Ad

HAa

b

c

d

e

a

ARAA

KDADS

1

mcucctot(sif

1

0d

Journal of Materials Processing Technology 212 (2012) 523– 533

Contents lists available at SciVerse ScienceDirect

Journal of Materials Processing Technology

jou rna l h om epa g e: www.elsev ier .com/ locate / jmatprotec

dhesion analysis and dry machining performance of CVD diamond coatingseposited on surface modified WC–Co turning inserts

umberto Gomeza,e, Delcie Durhama, Xingcheng Xiaob,∗, Michael Lukitschb, Ping Luc, Kevin Chouc,nil Sachdevb, Ashok Kumara,d,∗∗

Department of Mechanical Engineering, University of South Florida, Tampa, FL 33620, USAChemical Sciences & Materials Systems Laboratory, General Motors R&D Center, 30500 Mound Road, Warren, MI 48090, USAMechanical Engineering Department, The University of Alabama, Tuscaloosa, AL 35487, USANanotechnology Research and Education Center (NREC), University of South Florida, Tampa, FL 33620, USADepartamento de Ingeniería Mecánica, Universidad del Norte, Barranquilla, Colombia

r t i c l e i n f o

rticle history:eceived 11 October 2011ccepted 21 October 2011vailable online 26 October 2011

eywords:iamonddhesionry machining

a b s t r a c t

This paper investigates the effects of different surface pretreatments on the adhesion and performanceof CVD diamond coated WC–Co turning inserts for the dry machining of high silicon aluminum alloys.Different interfacial characteristics between the diamond coatings and the modified WC–Co substratewere obtained by the use of two different chemical etchings and a CrN/Cr interlayer, with the aim toproduce an adherent diamond coating by increasing the interlocking effect of the diamond film, andhalting the catalytic effect of the cobalt present on the cemented carbide tool. A systematic study isanalyzed in terms of the initial cutting tool surface modifications, the deposition and characterizationof microcrystalline diamond coatings deposited by HFCVD synthesis, the estimation of the resulting

urface modification diamond adhesion by Rockwell indentations and Raman spectroscopy, and finally, the evaluation of thedry machining performance of the diamond coated tools on A390 aluminum alloys. The experimentsshow that chemical etching methods exceed the effect of the CrN/Cr interlayer in increasing the diamondcoating adhesion under dry cutting operations. This work provided new insights about optimizing thesurface characteristics of cemented carbides to produce adherent diamond coatings in the dry cuttingmanufacturing chain of high silicon aluminum alloys.

. Introduction

There is an uprising trend in the dry machining of compositeetals such as aluminum–silicon alloys and aluminum matrix

omposites as the result of the environmental impact of coolantssed in traditional wet machining operations. The amount ofoolants disposed in the form of mist, waste, and coolant-coatedhips have been reported to produce a significant harmful effect tohe environment (Adler et al., 2006). The estimated global marketf over $1500 million in 2007 (increasing 6% annually) relatedo equipment used in filtration and separation of cutting fluidsSutherland, 2008), provides an important reason to companies in

eeking new strategies to reduce fluids consumption by using min-mum quantity lubrication (MQL) systems or remove them entirelyrom the machining operation. Additionally, the development of

∗ Corresponding author. Tel.: +1 248 912 8132.∗∗ Corresponding author at: University of South Florida, 4202 East Fowler Ave, ENB18, Tampa, FL, USA. Tel.: +1 813 974 3942; fax: +1 813 974 3610.

E-mail addresses: xingcheng.xiao@gm.com (X. Xiao), kumar@usf.edu (A. Kumar).

924-0136/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.jmatprotec.2011.10.020

© 2011 Elsevier B.V. All rights reserved.

new automated high-speed machine centers and the use of novelcomposite materials in complex designs, create new challenges forcoated cutting tools that need to function under these aggressivemachining conditions.

Roy et al. (2009) found that the chemical inertness of CVD dia-mond coatings is the key factor to enhancing the performance ofcutting tools in the dry machining of Al–Si alloys, outperforminguncoated tools along AlON, TiC, TiB2, TiN, and Al2O3 coated tools.Particularly, aluminum alloys are very abrasive and extremely dif-ficult to dry machine with conventional TiN PVD coated materialsdue to the formation of built-up layer (BUL) or built-up edge (BUE)over the rake surface of the tool as concluded by Gangopadhyayet al. (2010).

Köpf et al. (2006) discussed the initial substrate pretreatmentsrequired to deposit adherent diamond coatings in WC–Co tools forthe machining of non-ferrous metals and fiber reinforced plastics.However, the machining performance of diamond coated tools

is not yet robust due to a non-optimized adhesion between thecarbide tool and the deposited diamond film. The insufficientdiamond adhesion with the cutting tool substrates would renderthem inadequate or lead to unpredictable behavior and even

5 Proces

pocpsc

fcdsbCn

mdiwbaitdttc

Wdadloca

muWd

ttdtlept

uutcetipdc

lisc

24 H. Gomez et al. / Journal of Materials

ossible catastrophic failure during dry cutting operations. More-ver, the adhesion and machining performance of CVD diamondoated tools need to be optimized based upon considering thearticular manufacturing chain in terms of the substrate condition,urface pretreatments, workpiece materials, and cutting operationonditions (Uhlmann and Koenig, 2009).

