How to Learn X-Ray Photoelectron Spectroscopy and Surface Analysis: From the Photoelectric Effect to Chemical States, Depth Profiles and Operando Interfaces

Learning goal: Build XPS reasoning from photoemission energy conservation to core-level spectra, surface sensitivity, chemical shifts, spin–orbit structure, satellites, charging, energy referencing, background/line-shape models, quantification, angle-resolved and sputter depth profiling, hard/ambient-pressure photoelectron spectroscopy, operando interfaces, standards and defensible spectral fitting.
Scope boundary: Spectroscopy remains the canonical owner of general spectral fingerprints and light–matter measurement; Atomic Structure and the Periodic Table owns electron shells/orbitals; Surface Tension, Capillarity and Wetting owns interfacial thermodynamics; Vacuum Science and Thin-Film Deposition owns vacuum/deposition processing; Corrosion and Electrocatalysis own their reaction systems. This article owns how the kinetic energies of emitted photoelectrons are used to infer the elemental and chemical state of the outermost solid surface, and why charging, contamination and fitting choices can change that inference.
Reader-safety boundary: Educational surface/materials analysis only.

Wait, What? X-Rays Can Penetrate Deeply—but XPS Sees Only the Surface

X-rays can enter far into a solid.

So why is X-ray photoelectron spectroscopy surface sensitive?

Because the measured particle is not the X-ray.

It is the electron trying to escape.

Electrons created deep inside usually lose energy before reaching vacuum.

Only those born near the surface can escape with their characteristic kinetic energy intact.

So:

the surface sensitivity of XPS comes mainly from electron escape, not shallow X-ray penetration

The One-Sentence Answer

Learn XPS by tracing photon energy → photoelectron kinetic energy → binding-energy spectrum → chemically constrained peak model, while continuously checking surface contamination, sample charging, electron escape depth, energy referencing and whether the proposed chemical state is unique.

Stage 1: XPS Uses the Photoelectric Effect

An X-ray photon can eject an electron.

A simplified energy relation is:

E_binding = hν − E_kinetic − φ

where:

hν = photon energy;
E_kinetic = measured electron kinetic energy;
φ = spectrometer work-function term.

Measure kinetic energy.

Infer binding energy.

Stage 2: Core-Level Binding Energies Identify Elements

Atoms have characteristic core-electron energy levels.

An XPS spectrum can contain peaks such as:

C 1s;
O 1s;
Fe 2p;
Si 2p.

Peak identity constrains elemental composition.

Hydrogen and helium are generally not ordinary XPS targets.

Stage 3: The Survey Spectrum Answers “What Is Present?”

A broad energy scan reveals major elemental peaks.

It is useful for:

unexpected contamination;
approximate composition;
choosing high-resolution regions.

A survey spectrum is a map.

It is not yet detailed chemical-state analysis.

Stage 4: High-Resolution Scans Answer “In What State?”

Zoom into one core level.

Small binding-energy shifts can reflect different chemical environments.

Examples include changes in:

oxidation state;
bonding;
screening.

This chemical-state sensitivity is one of XPS’s major powers.

Stage 5: Chemical Shift Is Not a Simple Charge Meter

More positive oxidation state often shifts selected core levels to higher binding energy.

But binding energy also depends on:

final-state screening;
bonding;
local potential;
referencing.

“Peak moved 1 eV, therefore atom gained exactly this charge” is too literal.

Stage 6: Photoemission Creates a Final-State Problem

Removing a core electron leaves a hole.

The remaining electrons respond.

The measured peak therefore reflects both:

the initial electronic environment;
the final ionised state.

XPS is not a passive readout of isolated atomic orbital energies.

Stage 7: Spin–Orbit Coupling Splits Many Core Levels

For orbitals with non-zero angular momentum, peaks can split into components such as:

2p₃/₂;
2p₁/₂.

Their:

energy separation;
relative area

follow physical constraints.

Fitting the two peaks independently without those constraints can create nonphysical chemistry.

Stage 8: Satellites Are Real Physics, Not Automatically Extra Chemical Species

Spectra can contain:

shake-up peaks;
multiplet structure;
plasmon-loss features.

Transition-metal oxides are especially rich in such features.

Adding one chemical-state label per visible bump is often wrong.

