Wear behaviour of CVD diamond-coated tools in the drilling of woven CFRP composites

Author’s Accepted Manuscript Wear behaviour of CVD diamond-coated tools in the drilling of woven CFRP composites Chunliang Kuo, Chihying Wang, Shunkai Ko

www.elsevier.com/locate/wear

PII: DOI: Reference:

S0043-1648(16)30278-2 https://doi.org/10.1016/j.wear.2017.11.015 WEA102297

To appear in: Wear Received date: 9 September 2016 Revised date: 22 September 2017 Accepted date: 21 November 2017 Cite this article as: Chunliang Kuo, Chihying Wang and Shunkai Ko, Wear behaviour of CVD diamond-coated tools in the drilling of woven CFRP composites, Wear, https://doi.org/10.1016/j.wear.2017.11.015 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Wear behaviour of CVD diamond-coated tools in the drilling of woven CFRP composites Chunliang Kuo*, Chihying Wang, Shunkai Ko School of Mechanical Engineering, National Taiwan University of Science Technology, 43 Keelung road, Sec. 4 Taipei, Taiwan, 10607 * Corresponding author. Tel.: +886 2 2733 3141#6448; fax: +886 (0)2 2737 6460. [email protected]

Abstract The high specific strength/modulus of carbon fibre reinforced polymer composites (CFRPs) has led to their widespread use in the aerospace industry. Unfortunately, woven CFRP laminate can be exceedingly difficult to drill, resulting in rapid tool wear and fracturing. This work examines tools with novel geometric designs, such as double-point and multi-facet drills with diamond coatings. Experiments were conducted to characterize tool wear and fracture modes at two cutting speeds (50 and 75 m/min) and three feed rates (0.05, 0.1, 0.15 mm/rev) under dry conditions. This produced thrust forces of up to 128.28 N and torque values reaching 0.33 Nm, with a measuring temperature of ~135 °C. The double-point drill presented various tool wear patterns, including chipping and delamination of the diamond-coated layer. The tool wear patterns on the multi-facet drill included progressive abrasion wear, scoring and severe three body abrasion along the cutting lips on the first, second and the third facets respectively. The double-point drills presented fatigue ruptures, whereas the multi-facet drills showed chipping and micro-cracks under the chattering and excessive vibration during drilling. When the thermally induced static stress was imposed on the effects of the compressive residual stress, the observed Raman shift of the diamond structure can be the result.

1

Keywords Woven CFRP laminate, CVD diamond-coated tool, drilling, tool wear, fracture mode

1. Introduction Carbon fibre reinforced polymer composites (CFRPs) are increasingly being used in the aerospace industry for their high specific strength/modulus [1, 2]. When assembling CFRP components in structures, these laminates are held together mechanically by bolting, screwing, or riveting [3]. This involves fitting mechanical fasteners into machined holes drilled through the structural laminate materials with highly effective drilling methods; a new strategy of implementing CVD diamond coating on the tools was recommended [4]. The quality of the resulting holes usually depends on the cutting forces, in terms of thrust force and torque, which are intermediate variables that vary with cutting conditions [5]. Thus, the wear resistance of the cutting tool is the dominant factor for prolonging the tool life and achieving a quality drilled hole. During machining of CFRP laminates, some of the problems encountered included rapid tool wear, delamination, cracking, and thermal degradation of the underlying matrix [6]. The wear modes of hard coatings, such as CVD diamond, do not easily fit into the wear modes (ISO 3685) or associated fracture patterns in metal cutting. Cardorin and Zitoune [7] suggested that the tool wear mechanism in diamond-coated twist drills is primarily abrasion on the flank face and delamination of the coating layers. Raghuveer et al. [8] suggested that delamination of the coating layer was caused by chipping failure; whereas the rounding of the cutting edges followed the detachment of the diamond layer from the tungsten carbide substrate, which subsequently decreased the hardness and led to rapid wear. Rawat and Attia 2

[9] reported that the progression of tool fractures starts with chipping, followed by abrasion wear and the possible adhesion of carbon fibre fragments on the flank face in the high-speed drilling of woven carbon fibre composites. Salgueiredo [10] proposed that the self-mated tribological system in the multilayer micro/nanocrystalline CVD diamond coatings could reduce the friction coefficient (0.06) and retard the wear rate (~2.4×10-7 mm3 N-1 m-1). This is due to the soft (sp2) diamond structure creating a state of low residual stress for the hard (sp3) underlying diamond coating, thereby preventing it from delaminating. Dumpala et al. [11] suggested that the Raman peak shift at 1332 cm-1 was associated with a stress-free crystalline diamond coating; whereas the shift in the Raman peak towards 1336 cm-1 was due to the presence of compressive residual stress on the surface of the coating. The wear patterns found in conventional drilling of CFRP composites are dominated by the tools’ material, geometry and operating parameters. The application of hard coatings, such as CVD diamond, has been shown to prolong tool life without compromising hole quality. Murakawa and Takeuchi [12] reported that thin film coatings produced via CVD technology were composed of fine crystals with random orientation, resulting in a high degree of hardness (85 GPa), high strength adhesion, a low friction coefficient and low manufacturing cost. In contrast, Söderberg et al. [13] reported that stacked columnar structures can form voids and dislocations during film growth, leading to the bonding of a non-diamond structure at the interface, which could produce residual compressive stress in the film. This phenomenon can lead to plastic deformation and/or delamination under heavy loading, which invariably leads to a reduction in strength. Iliescu et al. [14] suggested that double-point angle geometry (125º × 90º) produces the lowest thrust force, despite the fact that the diamond coating is not used. When diamond coatings are utilized, the tribological performance determined by the coating’s film structure has a greater impact than the tool’s geometry. When tool wear patterns varied with the point geometry, Faraz et al. [15] found that cutting edge rounding (CER) patterns are similar among twist drills and distinct from the flank wear 3

