_a review of micro and nano machining from a materials perspective

Journal of Materials Processing Technology 167 (2005) 316–337

A review of micro and nanomachining from a materials perspective

G.M. Robinson, M.J. Jackson∗

Center for Advanced Manufacturing, College of Technology, Purdue University, P.O. Box 2021,West Lafayette, IN 47907-2021, USA

Received 15 March 2005; received in revised form 28 May 2005; accepted 8 June 2005

Abstract

This paper reviews the current state-of-the-art surrounding the development of micro and nanomachining processes viewed from a materials’perspective. The paper begins by reviewing the theory of machining at the micro and nanoscale, and then introduces the reader to the advantagesassociated with ultra high speed machining. However, the heterogeneous nature of the material being machined has a profound effect on toolwear. Therefore, an extensive part of the paper is devoted to the development of cutting tool materials, coatings and extremely high speedmachining regimes. Specially constructed machine tools are required to use these cutting tools at speeds in excess of one million revolutionsper minute. This review provides a timely explanation of the literature that surrounds the advances made in micro and nanomachining from am©

K

1

dfa[aa[sTcppd(wubt

tified

lfn

ant’serated.

facele,s

0d

aterials’ perspective.2005 Published by Elsevier B.V.

eywords:Micromachining; Nanomachining; High speed machining; Tool wear; Meso machine tools; Machining theory

. Introduction—machining theory

In the 1940s, Ernst and Merchant[1], Merchant[2,3],eveloped models for orthogonal cutting (Fig. 1). Unde-

ormed material was shown to shear after passing throughprimary shear zone (Fig. 2). Earlier work by Piispanen

4] likened this shearing to a deck of stacked cards inclinedt ϕ, the shear angle (Fig. 3). Idealized cutting conditionsre considered based on the assumptions made by Merchant

2] and presented by Shaw[5]: (a) The tool is perfectlyharp and there is no contact along the clearance face; (b)he shear surface is a plane extending upward from theutting edge; (c) The cutting edge is a straight line extendingerpendicular to the direction of motion and generates alane surface as the work moves passed it; (d) The chipoes not flow to either side; (e) The depth of cut is constant;f) The width of the tool is greater than the width of theorkpiece; (g) The work moves relative to the tool withniform velocity; (h) A continuous chip is produced with nouilt-up edge; and (i) The shear and normal stresses alonghe shear plane and tool are uniform.

The forces acting between the chip and tool are idenin a free body diagram (Fig. 4). At equilibrium the forcebetween the tool and chip R, consisting ofNC andFC, is equato the force between the workpiece and chip, R′ consisting oFP andFQ and also ofFS andNS (Fig. 5). These forces cabe rearranged and applied at the tool tip; this is Merch[2,3] circle of cutting forces (Fig. 6). From this circle, a rangof equations describing the cutting process can be gene

FS = FP cosϕ − FQ sinϕ (1)

NS = FQ cosϕ + FP sinϕ (2)

NS = FS tan(ϕ + β − α) (3)

whereα is the angle between the vertical and the toolcalled the rake angle consisting ofϕ the shear plane angβ the friction angle andµ the coefficient of friction and igiven by;

m = tanβ (4)

FC = FP sinα+ FQ cosα (5)
∗ Corresponding author. Tel.: +1 765 494 0365; fax: +1 765 494 6219.E-mail address:[email protected] (M.J. Jackson). NC = FP cosα− FQ sinα (6)
924-0136/$ – see front matter © 2005 Published by Elsevier B.V.
oi:10.1016/j.jmatprotec.2005.06.016

G.M. Robinson, M.J. Jackson / Journal of Materials Processing Technology 167 (2005) 316–337 317

Fig. 1. Orthogonal cutting conditions[5].

µ = FC

NC(7)

µ = FP sinα+ FQ cosα

FP cosα− FQ sinα(8)

µ = FQ + FP tanα

FP − FQ tanα(9)

The shear stressτ is given by;

τ = FS

AS(10)

where

AS = bt

sinϕ(11)

Here,b is the width of cut andt is the depth of cut, thereforethe shear stress is;

τ = (FP cosϕ − FQ sinϕ)sinϕ

bt(12)

Similarly the normal stressσ is given by:

σ = NS

AS(13)

Fig. 3. Piispanen’s[4] stacked card analogy of metal cutting.

Fig. 4. Isolation of forces on a free body diagram, after Shaw[5].

σ = (FP sinϕ + FQ cosϕ)sinϕ

bt(14)

An equation forϕ is still required. It is found experimentallythat when certain metals are cut there is no change in density.The subscript C refers to the chip andl is the length of cut.Therefore;

tbl = tCbClC (15)

ial passes through the shear zone[5].
Fig. 2. Undeformed mater

318 G.M. Robinson, M.J. Jackson / Journal of Materials Processing Technology 167 (2005) 316–337

Fig. 5. Identification of the constituent components of force[5].

It is found experimentally that ifb/t≥ 5, the width of the chipis the same as the workpiece thus

t

tC= lC

l= r (16)

wherer is the cutting ratio, or chip thickness ratio;

r = t

tC= AB sinϕ

AB cos(ϕ − α)(17)

where AB refers to the length of the shear plane. Solving forthe shear angleϕ.

tanϕ = r cosα

1 − r sinα(18)

The work length may be determined by weighing the chip, ifthe chip weighswC andγ ′ is the specific weight of the metalthen.

l = wC

tbγ ′ (19)

The shear strainγ is given by;

γ = cosα

sinϕ cos(ϕ − α)(20)

Expressions can also be developed for the cutting velocityV, which is the velocity of the tool relative to the work anddirected parallel toFP. The chip velocityVC is the velocity ofthe chip relative to the tool and directed along the tool face.The shear velocityVS is the velocity of the chip relative tothe workpiece and direct along the shear plane.

VC = sinϕ

cos(ϕ − α)V (21)

or,

VC = rV (22)

VS = sinα

cos(ϕ − α)V (23)

or,

VS = γ sinϕV (24)

The shear strain rateγ ′′ is given by:

γ ′′ = cosα

cos(ϕ − α)

V

�y(25)

w ftena cingb surea fore,M hst thoutt

izedc le,o thet ht, h,c

h

w dn

fromt fric-t art oft reda fin-i hearf edgeb

nedt totalp um.H ty ofe er
Fig. 6. Merchant’s circle of cutting forces[2,3].
here�y is the thickness of the shear zone and is opproximated by assuming its value is equal to the spaetween slip planes. It is not always convenient to meall the quantities required to apply the equations. Thereerchant and Zlatin[6] produced a number of nomograp

hat can be used to obtain some of these quantities wihe need for extensive experiments (Fig. 7).

Not all machining operations are similar to these idealutting conditions. For example, during milling, multipblique, cutting points remove material. The depth of

rough to peak surface marks produced or scallop heigan be determined by:

= f

4(D/f ) ± (8n/π)(26)

hereD is the diameter of the cutter,f the feed per tooth anis the number of cutting teeth.The built-up edge is a situation that seriously deviates

he idealized cutting conditions. The heat generated byion at the secondary shear zone can be enough to weld phe chip to the tool. Tool chip contact conditions are altend this affects the depth of cut and quality of surface

sh. With time the built-up edge grows so large that the sorces are high enough to remove it and another built-upegins to form, the process is therefore dynamic.

Piispanen[4] and later Merchant independently reasohe shear plane angle would take a value such that theower consumed during machining would be a minimowever, depending on the assumptions made, a variequations can be generated for the shear angle. Stabl[7],


Fig. 7. (a and b) Merchant and Zlatin’s nomographs[6].


Lee and Shafer[8] and Oxley [9] developed such equa-tions. A comprehensive review of the machining processat the microscale that takes account of the size effect hasbeen provided by Shaw[10] and Shaw and Jackson[11]. Acomprehensive review of metal cutting at the macroscale hasrecently been published by Shaw[5].

2. High speed machining

High speed machining is usually defined by spindle speedsbetween 30,000 and 100,000 revolutions per minute (rpm)and usually has the advantages of an increased metal removalrate, reduced cutting forces, increased dissipation of heatand better surface roughness. Schulz[12] summarized theadvantages inTable 1. A detrimental effect of high speedmachining can be that the rate of tool wear increases, how-ever, Schulz[12] found that optimizing the cutting parame-ters, contact conditions, workpiece material, the cutting tooland the machining strategy extend tool life.

