Advertisement

Electrochemical Energy Reviews

, Volume 1, Issue 3, pp 433–459 | Cite as

In Situ and Surface-Enhanced Raman Spectroscopy Study of Electrode Materials in Solid Oxide Fuel Cells

  • Xiaxi Li
  • Kevin Blinn
  • Dongchang Chen
  • Meilin Liu
Review article
  • 591 Downloads

Abstract

Solid oxide fuel cells (SOFCs) represent next-generation energy sources with high energy conversion efficiencies, low pollutant emissions, good flexibility with a wide variety of fuels, and excellent modularity suitable for distributed power generation. As an electrochemical energy conversion device, the SOFC’s performance and reliability depend sensitively on the catalytic activity and stability of electrode materials. To date, however, the development of electrode materials and microstructures is still based largely on trial-and-error methods because of the inadequate understanding of electrode process mechanisms. Therefore, the identification of key descriptors/properties for electrode materials or functional heterogeneous interfaces, especially under in situ/operando conditions, may provide guidance for the design of optimal electrode materials and microstructures. Here, Raman spectroscopy is ideally suited for the probing and mapping of chemical species present on electrode surfaces under operating conditions. And to boost the sensitivity toward electrode surface species, the surface-enhanced Raman spectroscopy (SERS) technique can be employed, in which thermally robust SERS probes (e.g., Ag@SiO2 core–shell nanoparticles) are designed to make in situ/operando analysis possible. This review summarizes recent progresses in the investigation of SOFC electrode materials through Raman spectroscopic techniques, including topics of early stage carbon deposition (coking), coking-resistant anode modification, sulfur poisoning, and cathode degradation. In addition, future perspectives for utilizing the in situ/operando SERS for investigations of other electrochemical surfaces and interfaces are also discussed.

Keywords

SOFC Raman spectroscopy Surface enhanced Raman spectroscopy (SERS) In situ Operando 

List of Nomenclatures and Acronyms

BZCYYb

BaZr0.1Ce0.7Y0.1Yb0.1O3−δ

BZY

BaZr1−xYxO3−δ

DFT

Density functional theory

EDX

Energy dispersion X-ray spectroscopy

EIS

Electrochemical impedance spectroscopy

EF

Enhancement factor (of SERS)

GDC

Gadolinium doped ceria

IR

Infrared spectroscopy

LNF

LaNi1−xFexOδ

LSM

La1−xSrxMnO3−δ

LSCF

La1−xSrxCo1−yFeyO3−δ

LSP

Localized surface plasmon

LSPR

Localized surface plasmon resonance

OCV

Open circuit voltage

ORR

Oxygen reduction reaction

PEMFC

Proton exchange membrane fuel cells

PLD

Pulsed laser deposition

PNM

PrNi1-xMnxOδ

R6G

Rhodamine-6G (a chemical to evaluate SERS enhancement factor)

SDC

Scandium doped ceria

SEM

Scanning electron microscopy

SERS

Surface enhanced Raman spectroscopy

SOEC

Solid oxide electrolysis cell

SOFC

Solid oxide fuel cell

SSC

Sm1−xSrxCoO3−δ

TEM

Transmission electron microscopy

TEOS

Tetraethyl orthosilicate

TPB

Triple phase boundary (of electrolyte, electrode, and gas phase)

UV-Vis

Ultraviolet-visible spectroscopy

XPS

X-ray photoelectron spectroscopy

XRD

X-ray diffraction

YSZ

Yttrium stabilized zirconia

PACS

88.30.pd 88.30.pn 82.80.Gk 

1 Introduction

The research and development of solid oxide fuel cells (SOFCs) is stimulated by the need for energy conversion systems with higher efficiencies and lower greenhouse gas emissions. SOFCs, utilizing an ion-conducting ceramic membrane [(e.g., yttrium-stabilized zirconia (YSZ) or scandium-doped ceria (SDC)] and two ceramic or ceramic–metal electrodes with catalytic materials, can convert chemical energy in hydrogen or fossil fuels into electricity through electrochemical reactions [1]. And unlike thermo-mechanical power generation mechanisms such as gas turbines, fuel cells can eliminate the intermediate steps of transforming chemical energy into heat and mechanical energy before the conversion to electricity, therefore bypassing the Carnot cycle limitation of energy conversion efficiencies. SOFCs can also provide easier carbon sequestration due to their high CO2 concentrations in the fuel exhaust as compared with other carbon-bearing molecules, adding the additional benefit of lowered greenhouse gas emissions.

Compared with other lower-temperature fuel cells such as proton exchange membrane fuel cells (PEMFCs), SOFCs also demonstrate significant economic benefits in terms of material costs and fuel flexibility. Here, because SOFCs are designed to operate at elevated temperatures (500~800 °C), they can use Ni and perovskites electrode materials as opposed to low-temperature electrochemical reactions which require noble metals such as Pt as catalysts to achieve acceptable power densities, making SOFC technologies economically feasible [2]. In addition, SOFCs possess high fuel flexibility, being able to utilize a wide variety of carbon-based fuels such as coal gas [3], methane [4], propane [5, 6], methanol [7], and even octane [8, 9],rather than operating solely on pure hydrogen, allowing for immediate inclusion into current electricity generation infrastructures [10].

SOFC technologies can also be promising components in smart grids by virtue of its capability for distributed power generation and energy storage. In addition, sub-MW-level SOFC stacks can operate at optimal fuel efficiencies, making them ideal for use in distributed power generation applications as compared with conventional gas turbines which require GW-level operations to reach optimum efficiencies [2, 11, 12, 13]. Furthermore, the reverse operation of SOFCs as solid oxide electrolysis cells (SOECs) is a versatile method for energy storage and is critical to complement power grids with high proportions of renewable energy such as wind or solar.

The design of high-performance electrode materials that remain stable over long service periods is a perennial topic of SOFC research, however, with the most critical component impacting the performance and stability of SOFC electrodes being electrochemical interfaces. Figure 1 illustrates the main electrochemical interfaces of an SOFC, along with key reaction sequences and potential degradation mechanisms. Here, the understanding of reaction sequences and degradation mechanisms of electrochemical interfaces is essential for the design of optimal electrode materials with high activities and stabilities.
Fig. 1

Main components of an SOFC and possible reaction sequences on the electrodes. Call-outs illustrate effective electrode modifications and possible degradation of fuel cell components

SOFC electrode materials are susceptible to reactions and contaminants from operating environments that produce undesirable phases and species which in turn degrade electrode performances. For example, nickel-based anodes are vulnerable to the carbon deposition in cases where hydrocarbons are used as fuels [14, 15, 16, 17]. In addition, if fuels containing ppm levels of H2S are fed to nickel-based anodes, significant performance losses are also observed [18, 19, 20]. In addition, cathode materials are also susceptible to contaminations and material degradation over long-term operations. For example, the use of Cr-containing alloys as interconnect materials can lead to Cr poisoning, degrading cathode performances [21, 22, 23]. Furthermore, cathode materials such as La1–xSrxCo1–yFeyO3–δ (LSCF) can degrade due to elemental leaching and surface segregation, with the presence of CO2 and water vapor in air accelerating this process [24, 25].

Surface modification is a promising strategy to resolve the degradation issues facing electrochemical interfaces of SOFCs. For example, decorating La1–xSrxCo1–yFeyO3–δ (LSCF) porous electrodes with certain nanoparticles such as Sm1–xSrxCoO3–δ (SSC) and Ce1–xSmxO2–δ (SDC) has been reported to reduce polarization resistances [26, 27], whereas coating LSCF surfaces with La1–xSrxMnO3–δ (LSM) can improve long-term stabilities [28]. Carbon deposition on nickel-based anodes can also be alleviated through modifications such as BaO, BZCYYb, or BZY, which are catalysts that can facilitate the gasification of deposited carbon atoms [3, 9, 29, 30].

Due to the complexity of the electrochemical interfaces of SOFCs, however, conventional characterization tools cannot always identify the factors responsible for electrode function and degradation. And although the electrochemical testing of full fuel cells or symmetric cells can provide overall performance evaluations of electrode materials, the functional nature of electrode surfaces is often entangled with other factors such as microstructures, electrode thicknesses, ohmic resistances of electrolyte materials, and polarization resistances of other interfaces. Here, although ex situ characterization tools such as SEM and XPS can allow for the analysis of microstructures and elemental compositions on electrode surfaces, surface morphologies and compositions obtained from these postmortem analyses may not be representative of what is occurring during fuel cell operations. Therefore, the in situ/operando probing and mapping of SOFC electrodes are crucial to gain knowledge related to electrochemical interfaces under exposure to operational environments.

Raman spectroscopy, by virtue of its ability to conduct in situ analyses, is a unique approach to study SOFC electrodes. Unlike electron-based characterization tools such as XPS and EDX, Raman scattering is a vibrational spectroscopy technique based on laser interactions with sample surfaces and thus does not require a vacuum environment, enabling the analysis of samples under exposure to gas atmospheres at high temperatures. To date, many studies of SOFC electrodes using Raman spectroscopy have been reported, covering topics including coking, sulfur poisoning, cathode degradation, and the remediation of these issues. In addition, enabled by the development of thermally and chemically robust SERS probes, our group has even implemented in situ/operando SERS analysis of SOFC materials at 450 °C.

Overall, this review will provide a systematic account of the use of the in situ Raman and the SERS for the investigation of electrochemical processes on SOFC electrodes. In this review, following the current introductory section, Sect. 2 will focus on the development of in situ Raman and SERS capabilities for the analysis of SOFC electrodes, focusing especially on the design of thermally robust SERS probes. Section 3 will present our studies on the early stage carbon deposition using the SERS-enabled probing and mapping. Section 4 will be devoted to the investigation of Ba-containing catalysts as coking-resistant modifications through the in situ Raman and the SERS. Section 5 will present SERS studies of nickel surfaces after exposure to H2S, and Sect. 6 will discuss the SERS analysis of LSCF degradation in the presence of Cr species and humidified air. Finally, Sect. 7 will discuss the limitations and remedies of the current SERS and in situ/operando Raman spectroscopy methods along with new SERS architectures for general high-temperature electrochemical studies.

2 In Situ Raman and SERS Techniques for SOFC Electrodes

2.1 Raman Spectroscopy

This review will not elaborate in great depth on the fundamental theories of Raman spectroscopy, but basic principles concerning its application in SOFC electrode environments will be discussed. More information concerning Raman theory can be found in textbooks [31, 32, 33] and other widely available literature sources.

Raman scattering is an inelastic scattering process in which incident photons interact with scattering media and change energy. This loss (Stokes mode) or gain (Anti-Stokes mode) in energy corresponds to vibration modes of scattering media and can be used for the fingerprinting of phases or functional groups. For the analysis of SOFC electrode materials, a single wavelength laser is used to illuminate sample surfaces, and Raman spectra can be collected from the backscattered photons. This spectra can subsequently provide characteristic information on bulk phases and adsorption species.

The Raman spectra of crystalline materials are characteristic of their lattice vibration modes, often referred as phonon modes, which directly correlate to the space groups of the materials. However, not all material phonon modes are Raman active because Raman scattering requires polarizability tensors corresponding to vibration modes to change with position. Therefore, centrosymmetric crystal structures, such as cubic structures (BCC or FCC) of metals, rock salts (AX), or perfect perovskites (ABO3), cannot generate Raman scattering. However, although many SOFC electrode materials are based on centrosymmetric crystal structures, these materials can possess Raman-active modes as a result of crystal distortion. For example, La1–xSrxMnO3–δ (LSM), a widely studied SOFC cathode material, is an orthorhombically distorted perovskite belonging to the space group Pnma (D2h point group); it theoretically has 24 Raman-active vibrational modes (seven Ag + 5B1g + 7B2g + 5B3g). In addition, defects in crystal lattices can reduce symmetry and generate active Raman bands. For example, CeO2 lattices doped with elements with +3 valences (e.g., Sm, La, Pr, etc.) can result in the appearance of an additional peak between 500 and 600 cm−1 as a result of the presence of oxygen vacancies that reduce crystal symmetry [34, 35, 36, 37]. This same spectral feature can also appear if CeO2 is exposed to reducing atmospheres [38]. Therefore, the capability of Raman spectroscopy to detect changes in lattice symmetry as a result of doping, element segregation, contamination, and redox reaction allows for versatile application in the analysis of SOFC electrode materials.

