Handbook of Advanced Ceramics Machining Episode 3 pps

microholes, which are considered to be pores in HIPSN itself or fracturetraces of fragile portions inside HIPSN, becomes very smooth.Thus, a smooth ductile-mode ground surface can also b

microholes, which are considered to be pores in HIPSN itself or fracturetraces of fragile portions inside HIPSN, becomes very smooth.

Thus, a smooth ductile-mode ground surface can also be obtained withplunge grinding of HIPSN ceramic with the coarse #140-mesh wheel

2.3.1 Ultra-smoothness Grinding Method

As stated above, ductile-mode grinding of fine ceramics is possible withplunge grinding with the coarse #140-mesh grain-size diamond wheel The3D surface roughness in ductile-mode plunge grinding, however, is limited

to about 200 nm (Rmax) in the above-mentioned experiments because of theformation of grinding grooves The 3D surface roughness of 200 nm (Rmax)

is much better than the well-known data obtained in the usual grindingoperation However, the ultrasmooth roughness below 10 nm (Rmax), which

is almost the same as that after lapping, could not be obtained in plungegrinding To obtain ultra-smoothness roughness with the coarse-grain-sizewheel, it is necessary to diminish and remove the grinding grooves

Now, the ultra-smoothness grinding method has been newly devised [5,6].The schematic of the new method is shown in Figure 2.21 In the method, theworkpiece is fed simultaneously toward the directions normal and parallel tothe grinding direction The method is different from the usual traversesurface grinding in which the crossfeed normal to the grinding direction is

Orientation (Front)

U: 98 nm L: − 91 nm WYKO

50 3

− 10.00

− 5.00 0 5.00 10.00

98 51

q ⬚

RA: 45.8 nm P—V: 238 nm

RMS: 1.85 nm

RA: 1.47 nm P—V: 9.71 nm

FIGURE 2.19

A WYKO 3D image and 2D profile of the HIPSN surface ground at n w ¼ 0.05 mm=sec with the

#140-mesh wheel [SD140Q50M, V g ¼ 20 m=sec, t t ¼ 5 mm, Soluble (1=50)].

Ductile-Mode Ultra-Smoothness Grinding of Fine Ceramics 47

done after grinding the whole width of the workpiece in a pass or stroke ofthe table In the method, the crossfeed (vw)nnormal to the grinding direction

is faster than the plunge feed (vw)pparallel to the grinding direction Theexperiments are carried out with the NC grinding machine as shown sche-matically and photographically in Figure 2.22 and Figure 2.23, respectively

400 (nm)

0

0 ( µ m)

30 40

250 (nm)

RMS

Rmax

40 50

0 10

FIGURE 2.21

Ultra-smoothness grinding method.

In the experiments, (vw)nand (vw)pare set by determining the resultant feed(vw)c ¼ {(vw)nþ (vw)p}1, crossfeed width Bswand feed width f The machine

is reconstructed with a conventional NC equipment, which has an accuracy

of 1 mm for each movement of the X, Y, and Z directions The experimentalconditions are summarized in Table 2.2

2.3.2 Ultra-smoothness Grinding Results

Figure 2.24 shows the microscopic photographs of the HPSC surface ground

at (vw)c ¼ 3.33 mm=sec with the #140-mesh wheel From the figure, it is

1

1

: NC grinding machine : Motor

: Spindle : CNC controller : Inverter : Operation board : Coolant supply system : Nozzle

2 2

3 3

4

4 Z

X Y

5 5

6 6

7 7

8 8

Coolant

supply system

CNC controller (YASNAC MX1)

Operation board of Electromagnetic chuck

FIGURE 2.23

Appearance of NC grinding machine.

Ductile-Mode Ultra-Smoothness Grinding of Fine Ceramics 49

obvious that no grinding cracks and no grinding grooves are found over theobserved workpiece area.

Figure 2.25 shows a WYKO 3D image of 256 mm2 of HPSC surfaceground at (vw)c ¼ 3.33 mm=sec From the figure, it is found that theworkpiece surface is not formed by continuous regular grooves as shown

in Figure 2.5 but by some discontinuous short grinding grooves The 3Dsurface roughness is about 26 nm (P-V), 3.7 nm (RMS), and 3 nm (Ra),corresponding to near ultra-smoothness surface roughness The height ofthe grinding groove is considered to be below about 26 nm (P-V), which ismuch lower than that in the plunge ductile-mode grinding as shown inFigure 2.5 The pitch of the grinding groove is also not regular as observed

FIGURE 2.24

Microscopic photograph of the HPSC surface ground at (y w ) c ¼ 3.33 mm=sec with the

#140-mesh wheel by ultra-smoothness grinding method [SD140Q50M, V g ¼ 20 m=sec,

t ¼ 5 mm, Soluble (1=50)].

