Cryogenic PCBN turning of ceramic (Si 3N 4)

Cryogenic PCBN turning of ceramic (Si 3N 4)

WEAR Wear 195 (1996) 1-6 Cryogenic PCBN turning of ceramic (S&N,) Z.Y. Wang, K.P. Rajurkar, M. Murugappan Industrial and Management Systems Engineeri...

636KB Sizes 0 Downloads 14 Views

WEAR Wear 195 (1996) 1-6

Cryogenic PCBN turning of ceramic (S&N,) Z.Y. Wang, K.P. Rajurkar, M. Murugappan Industrial and Management Systems Engineering Department,

Universiry of Nebraska-Lincoln,

Lincoln, NE 68588-0518,


Received 3 November 1994; accepted 31 March 1995


This paper presents a technique for machining advanced ceramics with liquid nitrogen (LN) cooled polycrystalline cubic boron nitride (PCBN) tool. A LN circulation system has been designed to control the cutting tool temperature. The tool wear of PCBNSO during the machining of the reaction bonded silicon nitride (Si,N4) reduced significantly with LN cooling. The surface roughness due to LN cooling assisted cutting was about six times better than normal cutting. Keywords:


turning; Ceramics;

Boron nitride; Tool wear

1. Introduction Advanced structural ceramics, such as silicon nitride, are attractive for many applications due to their high strength at elevated temperatures, resistance to chemical degradation, wear resistance, and low density. Despite these advantages, the cost of machining can be as high as 90% of the total cost of some high precision components; and damage caused during machining can be detrimental to the performance of the component and can result in premature failure [ 11. A wide variety of methods have been proposed and tested for machining of advanced ceramic components [ 2,3]. Some of these methods such as grinding [ 41, honing, lapping, polishing [ 51 and ultrasonic machining [ 61 are extensively used for machining ceramics. Others that have been proposed, such as electrical discharge machining [ 71, laser beam cutting [ 81, electron beam and ion beam cutting [ 9, lo] and microwave cutting [ 111 require additional research and development. Turning and milling methods, although extensively used in metal machining, have only limited application to advanced ceramics in their fully sintered state because of high rates of tool wear and severe workpiece surface damage. Plasma torch heating of workpiece during turning has been reported to be effective for some engineering ceramics [ 121. This method involves heating the workpiece material to temperatures as high as 1000 “C when it enters the cutting zone and is machined using polycrystalline diamond compact or PCBN cutting tool. Although, the tool wear was reduced by a factor of eight, in turning silicon nitride by plasma torch heating, from what it would have been without the plasma torch heating, it was still unacceptably high. Materials sen0043-1648/96/$15.00 0 1996 Else&x SSDIOO43-1648(95)06645-4

Science S.A. All rights reserved

sitive to thermal shock, for example alumina and zirconia, did not show improved machinability by plasma torch heating. The hot-machining of mullite and silicon nitride investigations have been reported by Uehara and Takeshita [ 131 and Akasawa et al. [ 141. Koenig et al. [ 151 have also reported a method of laser-assisted turning of hot-pressed silicon nitride with PCBN cutting tools. The thermally assisted turning does not lend itself to large-volume production, because of short tool life and poor surface finish. The extremely high tool wear rate and surface damage of ceramics produced during machining are mainly due to the high cutting temperature and the extreme hardness of work materials. Hot-machining softens the workpiece and hence reduces its hardness, thereby making the machining relatively easy. However, the hot-machining inevitably increases the temperature in the cutting zone, and as a result, the temperature-dependent wear on the tool increases rapidly. In a different approach [ 161, the temperature in the cutting zone is reduced to a lower range by cryogenic cooling in turning of stainless steel with diamond tool, the experiments show that the temperature-dependent wear reduces significantly. If the temperature on the tool can be controlled in turning of advanced ceramics, it is also possible to prolong the tool life to a reasonable range. This paper proposes liquid nitrogen (LN) cooling to control the temperature in the cutting zone so that the temperature dependent tool wear can be reduced significantly. In this method the workpiece is not subjected to any preheating and hence the properties of the workpiece material are not altered before it enters the cutting zone. The selection of the work-

