Lightweight Design worldwide

, Volume 11, Issue 1, pp 6–11 | Cite as

Epoxy resins reinforced with carbon nanotubes

  • Jan Benra
  • Stefan Forero
Cover Story Innovative Composites

Carbon nanotubes can improve the mechanical properties of CFRP materials, thereby increasing the cycles-to-failure of components. A new type of carbon nanotube equipped with elastomer side chains also significantly increases maximum bending strain and the critical stress intensity factor, as Future Carbon shows. The required process steps can already be applied on an industrial scale.

Exceptional Properties

Carbon nanotubes (CNT) possess exceptional mechanical, electrical and thermal properties. For example, their elastic modulus, also known as Young’s modulus, for tension, bending, pressure and vibration is around 1 TPa [1, 2, 3], tensile strength around 10 GPa [1], thermal conductivity around 3000 to 6000 W/mK [3, 4], and electrical conductivity around 106 S/m [5].

With regard to their chemical structure, CNTs can be seen as layers of graphene rolled into cylinders. Two-dimensional graphene consists of a layer of carbon atoms that have been sp2-hybridized and exhibit a densely packed, hexagonal, honeycomb structure [22]. Figure 1 (left) is a schematic diagram of the structure of graphene.

Structure of graphene [22] (left) (© Antonio Castro Neto) and schematic diagram of an MWCNT [23] (right) (© Nanotechnology Now)

A nanotube can be made up of one or more stacked layers of graphene that have been rolled up [5, 6]. They are therefore termed single-walled carbon nanotubes (SWCNT), or multi-walled carbon nanotubes (MWCNT).

Ever since the rediscovery of CNTs by Iijima in 1991 [6], a large number of teams all over the world have attempted to transfer their outstanding mechanical properties to epoxy resin matrices [1, 5, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21] — in some cases with success. There is a wide range of challenges associated with this.

For example, the diameter d of these carbon nanotubes is in the nm range, while their length can be several μm. They therefore have a high aspect ratio and are not generally arranged as individual, straight cylinders but exist as warped and twisted clusters that are sometimes several μm long [3]. For this reason, the dispersion and adjustment of particle size plays a crucial role in epoxy resin dispersion. Shortening the CNTs during dispersions should be avoided if possible.

A further aspect is the connection of the CNTs to the epoxy resin matrix. Gojny set up various failure mechanisms in 2005, which can be seen in Figure 2 [10]. Figure (a) shows the initial state of the CNT in the polymer. If the binding forces between CNT and matrix are too weak, the CNT will slip out if a fracture occurs (b). Case (c) shows the CNT breaking. The connection to the matrix is very strong. The CNT tears in combination with extensive, fast and local deformation. A further possibility of failure is the telescope effect with MWCNTs (d). The outer wall of the MWCNT cracks open and the inner shell slips out. This is caused by a close connection to the matrix. Case (e) shows extremely effective crack bridging. A spatial connection of the reactive groups to the boundary surface in this case allows the boundary layer to be partially debonded. The resulting local connection to the matrix at the ends of the CNT combined with the boundary surface failure contributes to this crack bridging. It prevents the crack from opening up further. Any more stress would ultimately lead to failure of the CNT according to (c) and/or (d) [10].

Schematic diagram of possible failure mechanisms of CNTs in a matrix: (a) initial state, (b) slipping out, (c) CNT fracture, (d) telescope effect, (e) crack bridging [10] (© Florian H. Gojny)

Procedural Method

Figure 4 shows a flow diagram illustrating the individual steps in processing CNTs with the epoxy resin matrix. The untreated CNTs are first functionalized. In this process, oxygen groups such as -COOH or -OH are applied to the outer shell by means of oxidation. This step increases the compatibility between CNT and epoxy resin. It suppresses the subsequent formation of aggregates and minimizes separation effects during the hardening cycle. In addition, this step also involves setting the bonding in such a way that crack bridging can take place as shown in Figure 2 (e).

Crack bridging using elastomer side chains (© Future Carbon)

In the next step, the functionalized CNTs are introduced into the epoxy resin matrix. This is effected using conventional stirring or dissolving methods. In the dispersion steps, the clusters are gently broken up to ensure that the CNTs are distributed throughout the matrix as homogeneously as possible. Depending on the resin used, the result is a tar-like to hard master dispersion material that can be thinned to the appropriate concentration using standard methods. This thinned dispersion material can then be processed further by the processor by adding a hardening agent.