Carbides enriched with 3–13 wt% of cobalt binder provide highracture toughness and are the most common substrate materials inoated tools used for dry machining applications. However, wheniamond is deposited on these substrates, the diamond (carbon)olubility of 0.2–0.3 wt% in cobalt degrades the adhesion of the filmy forming a graphitic layer at the interface at the conventionalVD deposition temperatures of ∼800 ◦C, which prevents diamonducleation.

Several studies have been reported on enhancing the growthechanisms and behavior of adherent diamond films deposited on

ifferent kinds of substrates, however, the majorities do not takento account the practical substrate surface conditions encountered

hen using commercial carbide tools existing in the market. It haseen shown by Li and Hirose (2007) that film and substrate operates a “composite” system and the interface between them plays anmportant role in the durability of the coating. In order to optimizehe final performance of diamond coated tools under the harsh con-itions developed during dry machining operations, the resultingribological interface must be understood as a system in terms ofhe fundamental coating adhesion and the wear mechanisms at theutting edge of the tool.

A systematic study between the shape and characteristics of theC–Co cutting tool, the CVD diamond deposition process, and the

ry machining parameters, was proposed by Chou et al. (2010) asn optimal approach to achieve adherent diamond coatings for dryrilling applications. Haubner and Kalss (2010) concluded that the

ifetime of diamond coating tools is influenced by the interactionf many factors. As a consequence, an optimization of diamondoated tool performance is needed for particular manufacturingpplications.

With the aim to analyze the dry cutting behavior of CVD dia-ond coated tools in specific manufacturing chains, this study

tilized different surface pretreatments applied to commercialC–Co 6% turning inserts to modify their surfaces prior to the

iamond deposition.Substrate pretreatments were focused on providing a surface

hat facilitates the diamond film interlocking effect and at the sameime eliminates the effect of the cobalt by increasing nucleationensity, both of which are reported to improve film adhesion andool life. Two chemical etching methods and a CrN/Cr buffer inter-ayer were evaluated in the present study. The pretreated tools willxhibit differences in their substrate surface textures and integrity,roviding different interfacial characteristics with a direct effect onhe diamond adhesion and dry machining performance.

The wear behavior and adhesion improvement of diamond bysing a Cr–N interlayer was evaluated by Glozman et al. (1999),sing a fretting test rig and compared with indentation and scratchests. Tribological and mechanical properties of HFCVD diamondoatings deposited on WC–Co substrates with different Cr interlay-rs were measured by indentation techniques and correlated withhe substrate roughness and hardness by Chou et al. (2008). Flu-dized Bed modified Cr/CrN interlayers on WC–Co substrates wereroposed by Polini et al. (2010), as a mechanism to enhance theiamond film nucleation by forming a highly adherent diamondoating and evaluated by dry ‘pin-on-disk’ tribological tests.

Of particular interest is the behavior of the CrN/Cr inter-

ayer when compared with chemical etching as diamond adhesionmprovement methods in dry machining conditions. In the presenttudy, diamond coatings were deposited using the same growthharacteristics and thicknesses (25–30 �m) than the commercial

sing Technology 212 (2012) 523– 533

microcrystalline diamond (MCD) tools found in the market. Theadhesion characteristics of the diamond coated tools were eval-uated by indentation techniques and Raman spectroscopy, andcompared with the diamond wear failure under a particular dryturning machining operation on A390 aluminum workpiece. It wasfound that chemical etching methods surpassed the effect of theCrN/Cr interlayer in increasing the diamond coating adhesion dur-ing dry cutting operations. The wear failure mechanism of theCrN/Cr interlayer on the diamond adhesion is discussed, thus, pro-viding insights about the use and optimization of these interlayersto increase the dry machining performance of WC–Co diamondcoated tools.

2. Materials and experiments

Commercial tungsten carbide–6 wt% cobalt square positiveangle turning inserts (WC–6%Co/SPG-422) were used as tool sub-strates for further MCD coating depositions. Different surfacepretreatments were selected in order to evaluate the most reportedcommon technical approaches, which have been claimed to be ableto overcome the detrimental effect of the cobalt binder, includ-ing the removal of the cobalt at the surface by chemical etching(Zhang et al., 2000) and by pre-depositing an inter-diffusion bar-rier layer (Polini and Barletta, 2008) to suppress the interactionbetween cobalt and carbon.