TABLE 2.2

Ultra-smooth Grinding Experimental Conditions

Resultant feed (v w ) c ¼ 3.33 mm=sec

Flow rate: 12 L=min

in plunge grinding Consequently, the grinding grooves are considered to

be removed to some extent

Figure 2.26 shows an AFM 3D image and 2D profiles of 50 mm2of HPSCsurface ground at (vw)c ¼ 3.33 mm=sec The upper and middle 2D profilesparallel to grinding direction are measured at the places excludingand including the microhole, respectively Obvious microholes can hardly

be found on the observed 3D image From the 2D profiles, the surfaceroughness values parallel to the grinding direction are about 3 nm (P-V)and 0.8 nm (RMS) excluding microholes and about 9 nm (P-V) and 1 nm(RMS) including microholes, respectively In addition, the surface roughnessnormal to the grinding direction is about 8 nm (P-V) and 1.5 nm (RMS).Accordingly, in the measurement of AFM accuracy order, it is reasonable tosay that the ground surface with the newly devised method can be ultra-smooth The entire ground workpiece area of 10 mm2, which is furtherobserved widely, however, also consists of the ultrasmooth surface formed

by the low height of grooves with no grinding cracks

Ductile-mode grinding of fine ceramics with the coarse #140-mesh bonded diamond wheel has been shown to be possible in plunge surfacegrinding It is estimated that the thermal effect due to the large wear land ofthe cutting edge on the #140-mesh wheel surface has considerable effect on

Orientation (Front)

U: 9.6 nm L: − 16.5 nm WYKO

P—V: 26.1 nm

FIGURE 2.25

A WYKO 3D image of the HPSC surface ground at (y w ) c ¼ 3.33 mm=sec with the #140-mesh wheel

by ultra-smoothness grinding method [SD140Q50M, V g ¼ 20 m=sec, t t ¼ 5 mm, Soluble (1=50)].

Ductile-Mode Ultra-Smoothness Grinding of Fine Ceramics 51

ductile-mode grindability The various kinds of grinding parameters, that is,table speed, wheel speed, workpiece material property, and so on, influenceductile-mode grindability As a result, optimum grinding conditions should

be selected for ductile-mode grinding Based on the result obtained in plungesurface grinding, ultra-smoothness grinding method has been newly devisedfor obtaining the ultra-smoothness roughness below 10 nm (Rmax), which isalmost the same as that after lapping According to the experimental inves-tigation using the new ultra-smoothness grinding method, the surfaceroughness of the ground HPSC ceramic surface of 50 mm2measured with

( µ m)

20

20 30

30 40

40 (Invert image) (Normal image)

RMS

Rmax

Rz

L : 49.609 µ m : 1.053 nm : 9.103 nm : 3.100 nm

RMS

Rmax

Rz

L : 49.609 µ m : 1.511 nm : 8.109 nm : 6.374 nm

RMS

Rmax

FIGURE 2.26

An AFM 3D image and 2D profiles of the HPSC surface ground at (y w ) c ¼ 3.33 mm=sec with the

#140-mesh wheel by ultra-smoothness grinding method [SD140Q50M, V g ¼ 20 m=sec, t t ¼ 5 mm, Soluble (1=50)].

AFM is as smooth as about 9 nm (P-V) and 1.5 nm (RMS) Therefore, in themeasurement of AFM accuracy order, it is reasonable to say that the groundsurface with the newly devised method can be ultrasmooth The entireground workpiece surface of 10 mm2, however, is also ultrasmooth.

The new method is inadequate for productive ultra-smoothness grinding

of fine ceramics When the method is used after the rough grinding offine ceramics with the same coarse-grain-size grinding wheel, this may

be obtained Nevertheless, it is considered important that the method isimproved toward high productivity due to the investigation of the optimumgrinding condition [7–10]

References

1 Yoshioka, J., Hashimoto, F., Miyashita, M., Kanai, A., Abo, A., and Daito, M.,Ultraprecision grinding technology for brittle materials ASME, Shaw, M.C.,Grinding Symposium PED, Vol 16, 1985, 255

2 Namba, Y., Yamada, Y., Tsuboi, A., Unno, K., and Nakao, H., Surface structure ofMn–Zn ferrite single crystals ground by an ultra-precision surface grinder withvarious diamond wheels Annals of the CIRP, Vol 41, No 1, 1992, 347

3 Ichida, Y et al., Mirror finish grinding of silicon nitride ceramics Proceedings 1stInternational Conference on New Manufacturing Technology, 1990, 317

4 Yasui, H., Arino, Y., and Matsunaga, K., Ductile-mode high smoothness grinding

of fine ceramics by diamond wheel of coarse grain size (1st Report), Journal of theJapan Society for Precision Engineering, Vol 63, No 9, 1997, 1270