During functionalization, oxidative methods were investigated, and the oxygen content, Table 1, and the type of connection were determined using X-ray photoelectron spectroscopy (XPS analysis). The selected methods were plasma-treated CNTs (CNTP), acid-treated CNTs (CNTO), and thermally treated CNTs (CNTD) from untreated CNTs (CNTK). The advantages and disadvantages of the methods are listed in Table 2. As can be seen from the table, the thermal method only managed to increase the oxygen content marginally.

Extract of XPS measurements (© Future Carbon)


Carbon [Atom-%]

Oxygene [Atom-%]














Advantages and disadvantages of CNTOs, CNTDs, and CNTPs (© Future Carbon)



CNTO, acid oxidation

High degree of functionalization possible

Use of acid

Degree of functionalization can be variably set

Complex equipment required

High time investment

Clusters --> difficult to disperse

Batch process

CNTD, thermal oxidation

Dry processing

Low degree of functionalization

High throughput

No acid

Easily dispersible CNTs

— Continuous processing

CNTP, oxidized in plasma

High degree of functionalization

Low penetration depth of plasma

Dry processing

No acid

High throughput

Easily dispersible CNTs

Continuous processes possible

In contrast, acid treatment achieved high levels of oxygen, but proved to be very complex for industrial-scale use, and required large quantities of acid. In addition, very hard clusters resulted in the washing and drying process that were then very difficult to disperse.

Plasma treatment showed good levels of functionalization with dry processing and no acid needing to be used. It also allows high throughout rates, and the CNTPs proved to be easily dispersible. This method is very well suited to functionalizing CNTs on an industrial scale.

CNTs were equipped with elastomer side chains (CNTN) in a further functionalization step. The theoretical model here is that the elastomer side chains in the CNTNs work like a “rubber” when cracks form that redirects the energy of tearing into the CNT, thus having a mechanically reinforcing effect. The advantage here is that the bonding of the CNTs no longer needs to be adjusted to the matrix, and the reinforcing mechanisms as shown in Figure 2 can be extended to include an additional model. This is shown in Figure 4. The energy of tearing is transferred via the elastomer side chains to the CNTs similar to a spring mechanism.


In recent years, various methods for dispersion have been established in different publications [5, 7, 9, 24, 25, 26, 27, 28]. The methods include:
  • ▸ mechanical stirring

  • ▸ milling using a three-roll mill

  • ▸ ultrasonic techniques

  • ▸ high-pressure dispersers.

These methods have their advantages and disadvantages. The methods fail if CNTs are to be introduced into epoxy materials in industrial quantities affordably and with an adequate dispersion quality. One suitable method is extrusion using a twin-screw extruder. Tests using a laboratory extruder with a diameter of 2 cm exhibited adequate dispersion quality with relatively high throughput rates that are ten times those of a three-roll mill. Furthermore, the resin and the CNTs can be measured out and dispersed directly in the extruder using appropriate metering equipment. There is no need for the pre-mix step shown in Figure 3, and the method can be used as a continuous process on an industrial scale.

Flow diagram showing introduction of CNTs into an epoxy resin matrix (© Future Carbon)

The dispersion and adjustment of particle size plays a crucial role.

Thinning Dispersion

After dispersion, the dispersion material is first thinned using a stirring method. This step already shows whether mechanical reinforcement is promising. Important characteristics for this are:
  • ▸ homogeneous dispersion,

  • ▸ even after warming

  • ▸ small particle size (< 5 μm)

  • ▸ no aggregation or

  • ▸ separation after warming

  • ▸ good CNT-matrix compatibility.

Fast tests such as microscope inspections, Figure 5, or phase separation, Figure 6, can be used to determine these characteristics. The desired aim is small particle size with little aggregation. The two microscope images in Figure 5 show, on the right, that the CNTs are well dispersed. There is, however, a tendency to aggregation. The CNTs have agglomerated and there are areas with virtually no CNTs. This phenomenon cannot be observed in the image on the left. The CNTs are distributed throughout the dispersion material.

Microscope images of dispersion: good dispersion and low aggregation (left), high aggregation (right) (© Future Carbon)


Phase separation on a slide at 60 °C: high (left) and low phase separation (right) (© Future Carbon)

In order to perform a fast test for phase separation, one drop of the thinned dispersion material is placed on a slide and held at a temperature of about 80 °C for 30 min. Phase separation can be observed during this time. This can be seen in Figure 6. A high degree of separation can be seen in the image on the left. No separation can be seen in the image on the right.

