How do you read operando X-ray absorption spectroscopy plots during electrochemical cycling? Begin by finding the selected absorption edge, its energy scale, the corresponding applied potential or battery state, and the measurement mode. XANES and EXAFS answer related but different structural questions: the near-edge region is sensitive to electronic structure and local geometry, while the extended fine structure carries information about neighbouring atoms.
Operando XAS figures often display normalized absorption against photon energy, first derivatives of the XANES edge, k-space EXAFS oscillations, or Fourier-transformed EXAFS magnitude versus radial distance. A caption may show spectra collected while the device charges, discharges or operates at selected electrode potentials. These formats cannot be interpreted as ordinary diffraction peaks, and a visibly shifted edge is not automatically a direct measurement of one exact oxidation state.
This guide explains how to translate operando XANES and EXAFS data into careful statements about redox trends, coordination and local structure. It covers reference spectra, energy calibration, fitting uncertainty, beam effects, sample heterogeneity and the difference between measured changes and assigned mechanisms. The search intent is specifically electrochemical state-resolved XAS, not a generic introduction to static X-ray spectroscopy.
What makes a measurement operando
An operando XAS study collects element-specific absorption data while an electrochemical system performs its relevant function, often with simultaneous current, voltage or capacity records. The state of the working material may evolve during spectral acquisition. An in situ study also examines material in a controlled environment, but its operating conditions may differ from real use. Names alone do not tell the reader how faithfully operation was reproduced.
Ask whether a spectrum belongs to a dynamic sweep, a potentiostatic hold, charging, discharge or open circuit. If the beam was switched on only after long pauses, the observation describes a relaxed state rather than a true instantaneous operating state. Report the actual procedure and do not infer that every transient was resolved.
Understand element specificity
X-ray absorption near a selected elemental absorption edge preferentially probes atoms of that element, making XAS useful in complex materials. Yet the measured signal is still averaged over the illuminated atoms contributing to the detection mode. It cannot inherently reveal the unique behaviour of one catalytic surface site among many chemically different atoms of the same element.
For a mixed metal oxide, a nickel edge provides information about nickel-containing environments, not direct proof of how neighbouring iron atoms changed. To examine a second element, the experiment normally needs its own edge and compatible energy range. Translate element-resolved evidence precisely rather than treating the spectrum as a universal picture of the whole compound.
Read the photon-energy axis correctly
XANES spectra typically plot absorption-related response against incident photon energy, usually in electronvolts. A shift toward higher energy may occur with changes in effective oxidation state or local environment, but not every edge shift is a pure oxidation-state scale. Normalization, chemical bonding, geometry and final-state effects influence the observed spectrum.
Do not rewrite photon energy as electrode voltage. These are different axes with different physical meanings. The photon energy selects an X-ray transition; electrode potential controls the electrochemical system. A figure with several curves at different applied potentials overlays one variable on another, and translation must keep them separate.
Identify which absorption edge is measured
K, L and other absorption edges arise from different core-electron transitions and may show different information and selection-rule effects. A material’s nickel K-edge cannot be interpreted with the same diagnostic expectations as its L-edge without attention to the transition. Edge energy also determines accessible experimental geometry, beamline and detection constraints.
Read the edge name, element and energy window before naming a feature. Near-edge peaks sometimes called white lines are not generic chemical product peaks. Their shape and intensity are sensitive to electronic and geometric factors. A credible description identifies the measured edge and avoids transferring an assignment from a different edge solely because two plots look alike.
Separate XANES from EXAFS
XANES concerns the near-edge structure associated with absorption threshold and electronic or geometrical changes. EXAFS refers to oscillatory absorption modulation at energies farther beyond the edge, linked to photoelectron scattering from neighbouring atoms. The mathematical representation and models differ. One cannot simply read EXAFS oscillation amplitude as a concentration curve or XANES intensity as a diffraction peak height.
Explain whether an observation concerns edge position, near-edge shape, oscillation phase, fitted coordination or a Fourier-transform feature. A paper may use both regions to support a coherent interpretation, but each measurement has independent uncertainties. Accurate translation states which region underpins each claim.
Understand normalization and background removal
Raw absorption measurements are commonly corrected and normalized to allow comparisons across conditions. Pre-edge baseline subtraction and post-edge normalization influence the visual amplitude and shape. Different conventions can make otherwise similar spectra look surprisingly different. Poor background choice can shift apparent derivative maxima or amplify features that are not chemical changes.
