer Nature

a STM image of a TMA-Fe network on Au(111). b High-resolution STM image of a TMA-Fe network with the model superimposed. c STM image of a PBP-Fe network on Au(111). d High-resolution STM image of a PBP-Fe network with the model superimposed. e Large-scale STM image of a PBP-Fe network on Au(111) after EC tests. f Zoomed in STM image of (e). g EC ORR test in 0.1 M NaOH solution: bare Au(111) (gray solid line), TMA-Fe (pink solid line), PBP-Fe (turquoise solid line) and TMA-Fe after the addition of 10 mM H2O2 (pink dashed line). h Bare Au(111) (gray line), TMA-Mn (orange) and TMA-Mn after the addition of 10 mM H2O2 (orange dashed line). Insets in the graphs show the different onset potentials for the shoulder at − 0.2 V.

Full size image

3 EC-STM

EC-STM operates in electrolyte solutions and combines STM with EC cells. As a result, EC-STM can provide atomic- or molecular-level surface structures of electrocatalysts and chemical processes at the solid–liquid interface, thus allowing for the in situ monitoring of electrocatalyst surface structures and electrode processes including adsorption/desorption and dissociation of adsorbates under reaction conditions [16, 102]. Unlike STM operating under UHV conditions, the typical setup of EC-STM consists of two independent working electrodes (tip and sample), a reference electrode and a counter electrode, which allows for the simultaneous measuring of electrochemical currents (between sample and counter electrode) and tunneling currents (between tip and sample). Aside from tunneling currents, however, faradaic currents generated during electrochemical reactions at the tip–electrolyte interface also need to be minimized during measurements, and currently, the most effective method is to insulate the tip immersed in solution by coating the entire tip except the apex of the tip end, which can effectively minimize faradaic currents through the reduction in the exposed tip area with respect to the electrolyte [103, 104].

Due to continuous efforts in the development of instruments and methodologies for EC-STM in in-situ and in-operando studies, EC-STM has become a set of indispensable experimental tools in the field of electrocatalysis for a wide range of reactions under harsh electrochemical conditions. The significance of EC-STM in the field of electrocatalysis began with the acquisition of atomic-resolution STM images of abundant single-crystal metals and metal alloys in electrolyte solutions [105,106,107,108,109,110,111,112], followed by the correlation of these potential-dependent surface structures to corresponding electrocatalytic properties. And with the further development of EC-STM, deeper understandings of key properties relating surface structures to electrocatalytic properties including surface adsorbate sites, activity, selectivity and stability can be acquired, all of which can greatly benefit the optimization and design of desirable electrocatalyst materials.

3.1 Metal Surfaces and Their Alloys

Earth-abundant Cu has received tremendous attention in the electrochemical reduction of CO2 because it can produce more reduced products such as hydrocarbons and oxygenates that require more than two-electron reduction due the intermediate binding energy of Cu for CO [113, 114]. Here, EC-STM is an important tool in the identification of the electrochemical properties of the different facets and surface structures of Cu metal. For example, Hahn et al. [115] used EC-STM to identify the atomic structures of three different Cu surfaces, including Cu(111), Cu(100) and Cu(751), in which combined with electrochemical CO2 reduction tests, these researchers found experimental evidence that Cu(100) and Cu(751) surfaces were both more active and selective for C–C coupling than Cu(111) and that the Cu(751) surface was the most selective for > 2e− oxygenate formation at lower overpotentials. Baricuatro et al. [116, 117] also used a combination of EC-STM with differential electrochemical mass spectroscopy (DEMS) to demonstrate that CO was only adsorbed at potentials more negative than − 0.8 V to form a fully vertical c(2 × 2) structure (Fig. 6). In another example, Bae et al. [118] applied EC-STM to study nitrate adsorption and reduction on Cu(100) and foundthat nitrate ions formed a (2 × 2) adlattice from the open-circuit potential to − 0.22 V versus Ag/AgCl on Cu(100) surfaces, but a predominant c(2 × 2) structure from − 0.25 to − 0.36 V.

Fig. 6
figure 6

Reprinted with permission from Ref. [116]. © 2018 Elsevier B.V

EC-STM images of Cu(100) in 0.1 M KOH at − 0.9 V (a, b) and − 0.8 V (c), with (b, c) and without (a) CO in solution. The adsorption of CO on Cu(100) is schematically illustrated in the bottom panel.

