Author: [Bruce Zhou]
Affiliation: [Jota Machinery Composites Material Prepreg Solution]
Corresponding Author: [jotamachinery@gmail.com]
Published : December 08 , 2025

Abstract

Aerospace-grade epoxy matrices offer high stiffness and strength but remain intrinsically brittle, limiting damage tolerance and delamination resistance in unidirectional (UD) carbon-fibre laminates. Carbon nanotubes (CNTs) are an attractive nano-toughener, yet their practical use is constrained by dispersion difficulties, filtration during liquid moulding, and poorly controlled surface chemistry. This paper reviews and re-interprets a recent experimental study on multi-walled carbon nanotubes (MWCNTs) in a DGEBA/TGMDA–DDS epoxy system processed as UD carbon-fibre prepreg. The work compares neat MWCNTs with low-pressure plasma-treated MWCNTs, using extruded masterbatches and hot-melt prepregging to eliminate filtration effects and isolate the role of surface modification.

X-ray photoelectron spectroscopy shows that plasma treatment increases oxygen content on CNT surfaces from about 1.25 to 3.7 at.% and enriches carbonyl groups, while transmission electron microscopy reveals a marked reduction in agglomeration. In the neat resin, 0.5 wt% plasma-treated MWCNTs raise the mode I fracture energy GIc by approximately 38 % relative to the unmodified epoxy. In UD laminates, the same loading yields increases of about 40 % in mode I interlaminar toughness GIc and 26 % in mode II toughness GIIc, with only a slight decrease in short-beam interlaminar shear strength (ILSS). In contrast, neat MWCNTs at identical loading cause significant ILSS degradation (~33 %) and reduce both GIc and GIIc despite modest gains at resin level. Fractographic analysis links these trends to CNT morphology: untreated MWCNTs form large agglomerates that act as crack initiators, whereas plasma-treated MWCNTs are dispersed as primary tubes that contribute to crack bridging, pull-out and enhanced hackle formation.

The results demonstrate that nano-toughening of aerospace laminates with CNTs is not governed by filler content alone but by the interplay of surface chemistry, dispersion, and processing route. Design implications are drawn for prepreg manufacture and interlaminar tailoring, and potential extensions toward fatigue, multifunctionality and hybrid nanofillers are outlined.

Keywords

plasma-treated carbon nanotubes; multi-walled CNTs; aerospace epoxy; unidirectional prepreg; interlaminar fracture toughness; mode I and mode II delamination; interlaminar shear strength; nano-toughening; fracture micromechanisms

1. Introduction

Crosslinked epoxies such as DGEBA/TGMDA systems cured with aromatic diamines form the backbone of aerospace structural composites. They deliver high glass transition temperatures and good hot/wet performance but suffer from low fracture toughness. At laminate level, this brittleness translates into limited resistance to delamination under out-of-plane and mixed-mode loading, making interlaminar enhancement a persistent design target.

Nano-scale modifiers, especially carbon nanotubes, promise to toughen epoxy matrices through crack deflection, crack bridging and pull-out, while also offering potential gains in electrical and thermal conductivity. However, three recurring obstacles complicate their use in real aerospace laminates:

1. Dispersion – CNTs strongly agglomerate; high shear mixing raises viscosity and can damage fibre sizing.
2. Filtration – in liquid moulding (RTM, infusion), CNTs are filtered by fibre beds, leaving the interlaminar region depleted of nanofillers.
3. Surface chemistry – the balance between covalent functionalisation and structural integrity is delicate, and its influence on laminate-scale toughness is not straightforward.

The study considered here addresses these obstacles by combining extruded CNT–epoxy masterbatches with a hot-melt prepreg process, and by directly comparing neat and plasma-treated multi-walled CNTs at a fixed loading. This configuration avoids filtration effects and allows the role of surface modification to be isolated at both resin and laminate levels.

The following sections reorganise the published data into a process–structure–property framework. We begin with materials and processing details, then analyse how plasma treatment alters CNT surface chemistry and dispersion, and finally connect these changes to fracture toughness and interlaminar performance in UD laminates.