Cutting speeds can be significantly increased if ceramictools are used. Usui et al.[13] constructed a cutting modelbased on an energy approach; later Usui and Hirota[14]extended the model to examine chip formation and cuttingforces with a single point tool. Finally, Usui et al.[15] investi-g ools.E ltt per-a ingt ssesa eightd nd toia forc

ofh fric-t greeo tings ness.O ingr peeda g thet

Fig. 8. A graph of the work of Kitagawa et al.[16] showing that an increasein cutting speed produces an increase in temperature.

Fig. 9. A graph of the work of Kitagawa et al.[16] showing the variation oftemperature on the tool with increasing cutting speed.

Finite element models produced by Moufki[18] show forsteel, that cutting temperatures are predicted to reach between500 and 1000◦C. The coefficient of friction has a significanteffect on heat generation but its value is not constant, makingpredicting the tool–chip interface temperature challenging,

TS chulz12]

B Application examples

L teel and cast iron Aircraft and aerospace production, die andmould manufacturing

H special workpieces Optical industry, fine mechanical partsL d workpieces Aircraft and aerospace industry, automotive

industryH cal frequencies Precision mechanisms and optical industryH ieces with critical heat influence Precision mechanisms magnesium alloys

ated thermally activated wear mechanisms for ceramic txperimental work by Kitagawa et al.[16] showed that too

emperature increases with cutting speed (Figs. 8 and 9) andhat rake face temperature is higher than the flank face temture. Usui et al.[13,14] point towards evidence suggest

hat tool wear is triggered by thermally activated procend to a lesser extent by abrasive processes. Burr hecreases and the quality of surface roughness is fou

ncrease with increasing cutting speed, Kitagawa et al.[16]lso found that Taylor’s tool life equation could be usedalculating the life of ceramic tools.

Ozel and Altan[17] produced finite element modelsigh speed machining by using a variable coefficient of

ion to account for the dynamic cutting situation. The def chip curl is found to be heavily dependent on the cutpeed and higher cutting speeds reduce the chip thickzel’s model[17] predicts that in the high speed machin

egime cutting forces decrease with increasing cutting snd a rise in feed rate increases the cutting forces alon

ool edge.

able 1ummary of the advantages of high speed machining discussed by S[

enefit Application field

arge cutting volume per machining time Light metal alloys, s

igh surface quality Precision machining,ow cutting forces Processing thin walle

igh frequencies of excitation No machining in critieat transported by the chips Machining of workp


Fig. 10. The finite element mesh produced by Kim and Sin[24] to modelthe cutting process.

Bailey [19]. Montgomery[20] showed that at large slidingvelocities a reduction in the coefficient of friction is observedwhen the normal pressure or sliding velocity is similarlyincreased. Therefore, no existing model adequately accountsfor all aspects of cutting and results are not always satisfac-tory, in particular, predictions of temperature. The models arevalidated by comparing predicted temperatures to directly

measured experimental data obtained using thermocouplesdemonstrated by Groover and Kane[21], infrared sensorsdemonstrated by Wright and Trent[22] and chip microstruc-ture analysis demonstrated by Fourment et al.[23]. Finiteelement analysis by Kim and Sin[24] can accurately pre-dict the way in which chips form. However, it is difficultto quantify the contributions made by different mechanismsbecause the situation is dynamic, i.e. the coefficient of fric-tion is dependent on the sliding velocity or normal pressure,which changes as the tool advances (Figs. 10 and 11).

Trent and Wright [25] observed during slow speedmachining that the condition of chip sliding is dominant,during high speed machining the condition of seizure is domi-nant. Seizure occurs when the apparent area of contact equalsthe actual area of contact; this is in agreement with the workon friction carried out by Doyle et al.[26]. Seizure is charac-terized by craters located near to the tool edge and Gekondeand Subramanian[27] predicted that these craters have a max-imum depth that correlates to the phase change temperature.The phase change temperature is sufficient to cause disloca-tion generation, which leads to diffusion wear. Gekonde andSubramanian[27] also found that total tool wear is the netresults of mechanical and chemical wear. Mechanical wearremains constant, independent of the cutting speed, whereaschemical wear in the form of diffusion increases with cuttingspeed. Metal cutting theory is well established but its appli-cf kesr

Fig. 11. The cutting temperature distribution produced by Kim a

ation, particularly to milling, can be difficult; Gygax[28]ound the rapid periodic impact of each cutter often maeading a dynamometer challenging. Rotberg[29] decided

nd Sin’s[24] finite element analysis to model the cutting process.


Fig. 12. The hexapod machine structure considered by Schmitt[30].

the most efficient method to apply cutting models is to uti-lize a large number of sensors, a few experimental resultsand idealized cutting conditions. Schmitt[30] points out thathigh-speed milling is not always the most suitable methodfor machining. It is usually employed specifically where itsattributes are critical for successful machining of the part.

Schmitt [30], Weck and Staimer[31] and Ibaraki et al.[32] considered different methods of creating the workingenvelope. They found a hexapod structure (Fig. 12) is morestable than a conventional milling machine structure. Thereduction in vibrations and jerk yields an improved surfaceroughness. Moller[33] summarized the special requirementsof high-speed spindles. He found that the best designs hadno transmission, achieved high speeds with low vibration,were liquid cooled leading to high thermal stability and had

low inertia which allows high accelerations and decelera-tions. Another significant challenge identified by Moller[33]is the problem of selecting bearings suitable for operationat high speed. Moller[33] also showed the advantages ofcutting at high speed include a reduction in force (Fig. 13).Cohen and Ronde[34] have proposed that hydrostatic bear-ings could operate successfully at high speed. Hydrostaticbearings eliminate run out, hence they are often used onmachine tools that machine optical materials.

3. Tool wear

Establishing the point of tool wear is important because thetool can be changed prior to unacceptable machining results.

by us
Fig. 13. Force reduction ing high speed cutting[33].


Fig. 14. The contribution made to overall wear by chemical and mechanicalsources[36].

Tansel et al.[35] trained neural networks to monitor cut-ting forces, and based on past failures, the point of criticaltool wear could be identified. Ingle et al.[36] investigatedcrater wear, which has chemical and mechanical components(Fig. 14). Ingle et al.[36] machined calcium treated AISI1045 grade steel with a cemented carbide tool containing97% tungsten carbide, 0.4% Ta(Nb)C and 2.6% cobalt. Anexpression based on the work of Bhattacharyya and Ham[37] and Bhattacharyya et al.[38] was developed to estimatethe amount of tungsten transported by diffusion during thecutting time.

W = 1.1284C0ft(D/τ)1/2 (contact area) (27)

whereW is the amount of tungsten dissolved,C0 the equi-librium concentration ofWat interface,F the ferrite volumefraction, t the cutting time,D the diffusivity of tungsten inferrite, andτ is the tool chip contact time. It was found that ifa tungsten carbide tool was used to machine steel, a TiN coat-ing decreased the thermodynamic potential for dissolution bysix orders of magnitude compared to the uncoated case. Ingleet al. [36] also found that cutting speed affected the cuttingtemperature; at 50 m/min the temperature was 1337 K, andat 240 m/min the temperature was 1512 K. In related work,Subramanian et al.[39] found diffusion to be the main sourceof crater wear when machining AISI 1040 grade steels witha gher.N ento thec r-i ntedt and1 ork-p gstenc ini-tp toola chipi pm.H chip

is approximately constant, regardless of the cutting speed,around 0.8 ppm.

To test the effectiveness of a 10�m thick CVD depositedHfN coating, experiments were performed at cutting speedsof 220 and 240 m/min and the concentrations of tungsten inthe chips was found to be 11.3± 0.1 and 11.2± 10.1 ppm,respectively. It was therefore concluded that at high cuttingspeeds a thin HfN coating was successful at preventing dis-solution of tungsten from the workpiece. During testing ofthe HfN coated tools, Subramanian et al.[39] employed atechnique suggested by Hastings et al.[41] that was basedon a method developed by Boothroyd[42] to determine thechip temperature.