Full size image

Alloying, especially surface alloying, with low-cost metals can decrease precious metal loading in corresponding materials and greatly benefit large-scale industrial applications. In addition, alloying can allow for the tuning of surface properties related to species adsorption, surface activity, selectivity and stability in which EC-STM is playing an increasingly important role in revealing these surface-related properties. For example, abundant metal alloys on noble metal surfaces such as Cd/Au, Cu/Au and Pd/Au on Au surfaces have been intensively investigated using EC-STM in terms of both synthesis by electrode position [119,120,121,122,123,124,125] and corresponding adsorption behaviors toward reactive species [126,127,128,129,130].

3.2 Metal-Supported Organic Molecule-Based Electrocatalysts

EC-STM has also been recognized as a powerful tool for the in situ study of organic molecule-based electrocatalysts supported by noble metal surfaces in terms of both essential properties including formation and structure and electrocatalytic properties. And to date, numerous studies have been conducted on organic molecule-based structures using EC-STM, not only in the field of electrocatalysis, but also in other applications such as gas sensors, optoelectronic devices, field effect transistors and solar cells [131,132,133,134,135,136,137,138,139,140,141]. Due to the focus of this review, however, this section will focus on electrocatalysis-related studies carried out using EC-STM. For example, Yoshimoto et al. [142] used a combination of CV measurements and EC-STM (in a benzene solution, 0.1 M HClO4) to demonstrate that in electrochemical ORR, iron octaethylporphyrin (FeOEP)/Au(111) enabled the two-step four-electron reduction of O2 to H2O, whereas cobalt octaethylporphyrin (CoOEP)/Au(111) enabled the two-electron reduction of O2 to H2O in which in situ EC-STM measurements revealed the formation of a highly ordered adlayer of FeOEP on the reconstructed Au(111) surface during the first electrochemical reduction of O2 to H2O2 and that under more negative electrode potentials, H2O2 can be further reduced to H2O. Here, the in situ EC-STM images of this reaction suggested that FeOEP molecules on the terrace possessed high mobility based on the unclear surface morphology.

EC-STM can also be used to conduct the single-molecule imaging of adsorbates on electrocatalysts under solution conditions at liquid–solid interfaces in individual metal porphyrin molecules on metal substrates [143,144,145,146]. For example, Hulsken et al. [143] were able to demonstrate by capturing high-resolution STM images in a liquid cell with and without O2 and recording ultraviolet–visible (UV–Vis) solution spectra that oxygen atoms in O2 molecules were bound to adjacent porphyrin catalysts on the surface before incorporation into alkene substrates. Den Boer et al. [147] also used EC-STM to reveal the multi-step reaction of individual manganese porphyrin molecules on graphite with molecular oxygen at the solid/liquid interface in which at least four distinct reaction species with different configurations in the STM image were identified with their reaction dynamics being observed and revealed by EC-STM at the liquid/solid interface and under different gas environments.

Numerous electrocatalytic processes involving metal porphyrins under varying electrode potentials have also been revealed using in situ EC-STM [148,149,150]. For example, Wang et al. [148] obtained high-resolution STM images under saturated O2, air and N2 and were able to identify the in termediates of FePc-O2 and bare FePc monolayer on Au(111) based on the appearance of bright and dim points (Fig. 7a). And based on the reversible switch between FePc-O2 and FePc from substrate potentials of 350 mV to 50 mV (Fig. 7b), these researchers were also able to demonstrate the ORR process for O2reduction to H2O2, which was further validated by theoretical calculations and CV measurements. In a more recent study, these researchers also found that the OER activity of a 5,10,15,20-tetraphenyl-21H,23H-porphyrin cobalt(II) (CoTPP)/Au(111) electrode can be enhanced by increasing the alkalinity of the electrolyte in CV tests, and by combining UV–Vis absorption spectra with in situ EC-STM, these researchers found that OER in alkaline media occurred through the formation of CoTPP-OH-species intermediates to CoTPP-O2 product [149].