With regard to surface analyses, active Raman bands of adsorbed surface species correlate with vibration modes of functional groups, such as OH (~ 3500 cm−1), \( {\text{CO}}_{3}^{2 - } \) (1059 cm−1), and \( {\text{SO}}_{4}^{2 - } \) (980 cm−1). And by relying on the capability of these surface analyses, reaction mechanisms can be proposed and verified based on surface species dynamics. For example, the in situ Raman detection of surface functional groups has been demonstrated for oxygen adsorption on CeO2 surfaces in which different peak positions can be used to distinguish between peroxide and superoxide adsorption on surfaces [39]. In addition, in situ Raman can also be used to detect water adsorption on BaO and BZCYYb surfaces, helping to elucidate the mechanisms of water-mediated carbon removal through surface modification [3, 30]. Space groups and correlated Raman modes of typical SOFC electrode materials and adsorption species are presented in Table 1.
Table 1

Space groups and Raman modes of common SOFC electrode materials and adsorption species

Phase/species

Group

Raman modes

Main peak(s)a

YSZ

\( {\text{O}}_{\text{h}}^{6} \)

F2g

620

SDC

\( {\text{O}}_{\text{h}}^{6} \)

F2g

460

LSM (La/Sr: 0.85/0.15)

\( {\text{D}}_{{ 2 {\text{h}}}}^{16} \)

7Ag + 5B1g + 7B2g + 5B3g

535

LSM (La/Sr: 0.8/0.2)

\( {\text{D}}_{{ 3 {\text{d}}}}^{6} \)

Ag + 4Eg

697. 520

LSCF

\( {\text{D}}_{{ 3 {\text{d}}}}^{6} \)

Ag + 4Eg

646. 525

BZCYYb

\( {\text{D}}_{{ 2 {\text{h}}}}^{16} \)

7Ag + 5Blg + B2g + 5B3g

335

BZY

\( {\text{O}}_{\text{h}}^{2} \)

None

720b

BZYb

\( {\text{O}}_{\text{h}}^{2} \)

None

725b

C (Graphite)

\( {\text{D}}_{{ 6 {\text{h}}}}^{5} \)

E2g

1585

H2O

C2

2A1 + B2

3716.3368. 1594

O2

D2

Ag

1598

\( {\text{SO}}_{4}^{2 - } \)

Td

A1 + E + 2F2g

980

\( {\text{CO}}_{3}^{2 - } \)

D3h

A1g + 2E

1059

C3Hs

C2v

xA1 + yB2

1528

C2H4

D2h

3Ag + B1g + 2B3g

1425

COH

Cs

2A1 + A2g

2483c

BaO2

\( {\text{D}}_{{ 4 {\text{h}}}}^{17} \)

Alg + Eg

240. 800c

aApproximate positions of experimentally observed peaks

bPeaks were observed but not predicted by group theory; bands are likely defect induced

cNot experimentally observed; cited from Refs. [31, 40]

2.2 In situ/Operando Raman Spectroscopy

Typically, Raman spectroscopic analyses of SOFC electrode materials can be performed by using commercially available Raman spectrometers, and in our group, Raman spectroscopic analyses are performed by using a Renishaw RM 1000 spectromicroscopy system (~ 2 µm spot size) in which an air-cooled Ar laser (CVI Melles Griot) emitted at 488 nm and 514 nm and a solid-state diode laser (Thorlabs) with a 633-nm emission line are used for the excitation of Raman signals with a total power of 30 mW, 5 mW, and 10 mW, respectively. Here, data handling is routinely conducted by using the built-in WIRE software. In addition, MATLAB scripts were also developed in our group for the batch processing of line scans, mappings, and time-series data. Furthermore, these scripts also allow for standard background subtractions and more versatile information extractions.

As for in situ Raman analyses, a customized Harrick environmental chamber (Fig. 2) is used to control temperatures, gas atmospheres, and apply electrical biases on model fuel cells with patterned electrodes. A quartz window sealed with a BUNA O-ring is also used to allow for the transmission of the excitation laser and Raman-scattered photons. Here, lenses with focal lengths of > 8 mm are used to allow for sufficient working distances between the lens and the sample surface. Furthermore, the stainless steel jacket of the environmental chamber is cooled with running water. And because the surface temperature of the electrode host in the environmental chamber deviates from nominal temperatures, calibration curves are needed. In addition, a motorized stage (Prior Scientific) with 1 µm spatial resolution is used to enable the mapping of samples.
Fig. 2

Schematic and picture of the environmental chamber used for in situ Raman spectroscopy

This design of in situ Raman chamber offers unique advantages to our investigation of SOFC electrode materials. First, it allows for characterization of electrode materials under practical fuel cell operating conditions - exposure to reaction gases (and contaminants) while the cell is subject to an applied current or voltage at high temperatures. Second, it allows for mapping of reaction species on electrode surfaces or across the electrochemically active 3-phase boundary lines (e.g., Ni-YSZ-gas), which can be fabricated via microfabrication of patterned electrodes to help visualize reaction ‘hotspots’. Third, continuously acquisition of Raman spectra as a function of time under various electrochemical testing conditions provides information vital to gaining insight into the kinetics of electrode processes.

2.3 Surface-Enhanced Raman Spectroscopy

Raman scattering possesses lower cross sections (probability) compared with elastic scattering (Rayleigh) and characteristic absorption (IR), resulting in low yields of Raman signals. This low Raman spectroscopy signal strength poses particular challenges for the in situ analysis of SOFC electrodes. One challenge is that although Raman spectroscopy can identify adsorbed species, many key reactions intermediates of SOFC electrodes are present in trace amounts and are therefore undetectable. For example, sulfur contamination on nickel anodes is not readily detected by normal Raman spectroscopy, especially at operating temperatures [42, 43]. In another example, although severe carbon deposition on anodes can be easily detected by Raman spectroscopy, early stage identification of the carbon deposition is challenging [44]. Another challenge of low Raman spectroscopy signal strengths is that some electrode materials exhibit weak Raman modes as a result of the weak polarizability of their bonding. For example, although LSM and LSCF electrodes possess orthorhombic and rhombohedral distortions to their perovskite structures and can provide active Raman modes, their polarizability is low and produces weak Raman signals. Finally, high-temperature in situ analyses pose further challenges to detection sensitivities because the variation of energies for all Raman modes increases at elevated temperatures, resulting in decreased signal-to-noise ratios.

To overcome these problems, surface-enhanced Raman spectroscopy (SERS) is a technique that can be used to amplify the yield of Raman signals. First discovered on molecules adsorbed on electrochemically roughened silver substrates [45, 46], the SERS can significantly amplify Raman signals of adsorbed molecules and can be observed on Au, Ag, and Cu nanoparticles with a variety of geometries [47, 48, 49]. This SERS enhancement effect is attributed to the coupling of excitation lasers with localized surface plasmons (LSP) of Ag or Au nanoparticles, in which the electron clouds of metal particles oscillate with the excitation laser [50, 51]. Two review articles conducted by Kerker et al. [52] and Stiles et al. [52, 53] further elaborate the mechanistic interpretations of the SERS and its relationship with the LSP.

In the past several decades, despite the tremendous success of the application of the SERS in other fields (e.g., biomedical engineering), its powerful capabilities have not been exploited in the field of high-temperature catalysts where the fine surface probing is critically needed [54, 55]. Because of this, our group has systematically studied the application of the SERS on the analysis of SOFC electrode conditions and demonstrated its effectiveness in identifying surface contaminations, thin films, redox-induced vacancy structure changes, and adsorption species. In addition, both ex situ and in situ SERS probes have been developed to enhance signals for SOFC electrode analyses. In terms of ex situ analysis, SERS probes are applied to electrodes after the completion of chemical/electrochemical reaction processes on the electrode, whereas for in situ analysis, SERS probes need to be introduced prior to the reaction step of interest, with Raman signals being collected from the electrode during the reaction. Schematics for both analysis methods are presented in Fig. 3.
Fig. 3

Schematic of a ex situ and b in situ SERS analysis of SOFC electrode materials

2.4 SERS Probes for Ex Situ Analysis

As an example of our SERS probes for ex situ analysis, the magnetron sputtering of Ag was used to fabricate SERS probes for ex situ analysis. Here, Ag sputtering was found to produce SERS probes free from contamination with an enhancement factor (as evaluated by rhodamine-6G) on the order of 105, which is ideal for the analysis of surface contaminations and secondary-phase developments on electrode surfaces. In addition, the sputtering process of Ag probes was optimized to maximize the enhancement factor and the process can be well controlled [56]. Figure 4 shows that SERS capabilities can be tuned through the adjustment of sputtering durations in which optimal sputtering durations were found to be 180 s and 240 s, respectively, for excitation wavelengths of 514 nm and 633 nm. Here, the enhancement factor provided by R6G was several orders of magnitude higher than that by carbon film, and this is because the bulk of the carbon film contributes to a majority of the original Raman signal and because carbon does not selectively adsorb onto Ag nanoparticles.
Fig. 4

Optimization of net enhancement factors (EFnet) of sputtered Ag. a SEM images of sputtered Ag nanoparticles with systematically varied durations under a working pressure of 2.5 × 10−2 mBar and b their corresponding enhancement factors evaluated by using a standard carbon film with lasers emitted at 514 nm and 633 nm. Here, the carbon films were fabricated through the evaporation of a 5-mm-long carbon cord by using a Quorum coater. Reproduced with permission from [57]

As a critical advantage of our method, sputtered Ag nanoparticles also demonstrate excellent uniformity across the surface, allowing for the semiquantitative mapping analysis of the enrichment of species on sites of interest, such as the interface between the electrode and electrolyte. Figure 5 presents the SERS spectra acquired at different locations across the surface of a GDC thin film sputtered onto a 1 cm2 silicon wafer, with the sputtered Ag as SERS probes. Here, not only did the SERS probes enhance the weak F2g mode of the GDC, the standard deviation of the signal intensities accounted for only 3% of the total intensity.
Fig. 5

Surface uniformity of sputtered Ag as the SERS probes on GDC thin films. a Ordinary Raman SERS spectra of a GDC thin film. b SERS spectra of the GDC thin film collected from nine different spots

2.5 Thermally Robust SERS Probes for In Situ Analysis

For in situ analysis, thermally stable SERS probes can be fabricated by coating Ag nanoparticles with a SiO2 shell (Ag@SiO2 NPs) through a revised polyol method. Here, during the seeding process, AgNO3 is reduced by ethylene glycol to generate Ag nanocrystals in a solution-based nucleation and growth process and the SiO2 shell is introduced by adding a tetraethyl orthosilicate (TEOS) solution.