patterns that form when drilling CFRP laminate. They obtained a correlation between the delamination factor and the CER, with the validity made against conventional flank wear (VB). Similarly, Gaugel et al. [16] quantitatively defined tool wear by measuring the cutting edge rounding and delamination factors at the hole’s entry/exit in the drilling of unidirectional carbon/epoxy using diamond-coated/uncoated drills. They reported that the preliminary sharpening at the cutting edges was possibly ground by the stripped diamond residues; however, it did not effectively influence the tool wear. Whereas the delamination factor at the hole’s exit consistently reflected a strong correlation to the tool wear. In contrast, Wang et al. [17] suggested that the cutting edge rounding and abrasive wear did not effectively correlate to the tool wear in the drilling of CFRP laminates. They validated the correlations of the sliding wear rates between the ball-on-disk tribometer tests to those from drilling tests under a fixed feed rate and cutting speed of 0.0762 mm/rev and 180 m/min respectively. Alternatively, Poulachon et al. [18] highlighted that the progressive tool wear at the drill corner/margin could effectively influence the machined hole diameters in the drilling of unidirectional and multidirectional CFRP composite plates, with a feed rate and cutting speed of 0.05 mm/rev and 100 m/min respectively. They reported that the different wear regimes were related to the ratio between the surface of the uncut fibres and the total hole surface. Knowing the tool wear patterns were influenced by the cutting parameters, Karpat et al. [19] reported that severe fractures of the diamond coating were observed on the flank face of the drills under low feed rates (0.4-0.15 mm/rev) and cutting speeds (100-200 m/min), due to the cutting edge rounding and ineffective cutting performance. However, when the feed rate and cutting speed was increased to 0.225 mm/rev and ~299 m/min respectively, the fracture area on the diamond coating was reduced. Similarly, Ramirez et al. [20] investigated tool wear when drilling unidirectional CFRP composite plates under a constant feed rate and cutting speed of 0.05 mm/rev and 100 m/min respectively. They found that the wear at the cutting edge did not consistently withdraw along the flank face, particularly when chipping occurred 4

on the rake face. Chen [21] conducted experiments to examine the effects of cutting temperature on tool wear. He concluded that cutting temperature on the flank face increases with cutting speed (40-200 m/min), but decreases with feed rate (0.05-0.4 mm/rev), due to a reduction in the time available for the accumulation of heat (350-60 ºC). The designed fibre orientation in the CFRP composite materials could detrimentally influence the cutting tools. The inhomogeneous and anisotropic nature of the materials configured by the fibre orientations could vary the cutting force, leading to discontinuous machining and an inconsistent cutting force [22-25]. Wang et al. [22] classified the chip formation process in orthogonal cutting of FRP’s into five typical models (Type I to V) with respect to fibre orientation and tool rake angle, giving insights of possible wear patterns in the cutting tools. In addition, the greater cutting edge radius (depending on the ratio to the depth of the cut) could suppress the matrix instead of being sheared, resulting in a ‘bouncing-back’ effect and increased contact force. Isbilir and Ghassemieh [23] reported that the abrasive carbon particles produced a three-body grinding effect at the interface of the machined surface, tool rake face and flank face when drilling CFRP using TiAlN/TiN-coated tungsten carbide tools. This phenomenon increased the crater and flank wear, rounded the cutting edge and increased the thrust force, torque and surface roughness. In this work, prepared thin film (~19 µm) diamond-coated drills were utilized to investigate the tool wear and failure mode in the dry drilling of woven composite at the moderate cutting speeds (50-75 m/min) and feed rates (0.05-0.15 mm/rev) currently used in industrial production. Cutting forces and tool wear were investigated quantitatively; whereas wear mechanisms and fracture modes were identified via a scanning electron microscope and energy dispersive X-ray spectroscopy. Allotrope transformation in the diamond coating was identified using Raman spectroscopy.

5

2. Analysis of cutting forces and tool wear mechanism in CVD diamond-coated tools During cutting, the fracture modes associated with diamond-coated drills may involve cyclic flaking and delamination at the cutting lips. This phenomenon usually begins with build-up along the cutting edges, which unbalances forces along the cutting lips, and is followed by fracturing along the chisel edge and peripheral corners. Differences in the thermal expansion coefficient along the interface between the coating and substrate produce unbalanced internal stress, resulting in shear force. Shear stress exceeding the threshold of fracture toughness in the coating layer can trigger flaking and delamination. Low frequency load can also be the source of fatigue failure [27]. Cyclic loading at the cutting lip can lead to the initiation of cracks in the rim area, which tend to propagate rapidly as micro-cracks coalesce into striations, leading to catastrophic failure. Fig. 1(a) presents a schematic illustration of a standard drill with a single cutting edge. The dynamic rake angle (αd) and dynamic relief angle (Ωd) clearly vary with the feed rate and cutting speed, resulting in deviations by adding the velocity angle (ψ) to the normal rake angle (αn) and subtracting the feed angle (η) to the normal relief angle (Ωn) as shown in Eqs. (1), (2) and (3), respectively [28]. tanαn =

R R pheriphery

×

tanλ χ tan(2)

(1)

𝜒

tanψ = tanβ × cos( 2)

tanη =

𝑓 𝜋𝐷

(2)

(3)

where R is the variation in radius; Rpheriphery is the radius of the drill; λ is the helix angle, χ is the point angle; β is the friction angle; f is the feed rate; and D is the diameter of the drill. Theoretically, thrust force varies with tool geometry and feed rate. It increases with a reduction in the effective rake angle (α) and shear angle (Ø) [29], but decreases with a 6

reduction in the friction angle (β). In the chisel zone, thrust force was greatly exerted due to the negative rake angle at the chisel edge and nearly zero cutting speed at the chisel centre, leading to ineffective cutting [30, 31]. In contrast, the greatest torque is caused by the cutting force acting on the cutting edge and the frictional force on the tool periphery. Thus, extreme cutting force values were obtained either along the chisel edge or in peripheral corners. Fig. 1(b) presents the meshed model (Linear Hexahedron: C3D8R) with 2666 and 9720 elements in the coating layer and body respectively, for the analysis of the three-dimensional stress. In the simulation, materials in the coating layer (E = 1200 GPa and Poisson’s ratio: 0.03) and the body (E = 700 GPa and Poisson’s ratio: 0.3) were elastically and isotopically deformed, based on the input loads. Aside to the cutting force induced fractures, the seizure zone on the tool rake face bears intermittent loads due to the inhomogeneous and anisotropic properties in CFRP composites. The loads acting on this seizure zone in the very short chip-tool contact length (~20 μm) could initiate a fatigue crack.

