SEM vs TEM: Why Do Nanoparticles Look Uniform but Agglomerated?

For battery materials, catalysts and ceramic nanopowders, it is a widespread characterization paradox: SEM presents a homogeneous powder surface, while TEM on the identical sample displays obvious particle clusters or agglomerates. This conflict does not mean either microscope delivers wrong results or the powder changed during testing. SEM and TEM rely on distinct electron signals, different observation scales, and separate sample preparation workflows. Critically, nanopowders have a multi-level hierarchical structure at the same time.

To interpret results correctly, we must answer four core questions: What structural level are we observing? What signal forms the image? How is the specimen prepared? What definition of particle size is adopted?

This article summarizes the root causes of SEM-TEM visual divergence, artifacts introduced by drying and dispersion, and how to combine electron microscopy with BET, XRD, DLS, SAXS to judge the real dispersion status of nanopowders.

1. Clarify Definitions: Primary Particles, Aggregates and Agglomerates

Misunderstanding of terminology is the primary source of confusion. Nanopowders often contain multiple structural levels, including primary particles, aggregates and agglomerates. Depending on synthesis, surface chemistry and handling history, one or more of these levels may coexist in the same powder.

• Primary Particle: The basic identifiable unit formed during material synthesis. It can be single-crystal, polycrystalline, porous or surface-coated. Its size does not equal crystallite size.
• Aggregate: Particle assembly with strong bonding via necking, fusion or chemical bonds. Ordinary stirring or sonication cannot easily separate its subunits. It acts as an integrated structural unit.
• Agglomerate: Loose assembly of particles/aggregates held by van der Waals force, electrostatic force or capillary force. These clusters can be partially broken during dispersion, subject to surface chemistry, solvent, dispersant and input energy.

Key takeaway: Primary particle, aggregate and agglomerate belong to different structural hierarchies. One single average particle size cannot describe all three.

2. Core Reason: Micrometer Uniformity ≠ Nanoscale Dispersion

Both SEM and TEM observations can be true at the same time. At low SEM magnification, a “uniform” appearance may mainly reflect relatively uniform coverage at the micrometer scale, while nanoscale primary-particle contacts remain unresolved. It does not automatically mean all nanoscale primary particles are fully separated. TEM operates at far higher resolution and reveals that this seemingly uniform powder layer is built from closely contacted, overlapped nanoparticles.

For battery electrode powders, this distinction matters greatly. Macroscopically homogeneous powder may still contain hidden nanoclusters, which degrade slurry viscosity, conductive network construction and coating quality.

3. SEM and TEM Work on Different Imaging Contrast Mechanisms

SEM: Surface topography imaging

Standard SEM collects secondary electrons to visualize exposed contours, edges and surface roughness. The signal originates from the near-surface region. Strength: characterize powder shape, large agglomerates, particle packing and surface defects. Limitation: Subsurface particles are blocked by upper particles. Its large depth of field makes stacked powder look more continuous than its real 3D structure.

TEM: 2D projection through electron-transparent samples

Bright-field TEM contrast can arise from mass-thickness contrast, diffraction contrast and imaging conditions. In many nanoparticle specimens, thicker or more strongly scattering regions appear darker, but crystal orientation, diffraction condition and specimen thickness can also strongly modify image intensity. A TEM micrograph is a planar projection of 3D material along the electron beam.

Critical artifact: Two particles stacked along beam direction will overlap in projection and form a dark continuous region, even if they have no physical fusion. Tilt-series imaging or electron tomography can provide much stronger evidence for distinguishing true 3D contact from projection overlap.

4. TEM and SEM Each Have Their Own Hiding/amplifying Effects

1. TEM exaggerates clustering Particle stacking within the thin specimen enhances scattering and creates dense dark clusters. The effect worsens with high-Z materials, thicker deposits and poor contrast separation. Besides, operators tend to image thinner transparent regions, bringing sampling bias.
2. SEM conceals nanoscale agglomeration Low magnification cannot resolve tiny boundaries between primary particles. Conductive coating (Au/Pt/carbon) adds extra thickness and masks nano gaps. For nanoparticle boundary measurement, the conductive coating layer should be kept as thin as possible.