Fatigue Tests

Load change tests were performed on bisphenol A standard resin with anhydride hardener. For this purpose, winding reels, Figure 7, were produced with and without CNTOs, and load changes were subsequently performed with an amplitude of 9.0kN at a frequency of 2 Hz. Figure 8 shows the thermal images with and without CNTs. The struts without CNTs failed after around 55,000 load changes with heat generation. The struts with CNTOs failed after around 150,000 load changes, thus withstanding about three times the number of load changes.

CFRP winding reel for load change tests (© Future Carbon)


Thermal image of struts without CNTs (left) and with CNTs (right) (© Future Carbon)

Interlaminar Energy Release Rate

In tests for mode I (GIC) and mode II (GIIC), an epoxy resin mixture comprising bisphenol A standard resin and TGMDA resin (60:40) with an amine hardener was used. Plasma-treated CNTPs were used as CNTs. The prepregs were produced following dispersion and sheets manufactured in an autoclaving process. Test specimens were then sawn from these to measure the interlaminar energy release rate in modes I and II. It proved possible to increase GIC by 52 % and GIIC by 40 %. Furthermore, it was possible to reduce the speed of crack propagation by an order of 3 compared with the unfilled system, from 10−3 mm/vibration 10−6 mm/vibration (at ΔK = 0.45 MPa√m) [29].

CNTs with Elastomer Side Chains

These tests were performed in a hardened resin curing system without carbon fibers. A cycloaliphatic epoxy resin combined with an anhydride hardener was used. The system without CNTs was compared with three systems of various types of CNT. The results are shown in Figure 9. The CNTNs were able to achieve very good mechanical characteristic data with elastomer side chains. Compared with the unfilled system, maximum bending strain εfM could be increased by 104 %, maximum bending stress σfM by 45 %, and the critical stress intensity factor KIC by 45 %. In contrast, the flexural modulus fell by around 7 %. The plasma-treated CNTPs also exhibit very good characteristic values.

Results of fracture toughness (KIC) and three-point bending tests (© Future Carbon)



The findings presented here came from the publicly funded projects CarboDis (Inno.CNT) and NanoPreg (AiF).