When two XANES traces are overlaid, confirm that they received compatible processing. A tall normalized peak should not be described as twice as much absorbing material without quantitative calibration. Keep the difference between raw intensity, absorption coefficient and normalized absorption visible in the prose.
Read edge position as a trend, not an oxidation-state ruler
Edge energy can correlate with formal oxidation state when suitable references share similar chemistry, but its value also responds to ligand identity, covalency and geometry. A shift toward higher energy under charge may support oxidation of the absorbing element. It does not establish an exact integer valence or exclude compensating changes elsewhere.
Use known reference compounds and examine their comparability before assigning a nominal oxidation state. State ‘consistent with an increase in average oxidation state’ when justified. Avoid claiming every atom changes from one integer valence to another unless multiple pieces of evidence support that more precise model.
Check first-derivative XANES plots
First derivatives highlight rapid spectral changes and can help define an operational edge position. But the peak of a derivative depends on smoothing, interpolation and the spectral region selected. Multicomponent edges may exhibit shoulders that complicate a single peak-based estimate. Derivative plots magnify noise as well as genuine shifts.
If a study gives an edge-position change of a small fraction of an electronvolt, ask how the energy calibration and fitting uncertainty were established. A derivative maximum is a method-dependent descriptor, not a unique universal chemical threshold. Translate the operational definition instead of pretending every analysis extracts identical edge values.
Interpret white-line intensity with restraint
A strong near-edge resonance, often called a white line at some edges, can change with electronic occupancy, bonding, coordination and many-body effects. Its intensity sometimes relates to electronic vacancies under an appropriate model, but the relationship is not universal across different elements or chemical environments.
When a white line becomes more intense during battery charging, report the observed change before claiming a particular electron count. Compare a reference series and include independent electrochemical charge balance when possible. A simple rule such as ‘taller line always means more oxidation’ can be misleading in chemically complex systems.
Distinguish spectral shape changes from rigid shifts
An absorption edge can move as a whole, broaden, change relative peak heights or develop a new shoulder. These are not equivalent behaviours. A two-phase mixture may alter overall line shape differently from a homogeneous environment undergoing a gradual shift. A difference spectrum may appear derivative-like when one feature merely moves.
Overlay normalized spectra on a common calibrated energy axis and inspect more than one landmark. If a shoulder appears while the main edge barely shifts, report both facts. A useful translation does not compress every change into ‘the edge shifted’, because doing so may erase the strongest clue about heterogeneous chemistry.
Understand absorption measurements under electrical load
Electrochemical devices can contain current collectors, separators, electrolytes, windows and active layers that absorb or scatter X-rays. Experimental cell design determines which portion of a working electrode contributes most strongly to the signal. A configuration optimized for transmission may behave differently from fluorescence detection in a highly absorbing sample.
When comparing operando spectra, identify whether the beam viewed a thin film, packed electrode or complete battery cell. The measured signal may average both electrochemically active and inactive material. A spectral feature persisting through a cycle could originate partly from regions not taking part in the process.
Recognise transmission and fluorescence modes
Transmission XAS records attenuation through a sample, whereas fluorescence mode detects emitted X-ray photons after absorption. Each can be useful under different concentration and thickness conditions. Self-absorption, saturation and geometry-dependent distortions can alter fluorescence spectra, especially for concentrated or thick materials. These issues affect quantitative comparison of peak intensities.
State the detection mode when it matters to the interpretation. Do not directly compare fluorescence and transmission white-line heights without checking the processing and known distortions. The apparent strength of a spectral feature can reflect measurement geometry as well as material chemistry.
Read EXAFS oscillations in k-space
EXAFS data are often expressed as χ(k), an oscillatory signal against photoelectron wave number k, commonly in inverse ångströms. Analysts may multiply by k, k² or k³ to emphasize different portions of the spectrum. That weighting changes the appearance and relative prominence of the oscillations without changing the underlying measured sample.
A k³χ(k) plot cannot be described as a count of three-dimensional atoms. It is a weighted representation used for structural analysis. State the weighting before comparing amplitudes across papers, and retain the energy-to-k conversion assumptions. Higher-weighted noisy regions can become visually prominent even when their underlying signal is uncertain.
