Towards understanding the hole making performance and chip formation mechanism of thermoplastic carbon fibre/polyetherketoneketone composite

9 Here, we report the first study on the hole making performance of thermoplastic carbon 10 fibre/polyetherketoneketone (CF/PEKK) composite. Different hole making methods (conventional 11 drilling vs. helical milling) have been compared and the effect of different feed rates has been 12 investigated. The effect of thermal-mechanical interaction on the resulting hole damage has been 13 elucidated for the first time for carbon fibre reinforced thermoplastics (CFRTPs) hole making. In the 14 material science dimension, advanced material characterization techniques have been deployed to 15 reveal the material removal mechanisms at microscopic scale and unveil the underlying material 16 structural change at a molecular level. Results show that the delamination damage of CF/PEKK is a 17 result of the thermal-mechanical interaction. For conventional drilling, the high machining temperature 18 (at low feed rate < 0.1 mm/rev ) has a stronger influence on the delamination damage and the 19 delamination starts to show stronger dependence on the thrust force at high feed rate > 0.1 mm/rev. In 20 contrast, helical milling generates a much higher machining temperature which plays a more 21 predominant role in the associated delamination damage. Microstructural analysis shows that all the 22 hole surfaces feature matrix smearing, as a result of combined in-plane shear stress and high machining 23 temperature. Conventional drilling leads to more severe hole wall microstructural damage (matrix loss 24 and surface cavity) as compared to helical milling. Finally, thermal analysis reveals that the hole 25 making process has led to significantly increased crystallinity in the PEKK matrix as a result of the 26 strain-induced crystallization under the combined effect of shear stress and high temperature.


Introduction
1 Carbon fibre reinforced plastics (CFRPs) have been intensively deployed in aerospace and automotive 2 industries nowadays, as their high specific strength favours the lightweight design and energy saving 3 requirements of the modern industry. In recent years, there is a surge of interest in the application of 4 novel carbon fibre reinforced thermoplastics (CFRTPs) in the manufacturing sector, as these materials 5 demonstrate several prominent advantages. Compared to the conventional thermosetting CFRPs, 6 CFRTPs require less stringent storage condition and have infinite shelf-life under ambient conditions. 7 Their out-of-autoclave processing can achieve much shorter manufacturing cycle and requires less 8 energy consumption [1]. The thermoplastic nature of the material also endows CFRTP with better 9 recyclability and greater reparability [2], which can contribute greatly to the carbon emission reduction 10 in sustainable manufacturing . 11 Amongst the available CFRTPs, carbon fibre reinforced polyetherketoneketone composite (CF/PEKK) 12 stands out for its exceptional mechanical properties (tensile strength ~2.4 GPa), strong chemical 13 resistance, high thermal stability (glass transition temperature Tg ~160 ℃ ) and wide processing 14 window (330 -380 ℃) [3]. The roadmap for development of thermoplastic composites in Europe, 15 supported by Airbus and a variety of national aerospace consortia (see Fig. 1), suggests that CF/PEKK 16 will be top choice of composite in high end applications such as primary and secondary structural 17 components such as leading edges, floor panels, wing spars and engine pylons in future aircrafts [4-18 7]. Like many other CFRTPs, CF/PEKK parts can be manufactured into net shape through compression 22 moulding or assembled by means of welding [9]. However, secondary processing such as mechanical 23 riveting through fastener holes still remains an imperative manufacturing process, particularly in 24 joining of dissimilar materials (e.g. metal/CFRTP joining) and load bearing components [10]. In 25 composite part assembly, conventional drilling (CD) is the most commonly deployed hole making 26 technique due to its great efficiency. Other emerging hole making technology, such as helical milling 27 (HM), has also attracted increasing attention [11]. In HM, the cutter proceeds in both the tangential 28