Fig. 1. Schematic illustration of (a) velocity angle and feed angle and (b) static stress on a single cutting edge tool 7

According to the cutting mechanism, a reasonable wear process of a coated diamond tool during drilling of CFRP could be proposed as in Fig. 2. In the schematic illustration of the tool wear evolution, the wear behaviour could be classified into four stages. Fig. 2(a) shows the continued adhesion of polymeric residue on the rake and flank faces of the cutting tool due to high temperature induced by high cutting speed. The varied geometry at the cutting edge reduces the relief angle thereby diverting the cutting actions into ploughing and rubbing actions. In Fig. 2(b), abrasion wear on the flank face increases steadily due to the compression stress on the sliding zone being high and also due to the increase in friction heat and temperature. Scoring wear on the successive flank face may occur due to the removed abrasive mixtures pushing by the fibres bouncing back on the machined surface. The wear process may involve graphitization in the diamond under a high cutting temperature, attrition in the columnar diamond grains and the three body abrasion wear which is produced by the mixture of the diamond debris (HV~10000) and carbon fibre fragment (HV~500). As shown in Fig. 2(c), due to the increased intermittent loads on the seizure zone, diamond film failure in terms of flaking and delamination occurs along the cutting edge, which exposes the sharp edge of the tungsten carbide substrate which is simultaneously re-sharpened under the grinding effects of the carbon fibre abrasives. As seen in Fig. 2(d), when the CVD diamond-coated layer had entirely flaked off, the exposed tungsten carbide edge was rapidly worn away; thereby greatly increasing the cutting force. This was due to the rounding of the cutting edge as well as a loss of clearance and rake angle during cutting.

8

Fig. 2. Schematic illustration showing tool wear patterns in (a) adhesion and abrasion wear, (b) extended abrasion wear on the flank face, (c) flaking and delamination and (d) breakage of substrate material.

3. Experimental work 3.1 Some properties of workpiece, cutting tool, and test rig set-up The workpieces in Fig. 3(a) are woven CFRP laminates, comprising carbon fibre (vf=58% vol.) and epoxy resin forming woven fabric prepreg (CL2 STY 3K-70-PW). The laid-up prepreg was 16 ply in a 0˚/90˚ orientation, which resulted in an overall thickness of 3.5 mm. The laminates presented a tensile strength of 75 MPa and elastic modulus of 7-9.5 GPa, respectively. The composite materials used for the measurement of the cutting forces, tool wear and fracture patterns were in 120 mm×20 mm×3.5 mm strips. Whereas the tool life was provided in the mainstream tests using the square plates measuring of 320 mm×240 mm×3.5 mm. Fig. 3(b) presents the geometrical features of the different, i.e., multi-facet and double-point angle designs with a CVD diamond coating layer of 19 μm. The CVD diamond coating was characterised with the low coefficient of thermal expansion (CTE) of 1.0 µm/K and high elastic modulus (E) of ~1200 GPa; whilst those of the tungsten carbide body were 9

5.8 µm/K and ~610 GPa respectively. The tool geometry data are detailed in Table 1.

Fig. 3. (a) Woven CFRP workpiece material and (b) drills utilised in the experiment Table 1 Tool geometry Specific angle

Double-point drill Multi-facet drill

1st point angle (°) 2nd point angle (°) 3rd point angle (°) 1st rake angle (°) 2nd rake angle (°) 3rd rake angle (°) Helix angle (°)

120° 20° ~26° ~78° 40°

120° 30° 6° 0° 0° 0° 0°

Relief angle (°)

14°

12°

Fig. 4(a) presents the machine set-up for the mainstream tool life testing on the woven CFRP composite plates. Experiments were conducted using an Akira Seiki SR3XP with a maximum power of 16 kW and a maximum spindle speed of 11,000 rpm. Fig. 4(b) details the set-up of the machining fixture assembled with the dynamometer when drilling woven CFRP composite strips. The infrared sensor manufactured from Calex Electronics can measure temperatures ranging from -40 °C to 1000 °C with the associated response time of 240 ms and 10

the accuracy of ±1% of 1 °C.

Fig. 4. Set-up of the experiment for the (a) mainstream testing and (b) measurements of cutting forces and temperatures 3.2 Design of experiment In this study, a full-factorial experiment was used to design the drilling processes. The experiment was based on variations in tool geometry (two levels), cutting speed (two levels), and feed rate (three levels), with two repetitions which entailed a total of 24 tests, as shown in Table 2. The laid-up orientation of the workpiece, drill diameter and diamond coating were unchanged, as shown in Table 3. The parameters implemented in the 12 tests were formulated as a full-factorial array (2×2×3), as shown in Table 4. The cessation criteria in the tests was 140 holes to determine flank wear in the coating against ISO standard (3685), which specified flank wear as 300 μm (VB=300) in the substrate.

11

Table 2 Variable parameters Factor Tool geometry

Level 1 Level 2 Double-point drill Multi-facet drill

Cutting speed (m/min) Feed rate (mm/rev)

50 0.05

75 0.1

Level 3

0.15

Table 3 Fixed parameters Factor Workpiece material Drill diameter Diamond coating/ Coating thickness

Description CFRP ([0º/90º]8S) 6.35 mm Thin film CVD coating/ 19 μm

Table 4 Full-factorial (2×2×3) array Cutting Feed rate Response Test Tool geometry speed (mm/rev) 1/2/3 No. 1 Double-point (m/min) 50 0.05 2 3 4 5 6 7 8 9 10 11 12

Double-point drill Double-point drill Double-point drill Double-point drill Double-point drill Multi-facet drill Multi-facet drill Multi-facet drill Multi-facet drill Multi-facet drill Multi-facet drill drill

50 50 75 75 75 50 50 50 75 75 75

0.1 0.15 0.05 0.1 0.15 0.05 0.1 0.15 0.05 0.1 0.15

3.3 Data acquisition system Thrust force and torque were measured on a platform mounted with a Kistler 9272 piezoelectric dynamometer. The force signals were transmitted to a Kistler 5070A charge amplifier and analyzed using Dynoware. Tool flank wear was determined according to ISO3685 standards. The fracture modes and worn surface of the tools were characterized using a scanning electron microscope (SEM: Jeol-6390LV) with a focusing field of 5–200 nm 12

and maximum magnification of 300,000×. The ingredients of the residue on the cutting lips were identified using energy dispersive x-ray (EDX: Inca 4.07) spectroscopy. The allotrope transformation in the diamond coating was examined by Raman spectroscopy.