Interpret Fourier-transform EXAFS peaks
Fourier transformation of EXAFS oscillations provides a radial-distance-like representation of local scattering paths. The plotted magnitude versus R is convenient for visualising shells of neighbouring atoms. However, without phase-shift correction, a Fourier-transform peak position generally does not equal the actual interatomic distance. Calling every plotted peak a literal bond length creates a systematic interpretation error.
If a peak appears around one apparent radial coordinate, inspect whether the figure uses phase-corrected distance or uncorrected transform magnitude. The proper wording is ‘a Fourier-transform feature near the stated apparent R’ until the fitting model identifies corresponding distances. Different scattering atoms and paths can contribute in overlapping regions.
Coordination number is model-derived
EXAFS fitting may estimate an effective coordination number, the average number of neighbouring atoms in a specified scattering shell. This value depends on assumptions about scattering amplitude, disorder, energy offset and fitting constraints. It is correlated with other parameters and need not represent a clean integer, particularly in heterogeneous or nanoscale materials.
A fitted coordination of approximately five does not necessarily mean every metal atom has exactly five identical neighbours. Report a model-dependent average with uncertainty and note the relevant shell. A chemical model can describe a mixture of distinct environments even when the fitted average looks deceptively simple.
Bond lengths require phase-aware fitting
EXAFS can constrain local interatomic distances through fitting scattering models to oscillation phase and amplitude. The result should be distinguished from the raw Fourier-transform peak location. Distance estimates depend on usable k-range, data quality, model and whether neighbouring shells can be resolved.
When a study reports a small expansion during discharge, inspect the fit uncertainty and whether the same assumptions were applied across states. Translate the change as a fitted mean bond-distance trend where supported. Avoid turning a few hundredths of an ångström into an exact structural rearrangement if uncertainties overlap.
Debye–Waller factors describe disorder with caveats
An EXAFS fit often includes a mean-square relative displacement parameter related to disorder in a scattering path. It can reflect thermal motion, static variability or a mixture. Higher apparent disorder can suppress oscillation amplitude and correlate with coordination number. A changing fit parameter is not a direct photograph of atoms becoming random.
Discuss whether measurements were made at comparable temperature and whether alternative fitting choices could explain the amplitude change. In a working battery, heat and structural disorder may evolve together. Accurate wording distinguishes a fitted disorder parameter from a uniquely identified microscopic defect type.
Multiple scattering can matter
Photoelectron scattering may involve more than one neighbouring atom along a path, generating multiple-scattering contributions in EXAFS and near-edge structure. Ignoring such paths can lead to incorrect shell assignments. Their strength depends on geometry, particularly near-linear arrangements that can enhance certain contributions.
If a fit requires a feature beyond the first coordination shell, do not automatically label it a direct next-nearest neighbour. Consult the modelled scattering paths and structural references. Translate the fitted pathway at the level supported by the calculation. A radial feature may involve several atomic interactions rather than one straightforward bond.
Treat mixed phases as mixtures
Operando electrodes may contain several coexisting crystallographic or chemical phases. XAS averages contributions from absorbing atoms in those environments, making a mixed spectrum plausible. Linear-combination fitting against standards can estimate contributions if appropriate endmember references span the chemistry, but an excellent algebraic fit does not guarantee that the endmembers are chemically complete.
Report fitted component fractions as model-dependent estimates and mention the standards used. If a spectrum could also be explained by a continuously changing local environment, do not announce a sharp two-phase transition from fitting alone. Complementary diffraction, microscopy or thermodynamic evidence helps distinguish the scenarios.
Be cautious with linear-combination fitting
Linear-combination fitting assumes the sample spectrum can be represented as a weighted mixture of selected reference spectra after suitable processing. It can be extremely useful for tracking an evolving ensemble. The interpretation fails if references differ in relevant chemistry, spectral resolution, normalization or geometric effects, or if unknown components contribute materially.
When a paper claims a certain percentage of oxidized phase, check whether that percentage is a fitted spectral contribution or a directly measured mass fraction. Those quantities are not automatically identical. Translate the fitting procedure and its constraints, then explain the degree to which independent data support the composition estimate.
Connect state of charge to XAS observations
Battery figures may label spectra by state of charge, delivered capacity or cell voltage. These variables correlate with electrochemical progression but are not synonyms. A given voltage can occur on both charge and discharge with different internal states, especially when hysteresis or aging is present. Element-specific XAS reveals one part of the electrode chemistry alongside those global device measurements.