Z.Y. Wang etal. /Wear 195 (1996) 14


Table 1 Properties of reaction bonded silicon nitride (RBSN) Properties




Density Grain size Tensile strength Compressive strength

2.4 g cme3 75 pm 18 kpsi 105 kpsi

Fracture toughness Hardness Young’s modulus Thermal conductivity

3.4 MPa ml’* 60 HRC 172.22 GPa 13Wm-‘K-‘s-’

piece material and the tool material is described in Section 2. The experimental results of PCBNSO are presented and discussed in Section 3 and Section 4, respectively. Section 5 describes the chip formation, tool failure and machined quality. Section 6 summarizes the work.

2. Work material properties and tool material selection The work material selected for the experiments is reaction bonded silicon nitride (Si,N,) , which is an ideal material for direct contact with molten salts, aluminum, zinc, sodium, lithium, and potassium. It has been widely used for pumps, pipes, shafts, protection tubes and valves. The reaction bonded silicon nitride (RBSN) offers many engineering properties that are not found in other ceramics or refractory materials, such as high strength even at 1400 “C, low thermal expansion, excellent shock resistance and chemical inertness. Some of the important properties, as provided by the manufacturer, of the workpiece used in the experiments are listed in Table 1. In machining advanced ceramics, it is also crucial to select the right tool material which is thermally stable and which has superior mechanical properties to those of the workpiece materials. For the present study, the tool material was selected according to an empirical relationship which relates wear resistance W, to fracture toughness KIc, Young’s modulus E and hardness H values [ 171 Wa = @&sE-0.8Hi.43


3 ____________________________________~~~ Zi




Fig. 1. Estimated tool wear resistance.

The wear resistance of RBSN ( Si3N4) and five other different tool materials was calculated and shown in Fig. 1. The materials were: white ceramic (A&O,), tungsten carbide (WC), black ceramic ( A1,03 + TIC), polycrystalline cubic boron nitride (PCBN), and polycrystalline diamond (PCD) . Fig. 1 shows that PCD has the highest hardness and the best wear resistance. However, at high temperature it is susceptible to phase transformation and transforms into graphite. PCBN’s wear resistance is second only to PCD. Most importantly, PCBN is thermally stable; there is little tendency for the PCBN to revert back to hexagonal boron nitride, even at temperatures as high as 1400 “C. If the cutting temperature is controlled in a certain range, the temperature-dependent tool wear reduces. Therefore, PCBN seems to have the potential to machine advanced ceramics.

3. Experimental results for PCBNSO Experiments were carried out to investigate the machinability of RBSN (S&N,), with PCBN 50. The parameters for fine cutting were: speed= 840 rev min-‘; feed rate=O.l mmrev-’ (0.004 in rev-‘); and depth of cut =OS mm (0.0195 in); rake angle yo= -6”; inclination angle A,= -6”; clearance angle %=6”; and nose radius = 0.8 mm (0.3 1 in). The machine used in the experiments was Clausing Colchester 15 lathe. The workpiece was a cylindrical bar of dimension 80 X 80 mm. A LN circulation system (shown in Fig. 2(a)) was designed to keep the tool temperatures at a lower range. An unsheathed fine gage K type thermocouple of wire diameter 0.032 in was mounted on the rake face at a distance of 2 X 1 mm from the cutting point (Fig. 2(b)). The thermocouple along with a miniature digital panel and connecting wires constitute the temperature measurement system. After each cut, the tool was studied under a microscope. Both, the flank wear depth (VB) and length were measured and photographed along with the ruler (Fig. 6). The procedure was repeated five times for the tool used with LN cooling and two times for the tool used without LN. The points in Fig. 3 corresponds to the wear measured after each cut. In Fig. 3, it can be seen that the tool wear of PCBNSO used without LN cooling was much higher than that of PCBNSO used with LN cooling. The tool wear (both VB and the length of wear along the cutting edge) increased sharply with the cutting distance for the tool used without LN cooling and after 40 mm of cylindrical turning, the tool wear VB was

