Introduction
What are Carbon Fibre Reinforced Polymers?
Carbon fibre reinforced polymers (CFRPs) consist of carbon fibres bound together by a resinous matrix. They are becoming more widely used across a number of sectors such as in the aerospace, automotive and wind energy industries. However, the multi-component nature of these materials make them particularly difficult to recycle [1].
What are Supercritical Fluids?
Supercritical fluids (SCFs) form when the critical point (a certain temperature and pressure) is exceeded. They exist as neither a gas nor a liquid but exhibit properties between the two phases. They diffuse like a gas but dissolve materials like a liquid thereby facilitating excellent mass transfer and enhancing reaction rate [2].
Research Aims
Currently, waste CFRPs are pyrolysed and the resin is wasted. Solvolytic technology has the potential to recover fibres and organic compounds, improving resource efficiency. This work aims to study the degradation of a CFRP using a supercritical acetone/water solvent and assess the impact of the recycling conditions on the fibre properties.
Materials & Methods
An acetone/water solvent was supplied in a ratio of 80:20 v/v. The temperature range investigated was 300 to 380°C in reaction times of 0 to 2.5 h. The resin removal yield (RRY) was quantified through measuring the difference in mass and by calcination.
Resin Degradation
Resin Degradation
Results: Resin Degradation
Figure 1 shows that there is a strong dependence of temperature on RRY; an increase from 300 to 320°C more than doubles the RRY from 41 to 90% after processing for 120 min. A further increase of 20°C completely strips the resin from the fibre surface in just 45 min.
XRCT imaging of partially degraded samples show that when the RRY is 79%, the solvent molecules have penetrated through the plies in the X and Z directions. This suggests that the overall degradation process is not limited by the mass transfer of the solvent or degradation products but depends on the rate of the reaction.
Upon increasing the reactor loading (Figure 3), there is little difference observed for each of the conditions investigated. This suggests that the solvent is always in sufficient excess and the rate therefore does not depend on the concentration of resin in the solvent.
A reaction-rate limited, shrinking core model [3] was fitted to the data using Equation 1. The rate constant, k, was obtained for each temperature investigated. As there was a significant heating phase, both reaction, tR, and heating, tH, times were included.
As rate constants at various temperatures have been determined, it is possible to relate this to Arrhenius reaction kinetics. Figure 5 shows the resulting Arrhenius plot of ln(k) vs. 1/T which gives an activation energy, EA, and frequency factor, k0, of 164 kJ.mol-1 and 1.29 x1012 min-1 respectively.
Fibre Characterisation
XRCT
Partially degraded samples were imaged with X-Ray Computer Tomography using a Bruker Sky Scan 1172 and an acceleration voltage of 25 kV.
SEM
Fibres were coated with platinum and inspected with a Philips XL30 ESEM at magnifications from 2,000 to 17,500 using a voltage of 20 kV.
Figure 1 shows that there is a strong dependence of temperature on RRY; an increase from 300 to 320°C more than doubles the RRY from 41 to 90% after processing for 120 min. A further increase of 20°C completely strips the resin from the fibre surface in just 45 min.
Figure 1. Effect of changing time and temperature on resin removal yield
XRCT imaging of partially degraded samples show that when the RRY is 79%, the solvent molecules have penetrated through the plies in the X and Z directions. This suggests that the overall degradation process is not limited by the mass transfer of the solvent or degradation products but depends on the rate of the reaction.
Figure 2. XRCT images where RRY = 79%
Upon increasing the reactor loading (Figure 3), there is little difference observed for each of the conditions investigated. This suggests that the solvent is always in sufficient excess and the rate therefore does not depend on the concentration of resin in the solvent.
Figure 3. Effect of increasing reactor loading on resin removal yield
A reaction-rate limited, shrinking core model [3] was fitted to the data using Equation 1. The rate constant, k, was obtained for each temperature investigated. As there was a significant heating phase, both reaction, tR, and heating, tH, times were included.
Figure 4. Model (line) and experimental data points at various process conditions
As rate constants at various temperatures have been determined, it is possible to relate this to Arrhenius reaction kinetics. Figure 5 shows the resulting Arrhenius plot of ln(k) vs. 1/T which gives an activation energy, EA, and frequency factor, k0, of 164 kJ.mol-1 and 1.29 x1012 min-1 respectively.