Here, the thermal robustness of the Ag@SiO2 SERS probes was evaluated by collecting localized surface plasmon resonance (LSPR) spectra after heat treatment at different temperatures because the extinction capability from the LSPR spectra directly correlates with SERS capabilities of the probes. And as shown in Fig. 6, extinction capabilities decrease as heat treatment temperatures increase, and is completely diminished after heat treatment at 600 °C. However, for temperatures equal to or below 500 °C, a considerable portion of the extinction capability remains. In contrast, the extinction capability of the sputtered Ag probes disappears after heat treatment. Here, the thermal robustness of the SERS probes was further confirmed by TEM images of the probes before and after heat treatment, as shown in Fig. 7 in which the core–shell structure of the probes remained largely intact after heat treatment at 450 °C.
Fig. 6

UV–Vis extinction spectra of SERS probes after heat treatment. a Ag@SiO2 nanoparticles deposited on glass coverslips before and after heat treatment at different temperatures (200~600 °C) for 30 min in air. b Sputtered Ag nanoparticles before and after heat treatment at 450 °C for 1 h in air

Fig. 7

TEM images of Ag@SiO2 nanoparticles a before heat treatment and b after annealing at 450 °C for 1 h in air and c in 4% H2

The capability of thermally robust SERS can also be demonstrated by Ag@SiO2 probes with different excitation wavelengths, as shown in Fig. 8. Here, the substrates onto which SERS probes are loaded are GDC thin films on silicon wafers, and SERS probes are heat-treated at 400 °C in air for 1 h and allowed to cool down to room temperature before analysis. And with excitation wavelengths of 488 nm, 514 nm, and 633 nm, respectively, enhancement factors of ~ 40 can be obtained and can be even higher if a thinner coating or R6G was used in enhancement factor calculations. It was also noticed that the 514-nm laser showed the strongest fluorescence resonance in the presence of the Ag@SiO2 nanoparticles, but under 488- or 633-nm excitation, such fluorescence background resonance was significantly reduced. This suggests that the changing of wavelengths for incident lasers may mitigate fluorescence that interferes with Raman band analyses.
Fig. 8

Effect of incident wavelengths of lasers on the SERS capability of Ag@SiO2 probes. a Pure Ag@SiO2 nanoparticles loaded on silicon wafer (Ag@SiO2|Si, dash-dot line), blank GDC thin film deposited on silicon wafer (GDC|Si, solid black line), and Ag@SiO2 nanoparticles-loaded GDC thin film after heat treatment in 4% H2 (Ag@SiO2|GDC|Si, solid colored line), inspected with a 30 mW blue laser (488 nm). b The same samples inspected with a 10 mW green laser (514 nm). c The same samples inspected with a 15 mW red laser (633 nm). Each spectrum represents an average of the spectra collected from nine random points

3 Study of Coking on Nickel-Based Anodes

Coking is a common degradation mode of nickel-based SOFC anodes running on hydrocarbon fuels (i.e., methane, propane, etc.) [58]. For the rational design of SOFC anode materials, the probing and mapping of interactions between hydrocarbon fuels and anode surfaces under operating conditions are imperative. In particular, the understanding of the kinetics of the carbon deposition at early stages can help to identify the intrinsic coking propensity of material surfaces. This section reviews the ex situ, in situ, and operando surface-enhanced Raman spectroscopy (SERS) studies to probe and map the early stage carbon deposition on nickel surfaces.

3.1 Early Stage Carbon Deposition on Nickel Surfaces

Ordinary Raman spectroscopy is capable of capturing carbon depositions on nickel surfaces both ex situ and in situ. Figure 9 shows the mapping of a nickel coupon after exposure to propane-containing gas at 550 °C. Here, due to differing crystal orientations, nickel grains on the same coupon demonstrate different degrees of coking, resulting in colored patches. To study this, the in situ mapping was used and the results revealed that more discolored areas match exactly with the areas with stronger carbon signals. The sensitivity of ordinary Raman was so limited in this case, however, that further information could not be obtained than what was obtained by using the optical microscopy and the SEM in the study of coking.
Fig. 9

a Optical micrograph of a Ni coupon after exposure to C3H8-containing gas at 550 °C along with b a map of the carbon D-band Raman intensity (integrated from 1270 to 1445 cm−1) collected from the same area. c Actual spectra from selected spots in the map. d SEM micrograph of the same Ni coupon with grain boundaries marked by dotted lines. Reproduced with permission from [41]

In the case of SERS probes, the detection sensitivity of carbon depositions can be significantly improved. For example, Fig. 10 presents the analysis of a nickel coupon after exposure to propane-containing gas for a short period of time. Whereas ordinary Raman can only detect carbon in the discolored regions, after the introduction of SERS probes, carbon signals can be detected from non-discolored regions as well in which SEM analysis of such regions showed a small nuclei, possibly associated with the early stage carbon deposition.
Fig. 10

Ex situ SERS identification of the early stage carbon deposition on nickel surfaces. a Optical micrograph of a hydrocarbon-exposed nickel surface and d SEM image with low magnification, both of which show patches related to different grades of the carbon deposition. Raman spectroscopy and SERS analysis of b the mildly coked region and c the heavily coked region. High-magnification SEM of e the mildly coked region and f the heavily coked region. SEM images (d)–(f) were taken before any silver deposition, and all Raman spectra were collected by using a 633-nm laser with silver deposition duration being kept at 180 s. Reproduced with permission from [57]

The SERS can also allow for the in situ monitoring of the carbon deposition process with Ag@SiO2 nanoparticles acting as SERS providers. For example, Fig. 11 shows the time-resolved study of coke formation and elimination on nickel surfaces at 450 °C by using the ordinary Raman and the SERS. Here, although the intensities of the carbon D-band and G-band increased over time when exposed to a propane-containing gas, the SERS enabled sample showed much higher intensities and signal-to-noise ratios. As a result, the signal intensity from the SERS-activated sample showed better-resolved trends, revealing that carbon deposition rates onto nickel surfaces were initially rapid but decreased over time. In this study, the coked sample was also exposed to a gas mixture of 1% O2, 3% water vapor and 96% Ar, and the subsequent SERS analysis also showed a clear trend of coke removal from the sample surface.
Fig. 11

In situ normal Raman (NR) and SERS analysis of the carbon deposition and removal. ac The monitoring of the carbon deposition process on nickel surfaces through wet propane exposure at 450 °C on a blank Ni and b Ag@SiO2-loaded Ni, and c the integrated intensity of the carbon D-band at 1350 cm−1. de The monitoring of the subsequent removal of the carbon deposition through exposure to wet 1% O2 on d blank Ni and e Ag@SiO2-loaded Ni, and f the integrated intensity of the carbon D-band at 1350 cm−1. Each individual spectrum was taken with a 514-nm laser at 5 mW with an acquisition time of 10 s. Reproduced with permission from [59]

Although the in situ SERS can provide significantly enhanced carbon signals that develop on nickel surfaces, continuous exposure to wet propane results in carbon deposition rates which are too fast, preventing the collection of representative spectra for the amount of carbon depositions at very early stages. Therefore, to quantify the early stage carbon deposition, our group designed our experiment based on a method in which short pulses of propane are delivered into the environmental chamber for the controlled growth of carbon deposits [60]. Here, during each 10~15 s pulse, 50% propane (2% H2 and 48% Ar) is flown into the chamber with an equivalent propane volume of 10~20 mL. Following this, the chamber is switched to a subsequent purging mode for 2 min with dry 4% H2 (96% Ar) for spectra collection. And by contrast to the continuous flow of high-concentration propane, this method can register carbon depositions to the precise amount of propane delivered into the chamber.

In this study, the SERS signal for carbon depositions was shown as a function of the amount of propane delivered to the chamber and various stages of the carbon deposition can be identified in which the first stage is a quick growth of carbon signals in the first 20 mL pulse, as shown in Fig. 12a. Following the first few pulses, the carbon deposition proceeds, but growth rates slow down and continuous exposure to more hydrocarbon molecules results in a linear increase of carbon G-band intensities as a function of the exposure amount of propane, as displayed in Fig. 12b. Here, the quick development of carbon peaks during the first 20 mL of exposure suggests the quick initiation of the carbon deposition on nickel surfaces. However, during these first few pulses, nickel regions active for the carbon deposition become covered, and further deposition needs to overcome higher energy barriers and is therefore slowed down. In contrast to the SERS analysis, ordinary Raman spectroscopy studies of carbon depositions on the same type of nickel foil cannot produce any signals until 1000 mL of propane has been introduced, but these signals quickly develop afterward, indicating a self-catalyzed growth of bulk carbon in forms such as encapsulation, whiskers, and/or filaments.
Fig. 12

In situ SERS analysis of carbon deposition at different stages. a SERS intensity of carbon G-bands as a function of propane exposure during initial carbon deposition; each pulse contains 10 mL of propane. b SERS intensity of carbon G-band intensities as a function of propane exposure as the carbon steadily deposits; each pulse contains 150 mL of propane. c SERS and ordinary Raman analyses of carbon G-band intensities with a continuous flow of propane. Reproduced with permission from [60]

With the control of pulse exposure to introduce the early stage carbon deposition to nickel along with high-temperature SERS probes to enhance signals introduced by these early stage carbon depositions, a series of coking scenarios for SOFC anodes can be investigated which are of interest to the study of coking resistances. The details of these are presented in our previous paper [60], and the key findings are summarized below:

  1. 1.

    GDC decorated Ni surfaces with nominal thicknesses of ~ 3 nm show significant coking resistances as compared with pure Ni surfaces (Fig. 13a).

     
  2. 2.
    Humidified propane-containing gas mixtures are much less prone to coking on nickel surfaces (Fig. 13b).
    Fig. 13

    Impact of GDC modification and oxygen regeneration on the initiation of coking. a Intensity of carbon G-bands as a function of propane exposure volumes, collected from blank Ni foils and GDC-modified Ni foils, respectively, during the initial round of exposure. b Intensity of carbon G-bands upon exposure to dry/wet propane collected from GDC-modified Ni foils after the initial round. After each round of propane exposure, the sample was regenerated by air. Reproduced with permission from [60]

     
  3. 3.

    On Ni–YSZ interfaces, intermediate hydrocarbon species (characteristic of many Raman modes between ordinary D-band and G-band) can form (Fig. 14a).

     
  4. 4.
    An anodic bias applied to the Ni–YSZ interfaces can facilitate the removal of intermediate hydrocarbon species, but not all carbon depositions (Fig. 14b).
    Fig. 14

    Time-resolved operando SERS analysis of carbon deposition and electrochemical removal on Ni–YSZ interfaces. a Spectra acquired as a function of time under 50% C3H8, 2% H2, and 48% Ar. b Spectra acquired under 50% C3H8, 2% H2, and 48% Ar with anodic biases (OCV, 0.1 V, 0.2 V, 0.5 V, and 1 V) applied consecutively, showing the carbon removal process. Arrows beside the legend indicate the order of spectra acquisition. Each spectrum reflects an acquisition time of 50 s. The band intensity was computed by integrating the Raman spectral line over a 20 cm−1 bandwidth. Reproduced with permission from [60]

     

The in situ/operando probing and mapping of the carbon deposition are key to the understanding of the coking process and the development of remediation methods. Although the ordinary Raman can only provide limited information for coking at later stages, the in situ SERS can significantly enhance signal strengths and allow for the study of the carbon deposition at its incipient stage. In addition, the carbon deposition is very sensitive at early stages to surface modifications, gas compositions, electrical biases, and synergies on interfaces. Here, the in situ SERS allows for the capture of the subtle influences of these factors before electrode surfaces are overwhelmed by the growth of bulk-phase carbon, thus providing information not accessible by conventional tools.

4 Study of Coking-Resistant Catalysts

Before discussions of coking-resistant materials, a brief discussion of the mechanisms of the interaction between hydrocarbon molecules and nickel surfaces is necessary. As SOFC anodes are fed with hydrocarbon fuels, hydrocarbon molecules adsorb on Ni surfaces and undergo a series of dehydrogenation and C–C bond cleavage steps [61, 62], releasing H2 and forming surface carbonaceous groups (e.g., –C3H7, –C2H5, –CH3, and adsorbed atomic carbon Cad) on nickel surfaces. In one direction, generally referred to as “coking”, adsorbed carbonaceous groups coalesce and form surface carbon patches, filaments, or whiskers and are detrimental to anode performance and stability. In the other direction, adsorbed carbonaceous groups react with adsorbed H2O, CO2, and O2 to escape from the nickel surface without causing the carbon deposition. Here, the fine-tuning of the surface properties of nickel-based anodes and catalysts can promote the removal of adsorbed carbonaceous groups before the formation of coke and is a key approach to minimize coking [63, 64, 65].

Recently, several Ba-containing oxides, including BaO, BaZr1–xYxO3–δ (BZY), and BaZr0.1Ce0.7Y0.1Yb0.1O3–δ (BZCYYb) have been found to provide superior coking resistances to nickel-based SOFC anodes [3, 9, 30]. Here, Raman spectroscopy revealed the presence of surface –OH groups on these Ba-containing oxides, and the coking resistance capabilities of these oxides can generally be attributed to the water adsorption capacities on these materials. In addition, earlier studies of reforming catalysts also reported that alkaline earth oxides (e.g., MgO, CaO, SrO, and BaO) have a high tendency for water and CO2 adsorption and therefore can promote the reformation of carbonaceous species [66, 67, 68]. Furthermore, DFT simulation studies validated the fact that Ba-based perovskites are capable of the dissociative adsorption of water as a result of the low work function and high Fermi basicity of the BaO-terminated crystal faces [9, 69, 70].