If there is no coating the tool can loose a large amountof tungsten by dissolution to the chips. Sherby et al.[43]found that creep resistance of a metal is higher than its self-diffusion activation energy. Sherby[44] found during hightemperature deformation that the formation of sub bound-aries, grain boundary shear, and fine slip are related to inter-atomic diffusion. These static diffusion studies by Sherby[44] can help explain the dynamic diffusion present at thetool–chip interface. Gregory[45] studied diffusion betweena cemented carbide tool of composition WC 74%, Co 10%and TiC 16% and a workpiece material of Armco iron ofcomposition C 0.05%, Mn 0.05%, S 0.028%, P 0.01% andFe balance. Specimens were diffused under vacuum at tem-p .T ogo-n edr undt fol-l ne.T d tob ter-mp nsid-e st niumc eelc eavet moli thei l aren byt hea l fort ed ut-w esee ismsb ne

edme singc the

tungsten carbide tool at speeds of 175 m/min and hiaerheim and Trent[40] observed a concentration gradif tool material in the chip, specifically W and Co nearhip–tool interface. Subramanian et al.[39] conducted expements coated and uncoated tools made from K1 cemeungsten carbide containing 85% WC, 4% (TaC/NbC)1% Co. A loss of tungsten carbide particles from the wiece characterizes mechanical wear and a loss of tunarbide by dissolution characterizes chemical wear. Theial amount of tungsten in the workpiece was 11± 0.5 partser million (ppm). For the uncoated tungsten carbidet 150 m/min, the amount of tungsten dissolved in the

s 1.5 ppm whereas at 240 m/min this amount is 10.5 powever, the amount of tungsten carbide found in the

eratures of 1100, 1175, 1250 and 1325◦C (Fig. 15a and b)hese results were compared to a tool subjected to orthal cutting of the Armco iron at 370 ft/min for 1 h with a feate of 0.0015 in./min. The first stage of diffusion was foo be the outward migration of the cobalt binder phaseowed by a build-up of titanium carbide in the reaction zohe activation energy triggering cobalt diffusion was foune 83 kcal/mol, which is comparable to 72.9 kcal/mol deined in another study by Suzuoka[46]. After the diffusionhase, a period of stable of wear sets in; this can be cored comparable to the run-in of a tool. Trent[47] suggest

he stable period of wear is due to the presence of titaarbide, which is more difficult to take into solution in stompared to tungsten carbide. For titanium carbide to lhe tool material a thermal activation energy of 124 kcal/s required. However, for titanium carbide to diffuse intoron–cobalt reaction zone energy levels of 150 kcal/moeeded. Gregory[45] argued that final wear is explained

he continual diffusion of the titanium carbide zone. Tctivation energies of its constituents are, 98 kcal/mo

he inward diffusion of iron, 142 kcal/mol for the volumiffusion of iron in tungsten and 77.5 kcal/mol for the oard diffusion of tungsten during the whole process. Thnergies point towards grain boundary diffusion mechanecause Danneberg[48] found that the volume self-diffusionergy of tungsten was between 110 and 121 kcal/mol.

Nayak and Cook[49] reviewed some thermally activatethods of tool wear. Nayak and Cook[49] found mod-ls predicting thermally triggered wear mechanisms uontinuum diffusion theory experience difficulties when


Fig. 15. (a and b) Examples of diffusion couples tested by Gregory[45].

workpiece material and tool material are similar. This isbecause the model assumes diffusion is driven by a con-centration gradient of the diffusing material. An alternateapproach is to assume that vacancy concentration is the mech-anism driving diffusion. Characteristics identifying the wearmechanism were determined so that an examination of exper-imental results would reveal the process responsible. Nayakand Cook[49] conducted a series of experiments and data wasobtained using M2 tool steel hardened to HRC 65. There wasno evidence that an asperity fracture model can predict wear.A creep model could explain the triggering of wear but thewear rate is calculated at 4× 10−12 in./s, which is six ordersof magnitude less than that observed. Depletion of interstitialcarbon atoms was also investigated; however this mechanismwas rejected because diffusion was in the opposite directionthe theory predicted. Wear by then removal of oxide films isrejected because no change is observed when machining inan oxygen free environment and there is not enough oxygenin the material to cause the required effect. Nayak and Cook[49] also considered tempering wear. At critical tempera-tures transition carbides, such as W2C and Mo2C can form,thus reducing tool hardness due to tempering. The equation,V= 1.4× 1015 e−90000/RT in./s can be derived for the tool wearrate,V, which agrees reasonably well with experiments con-ducted.R is the universal gas constant (cal/mol) andT is theabsolute temperature (Kelvin). However, equations predict-i eriva-t tot causet over-a we ctorsa , canba g( ol),R -

lute temperature (Kelvin).Q andA variables are dependenton the work material. The tool–chip interface is consideredsimilar to a grain boundary interface and therefore the acti-vation energyQ required for an atomic jump is also similar;this is in the order of 40–45 kcal/mol as determined by Ley-monie and Lacomb[50]. The atomic vibration frequencyis approximately 1014 s−1 and the atomic spacing approx-imately 10−8 in., yielding, v= 106 e−42000/RT in./s which isin excellent agreement when cutting 1018, 1045, 4140 and10C–40C steels.

A statistical model for the predictions of tool wear can bedeveloped; Davis[51] suggested a Monte Carlo approach formodeling mechanical wear based on the following assump-tions. Quanta of energy are added to the rubbing particles allof which leave the same site. The quanta of energy are addedto the particles at domains randomly spaced over the surfaceat moments randomly spaced in time. The quantized energyof a particle over the surface diffuses continuously into thematerial at a known rate. Finally the surface energy of a par-ticle causes it to detach as a wear particle. Bhattacharya andHam[37] extended this approach developing expressions forthe width of flank wear due to mechanical wear from abrasiveand adhesive sources. The width of the flank wears,Hf (T),at timeT is given by:

H (T ) = K′′V T 1−α (28)

wc

K

w

C

ng other machining parameters generated during the dion of this equation predict significantly different resultshose observed. Atomic wear was also considered, beool atoms are in a higher energy state than chip atoms,ll migration is from tool to chip where atoms flow to lonergy sites in the chip, such as vacancies. If these fare considered an equation describing the wear rate Ve developed, thusV is equal toυAe−Q/RT, whereυ is thetomic vibration frequency (s−1), a is the atomic spacinin.),Q is the energy barrier for the atomic jump (kcal/m

is the universal gas constant (cal/mol) andT is the abso

f C

hereVc is the cutting velocity,T the cutting time,α thelearance angle andK′′ is given by:

′′ = 3K′

2C′ (1 − α) (29)

hereC′ is given by:

′ = C2

σ2 (30)


whereC is a constant,σ the deviation from the mean andK′is given by:

K′ = Km

ρ(

tanα1−tanγ tanα

) (31)

whereρ is the density of tool material, m is the mass of thewear particle,γ the true rake angle andK is the constantgoverning decay rate. The height of flank wear,hf , can alsobe determined by:

hf =[

5

4

KVcA3/4(1 − tanα tanγ)T

b tanα

](32)

whereA is a constant based on the tool and workpiece com-bination andb is the width of cut. This model worked wellwhen applied to experiments conducted by Bhattacharyyaet al.[38] and experiments conducted by the internationallyrecognized tool wear collection body OECD/CIRP. How-ever, a large amount of data for specific tool and workpiececombinations must be collected before this approach can beused. Bhattacharyya et al.[38] points out that the reliabil-ity of the model depend on the accuracy of the data, makingthis approach susceptible to compounded errors. It is alsoinconvenient to collect a large amount of data every timenew machining conditions are encountered.

4

ia-m ble tor toold ondc ol andc /TiNc en thet hash inert.H tedc dia-m t be

Fig. 17. Fracture of the coating[52].

etched away or a diffusion barrier introduced, Jackson et al.[53].

