Microwave curing of carbon–epoxy composites: Penetration depth and material characterisation

TL;DR: In this paper, the authors present some evidence which suggests that with the correct hardware and operating procedure/methodology, consistent and high quality carbon-epoxy laminates can be produced, with the possibility of scaling up the process, as demonstrated by the micro and macro-scale mechanical test results.

Abstract: Microwave heating has several major advantages over conventional conductive heating when used to cure carbon–epoxy composites, especially in speed of processing. Despite this and many other well-known advantages, microwave heating of carbon–epoxy composites has not taken off in industry, or even academia, due to the problems associated with microwave energy distribution, arcing, tool design and (ultimately) part quality and consistency, thus leading to a large scepticism regarding the technique/technology for heating such type of materials. This paper presents some evidence which suggests that with the correct hardware and operating procedure/methodology, consistent and high quality carbon–epoxy laminates can be produced, with the possibility of scaling up the process, as demonstrated by the micro- and macro-scale mechanical test results. Additionally, the author proposes a methodology to practically measure the maximum microwave penetration depth of a carbon–epoxy composite material.

Summary

1. Introduction

• The production of quality parts, lower cost and time has been a priority for manufacturing companies, and increasingly so in today’s very competitive global market, particularly for companies in developed countries where costs are generally higher.
• Assess the mechanical properties of MW cured composites under tension, compression, in-plane shear (IPS) and indentation loading, and compare the results with conventionally cured samples.
• In the past, samples produced using MWs were typically less than one wavelength (i.e. 125mm), and smaller than the dimensions recommended by test standards such as ASTM D3039 [24], ASTM D6641 [25] and 5 ASTM D3518 [26], possibly due to the difficulty in obtaining a highly homogeneous MW field over the specimen volume.
• Likewise, knowing that MW heating of CFRPs in the past was neither consistent, homogeneous, nor followed a suitable procedure, it is difficult to assume the results in the literature are accurate or consistent.

2.1 Materials and Equipment

• The materials and MW equipment employed in the present study are consistent with those used in [23], i.e. 600g/m2 uni-directional (UD) out-of-autoclave (OoA) carbon fibre reinforced epoxy from Gurit, which has a PAN -based carbon fibre with an elastic modulus of 255GPa, tensile strength of 4.3GPa, fibre density of 1.8g/cm3, and cured ply thickness (CPT) of 0.6mm [27].
• The MW curing methodology was consistent with that presented by Kwak et al [3,23].
• Two laminates were produced for each test case, i.e. one oven cured, and one MW cured.
• The establishment of the MW penetration depth of a material at a specific MW frequency is of much importance as this will determine whether the material under investigation will heat evenly through the thickness of the material.
• A thermal imaging camera was used to measure the temperature gradient along the side of the enclosure as the water in the container was heated using MWs.

2.4.1 Tension

• The tensile strength and elastic modulus were determined based on ASTM D3039M [24].
• Glass fibre reinforced polymer (GFRP) composite end-tabs were bonded using a room temperature cure adhesive on the 0° and 90° tensile test coupons.
• The results of a minimum of six coupons (out of ten coupons) with acceptable failure modes were considered and analysed.
• The compression strength was calculated from Eq. 4 [25], and the compression modulus was calculated from Eq. 5 [25].
• 𝜎𝜎𝑐𝑐,𝑚𝑚𝑚𝑚𝑚𝑚 = 𝑃𝑃𝑐𝑐,𝑚𝑚𝑚𝑚𝑚𝑚 𝐴𝐴 (Eq.4) where σc,max is the maximum compressive strength (MPa), Pc,max is the maximum failure load (N) and A is the cross-sectional area (mm2).

2.4.3 In-plane shear

• The IPS strength and shear modulus properties of the material were determined based on ASTM D3518 [26].
• The test specimens had a balanced and symmetric (±45°)2S layup, with dimensions identical to those used for the tensile tests (i.e. 230x25x2.4mm), and the cross-head speed was 2mm/min.
• The adhesive layer had a chamfer on the gauge section.
• The results of a minimum of six coupons with acceptable failure modes were considered and analysed.