However, due to the lack of the direct probing of interactions between hydrocarbon fuels and coking-resistant materials, current understandings of coking resistance mechanisms are largely based on thermodynamic theories and phenomenological interpretations based on material screening. Therefore, to advance the understanding of coking resistance mechanisms of hydrocarbons, the in situ Raman spectroscopy and the SERS have been employed to identify phase transformations and surface functional group evolutions on material surfaces upon exposure to systematically varied gas atmospheres. Here, the in situ spectroscopic fingerprinting of surface species can provide direct evidence that is critical to the interpretation of coking resistance mechanisms and allow for the study of kinetics through time-resolved analyses.

Coking can be inhibited on BaO-modified Ni anodes, and the same phenomena were ascertained by the in situ Raman analysis of BaO-modified Ni powders in exposure to hydrocarbons. As shown in Fig. 15, although peaks related to carbon formation initiated on plain Ni powders shortly after the introduction of propane-containing gases, it was largely inhibited on BaO-modified Ni powders over the inspected time frame. And in addition to the inhibition of coking, in situ Raman spectra also showed the emergence of a strong Raman peak at 1060 cm−1 on the BaO-modified Ni powder, suggesting the formation of carbonate (–CO3) species.
Fig. 15

In situ Raman analysis of Ni powders and BaO-modified Ni powders during propane exposure at ~ 450 °C. a Time-resolved Raman spectra collected on Ni powders and b BaO-modified Ni powders after wet propane were introduced. c Intensity of the carbon G-band (1580 cm−1) as a function of the duration of propane exposure. Peak intensities integrated within a 20 cm−1 bandwidth around the peak position. Reproduced with permission from [59]

Furthermore, to investigate the possible relationship between carbonate formation and coking-resistant capability, pure BaO and BZCYYb powders were exposed to propane-containing gases, and in situ Raman spectra were collected (shown in Fig. 16). Here, although no carbonate species were initially present, BaO formed strong carbonate species upon exposure to propane and concomitantly, the inherent –OH groups on BaO diminished. Detailed analyses of the results suggested that the appearance of carbonate species and the disappearance of the –OH groups were perfectly synchronized and that these two changes were irreversible if the BaO samples were re-exposed to wet 4% H2. In contrast, the BZCYYb sample showed both surface carbonate species and –OH groups initially and during exposure to propane, the Raman peak associated with carbonate groups decreased, along with an increase in carbon signals. In contrast to BaO, the change in surface species was found to be reversible on BZCYYb surfaces and after the reintroduction of wet 4% H2, carbonate peaks regained intensity, and carbon peaks disappeared. Furthermore, surface –OH groups on BZCYYb remained after the coking–regeneration cycle.
Fig. 16

In situ Raman spectroscopy study of BaO and BZCYYb powders upon exposure to different gas atmospheres. a Raman spectra collected from BaO at ~ 450 °C upon exposure to (1) dry 4% H2 (with 96% Ar), (2) dry 50% C3H8 (with 2% H2, 48% Ar), (3) wet 4% H2 (with 3% H2O, 93% Ar), (4) dry CO2, and (5) dry air. b The intensities of key spectroscopic features with respect to different gas species. c Raman spectra collected from BZCYYb powders at ~ 450 °C upon exposure to a variety of gas atmospheres, and d the corresponding intensities of key functional groups. All spectra were excited by using a 15 mW argon laser emitted at 514 nm. Feature intensity was calculated by averaging the signal count around the peak within the 20 cm−1 bandwidth. Reproduced with permission from [71]

In this study, in situ Raman spectroscopy data revealed different possible mechanisms for BaO and BZCYYb in terms of how they convey coking resistance to anodes. Here, BaO exhibited irreversible transformation from a –OH-rich surface to a –CO3-rich surface, whereas BZCYYb exhibited a more reversible change of surface functional groups. Synchrotron-based XRD validated this hypothesis by demonstrating that after exposure to propane, the BaO-modified Ni produced prominent patterns associated with the witherite structure of bulk BaCO3, whereas the mixture of Ni-BZCYYb did not form bulk carbonate after exposure to propane. This could be because the inert nature of the BaCO3 phase can prevent BaO from regenerating upon exposure to wet 4% H2. In contrast, carbonate species on BZCYYb naturally adsorb on surfaces and can participate in the reforming process of propane, corroborating with observed intensity decreases upon propane exposure.

The evolution of spectroscopic features on BaO and BZCYYb upon exposure to different atmospheres is summarized in Table 2. And based on the analysis of in situ Raman spectroscopy data, BaO shows an irreversible reaction with propane, whereas BZCYYb shows reversible fluctuations of surface functional groups. For BaO, inherent –OH groups could be surface water adsorbents or Ba(OH)2, which can desorb/decompose upon exposure to dry atmospheres, but regenerate in wet 4% H2. During exposure to propane, these –OH groups can assist in the reformation and production of CO2, which can react with BaO to form bulk BaCO3. This bulk BaCO3 is very stable and cannot be regenerated up to 700 °C. As for BZCYYb, –OH and –CO3 groups are inherently present on material surfaces. And because both species are surface adsorbents, they can participate in the reformation of propane. In addition, instead of forming bulk BaCO3, the –CO3 species on BZCYYb remain as surface adsorbents and can quickly regenerate upon exposure to wet 4% H2.
Table 2

Evolution of surface species upon exposure to systematically varied atmospheres as determined by the in situ Raman spectroscopy

 

Initial state

Dry–wet 4% H2 cycling

C3H8 exposure and regeneration

on BaO surface

Carbon D-band and G-band

None

None

None

H2O/–OH group

Present

Regenerate

Diminish irreversiblyd

Carbonate group

None

None

Form bulk carbonate irreversiblya

on BZCYYb surface

Carbon D-band and G-band

None

None

Increase then decrease

H2O/–OH group

Present

Regenerateb

Remains

Carbonate group

Present

Present

Decrease, then increase, no bulk carbonate formationa

All spectra information, unless otherwise specified, can be found in the Ref. [72]

aFormation of bulk carbonate is determined by the synchrotron-based X-ray diffraction [72]

bSpectra information can be found in the Ref. [30]

c–OH group signals during hydrocarbon exposure were interfered due to fluorescence, but a significant amount remained after re-exposure to wet 4% H2

dSame phenomena observed during exposure to CO2

The in situ Raman study of Ba-containing catalysts upon systematic exposure to coking and regeneration atmospheres assisted in the unraveling of mechanisms for their interaction with hydrocarbon fuels and provided guidance for the design of surface modifications for Ni-based SOFC anodes. Here, although strong basicity is beneficial for water adsorption, promoting the regeneration of hydrocarbons, strong basic substances such as BaO can also easily react with CO2, form inert carbonates and pose challenges to long-term performance. However, if Ba elements are embedded into perovskite structures, such as in the case of BZY and BZCYYb, they can provide surface –OH and –CO3 species that remain reversible upon exposure to hydrocarbon-containing fuels.

5 Study of Sulfur Poisoning on Nickel Surfaces

Sulfur poisoning is another important degradation mechanism of nickel-based SOFC anodes. If nickel anodes are exposed to fuels containing ppm levels of H2S, fuel cells will suffer significant performance losses. This poisoning effect of H2S is usually attributed to the adsorption of sulfur on nickel which blocks active sites for the adsorption of fuel molecules. And although the adsorption of sulfur is reversible and cell performances can recover over the short term, extended exposure to H2S-containing fuels will lead to the non-recoverable degradation of cell performances, which can be correlated with reactions between materials and contaminants [19, 73, 74, 75, 76]. To mitigate the effects of sulfur poisoning, alternative anode materials have been proposed and surface modifications to nickel anodes have been designed [30]. However, challenges still exist in the understanding of key fundamental mechanisms of sulfur poisoning, such as the concentration levels of H2S that induce sulfur adsorption, the key venues of sulfur adsorption that impair cell performance, and the mechanisms through which the poisoning effects can be mitigated.

The study of sulfur poisoning is difficult however, because H2S contamination levels of interest are usually low in the range of 0.1~10 ppm. And at such concentrations at operating temperatures (750 °C), the formation of nickel sulfides is not thermodynamically feasible, and therefore sulfur is present mainly in the form of surface adsorbents that can quickly desorb with pure H2 purging [20, 77]. This makes the identification and quantification of species related to S poisoning either in situ or ex situ difficult.

Previous in situ Raman spectroscopy studies of sulfur poisoning over Ni anodes lack spectral features related to the formation of sulfur or sulfides at operating temperatures. This is because adsorbed atomic sulfur on nickel surfaces may not generate enough Raman scattering upon laser excitation [78]. In addition, ex situ analysis may also be distorted by the route through which contaminated nickel foils cool to room temperature. As reported by Cheng et al. [79], if nickel foil cools slowly to room temperature in 100 ppm H2S, strong signals related to NiS and adsorbed sulfur can be resolved. However, if nickel foil cools to room temperature under H2 or quenched in H2S-containing gas, a better representation of operation conditions, species related to sulfur poisoning are absent. And given the fact that sulfur poisoning occurs at concentrations as low as 0.1 ppm, the sensitivities of ordinary Raman are insufficient for this study.

However, SERS can dramatically enhance sensitivities to surface sulfur species after H2S exposure, as shown in Fig. 17. Here, Ni foil was exposed to 0.5 ppm H2S at 767 °C for 12 h before being quickly quenched to RT (~ 300 °C/min) under the same H2S-containing condition. And prior to SERS treatment, the Ni foil showed no peaks related to sulfur deposition, but after Ag sputtering, the SERS phenomena revealed several Raman bands on the Ni foil related to the presence of elemental sulfur, NiS and Ni3S4. Such high sensitivities allow SERS to probe mechanisms related to sulfur poisoning.
Fig. 17

SERS identification of species related to S poisoning on Ni foil after heat treatment in H2S. a Ordinary Raman spectra of H2S-poisoned Ni foil (exposed to 0.5 ppm H2S at 767 °C for 12 hr, quenched in the same H2S-containing atmosphere). b SERS spectra of Ni foil annealed in pure H2. c SERS spectra of the same sample as in a. Ni foil exposed to different H2S levels (0.01 to 0.5 ppm) at 767 °C for 12 h and quenched in pure H2. d Exposure to 0.01 ppm H2S, e 0.1 ppm H2S, and f 0.5 ppm H2S, where ● designates peaks associated with NiS, and ▲ designates peaks associated with Ni3S4

Furthermore, SERS is capable of identifying sulfur deposition on nickel surfaces even at lower levels of H2S exposure. As presented in Fig. 17d–f, the SERS spectra from Ni coupons exposed to trace amounts of H2S (0.01 ppm, 0.1 ppm, and 0.5 ppm) at 767 °C for 12 h and the peaks related to sulfide formation can be clearly resolved. In addition, to eliminate the possibility for H2S to react with Ni during the cooling step, these three samples were purged with pure H2 for 2 min before being rapidly cooled to RT under pure H2. Here, the presence of Ni3S4 and NiS, even in samples purged with H2 during cooling, suggests that adsorbed sulfur was not fully desorbed before quenching and therefore can react with Ni to form NiS. Another non-intuitive observation was that the quantity of sulfur did not correlate with the concentration of H2S exposure (e.g., the 0.5 ppm exposed Ni coupon showed less NiS than the 0.01 ppm H2S-exposed sample), indicating factors that were not well controlled (such as quenching rates and crystal orientations on Ni) can impact the quantity of sulfur adsorption more than H2S concentrations.