Fig. 7
figure 7

Reprinted with permission from Ref. [148]. Copyright © 2016 American Chemical Society

a EC-STM images and height profiles of FePc on Au(111) under 0.1 M HClO4 saturated by (i) and (iv) oxygen, (ii) and (v) air and (iii) and (vi) nitrogen. Insets show enlarged STM images. b Sequential EC-STM images and height profiles of FePc monolayer on Au(111) in 0.1 M HClO4 saturated by oxygen at different potentials. Insets show the high-resolution STM images of FePc molecules. Yellow dotted lines in (ii) display the potential boundary. (i) and (v) E = 350 mV, Ebias = − 300.0 mV, (ii) and (vi) upper region: E = 350 mV, Ebias = − 300.0 mV, lower region: E = 50 mV, Ebias = − 300.0 mV, (iii) and (vii) E = 50 mV, Ebias = − 300.0 mV, (iv) and (viii) E = 350 mV, Ebias = − 300.0 mV.

Full size image

4 Others

4.1 High-Speed EC-STM

The time required to acquire conventional EC-STM images is usually on the order of seconds or minutes and is much slower than the highly dynamic processes of electrochemical catalytic reactions that are in the millisecond range. As a result, increases in scanning speed can greatly aid in the recording of a wide range of important dynamic phenomena at electrochemical interfaces, such as electrode surface variation, adsorbate diffusion and surface species transition between reactants, intermediates and products, all of which are essential in the study of electrocatalytic mechanisms. To address this, Magnussen et al. [151, 152] demonstrated the use of high-speed EC-STM (also known as video EC-STM) in their study and used this technique to study the electrochemical deposition and dissolution of Cu(100) in diluted HCl solution in which the high time resolution and high stability of the video EC-STM as well as the fast data acquisition and processing systems, high-gain high-bandwidth tunneling current preamplifiers, rapid feedback electronics and better mechanical stability are all indispensable [153,154,155,156].

In the last two decades, great progress has been made to video EC-STM techniques to enable the study of a variety of adsorbate dynamics, mainly on noble metal surfaces. These direct microscopic observations of electrochemical processes either on electrode surfaces or on adsorbates at electrochemical interfaces can help in the interpretation of fundamental mechanisms of electrocatalytic reactions and electrochemical growth. For example, Matsushima et al. [157] took advantage of the high time resolution of video STM to monitor the surface structure of Cu(100) during HER in 0.1 M HClO4 solution and were able to find that during HER, Cu(100) can undergo surface reconstruction from an unconstructed surface to a stripe-like structure (parallel to the Cu lattice) in less than 0.1 s, which can increase HER rates. Adsorbate dynamics, especially in the case of ordered halogen (Cl and Br) c(2 × 2) structures on Cu(100) and Ag(100) with low coverage of other adsorbate species, have also been intensively studied using high-speed EC-STM [158,159,160,161,162,163,164]. For example, Rahn et al. [162] obtained a sequence of video EC-STM images taken every 100 ms on an Ag(100) surface in 1 mM HClO4 and 1 mM KBr at − 0.12 VSCE (SCE = saturated calomel electrode) and were able to directly record the spontaneous transition of sulfur adsorbate between surface sites and subsurface sites (Fig. 8).

Fig. 8
figure 8

Reprinted with permission from Ref. [162]. Copyright © 2018 American Chemical Society

Sequence of in situ video STM images obtained on Ag(100) electrodes in 1 mM HClO4 + 1 mM KBr showing an ordered c(2 × 2)-Br lattice. Two sulfur adsorbates on the surface (bright protrusions) are indicated by white and black arrows.

Full size image

4.2 Scanning Noise Microscopy

Scanning noise microscopy (SNM) was developed to measure the noise parameters of tunneling current in STM and can be applied in monitoring dynamics at the nanoscale, especially in the case of molecular dynamics [165,166,167,168]. For example, Pfisterer et al. [169] used this technique in an electrochemical environment to directly reveal catalytically active surface sites under reaction conditions by monitoring relative changes in tunneling current noise. Here, these researchers proposed that tunneling currents under constant bias can vary with time on active sites (edge sites) as compared with those over non-active sites (terrace sites), suggesting local catalytic processes (Fig. 9ai, ii). This concept was also demonstrated on Pt(111) surfaces (Fig. 9aiii, iv) in which the spatial noise of tunneling currents appeared on the step edge of Pt(111) under HER conditions to reveal that active HER sites were likely to be located at the step concavities of Pt electrodes. The high spatial resolution of this method was further demonstrated on a Pd/Au(111) surface with more HER active Pd nanoislands being supported by less active Au(111) (Fig. 9b) in which if a working electrode potential was negative enough to initiate HER on Pd nanoislands but insufficient to do so on the less active Au(111) substrate, noticeably higher noises were observed on the Pd nanoislands, especially near the Pd/Au boundary as compared with Au(111).