  1. [1]
    Miyagawa, H. et al.: Mechanical properties of Carbon Nanotubes and Their Polymer Nanocomposites. In: Journal of Nanoscience and Nanotechnology (2005), No. 5, pp. 1593–1615CrossRefGoogle Scholar
  2. [2]
    Wong, E. W. et al.: Nanobeam Mechanics: Elasticity, Strength, and Toughness of Nanorods and Nanotubes. In: Science (1997), No. 277, pp. 1971–1975CrossRefGoogle Scholar
  3. [3]
    Rao, C. N. R.: Nanotubes. In: Chemphyschem (2001), Nr 2, pp. 78–105CrossRefGoogle Scholar
  4. [4]
    Krüger, A.: Neue Kohlenstoffmaterialien. Wiesbaden: Teubner, 2007Google Scholar
  5. [5]
    Ma, P.-C. et al: Dispersion and functionalization of carbon nanotubes for polymer-based nanocomposites. In: Composites (2010), No. A 41, pp. 1345–1367CrossRefGoogle Scholar
  6. [6]
    Iijima, S.: Helical microtubles of graphitic carbon. In: Nature (1991), No. 354, pp. 56–58CrossRefGoogle Scholar
  7. [7]
    Thostenson, E. T. et al.: Processing-structure-multi-functional property relationship in carbon nanotube/epoxy composites. In: Carbon (2006), No. 44, pp. 3022–3029CrossRefGoogle Scholar
  8. [8]
  9. [9]
    Gojny, F. H. et al.: Carbon nanotube-reinforced epoxy-composites: enhanced stiffness and fracture toughness at low nanotube content. In: Composite Science and Technology 64 (2004), pp. 2363–2371CrossRefGoogle Scholar
  10. [10]
    Gojny F. H. et al.: Influence of different carbon nanotubes on the mechanical properties of epoxy matrix composites — A comparative study. In: Composite Science and Technology 65 (2005), pp. 2300–2313CrossRefGoogle Scholar
  11. [11]
    Ganguli, S.; Aglan H.: Effect of Loading and Surface Modification of MWCNTs on the Fracture Behavior of Epoxy Nanocomposites: Journal of Reinforced Plastics and Composites 25 (2006), pp. 175–188Google Scholar
  12. [12]
    Cooper, C. A. et al.: Detachment of nanotubes from a polymer matrix. In: Applied Physics Letters 81 (2002), No. 20, pp. 3873–3875CrossRefGoogle Scholar
  13. [13]
    Lachman, N.; Wagner, H. D.: Correlation between interfacial molecular structure and mechanics in CNT/epoxy nano-composites. In: Composites A (2010), No. 41, pp. 1093–1098CrossRefGoogle Scholar
  14. [14]
    Chatterjee, S. et al.: Size and synergy effects of nanofiller hybrids including graphene nanoplatelets and carbon nanotubes in mechanical properties of epoxy composites. In: Carbon 2012, No. 50, pp. 5380–5386CrossRefGoogle Scholar
  15. [15]
    Kepple, K. L.: Improved fracture toughness of carbon fiber composite functionalized with multi walled carbon nanotubes. In: Materials Science and Engineering A (2006), No. 435-436, pp.145–149CrossRefGoogle Scholar
  16. [16]
    Lubineau, G.; Rahaman A.; A review of strategies for improving the degradation properties of laminated continuous-fiber/epoxy composites with carbon-based nanoreinforcements. In: Carbon 50 (2012), pp. 2377–2395CrossRefGoogle Scholar
  17. [17]
    Ma, P. C. et al.: Effects of silane functionalization on the properties of carbon nanotube/epoxy nanocomposites. In: Composite Science and Technology 67 (2007), pp. 2965–2972CrossRefGoogle Scholar
  18. [18]
    Ma, A. W. K. et al.: The microstructure and rheology of carbon nanotube suspensions. In: International Journal Materials Forming (2008), No. 1, pp. 75–81CrossRefGoogle Scholar
  19. [19]
    Ma, P. C. et al.: Dispersion, interfacial interaction and re-agglomeration of functionalized carbon nanotubes in epoxy composites. In: Carbon 48 (2010), pp. 1824–1834CrossRefGoogle Scholar
  20. [20]
    Yaping, Z. et al.: Functionalized effect on carbon nanotube/epoxy nano-composites. In: Materials Science and Engineering A (2006), No. 435-436, pp. 145–149Google Scholar
  21. [21]
    Yue, L. et al.: Epoxy composites with carbon nanotubes and graphene nanoplatelets — Dispersion and synergy effects. In: Carbon (2014), pp. 268–278Google Scholar
  22. [22]
    Neto A. C.; Guinea F.; Peres, N. M.: Drawing conclusions from grapheme. In: Physics World (2006), No. 19, pp. 32–37CrossRefGoogle Scholar
  23. [23]
    Pillai, S. K.; Ray, S. S.: Epoxy-based Carbon Nanotubes Reinforced Composites. In: Advances in Nanocomposites - Synthesis, Characterization and Industrial Applications (2011), No. 32, pp. 729–792Google Scholar
  24. [24]
    Sandler, J. et al.: Development of a dispersion process for carbon nanotubes in an epoxy matrix and the resulting electrical properties. In: Polymer (1999), No. 40, pp. 5967–5971CrossRefGoogle Scholar
  25. [25]
    Li, B. J. et al.: Correlations between Percolation Threshold, Dispersion State, and Aspect Ratio of Carbon Nanotubes. In: Advanced Functional Materials (2007), No. 17, pp. 3207–3215CrossRefGoogle Scholar
  26. [26]
    Schultz, S. et al.: High-Pressure Homogenization as a Process for Emulsion Formation. In: Chemical engineering and technology (2004), No. 27, pp. 361–368CrossRefGoogle Scholar
  27. [27]
    Azoubel, S., Magdassi, S.: The formation of carbon nanotube dispersions by high pressure homogenization and their rapid characterization by analytical centrifuge; In: Carbon (2010), No. 48, pp. 3346–3352CrossRefGoogle Scholar
  28. [28]
    Panagiotou, T. et al.: Deagglomeration and Dispersion of Carbon Nanotubes Using Microfluidizer High Shear Fluid Processors. In Nanotech (2008), No. 1, pp. 39–42Google Scholar
  29. [29]
    Zeiler, R. et al: Mit Carbon Nanotubes modifizierte Epoxidharze als Matrix in Faser-Kunststoff-Verbundwerkstoffen. In: Kunststoffe (2016), No. 2, pp. 35–39Google Scholar

Copyright information

© Springer Fachmedien Wiesbaden 2018

Authors and Affiliations

  • Jan Benra
    • 1
  • Stefan Forero
    • 1
  1. 1.Future Carbon GmbHBayreuthGermany

Personalised recommendations