The impact of grain orientations on sulfur deposition tendencies can also be studied by the SERS mapping of nickel coupons after exposure to H2S-contaminated fuels. Here, the intensity of the 520 cm−1 peak across a grain boundary on nickel coupons can be used to create a map of NiS deposition, as shown in Fig. 18 in which two grains showed drastic differences in NiS signals, suggesting different tendencies of sulfur adsorption. These results also suggest that if H2S is present in trace amounts, its adsorption on nickel surfaces is more dependent on surface conditions than thermodynamic isotherms. This was also suggested by previous DFT calculations [80, 81].
Fig. 18

SERS analysis of Ni foil after H2S exposure. a Scanned regions indicated in the optical microscope image. b Intensity map of the 520 cm−1 peak of the scanned region. Ni foil used for analysis was exposed to 0.5 ppm H2S for 12 h and quenched under pure H2

Furthermore, the in situ SERS analysis can also be deployed to analyze the formation of NiS on nickel surfaces, as displayed in Fig. 19. Here, the nickel coupon was loaded with thermally stable SERS probes and first heated in pure H2 to 450 °C. The sample was subsequently exposed to 0.5 ppm H2S-containing hydrogen for 12 h, but yielded no signals related to sulfur adsorption. Afterward, the environmental chamber was purged with pure H2, and the nickel foil was quenched to room temperature. Here, spectra collected during the quenching step showed that as temperatures drop, a peak at 510 cm−1 evolves, indicating the formation of NiS. This in situ SERS analysis result corroborates the implication from the ex situ analysis that adsorbed sulfur on nickel surfaces cannot be completely removed with hydrogen purging and will subsequently convert into NiS and Ni3S4 as temperatures quickly drop.
Fig. 19

In situ SERS analysis of Ni foil being exposed to H2S. a Raman spectra collected on nickel foil in situ, (1) before, and (2) after exposure to 0.5 ppm H2S at 400 °C, and (3) after quenching in pure H2. b Schematic showing the pathways of NiS, Ni3S4, and elemental sulfur formation

6 Study of the Surface Degradation of Cathode Materials

Long-term stability is another challenge for conventional cathode materials used in SOFCs. In particular, LSCF electrodes show steady degradation over long-term operations [82] caused by exposure to CO2-, H2O- and Cr-containing species, contaminating electrodes and slowing down electrochemical processes [83, 84, 85, 86]. In addition, the formation of cation segregation (e.g., La2O3, SrO, Co3O4, and Fe2O3) also correlates with performance degradation [85, 87].

Therefore, the identification of degradation modes for LSCF electrodes is crucial to the design of appropriate coatings or modifications for the alleviation of degradation. However, the characterization of surface species and incipient phases related to LSCF degradation is difficult due to the low concentrations of these surface species. Here, the surface-enhanced Raman spectroscopy can be used to significantly boost the sensitivity to perovskite phases of SOFC cathode materials and can identify phase distortions and surface contaminations related to the degradation process of cathode materials.

6.1 Identification of Bulk Phases and Surface Modifications

The SERS can enhance the inherently low-intensity signals of perovskite-based cathode materials. For example, the application of the SERS on cathode materials is displayed in Fig. 20 and demonstrates that significantly enhanced Raman spectra can be collected from both LSCF powders and LSM thin films after SERS conditioning. Here, it is worth noting that normal Raman signals for both materials are very weak because LSCF and LSM belong to \( {\text{R}}\bar{3}{\text{c}} \) and Pbmn space groups, respectively [88, 89], which represent near-cubic perovskite phases that have little Raman activity. As shown in Fig. 20a, the normal Raman spectra of pristine LSCF pellets show three strong modes (180, 320, and 720 cm−1) and three weak modes (540, 950, and 1130 cm−1) within the inspection range and that after the application of Ag nanoparticles, all Raman modes were enhanced by ~ 100 times. In addition, the SERS analysis of the 100 nm LSM thin coating on the YSZ substrate, as shown in Fig. 20a, not only demonstrates the capability of SERS to enhance LSM signals, but also demonstrates the strong surface specificity in which only LSM coatings showed enhanced signals.
Fig. 20

SERS capabilities for the analysis of cathode materials. Normal Raman and SERS spectra collected from a pressed LSCF powders, and b LSM films sputtered on a YSZ pellet. Argon laser emission lines at 514 nm were used for the excitation of Raman signal in this study

The unique sensitivity of the SERS toward surface species allows for identification of surface modifications. For example, to enhance the oxygen reduction activity and stability of LSCF porous electrodes, other perovskites, such as Pr(Ni, Mn)O3 (PNM), (La, Sr)MnO3 (LSM), and PrMnO3 (PM), are applied [90, 91]. And as shown in Fig. 21, the surface modification layer, despite the low concentration, manifests prominent spectral features under SERS analysis. Here, the LSCF alone does not show strong stretching modes because it belongs to the \( {\text{R}}\bar{3}{\text{c }} \) structure, which has higher crystal symmetry and the bands around 550 cm−1 and 660 cm−1 are assigned to antisymmetric stretching (AS) and symmetric stretching (SS) modes, respectively, of the Pbmn structure. Furthermore, the Pbmn structure, which has a lower symmetry than R3c, is characteristic of perovskites with Mn occupying B sites, whereas the change in the AS mode on the other hand is influenced by the doping of the host lattice by other cations [92].
Fig. 21

SERS analysis of cathode infiltration. The surface-enhanced Raman spectroscopy study of porous LSCF electrodes after surface modifications with Pr(Ni, Mn)O3 (PNM), (La, Sr)MnO3 (LSM), and PrMnO3 (PM). Argon laser emission lines at 514 nm were used for the excitation of Raman signals in this study

6.2 Analysis of LSCF Surface Degradation Over Long-Term Operations

The Raman signature introduced by lattice distortion is also useful in the characterization of LSCF cathode degradation. Here, LSCF powders are pressed into pellets, fired at 1080 °C for 10 h, and subsequently annealed at 750 °C for 500 h in a variety of atmospheres that simulate cathode operation conditions, including pure Ar, 1% O2, pure O2, 1% CO2, and 3% H2O. To enable ex situ SERS analysis, samples are loaded onto glass slides, followed by deposition of a thin layer of Ag nanoparticles (180 s at 8 W).

Presented in Fig. 22 are the SERS analyses of LSCF pellets after exposure to a variety of atmospheres including pure Ar, 1% O2, pure O1, 2% CO2, and 3% H2O. Here, SERS spectra of the pristine LSCF sample were subtracted from the atmosphere-treated samples to visualize differences, and from the results, it can be seen that annealing in Ar, 1% O2, and pure O2 does not result in any identifiable changes on LSCF surfaces, whereas annealing in 1% CO2 and 3% H2O produced slight increases in the 500~600 cm−1 range, suggesting the formation of a Pbmn phase [92]. This observed formation of the Pbmn phase is related to surface SrO segregation upon exposure to H2O- and CO2-containing atmospheres. And because SrO itself possesses no active Raman modes, the phase segregation can be identified indirectly. For example, according to the extensive XRD analysis of the bulk phase by Tai et al. [93], the perovskite structure of LSCFs shows orthorhombic distortion \( ( {\text{R}}\bar{3}{\text{c)}} \) at higher Sr content, which can transform into rhombohedral distortion (Pbnm) at lower Sr content. In addition, the emergence of small humps around 500~600 cm−1 related to the Pbnm distorted structure suggests the depletion of Sr from the LSCF lattice and thus is linked to SrO segregation [94].
Fig. 22

SERS analysis of LSCF pellets thermally treated in different atmospheres. a SERS spectra and b differential analysis of the SERS spectra with respect to pristine conditions

6.3 Study of Cr Poisoning on LSCF Cathodes

SERS probes can also be used in the detection of Cr poisoning on LSCF electrodes which are present in trace amounts. LSCF porous electrodes in physical contact with Cr alloys will result in strong Cr poisoning, and the degree of poisoning scales with the concentration of steam, as presented in Fig. 23. In this study, different water concentrations (3%, 5%, and 10%) were used during heat treatment of the samples at 750 °C for 66 h. And as the water concentration increased, obvious growths of signal intensities for SrCrO4 can be observed in which the 3% H2O + Cr-treated sample yielded a \( {\text{CrO}}_{4}^{2 - } \) peak of ~ 500 cts, and the 5% H2O + Cr sample showed ~ 2000 cts. Furthermore, the chromate in the 10% H2O + Cr sample yielded ~ 4500 cts. And because the SERS enhancement factor is uniform on different regions of the sample and each SERS spectra line is the average of the signals collected from ten spots, the intensities of the SERS signal can represent the quantity of \( {\text{CrO}}_{4}^{2 - } \) species on the topmost layer of LSCF electrodes.
Fig. 23

Impact of water concentration on the Cr poisoning of LSCF porous electrodes. a Schematic of the Cr-poisoning test. b SERS analysis of the LSCF porous electrode which is in physical contact with Cr alloy under atmospheres with different H2O concentrations

Even without direct contact with Cr sources, LSCF porous electrodes can show evidence of Cr deposition by using SERS analysis as displayed in Fig. 24. Here, on the sample treated with dry air, the working electrode showed a significant peak at 860 cm−1 that is related to the formation of \( {\text{CrO}}_{4}^{2 - } \), whereas the counter electrode of the same sample showed only a slight signal at 840 cm−1 that is also a \( {\text{CrO}}_{4}^{2 - } \)-related band. The 1060 cm−1 peak, on the other hand, produced similar sizes for both electrodes. In addition, on the working electrode (WE) of the sample treated with wet air, the Raman peak at 860 cm−1 was significantly stronger than that on the WE treated with dry air, whereas on the counter electrode (CE) of the H2O-treated sample, \( {\text{CrO}}_{4}^{2 - } \)-related peaks appeared at 840 and 860 cm−1, suggesting fine differences of chromate phases evolving on the surface.
Fig. 24

SERS spectra of porous LSCF electrodes after long-term operations without direct Cr source. Porous electrodes tested in a dry air and b wet air under polarizations. WE (working electrode) is biased with negative potential (cathodic bias), whereas CE (counter electrode) is biased with positive potential. ~ Ten data points were collected and averaged. The excitation laser was 514 nm at a power of 15 mW

6.4 Study of Cathode Modifications for Cr-Poisoning Resistance

Surface modifications of LSCF porous electrodes have been found to provide resistances to Cr poisoning, and two compositions were explored as surface modifiers: La0.8Ca0.2Ni0.4Fe0.6O3–δ (LCNF) and La0.8Ca0.2FeO3–δ (LCF). In this study, precursors of surface modifiers were fabricated by dissolving stoichiometric amounts of metal nitrates in a mixture of deionized water and ethanol, together with polyvinyl pyrrolidone (PVP) as a surfactant and glycine as a complexing agent. Following this, 5 μL of the stock solution was deposited onto the as-prepared LSCF surface and a porous cathode and dried overnight in ambient air followed by heat treatment at 900 °C in air for 1 h [90]. The resulting modified and blank LSCF porous electrodes were subsequently tested by using a Crofer 22 APU foil as the interconnect, and all cells were tested at 750 °C with 3% H2O and 1% CO2 being fed along with air to the cathode. As shown in Fig. 25, the cathodic overpotential of the blank LSCF and the catalyst-infiltrated LSCF cathode as a function of testing times under testing conditions is prone to Cr poisoning and that LSCF electrodes infiltrated with LCNF and LCF exhibited mitigated effects toward Cr poisoning.
Fig. 25

Electrochemical performance evaluations of Cr-poisoning resistance through surface modifications. Time dependence of cathodic overpotential for blank LSCF and catalyst-infiltrated LSCF cathodes in contact with Cr alloys at 750 °C at a constant voltage of 0.25 V in which air containing 3% H2O and 1% CO2 was fed to the cathode. Reprint from [90]

This mitigation of Cr poisoning by surface modifications can also be investigated with the SERS, and Fig. 26 displays the averaged SERS spectra collected from porous LSCF with and without surface modifications after accelerated Cr-poisoning tests. Here, all samples displayed characteristic Raman features associated with SrCrO4 (857 cm−1) and Cr2O3 (350 cm−1) and the intensities of the SrCrO4 bands decrease in the order of Blank > LCF > LCNF, suggesting less SrCrO4 formation on modified LSCF porous electrode surfaces. In addition, although the quantitative order of SrCrO4 concentrations was consistent with electrochemical testing results in which modified LSCFs suffered less Cr poisoning, the contrast was not significant.
Fig. 26

SERS analysis of Cr deposition on LSCF porous electrodes with and without modification. Spectra collected from the top surface of working electrodes (blank LSCF, LCF-modified LSCF, and LCNF-modified LSCF) of the symmetric cell. Each spectrum is an average of the spectra collected from 27 random points on each electrode surface

To estimate the distribution of Cr poisoning effect on electrode surfaces, SrCrO4 peak intensities were probed at 27 locations (uniformly distributed) on the electrode surface. Figure 27a shows the cumulative percentage of sampled points plotted with respect to SrCrO4 peak intensities, and compared with the blank LSCF electrode, the LCNF- and LCF-modified electrodes showed higher percentages of sampled points with lower SrCrO4 intensities. In addition, if sampled points with less than 200 counts in SrCrO4 peak intensity were deemed “Cr-free”, the resulting percentage of estimated “Cr-free” regions on the modified samples can reach about five times more than that on the blank LSCF electrode, as shown in Fig. 27b.
Fig. 27

Statistical analysis of the coverage of Cr-related species on LSCF porous electrodes with and without modifications. a Cumulative percentage of sampled points as a function of SrCrO4 band (857 cm−1) intensity. b Percentage of “Cr Free” regions in which less than 200 counts were detected around the SrCrO4 peak

Because the SERS is a surface-specific technique, any trace amounts of Cr deposition will register high-intensity SrCrO4 peaks on the Raman spectrum. Therefore, the analysis of SrCrO4 peak distribution statistics can help determine the coverage of Cr-poisoning species. And based on these statistical analyses, the LCNF- and LCF-modified LSCF porous electrodes were found to be resistant to Cr deposition on large portions of the surface. Here, the existence of SrCrO4 in certain regions of the modified samples can be attributed to the non-uniform distribution of the LCNF and LCF coating, which can be improved through optimizations in future studies.