2. Materials and Processing Route

2.1 Epoxy system and reinforcement

The matrix is a typical aerospace toughened epoxy based on:

• Resin blend: DGEBA/TGMDA at 40/60 phr,
• Hardener: 4,4′-diaminodiphenyl sulfone (DDS),

producing a network with dry glass transition temperature Tg,dry around 219 °C and hot-wet Tg near 201 °C. This combination is commonly used in high-temperature prepregs where hot/wet retention is critical.

The fibre reinforcement is a standard high-strength carbon fibre:

• Fibre: Tenax HTS40 F13 12K, 800 tex unidirectional roving,

delivering the axial stiffness and strength expected for primary structures.

2.2 CNTs and plasma treatment

Multi-walled carbon nanotubes are introduced via:

• Neat MWCNTs (CNT-n) in epoxy masterbatch, and
• Plasma-treated MWCNTs (CNT-p), produced by low-pressure rotary plasma in an oxygen-containing environment.

The plasma step introduces polar oxygenated groups on the CNT surface without resorting to liquid-phase acid treatments that can cut or overly damage the tubes. The final CNT contents in the systems studied reach up to 0.5 wt% relative to resin.

2.3 Masterbatch extrusion and hot-melt prepregging

To ensure robust dispersion and avoid filtration:

1. CNT–epoxy masterbatches are produced by twin-screw extrusion, supplying high shear to deagglomerate the nanotubes and distribute them in the resin.
2. A hot-melt prepreg route is used:
   • CNT-modified resin is film-coated at ~70 °C,
   • UD carbon fibre tows are impregnated via calendering at ~100 °C under 5 bar,
   • Approximately 20 rovings are combined to form ~200 mm wide prepreg tapes.

This path keeps CNTs in the resin film and avoids their loss into breather or tooling, in contrast to liquid infusion.

2.4 Laminate configuration and cure cycle

Laminates are laid up as:

• 26-ply UD configuration, mostly 0°, with the 2nd and 25th plies at 90° to stabilise handling,
• Total thickness around 3 mm,
• A mid-plane Teflon insert to pre-define the delamination plane for fracture testing.

Curing takes place in an autoclave under ~7 bar pressure, with vacuum to gel, at 180 °C followed by a 200 °C post-cure, typical of aerospace prepregs. This ensures full crosslinking and consistent fibre volume fraction across formulations.

3. Effect of Plasma Treatment on CNT Surface and Dispersion

3.1 Surface chemistry from XPS

X-ray photoelectron spectroscopy reveals a clear difference between the two CNT variants:

• Neat CNT-n
  • Carbon ≈ 98.75 at.%, oxygen ≈ 1.25 at.%.
  • Among oxygenated species, about 82 % are assigned to C–O, 18 % to C=O environments.
• Plasma-treated CNT-p
  • Carbon ≈ 96.3 at.%, oxygen ≈ 3.7 at.%.
  • The fraction of carbonyl-type C=O species increases to about 30 % of the oxygenated bonds.

Thus plasma treatment roughly triples the oxygen content and enriches polar carbonyl functionality. From an interfacial design standpoint, this implies stronger potential for secondary bonding or chemical interaction with the DGEBA/TGMDA–DDS network.

3.2 Dispersion state from TEM

Transmission electron microscopy of the cured nanocomposites shows:

• In CNT-n systems, notable agglomerates up to ~750 nm in diameter coexist with dispersed tubes; clusters are frequent in resin-rich regions.
• In CNT-p systems, the majority of CNTs appear as primary nanotubes, with only occasional agglomerates around the micrometre scale.

Even after high-shear extrusion, surface modification still exerts a measurable influence: plasma-modified CNTs are less prone to restacking and form a more uniform network in the cured resin.

4. Fracture Toughness of the Neat Epoxy

4.1 Baseline behaviour

The unmodified epoxy shows:

• Mode I fracture toughness KIc ≈ 0.48 MPa·m½,
• Fracture energy GIc ≈ 71 J/m²,

representative of a relatively brittle aerospace matrix.

4.2 Influence of CNT content and surface treatment

When CNTs are added:

• At 0.25 wt% CNT-n, KIc rises by approximately 10 %, and GIc reaches about 90 J/m² (~27 % gain).
• At 0.5 wt% CNT-n, GIc drops slightly compared to the 0.25 wt% case, indicating a plateau and possible early agglomeration effects.
• At 0.25 wt% CNT-p, both KIc and GIc improve over the neat system, showing a steady trend.
• At 0.5 wt% CNT-p, KIc increases by around 20 %, while GIc reaches ~98 J/m², representing a ~38 % improvement over the baseline.