2000 2010 (Year)
and the axial directions thus providing combined frontal and peripheral cutting. Compared with CD, 1 reduced burr formation, improved hole geometrical accuracy and smaller cutting force have been 2 reported in HM of aeronautical alloys [12] and thermosetting CFRPs [13]. 3 To date, a large body of literature has been dedicated to the study of hole making performance of 4 conventional thermosetting CFRP, covering wide ranging research topics including process 5 optimization [14][15][16], cutting tool design [17,18], cutting force modelling [19][20][21] and delamination 6 damage modelling [22][23][24]. The readers are referred to the review papers [18,25,26] in this field for 7 more information. In comparison, the number of studies concerning hole making performance of 8 CFRTP is very limited. The first study on hole making of CFRTP was conducted by Hocheng and Puw 9 [27] on drilling machinability of CF/acrylonitrile butadiene styrene (CF/ABS). They found that the 10 thrust force in drilling of CF/ABS was proportional to feed rate and highly finished surface (Ra < 1μm) 11 can be achieved for a wide range of drilling parameters (cutting speed = 1.96 -50 m/min and feed rate 12 = 30 -3000 mm/min). In a later study by Hocheng et  It is known that manufacturing of the thermosetting CFRPs involves a curing process. The resin 34 undergoes a non-reversible chemical reaction and form rigid, covalently boned crosslinks which 35 maintain their rigidity under elevated temperature [32]. In contrast, CFRTPs show typical temperature-36 and the deformation behaviour of the matrix. Ahmad et al. [36] and Xu et al. [30] found that the 23 elongated CFRTP chips adhered to the drill bit main cutting edge, which led to severe tool clogging, 24 deteriorated tool cutting performance and accelerated tool wear [36]. So far, no work has been done to 25 investigate the chip microstructure and the associated material removal mechanism of CF/PEKK. The 26 polymer chains within CFRTP matrix will be highly mobile when subjected to strong mechanical 27 shearing and high temperature during the hole making process. There has been no study dedicated to 28 understanding the microstructural evolution of machined CFRTP at a molecular level. 29 In this paper, the hole making performance of advanced thermoplastic CF/PEKK composite is reported 30 for the first time. The effect of different hole making methods (CD and HM) and feed rates on the 31 resulting delamination and hole wall microstructural damage will be discussed, considering the 32 thermal-mechanical interaction arising from the hole making process. The microstructure of the 33 formed chips will be analysed in detail to reveal the fundamental material removal mechanism. In 34 addition, thermal analysis will be carried out for the chips to elucidate the material molecular structural 35 evolution during the machining process. This study will not only reveal the fundamental material 36 science involved in the hole making process, but also provide a parametric and methodological 1 guidance for assembly of load-bearing aircraft parts involving CFRTP components (such as 2 composite-metal stacked structures in frames and wings) [37,38] Computer Numerical Control (CNC) machine, which is equipped with an in-house chip extraction 23 system to comply with health and safety regulations. Fig. 2 illustrates the details of the experiment 24 setup. A 3 mm thick steel plate with 10 mm diameter pre-drilled holes was placed on the CFRTP drill-25 exit side, to ensure consistent supporting stiffness for each hole made [42]. During the hole making 26 The hole-exit delamination was imaged by Alicona infinite focus G5 microscope (Bruker, UK, 2.5X 15 magnification). The exit delamination ( Fig. 3

grey area) can be quantified by the delamination factor 16
Fda [43], which can be calculated following Eq. 1-3: 17 where is the maximum diameter of damaged area, is the nominal diameter of the hole, 18 is the area related to the maximum diameter of the delaminated zone ( ), is the 19 nominal hole area and is the damaged area. The delamination factor was calculated through 20 image analysis using MATLAB R2019b software. 1 The hole wall microstructures and chip morphology were inspected by Scanning Electron Microscopes 2 (SEMs) FlexSEM 1000 (Hitachi Ltd., Japan) under an acceleration voltage of 5 kV. The samples were 3 sputter-coated with gold before SEM observation. The thermal properties of the CF/PEKK samples were measured using a thermogravimetric analysis 8 (TGA) and differential scanning calorimetry (DSC) thermal analysis system (TGA/DSC2, Switzerland) 9 under a N2 atmosphere in a temperature rage 30 to 550 ℃ and a heating rate of 10 ℃/min.

×100% (4)
Where Δ is the melting enthalpy, Δ is the cold crystallization enthalpy, Δ 100% = 130 J/g is 12 the melting enthalpy of the theoretical 100% crystalline PEKK [45] and = 66% is the fibre 13 weight fraction in the CF/PEKK composite. 14 15 The average stable cutting phase thrust force generated under different feed rates is shown in Fig. 5. 6 HM generated a consistently low thrust force (~ 50 N), which is independent of the feed rate used. 7