Faure et al.[52] assessed the coatings with a ‘Revetest’test device and drilling experiments. The ‘Revetest’ deviceapplies a constant force to an indenter during constant veloc-ity displacement of the sample and critical loads are identifiedby an acoustic signal (Fig. 16a and b). Drilling tests wereperformed at 69,000 rpm at 3.5 m/min with a 1 mm diameterdrill. It is common for a soft substrate to deform plasticallywhile its coating does not because it has a high Young’smodulus. The coating alone bears the load and fails thusexposing tool material to wear mechanisms (Fig. 17). Dia-mond coatings with a TiN interlayer (Fig. 18) can withstand10 times more force compared to diamond coatings withoutthe interlayer; adhesion is not necessarily better but tough-ness is improved. Both cases with and without an interlayeroffer a significant improvement over tools with no coatingwhere rounding of the tool edge is observed after drillingonly one hole (Fig. 19). If a TiC/TiN/TiC multilayer is appliedlarge forces, around 100 N, are required to break the surface.Uncoated tools were able to drill 10,000 holes and coatedtools drilled 20,000.

Bell [54] categorized tool materials into three basicgroups, high-speed steels, cemented carbides and ceramicand super hard materials including alumina based compos-ites, sialons, diamond and cubic boron nitride (Fig. 20). To

aure et[5

. Tool coatings

Faure et al.[52] discussed the main issues involving dond coatings. Thin uncoated tool edges are suscepti

ounding and therefore must be protected. Coating theirectly with diamond can result in shattering of the diamoating because of the hardness gradient between the tooating. To reduce this effect an interlayer, such as TiCan be introduced which also enhances adhesion betweool and coating. Diamond is an ideal coating material; itigh hardness, high wear resistance and is chemicallyowever it is difficult to coat steels, Ni alloys, cemenarbides and alloys containing transition metals withond. It is possible to coat WC–Co, but the cobalt mus

Fig. 16. (a and b) Scratch tests performed by F
al.2] to determine the adhesion of diamond coatings.


Fig. 18. A drill produced by Faure et al.[52] with an interlayer and a diamondcoating before testing.

prevent failure, Mills[55] suggests modifying the surface oftool material by CVD or PVD techniques. Different machin-ing processes are characterized by different wear mechanisms(Fig. 21) and the choice of tool coating should be selected tooffer the best protection for a particular set of machining con-ditions; e.g. the combination of wear rate, bearing pressureand tool material (Fig. 22). There are four main types of toolcoatings, titanium based, e.g. TiAlCrN, ceramic based, e.g.Al2O3, super hard, e.g. CVD diamond and solid lubricantcoatings, e.g. Me–C:H.

Kubaschewski and Alcock[56] concluded that to preventthe onset of diffusion the enthalpy of the coating must beas negative as possible to increase the temperature at whichdiffusion is triggered. From this point of view most carbidecoating materials, such as TiC, HfC, ZrC are more suitable forcutting steel than tungsten carbide, similarly for the nitridesexcept CrN up to a temperature of 1500◦C. Eventually,

Fig. 20. Thickness of surface layers and their method of manufacture[54].

Fig. 21. Different wear mechanisms on a cutting tool[55].
Fig. 19. An uncoated drill after drilling one hole[52].


Fig. 22. Normalized wear rate,Ka, plotted against bearing pressure for arange of tool materials[55].

multilayer coatings were developed. Klocke and Krieg[57]summarized their three main advantages, better adhesion tothe tool, improved mechanical properties and different layerscan provide different functions. The final coating incorpo-rates properties from each layer, e.g. a TiN–NaCl multilayerhas hardness 1.6 times greater than a single TiN layer.

In some cases, such as machining hypereutectic Al–Sialloys, only diamond coatings can offer improvements intool life due to the abrasiveness of Si particles. Quinto etal. [58] investigated coatings deposited by CVD and PVDtechniques on two alloys A and B (Figs. 23–26). Coatingswith an Al content tend to perform better regardless of theapplication, coating process or chemical content of the othercoating constituents. This is because abrasive resistance, oxi-dation resistance and hardness are all improved. Quinto et al.[58] found that PVD coatings outperform CVD coatings thatoutperform uncoated tools (Figs. 27–29).

Dry machining of steels in the range of 55–62HRCat 15,000–25,000 rpm generates cutting temperatures of1000◦C and the tool must be protected from oxidation wear.Previous work by Munz et al.[59] has shown that above

Fig. 24. CVD deposited TiN on alloy B[58].

Fig. 25. PVD deposited TiN on alloy A[58].

Fig. 26. PVD deposited TiN on alloy B[58].

Fig. 27. Appearance of an uncoated insert at the end of tool life after machin-ing 1045 AISI steel[58].
Fig. 23. CVD deposited TiN on alloy A[58].


Fig. 28. Appearance of an TiN CVD coated insert at the end of tool life aftermachining 1045 AISI steel[58].

Fig. 29. Appearance of an TiN PVD coated insert at the end of tool life aftermachining 1045 AISI steel[58].

800◦C diffusion of stainless steel is triggered and cavitiesbegin to form between the substrate and coating, althoughthis can be prevented by adding 1% yttrium. One methodused by Constable et al.[60] to analyze the integrity of coat-ings is Raman microscopy. Constable et al.[61] demonstratedthe usefulness of Raman microscopy when a PVD combinedcathodic arc/unbalanced magnetron deposition system wasused to coat high-speed steel and stainless steel for abrasiontests. The coatings had a thickness between 2.5 and 4�mwith a surface roughness of 0.02–0.03�m roughness aver-

age (Ra). Deeming et al.[62] investigated the effect differentcoatings have on delaying the onset of oxidation. Duringhigh speed machining, temperatures regularly exceed 900◦C.Deeming found a TiN coating delayed oxidation until 500◦C,a TiAlN coating delayed oxidation until 700◦C and a multi-layered system delayed oxidation formation until 950◦C. Thefinal coating tested was TiAlCrYN. The addition of yttriumincreases wear at low temperatures, however at higher tem-peratures yttrium causes maximum wear to occur at 600◦Cand minimal wear to occur at 900◦C. Without the additionof yttrium the wear rate continually increases with temper-ature. Deeming et al.[62] suggests yttrium diffuses into thegrain boundaries and at high temperatures there is some stressrelaxation. TiAlCrYN has a lower coefficient of friction withincreasing temperature compared to TiN (Fig. 30) and alsohas a lower wear rate at elevated temperatures (Fig. 31).

A problem with PVD coatings identified by Creasey et al.[63] is when ion etching is used to evaporate target materi-als; subsequent deposition by magnetron sputtering can leadto the formation of droplets which adhere badly to the sur-face and cause weaknesses in the coating. This is particularlyproblematic when depositing TiAlN. It has been shown byMunz et al.[64] the melting temperature of the cathode mate-rial influences the number and size of these droplets. Gahlinet al. [65] have demonstrated cathodic arc deposition andWang and Oki[66] have demonstrated low voltage beame t-i on-arcb t lowa tech-n BMfi forcef

m-p atingf ucedu withaa steelhl thefi on

Fig. 30. Comparison of the coefficient of friction

vaporation deposition. Hurkmans et al.[67] deposited coangs using a combined steered arc/unbalanced magnetrond sputtering (ABS) technique. It has been found thadhesion problems can be overcome by using the ABSique and conducting a pre-etch at 1200 eV Cr prior to Ulm deposition; this can more than double the adhesiverom 25 to 50 N.

Wadsworth et al.[68] reasoned that a multilayer coosed of several materials would produce an optimal co

or preventing tool wear; such a coating can be prodsing PVD techniques. Wadsworth deposited coatingsn ABS system at 450◦C, the targets were Ti0.5Al0.5 andCr plate, the substrates were 304 austenitic stainless

eat treated M2 high speed steel. A 0.2�m thick TiAlN baseayer was deposited prior to the TiAlN–CrN multilayer;nal coating thickness was 3.5�m. Tests were conducted

produced by TiN and TiAlCrYN coatings[62].


Fig. 31. Comparison of the wear rate produced by TiN and TiAlCrYN coatings at elevated temperatures[62].

these coatings at 750◦C it was found after 16 h the hardnessdropped to 475 HK0.01. The addition of Cr reduces the onsetand rate of oxidation.