2.1. Chemical etching

WC–Co inserts were first cleaned in an acetone ultrasonic bathfor 10 min, followed by an ultrasonic rinse for 5 min in methanolto remove any contamination from previous processes. After that,samples were ultrasonically treated with Murakami’s solution(1:1:10 KOH + K3[Fe(CN)6] + H2O) for 10 min and then rinsed withdeionized water. In Method E-1, Murakami’s step was followedby immersion of the samples in an ultrasonic bath containing 10%HNO3 + 90% H2O2 for 60 s. Method E-2 included the initial etchingwith the same Murakami’s solution followed by immersion in anultrasonic bath containing 3 ml of H2SO4 and 88 ml of H2O2 for 60 s.After chemical etching, samples were ultrasonically rinsed withdeionized water and dried with nitrogen gas.

2.2. Cobalt inter-diffusion barrier interlayer

A buffer interlayer was deposited to prevent the diffusion ofcarbon into the underlying cobalt binder phase and to act as astress relaxation layer to reduce the thermal expansion coefficientmismatch between the diamond and the substrate material as con-cluded by Sarangi et al. (2008). In this work, as-received WC–Co (6%)inserts were coated with an initial layer of CrN (1.5 �m) followed bya top layer of Cr (1.5 �m) using a commercial cathodic-arc PhysicalVapor Deposition (PVD) system. Then, the interlayers were furtherroughened to promote diamond nucleation and interlocking of thediamond film. Two methods were used to nucleate the diamond:Method I-1 is an additional step of media blasting on the top Crsurface during 1 min with 50 �m diamond particles at a pressureof 40 psi as reported by Xu et al. (2007); Method I-2 is an additionalsurface scratching process to the Cr interlayer during 60 min in theultrasonic bath containing a solution of 2.4 g of 50 �m diamondpowders dispersed in 50 ml of methanol.

The surface roughness of the substrates and the morphologybefore and after each surface pretreatment were measured by aVecco Wyko-NT9100 white-light interferometer and a Hitachi S-

800 Scanning Electron Microscope in conjunction with an EnergyDispersed Spectroscopy (EDS) system attached to the SEM. Thecross section before and after each treatment was prepared by dic-ing the inserts with a refrigerated diamond saw, then followed by

H. Gomez et al. / Journal of Materials Processing Technology 212 (2012) 523– 533 525

Table 1Surface modification treatments applied to the WC–Co(6%) turning inserts.

Surface denomination Treatment characteristics

As-received (AG) Ultrasound cleaning in acetone and methanolMurakami (M) Ultrasound bath in (1:1:10 KOH + K3[Fe(CN)6] + H2O)Method E-1 Murakami + ultrasound bath in 10% HNO3 + 90% H2O2

Method E-2 Murakami + ultrasound bath in a solution of 3 ml ofH2SO4 and 88 ml of H2O2

PVD coated (I-AG) Deposition of 1.5 �m of CrN and 1.5 �m of Cr (top)

ge2mtKfip

2

HpoafabD4u2waodosfKwtwb

2

naissFpposortpi

Method I-1 PVD coated + shoot peening (diamond powder)Method I-2 PVD coated + scratching in diamond solution

rinding and polishing the resulting surface mounted in conductivepoxy resin with an automatic preparation Struers Prepmatic-

system. Supplementary SEM analysis was conducted on theetallographically prepared samples by immersing the cross sec-

ioned samples for 5 min in a solution of Murakami reagent (10 g3Fe(CN)6 + 10 g NaOH in 100 mL H2O) in order to reveal the modi-ed cemented carbide microstructure. All abovementioned surfaceretreatments are summarized in Table 1.

.3. Diamond deposition and characterization

Commercial diamond films were synthesized in the SP3 M650ot Filament CVD system using hydrogen and methane as gasrecursors at a pressure of ∼40 Torr and a substrate temperaturef ∼850 ◦C to form a continuous microcrystalline diamond filmpproximately 25 �m for all pretreated inserts. The diamond sur-ace characteristics were measured by white-light interferometernd SEM analysis. X-ray diffraction (XRD) patterns were recordedefore and after the diamond deposition with a Bruker-AXS D8iscover diffractometer (Cu K� wavelength ∼1.544 A) operated at0 kV and 40 mA. Data were collected between 25◦ and 85◦ 2�,sing an integration time of 7 s per step and a step size of 0.01◦

�. Raman spectroscopy of the diamond coated-pretreated toolsas performed with a Renishaw 1000 Raman spectrometer with

n Argon laser at a wavelength of 514.5 nm, and a laser spot sizef 1 �m at a power of 25 mW. Rockwell C indentations were con-ucted with a Wilson/Rockwell Hardness Tester (Series 500) tobtain a qualitative representation of the diamond coating adhe-ion at different load levels. The resulting diamond delaminationsrom the center of the indentation zone were measured with aeyence VHX-500 digital microscope while the fracture patternsere analyzed with an SEM. The film compositions in the indenta-

ion zone were determined by electron probe microanalysis (EPMA)ith a Cameca Instruments model SX100 electron probe. Electron

eam conditions were typically 15 kV and 40 nA during the analysis.