7 Concluding Remarks

7.1 Limitations of Characterization Methods

Although the in situ Raman and the SERS can provide information that allows for the deeper understanding of chemical and electrochemical processes of SOFC electrodes, further research is needed to improve the reliability, sensitivity, and versatility of these analytical techniques. By improving these key aspects, the in situ Raman and the SERS can become powerful tools for the study of electrochemical, catalytic, and electrocatalytic systems.

7.1.1 Limitations of Fluorescence Artifacts

Fluorescence is a common challenge for the Raman spectroscopy. Because the fluorescence process has higher cross sections than Raman scattering, the background interference it brings forth often eclipses Raman spectral features of interest. In addition, the fluorescence process also impacts SERS analysis because the LSPR of Ag nanoparticles also enhances the fluorescence process.

Here, several strategies can mitigate the influence of the fluorescence background, in which the first relies on the ps time-resolved spectrometer for fluorescence rejection. The Raman scattering is a coherent process in which scattered photons are generated immediately when probed molecules interact with excitation photons, whereas the fluorescence is an incoherent process that involves the adsorption of excitation photons, the relaxation of excited molecules, and the emission of fluorescent photons [95]. Therefore, because the Raman scattering occurs within 1 ps upon excitation, and fluorescence is generated in the timescale of ns, these two processes can be separated temporally by using a pulsed excitation source and a picosecond gated detector [96, 97].

Another strategy for the reduction of fluorescence is to change the excitation laser energy so that the Raman spectra and fluorescent spectra do not overlap. This is possible because the energy of Raman scattered photons changes with the excitation photon energy while that of fluorescent photons does not change, which is determined largely by the electronic structure of the materials or the molecules under study. And based on this principle, researchers have developed Shifted Excitation Raman Difference Spectroscopy (SERDS), which extracts Raman bands by comparing spectra collected through different excitation sources [98, 99].

The above two strategies require significant modifications to spectrometers. When using common Raman spectrometer, post-processing of collected spectra can also assist fluorescence rejection [100, 101]. This is because fluorescence bands are usually broader than Raman spectra, and polynomial background fitting can often provide sufficient background cleaning. And in this study, the polynomial background fitting was adopted for the extraction of Raman spectral features. This method is useful and easy to apply, although some irregular fluorescence features cannot be removed and can obscure Raman features of interest.

7.1.2 Limitations of Spectral Feature Interpretation

Another limitation of the Raman spectroscopy is the assignment of spectral features for unknown species. Because vibration bands of material surfaces cannot be precisely predicted, assignments of spectral features to certain phases are empirical in general. In addition, existing databases for Raman spectra are not as comprehensive as those of XRD. As a result, many peak assignments need to be conducted through comparisons against literature figures. And for the same functional groups, vibration band energies vary when molecular bonding conditions change, making the assignment of functional groups even more challenging. Therefore, clear identification of material surface phases and intermediate species during catalytic reactions such as hydrocarbon reforming can be overwhelmingly complicated.

To address the difficulty of the phase identification in Raman spectroscopy studies of fuel cell electrode materials, data handling techniques such as principal component analysis (PCA) can be incorporated with current spectral analysis processes. By comparing spectra upon exposure to systematically varied atmospheres and temperatures, spectral features related to the change of surface phases can be identified and thus provide guidelines for mechanistic interpretations. Furthermore, databases of electrode materials under different environments can be constructed for the benefit of the research community of SOFCs and other electrocatalytic devices.

7.1.3 Limitations of Thermally Robust SERS Probes

Although Ag@SiO2 nanoparticles can serve as efficient, the thermally stable SERS probes at temperatures below 500 °C, they are not suitable for probing reactions at higher temperatures. Therefore, the design of SERS probes that are thermally and chemically stable at even higher temperatures is required. To achieve this goal, shell materials such as TiO2 and ZrO2, core materials such as gold, and larger particle sizes need to be tested.

In addition, although SiO2 shells are inert to most catalytic reactions, the presence of SERS probes can change the performance of model fuel cells because these nanoprobes can physically block some areas from accessing reacting species. Therefore, electrochemical testing results acquired in the presence of SERS probes may deviate from those under the conditions without the SERS probes. However, this effect can be corrected with careful calibration and analysis; it may not be a detrimental factor for in situ/operando SERS analysis because the main results in these tests are the Raman spectra rather than the electrochemical response. Regardless, caution is warranted when it is tried to extract quantitative electrochemical responses.

7.2 New Architectures for SERS study

7.2.1 Combinatorial Studies of Interfaces

The Operando SERS has shown great promise in the identification of critical surface species and reaction intermediates related to the hydrocarbon reforming and carbon deposition on nickel surfaces. With the high surface sensitivities provided by the SERS, time-resolved analysis can be performed on surface electrochemical processes. Here, patterned electrode and catalyst-patterning techniques can create multiple interfaces of interest on a single sample, facilitating the monitoring of reaction hot spots. As shown in Fig. 28, different venues corresponding to different electrochemical functional regions can be monitored on a model cell with patterned electrodes, as defined by the substrates [the electrode, the electrolyte and the triple phase boundary (TPB)], the surface modifications (modified and bare regions), and the electrochemical connectivity (with and without bias). In addition, the spatial resolution provided by confocal Raman spectroscopy systems allows for the mapping of critical species across interfaces, providing additional information on electrochemical processes.
Fig. 28

Critical surface and interface regions for the operando study of electrode behaviors

7.2.2 In Situ SERS with Embedded Probe Architecture

The SERS experiments conducted in this paper, either ex situ or in situ, are based on the configuration of “external probes”. In comparison, the architecture of “embedded probes” can provide higher enhancement factors. This is because catalyst materials can form uniform coatings on Ag (or Au) nanoparticle surfaces, allowing for more surface intermediate species within the region of the enhanced LSPR field. For example, YSZ-coated Ag nanoparticles can be used to study the oxygen reduction reactions on YSZ surfaces. In addition, the loading of a small amount of LSM or LSCF particles on Ag@YSZ probes can provide the SERS for the study of ORRs on cathode–YSZ interfaces. Similar architectures can be used for the study of hydrocarbon-reforming processes on Ni–catalyst interfaces. Here, by coating Ag or Au nanoparticles with Ni, followed by the deposition of nanosized catalysts such as doped ceria, the heterogeneous interface between nickel and CeO2 can be probed. Embedded SERS probes can also be used to monitor the chemical reactions during material fabrication processes. Overall, atomic layer deposition is an effective method to develop embedded probes, as demonstrated by Tian’s group and van Duyne’s group, who created thin film coatings using a variety of metals and ceramics on Ag nanoprobes. However, the uniform coating of Ag surfaces with other materials of interest (e.g., YSZ, GDC or BZCYYb) is challenging because the precursors of these metal oxides for the ALD process are underdeveloped (Fig. 29).
Fig. 29

In Situ Raman spectroscopy based on an “embedded probe” architecture. a For the study of cathode–electrolyte interfaces. b For the study of catalyst–Ni interfaces

Notes

Acknowledgements

This work was supported by the US Department of Energy (DOE) SECA Core Technology Program under Award No. DE-NT0006557 and DE-FE0031201, ARPA-E REBELS Program under Award No. DE-AR0000501, and by the HetroFoaM Center, an Energy Frontier Research Center funded by the US DOE, Office of Science, Office of Basic Energy Sciences (BES) under Award No. DE-SC0001061.