2.4.4 Failure mode assessment

• Scanning electron microscopy (SEM) was used to assess the differences in failure modes between the oven and MW cured composites tested under 90° tensile.
• The areas of interest included degree of matrix remaining on the fibre surface, fibre bridging/disbonding and matrix fracture surface.

2.4.5 Indentation

• Indentation testing is a proven technique used to assess a material’s hardness, as demonstrated by the standardisation of such technique in international test standards such as ASTM D2583 [33], ASTM E2546 [34] and ISO 14577-4 [35].
• The difference between standard indentation methods and the more recent instrumented methods is mainly on the size and displacement of the indentations, which can be ‘a few’ nanometres.
• The tests were carried out using load control, with a load limit of 5mN. 2.4.6 Degree of Cure, Void Volume (Vv) and Fibre Volume (Vf) Content 11 A Perkin Elmer DSC 6000 was used to identify the material’s Tg and degree of cure.
• The degree of cure was calculated by comparing a reference enthalpy value obtained from a semi-cured (the term ‘semi-cured’ describes the prepreg’s stage of cure) sample with a second enthalpy value obtained from a cured sample.
• Three sections of each of the laminates were prepared, and six non-overlapping images were taken, i.e. a total of 18 images were used from each laminate.

3.1 MW Penetration Depth

• The average and standard deviations are from three tests.
• The reference tests showed a high (~19°C and ~11°C) and a low (<0.3°C) temperature change for the R1 (perforated aluminium plate only, i.e. full MW penetration) and R2 (solid aluminium plate only, i.e. no MW penetration) tests respectively.
• The change in temperature as a function of laminate thickness showed approximately an exponential decay pattern, thus agreeing well with the theory described in §2.3.

3.2 Degree of Cure, Void Volume (Vv) and Fibre Volume (Vf) Content

• The oven cured sample’s degree of cure for a 45 minute dwell at 120°C was ~95%, thus agreeing with the values provided by the manufacturer.
• A 40 minute MW dwell at 120°C provided very similar longitudinal and transverse tensile performance compared with the oven cured samples.

3.3.1 Tension

• The tensile test results (Fig. 3) showed that the maximum longitudinal ultimate tensile strength (UTS) of MW cured samples was similar compared to oven-cured ones, which is similar to what was presented previously by Kwak et al [3].
• On the other hand, the average transverse UTS of oven cured composites was slightly greater (~11%) than the highest average of the MW cured composites.
• Little change in longitudinal and transverse elastic moduli was observed across all the samples regardless of the curing method and curing cycle.
• There were no indications of large variations in strength or modulus across the laminate for both the MW and oven cured samples (Fig. 4).

3.3.2 Compression

• All four sets of MW cured samples showed greater average ultimate compression strengths (UCS) compared with the oven cured samples, with the MW cured samples showing a slightly greater standard deviation (Fig. 5).
• As with the modulus values obtained under tensile loading, no significant changes were observed.
• The elastic modulus of oven and MW cured laminates were calculated from the reduced/indentation modulus using Eq. 10.
• There was little variation in results irrespective of the heating method.
• By applying the rule of mixtures to the elastic modulus values obtained in §3.3.1 for a Vf of 50%, the elastic modulus of the matrix is 4GPa, i.e. 15% lower than the values obtained using the indentation method.

3.4 Failure Mode Assessment

• Some similarities and differences were observed between the conventional oven cured and MW cured samples when the failure mode was assessed using SEM.
• Due to the selective heating nature of MWs (i.e. predominant heating from carbon fibres to the matrix, potentially providing a higher temperature at the fibre-matrix interface), it can be deduced that all MW cured coupons produced greater compressive strengths because the matrix close to the carbon fibres had, assuming an Arrhenius relationship, a relatively high degree of cure.
• This is evidenced by the fact that; i) all MW cured samples performed better under compression than the oven cured samples, even the sample with 82% degree of cure, ii) indentation tests demonstrated there is very little difference in matrix modulus, and, iii) SEM shows significantly higher degree of matrix remaining after testing of MW cured samples.
• The suitability of MW-heating of a material (particularly CFRPs) need to be investigated on a case-bycase basis, i.e. pre-work is required prior to attempting to process materials using MWs, since the global and local MW field will vary depending on factors such as the part’s geometry, temperature, dielectric and conductivity properties.