5. Sample Preparation Is Often the Biggest Variable

Dry SEM samples and drop-cast TEM samples are not identical specimens.

• SEM dry powder preparation: Mount powder on conductive tape. Spreading, tapping or pressing compacts weak agglomerates and reduces voids. The final image reflects mounted dry powder status, not dispersion state in battery slurry.
• TEM drop-casting preparation: Powder → solvent dispersion → sonication → drop on grid → solvent evaporation. Sonication may break loose agglomerates but leave strong aggregates intact. If colloidal stabilization is insufficient, particles re-cluster during solvent evaporation.

6. Drying Artifact: Coffee-Ring Effect Creates Artificial Agglomerates

Even well-dispersed suspension can form clusters after drop casting. During droplet drying, outward capillary flow associated with the coffee-ring effect may transport particles toward the contact line. Dewetting and capillary attraction can further promote clustering, while Marangoni circulation may either redistribute particles or suppress/enhance edge accumulation depending on the surface-tension gradient.

Important rule: Agglomerates observed on dried TEM grids do not prove the same clusters existed in original liquid slurry. This is the major limitation of conventional drop-cast TEM.

7. How to Distinguish Real Aggregation vs Drying Artifacts

No single visual rule works, but morphology gives clues:

• Drying-induced artificial clusters: particle chains, edge accumulation, 2D irregular islands; individual primary particle contours remain intact inside clusters.
• Strongly bonded or fused aggregates: may show particle necks, fused interfaces or other persistent interparticle connections that remain after controlled dispersion.

Single TEM micrograph cannot confirm whether clusters exist in liquid phase. The reliable solution is combining microscopy with ensemble liquid-state testing before drying.

8. Common Misconceptions Between Particle Characterization Techniques

1. DLS hydrodynamic size ≠ SEM/TEM geometric size SEM/TEM measures projected geometric diameter from static dried particles. DLS calculates hydrodynamic diameter of moving scattering entities in liquid via Brownian motion. DLS intensity signal heavily weights larger clusters. A 30 nm primary particle in TEM and 200 nm DLS size are not contradictory.
2. BET gives specific surface area, not direct particle size The BET equivalent diameter is calculated by formula with assumptions of density, shape and nonporous structure. Porosity, rough surface or particle fusion will make BET-derived value deviate from microscopy size. Treat BET as a surface-area indicator for battery powders.
3. XRD crystallite size ≠ particle size XRD estimates coherent diffraction domain size of crystal lattice. One primary particle can contain multiple crystallites. XRD peak broadening also includes instrument broadening and microstrain contributions.
4. SAXS captures nanoscale structure in liquid suspension SAXS detects electron density fluctuation in bulk sample and can obtain particle distribution and aggregate structure in suspension. However, SAXS results rely on structural models and cannot be treated as absolute “true size”.

Laser diffraction is mainly suitable for larger submicron-to-micron particle populations and agglomerates; it should not automatically be interpreted as primary nanoparticle size.

9. Systematic Framework to Evaluate Real Agglomeration Status

Do not ask “trust SEM or TEM”. Instead define what each technique measures and sample conditions.

Question	Recommended Characterization
Micrometer morphology of dry powder	Low/medium magnification SEM
Primary particle shape	HR-SEM / TEM
Fused aggregate or projection overlap	HR-TEM, tilt imaging / tomography
Dry-powder/submicron–micron particle distribution	Laser diffraction, where applicable
Nanoscale particles/clusters in dilute suspension	DLS / SAXS / suitable nanoparticle sizing method
Accessible specific surface area	BET
Crystal coherent domain size	XRD peak fitting
Long-term dispersion stability	Sedimentation test, rheology

The target is not to unify all numerical results, but to build a consistent structural interpretation from multi-dimensional measurements.

10. Practical Impact on Battery Electrode Manufacturing

Nanopowder agglomeration is not merely a microscopy puzzle. It can strongly influence slurry processing and electrode quality. Slurry preparation follows three sequential steps: Wetting → Dispersion → Stabilization. Poor dispersion causes coarse agglomerates, coating streaks and broken conductive networks. Excessive shear cannot solve the issue and may damage binders or conductive additives.