4. Results and discussion 4.1 Analysis of cutting force Fig. 5(a) presents the variation of the cutting force when drilling for 2.935-6.525 sec in Test 1. Thrust force and torque were continuously recorded from stage I to V with the corresponding drilling actions taken in order: the full engagement, intermediate position, the onset of the breaking-through and fully breaking-through the CFRP workpiece. The double-point drill was new (Hole 1), the feed rate was set to 0.05 mm/rev and the cutting speed was 50 m/min under dry conditions. The recorded maximum thrust force was 76.03 N; while the first facets in the cutting lips were entirely engaged in Stage I. The maximum torque dropped to 0.2 Nm after the drill penetrated the workpiece, as seen in Stage III. The former situation is not unusual in drilling; whereas the latter situation is worthy of further discussion. The maximum thrust force was obtained when the cutting lips were entirely engaged with the CFRP composite material, due to maximization of the shear plane; however, surprisingly the maximum torque did not occur in the same position. This can be attributed to the undersized diameter in the machined hole due to fibre pull-out and the interplay delamination on the machined surface. This phenomenon resulted from the varied chip formation process which configured with the fibre orientation [22, 25]. As a result, the contact force at the margins of the drill body will be greatly drawn. When the drill advances to the last plies of the CFRP composite workpiece, the lack of support beneath can lead to deformation and bending, resulting in excessive contact between the margins of the hole and drill body. This can lead to a dramatic increase in torque, as indicated by the peak value in Stage III. After the drill bit penetrated the workpiece, thrust force and torque dropped to nil. 13

In contrast, Fig. 5(b) shows that the multi-facet drill drew 70% thrust force after the first facets were fully engaged in Stage I. The high thrust force benefited the drilling action by preventing the drill point from chattering and vibrating. In Stage II, when the 2nd facet of the cutting lips advanced into the workpiece and became entirely engaged, the additional contact further limited vibration and produced the highest thrust force. In Stage III, when the drill broke through the last plies in the workpiece, the thrust force was partially released; however, torque increased. This indicates that the decrease in feed force reduced damping and stabilization at the centre point; leading to a resumption of chatter and vibration, which increased contact at the periphery corners and margins in the drill body, as shown in Stages IV and V. The torque dropped entirely when the cutting lips of the 3rd facet passed through the workpiece, as shown in Stage VI.

Fig. 5. Cutting force measurement in the first hole using (a) double-point drill in Test 1 and (b) multi-facet drill in Test 7 Fig. 6 presents the maximum cutting force measured in the last hole drilling for different 14

experiment conditions. As shown in Fig. 6(a), the recorded thrust forces and torques at the last hole ranged from 72.58–96.34 N and 0.18–0.30 Nm respectively. Amongst these tests, only in Test 3 could the double-point drills (at a low cutting speed of 50-75 m/min) sustain drilling to the end of the task (140 holes) with less than VB300 in the flank face. Thrust force was measured at 94.39 N while cutting the first hole, which remained nearly the same (94.4 N) until the last hole (Hole 140); whereas the corresponding torque values were 0.334 Nm and 0.309 N-m at the first and last holes, respectively. Regardless of cutting speed (50-75 m/min), the high thrust force was shown to be produced by the high feed rate (0.10-0.15 mm/rev). However, when the feed rate was low (0.05 mm/rev), the high thrust force was produced under a low cutting speed (50 m/min). This phenomenon was due to the indentation and ploughing actions across the cutting lips, leading to an uneven distribution of the thrust force under a low cutting speed. The uneven distribution of the thrust force along the cutting lips resulted in unbalanced compression stress, which led to unbalanced strain at the interface between the coating layer and substrate material, particularly when the load was high. The highest axial compression stress was in the peripheral corner, due to the limited projected area along the cutting lips. The highest torque was simultaneously produced due to the greatest contact radius being located at the peripheral corners. This compression stress made it possible for internal stress-induced delamination to occur between the diamond-coated layer and carbide (WC-6Co) substrate. Similarly, when the cutting speed was increased to 75 m/min, cutting tool fails were observed in all tests. Although the friction heat resulted in a measuring temperature of ~135 °C for the thermal expansion in the diamond coating, creating conditions amenable to crack initiation, the recorded fluctuations of thrust force (~25 N) and torque (~0.2 Nm) produced by chattering and vibrations were more prominent and dominant to the fracturing of this coating. Finally, under the combined effects of the fluctuations of the compression stress and the induced thermal stress the coating layer underwent delamination and flaking. In contrast, when the feed rate was increased to 0.15 mm/rev, thrust force slightly 15

increased (~10.3%) but fluctuations in the thrust force (~27.2 N) and torque (~0.21 Nm) almost remained the same. This phenomenon is not unusual, since a high feed rate usually draws a high feed force and thereby produces the damping effect and reduces the elastic instability, which is produced by the unbalanced cutting forces acting on the cutting lips. In addition, the high feed rate reduces the contact time (down to 30%) between the interface of the cutting tool and the workpiece; thereby reducing the cutting temperature (84.7 °C) and the friction heat produced in the cutting zone. Fig. 6(b) shows the cutting force variation in different tests using a multi-facet drill. The maximum thrust force ranged from 77.1 to 128.28 N and torque values ranged from 0.172 to 0.329 Nm, respectively. Comparisons within the same feed rates, it was observed that increasing the cutting speed (50%) only slightly decrease the thrust force (~10%) and torque (~15%). However, under a set cutting speed of 75 m/min, a 300% increase in feed rate from 0.05 mm/rev to 0.15 mm/rev increased thrust forces by ~60.7% and torque by ~91.2%. Under a set feed rate and cutting speed, the double-point drill outperformed the multi-facet drill due to the geometrical effect. The double-point drill had a shorter cutting lip length, which reduces the size of the shear zone and thereby produces less thrust force. In contrast, the straight flutes found in the multi-facet drill provide a longer contact length, which produces greater thrust force.