.4. Dry machining performance test

Dry machining experiments were performed using a computerumerical control (CNC) lathe Hardinge Cobra 42 equipped with

Kistler dynamometer (9257B) to monitor the cutting forces dur-ng the machining, and a Kistler 8152B piezotron acoustic emissionensor to collect AE-RAW (raw data) and AE-RMS (root-mean-quare) values at a sampling rate of 500 kHz during the machining.ig. 1 shows the dry machining performance test setup. Work-ieces were round bars made of A390 aluminum alloy. Machiningarameters were kept constant at a cutting speed of 10 m/s, feedf 0.8 mm/rev, and a depth of cut of 1 mm. These conditions wereelected in accordance to previous experiments performed by onef the coauthors on diamond coated tools (Hu et al., 2007a) which

epresent the most aggressive parameters in terms of wear andool life out of a set of machining conditions on a similar work-iece material. During the machining test, the diamond coated

nserts were periodically inspected to measure the flank wear-land

Fig. 1. Dry machining performance setup used for the MCD coated tools experi-ments.

with optical microscopy. Worn tools after the machining tests werecleaned with 10 vol% hydrochloric acid to remove the aluminumalloy deposited on the cutting area and then examined by SEM andEPMA.

3. Results and discussion

3.1. Pretreated surfaces before diamond deposition

Fig. 2 depicts the surface maps and roughness parametersobtained by white-light interferometry corresponding to the sur-face of as-received WC–Co (6%) commercial turning inserts (AG)and the surface after the Murakami treatment (M), methods E-1, E-2, I-1, and I-2, respectively. Roughness parameters values such as Rz,Rt, Rp, and Ra were recorded and averaged from six measurementson different top surface points. The scan size was determined by theoptical magnification of the system (50×) in a resulting surface areaof 126 �m × 94 �m containing 640 × 480 data points. These val-ues were kept constant for all roughness measurements in order todecrease the systematic error in the resulting roughness parametervalues as suggested by Poon and Bhushan (1995).

The surface map AG represents the as-ground original surfacedepicting the directional marks (feed marks) left by the grindingprocess on the surface of the tool, which also confers a non uni-form texture in terms of the distribution of the WC grains and theCo binder due to the amount of surface damage after the grindingprocess. These feed marks display an average surface roughness(Ra) of 0.16 �m measured perpendicular to them across a lengthof 126 �m. Murakami (M) treated surfaces displayed an increasedsurface roughness parameter Ra of 0.24 �m with a partial removalof the original surface features. Previous experiments revealed anincrease in the surface roughness with an increase in the exposuretime in the Murakami solution; Ra values of 0.45 �m were obtainedfor 30 min under this treatment.

The surface topography after treatment method E-1 revealsa uniform surface with a Ra value of 0.51 �m, with a completeremoval of feed marks compared with Method E-2, which exhibitssome directional surface features with a corresponding Ra valueof 0.45 �m. The increase in roughness after the acid treatment formethods E-1 and E-2 is attributed to the Murakami initial treat-

ment, which attacks the WC grains, thus reconstructing the initialsurface, while the final etching of the acid (for both etching meth-ods) oxidizes the cobalt binder to soluble Co2+ compounds thatget washed out during ultrasonication (Polini, 2006). This washing

526 H. Gomez et al. / Journal of Materials Processing Technology 212 (2012) 523– 533

F ceives I-2, re

apa

cafpoo(Cfep

psm

vosrcfifw

tt

ig. 2. Surface maps obtained by white-light interferometry corresponding to as-reurface textures after treatment method E-1, method E-2, method I-1, and method

ction may also leach weakly bonded WC grains from the surfaceromoting an additional reconstruction at the surface and creating

new surface texture.The effect of media blasting and diamond scratching of the

hromium after the deposition of the initial PVD CrN/Cr interlayerre also characterized in Fig. 2. The PVD process produced a sur-ace (surface map not shown) that still displays features with areferential direction resulting from the conformal PVD coatingver the initial feed marks in the as-ground surface. Furthermore,ther features resulting from the original cathodic-arc PVD processmicrodroplets) were observed and will be discussed later in Fig. 4c.onsequently, the roughness value for the PVD top chromium inter-

erometry surface map are considerably affected by the combinedffect of these features, and are not suitable for a quantitative com-arison purpose.

Method I-1 produced a uniform surface roughness with a Ra

arameter of 0.28 �m, few visible directional marks (at a biggerurface area scale of 250 �m × 190 �m) and complete removal oficrodroplets on a microscale.Method I-2, on the other hand, generated a surface with a Ra

alue of 0.11 �m, with less Cr microdroplets compared with theriginal PVD coating, and did not completely reconstruct the topurface in terms of the initial as-ground feed marks, with resultingoughness parameter values close to the original as-ground surfaceharacteristics, as shown in Fig. 3. The data presented in this lastgure summarizes all surface roughness parameters for each sur-

ace denomination, including the maximum and minimum values

ith respect to the average.