Figures

Table 3

Table 2

Table 1 (cont)

Table 1 (cont)

Full text

1
Microwave Curing of Carbon-Epoxy Composites: Penetration Depth and Material
Characterisation.
M. Kwak*(1), R. Wise(2), P. Robinson(3), A. Bismarck(3).
Affiliations:
1 Integrity Management Group, TWI Ltd, Great Abington, CB21 6AL, UK
2 Joining Technologies Group, TWI Ltd, Great Abington, CB21 6AL, UK
3 The Composites Centre, Imperial College London, London, SW7 2AZ, UK
*Corresponding author M. Kwak, Tel. +44(0)1223 899000; fax: +44(0)1223 892588
m
useok.kwak@twi.co.uk (M. Kwak), p.robinson@imperial.ac.uk (P. Robinson),
a.bismarck@imperial.ac.uk (A. Bismarck), roger.wise@twi.co.uk (R. Wise)

2
Abstract
Microwave heating has several major advantages over conventional conductive heating when used to cure
carbon-epoxy composites, especially in speed of processing. Despite this and many other well-known
advantages, microwave heating of carbon-epoxy composites has not taken off in industry, or even
academia, due to the problems associated with microwave energy distribution, arcing, tool design and
(ultimately) part quality and consistency, thus leading to a large scepticism regarding the
technique/technology for heating such type of materials. This paper presents some evidence which
suggests that with the correct hardware and operating procedure/methodology, consistent and high quality
carbon-epoxy laminates can be produced, with the possibility of scaling up the process, as demonstrated
by the micro- and macro-scale mechanical test results. Additionally, the author proposes a methodology
to practically measure the maximum microwave penetration depth of a carbon-epoxy composite material.
Keywords
A. Polymer-matrix composites (PMCs)
B. Mechanical properties
B. Interface/interphase
E. Cure

3
1. Introduction
The production of quality parts, lower cost and time has been a priority for manufacturing companies, and
increasingly so in today’s very competitive global market, particularly for companies in developed
countries where costs are generally higher. Additionally, the increasing demand of composite-intensive
aircraft such as Boeing’s 787, Airbus’ A350 and Bombardier’s C-series, as well as the expansion of
composites into applications which were previously considered unsuitable (e.g. automotive, electronic
packaging, etc.), has meant that increased productivity at a lower cost is key.
The production of parts made of composites typically requires the purchasing of costly materials – cost of
carbon fibre is estimated to be better than 500x greater than that of steel [1] – followed by a lengthy and
energy-intensive heating process. When producing parts made of polymer matrix composites (PMCs), the
low thermal conduction/diffusivity of the matrix leads to an inherent limitation in cycle reduction using
conventional heating methods, thus a 24 hour cure cycle is sometimes necessary for curing thick parts.
One possibility to reduce production time and its associated costs is to use alternative heating methods
such as microwave (MW) heating. The advantages of MW heating are well-known [2-5], but there are
some major challenges remaining, such as even energy distribution and consistency, arcing, tooling
design and part quality. These challenges need to be addressed before MW heating/curing can be
considered for (structural) industrial applications.
In the present investigation, carbon-epoxy composites were cured in a highly homogeneous MW field,
employing a suitable heating/curing methodology – which differs from the work reported previously by
other authors as described by the discrepancies in the results obtained which are explained later. These
samples were then tested under different loading conditions and the performance was evaluated against
conventionally cured composites. Additionally, the importance of MW penetration depth is presented, and
a practical method for measuring this property is introduced. The main objectives of the current work are:
• Present the current state of the art in MW curing of carbon-epoxy composites, and clarify the
discrepancies in the physical and mechanical test results obtained by previous investigators.
• Provide a methodology to measure MW penetration depth in composites.
• Assess the mechanical properties of MW cured composites under tension, compression, in-plane shear
(IPS) and indentation loading, and compare the results with conventionally cured samples.