Smith et al.[69] coated various substrates and performeda series of tests. The two coatings tested were TiAl and TiAlNwith a Cr etch. The growth defects previously discussedwere highlighted in the TiAl coating (Fig. 32) while the Cretched TiAlN coatings was defect free (Fig. 33). Drills werecoated and were initially tested at 835 rpm and a feed rate of0.28 mm/rev, for each subsequent test the spindle speed wasincrementally increased. It was found TiAlN coated drills outperformed commercially available drills due to the smootherTiAlN surface, which was enhanced by the Cr etch.

Fig. 34. Coatings deposited at−25 V bias[71].

Petrov et al.[70] also experienced the droplet forma-tion described by Munz during the coating procedure. UnderUBM deposition conditions the films exhibited columnargrowth with a compressive stress around 3 GPa and around1.5% trapped Ar. For a thickness of 3�m peak-to-peak rough-ness was observed at 50 nm. Using a UBM/CA depositiontechnique a thickness of 3�m peak-to-peak roughness wasobserved at only 5 nm. Higher compressive stresses wereobserved, 9 GPa with an Ar concentration of 0.5%.

Salagean et al.[71] also deposited such coatings using anarc and unbalanced magnetron-sputtering cathode equippedwith an Nb target; base pressure 5× 10−5, 10 kW and a6 A coil at 400◦C. It was found that prior etching treatmentdefined the quality of the new surface as well as the voltagebias during deposition (Figs. 34–36). Donohue et al.[72] pre-ferred TiAlN to TiN coatings due to their greater resistanceof oxidation (Figs. 37 and 38). Donohue et al.[72] also inves-tigated the oxidation temperature, for TiN, and found it was600◦C, for Ti0.46Al0.54N, 870◦C, for Ti0.44Al0.53Cr0.03N,920◦C and 950◦C, for Ti0.43Al0.52Cr0.03Y0.02N. They there-fore concluded that extra alloying elements were significant.

5. Micromachining

Inamura et al.[73] found it difficult to apply finite element

Fig. 32. A cross-section showing growth defects in a TiAl coating[69].

Fig. 33. Cross-section of a Cr etched TiAlN coating[69].
a nalysis to micromachining processes around 1 nm (Fig. 39).




They found that a more successful approach is to use molec-ular dynamics simulations, where the position of each atomis resolved by Newtonian dynamics. Such simulations pre-dict that blunt tools generate a large shear area, faster cuttingspeeds produce thinner chips, and that sharp tools operatedat high cutting speeds impart low forces to the tool. However,

F ing al

Fig. 38. Thermo-gravimetric oxidation rate measurements in air during anisothermal anneal at 900◦C [72].

Kim and Moon[74] found that when considering the cuttingzone (Fig. 40), if the tool becomes blunt the forces generatedby a blunt tool at high speed are significantly higher thanforces produced by a blunt or sharp tool at low speed. There-fore, to exploit the advantages of micro cutting, moleculardynamics simulations of chip flow (Fig. 41), show that thecutting tool must be sharp and cutting speeds must be high.

Gillespie[75] observed burr formation at the micro scaleand discovered macro scale burr removal techniques couldnot be applied at the micro scale. Gillespie and Blotter[76]stated there are three generally accepted burr formation mech-anisms: lateral deformation; chip bending and chip tearing.Kim [77] found that accuracy and repeatability of macro burrremoval techniques are lost at the micro scale. Work by Koand Dornfield[78] described a three-stage process for theburr: initiation; development; and formation. However, at thistime burr formation is not well understood but correlationsbetween machining parameters can identify key variablesthat help reduce burr size; for example, observations of burrs(Fig. 42) by Lee and Dornfeld[79], indicate that up-millinggenerally creates smaller burrs than down milling and theseobservations along the length of a track led to the categoriza-tion of different burr types (Fig. 43).

The results of Ikawa et al.[80] suggest there is a criticalminimum depth of cut, below which chips do not form. Theanalysis of Yuan et al.[81] indicates chip formation is notp ttinge hipss chipco sarilyf abin

ig. 37. Thermo-gravimetric oxidation rate measurements in air durinear temperature ramp at 1◦C min−1 [72].

ossible if the depth of cut is less than 20–40% of the cudge radius. Micro machining at 80,000 rpm produces cimilar to those created by macro scale machining, whereurl and helix effects are observed by Kim et al.[82]. Kim alsobserves that if the feed rate is too low a chip is not neces

ormed by each revolution of the tool. Sutherland and B


Fig. 39. Finite element mesh to simulate micromachining[73].

Fig. 40. Molecular dynamics simulation of the cutting zone[74].

[83] found that feed, or machining marks, are separated by aspacing equal to the maximum uncut chip thickness. Resultsshow at small feeds per tooth the distance between feed marksis larger than the uncut chip thickness indicating no chip has

Fig. 42. An exit burr[79].

been formed. Kim et al.[82] conclude that a tool rotationwithout the formation of a chip is due to the combined effectsbetween the ratio of cutting edge radius to feed per tooth andthe lack of rigidity tool of the tool.

Work by Ikawa et al.[84] and Mizumoto et al.[85] showedsingle point diamond turning can machine surface roughnessto a tolerance of 1 nm. It is useful to model the cutting process

Fig. 41. Chip formation predicted by m
olecular dynamics simulation[74].


Fig. 43. Categorization of burr types[79].

using a molecular dynamics simulation approach. Ikawa et al.[84] found that the behavior and flow of the material could bepredicted by the laws governing interactions between neigh-boring atoms. Ikawa’s simulations[85] suggest the depth ofcut and cutting edge radius are critical parameters that deter-mine chip formation (Fig. 44). Shimada et al.[86] comparedtheir predicted results with experimentally determined valuesfor a copper workpiece and found them to be accurate.

If the strain applied to the workpiece by the tool is largeenough, dislocations can be generated. As the tool advancesmore dislocations form at the tool–chip interface and ifenough join at the primary shear zone then a chip is formed.After the tool has finished cutting, dislocations that pene-trated the workpiece migrate out towards the surface becausethe lattice can relax. This phenomenon can be observed asatomic sized steps on the surface, which represents the bessurface roughness possible, Shimada[87].

In macro scale machining the tool edge sees bulk prop-erties of the workpiece, however in micromachining the tooledge sees features of the material matrix, such as grain boundaries. Shimada et al.[88] used molecular dynamic models tosimulate this interaction. Simulations running cutting speedsat 2000 m/s show the kinetic energy imparted to the work-piece is far greater than the cohesive energy of the workpiece.Komanduri et al.[89] conducted simulations that help high-light differences between macro and micro scale cutting, fore sili-c cturef 3%d r-e ler eta ainsb mes as arge

ploughing forces. The tool’s slenderness ratio reaches a pointwhere tool stiffness is reduced. Vogler’s model predicted aworn tool edge could produce 300% more cutting force andresults in poor surface roughness and increased burr forma-tion. Cutting conditions of 120,000 rpm, a feed per tooth of4�m and a depth of cut of 100�m produced a peak cuttingforce of 3 N. Separation between feed marks was 0.004 mmand the associated waviness wavelength was 0.02 mm.

6. Meso machine tool design

If high speed machining is to be successful at the microand nanoscale, high spindles speeds must be employed toensure materials are processed at their recommended cuttingspeed. Popoli[91] has considered design problems of high-speed spindles and the limitations of adapting current spindledesign. An integral motor is found to be the best way ofpowering the tool. The system consists of a shaft that holdsthe tool, and bearings that hold the shaft and a method ofinternally applying the power, e.g. windings on the shaft anda direct current. For the case of an ac spindle, speed is givenby the following equation.

speed (rpm)= frequency (Hz)× 120

number of motor poles(33)

T ings.T reci-s ore,a s ares n andG ot-i d. Ah ighs mayn otingp

proxi-m ds upti me-t cano olw toolsf whent diusa pro-d yw otatet alsob It hasb r tips macros 0 m/sb ngst d

xample, a volume change is observed when machiningon. A pressure induced phase change modifies the strurom cubic to body centered tetragonal resulting in a 2enser chip. Vogler et al.[90] have also examined the diffences between macro and micro scale machining. Vogl. [90] observe the tool edge and workpiece material grecome comparable in size. The tool edge radius becoimilar size to the uncut chip thickness; this results in l

t

-

he next design issue to consider is the choice of bearhe fastest rated bearings commercially available are pion bearings; but these are limited by friction. Therefir bearings must be considered, however air bearingensitive to external debris, such as dust. Fredericksorimes[92] highlight a problem of rating spindles by qu

ng the power, which is the product of torque and speeigh power motor could be the result of low torque and hpeed or high torque and low speed; maximum powerot be available at the maximum speed. Therefore, quower yields little information about the motor.