Stage 9: Metallic Peaks Can Be Asymmetric

Metals contain low-energy electronic excitations.

Core-level peaks can develop asymmetric tails.

Line-shape choice therefore depends on electronic structure.

A symmetric Gaussian-like fit to every material is not physically justified.

Stage 10: Surface Sensitivity Comes From Inelastic Electron Scattering

Photoelectrons travelling through a solid can lose energy.

Their inelastic mean free path is often only nanometres at typical XPS kinetic energies.

Therefore the characteristic core-level peaks mainly represent the top few nanometres.

Stage 11: XPS Measures the Surface You Actually Prepared

Touch a sample.

Expose it to air.

Clean it aggressively.

All can change the outermost nanometres.

Because XPS is surface sensitive, preparation history can dominate the result.

The “bulk material composition” may not describe the XPS receiver.

Stage 12: Adventitious Carbon Is Common—but Energy Referencing Is Not Trivial

Air-exposed surfaces often carry hydrocarbon contamination.

Many analysts reference the C 1s adventitious-carbon peak to a conventional binding energy.

But the practice is not universally reliable.

Charging, substrate interactions and carbon chemistry can shift the peak.

Modern XPS practice treats referencing as an explicit methodological choice.

Stage 13: Insulators Can Charge Positively Under X-Ray Illumination

Photoelectrons leave the sample.

If electrons cannot flow back efficiently, positive charge accumulates.

This can:

shift peaks;
broaden peaks;
distort spectra.

A large binding-energy shift may be electrostatic, not chemical.

Stage 14: Charge Neutralisation Does Not Automatically Solve Referencing

Electron flood guns or combined electron/ion neutralisers can reduce charging.

But:

undercompensation;
overcompensation;
spatially differential charging

can remain.

A charge-corrected spectrum still needs a defensible energy reference.

Stage 15: The Background Contains Inelastic-Scattering Information

Electrons that lose energy contribute to a background under peaks.

Common background models include:

Shirley;
Tougaard.

They make different assumptions.

Background choice changes peak area and therefore quantification.

Stage 16: Peak Fitting Is Constrained Physical Modelling

A fit should use justified constraints from:

chemistry;
spin–orbit splitting;
area ratios;
known line shapes;
plausible widths.

A software optimiser can fit nonsense extremely well.

Residual minimisation is not chemical reasoning.

Stage 17: More Peaks Always Improve Mathematical Fit

Add enough components and the residual becomes smaller.

That does not mean the new species exist.

This is overfitting.

A peak component should have a physical reason to exist.

Stage 18: FWHM Carries Information—but Also Instrument Broadening

Peak width can reflect:

lifetime broadening;
chemical heterogeneity;
disorder;
charging;
instrumental resolution.

A broader peak does not uniquely prove “more disorder”.

Several mechanisms can produce the same visual symptom.

Stage 19: Quantification Uses Relative Sensitivity Factors

Peak areas depend on:

photoionisation cross-section;
analyser transmission;
electron escape.

Relative sensitivity factors convert measured intensity into approximate composition.

The 2024 edition of ISO 18118 formalises standard approaches to RSF-based quantification.

Stage 20: Atomic Percent Is a Model Output

Reported XPS atomic percentages assume:

a sampling model;
sensitivity factors;
background subtraction;
peak assignment;
reasonably uniform sampled region.

They are not direct atom counts.

Stage 21: Surface Layers Can Hide the Bulk

A thin oxide or contamination film can dominate the surface spectrum.

The substrate signal is attenuated exponentially.

This allows XPS to estimate overlayer thickness under suitable models.

But roughness and nonuniform coverage complicate the calculation.

Stage 22: Angle-Resolved XPS Changes Sampling Depth

Collect electrons at different emission angles.

More grazing emission gives greater surface sensitivity.

ARXPS can infer depth-dependent composition without sputtering.

But the reconstruction is model-dependent.

Stage 23: Sputter Depth Profiling Is Destructive

Ion bombardment removes surface material.

Repeated XPS scans can create a composition-versus-depth profile.

But sputtering can also cause:

reduction;
oxidation changes;
preferential sputtering;
atomic mixing;
roughening.

The measurement can create the chemistry it reports.