5 Yasui, H., Yamazaki, G., Hiraki, Y., Sakamoto, S., Sakata, M., Saeki, M., andHosokawa, A., Ultra-smoothness grinding of fine ceramics with #140-mesh grainsize diamond wheel, Proceedings of 14th American Society for Precision Engin-eering, Annual Meeting, 1999, 125

6 Yasui, H and Yamazaki, G., Possibility of ultra-smoothness grinding of fineceramics using a coarse grain size diamond wheel, Journal of the Japan Societyfor Precision Engineering, Vol 69, No 1, 2003, 115

7 Yasui, H., Development of polishingless ultra-smoothness grinding method(1st Report), Journal of the Japan Society for Precision Engineering, Vol 69, No 12,

2003, 1713

8 Yasui, H and Sawa, T., Effect of grinding fluid supply on ultra-smoothnessgrinding of fine ceramics, Proceedings of the ASPE 18th Annual Meeting,

2003, 447

9 Yasui, H and Sawa, T., Influence of fluid supply on ultra-smoothness grinding

of silicon nitride ceramic with #140 metal bond diamond wheel, Proceedings ofthe ASPE 19th Annual Meeting, 2003, 565

10 Yasui, H and Sawa, T., Ultra-smoothness grinding of a glass with #140 metalbond diamond wheel, ICPMT2004, 2004

Ductile-Mode Ultra-Smoothness Grinding of Fine Ceramics 53

Mechanisms for Grinding of Ceramics

S Malkin and T.W Hwang

CONTENTS

3.1 Introduction 55

3.2 Indentation Fracture Mechanics Approach 56

3.2.1 Median=Radial Cracks: Static Indentor 57

3.2.2 Median=Radial Cracks: Moving Indentor 62

3.2.3 Lateral Cracking and Crushing 65

3.3 Machining Approach 67

3.3.1 Grinding Debris 67

3.3.2 Microscopy of Scratches and Ground Surfaces 67

3.3.3 Grinding Energy and Mechanisms 70

3.3.3.1 Specific Grinding Energy 72

3.3.3.2 Brittle Fracture Energy 74

3.3.3.3 Plowed Surface Area Analysis 78

3.3.3.4 Plowed Surface Energy and Workpiece Properties 79

3.4 Concluding Remarks 81

References 83

3.1 Introduction

Despite the development of advanced ceramic materials possessing enhanced properties, the widespread use of these materials for structural applications has been limited mainly because of the high cost of machining

by grinding In the manufacture of ceramic components, grinding can comprise up to 80% of the total cost [1] Efficient grinding requires selecting

55

operating parameters to maximize the removal rate while controllingsurface integrity [2,3] Lowering grinding costs by using faster removalrates is constrained mainly by surface damage to the ceramic workpiece,which causes strength degradation Any attempts to optimize the grindingparameters should take into account the prevailing grinding mechanismsand their influence on the resulting surface damage and mechanicalproperties.

The present chapter is concerned with what happens during grinding asabrasive grits interact with the ceramic workpiece Most past research ongrinding mechanisms for ceramics has followed either the ‘‘indentationfracture mechanics’’ approach or the ‘‘machining’’ approach [2] The inden-tation fracture mechanics approach models abrasive–workpiece interactionswith the idealized deformation and crack systems produced by an indentor.The machining approach typically involves measurement of forces forsingle-point and multipoint cutting coupled with microscopic observations

of surface morphology and grinding debris Both of these approaches provideimportant insights into the nature of the grinding process for ceramics

3.2 Indentation Fracture Mechanics Approach

The indentation fracture mechanics approach likens abrasive–workpieceinteractions for grinding of ceramics to small-scale indentation events Thedeformation and fracture patterns observed for normal contact with aVickers pyramidal indentor under an applied load P are illustrated inFigure 3.1 A zone of plastic deformation is found directly under theindentor Two principal crack systems emanate from the plastic zone: med-ian=radial and lateral cracks Median=radial cracks are usually associatedwith strength degradation and lateral cracks with material removal

FIGURE 3.1

Plastic zone, median=radial cracks

(R), and lateral cracks (L) for Vickers

indentation (From Lawn, B.R and

Swain, M.V., J Mater Sci., 10, 113,

Though originally developed for static normal loading, this approach hasalso been extended to include the effect of a tangential load (movingindentor).