16

Fig. 6. Maximum cutting forces at the first and last holes: (a) double-point drill and (b) multi-facet drill Fig. 7 presents the main effects plots of thrust force and torque with the operating parameters of cutting speeds (50, 75 m/min) and feed rates (0.05-0.15 mm/rev) for both double-point and multi-facet drills in all tests. The parametric combinations which produced high or low cutting forces are also both presented. Surprisingly, the combination best suited to minimizing thrust force was a feed rate of 0.05 mm/rev and lower cutting speed of 50 m/min, regardless of tool geometry. To minimize torque at a given feed rate (0.05 mm/rev), the double-point drill bit requires a high cutting speed (75 m/min); whereas the multi-facet drill bit requires a low cutting speed (50 m/min). Clearly feed rate dominates torque, which is a component of the cutting force exerted at the periphery of the drill. Friction force at the rake face and on the margins also proved to be an important factor because the maximum rake angle calculated in the double-point drill is expressed as the sum of the rake angle and velocity angle. In contrast, the multi-facet drills have straight flutes (α=0°), which means that 17

the rake angle is determined entirely by the velocity angle. Thus, the multi-facet drills exerted high friction force on the rake face, thereby increasing torque to a level beyond that of the double-point drills. Table 5 presents the ANOVA results illustrating the importance of operating parameters with regard to cutting forces under the two drill geometries. Feed rate significantly influenced the thrust force of the double-point drill with a percentage contribution rate (PCR) of 83.5%; whereas torque was strongly influenced by cutting speed as well as feed rate with PCRs of 5% and 94.4%, respectively. Conversely, the feed rate was the only significant factor affecting the thrust force and torque of the multi-facet drills.

Fig. 7. Main effect plots of thrust force and torque associated with double-point drills and multi-facet drills Table 5 ANOVA table of cutting force Double-point drill Source of variances

Thrust force

Torque

D.F.

F

P

PCR

F

P

PCR

Cutting speed

1

0.4

0.61

0.0%

37.7*

0.025

5.0%

Feed rate

2

12.02

0.08

83.5% 350.1*

0.003

94.4%

Residual

2

Sum of square variances

5

Multi-facet drill Source of variances Cutting speed

16.5%

0.7%

Thrust force

Torque

D.F.

F

P

PCR

F

P

PCR

1

4.8

0.16

2.2%

0.3

0.62

0.0% 18

Feed rate

2

Residual

2

Sum of square variances

5

82.22*

0.01

94.8% 15.87 2.9%

0.06

87.3% 12.7%

*Significant at the 5% level, F0.05 1,2 = 18.5, F0.05 2,2 = 19.

4.2 Tool wear and fracture modes Due to the cutting forces varying with the point angles and facets of the cutting edges in the oblique cutting, tool wear patterns were presented in many cases. Fig. 8 illustrates the progression of the tool wear or fractures of cutting edge in all tests involving double-point drills. Most of the tests resulted in progressive failure, such as chipping in peripheral corners and gross fractures along one or the other cutting lips. The tool wear in Test 3 (VB <200 µm) was far less obvious than in the other tests; due to a high feed rate retarding the tool wear along the edges because of the short contact time and a reduction in friction heat reducing the abrasion. When the cutting speed was increased to 75 m/min, tool failure occurred as follows: 0.05 mm/min (before the 60th hole); 0.10 mm/rev (before the 20th hole); and 0.15 mm/rev (before the 20th hole). These failures may have been initiated from an increase in friction heat in the cutting zone, which resulted in a softening of the polymeric matrix and a relaxation of the fibre arrangement. The combination of bending, shearing and compression may have interfered with the cutting mechanism in the shear zone, as indicated by the orientation of the relaxed fibres. In such cases, tool wear is dominated by cutting status such as ploughing and chattering.

19

Fig. 8. Progression of tool wear and fracturing on a double-point twist drill Fig. 9 shows the simulated results of the loading stress, strain and deflection at the cutting edge which reflected the linear increases of 24.1, 24.2, and 24.26 % when comparing the low feed rate (0.05 mm/rev) to that with two-fold increase of 0.15 mm/rev under the constant cutting speed (50 m/min). It was obvious that the increase of feed rate did not increase the thrust force with the same proportion. The low cutting speed (50 m/min) and low feed rate (0.05 mm/rev) in Test 1 resulted in the progressive tool breakage as shown in Fig. 10(a). SEM analysis shows that the fractography possessed fatigue characteristics, such as numerous striations and micro-cracks. This is in line with the calculated results, in which tensile stress in the coating inter-layer, shear force in the interface of the coating layer and tungsten carbide substrate, and the cutting forces produced in the shear zone combined to produce a force directed toward the peripheral corners. The resulting high stress produced gross fracturing in the peripheral corners, where wholesale dislodgement of the diamond coating and the removal of substrate material were observed. On the coating layer, the attrition effects on the delaminated diamond coating were observed, most likely produced by abrasion following the removal of the wholesale diamond coating layer. In contrast, Fig. 10(b) shows that the abrasion wear on the tool flank face in Test 3 was evenly distributed across the cutting lips within the chisel circle, which has a negative rake angle. An increase in torque 20

was found to be proportional to the radius near the peripheral corners, despite a slight decrease in thrust force. Thus, abrasion wear was shown to increase proportionally with cutting speed, leading to flaking/delamination of the coatings parallel to the direction of force. Moreover, interlayer delamination was introduced on the superficial surface of the diamond coating whilst micro-cracks were formed in the tungsten carbide substrate. Voids and cracks were encountered between the coating layer and substrate, possibly due to the products of the coalescence of the intergranular cracks in the plastic deformation and the relaxation in the grain boundary after cyclic loading in cutting actions.