These results imply that both CNT types provide some degree of matrix toughening, but plasma-treated CNTs support a monotonic increase with loading, consistent with better dispersion and interfacial bonding. The neat CNTs show evidence of an optimum at intermediate loading, after which agglomeration offsets the benefits.

5. Interlaminar Properties of UD Laminates

5.1 Interlaminar shear strength (ILSS)

Short-beam tests provide a first indicator of how the nanofiller network interacts with fibre–matrix interfaces and resin-rich regions.

• Reference laminate: ILSS ≈ 91 MPa.
• 0.5 wt% CNT-n laminate: ILSS decreases by roughly 33 %, a severe reduction.
• 0.5 wt% CNT-p laminate: ILSS is slightly lower than the reference, but the reduction is modest.

The sharp drop for CNT-n indicates that agglomerates behave as distributed defects under the complex stress state of short-beam bending. By contrast, the better-dispersed CNT-p population does not significantly undermine the shear load-carrying capability.

5.2 Mode I interlaminar fracture toughness GIc

Mode I delamination tests (DCB) give:

System	GIc resin (J/m²)	GIc laminate (J/m²)
Neat	71	250
0.5 wt% CNT-n	82	216 (≈ –13 %)
0.5 wt% CNT-p	98	356 (≈ +40 %)

Two points stand out:

1. The laminate GIc of the unmodified system is about three times the neat resin GIc, owing to fibre bridging and other laminate-scale mechanisms.
2. For CNT-n, the slight resin-scale gain is more than cancelled by adverse laminate effects (agglomerates and weakened interfaces), leading to a drop in GIc.
3. For CNT-p, the resin-scale improvement transfers effectively to the laminate, yielding a ~40 % higher GIc than the reference.

This confirms that neat resin metrics cannot be used alone to predict delamination resistance; the quality of the nano-modified interlaminar region is decisive.

5.3 Mode II interlaminar fracture toughness GIIc

Under Mode II loading (ENF):

• Neat laminate: GIIc ≈ 754 J/m².
• 0.5 wt% CNT-n laminate: GIIc ≈ 641 J/m² (~15 % decrease).
• 0.5 wt% CNT-p laminate: GIIc ≈ 952 J/m² (~26 % increase).

Again, untreated CNTs are detrimental at laminate level, while plasma-treated CNTs provide a meaningful improvement in shear-driven delamination resistance.

Taken together, the interlaminar results show that, at equal loading:

• Neat MWCNTs: moderate resin-level toughening but reduced GIc, GIIc and ILSS in laminates.
• Plasma-treated MWCNTs: significant gains in GIc and GIIc with only minor ILSS penalties.

6. Fractographic Micromechanisms

6.1 Mode I fracture surfaces

Scanning electron micrographs of DCB fracture surfaces reveal distinct failure morphologies:

• Neat laminate
  • Resin regions exhibit microcracking and river patterns typical of brittle epoxy.
  • Fibre surfaces show evidence of pull-out and moderate fibre–matrix debonding, contributing to the laminate’s higher GIc compared to neat resin.
• CNT-n laminate
  • Large CNT agglomerates are visible in resin-rich zones.
  • Microcracks originate or propagate through these clusters.
  • Fibre surfaces appear coated with CNTs in some locations, but the interfacial region is locally weakened, leading to relatively easy debonding without substantial additional energy absorption.
• CNT-p laminate
  • Agglomerates are smaller and less frequent; most CNTs are embedded as individual tubes or small bundles within the matrix.
  • Fracture surfaces show clear CNT pull-out and bridging features, superimposed on matrix cracking and fibre pull-out.
  • This combination of mechanisms accounts for the higher energy dissipation and increased GIc.

The micrographs support a simple picture: poorly dispersed CNTs act as defects, whereas well-dispersed, surface-modified CNTs operate as crack-bridging and pull-out elements in the interlaminar region.

6.2 Mode II fracture surfaces

In Mode II, the fracture surfaces of the reference laminate show pronounced hackle patterns, associated with repeated microcrack nucleation and coalescence under shear.