Recommended micro scale cutting speeds can be apately 500,000 rpm, current dental drills can reach spee

o 300,000 rpm but have a run out of 10�m; a figure whichs usually greater than the chip thickness. Tools with diaers of 25�m have been used to mill at 30,000 rpm butnly achieve feeds of 5–14 in. h−1. At the macro scale toear is usually due to edge wear, at the micro scale

ail because bending strength is exceeded. This occurshe chip thickness ratio is larger than the tool’s edge rand cutting forces are large. Owing to the small chipsuced at the micro scale Zelinski[93] concluded the onlay to achieve a reasonable material removal rate is to r

he tool faster. The problem of high-speed bearings haseen encountered in the design of MEMS components.een shown that rotating micro devices must have similapeeds to their macro scale counterparts. For example,cale turbo machinery typically has tip speeds around 50ut current MEMS tip speeds are limited by MEMS beario approximately 2 m/s. Frechette et al.[94] have considere


Fig. 44. Chip formation in relation to the depth of cut and cutting edge radius[80].

this problem and designed a micro gas bearing. The rotorsits on a fluid film avoiding solid contact thereby minimiz-ing friction. Gas bearings are used to support radial motionand gas thrust bearings support axial motion (Fig. 45). Thedevice was designed to overcome viscous drag produced bythe gas bearings and fluid membrane; at 500 m/s this drag wascomputed to be 13 W. The radial bearing (or journal bearing)separates the rotor and housing, it is 300�m deep and hasan average separation of 15�m, which is maintained by apressure differential. If the rotor becomes dislodged the pres-sure differential restores its center position, which was firstdemonstrated by Orr[95]. However, the hydrostatic journalbearing acts like a spring and at certain speed coincides withthe natural frequency of the rotor producing unwanted oscil-lations. Rotating micro devices, such as these have reached

1,400,000 rpm before failure; this is equivalent to a tip speedof 300 m/s.

7. Future applications and research directions

The study of micro and nanomachining from a materialsperspective has demonstrated a new outlook on the problemsassociated with machining with conventional cutting toolsat these scales. The economies of scale can certainly begained if a number of emerging problems can be solved.One dominant problem that occurs at the microscale is thebending of the cutting tool and the inability to cut chips atlow speeds. Therefore, special attention must be applied toconstructing stiffer tools and to prevent the rounding of the


Fig. 45. Gas bearings designed and tested by Frechette et al.[94].

tool edge when the tool is initially straight. This may beovercome by coating small cutting tools with nanocrystallinediamond that has many cutting points, which are in contactwith the workpiece even when tool bending takes place. Theincompatibility between diamond and ferrous materials canbe overcome by coating the nanocrystalline diamond with athin coating of a compound that has the least thermodynamicpotential for dissolution. This direction will give rise tousing multilayered coatings that have beneficial advantages,such as thermally conducting heat away from the zone ofcutting and reducing the generation of frictional heat byusing a carbon-based soft layer lubricating coating. In termsof the construction of meso machine tools, future researchdirections include designing spindles that rotate the cuttingtool at extremely high speeds that use thin layer diamond

coatings on bearing surfaces. Air turbine spindles withintegrated gas bearings are the possible solution to achievingextremely high spindle speeds. In addition to using small-scale machine tools, machine tools must be axisymmetric inconstruction so that mechanical and thermal disturbances areminimized.

References

[1] H. Ernst, M.E. Merchant, Chip Formation, Friction and High QualityMachined Surfaces, Surface Treatment of Metals, vol. 29, AmericanSociety of Metals, New York, 1941, pp. 229–378.

[2] M. Merchant, Mechanics of the metal cutting process 1: orthogo-nal cutting and a type 2 chip, J. Appl. Phys. 16 (5) (1945) 267–275.


[3] M. Merchant, Mechanics of the metal cutting process 2: plastic-ity conditions in orthogonal cutting, J. Appl. Phys. 16 (6) (1945)318–324.

[4] V. Piispanen, Lastunmuodostumisen Teoriaa, Teknillinen Aiakausle-hti 27 (9) (1937) 315–322.

[5] M.C. Shaw, Metal Cutting Principles, second ed., Oxford Series onAdvanced Manufacturing, Oxford University Press, 2005.

[6] M. Merchant, N. Zlatin, Nomographs for analysis of metal-cuttingprocesses, Mech. Eng. 67 (11) (1945) 737–742.

[7] G.V. Stabler, Fundamental geometry of cutting tools, in: Proceedingsof the Institution of Mechanical Engineers, vol. 165, No. 63, 1951,pp. 114–121.

[8] E.H. Lee, B.W. Shaffer, Theory of plasticity applied to problem ofmachining, J. Appl. Mech. 73 (1951) 405.

[9] P.L.B. Oxley, Strain hardening solution for ‘shear angle’ in orthog-onal cutting, Int. J. Mech. Sci. 3 (1–2) (1961) 68–79.

[10] M.C. Shaw, The size effect in metal cutting, in: Proceedings of theIndian Academy of Sciences—Sadhana, vol. 28, No. 5, 2003, pp.875–896.

[11] M.C. Shaw, M.J. Jackson, The size effect of micromachining, in:Microfabrication and Nanomanufacturing, CRC Press, Taylor andFrancis Publishers, Florida, USA, 2005.

[12] H. Schulz, State of the art of high speed machining, in: A. Molinari,D. Dudzinski, H. Schulz (Eds.), High Speed Machining, Universityof Metz, France, 1997, pp. 1–8.

[13] E. Usui, A. Hirota, M. Masuko, Basic cutting model and energyapproach, Trans. ASME, J. Eng. Ind. 100 (2) (1978) 222–228.

[14] E. Usui, A. Hirota, Chip formation and cutting force with conven-tional single-point turning, Trans. ASME, J. Eng. Ind. 100 (2) (1978)229–235.

[15] E. Usui, T. Shirakashi, T. Kitagawa, Cutting temperature and crater78)

[ tingSn,

[ enttures(5)

[ .1),

[ ear

[ ear

[ tionrans.

[ ing211

[ ez,igh38–

[ tingf. 36

[ th-

[ eenond.,

[ ations.),

High Speed Machining (1, vol. 1), University of Metz, France, 1997,pp. 49–62.

[28] P.E. Gygax, Cutting dynamics and process–structure interactionsapplied to milling, Wear 62 (1) (1980) 161–184.

[29] J. Rotberg, Cutting force prediction in high speed machining the fastevaluation approach, in: A. Molinari, D. Dudzinski, H. Schulz (Eds.),High Speed Machining (1, vol. 1), University of Metz, France, 1997,pp. 63–74.

[30] I.T. Schmitt, High speed milling machines, in: A. Molinari, D.Dudzinski, H. Schulz (Eds.), High Speed Machining (1, vol. 1),University of Metz, France, 1997, pp. 75–83.

[31] M. Weck, D. Staimer, Parallel kinematic machine tools—current stateand future potentials, CIRP Ann.—Manuf. Technol. 51 (2) (2002)671–683.

[32] S. Ibaraki, T. Okuda, Y. Kakino, M. Nakagawa, T. Matsushita, T.Ando, Compensation of gravity-induced errors on a hexapod-typeparallel kinematic machine tool, JSME Int. J., Ser. C 47 (1) (2004)160–167.

[33] B. Moller, High speed and precision—features of motorised spin-dles, in: A. Molinari, D. Dudzinski, H. Schulz (Eds.), High SpeedMachining (1, vol. 1), University of Metz, France, 1997, pp. 116–128.

[34] G. Cohen, U. Ronde, Use of spindles with hydrostatic bearings inthe field of high speed cutting., in: A. Molinari, D. Dudzinski, H.Schulz (Eds.), High Speed Machining (1, vol. 1), University of Metz,France, 1997, pp. 129–141.