Fig. 4a–d is SEM micrographs of the substrate top surfaces andheir cross sections after metallographic preparation. Fig. 4a showshe feed marks on the surface of the tool, while the microstructure

d WC–Co (6%) turning inserts topography (AG), Murakami treated surface (M), andspectively. The surface area is 126 �m × 94 �m containing 640 × 480 data points.

(insert) of the cemented carbide corresponds to the WC grains withan average diameter of 1.4 �m embedded in the Co matrix. Fig. 4bshows the surface reconstruction after pretreatment E-1 with a Coremoval depth of ∼9.5 �m along the cross section of the tool.

The surface modification after pretreatment E-2 is shown inFig. 4c, and shows a cobalt binder depletion band of ∼8.0 �m belowthe cross section. It can be seen that both the E-1 and E-2 pre-treatments, reconstruct the surface at the microscale in terms ofeliminating the directional features from the as-received sample(in a rougher surface for method E-1), which are detrimental tothe final adhesion of the diamond coating due to the debonding ofthe film that will occur along the preferential direction of the feedmarks. The microstructure morphology of the chemically etchedsamples depicts the removal of the cobalt binder, which producessingle WC grains in the cross section of the tool for different Coremoval rates. The CrN/Cr interlayer surface deposited on top ofthe as-received WC–Co inserts, including its cross section, is shownin Fig. 4d, which depicts a uniform coating with the presence ofCr particles (microdroplets) entrained on the surface, typical for acathodic arc-PVD process described by Warcholinski and Gilewicz(2009), with heights ranging from 1.2 to 4.5 �m and diametersranging between 1.5 and 3.0 �m based on interferometer measure-ments. The cross section (inset) shows the conformal depositionand morphology of both, CrN and Cr interlayers on top of the WC–Cosubstrate.

3.2. Diamond coated surfaces after the surface pretreatments

Diamond, with a coating thickness of approximately 25 �m,was deposited on the different surface modified substrates.The diamond coatings on samples corresponding to method I-2

H. Gomez et al. / Journal of Materials Processing Technology 212 (2012) 523– 533 527

Fig. 3. Roughness characteristics for each surface designation represented by the ten-point height (Rz), maximum peak-to-valley height (Rt), highest peak (Rp), and arithmeticalmean deviation (Ra) texture roughness parameters.

Fig. 4. (a) SEM micrograph sowing the finishing feed-marks at the surface of WC–Co (6%) as received turning inserts and the WC grains distribution (inset) in the Co binder.(b) SEM micrograph at the surface of substrates after E-1 pretreatment including the Co depletion layer (inset) in the cross section. (c) SEM micrograph at the surface ofsubstrates after E-2 pretreatment including the Co depletion layer (inset) in the cross section. (d) SEM micrograph at the surface of substrates after the CrN/Cr PVD depositionand the cross section (inset) showing the layer interfaces and morphology.

528 H. Gomez et al. / Journal of Materials Processing Technology 212 (2012) 523– 533

Fig. 5. Morphology (SEM) of HFCVD grown diamond coatings deposited on surfacesp

dwIacd〈sata6cn

Fig. 6. XRD patterns of as-ground WC–Co (6%) samples (AG) prior to diamond depo-

(hexagonal) {1 0 1} peak normally present at ∼2� − 47.0◦ due to the

retreated samples corresponding to methods E-1, E-2, and I-1.

elaminated from the substrate immediately after deposition,hile no delamination was observed for samples E-1, E-2, and

-1. The surface characteristics of the various diamond coatingsre shown in SEM micrographs in Fig. 5. The diamond surfaceorresponding to pretreatment method E-1 shows a continuousiamond film with well defined polycrystal facets due to the1 0 0〉{1 1 1} texture on the surface. Optical interferometry mea-urements showed that the grains were ∼2–4 �m and had anverage roughness Ra of 0.50 �m. The surface characteristics ofhe diamond coatings on substrates from method E-2 displayed

smaller diamond grain size ∼1–2 �m with a roughness Rz of

.53 �m, Rt of 2.70 �m, Rp of 1.35 �m, and Ra of 0.38 �m. Theoating had sub-micron facetted crystals and a greater amount ofon-diamond carbon phases.

sition, diamond coated sample after method E-1, diamond coated sample aftermethod E-2, PVD CrN/Cr coated sample (I-AG), and diamond coated sample aftermethod I-1.

The diamond coatings on samples I-1 consisted of a combinationof small crystal facets in ball-like agglomerated deposits, a crystalsize of ∼1 �m, and roughness parameter values Rz of 4.80 �m, Rt of2.43 �m, Rp of 1.24 �m, and Ra of 0.31 �m. A comparison betweenthe surface morphology and roughness parameters of coated sam-ples from the three methods E-1, E-2 and I-1 confirms that thefinal roughness and morphology of the diamond coatings are cor-related to the initial roughness and characteristics of the substrateas conclude by Mallik et al. (2010). The higher Rz value (areal sur-face data corresponding to the average absolute value of the fivehighest peaks and the five lowest valleys) obtained in coated sam-ples E-2 and I-1, compared with sample E-1, confirms that diamondcoatings deposited on pretreated surfaces with methods E-2 and I-1partially contain some of the surface features related to the originalfeed marks of the as-ground (AG) sample, and are visible (figuresnot shown) in the interferometry results of the diamond coatedsurface texture maps.