Z.Y. Wang et al. /Wear

4. Temperatures Fig. 2. Schematic illustration of cryogenic treated turning tool. PCBN 50, WEAR VS. LENGTH OF CUT















Fig. 3. Effect of cryogenic treatment on tool wear workpiece. RBSN ( Si3N4), tool: PCBN 50, speed= 840 rev min-‘, feed rate=O.l mm rev-‘, depth of cut = 0.5 mm; y0 = - 6”; A,= - 6”; cu,= 6”; nose radius = 0.8 mm. PCBN50 : TEMP VS TIME COI.PbJ3lsoN OF EM= WITH & WllHOUl LN

300 ,


The PCBN tool in hot-machining of silicon nitride has been reported to have lasted only for 1 min (Uehara, 1986). The cutting conditions used in that test were: speed = 21 m min- ‘; depth of cut =OS mm; feed rate=O.O44 mm rev-‘. The cylindrical length of cut for which the tool could be used was 21 m. The cutting conditions of the cryogenic PCBN turning used were: speed = 201.66 m min- ‘; depth of cut = 0.5 mm; feedrate=O.l nunrev-‘; the workpiece diameter = 80 mm. The cylindrical length of cut was 422.4 m. Although a direct comparison cannot be made because of the possible differences in the properties of the workpieces, the tool life is much higher in cryogenic PCBN turning than in the hot-machining.

D~ltal Monltor


195 (1996) 14

in the cutting process

The cutting temperatures for turning RBSN (S&N,) with and without LN cooling are shown in Fig. 4. Since the thermocouple was mounted at a distance of 2 x 1 mm from the cutting point, the measured data do not represent the actual cutting temperature of the tool. Nevertheless, it still gives a clear picture of the effect of LN cooling on the cutting tool. Fig. 4 shows the measured cutting temperature for (i) the tool used with LN cooling (five cuts, total length of cut 160 mm), (ii) the tool used without LN (single cut for a length of 40 mm), and (iii) another tool used without LN (two cuts, total length of cut 40 mm). In all the experiments, the tool temperature increased with the cutting time when LN cooling was not applied. When the tool was used for the second cut, the temperature increased at a much higher rate than in the first cut because of the increased tool wear. When the turning process was subjected to LN cooling, the measured temperature ( - 160 to - 170 “C) was almost constant. Experiments show that the temperature has a significant influence on the tool wear. When the cutting process is not subjected to LN cooling, the PCBNSO tool life is too short to be used for turning of RBSN (S&N,).


5. Chip formation, tool failure and machined quality The formation of the chips involve shearing of the work material in the region of a plane extending from the tool edge -100 ._...___~~_____.____.______.~~~~_~~~~~~.~__.~~~~~~~~~~~~~-to the position where the upper surface of the chip leaves the , LNCOOLLD workpiece (OA in Fig. 5 (a) ) . A very large amount of strain -200 takes place in this region in a very short interval of time. As 5 10 15 20 25 0 lU&lNsEC a result, the RBSN ( Si3N4) was fragmented in the primary deformation zone, and chips come out in the form of fine +LNcooLED.ALLFlvscw~ NoLNcooLr4G.1slcur +NOC-.2M)EvT + NOCOOLlNQ.FlR~CUT powders. The diameters of the powder chips varied from a couple of microns to twenty or thirty microns (Fig. 5(b)). Fig. 4. Measured temperatures in the test. These chips were very different than continuous and discontinuous chips obtained during the most of the turning opera2.87 mm which is extremely high for a normal cutting opertions. The normal turning chips may adopt many shapesations and the tool was rendered unsuitable for further cutting. straight, tangled or different types of helix and often they However for the tool used with LN cooling the VB was only have considerable strength and cause crater wear, crack 0.39 mm after cutting for a length of 160 mm.

Z.Y. Wang et ul. /Wear

Fig. 5. Chip formation in turning of RBSN (S&N,) : (a) schematic representation; (b) scanning electron mmicroscopy picture of RSBN chips (orig-

inal magnification, 150X ).