[35] I.N. Tansel, T.T. Arkan, W.Y. Bao, N. Mahendrakar, B. Shisler, D.Smith, et al., Tool wear estimation in micro-machining. Part 1. Toolusage-cutting force relationship, Int. J. Machine Tools Manuf. 40 (4)(2000) 599–608.

[36] S.S. Ingle, S.V. Subramanian, D.A.R. Kay, Micromechanisms ofn then-

[ ticalges.

[ Partd-8,

[ ismss ofing,

[ ools977)

[ re-and980)

[ etal38–

[ ls at

[ poly-

[ the(490)

[ tals,

[ tools,166,

[ tudy

wear of carbide tools, Trans. ASME, J. Eng. Ind. 100 (2) (19236–243.

16] T. Kitagawa, A. Kubo, K. Maekawa, Temperature, Wear of cuttools in high speed machining of inconel 718 and Ti–6Al–6V–2Wear 202 (2) (1997) 142–148.

17] T. Ozel, T. Altan, Process simulation using finite elemmethod—prediction of cutting forces, tool stresses and temperain high-speed flat end milling, Int. J. Machine Tools Manuf. 40(2000) 713–738.

18] A. Moufki, Modelling of orthogonal cutting, in: A. Molinari, DDudzinski, H. Schulz (Eds.), High Speed Machining (1, vol.University of Metz, France, 1997, pp. 8–28.

19] J.A. Bailey, Friction in metal machining—mechanical aspects, W31 (2) (1975) 243–275.

20] R.S. Montgomery, Friction and wear at high sliding speeds, W36 (2) (1976) 275–298.

21] M.P. Groover, G.E. Kane, A Continuing studyin the determinaof temperatures in metal cutting using remote thermocouples, TASME, Ser. B J. Eng. Ind. 93 (2) (1971) 603–609.

22] P.K. Wright, E.M. Trent, Metallographic methods of determintemperature gradients in cutting tools, J. Iron Steel Inst. (Lond.)(5) (1973) 364–368.

23] L. Fourment, A. Oudin, E. Massoni, G. Bittes, C. Le CalvNumerical Simulation of Tool Wear in Orthogonal Cutting. HSpeed Machining (1, vol. 1), University of Metz, 1997, pp.48.

24] K.W. Kim, H.C. Sin, Development of a thermo-viscoplastic cutmodel using finite element method, Int. J. Machine Tools Manu(3) (1996) 379–397.

25] E.M. Trent, P.K. Wright, Metal Cutting, fourth ed., ButterworHeinemann, Woburn, MA, 2000.

26] E.D. Doyle, J.G. Horne, D. Tabor, Frictional interactions betwchip and rake face in continuous chip formation, Proc. R. Soc. LSer. A 366 (1725) (1979) 173–183.

27] H.O. Gekonde, S.V. Subramanian, Influence of phase transformon tool crater wear, in: A. Molinari, D. Dudzinski, H. Schulz (Ed

crater wear, in: Proceedings of the Second Conference oBehaviour of Materials in Machining, Institute of Materials, Lodon, 1994, pp. 112–124.

37] A. Bhattacharyya, I. Ham, Analysis of Tool Wear Part 1: TheoreModels of Flank Wear, ASME-Paper 68-WA/Prod-5, 1969, 9 Pa

38] A. Bhattacharyya, A. Ghosh, I. Ham, Analysis of Tool Wear2: Applications of Flank Wear Models, ASME-Paper 69-WA/Pro1969, 6 Pages.

39] S.V. Subramanian, K. Ramanujachar, S.S. Ingle, Micromechanof tool wear in high speed machining of steel, in: Proceedingthe First Conference on the Behaviour of Materials in MachinInstitute of Materials, London, 1989, pp. 223–235.

40] Y. Naerheim, E.M. Trent, Diffusion wear of cemented carbide twhen cutting steel at high speeds, Met. Technol. 4 (12) (1548–556.

41] W.F. Hastings, P. Mathew, P.L. Oxley, Machining theory for pdicting chip geometry, cutting forces, etc. from work materialcutting conditions, Proc. R. Soc. Lond., Ser. A 371 (1747) (1569–587.

42] G. Boothroyd, Photographic technique for determination of mcutting temperature, Br. J. Appl. Phys. 12 (5) (1961) 2242.

43] O.D. Sherby, R.L. Orr, J.E. Dorn, Creep correlations of metaelevated temperatures, J. Met. 6 (1) (1954) 71–80.

44] O.D. Sherby, Factors affecting the high temperature strength ofcrystalline solids, Acta Metall. 10 (2) (1962) 135–147.

45] B. Gregory, Diffusion study of titanium carbide effectivness inwear resistance of cemented carbide tools, Metallurgia 62(1970) 55–59.

46] T. Suzuoka, Lattice and grain boundary diffusion in polycrysTrans. Jpn. Inst. Met. 2 (1) (1961) 25–33.

47] E.M. Trent, Some factors affecting wear on cemented carbidein: Proceedings of the Institution of Mechanical Engineers, vol.No. 1, 1952, pp. 64–70.

48] W. Danneberg, Selbstdiffusionsuntersuchungen an Wolfram [Sof self diffusion of tungsten], Metall 15 (10) (1961) 977–981.


[49] P.N. Nayak, N.H. Cook, Evaluation of Some Models of Ther-mally Activated Tool Wear, American Society of MechancialEngineers—Papers, 67-Prod-15, 1967, 11 Pages.

[50] C. Leymonie, P. Lacombe, Mesure de l’energie d’activationd’autodiffusion Intergranulaire du fer [Measuring activation energyof intergranular self-diffusion of iron], Memoires Scientifiques de laRevue de Metallurgie 56 (1) (1959) 74–80.

[51] R. Davies, A tentative model for the mechancial wear process, in:ASME Symposium on Friction and Wear, Detroit, USA, 1957.

[52] C. Faure, W. Hanni, C.J. Schmutz, M. Gervanoni, Diamond-coatedtool, Diamonds Relat. Mater. 8 (2–5) (1999) 636–639.

[53] M.J. Jackson, M. Gill, H. Sein, W. Ahmed, Manufacture of diamondcoated cutting tools for micromachining applications, in: Proceedingsof the Institution of Mechanical Engineers, Part L—J. Mater. 217(2003) 77–82.

[54] T. Bell, Surface engineering: its current and future impact on tribol-ogy, J. Phys. D: Appl. Phys. 25 (1A) (1992) A297–A306.

[55] B. Mills, Recent developments in cutting tool materials, J. Mater.Process. Technol. 56 (14) (1996) 16–23.

[56] O. Kubaschewski, C.B. Alcock, Metallurgical Thermochemistry(fifth), Pergamon Press, Oxford, 1979.

[57] F. Klocke, T. Krieg, Coated tools for metal cutting—features andapplications, CIRP Ann.—Manuf. Technol. 48 (2) (1999) 515–525.

[58] D.T. Quinto, A.T. Santhanam, P.C. Jindal, Mechanical properties,structure and performance of CVD and PVD coated carbide tools,Int. J. Refract. Hard Met. 8 (2) (1989) 95–101.

[59] W.D. Munz, I.J. Smith, L.A. Donohue, A.P. Deeming, R. Good-win, TiAlN based PVD coatings tailored for dry cutting operations,in: Proceedings, Annual Technical Conference—Society of VacuumCoaters, Society of Vacuum Coaters, Albuquerque, NM, 1997, pp.

[ udies159.

[ W.D.slid-. Vac.689.

[ ncelideffield

[ nal-arc

[ tionching,

[ ans-t lowol. 76

[ lms90)

[ iumom-Surf.

[ sta-gs,

[ R.ith

.[ unz,

NbyN multilayers by combined cathodic-arc/magnetron-sputter depo-sition, Thin Solid Films 302 (1–2) (1997) 179–192.

[71] E.E. Salagean, D.B. Lewis, J.S. Brooks, W.D. Munz, I. Petrov, J.E.Greene, Combined steered arc-unbalanced magnetron grown niobiumcoatings for decorative and corrosion resistance applications, Surf.Coat. Technol. 82 (1–2) (1996) 57–64.