Fig. 9. Simulation results of stress, strain and deflection at the cutting edge for Test 1 and Test 3 using double-point drills

21

Fig. 10. Micrograph of tool fractures along cutting edges when using a double-point drill: (a) Test 1 and (b) Test 3 at cessation of tests

Fig. 11 presents a micrograph of the premature fracturing along the damaged cutting lips in Test 6. Chipping and delamination of the diamond coated layer were prevalent near to the corner in the first facet of the cutting lips. Superficial delamination on the diamond coated layer was encountered after de-bonding occurred in the local coating layer, with loss of wholesale substrate material. Fracturing clearly occurred in the peripheral corners in the second facet where the cutting temperature was most likely the highest. Tool failure was initiated by the fluctuated loading stress in the coating, followed by delamination associated with a difference in the expansion coefficient between the diamond-coated layer and carbide substrate. During cutting actions, excessive contact was enhanced between the cutting edge and the machined surface due to the bounce back of the elastic fibre and resin matrix material. In addition, interference produced by the altered orientation in the layers, resulted in intermittent cutting and fluctuations of the cutting forces, which may have led to fatigue 22

failure in the coating layer. In such cases, imperfections in the sintered tungsten carbide matrix can be exacerbated by the movement of the tungsten carbide grains in the relaxed binder matrix, creating conditions amenable to the propagation of the voids/cracks. When the intergranular micro-cracks aggregate as clusters in the shear plane, fractures can occur under a relatively low load (Test 4: 70.15 N). The surface damage in the peripheral corners of the cutting lips revealed a corrugated texture and the loss of carbide substrate material after flaking in the coating layer, thereby proving the movement of the intergranular cracks. The high cycle loads in chattering and vibrations produced the relaxation of the matrix. Thus, minor plastic deformation in the rim zone under the coating layer in the substrate could develop the micro-cracks transverse to the load of the resultant force, leading to the loss of the carbide substrate material. This is an indication that the fracture mode was not brittle fracturing, but rather fatigue fracturing.

Fig. 11. Tool fractures of the double-point drill in the cessation of Test 6 Fig. 12 illustrates the progression of tool wear on three facets (flank face): ranging from 54 µm to 72.3 µm on the first facet, 55 µm to 62 µm on the second facet and 30.4 µm to 43 µm on the third facet. All drills had successfully completed 140 holes without the occurrence 23

of gross fracturing. In Test 12, the first facet presented the most pronounced flank wear of 72.3 µm with a corresponding point angle of 120º, which accounted for 70% of the thrust force (120 N). The third facet presented the least pronounced flank wear of 30.4 µm, due to little thrust force exerted to the third point angle of 6º. The damping imposed by high thrust force on the cutting lips of the first facet stabilized the cutting action and retarded chattering. In contrast, a reduction in thrust force and friction force reduced tool wear on the cutting lips of the first facet. Conversely, in Test 10 with a cutting speed of 75 m/min and a feed rate of 0.05 mm/rev, the lowest flank wear (54 µm) was observed on the first facet; the greatest wear (42.2 µm) was observed on the third facet. This implies that insufficient feed force reduces the damping effect and increases the likelihood of chatter. Intermittent contact with the cutting lips of the third facet produces ploughing and flapping, resulting in excessive tool wear.

24

Fig. 12. Progression of tool wear on a multi-facet drill: (a) first facet, (b) second facet and (c) third facet 25

Fig. 13 presents a micrograph of tool wear patterns along the cutting lip of the three facets of a multi-facet drill after cutting the final hole in Test 7. These results demonstrate that the first facet underwent the most abrasion due to the main load of the thrust force. The dislodgement and attrition of the diamond grain left recess sites on the wear site on the flank face. In addition, the abrasion wear folded over the rake face leading to the cutting edge rounding. On the second facet, scoring on the successive flank face appeared due to the increased cutting speed proportional to the radius. Micro-cracks were identified transverse to the load of the cutting force. The third facet underwent the least abrasion due to the reduced thrust force, but experienced the highest torque due to the greatest peripheral radius. The wear site in the flank face was therefore composed of abrasion wear and smearing. In the wear site, dislodgement of diamond grains was observed aside to the micro-cracks, with severe scoring on the successive flank face.

Fig. 13. Micrographs of the tool wear on the multi-facet drill from Test 7 26

In contrast, Fig. 14 presents wear patterns along the cutting lips of the multi-facet drill in Test 10 and Test 12 using extreme feed rates. The SEM micrographs illustrate that a low feed rate induced ploughing, which caused scoring along the cutting lips of the second facet in Test 10 (Fig. 14(a)). Specifically, micro-cracking initiated on the flank face parallel to the cutting edge implies that load was applied normally to the cutting edge. This caused chattering and flapping between the cutting lips and the surface of the machined hole. The highest torque in the third facet of the cutting lips was synchronous with the most pronounced chattering, which resulted in sporadic chipping of the diamond coating. This is a demonstration that high cutting speed (75 m/min) in conjunction with low feed rate (0.05 mm/rev) produced sporadic chattering/vibrations, which resulted in micro-chipping along the cutting lips on the third facet. In contrast, when drilling with the same cutting speed but a high feed rate (0.15 mm/rev), chipping on the third facet was retarded. This demonstrates that a high feed rate increases the damping effect and reduces the time required for machining, thereby reducing the contact time on the tool-workpiece interface. In contrast, low feed rates (0.05-0.1 mm/rev) resulted in severe scoring on the flank face, as shown in the micrographs captured in Test 7. This suggests that extending the contact between the tool and workpiece can result in ploughing, leading to a scoring defect. Test 12 (Fig. 14(b)) showed fewer signs of scoring on the second facet, but severe three body abrasion on the third facet. This demonstrates that damping induced by a high feed force converts intermittent cutting into continuous cutting, which also greatly reduces friction heat. Thus, after the diamond coating wears out, thermo-chemical induced erosion begins.