References

  1. 1.
    Haile, S.M.: Fuel cell materials and components. Acta Mater. 51, 5981–6000 (2003)CrossRefGoogle Scholar
  2. 2.
    Singhal, S.C.: Advances in solid oxide fuel cell technology. Solid State Ion. 135, 305–313 (2000)CrossRefGoogle Scholar
  3. 3.
    Yang, L., Choi, Y., Qin, W., et al.: Promotion of water-mediated carbon removal by nanostructured barium oxide/nickel interfaces in solid oxide fuel cells. Nat. Commun. 2, 357 (2011)CrossRefPubMedPubMedCentralGoogle Scholar
  4. 4.
    Murray, E.P., Tsai, T., Barnett, S.A.: A direct-methane fuel cell with a ceria-based anode. Nature 400, 649 (1999)CrossRefGoogle Scholar
  5. 5.
    Zhan, Z.L., Barnett, S.A.: Use of a catalyst layer for propane partial oxidation in solid oxide fuel cells. Solid State Ion. 176, 871–879 (2005)CrossRefGoogle Scholar
  6. 6.
    Zha, S.W., Moore, A., Abernathy, H., et al.: GDC-based low-temperature SOFCs powered by hydrocarbon fuels. J. Electrochem. Soc. 151, A1128–A1133 (2004)CrossRefGoogle Scholar
  7. 7.
    Liu, M., Peng, R., Dong, D., et al.: Direct liquid methanol-fueled solid oxide fuel cell. J. Power Sources 185, 188–192 (2008)CrossRefGoogle Scholar
  8. 8.
    Murray, E.P., Harris, S.J., Liu, J., et al.: Direct solid oxide fuel cell operation using isooctane. Electrochem. Solid State Lett. 9, A292–A294 (2006)CrossRefGoogle Scholar
  9. 9.
    Liu, M.F., Choi, Y.M., Yang, L., et al.: Direct octane fuel cells: a promising power for transportation. Nano Energy 1, 448–455 (2012)CrossRefGoogle Scholar
  10. 10.
    Williams, M.C., Strakey, J.P., Surdoval, W.A.: The U.S. Department of Energy, office of fossil energy stationary fuel cell program. J. Power Sources 143, 191–196 (2005)CrossRefGoogle Scholar
  11. 11.
    Ikeda, K., Hisatome, N., Nagata, K., et al.: Development of 25 kW class SOFC module. ECS Trans. 7, 39–43 (2007)CrossRefGoogle Scholar
  12. 12.
    Day, M., Swartz, S.L., Arkenberg, G.: NexTech’s flexcell technology for planar SOFC stacks. ECS Trans. 35, 385–391 (2011)CrossRefGoogle Scholar
  13. 13.
    Badding, M., Bouton, W., Brown, J., et al.: Ultra-low mass planar SOFC design. ECS Trans. 35, 465–471 (2011)CrossRefGoogle Scholar
  14. 14.
    Krumpelt, M., Krause, T.R., Carter, J.D., et al.: Fuel processing for fuel cell systems in transportation and portable power applications. Catal. Today 77, 3–16 (2002)CrossRefGoogle Scholar
  15. 15.
    Zha, S., Moore, A., Abernathy, H., et al.: GDC-based low-temperature SOFCs powered by hydrocarbon fuels. J. Electrochem. Soc. 151, A1128–A1133 (2004)CrossRefGoogle Scholar
  16. 16.
    Atkinson, A., Barnett, S., Gorte, R.J., et al.: Advanced anodes for high-temperature fuel cells. Nat. Mater. 3, 17 (2004)CrossRefPubMedGoogle Scholar
  17. 17.
    Lin, Y.B., Zhan, Z.L., Liu, J., et al.: Direct operation of solid oxide fuel cells with methane fuel. Solid State Ion. 176, 1827–1835 (2005)CrossRefGoogle Scholar
  18. 18.
    Ray, E.R., Maskalick, N.J.: Contaminant effects in solid oxide fuel cells. In: Joint Contractors Meeting on Advanced Turbine Systems, Fuel Cells and Coal-Fired Heat, Morgantown, WV (1993)Google Scholar
  19. 19.
    Zha, S., Cheng, Z., Liu, M.: Sulfur poisoning and regeneration of Ni-based anodes in solid oxide fuel cells. J. Electrochem. Soc. 154, B201–B206 (2007)CrossRefGoogle Scholar
  20. 20.
    Wang, J.H., Liu, M.: Computational study of sulfur–nickel interactions: a new S–Ni phase diagram. Electrochem. Commun. 9, 2212–2217 (2007)CrossRefGoogle Scholar
  21. 21.
    Badwal, S.P.S., Deller, R., Foger, K., et al.: Interaction between chromia forming alloy interconnects and air electrode of solid oxide fuel cells. Solid State Ion. 99, 297–310 (1997)CrossRefGoogle Scholar
  22. 22.
    Jiang, S.P., Zhang, J.P., Zheng, X.G.: A comparative investigation of chromium deposition at air electrodes of solid oxide fuel cells. J. Eur. Ceram. Soc. 22, 361–373 (2002)CrossRefGoogle Scholar
  23. 23.
    Kurokawa, H., Kawamura, K., Maruyama, T.: Oxidation behavior of Fe–16Cr alloy interconnect for SOFC under hydrogen potential gradient. Solid State Ion. 168, 13–21 (2004)CrossRefGoogle Scholar
  24. 24.
    Finsterbusch, M., Lussier, A., Schaefer, J.A., et al.: Electrochemically driven cation segregation in the mixed conductor La0.6Sr0.4Co0.2Fe0.8O3–δ. Solid State Ion. 212, 77–80 (2012)CrossRefGoogle Scholar
  25. 25.
    Harrison, W.A.: Origin of Sr segregation at La1–xSrxMnO3 surfaces. Phys. Rev. B 83, 155437 (2011)CrossRefGoogle Scholar
  26. 26.
    Nie, L.F., Liu, M.F., Zhang, Y.J., et al.: La0.6Sr0.4Co0.2Fe0.8O3–δ cathodes infiltrated with samarium-doped cerium oxide for solid oxide fuel cells. J. Power Sources 195, 4704–4708 (2010)CrossRefGoogle Scholar
  27. 27.
    Lou, X.Y., Wang, S.Z., Liu, Z., et al.: Improving La0.6Sr0.4Co0.2Fe0.8O3–δ cathode performance by infiltration of a Sm0.5Sr0.5CoO3–δ coating. Solid State Ion. 180, 1285–1289 (2009)CrossRefGoogle Scholar
  28. 28.
    Lynch, M.E., Yang, L., Qin, W., et al.: Enhancement of La0.6Sr0.4Co0.2Fe0.8O3–δ durability and surface electrocatalytic activity by La0.85Sr0.15MnOδ investigated using a new test electrode platform. Energy. Environ. Sci. 4, 2249–2258 (2011)Google Scholar
  29. 29.
    Guo, J.H., Xie, C., Lee, K.T., et al.: Improving the carbon resistance of Ni-based steam reforming catalyst by alloying with Rh: a computational study coupled with reforming experiments and EXAFS characterization. ACS Catal. 1, 574–582 (2011)CrossRefGoogle Scholar
  30. 30.
    Yang, L., Wang, S., Blinn, K., et al.: Enhanced sulfur and coking tolerance of a mixed ion conductor for SOFCs: BaZr0.1Ce0.7Y0.2–xYbxO3–δ. Science 326, 126–129 (2009)CrossRefPubMedGoogle Scholar
  31. 31.
    Nakamoto, K.: Infrared and Raman spectra of inorganic and coordination compounds, theory and applications in inorganic chemistry. In: Chalmers, J.M., Griffiths, P.R. (eds.) Handbook of Vibrational Spectroscopy. Wiley, Hoboken (2008)Google Scholar
  32. 32.
    Smith, E., Dent, G.: Modern Raman Spectroscopy: A Practical Approach. Wiley, Hoboken (2013)Google Scholar
  33. 33.
    Weber, W.H., Merlin, R.: Raman Scattering in Materials Science. Springer, Berlin (2000)CrossRefGoogle Scholar
  34. 34.
    McBride, J.R., Hass, K.C., Poindexter, B.D., et al.: Raman and X-ray studies of Ce1−xRExO2−y, where RE = La, Pr, Nd, Eu, Gd, and Tb. J. Appl. Phys. 76, 2435 (1994)CrossRefGoogle Scholar
  35. 35.
    Ahn, K., Yoo, D.S., Prasad, D.H., et al.: Role of multivalent Pr in the formation and migration of oxygen vacancy in Pr-doped ceria: experimental and first-principles investigations. Chem. Mater. 24, 4261–4267 (2012)CrossRefGoogle Scholar
  36. 36.
    Peng, C., Wang, Y., Jiang, K., et al.: Study on the structure change and oxygen vacation shift for Ce1–xSmxO2–y solid solution. J. Alloys Compd. 349, 273–278 (2003)CrossRefGoogle Scholar
  37. 37.
    Hernandez, W.Y., Centeno, M.A., Romero-Sarria, F., et al.: Synthesis and characterization of Ce1–xEuxO2–x/2 mixed oxides and their catalytic activities for CO oxidation. J. Phys. Chem. C 113, 5629–5635 (2009)CrossRefGoogle Scholar
  38. 38.
    Mineshige, A., Taji, T., Muroi, Y., et al.: Oxygen chemical potential variation in ceria-based solid oxide fuel cells determined by Raman spectroscopy. Solid State Ion. 135, 481–485 (2000)CrossRefGoogle Scholar
  39. 39.
    Choi, Y.M., Abernathy, H., Chen, H.T., et al.: Characterization of O2–CeO2 interactions using in situ Raman spectroscopy and first-principle calculations. ChemPhysChem 7, 1957–1963 (2006)CrossRefPubMedGoogle Scholar
  40. 40.
    Efthimiopoulos, I., Kunc, K., Vazhenin, G.V., et al.: Structural transformation and vibrational properties of BaC2 at high pressures. Phys. Rev. B 82, 134125 (2010)CrossRefGoogle Scholar
  41. 41.
    Blinn, K.S., Abernathy, H., Li, X.X., et al.: Raman spectroscopic monitoring of carbon deposition on hydrocarbon-fed solid oxide fuel cell anodes. Energy Environ. Sci. 5, 7913–7917 (2012)CrossRefGoogle Scholar
  42. 42.
    Cheng, Z., Wang, J.H., Choi, Y., et al.: From Ni–YSZ to sulfur-tolerant anode materials for SOFCs: electrochemical behavior, in situ characterization, modeling, and future perspectives. Energy Environ. Sci. 4, 4380–4409 (2011)CrossRefGoogle Scholar
  43. 43.
    Cheng, Z., Abernathy, H., Liu, M.: Raman spectroscopy of nickel sulfide Ni3S2. J. Phys. Chem. C 111, 17997–18000 (2007)CrossRefGoogle Scholar
  44. 44.
    Pomfret, M.B., Owrutsky, J.C., Walker, R.A.: High-temperature Raman spectroscopy of solid oxide fuel cell materials and processes. J. Phys. Chem. B 110, 17305–17308 (2006)CrossRefPubMedGoogle Scholar
  45. 45.
    Jeanmaire, D.L., Van Duyne, R.P.: Surface raman spectroelectrochemistry: part I. Heterocyclic, aromatic, and aliphatic amines adsorbed on the anodized silver electrode. J. Electroanal. Chem. Interfacial Electrochem. 84, 1–20 (1977)CrossRefGoogle Scholar
  46. 46.
    Albrecht, M.G., Creighton, J.A.: Anomalously intense Raman spectra of pyridine at a silver electrode. J. Am. Chem. Soc. 99, 5215–5217 (1977)CrossRefGoogle Scholar
  47. 47.
    Marotta, N.E., Barber, J.R., Dluhy, P.R., et al.: Patterned silver nanorod array substrates for surface-enhanced Raman scattering. Appl. Spectrosc. 63, 1101–1106 (2009)CrossRefPubMedGoogle Scholar
  48. 48.
    McLellan, J.M., Siekkinen, A., Chen, J.Y., et al.: Comparison of the surface-enhanced Raman scattering on sharp and truncated silver nanocubes. Chem. Phys. Lett. 427, 122–126 (2006)CrossRefGoogle Scholar
  49. 49.
    Baia, L., Baia, M., Popp, J., et al.: Gold Films deposited over regular arrays of polystyrene nanospheres as highly effective SERS substrates from visible to NIR. J. Phys. Chem. B 110, 23982–23986 (2006)CrossRefPubMedGoogle Scholar
  50. 50.
    Mortazavi, D., Kouzani, A.Z., Kaynak, A., et al.: Developing LSPR design guidelines. Prog. Electromagn. Res. Pier 126, 203–235 (2012)CrossRefGoogle Scholar
  51. 51.
    Willets, K.A., Van Duyne, R.P.: Localized surface plasmon resonance spectroscopy and sensing. Annu. Rev. Phys. Chem. 58, 267–297 (2007)CrossRefPubMedGoogle Scholar
  52. 52.
    Kerker, M., Wang, D.S., Chew, H.: Surface enhanced Raman scattering (SERS) by molecules adsorbed at spherical particles: errata. Appl. Opt. 19, 4159–4174 (1980)CrossRefPubMedGoogle Scholar
  53. 53.
    Stiles, P.L., Dieringer, J.A., Shah, N.C., et al.: Surface-enhanced Raman spectroscopy. Annu. Rev. Anal. Chem. 1, 601–626 (2008)CrossRefGoogle Scholar
  54. 54.
    Qian, X.M., Nie, S.M.: Single-molecule and single-nanoparticle SERS: from fundamental mechanisms to biomedical applications. Chem. Soc. Rev. 37, 912–920 (2008)CrossRefPubMedGoogle Scholar