Figure 5. Arrhenius plot of rate constants obtained from shrinking core model
Fibre Characterisation
XRCT
Partially degraded samples were imaged with X-Ray Computer Tomography using a Bruker Sky Scan 1172 and an acceleration voltage of 25 kV.
SEM
Fibres were coated with platinum and inspected with a Philips XL30 ESEM at magnifications from 2,000 to 17,500 using a voltage of 20 kV.
SFTT
A filament with a gauge length of 20 mm was tested with an Instron 5566 and 10 N load cell at a crosshead speed of 0.2 mm.min-1 following ISO 11566.
Results: Fibre Characterisation
A filament with a gauge length of 20 mm was tested with an Instron 5566 and 10 N load cell at a crosshead speed of 0.2 mm.min-1 following ISO 11566.
Results: Fibre Characterisation
Figure 6. SEM images of fibres recovered after processing at a) 320°C, 2 h (mag = 2000x); b) 330°C, 1.5 h (mag = 2,500x); c) 340°C, 45 min (mag = 2,500x) and d) 340°C, 45 min (mag = 17,500)
Fibres recovered after processing at 320°C and 330°C appeared clean as illustrated by Figure 6. At 320°C, all plies were perfectly separated, however Figure 7 also shows that at 340°C, fibres started to exhibit a fluffy quality which limits future applications. Furthermore, minor surface fractures are also visible in Figure 6d after processing at these conditions.
Figure 7. Fibres recovered after processing at a) 320°C, 2 h; b) 340°C, 45 min and c) 360°C, 15 min
Figure 8 represents typical failure curves obtained during single fibre tensile testing. At least 30 fibres were examined for each of the given conditions and the failure stress for each was used to calculate the tensile strength using Equation 2.
Large variations in strength are expected due to fibres failing at a randomly distributed surface defect. To characterise these materials, Weibull distributions can be used where the scale factor (calculated using Minitab 2017 Statistical software) represents the tensile strength. Figure 9 shows there is little difference between the virgin and recovered fibres
The Weibull plot for all data recorded is shown in Figure 10. The strength distribution of fibres recovered after each process condition is represented by the linear regression lines. There is a high level of agreement between the data and corresponding curve.
Figure 8. Typical stress-strain curves of individual fibres recovered after processing at various conditions
Large variations in strength are expected due to fibres failing at a randomly distributed surface defect. To characterise these materials, Weibull distributions can be used where the scale factor (calculated using Minitab 2017 Statistical software) represents the tensile strength. Figure 9 shows there is little difference between the virgin and recovered fibres
Figure 9. Tensile strength (scale factor) of fibres recovered after processing at various conditions
The Weibull plot for all data recorded is shown in Figure 10. The strength distribution of fibres recovered after each process condition is represented by the linear regression lines. There is a high level of agreement between the data and corresponding curve.
Figure 10. Weibull distribution of all single fibre tensile tests
Conclusions
- A CFRP was fully degraded in 320°C, 2 h or 330°C, 1.5 h. Higher temperatures give shorter process times but compromises fibre quality.
- It is possible to fit the data to a shrinking core model and use this to predict the resin removal yield and calculate Arrhenius parameters.
- SEM images show clean fibres are recoverable, however, at 340°C, small fractures are visibile on the fibre surface.
- Tensile strength does not appear to be affected by the recycling process and data from SFTT can be fitted to a Weibull distribution.
1 School of Chemical Engineering, University of Birmingham;
2 School of Water, Energy & Environment, Cranfield University
The 10 largest coal producers and exporters in Indonesia:
- Bumi Resouces (BUMI)
- Adaro Energy (ADRO)
- Indo Tambangraya Megah (ITMG)
- Bukit Asam (PTBA)
- Baramulti Sukses Sarana (BSSR)
- Harum Energy (HRUM)
- Mitrabara Adiperdana (MBAP)
- Samindo Resources (MYOH)
- United Tractors (UNTR)
- Berau Coal