An edge moving during charge may indicate participation of that element in redox compensation. The extent of total battery charge also depends on other atoms, side reactions and electrode stoichiometry. Avoid assigning all stored charge to the measured element without a complete charge-balance argument.
Differentiate charge and discharge trajectories
A sequence of spectra collected during charge may not retrace the same path on discharge. Structural relaxation, kinetic lag, phase coexistence and irreversible processes can create differences at matched voltage or nominal state of charge. A hysteretic XANES feature alone does not prove permanent degradation; it may reflect rate-dependent and pathway-dependent conditions.
Compare matched electrochemical states with the same processing and clearly state which direction of operation each spectrum represents. If full recovery occurs after rest, a different conclusion follows than if the edge remains shifted after a complete cycle. Use specific recovery metrics instead of the single adjective ‘reversible’.
Interpret time resolution honestly
Operando XAS spectra require finite collection times and, in some modes, repeated scans. Fast structural transitions can occur during acquisition, yielding averages of evolving states. The ability to detect a transient depends on flux, acquisition settings, signal-to-noise and cell performance. A series of ten smooth spectra does not imply the sample remained constant while each was recorded.
If a current spike lasts less than a scan, avoid claiming the XAS frame shows its precise structure. Report a state averaged over the sampling interval and use time-aligned electrochemical records. Quick-scanning or time-resolved approaches may help, but the methods must demonstrate the relevant temporal resolution.
Consider beam-induced chemistry
Intense X-ray irradiation can change sensitive samples through heating, radiolysis or other beam-induced processes. This is especially relevant for electrolytes, redox-active species and certain soft X-ray measurements. A spectral change during repeated scans at fixed potential might be a beam effect rather than natural electrochemical aging.
Check whether the result depends on beam dose, exposure time or position. Replicate measurements at fresh spots or reduced flux can be informative. A strong mechanistic report should separate beam-driven changes from electrode-driven changes where possible. The X-ray probe is not always perfectly passive.
Watch energy calibration and monochromator drift
Meaningful edge shifts can be small, so energy stability is a central requirement. Beamline optics and monochromator calibration can drift over time. Reference foils or other standards measured concurrently or repeatedly allow correction. Spectra with unreliable relative energy calibration should not underpin fine oxidation-state comparisons.
Before treating a half-electronvolt shift as chemistry, inspect the reported calibration method and reproducibility. A figure may have many decimal places on the energy axis, yet uncertainty is set by measurement performance. Translate a modest edge movement with an appropriate qualifier when calibration limits are not clearly documented.
Explain pre-edge features carefully
Pre-edge peaks arise from transitions whose intensity and energy depend on electronic configuration, local symmetry and hybridization, among other factors. They can help distinguish coordination environments, but a stronger pre-edge feature is not automatically proof of higher oxidation state. Selection rules and symmetry breaking can have important effects.
Compare references sharing similar local geometry where possible and consider theoretical interpretations as models, not raw observations. If the study uses pre-edge intensity as a symmetry proxy, state that relationship with its assumptions. Avoid converting a subtle spectral shoulder into a definitive coordination geometry without complementary evidence.
Compare XAS with diffraction without confusing them
X-ray diffraction primarily probes periodic long-range structure, while XAS is sensitive to local absorbing-atom environments, including in materials lacking strong crystallinity. A material can show a clear change in local coordination before its average crystal structure changes appreciably. Conversely, a new diffraction phase can appear even when the average absorbing-atom edge changes only modestly.
Translate apparently different results as different sampling of structure rather than an automatic contradiction. If XRD identifies phase coexistence and XAS reports a redox trend, both can help explain the operating electrode. Refer to the eduKateSG XRD interpretation guide when interpreting diffraction alongside absorption.
Relate XANES to chemical valence vocabulary
Formal oxidation state is a bookkeeping concept useful for discussing electron transfer. XANES responds to actual electronic transitions and local chemical environment, which do not always map one-to-one to an integer bookkeeping label. Covalency and ligand-hole processes can make naive assignments misleading, especially in complex transition-metal oxides.
Prefer a phrase such as ‘the edge shift is consistent with a change in the average electronic environment associated with oxidation’ when reference calibration does not justify exact integer counting. State a formal valence only if the study gives credible supporting evidence. Precise terminology protects against a mechanistic story built on an oversimplified variable.
Account for inactive material in the beam
A heterogeneous electrode may contain electrically disconnected grains, unreacted cores or inactive phases that contribute to the XAS signal but not to electrochemical current. Since the measurement averages absorbing atoms, active changes can be diluted by an unchanged background. A small average edge shift does not necessarily imply every active site experiences a small change.