27

Fig. 14. Micrograph showing tool fractures along cutting edges when using a multi-facet drill: (a) Test 10 and (b) Test 12 at cessation of tests Fig. 15 presents Raman spectra and X-ray diffraction plots obtained from the flank face of the CVD diamond-coated double-point and multi-facet drills used in Test 3 and Test 12, respectively. A clear shift in the Raman peak was observed in both sampling areas (A and B) with the high side centred at 1336 cm-1, indicating the presence of the sp3 diamond structure or sp2 graphite structure [11]. A diamond sp3 structure generally presents a Raman peak at 1332 cm-1; however, the mixture of sp3 and sp2 presented features at 1332 cm-1 (diamond sp3 carbon, D band) and 1580 cm-1 (graphite sp2 carbon, G band) [28]. The conversion of sp3 to sp2 generally occurs under high temperatures (>700 ºC), producing a shift in the Raman spectra towards 1580 cm-1. However, under severe conditions, a shift in the Raman peak can be produced by residual compressive stress associated with a difference in the thermal expansion coefficient between the diamond-coated layer and tungsten carbide substrate [32]. 28

Despite none of the measuring temperatures recorded in these tests being higher than ~135 ºC, in fact the cutting temperature could be higher. Based on the law of one-dimensional steady state heat conduction, the brief estimation of the cutting temperature approached ~244 ºC, which could theoretically exert static stress of ~199.6 GPa between the interface of the CVD diamond and tungsten carbide. When the thermally induced static stress was imposed on the effects of the compressive residual stress, the observed Raman shift can be the result.

Fig. 15. Measurement results of Raman spectroscope and X-ray diffraction: (a) Sample A in Test 3 and (b) Sample B in Test 12 Table 6 presents the measurement results of the 2-theta, formula, intensity (counts) and deviations when compared to the bases in the new tool. Consequently, when the diamond structure (Sp3) was identified with the 2-theta value in the new tools, the correlated intensity (counts) would be the base for the comparisons with the used tools in Tests 3 and 12. In Table 6, a diamond structure with the 2-theta value of 43.92° in the used tools (Test 3) having an 29

intensity of 697 counts reflected an increase of 2.95% when compared to that of 677 counts in the new tool. This minute increase (< 5%) of the intensity showed the diamond structure in the used tool (Test 3) was not significantly varied. Whereas the similar 2-theta value of 44.16° in the used tools (Test 12) showing a reduced intensity of 456 counts, presented a significant decrease of 32.64%, meaning that the diamond structure reverted into a graphite structure. It was manifest that when the lattice structures were distorted by the internal stress, the diffraction of the X-ray was reduced due to the interferences of the absorption, reflection and scattering effects produced by the inconsistent lattice structures. As a result, the minute deformation at the lattice structure could disturb the X-ray diffraction by altering the peak values of the intensity and expanding the spans, despite the identified 2-theta remaining the same. In general, Raman spectrum shift can distinguish the allotropes of the carbon elements from graphite to diamond. However, the Raman shift driven by the residual stress needed to be identified by X-ray diffraction, since the distortion of the lattice structures disturbed the spectrum shift by reflecting on the decay of the intensity and the increase of the spans for the diamond structure and each identified formula. Table 6 Measurement results of the X-ray diffraction New tool

2θ, °

Formula

Intensity, Counts

Base

35.58

WC

7486

-

43.92

Diamond

677

-

48.23

WC

6301

-

Test 3

Deviation (%) 35.66

WC

7303

-2.45

43.92

Diamond

697

2.95

48.31

WC

6527

3.55

Test 12

Deviation (%) 35.82

WC

6920

-7.56

44.16

C

456

-32.64

48.47

WC

5946

-5.63

5. Conclusions In this study, double-point and multi-facet drills were utilized in the drilling of woven CFRP materials. The aim was to elucidate the correlation between tool wear and fracturing in diamond coatings subjected to cutting forces under two advanced geometrical designs at the drill points. The experimental set-up was based on standards of industrial applications. The results were as follows: 30

 Under the same cutting forces, multi-facet drills undergo limited abrasion wear due to the distribution of cutting forces along longer cutting lips (three facets) while exerting a damping effect capable of suppressing radial vibration. In contrast, double-point drills provide a shorter tool-workpiece contact length, which gives effective cutting but produces high abrasion.  Within the same cutting condition, tool wear does not appear to be proportional to the cutting force. The double-point drills produced less cutting force but were susceptible to premature tool fracture due to stress concentration at the peripheral corners. The multi-facet drill bits provided greater resistance to tool wear, due to the redistributed forces in the three facets along the cutting lips.  The failure modes of double-point drill bits under a low feed rate (0.05 mm/rev) and low cutting speed (50 mm/min) were fatigue fractures. At high feed rates (0.15 mm/rev), tool wear at low and high cutting speeds progressed through abrasion, chipping, and delamination of the CVD diamond-coated layer as well as gross fracturing.  The failure modes of multi-facet drill bits under a high feed rate (0.15 mm/rev) and high cutting speed (75 mm/min) were a combination of progressive abrasion, scoring and severe three body abrasion on the first, second and third facet of the flank faces. At a low feed rate (0.05 mm/rev), intermittent cutting induced chipping which eventually led to severe abrasive wear.  Fatigue fractures in the double-point drill bits were produced by chatter and vibration, resulting in the relaxation of the carbide matrix and leading to gross fracture. Similarly, fracturing in the multi-facet drills was due to vibrations, which manifested as micro-cracking and chipping in the flank face.

31

ACKNOWLEDGEMENTS The authors would like to thank the Aerospace Industrial Development Corporation (AIDC) for the supply of the workpiece material, tooling and funding (NTUST 104AL707) for this work. In addition, they also appreciated the support from the Ministry of Science and Technology (MOST) for providing funds (106-2221-E-011 -073) throughout this research.

REFERENCES [1]

M. Ramulu, T. Branson, D. Kim, A study on the drilling of composite and titanium stacks, Compos. Struct. 54 (2001) 67-77.

[2]

J. Xu, A. Mkaddem, M. El Mansori, Recent advances in drilling hybrid FRP/Ti composite: A state-of-the-art review, Compos. Struct. 135 (2016) 316-338.

[3]

D. Kim, M. Ramulu, Drilling process optimization for graphite/bismaleimide–titanium alloy stacks, Compos. Struct. 63 (2004) 101-114.

[4]

J.M. Park, D.J. Kwon, Z.J. Wang, G.Y. Gu, K.L. DeVries, A new strategy of carbon fiber reinforced plastic drilling evaluation using thermal measurement, J. Compos. Mater. 47 (2012) 2005-2011.

[5]

Y. Karpat, O. Bahtiyar, Comparative Analysis of PCD Drill bit Designs During Drilling of CFRP Laminates, Procedia CIRP 31 (2015) 316-321.

[6]

E. Brinksmeier, S. Fangmann, R. Rentsch, Drilling of composites and resulting surface integrity, CIRP Ann. – Manuf. Technol. 60 (2011) 57-60.