[72] L.A. Donohue, I.J. Smith, W.D. Munz, I. Petrov, J.E.Greene, Microstructure and oxidation-resistance of Ti(l− x− y− z)AlxCryYzN layers grown by combined steered-arc/unbalanced-magnetron-sputter deposition, Surf. Coat. Technol. 94–95 (1–3)(1997) 315–321.

[73] T. Inamura, N. Takezawa, Y. Kumaki, Mechanics and energy dissi-pation in nanoscale cutting, CIRP Ann. 42 (1) (1993) 79–82.

[74] J.D. Kim, C.H. Moon, Study on the cutting mechanism of micro-cutting using molecular dynamics, Int. J. Adv. Manuf. Technol. 11(5) (1996) 319–324.

[75] L.K. Gillespie, Deburring precision miniature parts, Prec. Eng. 1 (4)(1979) 189–198.

[76] L.K. Gilespie, P.T. Blotter, Formations and properties of machiningburrs, J. Eng. Ind., Trans. ASME Ser. B 98 (1) (1973) 66–74.

[77] J.S. Kim, Optimization and Control of Drilling Burr Formation inMetals (Doctoral Dissertation, University of California at Berkeley),Dissertation Abstracts International, 62(01) (2000) 496B (UMI No.3001895).

[78] S.L. Ko, D.A. Dornfield, American Society of Mechancial Engineers,Dynamics and Control Division DSC, Study Burr Form. Mech. 11(1988) 271–282.

[79] K. Lee, D.A. Dornfeld, An Experimental Study of Burr Formationin Micro Milling Aluminum and Copper, Technical Paper—Societyof Manufacturing Engineers, MR(MR02-202), 2002, pp. 1–8.

[80] N. Ikawa, S. Shimada, H. Tanaka, Minimum thickness of cut in

[ s onpre-–330.

[ a-ng

[ rated8.

[ is ofdia-

554.[ t-

ign96)

[ ecu-sses. 59

[ ssesgi-

[ micsSPIE,

[ ics81

[ vel-nical002,

[ ech-98-

83–93.60] C.P. Constable, J. Yarwood, W.D. Munz, Raman microscopic st

of PVD hard coatings, Surf. Coat. Technol. 116–119 (1999) 155–61] C.P. Constable, J. Yarwood, P. Hovsepian, L.A. Donohue,

Munz, Structural determination of wear debris generated froming wear tests on ceramic coatings using raman microscopy, JSci. Technol., Part A: Vac. Surf. Films 18 (4–11) (2000) 1681–1

62] A.P. Deeming, W.D. Munz, I.J. Smith, Dry High PerformaMachining (HPM) of Die and Moulds Using PVD Coated SoCemented Carbide Tools, Paper Presented at the Meeting of ShP.V.D. Research Group, 2001.

63] S. Creasey, D.B. Lewis, I.J. Smith, W.D. Munz, SEM image aysis of droplet formation during metal ion etching by a steereddischarge, Surf. Coat. Technol. 97 (1–3) (1997) 163–175.

64] W.D. Munz, I.J. Smith, D.B. Lewis, S. Creasey, Droplet formaon steel substrates during cathodic steered arc metal ion etVacuum 48 (5) (1997) 473–481.

65] R. Gahlin, M. Bromark, P. Hedenqvist, S. Hogmark, G. Hakson, Properties of TiN coatings and CrN coatings deposited atemperature using reactive arc-evaporation, Surf. Coat. Techn(1–3) (1995) 174–180.

66] D.D. Wang, T. Oki, The morphology and orientation of Cr–N fideposited by reactive ion plating, Thin Solid Films 185 (2) (19219–230.

67] T. Hurkmans, D.B. Lewis, J.S. Brooks, W.D. Munz, Chromnitride coatings grown by unbalanced magnetron (UBM) and cbined arc/unbalanced magnetron (ABS) deposition techniques,Coat. Technol. 86–87 (1–3) (1996) 192–199.

68] I. Wadsworth, I.J. Smith, L.A. Donohue, W.D. Munz, Thermalbility and oxidation resistance of TiAlN/CrN multilayer coatinSurf. Coat. Technol. 94–95 (1–3) (1997) 315–321.

69] I.J. Smith, D. Gillbrand, J.S. Brooks, W.D. Munz, S. Harvey,Goodwin, Dry cutting performance of HSS twist drills coated wimproved TiAlN, Surf. Coat. Technol. 90 (1–2) (1997) 164–171

70] I. Petrov, P. Losbichler, D. Bergstrom, J.E. Greene, W.D. MT. Hurkmans, et al., Ion assisted growth of Ti(1− x)AlxN/Ti(1 − y)

micromachining, Nanotechnology 3 (1) (1992) 6–9.81] Z.J. Yuan, M. Zhou, S. Dong, Effect of diamond tool sharpnes

minimum cutting thickness and cutting surface integrity in ultracision machining, J. Mater. Process. Technol. 62 (4) (1996) 327

82] C.J. Kim, M. Bono, J. Ni, Experimental Analysis of Chip Formtion in Micro Milling, Technical Paper—Society of ManufacturiEngineers, MR(MR02-159), 2002, pp. 1–8.

83] J.W. Sutherland, T.S. Babin, The geometry of a surface geneby the bottom of an end mill, Proc. NAMRC 16 (1988) 202–20

84] N. Ikawa, S. Shimada, H. Tanaka, G. Ohmori, Atomistic analysnanometric chip removal as affected by tool-work interaction inmond turning, CIRP Ann., Manuf. Technol. 40 (1) (1991) 551–

85] H. Mizumoto, S. Arii, A. Yoshimoto, T. Shimizu, N. Ikawa, Twisroller friction drive for nanometer positioning—simplified desusing ball bearings, CIRP Ann., Manuf. Technol. 45 (1) (19501–504.

86] S. Shimada, N. Ikawa, H. Tanaka, G. Ohmori, J. Uchikoshi, Mollar dynamics analysis of cutting force and chip formation procein microcutting, Seimtsu Koaku Kaishi/J. Jpn. Soc. Prec. Eng(12) (1993) 2015–2021.

87] S. Shimada, Molecular dynamics simulation of the atomic procein micromachining, in: J. McGeough (Ed.), Micromachining of Enneering Materials, Marcel Deckker, New York, 2002.

88] S. Shimada, R. Inoue, J. Uchikoshi, N. Ikawa, Molecular dynaanalysis on microstructure of diamond turned surfaces, Proc.Int. Soc. Opt. Eng. 2576 (1995) 396–405.

89] R. Komanduri, I. Chandrasekaran, L.M. Raff, Molecular dynamsimulation of the nanometric cutting of silicon, Philos. Mag. B(12) (2001) 1989–2019.

90] M.P. Vogler, X. Liu, S.G. Kapoor, R.E. DeVor, K.F. Ehmann, Deopment of Meso-Scale Machine Tool (mMT) Systems, TechPaper—Society of Manufacturing Engineers, MS(MS02-181), 2pp. 1–9.

91] W.F. Popoli, High Speed Spindle Design and Construction, Tnical Paper—Society of Manufacturing Engineering, MR(MR146)(1998) 23 Pages.


[92] P. Fredrickson, D. Grimes, Optimizing motor technology for spindleapplications, Motion Syst. Des. 46 (11) (2004) 24–32.

[93] P. Zelinski, Micro Milling at 1/2 Million RPM, Modern MachineShop, 1 (2003). Retrieved 10th March 2004. fromwww.mmsonline.com/articles/080306.html.

[94] L.G. Frechette, S.A. Jacobson, K.S. Breuer, F.F. Ehrich, R. Ghodssi,R. Khanna, et al., Demonstration of a Microfabricated High-Speed

Turbine Supported on Gas Bearings, Technical Digest. Solid-StateSensor and Actuator Workshop, TRF Cat No. 00TRF-0001, 2000,pp. 43–47.

[95] D.J. Orr, Macro-Scale Investigation of High Speed Gas Bearingsfor MEMS Devices, Doctoral Dissertation, Massachusetts Instituteof Technology, 2000.

http://www.mmsonline.com/articles/080306.html

http://www.mmsonline.com/articles/080306.html

_a review of micro and nano machining from a materials perspective

Documents