4
• Propose an explanation for any differences in the mechanical properties of MW cured and
conventionally cured composite materials.
Many papers have been published in the field of MW heating of materials, such as cement [6], rubber [7],
polymers [8-12] and polymer composites [13-23]. The current summary will only focus on MW heating
of carbon fibre reinforced polymer (CFRP) composites, more specifically carbon-epoxy composites, as
these present some specific challenges (e.g. arcing, selective heating, etc.) other types of materials (e.g.
thermosetting polymers, thermoplastics, glass-reinforced polymers) may not experience, and thus
possibly the reason why there are relatively few publications in this topic/material. As mentioned in §1,
there are some discrepancies in the results produced in the past [13-22], which is believed to be due to:
i) Differences in hardware design: MW systems require careful design, as achieving a high MW field
homogeneity is critical to achieve a highly homogeneous heating throughout the material. The MW
systems used in the past were relatively simple systems, and therefore it would seem unlikely these
could avoid cold/hot spots across the sample, as evidenced by the scepticism that has existed and
exists even today regarding MW heating of (CFRP composite) materials.
ii) Different methodologies used to define the process cycle: Most of the aforementioned MW systems
lacked temperature control, thus the processing methodology was typically ‘x’ MW power for ‘y’
time. Such a heating profile would have produced a variation in heating rate as a function of time, and
a fixed dwell temperature would have been unlikely. This would have inevitably led to a different cure
cycle to the one which thermosetting resins are normally designed to follow. Additionally, in
conjunction with point i), different MW applicator designs, and/or waveguide design (or lack of)
probably had a different effect on the material (even if they were set at the same MW power) due to
the different MW field distributions present in the cavities, i.e. the reproducibility of the heating
process and subsequent results are unlikely.
iii) Exposed carbon fibres cause arcing: Arcing causes three very undesirable effects; a) detrimental
damage on the material, b) vacuum bagging becomes unfeasible thus leading to high void content, c)
health and safety implications. The probability of arcing is greater in inhomogeneous MW systems.
iv) Mechanical tests carried out on samples with non-standardised dimensions: In the past, samples
produced using MWs were typically less than one wavelength (i.e. 125mm), and smaller than the
dimensions recommended by test standards such as ASTM D3039 [24], ASTM D6641 [25] and

5
ASTM D3518 [26], possibly due to the difficulty in obtaining a highly homogeneous MW field over
the specimen volume. The fact that testing of MW cured composites were only carried out for tension,
interlaminar shear (using short beam shear) and flexure tests (i.e. tests which do not require specific
test jigs and can be done with relatively small/short coupons) is an indication of the serious difficulties
past researchers experienced to produce large(-r) samples. Therefore, it may seem logical that tests
under compression loading for example were not carried out, even when compressive properties are
possibly, together with fracture, two of the most important mechanical properties of (composite)
materials. As an analogy, when CFRPs are cured in a conventional oven and undergoes excessive
thermal runaway for example, the material is thrown away rather than being tested, since the material
has undergone an unsuitable cure cycle and the material is not in an ‘acceptable’ condition. Likewise,
knowing that MW heating of CFRPs in the past was neither consistent, homogeneous, nor followed a
suitable procedure, it is difficult to assume the results in the literature are accurate or consistent.
Having these points in mind, Kwak et al’s [3] study may have been the first publication which described a
suitable methodology to heat CFRPs using MWs, producing laminate sizes large enough to follow the
relevant mechanical test standards with a high degree of confidence, reliability and consistency. This has
been a significant step forward as the results presented in the past were highly scattered, and little work
was done on process reliability [4]. Kwak et al’s subsequent study [23] was possibly the first publication
that produced a thick (50mm+) CFRP laminate with MWs using the procedure in [3]. A similar study was
carried out by Wei et al [21], where a laminate with dimensions of 76x76x38mm – again, dimensions of
less than one wavelength – was heated using MWs, however MW was used for post-curing only.
When assessing the main outcomes of the work carried out in the past by other investigators (Table 1)
[13-22], it can be seen that in terms of T
g
, Fang and Scola [14] reported an increase, Papargyris et al [18]
reported no significant changes, and Paulauskas [22] reported a decrease by using MW heating. In terms
of mechanical properties, various authors [14,15,18,19] reported similar or increased values, whereas
Paulauskas [22] reported a decrease with MW curing. In the most recent publications related to testing of
MW cured composites, Kwak et al [3] reported similar T
g
, similar 90° tensile strength, and an increase in
0° tensile strength by MW curing. Kwak et al [23] later demonstrated that the fracture toughness G
1C
indicated an apparent linear increase with fibre-matrix interfacial shear strength (IFSS), where the MW
cured G
1C
was greater than the oven cured G
1C
due to an increase in IFSS.