Results and discussion
Under CD however, the measured thrust forces are significantly higher than that of HM and shows a 8 clear increasing trend with the feed rate. 9 The difference in thrust force seen in CD and HM can be attributed to the distinctly different material 10 removal mechanisms involved in the two processes. For CD, the cutting speed near the borehole centre 11 approaches zero. The material removal is achieved by extrusion in the axial direction rather than 12 cutting [13], which induces a much higher axial thrust force. For HM, a large portion of material 13 removal is achieved by the peripheral cutting edge as the tool travels along its helical tool path. As a 14 result, less force is exerted in the axial direction. 3.2. Tool temperature 7 The heat generated during the hole making process is primarily due to the shear deformation of 8 CF/PEKK and the friction between the cutting tool and the workpiece [47]. The accumulated heat is 9 then distributed to the chips, the workpiece and the cutting tool. According to previous experimental 10 and numerical simulation studies by Thirukkumaran et al. [48], the workpiece temperature follows a 11 quasi-linear increasing trend with cutting tool temperature. Given the workpiece hole-exit temperature 12 cannot be measured directly with our existing experimental setup, hole-exit tool temperature was used 13 to inform the machining temperature. The tool temperature under different hole making methods and 14 feed rates is shown in Fig. 6 feed rate, the delamination seen for HM is more severe than CD. 14 The variation of delamination factor against the feed rate is depicted in Fig. 7. The corresponding tool 15 temperature and the thrust force was also included in the plots. For CD, the initial declining Fda (F < 16 0.1 mm/rev) correlates well with the declining tool temperature. Under a lower feed rate (F < 0.1 17 mm/rev), the machining temperature plays a more predominant role in the delamination damage, as 18 the much higher machining temperature under such condition can lead to significant loss of stiffness 19 of the CF/PEKK ply which makes it more prone to delaminate. At higher feed rate, the machining 20 temperature decreased significantly and the drastically increased thrust force would take over its 21 influence on the delamination damage. This is reflected by the increased Fda with increasing thrust 22 force for F > 0.1 mm/rev. For HM, the thrust force is independent of feed rate and remained at a constant low level (50 N). At 6 low feed rate (0.025 mm/rev), the tool temperature approaches the Tg of PEKK. The gradual declining 7 trend of tool temperature also correlates well with the decreasing Fda, implying the influence of tool 8 temperature on the HM delamination 9 The finding of this study is in contrast to the reports [25,46,51] on thermosetting CFRP, where thrust 10 force is considered the sole factor determining the delamination damage. In thermosetting CFRP, hole-11 exit delamination is mainly caused by mode I opening fracture as the cutting tool approaches the hole 1 exit and the resulting delamination factor increases with thrust force [18,25,26]. For CF/PEKK 2 however, the elevated temperature caused by the machining process also plays a significant role in the 3 resulting delamination. Given the thermoplastic nature of the PEKK matrix, high machining 4 temperature (approaching Tg) can result in softened matrix [52] and weakened matrix 5 support/encapsulation around the fibre bundles, which contribute to the more severe delamination 6 damage. 7

Hole wall microstructure 8
Microstructural damage such as fibre breakage, matrix loss and debonding can occur during hole 9 making of CFRP. The typical hole wall surface SEM images produced under different hole making 10 conditions is shown in Fig. 8 and Fig. 9. 11 For CD, low feed rate (F = 0.025 mm/rev) resulted in relatively smooth hole wall surface finish, with 12 negligible matrix loss, see Fig. 8 (a1). Increasing feed rate has led to more severe matrix loss and 13 surface cavity as shown in Fig. 8 (b1) and (c1). The matrix loss and surface cavity typically take place 14 at the obtuse fibre orientation where the material was removed by bending fracture [53]. The cracks 15 and tears caused by the bending stress can be easily transmitted to the hole wall subsurface, resulting 16 in distorted pits when the chip was separated from the workpiece [53]. As can be seen in Fig. 8 (a2-c2) 17 (a3-c3), for both 0° and 90° fibre orientation, the extent of hole wall damage increases with increasing 18 feed rate. The hole wall produced under low feed rate (F=0.025 mm/rev) mainly features matrix 19 smearing, where the fibre is well covered by the thermoplastic PEKK matrix. When the feed rate is 20 increased from 0.025 mm/rev to 0.2 mm/rev, more severe fibre debonding and breakage are evident 21 on the hole surface. 22 The hole wall microstructure under HM condition is significantly different from that of CD, see Higher feed rates tend to produce slightly thicker chips, as the volume of material removed per tooth 7 increases with the feed rate [13]. 8 Table 3 Morphology of chips produced in hole making of thermoplastic CF/PEKK composite 9