Fig. 6 shows the XRD patterns of the as-ground WC–Co (6%)samples (AG) prior to diamond deposition, and following diamondcoating with methods E-1, E-2, PVD CrN/Cr coating (I-AG), andafter method I-1. These XRD patterns present a significant peakdue to the WC (hexagonal) {1 0 1} reflection with a characteristichigh intensity peak at 2� = 48.2◦, which overlaps with the �-cobalt

scattering efficiency of W compared with that of cobalt, which isonly present at a level of 6% (Mallika and Komanduri, 1999). How-ever, Co peaks corresponding to the {1 1 1} and {1 0 1} reflections

H. Gomez et al. / Journal of Materials Processing Technology 212 (2012) 523– 533 529

F

ctaAtW1ioT

ig. 7. The Raman spectrum of samples E-1, E-2, and I-1 after diamond deposition.

an be distinguished at 2� = 44.4◦ and 2� = 76.9◦, respectively. Addi-ionally, the Co {1 1 1} reflection may overlaps with the peak from

possible presence of Cr in the AG sample, detected in sampleG during initial EPMA measurements, which is normally added

o the Co binder as Cr3C2 during the sintering process to limit theC grain growth (Okada et al., 1998). After deposition, samples E-

and E-2 show diamond peaks corresponding to the {1 1 1} that

s present at 2� = 43.85◦ and the {2 2 0} at 2� = 75.35◦ that mightverlap with the WC peak at 2� = 75.2◦ shown in the AG sample.he small Co peak present in samples E-1 and E-2 may be due to

Fig. 9. Backscattering SEM micrographs (left column) and W, Cr, and C Ka X-ray ma

Fig. 8. Lateral crack lengths present in diamond coated samples E-1, E-2, and I-1,resulting from discrete indentations levels at 45, 60, 100, and 150 kg.

small amounts of the binder as Co–Cr stable compound (Delanoëet al., 2004), not removed by the etching process at some points,providing some binding to the WC grain. The XRD spectrum of sam-ple I-AG shows broad peaks attributable to chromium nitride at2� = 38.2◦, 44.4◦, 64.65◦, and 77.65◦. Diamond coated samples afterpretreatment method I-1show XRD patterns with additional peaks

representing Cr3C2 and Cr7C3, suggesting intermediate chromiumcarbide compounds formed during the diamond depositionprocess.

ppings (center and right columns, respectively) for samples E-1, E-2, and I-1.

530 H. Gomez et al. / Journal of Materials Processing Technology 212 (2012) 523– 533

Fp

mspat2spetetcmpsfalv

c4altctd1daalTs14

smc

ig. 10. Flank wear-land (VB) time evolution of diamond coated turning insertsretreated under methods E-1, E-2, and I-1.

Raman spectra of diamond films deposited on the three afore-entioned pretreated samples are shown in Fig. 7. The Raman

pectrum of MCD samples E-1, E-2, and I-1 show sharp diamondeaks centered at 1335.1, 1336.6, and 1333.5 cm−1, respectively,nd shifted about 2.7 cm−1, 4.2 cm−1, and 1.1 cm−1 with respecto natural diamond (1332.4 cm−1 at atmospheric pressure and5 ◦C). These results correspond to biaxial residual compressivetresses of 2.91, 4.53, and 1.19 GPa, calculated using the approachroposed by Ager and Drory (1993) and used by Cristofanillit al. (2010) for diamond coatings deposited on WC–Co substratesreated with CrN interlayers. The obtained thermal stresses arexpected to be the same due to a uniform diamond depositionemperature for all samples by using a substrate heater and aonstant relative position of the samples with respect to the fila-ents during the HFCVD deposition process. The residual stresses

resent in the samples may be solely attributed to the intrinsictresses induced during the growth phase, responsible for the dif-erent diamond morphologies shown in Fig. 5. Furthermore, anpparent stress relief mechanism provided by the CrN/Cr inter-ayer can be observed from the lower biaxial compressive stressalue.