Fig. 9
figure 9

Reprinted with permission from Ref. [169]. Copyright © 2017, Springer Nature

a Revealing catalytic sites under reaction conditions using scanning noise microscopy. (i, ii) Schematic illustrations explaining the concept of using scanning noise microscopy to identify active sites showing that tunneling current noises change as local environments between STM tips and samples change (over a terrace, i, versus a step, ii, in the sample); (iii) STM line scans obtained over a Pt(111) surface in 0.1 M HClO4 if the potential of the sample is either sufficiently negative or too positive to initiate HER; (iv) typical STM line scans over a Pt(111) surface in 0.1 M HClO4 under HER conditions. b STM characterization of nanostructured surfaces under HER conditions. (i) Monoatomically high islands of Pd on Au(111) surface (top) and height profiles of Pd (bottom) (constant-current mode); (ii) STM image of the boundary between a Pd island and an Au(111) substrate under HER conditions in 0.1 M sulfuric acid (constant-height mode); (iii) detailed STM line scans for the case shown in (i; iv) statistical analysis of STM line scans shown in (iii), over Au, Pd and Au/Pd boundaries.

Full size image

4.3 Tip-Enhanced Raman Spectroscopy

Tip-enhanced Raman spectroscopy (TERS) is an integrated system involving STM and Raman spectroscopy and possesses a sharp, plasmonically active and conducting tip in which through the use of plasmonic resonance raised by a sharp noble metal tip under an illumination laser, local electromagnetic fields near the tip apex can be strongly enhanced [170, 171]. As a result, TERS can combine the sensitivity of Raman spectroscopy with the lateral resolution of scanning probe microscopy (SPM) and the sensitivity of surface-enhanced Raman spectroscopy (SERS), allowing for the nanoscale spatial resolution of Raman signals, particularly in UHV environments [172,173,174,175]. Because of this, TERS can probe individual surface molecule structures and their interaction with substrates at the nano/molecular scale, which is useful for the understanding of heterogeneous catalysis. In terms of application in the field of electrocatalysis, Nguyen et al. [176] were able to use UHV-TERS in their investigation of O2 adsorption on cobalt phthalocyanine (CoPc)/Ag(111), which is the initial step of ORR. Zhong et al. [177] also achieved 3 nm resolution in the real space of TERS in ambient environments by using phenyl isocyanide (PIC) as a probing molecule on an atomically well-defined Pd(sub-monolayer)/Au(111) bimetallic model catalyst in which through the recording of the identical Raman peaks of PIC on the surface of Pd/Au(111) (Fig. 10), these researchers were able to conclude that the Pd/Au(111) surface exhibited significantly less activity toward the catalytic oxidation of PIC than the Au(111) surface under ambient conditions.

Fig. 10
figure 10

Reprinted with permission from Ref. [177]. Copyright © 2016, Springer Nature

TERS results for a Pd/Au(111) bimetallic surface. a STM image of the Pd0.8ML/Au(111) surface. b Line-trace TERS spectra acquired along the dashed line in (a). c Color-coded intensity maps of three line-trace TERS images taken in succession. d Top three panels: intensity plots of the three main TERS peaks (1165, 1590 and 1995 cm−1) with respect to tip position. Bottom panel: topographic height profile of the surface along the dashed line in (a) superimposed with the atomic model of the surface atoms. e Calculated TERS enhancement factor if an Au tip is located at different positions on a Pd–Au–Pd surface (top inset: substrate model for calculation). f Calculated TERS enhancement distribution on the x–z plane if the tip occupies three different positions: Au (x = 0 nm), step edge (x = 8 nm) and Pd (x = 20 nm) surface.