• Neat laminate: well-developed hackles indicate substantial inelastic deformation and microcracking before full delamination.
• CNT-n laminate: hackles are muted; fracture appears more planar and brittle, with agglomerates visible at or near the crack path. The reduction in GIIc correlates with this more brittle morphology.
• CNT-p laminate: hackles are more pronounced than in the reference, and the matrix shows signs of enhanced microcrack activity, consistent with higher GIIc.

These observations underline that CNT morphology modulates whether Mode II fracture evolves toward a more brittle or more ductile-like pattern, and that plasma treatment shifts behaviour in the desired direction.

7. Design Implications for Aerospace Prepreg Laminates

7.1 Quantitative performance summary

At a CNT loading of 0.5 wt%, plasma-treated MWCNTs deliver:

• Neat epoxy:
  • GIc ≈ 98 J/m² (≈ +38 % over neat resin).
• UD laminate:
  • GIc ≈ 356 J/m² (≈ +40 % over reference laminate).
  • GIIc ≈ 952 J/m² (≈ +26 % over reference laminate).
  • ILSS slightly reduced, but still within acceptable bounds.

At the same loading, neat MWCNTs produce:

• Resin GIc only modestly improved,
• Laminates with lower GIc and GIIc than the reference,
• ILSS reduced by about one third.

The conclusion is straightforward: surface modification and dispersion control, not filler percentage alone, determine whether CNTs toughen or weaken an aerospace laminate.

7.2 Processing route as a design choice

The study also demonstrates the importance of the processing route:

• The use of extruded masterbatches ensures reproducible CNT distribution in the resin before prepregging.
• The hot-melt prepreg process prevents filtration and maintains consistent CNT content in the interlaminar region.

For designers and process engineers, this suggests that prepreg-based routes, rather than wet infusion, are preferable when targeting interlaminar toughening with CNTs in high-performance UD laminates.

7.3 Opportunities for further development

Building on these findings, several directions are attractive for future work:

1. Fatigue delamination – Extending the comparison between CNT-n and CNT-p to cyclic GIc and GIIc would align more closely with aerospace certification requirements.
2. Multifunctionality – Combining interlaminar toughening with improved through-thickness electrical or thermal conductivity could support lightning protection or structural health monitoring.
3. Hybrid nano-modifiers – Exploring combinations of plasma-treated CNTs with rubbery particles, nano-silica or graphene derivatives may offer synergistic effects on toughness and damping.
4. Micromechanical modelling – Incorporating measured CNT dispersion statistics and surface chemistry into cohesive zone or multiscale models could enable predictive design of nano-toughened interlaminar regions.

8. Conclusion

Plasma-treated multi-walled carbon nanotubes, introduced via masterbatch extrusion and hot-melt prepregging, provide a viable route to enhancing the delamination resistance of aerospace-grade UD carbon-fibre laminates without compromising processing or basic laminate quality. By enriching CNT surfaces with oxygen-containing groups and improving dispersion, the plasma treatment turns nanotubes from potential defect sites into active toughening agents.

At modest loadings around 0.5 wt%, the treated CNTs raise mode I and mode II interlaminar fracture toughness significantly, while leaving ILSS largely intact. Neat CNTs at the same loading, despite improving neat-resin fracture metrics to some extent, degrade laminate-scale toughness and shear strength due to agglomeration and interfacial weakening.

For aerospace designers, the main message is that nano-toughening must be engineered, not assumed. Surface chemistry, dispersion, and processing path must be jointly specified to ensure that CNTs actually reinforce the interlaminar region. With these conditions satisfied, plasma-treated CNTs can form part of a robust interlaminar design toolkit alongside traditional toughening strategies and emerging multifunctional concepts.

References

1. Bakis, G. et al. Mechanical properties of carbon nanotube-modified epoxy–carbon fibre unidirectional prepreg laminates. Polymers, 2021.
2. Gojny, F.H. et al. Carbon nanotube-reinforced epoxy composites: enhanced fracture toughness and influence of nanotube surface treatment.
3. Selected works on CNT functionalisation, interlaminar fracture mechanics and nano-toughened aerospace composites, providing broader context for the trends discussed here.