Diamond coatings adhesion was evaluated in terms of the lateralrack lengths resulting from discrete indentations levels at loads of5, 60, 100, and 150 kg. Three single indentations were performedt each load to ensure a delamination in the films, and ten crackength measurements were recorded from the center of each inden-ation. When the indentation force was sufficiently high, lateralracks were initiated, which propagated between the coating andhe substrate. Fig. 8 shows the average lateral crack lengths in theiamond coated samples for the various pretreatment methods E-, E-2, and I-1. Crack lengths were higher for sample E-1 and noelamination was observed at some load levels for samples E-1 (60nd 45 kg) and E-2 (45 kg). In contrast, delamination was observedt all indentation load levels for sample I-1, which displayed simi-ar crack length values with respect to sample E-2 at 60 and 100 kg.he relative error (with respect to the average) is different for allamples at each discrete indentation load, i.e. ∼10% for sample E-

at 150 kg, 7% for sample E-2 at 60 kg, and 15% for sample I-1 at5 kg.

Moreover, when no fracture or delamination was observed atome indentation loads for samples E-1 and E-2, white annulararks appeared on the diamond surface, which suggest that the

oating was still not adherent, and that the crack energy was

Fig. 11. SEM micrographs of the worn samples after the machining test.

insufficient to promote a complete delamination. As a result, itmay be inappropriate to make a quantitative comparison aboutthe interfacial toughness of the diamond coatings for the differentsurface pretreatment effects based on the data from Fig. 8, bycalculating the critical strain energy release rate Gc for crackpropagation as proposed in previous works as proposed by Droryet al. (1995). However, it is quite interesting to find that for sampleE-1, perhaps it has the highest value of crack lengths for a corre-sponding load of 150 and 100 kg, no delamination and no annularmarks appeared after indentations at load levels of 60 and 45 kg.Contrary, diamond delamination was achieved for all indentationloads in sample I-1 in conjunction with lower lateral crack lengths.

Fig. 9 presents the backscattered SEM micrographs (left col-

umn) of samples E-1, E-2, and I-1 after the Rockwell indentations,and shows the white spallation area of the diamond coatingsand the lateral cracks when an indentation load of 100 kg wasapplied to all samples. The Ka X-ray maps from W for samples

H. Gomez et al. / Journal of Materials Processing Technology 212 (2012) 523– 533 531

F s andc

Es

tlsbEditdcttifah

3

danlo

ig. 12. Digital microscopy images (left column), backscattering SEM micrographorresponding to the worn cutting edges of samples E-1, E-2, and I-1.

-1 and E-2, Cr for sample I-1, and carbon for all samples are alsohown.

The diamond coating surface features can be distinguished fromhe carbon maps shown at the right column of Fig. 9, where ball-ike agglomerates are present in sample I-1, directional marks inample E-2, and a uniform diamond surface in sample E-1. It cane also observed that diamond delamination for samples E-1 and-2 occurred from the Co-free WC substrate, whereas the diamondelamination for sample I-1 occurred at the diamond/interlayer

nterface, which suggested that the adhesion of the interlayer withhe substrate was stronger than the adhesion of the interlayer to theiamond coating. Some non Cr areas after delamination of diamondoated sample I-1 are depicted as blue regions, which correspond tohe highest stresses in the vicinity of the indentation, strong enougho promote the delamination of both, the diamond coating and thenterlayer. A semi-quantitative analysis of some of the failure sur-aces for sample I-1 show about 77 wt% Cr with less than 1% of Wnd Co, which suggests that the flaking of the diamond coating mayave occurred mainly at the diamond/CrN interface.

.3. Dry machining performance test

Fig. 10 shows the flank wear-land width (VB) time evolution ofiamond coated turning inserts pretreated with methods E-1, E-2,

nd I-1, including previous results of commercial MCD, PCD, andano-diamond coated inserts done by Hu et al. (2007b) in a simi-

ar workpiece material and machining conditions. The fluctuationsbserved in the wear-land width values (VB) in Fig. 10 are due to

EPMA chemical composition mappings (center and right columns, respectively)

some aluminum built-up edge (BUE) produced during the cutting.This results show that sample I-1 suffered a sudden increase ofwear-land in a short period of time (0.49 mm in 26 s), suggestingthat coating delamination occurred abruptly and resulted in rapidwear of the exposed interlayer/carbide substrate.

A better behavior was observed for samples E-1 and E-2, whichfailed at lower VB levels and higher cutting times; 0.3 mm in 293 sand 0.34 in 104 s, respectively. The failure of the diamond film wasalso detected by the change in the intensity of the AE-RAW signalcompared with the initial cutting pass. This data confirmed that thediamond coated turning inserts made with pretreatment methodE-1 are more effective and demonstrate better performance underdry machining conditions.

Fig. 11 shows SEM micrographs of the worn samples after themachining test and confirms the coating delamination for samplesE-2 and I-1. A gradual degradation of the diamond coating wasobserved for sample E-1 where flank wear-land was the major fea-ture observed, compared with sample E-2, which displayed sharptool wear growth. Sample I-1, on the other, hand suffered delam-ination at the diamond and CrN/Cr interlayer interface, leading tothe diamond coating becoming detached from the substrate duringthe very early stage of machining, which exposes the carbide as theonly remaining cutting surface.