A nice SEM image alone cannot predict powder coating performance. Material evaluation must link powder morphology to slurry rheology and electrode fabrication results.

11. Standardized Workflow for Reliable SEM-TEM Comparison

1. Characterize as-received dry powder by SEM under multiple magnifications.
2. Prepare controlled liquid dispersion and fully record solvent, dispersant, solid content, sonication parameters and standing time.
3. Acquire low-mag TEM for cluster distribution and high-mag TEM for primary particles. Capture multiple random fields rather than only the clearest area.
4. Pair electron microscopy with at least one ensemble characterization method.
5. Correlate all data to slurry and electrode performance, instead of stopping at powder morphology observation.

12. Typical Interpretation Pitfalls

1. Low-mag SEM uniform appearance = well-dispersed nanoparticles → It only shows evenly distributed large agglomerates.
2. Dark TEM clusters = severe inherent agglomeration of raw powder → Projection overlap and drying artifacts can exaggerate clustering.
3. DLS size should match TEM primary particle size → Two parameters measure different physical properties.
4. BET equivalent diameter equals real primary particle size → It is a model-calculated value.
5. XRD crystallite size equals particle size → XRD measures crystal coherent domains.
6. Longer sonication always achieves better dispersion → Surface stabilization and solvent chemistry are decisive; over-sonication may damage particles.
7. One TEM micrograph represents the whole batch morphology → EM samples tiny volume; multi-field and ensemble tests are mandatory.

FAQ

Q1: Why nanopowder looks uniform in SEM but agglomerated in TEM?

A: SEM and TEM observe different scales and imaging signals. Low magnification SEM records even coverage of micron-scale agglomerates, while TEM resolves contacting or overlapping primary nanoparticles inside those clusters.

Q2: Does dark region in TEM confirm real fused aggregate?

A: No. Stacked particles along electron beam create projection overlap and form dark areas without physical fusion.

Q3: Can TEM sample preparation create artificial agglomerates?

A: Yes. Drop casting and solvent evaporation induce particle assembly through coffee-ring effect, capillary force and loss of colloidal stabilization.

Q4: Which technique delivers the true particle size?

A: No single “true particle size”. SEM/TEM, DLS, BET, XRD and SAXS characterize different physical quantities. Select methods according to whether you need primary particle size, hydrodynamic cluster size, surface equivalent size or crystallite size.

Q5: How to verify agglomeration status inside battery slurry?

A: For dilute model dispersions, DLS or SAXS can provide useful ensemble information. For practical high-solid-content battery slurries, dispersion should preferably be evaluated by combining rheology, fineness/grind-gauge testing, microscopy, sedimentation/storage stability and controlled particle-size measurements. Any dilution-based measurement must consider whether dilution changes the original agglomeration state.

Q6: Why nanopowder agglomeration matters for electrode fabrication?

A: Agglomeration affects slurry viscosity, dispersion uniformity, conductive network, coating stability and electrode compaction. It must be assessed together with slurry and electrode processing performance.

Conclusion

SEM uniformity and TEM agglomeration of nanopowders can coexist without contradiction. The apparent discrepancy comes from four factors: multi-level particle hierarchy, different imaging signals, varied sampling scale and distinct sample preparation routes. SEM excels at surface morphology and large-range powder distribution. TEM reveals fine nano structures, yet it compresses 3D sample thickness into 2D projection; its liquid drop-casting procedure may alter particle arrangement.

For rigorous nanopowder characterization, do not force all test methods to output identical particle size. Build a complete evidence chain: powder morphology → primary particles → aggregates & agglomerates → suspension state → specific surface area → crystal domains → slurry processing behavior.

For battery materials, the core evaluation target is not simply whether particles look aggregated. We need to confirm if powder can be wetted, dispersed and stabilized under real electrode manufacturing conditions, and remain sufficiently dispersed and stable during slurry mixing, storage, transfer, coating and drying.