Stage 24: Cluster-Ion Sputtering Can Reduce Damage for Some Organics

Gas-cluster ion beams distribute impact energy across many atoms.

This can improve depth profiling of:

polymers;
organic films.

Damage is reduced, not eliminated.

Stage 25: Lateral Resolution Is Not Atomic

Microfocused XPS can map composition across a surface.

But spatial resolution is typically much coarser than electron or scanning-probe microscopy.

XPS owns chemical-state contrast.

It does not own every spatial scale.

Stage 26: Vacuum Protects Electron Transport

Conventional XPS operates under high or ultrahigh vacuum so emitted electrons can reach the analyser with limited gas-phase scattering.

Vacuum also slows fresh contamination.

The canonical Vacuum Science article owns the vacuum machinery.

Here, vacuum is part of the measurement path.

Stage 27: Vacuum Can Change the Surface

A hydrated, reactive or volatile surface under vacuum may not remain identical to its real operating state.

This creates a classic surface-science gap:

the easier a surface is to measure cleanly, the more artificial the environment may become

Stage 28: Air-Sensitive Materials Need Controlled Transfer

Battery electrodes, catalysts and reduced materials can change on air exposure.

Glovebox-compatible transfer systems can preserve surface state between preparation and XPS.

The sample’s journey is part of provenance.

Stage 29: Ultraviolet Photoelectron Spectroscopy Owns Valence States More Directly

UPS uses lower-energy ultraviolet photons and focuses strongly on:

valence electronic states;
work function.

XPS focuses heavily on core levels and chemical states.

The techniques are related but have different receivers.

Stage 30: Hard X-Ray Photoelectron Spectroscopy Looks Deeper

Higher-energy X-rays produce higher-energy photoelectrons with longer escape depths.

HAXPES can probe more deeply into:

buried interfaces;
multilayers;
devices.

Surface sensitivity becomes tunable rather than fixed.

Stage 31: Ambient-Pressure XPS Narrows the Pressure Gap

Special differential-pumping systems allow XPS at much higher gas pressures than conventional UHV.

This enables studies of surfaces under:

gases;
water vapour;
reactive environments.

The top few nanometres are observed closer to operating conditions.

Stage 32: Operando XPS Asks What the Active Surface Becomes

A catalyst before reaction may not equal the catalyst during reaction.

Operando/near-ambient-pressure XPS can track:

oxidation state;
adsorbates;
surface reconstruction

while chemistry is occurring.

This turns surface characterisation into state tracking.

Stage 33: Electrochemical Interfaces Are Especially Difficult

A working electrode is normally in contact with liquid electrolyte.

Electrons do not travel far through dense liquid.

Special thin-film, meniscus or membrane approaches are needed.

The measurement geometry can strongly constrain the chemistry being observed.

Stage 34: Battery Interfaces Reveal Why Surface Provenance Matters

The solid-electrolyte interphase can be only nanometres thick and highly reactive.

XPS is ideal in sampling depth.

But:

air exposure;
washing;
vacuum transfer;
sputtering

can change the SEI.

The analysis protocol can alter the object of interest.

Stage 35: Corrosion Films Are Layered Chemical Histories

XPS can distinguish selected oxide/hydroxide states near metal surfaces.

But the canonical Corrosion article owns the degradation mechanism.

This page asks:

what chemical surface state does the spectrum support?

Stage 36: NIST and ISO Standards Matter Because Peak Positions Are Not Enough

Reliable XPS depends on calibrated:

energy scale;
intensity response;
sensitivity factors.

The NIST XPS database and ISO surface-analysis standards support comparable interpretation.

Metrology is part of spectroscopy.

Stage 37: 2026 Practice Is Becoming More Explicit About Fitting Discipline

A 2026 Journal of Vacuum Science & Technology A paper on essential XPS practice emphasises:

photoemission physics;
transport;
data interpretation.

A 2026 open-source fitting project, KherveFitting, reflects growing interest in transparent fitting workflows.

The trend is away from invisible “click-to-fit” analysis.

Stage 38: Machine Learning Can Assist Spectral Analysis

AI can help with:

peak detection;
classification;
fitting suggestions;
large spectral datasets.

But training labels often come from human interpretations.

A model can reproduce systematic misassignment quickly.

AI needs physics constraints and independent validation.