Frequently asked questions

Q1. What are the contributions in this paper?

Affiliations: 1 Integrity Management Group, TWI Ltd, Great Abington, CB21 6AL, UK 2 Joining Technologies Group, TWI Ltd, Great Abington, CB21 6AL, UK 3 The Composites Centre, Imperial College London, London, SW7 2AZ, UK * Corresponding author M. Kwak, Tel. +44 ( 0 ) 1223 899000 ; fax: +44 ( 0 ) 1223 892588 

Q2. What is the effect of the increase in compression strength of MW cured samples?

The increase in 0° compression strength of MW cured samples could lead to changes in design allowables, which in turn could lead to thinner sections, thus offering cost and weight reduction. 

Q3. How much thickness of the UD SPARPREG CFRP will be heated?

The MW penetration depth tests suggest that approximately a maximum of 2mm will heat through the thickness of the UD SPARPREG CFRP. 

Q4. What was the effect of the thermal imaging camera on the temperature gradient?

As the enclosure wall was very thin, and was made of a material with a high thermal conductivity, the temperature gradient caused by MW heating was rapidly and easily picked up by a thermal imaging camera. 

Q5. What was the process methodology used in the composites industry?

The oven cured laminates were produced using a single-sided aluminium mould by means of OoA vacuum bagging (at -1.0bar) using typical consumables and curing methodologies used in the composites industry. 

Q6. What is the effect of the indentation depths on the surface of the sample?

Lower indentation depths may potentially minimise this effect, however at small depths the indentation may suffer from indentation size effects, where the results become very sensitive to the surface roughness caused by the preparation of the sample. 

Q7. What is the main reason why MW cured CFRPs performed better under compression than oven?

The suitability of MW-heating of a material (particularly CFRPs) need to be investigated on a case-bycase basis, i.e. pre-work is required prior to attempting to process materials using MWs, since the global and local MW field will vary depending on factors such as the part’s geometry, temperature, dielectric and conductivity properties. 

Q8. What is the importance of the MW penetration depth of a material?

The establishment of the MW penetration depth of a material at a specific MW frequency is of much importance as this will determine whether the material under investigation will heat evenly through the thickness of the material. 

Q9. What is the reason for the increase in compressive strength observed with increasing dwell time?

Therefore it can be further deduced that the further increase in compressive strength observed with increasing dwell time by the MW cured coupons was due to the increase in degree of cure of the matrix away from the fibres after the interface reached a high degree of cure. 

Q10. What are the main reasons why there are relatively few publications in this topic/material?

The current summary will only focus on MW heating of carbon fibre reinforced polymer (CFRP) composites, more specifically carbon-epoxy composites, as these present some specific challenges (e.g. arcing, selective heating, etc.) other types of materials (e.g. thermosetting polymers, thermoplastics, glass-reinforced polymers) may not experience, and thus possibly the reason why there are relatively few publications in this topic/material. 

Q11. How was the delta H value measured from the sample ‘x’ compared to the reference sample?

In order to obtain the degree of cure of sample ‘x’, the delta H value measured from the sample ‘x’ was compared to the reference sample’s delta H value (Eq. 11).