[7]

N. Cadorin, R. Zitoune, Wear signature on hole defects as a function of cutting tool material for drilling 3D interlock composite, Wear 332–333 (2015) 742-751.

[8]

M.S. Raghuveer, S.N. Yoganand, K. Jagannadham, R.L. Lemaster, J. Bailey, Improved CVD diamond coatings on WC–Co tool substrates, Wear 253 (2002) 1194-1206.

[9]

S. Rawat, H. Attia, Characterization of the dry high speed drilling process of woven composites using Machinability Maps approach, CIRP Ann. – Manuf. Technol. 58 (2009) 105-108.

32

[10] E. Salgueiredo, C.S. Abreu, M. Amaral, F.J. Oliveira, J.R. Gomes, R.F. Silva, Self-mated tribological systems based on multilayer micro/nanocrystalline CVD diamond coatings, Wear 303 (2013) 225-234. [11] R. Dumpala, N. Kumar, C.R. Kumaran, S. Dash, B. Ramamoorthy, M.S. Ramachandra Rao, Adhesion characteristics of nano- and micro-crystalline diamond coatings: Raman stress mapping of the scratch tracks, Diamond Relat. Mater. 44 (2014) 71-77. [12] M. Murakawa, S. Takeuchi, Mechanical applications of thin and thick diamond films, Surf. Coat. Technol. 49 (1-3) (1991) 359-364. [13] S. Söderberg, A. Gerendas, M. Sjöstrand, Factors influencing the adhesion of diamond coatings on cutting tools, Vacuum 41 (1990) 1317-1321. [14] D. Iliescu, D. Gehin, M.E. Gutierrez, F. Girot, Modeling and tool wear in drilling of CFRP, Int. J. Mach. Tools Manuf. 50 (2010) 204-213. [15] A. Faraz, D. Biermann, K. Weinert, Cutting edge rounding: An innovative tool wear criterion in drilling CFRP composite laminates, Int. J. Mach. Tools Manuf. 49 (2009) 1185-1196. [16] S. Gaugel, P. Sripathy, A. Haeger, D. Meinhard, T. Bernthaler, F. Lissek, M. Kaufeld, V. Knoblauch, G. Schneider, A comparative study on tool wear and laminate damage in drilling of carbon-fiber reinforced polymers (CFRP), Compos. Struct.155 (2016) 173-183. [17] X. Wang, P.Y. Kwon, C. Sturtevant, D. Kim, J. Lantrip, Tool wear of coated drills in drilling CFRP, J. Manuf. Processes 15 (2013) 127-135. [18] G. Poulachon, J. Outeiro, C. Ramirez, V. André, G. Abrivard, Hole Surface Topography and Tool Wear in CFRP Drilling, Procedia CIRP 45 (2016) 35-38. [19] Y. Karpat, B. Değer, O. Bahtiyar, Drilling thick fabric woven CFRP laminates with double point angle drills, J. Mater. Process. Technol. 212 (2012) 2117-2127. [20] C. Ramirez, G. Poulachon, F. Rossi, R. M'Saoubi, Tool Wear Monitoring and Hole Surface Quality During CFRP Drilling, Procedia CIRP 13 (2014) 163-168. [21] W.C. Chen, Some experimental investigations in the drilling of carbon fiber-reinforced plastic (CFRP) composite laminates, Int. J. Mach. Tools Manuf. 37 (1997) 1097-1108. [22] D.H. Wang, M. Ramulu, D. Arola, Orthogonal cutting mechanisms of graphite/epoxy composite. Part I: unidirectional laminate, Int. J. Mach. Tools Manuf. 35 (1995) 1623-1638. [23] D.H. Wang, M. Ramulu, D. Arola, Orthogonal cutting mechanisms of graphite/epoxy composite. Part II: multi-directional laminate, Int. J. Mach. Tools Manuf. 35 (1995) 1639-1648. [24] N. Bhatnagar, N. Ramakrishnan, N.K. Naik, R. Komanduri, On the machining of fiber reinforced plastic (FRP) composite laminates, Int. J. Mach. Tools Manuf. 35 (1995) 33

701-716. [25] X.M. Wang, L.C. Zhang, An experimental investigation into the orthogonal cutting of unidirectional fibre reinforced plastics, Int. J. Mach. Tools Manuf. 43 (2003) 1015-1022. [26] O. Isbilir, E. Ghassemieh, Delamination and wear in drilling of carbon-fiber reinforced plastic composites using multilayer TiAlN/TiN PVD-coated tungsten carbide tools, J. Reinf. Plast. Compos. 31 (2012) 717-727. [27] C. Kuo, S. Soo, D. Aspinwall, S. Bradley, W. Thomas, R. M'Saoubi, D. Pearson, W. Leahy, Tool wear and hole quality when single-shot drilling of metallic-composite stacks with diamond-coated tools, P. I. Mech. Eng. B-J. Eng. 228 (10) (2014) 1314-1322. [28] V. Astakhov, Geometry of Single-point Turning Tools and Drill bits, Springer, London East Lansing, 2010. [29] A.R. Watson, Geometry of drill bit elements, Int. J. Mach. Tool Des. Res. 25 (1985) 209-227. [30] E.J.A. Armarego, C.Y. Cheng, Drilling with flat rake face and conventional twist drill bits—I. Theoretical investigation, Int. J. Mach. Tool Des. Res. 12 (1972) 17-35. [31] T.C.S. Vandevelde, K. Vandierendonck, M. Van Stappen, W. Du Mong, P. Perremans, Cutting applications of DLC, hard carbon and diamond films1, Surf. Coat. Technol. 113 (1999) 80-85. [32] J. Gunnars, A. Alahelisten, Thermal stresses in diamond coatings and their influence on coating wear and failure, Surf. Coat. Technol. 80 (3) (1996) 303-312.

Highlights 

Tool wear is not proportional to cutting force but varies with point design.



The longer cutting lip of multi-facet drill bits provides damping.



The shorter cutting lip of double-point drills failed at the peripheral corners.



Multi-facet drill failed due to progressive abrasion wear, scoring, and chipping.



Double-point drill failure is due to delamination and premature tool fracturing.

34