195 (1996) l-6

zones along the cutting edge and on the flank face. Cracks were observed on the worn flank face and along the boundaries of the chipping. There was no plastic deformation observed on the cutting edge. The wear and the fracture were mainly on the Bank face. The wear pattern of PCBNSO used to turn RBSN (S&N,) for a length of 60 mm with LN cooling is shown in Fig. 6(b) . No chipping or fracture was observed on the rake face. Similar to the tool used without LN cooling, the wear was mainly on the flank face. There was no crater wear, surface damage or plastic deformation on the rake face. The reason is not clear yet why LN cooling does not cause chipping or fracture in the tool. Some tiny cracks were observed on the flank face. Fig. 7(a) is a scanning electron micrograph of PCBNSO used to turn RBSN without LN cooling in the experiments. It shows a crack developed at the boundary of the tool chipping which has occurred on the cutting edge. Fig. 7(b) shows a crack developed parallel to the cutting edge on the flank face. Investigating the development of cracks and chippings with scanning electron microscope, it was found that they all originated from the worn flank face or from the worn cutting edge of the tool. The severe the wear on the surface, the more the cracks and chippings on the tool. During the process of turning RBSN ( Si3N4) with LN cooling, the tool wear on the cutting edge and flank face was very slow as compared to the tool wear without LN cooling. Therefore, the crack growth became much slower and resulted in less chipping on the cutting edge and the flank face of the tool when LN cooling was used. The surface roughness of the workpiece was measured after turning RBSN (S&N,) with a PCBNSO tool. Fig. 8

Fig. 6. Wear patterns of PCBNSO in turning of ceramic (Si3N4) (each division in the ruler corresponds to 0.4 mm) : (a) flank face (without LN) ; (b) flank face (with LN).

development or other kinds of surface damage on the rake face of the tool. The analysis of the tool wear revealed that the contact between the tool and the workpiece was limited to the cutting edge. The work material was fragmented into powders in the primary deformation zone and therefore, no crater wear was observed on the tool in any of the tests. Fig. 6(a) shows PCBNSO insert used to turn RBSN ( Si,N4) for a length of 20 mm without LN cooling. The wear on the flank face was very high. There were some chipping

Fig. 7. Crack development on the tool: (a) rake tace (original magnification, 300 x ) ; (b) flank face (original magnification, 1000 X ).

Z. Y. Wang et al. /Wear

195 (1996) I-6


James McManis, Supervisor of Engineering Machine Shop, for providing the personnel and equipment to conduct the tests; Professor W.N. Weins and Dr. Yimin Xu, Department of Mechanical Engineering, for their technical support; Prof. Brian Robertson and Todd, Scanning Electron Microscopy Lab for their help in measuring and analyzing the tool wear.

References [ 11 S. Jahanmir,



120 mm


Length of Cut Fig. 8. Surface roughness

of machined workpiece.

shows that when the machining process was subjected to LN cooling, the R, value of the workpiece surface after being machined for a length of 160 mm is 3.2 p,m. However, when the process was not subjected to LN cooling, the R, value of the workpiece surface after being machined for just 40 mm is 20 p.m. The large difference was due to the variation in the tool wear due to the varied machining conditions. Since the tool wear increases with the length of cut and the surface roughness increases with tool wear, it could be deduced that if the surface roughness was measured after machining for 40 mm with LN cooling R, value would have been much less than 3.2 pm. The surface roughness of the hot-machined workpiece has been reported to be in the range lo-20 p,m [ 131. Although a direct comparison is not possible because of the difference in the machining conditions, workpiece properties, etc., the surface roughness obtained by the cryogenic PCBN turning seems to be much better. The possible changes in the properties of the workpiece induced by the cryogenic PCBN turning are yet to be investigated. 6. Conclusions The temperature in the cutting zone has a significant influence on tool wear in the machining of ceramic with a PCBN tool. When the process of turning RBSN with a PCBNSO tool was not subjected to LN cooling, the tool life was too short due to extensive tool wear. When the process was subjected to LN cooling, the tool wear of PCBNSO was less and it increased at a lower rate in the successive cuts. The wear on the tool was mainly attrition wear and abrasive wear. The surface roughness of the workpiece machined with LN cooling was much better than the surface roughness of the workpiece machined without LN cooling. Acknowledgements Support from the Nebraska Research Initiative Fund is gratefully acknowledged. The authors are also thankful to