Chip thermal properties 1
To evaluate the potential impact of the hole making process on the machined composite molecular 2 structure, thermal analysis was carried out for the machined chips using DSC technique, see Fig. 11. 3 The corresponding material thermal parameters obtained from DSC analysis were summarized in Table  4 4. From Table 4, it can be seen that the crystallinity data follows the trend CD > HM > control, 5 with CD under F = 0.2mm/rev giving the highest (28.35%), representing > 300% increase against 6 the control (unmachined CF/PEKK).  phenomenon was discovered for machined CFRTP. During machining, the tool cutting edge exerts a 5 significant in-plane shear stress on both the hole wall and the adjacent chip. PEKK is a semi-crystalline 6 polymer consisting of both crystalline and amorphous regions. Under the shear action, randomly 7 arranged molecular chains within the PEKK amorphous region will be stretched and re-aligned along 8 the shear direction, forming more orderly arranged crystalline region, as illustrated by Fig. 12. As shown in Fig. 13 (a), for 0° fibre orientation, the cutting edge exerts a compressive stress to the 2 workpiece along the fibre direction and an in-plane shear stress is generated along the fibre-matrix 3 interface [53]. With the advancement of the tool, the removed carbon fibre will be subjected to 4 extensive bending against the tool rake face. Brittle fibre fracture occurs when the external stress 5 exceeds the bending strength limit of carbon fibre. Despite the segmented fibres within, the chip 6 remains an overall intact morphology, as the ductile PEKK matrix can sustain a significant amount of 7 plastic deformation, holding the fractured fibres in place, see Fig. 13 (b). On the machined hole surface, 8 the high machining temperature can soften the matrix, and the highly viscous matrix can be re-casted 9 and smeared at certain regions of hole surface (see Fig. 13 (c, d)). The evident fibre breakage on the 10 machined hole surface is a result of the compressive action exerted by the tool cutting edge and rake 11 face in the workpiece thickness direction. For 90° fibre orientation as illustrated by Fig. 14 (a), the cutting speed is perpendicular to the fibre 18 direction. The carbon fibres undergo brittle fracture as a result of the shearing action of the cutting 19 edge and the fracture plane is perpendicular to the fibre direction. Again, despite its plastic deformation, 20 the highly ductile PEKK matrix has held the broken fibres in place, Fig. 14 (b). Similar matrix smearing 1 was found on the hole surface. Bending-induced fibre debonding and fibre breakage have been found 2 on the machined hole surface. Some of the broken fibres were found to be embedded in the smeared 3 matrix on the machined hole surface, Fig. 14 (c, d). It is expected the unique material removal mechanism and the associated material structural change 10 discovered in this study can be generalized to a wide of range of CFRTP materials, such as CF/PEEK, 11 CF/PPS, CF/ABS, etc. Our preliminary work also inspires the research into orthogonal cutting of 12 CFRTPs, through which more comprehensive understanding on the effect of cutting depth, cutting 13 speed and temperature on the material removal mechanism can be elucidated. Although in-depth 14 microstructural and thermal analysis has been carried out for chips in this study, measuring thermal  CD generates higher thrust force than HM. The CD thrust force and feed rate follow a quasi-linear 9 increasing relationship, whereas the thrust force of HM is independent on its feed rate. 10  The tool temperature for both CD and HM decreases with the feed rate. HM shows a much higher 11 tool temperature than CD, and this can be attributed to its prolonged tool engagement time and the 12 associated heat accumulation. 13  The CF/PEKK delamination damage is a result of combined thermal-mechanical interaction. For 14 CD under low feed rate, the high machining temperature plays a more predominant role whereas 15 under higher feed rate (lower temperature), the high thrust force would be more dominating. For 16 HM, the delamination damage formation is mainly influenced by the high machining temperature, 17 as the thrust force remains constant within the range of feed rate being investigated. 18  Matrix smearing has been observed on both CD and HM hole walls. This is due to the softening 19 and recasting of the highly ductile PEKK matrix onto the hole surface. Hole wall microstructural 20 damages such as matrix loss and surface cavity are more severe for CD. 21  CD leads to formation of continuous chips, which can be attributed to the continuous material 22 removal process during drilling and the excellent ductility of the PEKK matrix. Under lower feed 23 rate (< 0.1 mm/rev), the long and folded chips tend to clog the tools. Under higher feed rate (> 0. 1 24 mm/rev), shorter, spiral shaped chips can be effectively evacuated. In contrast, HM produced short 25 fragmented powdery chips as a result of the intermittent material removal process. 26  Chips produced by both CD and HM show increased crystallinity as a result of shear induced 27 crystallization, with CD (0.2 mm/rev) giving the greatest (300%) crystallinity enhancement. 28  Through advanced material characterization, a greater insight has been developed into the 29 machined CFRTP deformation characteristics, material removal mechanisms, and the associated 30 material structural evolution at microscopic and molecular levels. These findings would inspire 31 future researchers to better deploy the material process -structure -property relationship for 32 optimized CFRTP manufacturing. Declaration of competing interest 9 The authors declare that they have no known competing financial interests or personal relationships 10 that could have appeared to influence the work reported in this paper. 11 12

Acknowledgements 13
The funding support from EPSRC projects EP/P025447/1 and EP/P026087/1 is acknowledged. This