3.2.1 Median=Radial Cracks: Static Indentor

Now let us consider how median=radial cracks can affect the strengthdegradation due to grinding For this purpose, the normal applied load P

on the indentor (abrasive grit) is considered to cause a median=radial crack

of dimension c, as shown in Figure 3.1, which in turn leads to a reduction inthe fracture strength Larger cracks due to more severe grinding conditionsand bigger forces should cause a greater reduction in the fracture strengthafter grinding

Investigations of median=radial cracks using fracture mechanics started inthe 1970s [4–11] In one of the first studies, the median crack size waspredicted using the Boussinesq solution for the elastic stress field due topoint loading normal to the surface [4] The stress intensity factor obtained

by integrating the stress field around a median crack was used to predict therelationship between the applied load P and crack size c Because elasticitypredicts infinite stress at the contact point, the observed size of the deform-ation zone was taken as a lower limit on the integration A proportionalrelationship was predicted between the load and crack length (P / c), which

is consistent with experimental results for soda-lime glass [4] However, insubsequent work [7], median crack extension was observed to occur notonly during loading but also during unloading This was attributed tononuniform plastic deformation beneath the indentor, which causesresidual stresses Lateral cracking also occurred during unloading

Many researchers have investigated the influence of plastic deformationand residual stresses on median=radial and lateral cracks [10–17] An earlystudy focused on resolving median crack propagation into two parts, anelastic component and an irreversible (residual) component [10,11] Theresults suggest that the elastic component initiates the median crack andcauses it to extend downward during loading, whereas the residual com-ponent provides continued crack extension, as the indentor is withdrawn

By modeling the indentation under the indentor as an expanding plasticzone surrounded by an elastic matrix, the stable crack size c (Figure 3.1) for asharp pyramidal indentor after loading with force P and then unloadingwas obtained as [12]:

experimental results quite well for many materials [12] This relationshipwould apply only above a minimum threshold load P* below which mediancracks should not be initiated [9]:

P* ¼ 54:5(a=h2g4)(Kc4=H3), (3:2)where a, h, and g are constants (a ¼ 2=p for a Vickers indentor, h
1, and

g
0.2)

Although originally developed for a pyramidal indentor, a similar fracturemechanics analysis has also been developed for other indentor shapes.Assuming that the residual stress component is the main source of crackextension, the stress intensity factor for a penny-like crack was obtained as [17]:

Kr¼ xE(dV=V)1=3(dV)2=3=c3=2, (3:3)where x is a constant, V the plastic zone volume, and dV the indentationvolume To predict the relationship between load and crack size, it isnecessary to relate dV and dV=V to the applied load for different indentorshapes The problem becomes greatly simplified by assuming the hardness

to be independent of indentor shape, which is consistent with ments for several brittle materials [13,18,19] The hardness (indentationpressure) can be written as

where a0is a geometrical constant and a is a characteristic dimension of theindentation The ratio dV=V was also found to be independent of indentorshape and to follow a relationship of the form

For stable crack growth, combining Equation 3.3 and Equation 3.5 andequating Krto Kcleads to

Kc¼ x(EH)1=2(dV)2=3=c3=2: (3:6)For a pyramidal indentor, the indentation volume is

where a is half the diagonal length (Figure 3.1) Combining Equation 3.4,Equation 3.6, and Equation 3.7 would give the same load=cracklength relationship as Equation 3.1 with j ¼ {(2=3)2=3=a0}x For a sphericalindentor of radius R,

(dV)2=3¼ (p=4)2=3R2(a=R)8=3, (3:8)

where a is now the indentation radius In this case, the load=crack lengthrelationship becomes

P4=3=c3=2¼ (4p=R)2=3x1(KcE1=2H5=6): (3:9)Therefore, the load=crack length relations in Equation 3.1 and Equation 3.9predict P / c3=2for a pyramidal indentor and P / c9=8for a spherical indentor.This agrees quite well with experimental results in Figure 3.2 for ZnS

Spherical indentor Slope = 9/8

As stated above, strength degradation is usually attributed to ian=radial cracks and the influence of residual stresses on their extension[11,20–22] The residual stress distribution for static indentation has beenpredicted [13] by superimposing the Boussinesq elastic stress field with theelastic=plastic solution for a spherical cavity under an internal pressure Theresults indicate residual compressive stresses near the contact surface with asteep transition to tensile stresses reaching 0.1 to 0.15 H at the elastic=plastic boundary in the subsurface Similar residual stress distributionshave been reported for silicon nitride ceramics after grinding [23–25], butthe peak residual tensile stresses are much smaller For example, the peakresidual tensile stress in Figure 3.3 is only about 0.007 H after grinding ofhot-pressed silicon nitride (HPSN) [25] Such small s tresses may not causesignificant strength degradation However, critical stress intensity factorsevaluated from measurements of cracks produced by sharp indentors wereabout 30% less than those determined by double cantilever or doubletorsion tests [26,27], which suggests that the residual stresses due to inden-tation caused a 30% strength reduction.

med-According to fracture mechanics, the effect of crack size c on fracturestrength sfcan be written as

Depth from the surface ( m m)

200

Transverse to grinding direction Grinding direction HPSN