L.K. Ives, A.W. Ruff and M.B. Peterson, Ceramic Machining: Assessment of Current Practice and Research Needs in the United States, NIST Spec. Publ. 834, 1992. [2] R. Snoeys, F. Staelens, W. Dekeyser, Ann. CIRP, 35 (2) (1986) 467480. 131 W. Koenig, L. Cronjaeger, G. Spur, H.K. Toenshoff, M. Vigneau and W.J. Zdeblick, Ann. CIRP, 39 (2) (1990) 673681. [4] T. Nakagawa, K. Suzuki and T. Uematsu, Ann. CIRP, 35 (1) ( (1986) 205-210. [5] K. Subramanian, Ceram. Bull., 67 (6) (1988) 1026-1029. [6] H. Kamoun, M. Houbt, D. Kremer, B. Lecoco and G. Coffignal, Modelling the material removal in stationary mode for ultrasonic contour machining, Proc. ASME Winter Annu. Meet., New Orleans, LA, 1993, pp. 759-770. [7] W. Koenig, D.F. Dauw, G. Levy and U. Panten, Ann. CIRP, 37 ( 1) (1988) 623-631. [81 S.M. Copley, Handbook of High Speed Machining Technology, Chapman and Hall, New York, 1985, pp. 1771-1775. [9] M.E. Mochel, J.A. Eades, M. Metgger, J.I. Meyer and J.M. Mochel, Appl. Phys. Len., 44 (1984) 502-504. [ 101 B. Daudin and P.N. Martin, Mater. Sci. Eng. A, 125 ( 1989) 6366. [ 111 N. Taniguchi, K. Shimomura and T. Miyazaki, Ann. CIRP, 22 (1) (1972) 43-50. [ 121 T. Kitagawa and K. Maekawa, Wear, 239 (1990) 251-267. [ 131 K. Uehara and H. Takeshita, Ann. CIRP. 35 ( 1) (1986) 55-58. [ 141 T. Akasawa, H. Takeshita and K. Uehara, Ann. CARP, 36 (1) (1987) 3740. [ 151 W. Koenig and A. Wagemann, Machining of Ceramic Components: Process - Technological Potentials, NIST Spec. Publ. 847, 1993, pp. 3-20. [ 161 C. Evans, Ann. CIRP, 40 (1) (1991) 571-575. [17] A.G. Evans and D.B. Marshall, Wear mechanisms in ceramics, in Fundamentals of Fraction and Wear of Materials, Proc. ASM Mater. Sci. Semin., Pittsburgh, PA, 1980, pp. 439452.

Biographies K.P. Rajurkar: is a Mohr Professor of Engineering and Director of Nontraditional Manufacturing Research Center. He earned his Ph.D. in manufacturing and industrial engineering in 1982, from Michigan Technological University. His research interests includes theoretical and experimental investigations of process mechanisms, surface integrity, monitoring and control, and expert systems of advanced manufacturing processes such as electrodischarge, electrochemical, abrasive flow machining, ultrasonic machining, and agile manufacturing processes. Z.Y. Wang: received his Ph.D. degree in mechanical engineering from the Harbin Institute of Technology in 1991.


Z.Y. Wang et al. /Wear

Afterwards, he worked two years on ultra precision manufacturing at Aachen, Germany, supported by the Alexander von Humboldt foundation. Later on, he joined the faculty of the State University of New York and then the University of Nebraska Lincoln as a research scientist and a research associate respectively. His expertise is on agile manufacturing, ultra precision machining, ultrasonic machining, and traditional manufacturing.

19.5 (1996) 1-6

M. Murugappan: received the B .S degree in mechanical engineering from A.C. College of Engineering and Technology, TN, India in 1991. Then, he worked as a graduate engineer trainee at Tamil Nadu Newsprint and Papers limited, TN, India for one and half years. He is currently a graduate student at Industrial Engineering Department and a graduate research assistant at Nontraditional Manufacturing Research Center, University of Nebraska, Lincoln.