Stage 39: Professional XPS Is a Surface-Provenance-and-Model Problem

The professional question becomes:

Which surface layer produced this photoelectron feature, how were charging and energy scale controlled, which physically constrained peak model supports the proposed chemical state, and what alternative explanation remains possible?

Evidence: What Makes an XPS Chemical-State Claim Strong?

Stronger evidence combines:

calibrated energy scale;
reproducible sample preparation;
justified referencing;
appropriate charge control;
chemically constrained spin–orbit/satellite fitting;
known standards;
complementary diffraction, spectroscopy or microscopy.

A single fitted component colour is weak evidence by itself.

Misconceptions Worth Hunting

XPS measures the bulk because X-rays penetrate deeply.
Every visible peak is a different element.
Every shoulder is a different chemical compound.
Binding energy directly equals atomic charge.
Charge neutralisation automatically gives correct absolute binding energies.
Adventitious carbon referencing is universally exact.
Peak fitting is objective if the residual is small.
Sputter profiling reveals the original depth profile without changing it.
Atomic percent is a direct atom count.
Ambient-pressure XPS completely reproduces real operating conditions.

Transfer Check

An insulating oxide shows every peak shifted +4 eV compared with literature.

Did every element undergo a chemical change? Not necessarily. Charging is a first hypothesis.

A metal 2p spectrum has extra broad features.

Should each be assigned to a new oxidation state automatically? No. Satellites and multiplets may contribute.

After sputtering, a metal oxide appears more reduced.

Did the buried layer necessarily start reduced? No. Ion bombardment may have reduced it.

Two peak models fit equally well.

Can the lower residual alone choose the chemical interpretation? No. Physical constraints and complementary evidence are needed.

How We Know the Learning Has Held

A learner should be able to:

explain the photoelectric equation;
distinguish survey and high-resolution scans;
explain surface sensitivity through electron escape;
explain chemical shifts cautiously;
interpret spin–orbit splitting and satellites;
explain charging and referencing;
compare Shirley and Tougaard backgrounds conceptually;
explain constrained peak fitting;
explain sensitivity-factor quantification;
explain ARXPS and sputter profiling;
distinguish conventional, hard-X-ray and ambient-pressure XPS;
evaluate operando and AI-assisted claims.

Model Limits

XPS usually samples only a few nanometres.

Vacuum may alter reactive surfaces.

Charging can create nonuniform shifts.

Binding-energy assignments are not unique.

Peak fits can overfit.

Depth profiles can be damaged by sputtering.

Sensitivity-factor quantification assumes simplified geometry.

Professional XPS therefore keeps:

surface provenance + photon energy + electron escape + charging + reference scale + spectral model + sampling depth + complementary evidence

visible together.

Teaching Guide

Teach in this order:

photoelectric effect → kinetic energy → binding energy → core levels → surface sensitivity → chemical shifts → spin–orbit/satellites → charging → referencing → background → peak fitting → quantification → depth → ambient/operando → validation.

Begin with:

“If X-rays penetrate micrometres into a sample, why can XPS still be one of the most surface-sensitive techniques in materials science?”

Connect This to the eduKate Learning Estate

Research Foundations and Further Learning

Greczynski et al., Essential principles and practices in x-ray photoelectron spectroscopy — Journal of Vacuum Science & Technology A, 2026.
A step-by-step guide to perform x-ray photoelectron spectroscopy — Journal of Applied Physics.
NIST X-ray Photoelectron Spectroscopy Database: https://srdata.nist.gov/xps/
ISO 18118:2024 / summary: Standard Approaches to XPS and AES Quantification—A Summary of ISO 18118:2024 on the Use of Relative Sensitivity Factors.
Mendoza-Sánchez et al., Charge compensation and charge correction in X-ray photoelectron spectroscopy — Surface and Interface Analysis, 2024.
KherveFitting open-source XPS fitting framework — Surface and Interface Analysis, 2026.

The Quiet Ending

The beginner asks:

“What element made this peak?”

The developing surface scientist asks:

“What chemical state shifted it?”

The advanced learner asks:

“Could charging, satellites or fitting choices explain the same feature?”

And the professional asks:

Which surface state is uniquely supported after energy referencing, escape-depth physics, charge control and physically constrained fitting have all been accounted for?