Look at electrode thickness, illumination geometry and complementary spatial characterisation. A mechanistic description should distinguish average bulk response from the potentially sharper change at active surfaces. Statements about microscopic site behaviour require sampling and models capable of resolving that distinction.
Identify solid–electrolyte interface contributions
An electrochemical electrode interacts with an electrolyte and may acquire interphase layers during operation. Depending on the measured edge and detection geometry, some of those layers can contribute to XAS. Their chemical state can change during cycling and complicate assignment of features solely to the intended active electrode phase.
When a newly developed edge feature is attributed to bulk cathode oxidation, ask whether an interphase or contaminant containing the same element could contribute. Use surface-sensitive comparison, reference spectra and compositional data where available. The correct translation may need to mention both bulk and interfacial possibilities.
Do not treat fitted coordination changes as exact atom loss
A decline in fitted coordination number can be associated with reduced average order, particle-size effects, changing phase fractions or amplitude and disorder changes. Correlations among EXAFS fitting parameters mean that several models may yield similar curves. The headline ‘one atom disappeared from each site’ is usually much stronger than what ensemble fitting establishes.
Report the numerical estimate with model assumptions and confidence intervals if supplied. Ask whether the result persists under reasonable alternative fits. Keep atom-count statements at the scale actually measured: an average local coordination environment, not a universally identical microscopic unit.
Distinguish element migration from oxidation
During cycling, some materials can undergo cation redistribution or migration between crystallographic sites, altering local coordination. XANES changes may also reflect altered valence and bonding. A spectral feature compatible with migration does not independently prove that atoms moved to a specific new site. EXAFS modelling and complementary diffraction or microscopy can improve the case.
An effective explanation separates changes in electronic signatures from changes in neighbour distances and coordination. If both appear, a structural transformation hypothesis strengthens. Still, distinguish inference from direct imaging. A coordinated argument should show how each observation rules out alternatives rather than merely collecting visually striking plots.
Read difference-XANES spectra
Subtracting a reference XANES curve from an operating-state curve accentuates changes but can produce positive and negative lobes from one shifted edge. Those lobes should not be labelled as newly formed positive and negative compounds. The difference is defined by the subtraction convention, which must be explicit.
If one edge shifts slightly to higher energy, a difference trace can look like a derivative with opposite-signed features. Check absolute spectra first. A careful translation reports a potential-dependent change relative to the stated reference condition, then discusses whether the shape is most consistent with a shift, a phase mixture or both.
Understand fluorescence self-absorption
In fluorescence-detected XAS, emitted photons can be reabsorbed within a thick or concentrated sample, distorting relative intensities. The severity depends on composition, geometry and energy. Normalizing such spectra does not necessarily remove the distortion. Comparisons of white-line amplitude and fitted concentrations are especially vulnerable when sample geometry changes during operation.
Ask whether the study used an appropriate correction or design that limits self-absorption. If not, give greater weight to robust relative trends than to absolute intensities. Translation should not present detector-dependent amplitude differences as precise changes in atom population without verification.
Consider temperature and mechanical effects
Charging, catalysis and device operation can change temperature, pressure and mechanical strain. EXAFS disorder parameters and local distances respond to these variables, sometimes alongside redox state. A structural shift observed under load may therefore reflect more than one physical cause.
Read the operational environment: temperature control, current density, electrolyte flow and cell compression. A comparison between spectra taken at different temperatures should not be interpreted as a purely electrochemical response without accounting for thermal effects. State measured conditions clearly so a reader knows which variables were actually controlled.
Use an EXAFS model with realistic degrees of freedom
EXAFS fitting becomes unstable when too many free parameters compete for limited independent information. Coordination, distance, energy offset and disorder can compensate for one another. A low residual by itself does not establish that the chosen atomic structure is the true one. Physically constrained models and uncertainty reporting are essential.
A report should explain its k and R fitting ranges, path assumptions and constraints, not merely display a near-perfect overlaid fit. When two structural models fit similarly, acknowledge the ambiguity. An attractive model drawing is an interpretation built from the data, not a direct X-ray photograph of every atomic neighbour.
Distinguish statistical precision from model validity
Repeated scans can reduce random noise, but systematic errors in energy calibration, self-absorption or model choice persist. A parameter written as 2.001 Å can look exact even if the structural model is uncertain. The numerical precision of a fit output should not be confused with real physical accuracy.