  55. 55.
    Hu, M., Chen, J., Li, Z.Y., et al.: Gold nanostructures: engineering their plasmonic properties for biomedical applications. Chem. Soc. Rev. 35, 1084–1094 (2006)CrossRefPubMedGoogle Scholar
  56. 56.
    Li, X., Blinn, K., Fang, Y., et al.: Application of surface enhanced Raman spectroscopy to the study of SOFC electrode surfaces. Phys. Chem. Chem. Phys. 14, 5919–5923 (2012)CrossRefPubMedGoogle Scholar
  57. 57.
    Li, X.X., Blinn, K., Fang, Y.C., et al.: Application of surface enhanced Raman spectroscopy to the study of SOFC electrode surfaces. Phys. Chem. Chem. Phys. 14, 5919–5923 (2012)CrossRefPubMedGoogle Scholar
  58. 58.
    Kim, H., Lu, C., Worrell, W.L., et al.: Cu–Ni cermet anodes for direct oxidation of methane in solid-oxide fuel cells. J. Electrochem. Soc. 149, A247–A250 (2002)CrossRefGoogle Scholar
  59. 59.
    Li, X.X., Lee, J.P., Blinn, K.S., et al.: High-temperature surface enhanced Raman spectroscopy for in situ study of solid oxide fuel cell materials. Energy Environ. Sci. 7, 306–310 (2014)CrossRefGoogle Scholar
  60. 60.
    Li, X., Liu, M., Lee, J.P., et al.: An operando surface enhanced Raman spectroscopy (SERS) study of carbon deposition on SOFC anodes. Phys. Chem. Chem. Phys. 17, 21112–21119 (2015)CrossRefPubMedGoogle Scholar
  61. 61.
    Yang, M.L., Zhu, Y.A., Fan, C., et al.: DFT study of propane dehydrogenation on Pt catalyst: effects of step sites. Phys. Chem. Chem. Phys. 13, 3257–3267 (2011)CrossRefPubMedGoogle Scholar
  62. 62.
    Liu, H.Y., Yan, R.X., Zhang, R.G., et al.: A DFT theoretical study of CH4 dissociation on gold-alloyed Ni(111) surface. J. Nat. Gas Chem. 20, 611–617 (2011)CrossRefGoogle Scholar
  63. 63.
    Trimm, D.L.: Catalysts for the control of coking during steam reforming. Catal. Today 49, 3–10 (1999)CrossRefGoogle Scholar
  64. 64.
    Lee, W.Y., Hanna, J., Ghoniem, A.F.: On the predictions of carbon deposition on the nickel anode of a SOFC and its impact on open-circuit conditions. J. Electrochem. Soc. 160, F94–F105 (2013)CrossRefGoogle Scholar
  65. 65.
    Siahvashi, A., Chesterfield, D., Adesina, A.A.: Propane CO2 (dry) reforming over bimetallic Mo–Ni/Al2O3 catalyst. Chem. Eng. Sci. 93, 313–325 (2013)CrossRefGoogle Scholar
  66. 66.
    Asamoto, M., Miyake, S., Sugihara, K., et al.: Improvement of Ni/SDC anode by alkaline earth metal oxide addition for direct methane-solid oxide fuel cells. Electrochem. Commun. 11, 1508–1511 (2009)CrossRefGoogle Scholar
  67. 67.
    Gonzalez-Delacruz, V.M., Ternero, F., Pereñíguez, R., et al.: Study of nanostructured Ni/CeO2 catalysts prepared by combustion synthesis in dry reforming of methane. Appl. Catal. A Gen. 384, 1–9 (2010)CrossRefGoogle Scholar
  68. 68.
    Guo, J.J., Lou, H., Mo, L.Y., et al.: The reactivity of surface active carbonaceous species with CO2 and its role on hydrocarbon conversion reactions. J. Mol. Catal. A Chem. 316, 1–7 (2010)CrossRefGoogle Scholar
  69. 69.
    Shishkin, M., Ziegler, T.: Coke-tolerant Ni/BaCe1–xYxO3-δ anodes for solid oxide fuel cells: DFT plus U study. J. Phys. Chem. C 117, 7086–7096 (2013)CrossRefGoogle Scholar
  70. 70.
    Bandura, A.V., Evarestov, R.A., Kuruch, D.D.: Hybrid HF–DFT modeling of monolayer water adsorption on (001) surface of cubic BaHfO3 and BaZrO3 crystals. Surf. Sci. 604, 1591–1597 (2010)CrossRefGoogle Scholar
  71. 71.
    Li, X.X., Liu, M.F., Lai, S.Y., et al.: In situ probing of the mechanisms of coking resistance on catalyst-modified anodes for solid oxide fuel cells. Chem. Mater. 27, 822–828 (2015)CrossRefGoogle Scholar
  72. 72.
    Li, X., Liu, M., Lai, S.Y., et al.: In situ probing of the mechanisms of coking resistance on catalyst-modified anodes for solid oxide fuel cells. Chem. Mater. 27, 822–828 (2015)CrossRefGoogle Scholar
  73. 73.
    Brightman, E., Ivey, D.G., Brett, D.J.L., et al.: The effect of current density on H2S-poisoning of nickel-based solid oxide fuel cell anodes. J. Power Sources 196, 7182–7187 (2011)CrossRefGoogle Scholar
  74. 74.
    Cheng, Z., Wang, J.H., Choi, Y.M., et al.: From Ni–YSZ to sulfur-tolerant anode materials for SOFCs: electrochemical behavior, in situ characterization, modeling, and future perspectives. Energy Environ. Sci. 4, 4380–4409 (2011)CrossRefGoogle Scholar
  75. 75.
    Gong, M.Y., Liu, X.B., Trembly, J., et al.: Sulfur-tolerant anode materials for solid oxide fuel cell application. J. Power Sources 168, 289–298 (2007)CrossRefGoogle Scholar
  76. 76.
    Lohsoontorn, P., Brett, D.J.L., Brandon, N.P.: Thermodynamic predictions of the impact of fuel composition on the propensity of sulphur to interact with Ni and ceria-based anodes for solid oxide fuel cells. J. Power Sources 175, 60–67 (2008)CrossRefGoogle Scholar
  77. 77.
    Lee, K., Song, C.S., Janik, M.J.: Ab initio thermodynamics examination of sulfur species present on Rh, Ni, and binary Rh–Ni surfaces under steam reforming reaction conditions. Langmuir 28, 5660–5668 (2012)CrossRefPubMedGoogle Scholar
  78. 78.
    Cheng, Z., Wang, J.H., Choi, Y., et al.: From Ni–YSZ to sulfur-tolerant anode materials for SOFCs: electrochemical behavior, in situ characterization, modeling, and future perspectives. Energy Environ. Sci. 4, 4380–4409 (2011)CrossRefGoogle Scholar
  79. 79.
    Cheng, Z., Liu, M.: Characterization of sulfur poisoning of Ni–YSZ anodes for solid oxide fuel cells using in situ Raman microspectroscopy. Solid State Ion. 178, 925–935 (2007)CrossRefGoogle Scholar
  80. 80.
    Zha, S.W., Cheng, Z., Liu, M.L.: Sulfur poisoning and regeneration of Ni-based anodes in solid oxide fuel cells. J. Electrochem. Soc. 154, B201–B206 (2007)CrossRefGoogle Scholar
  81. 81.
    Hansen, J.B.: Correlating sulfur poisoning of SOFC nickel anodes by a Temkin isotherm. Electrochem. Solid State Lett. 11, B178–B180 (2008)CrossRefGoogle Scholar
  82. 82.
    Tietz, F., Haanappel, V.A.C., Mai, A., et al.: Performance of LSCF cathodes in cell tests. J. Power Sources 156, 20–22 (2006)CrossRefGoogle Scholar
  83. 83.
    Jiang, S.P., Zhen, Y.: Mechanism of Cr deposition and its application in the development of Cr-tolerant cathodes of solid oxide fuel cells. Solid State Ion. 179, 1459–1464 (2008)CrossRefGoogle Scholar
  84. 84.
    Simner, S.P., Anderson, M.D., Engelhard, M.H., et al.: Degradation mechanisms of La–Sr–Co–Fe–O3 SOFC cathodes. Electrochem. Solid State Lett. 9, A478–A481 (2006)CrossRefGoogle Scholar
  85. 85.
    Benson, S.J., Waller, D., Kilner, J.A.: Degradation of La0.6Sr0.4Co0.2Fe0.8O3–δ in carbon dioxide and water atmospheres. J. Electrochem. Soc. 146, 1305–1309 (1999)CrossRefGoogle Scholar
  86. 86.
    Lee, S.N., Atkinson, A., Kilner, J.A.: Effect of Chromium on La0.6Sr0.4Co0.2Fe0.8O3–δ solid oxide fuel cell cathodes. J. Electrochem. Soc. 160, F629–F635 (2013)CrossRefGoogle Scholar
  87. 87.
    Bucher, E., Sitte, W.: Long-term stability of the oxygen exchange properties of (La, Sr)1–z(Co, Fe)O3–δ in dry and wet atmospheres. Solid State Ion. 192, 480–482 (2011)CrossRefGoogle Scholar
  88. 88.
    Tai, L.W., Nasrallah, M.M., Anderson, H.U., et al.: Structure and electrical properties of La1–xSrxCo1–yFeyO3. Part 2. The system La1–xSrxCo0.2Fe0.8O3. Solid State Ion 76, 273–283 (1995)CrossRefGoogle Scholar
  89. 89.
    Mitchell, J.F., Argyriou, D.N., Potter, C.D., et al.: Structural phase diagram of La1–xSrxMnO3+δ: relationship to magnetic and transport properties. Phys. Rev. B 54, 6172–6183 (1996)CrossRefGoogle Scholar
  90. 90.
    Ding, D., Liu, M.F., Liu, Z.B., et al.: Efficient electro-catalysts for enhancing surface activity and stability of SOFC cathodes. Adv. Energy Mater. 3, 1149–1154 (2013)CrossRefGoogle Scholar
  91. 91.
    Lynch, M.E., Yang, L., Qin, W.T., et al.: Enhancement of La0.6Sr0.4Co0.2Fe0.8O3–δ durability and surface electrocatalytic activity by La0.85Sr0.15MnOδ investigated using a new test electrode platform. Energy Environ. Sci. 4, 2249–2258 (2011)CrossRefGoogle Scholar
  92. 92.
    Martín-Carrón, L., de Andrés, A., Martínez-Lope, M.J., et al.: Raman phonons as a probe of disorder, fluctuations, and local structure in doped and undoped orthorhombic and rhombohedral manganites. Phys. Rev. B 66, 174303 (2002)CrossRefGoogle Scholar
  93. 93.
    Tai, L.W., Nasrallah, M.M., Anderson, H.U., et al.: Structure and electrical-properties of La1–xSrxCO1–yFeyO3.2. The system La1–xSrxCO0.2Fe0.8O3. Solid State Ionics 76, 273–283 (1995)CrossRefGoogle Scholar
  94. 94.
    Barbero, B.P., Gamboa, J.A., Cadus, L.E.: Synthesis and characterisation of La1–xCaxFeO3 perovskite-type oxide catalysts for total oxidation of volatile organic compounds. Appl. Catal. B Environ. 65, 21–30 (2006)CrossRefGoogle Scholar
  95. 95.
    Weitz, D.A., Garoff, S., Gersten, J.I., et al.: The enhancement of Raman scattering, resonance Raman scattering, and fluorescence from molecules adsorbed on a rough silver surface. J. Chem. Phys. 78, 5324–5338 (1983)CrossRefGoogle Scholar
  96. 96.
    Everall, N., Hahn, T., Matousek, P., et al.: Picosecond time-resolved raman spectroscopy of solids: capabilities and limitations for fluorescence rejection and the influence of diffuse reflectance. Appl. Spectrosc. 55, 1701–1708 (2001)CrossRefGoogle Scholar
  97. 97.
    Matousek, P., Towrie, M., Ma, C., et al.: Fluorescence suppression in resonance Raman spectroscopy using a high-performance picosecond Kerr gate. J. Raman Spectrosc. 32, 983–988 (2001)CrossRefGoogle Scholar
  98. 98.
    McCain, S.T., Willett, R.M., Brady, D.J.: Multi-excitation Raman spectroscopy technique for fluorescence rejection. Opt. Express 16, 10975–10991 (2008)CrossRefPubMedGoogle Scholar
  99. 99.
    Li, Q., Wang, K.R., Wang, S.X.: A new approach for fluorescence subtraction in Raman spectroscopy. In: CLEO: 2011 - Laser Applications to Photonic Applications, OSA Technical Digest (CD), paper CFN7. OSA Publishing, Washington, D.C. (2011)Google Scholar
  100. 100.
    Lieber, C.A., Mahadevan-Jansen, A.: Automated method for subtraction of fluorescence from biological Raman spectra. Appl. Spectrosc. 57, 1363–1367 (2003)CrossRefPubMedGoogle Scholar
  101. 101.
    Ruan, H., Dai, L.K.: Automated background subtraction algorithm for raman spectra based on iterative weighted least squares. Asian J. Chem. 23, 5229–5234 (2011)Google Scholar

Copyright information

© Shanghai University and Periodicals Agency of Shanghai University 2018

Authors and Affiliations

  1. 1.School of Materials Science and Engineering, Center for Innovative Fuel Cell and Battery TechnologiesGeorgia Institute of TechnologyAtlantaUSA

Personalised recommendations