Full size image

Compared with UHV and ambient environments, TERS operating under electrochemical environments can clearly provide more detailed and accurate information on interfacial processes and structures under electrode potentials [178,179,180,181,182]. For example, Zeng et al. [179] reported a STM-based EC-TERS with its laser introduced horizontally into an EC cell through a long working distance microscope objective to illuminate the gap between the tip and sample surface (sample tilted by 10°) (Fig. 11a). As a result, a strong TERS signal on 4′-(pyridin-4-yl)biphenyl-4-yl)methanethiol (4-PBT)/Au(111) appeared and disappeared as the tip approached and retracted, respectively. These potential-dependent EC-TERS data with the additional Raman peaks as compared with SERS (Fig. 11b) were subsequently successfully applied to reveal the molecular configuration of 4-PBT on Au(111). By using EC-TERS, Kurouski et al. [180] were also able to observe steplike behaviors in TERS voltammograms that reflected the electrochemical reduction and oxidation of single or few Nile blue (NB) molecules.

Fig. 11
figure 11

Reprinted with permission from Ref. [179]. Copyright © 2015 American Chemical Society

a EC-TERS system: (i) schematic illustration of the setup; (ii) SEM image of the insulated gold tip; (iii) microscopic image of the tip as it approaches the substrate with laser illumination; (iv) TERS spectra of 4-PBT adsorbed on Au(111) surface with the tip approaching (top) and retracting (bottom), respectively. b EC-STM and EC-TERS measurements on a 4-PBT-adsorbed Au(111) surface: (i) EC-STM image of the surface; (ii) CVs of bare (black) and 4-PBT-adsorbed (red) Au(111) electrodes; (iii) potential-dependent EC-TERS spectra of a 4-PBT adsorbed on Au(111) surface; (iv) potential-dependent SERS spectra of a 4-PBT-modified Au NP surface.

Full size image

5 Summary and Perspective

STM is gaining increasing attention in the field of electrocatalysis and its ability to providing high-resolution information on surface structures, configurations and dynamics of surface adsorbates can significantly advance the fundamental understanding of electrocatalysts and electrocatalytic mechanisms, all of which are valuable for the rational optimization and design of excellent electrocatalysts. In this review, developments related to the study of electrocatalysts and electrochemistry using UHV-STM, EC-STM, high-speed EC-STM, SNM and TERS have been presented. And although challenges remain, research involving the application of these powerful techniques is rapidly growing and perspectives regarding UHV-STM and EC-STM can be listed as follows.

  1. 1.

    UHV-STM can provide atomic-resolution images of electrocatalyst surface structures that are highly important in the interpretation of electrocatalytic mechanisms because electrocatalytic reactions are atomic processes. And as compared with electron microscopy, STM can not only provide atomic-resolution images, but also provide information on local surface electronic structures through scanning tunneling spectroscopy (also known as dI/dV curves), which is critical for the interpretation of electrocatalytic mechanisms, for example, by cooperating with the well-known d-band theory developed by Hammer and Nørskovto describe the adsorption strength between metals and adsorbates depending on the position of their d-bands [183]. In addition to surface electronic structures, the other unique aspect of STM is the ability to acquire the configurations and dynamics of surface adsorbates through the measurement of real space atoms and molecules on the surface. And although UHV-STM can deviate more from actual electrocatalytic reactions as compared with EC-STM operating in electrochemical environments due to vacuum conditions and the requirement for surface cleanness, the combination of obtained information with other characterizations and theoretical calculations can generate pioneering studies of fundamental mechanisms and designs of novel catalyst systems.

  2. 2.

    EC-STM is a powerful and unusual tool for the in situ characterization of electrochemical interfaces and electrode processes in reaction environments because it can allow for the independent control of potentials at both the tip and working electrode. As a result, the obtained detailed information of electrode surfaces and adsorbate electrochemical processes under different working electrode potentials can allow for the direct recording of electrocatalytic reactions, which is highly advantageous as compared with STM operating in UHV or ambient environments. Because of this, EC-STM can provide indispensable insights into the fundamental mechanisms of electrocatalysis and electrochemistry. Despite this, the application of EC-STM is limited by strict sample requirements including high cleanness, flatness and conductivity, making this technique more applicable for model catalyst systems. In addition, the application of EC-STM in the study of gas/solution/solid interfaces is limited to a number of high-quality metal and metal alloy surfaces (or thin oxide islands and films supported by metals or metal alloys) and a limited number of gases. Here, theoretical simulations on electrocatalysts and electrochemical interfaces as well as characterizations using other in situ or in operando spectroscopic techniques such as Fourier transform infrared (FTIR), Raman and photoelectron spectroscopy (PES) can greatly assist EC-STM in revealing the properties of electrocatalysts and electrochemical mechanisms.