Translate fitted values with units and uncertainties and indicate which assumptions dominate the result. If uncertainty is not reported, avoid assigning significance to very small between-state differences. High-quality scientific prose is comfortable using ‘approximately’, ‘within the fitted model’ and ‘not resolvable from the present measurements’ when warranted.
Worked case: charging shifts the transition-metal edge
Imagine a cathode element’s XANES edge moves to higher energy as a cell charges, while the electrochemical record shows charge extraction. The observation is an element-specific spectral change correlated with charging. A reasonable hypothesis is increased average oxidation of that element, especially if chemically relevant references show similar trends.
Do not infer that every atom changed by the same integer valence or that all charge was compensated by the measured element. Review edge shape, white-line changes and any ligand-related evidence. The most defensible translation reports the trend, reference standard and possible redox contribution, reserving exact accounting for a full charge-balance analysis.
Worked case: EXAFS Fourier peak decreases
A Fourier-transform EXAFS magnitude peak becomes smaller after cycling. That may result from increased disorder, fewer average neighbours, phase mixing or geometry-dependent effects. It is not automatically a bond breaking event involving each absorber. Look at the corresponding χ(k) traces and fitted parameters before assigning a single cause.
If a credible model shows both altered disorder and a modest distance change, report those as fitted trends with uncertainty. Compare control samples and operating conditions. Avoid converting a lower Fourier peak into ‘the structure lost half its bonds’ unless a quantitative validated model actually supports such language.
Worked case: apparently reversible XANES but irreversible capacity
Suppose an electrode’s XANES edge returns near its initial position after discharge, yet the battery loses capacity over cycles. The spectroscopy suggests recovery of one average electronic signature, but capacity depends on many processes beyond the measured element’s edge. The observations are compatible rather than inherently contradictory.
Possible contributors include inaccessible active material, electrolyte side reactions, mechanical damage or transport limitations. XAS alone cannot allocate the cause. Translate the limited spectral recovery and the independent performance loss separately, then discuss which additional measurements would determine the dominant aging mechanism.
Worked case: a new shoulder under catalytic operation
A catalyst develops a small XANES shoulder at high anodic potential. The feature may indicate a modified local geometry or electronic environment, but identifying a unique transient species from one shoulder would be premature. Background choice, energy calibration and mixtures can create similar shapes.
Compare references, theoretical spectra and data at other operating states. Test whether the shoulder persists in repeated runs and whether it appears at a consistent potential. A scientifically careful summary calls it a potential-associated near-edge feature consistent with the proposed model, then states what independent evidence would strengthen the assignment.
Scientific language for XAS conclusions
‘Edge shifted’ is an observation. ‘Average oxidation state increased’ is an interpretation based on comparison. ‘A precise active site formed’ is a more specific structural claim. ‘That site caused the catalytic rate’ is a causal hypothesis. Moving between those phrases without added evidence is a common error in scientific communication.
A readable conclusion should name the element, edge, electrochemical state and supporting reference. It should differentiate model-fitted coordination and bond distance from raw spectral data. The word ‘operando’ establishes a measurement context, not a guarantee that the atomistic mechanism has been proven. Choose verbs and qualifiers according to the evidence available.
Related reading routes
To separate methods, compare the X-ray diffraction pattern guide, the battery charge–discharge guide and the GITT analysis guide. They support different observables: long-range crystal structure, device performance and diffusion-sensitive voltage response.
For precise scientific vocabulary and cross-topic reading, the eduKateSG vocabulary collection, retrieval and learning-transfer resources and English comprehension guide collection help readers practise matching a technical word to a supported claim.
Scientific sources and next checks
For detailed methodological grounding, see the open-access review of in situ and operando electrocatalyst characterisation using X-ray absorption spectroscopy. Its central caution is valuable: plausible-looking EXAFS fits and simple edge trends need careful modelling and experimental context. The study of phase-sensitive operando X-ray spectroscopy also illustrates why chemically active phases may not be represented faithfully by a single bulk-average curve.
Before accepting any interpretation, check reference standards, absorption edge and detector mode, energy calibration, acquisition timing, beam exposure, electrochemical state and uncertainty of fitted parameters. Use the data to constrain chemical stories rather than choosing a story first and making every spectral feature fit it. That is the core of responsible technical translation.