A detailed analysis of the worn tool cutting edges after dry

machining tests is shown in Fig. 12 for samples E-1, E-2, and I-1;the results from digital microscopy images, EPMA backscatter-ing electron (BSE) analysis, and chemical compositional maps areshown in columns from left to right, respectively. These results

5 Proces

emctieaiaititlsm

bttfd

ompDCabbsniwi

maCtd

flteatootttbeoc(fatdctt

32 H. Gomez et al. / Journal of Materials

videnced that the sample E-1 displays a gradual wear-land for-ation, no diamond debonding, and some Al built-up edge, typical

haracteristics of an adherent diamond coating. The results fromhe sample E-2 wear patterns represent a tool failure character-zed by an abrasion wear mechanism at the top of the cuttingdge followed by a diamond peeling at the bottom of the flakingrea. This wear pattern is associated to an early coating debond-ng at the top of the tool edge which promotes the subsequentbrasion of the WC substrate during the cutting operation, whichs also confirmed by the W compositional maps. In sample I-1,he failure sequence is similar to the sample E-2; however, thenitial diamond coating debonding begins at its interface withhe top Cr interlayer promoting an abrasion of the latter fol-owed by the CrN layer abrasion, and finally exposing the WC–Coubstrate as confirmed by the BSE image and the W chemicalap.The failure of sample E-2 under dry machining conditions might

e attributed to the undesirable directional surface features charac-erized in Fig. 2 which are perpendicular to the cutting direction athe tool edge. These directional features provide preferential pathsor crack propagations which finally promote the diamond coatingebonding as evidenced in Figs. 9 and 12.

The interlayer failure evidenced in sample I-1may be mainlyriginated from a weak chemical binding energy between the dia-ond coating and the top Cr surface where the blasted diamond

articles are not sufficient enough to provide a strong bonding.ue to the wide difference in machining performance between therN/Cr interlayer and the chemical etching pretreated tools, anydditional optimization in the distribution or amount of diamondlasted/impinged particles at the top of the interlayer could note sufficient to get closer to the performance of chemical etchedamples, and it may constitute a reason of why still there areo commercial solutions in the market of diamond coated tools

ntended for high silicon aluminum dry machining applicationshich incorporate the use of CrN/Cr interlayers as an adhesion

mprovement method.The CrN/Cr lack of adhesion can be also attributed to its mini-

al carbide formation ability due to its interdiffusion with carbonnd the formation of additional Cr multiphase compounds (Cr3C2,r7C3, and Cr23C6) discussed by Xiao et al. (2009) and evidenced inhe XRD patterns shown in Fig. 6, which also provide a preferentialirection for crack propagation.

The definition of adhesion refers to a system where the work ororce of detachment is measured by the application of an externaload capable of causing failure to the system under investiga-ion. Based on this concept, different methodologies, which aressentially destructive, have been developed to characterize thedhesion, including Rockwell C indentation, scratch, superlayerest, bulge, and blister test (Volinsky et al., 2002). These meth-ds are very useful for routine quality control. However, basedn our experimental results, it is not appropriate to correlatehe adhesion measured by the Rockwell C indentation evalua-ion with the dry machining performance. Furthermore, evaluatinghe adhesion of diamond coated cutting tools by pin-on-disc tri-ological tests (Polini and Barletta, 2010), impact tests (Bouzakist al., 2007), scratch, indentation, or sand abrasion testing meth-ds (Buijnsters et al., 2005) which are intended for relatively thinoatings may not be suitable for the diamond thickness range25–30 �m) normally used in CVD commercial diamond coatingsor dry machining applications. We believed the force appliednd stress field to cause the delamination in Rochwell C inden-ation test is different from the one which leads to the coating

elamination in the dry machining test. The former is a moreomplicated and more aggressive test. In the later, the delamina-ion is most due to the shear force and stresses generated duringhe machining operation. Our future work would be to develop

sing Technology 212 (2012) 523– 533

a more reliable approach to characterize the adhesion, which isbased on the specific applications and this particular manufacturingchain.

4. Conclusions

The surface characteristics of the WC–Co substrates in termsof texture features and roughness play an important role in thefinal performance of diamond coated tools, especially when com-mercial substrates are used for further diamond depositions. Theremoval of feed marks present in the as-ground samples leadingto a uniform surface texture and the cobalt removal depth are thekey factors to achieve an adherent diamond coating when chemi-cal etching methods are used. Perhaps smaller lateral crack lengthsvalues were obtained in sample I-1 after Rockwell indentations, nosignificant adhesion under dry machining conditions was achieved,as revealed by the sudden coating delamination. This suggests thatunder the present experimental conditions, no correlation can beestablished between a small crack length value and a better drymachining performance. In fact, sample E-1, exhibit higher cracklength values after Rockwell indentation under 100 and 150 Kgloads, and displays the best dry machining performance. The failureof the CrN/Cr interlayer is currently being analyzed by the authors inorder to promote a significant improvement under real dry machin-ing scenarios, and compared with other interlayer architecturesreported previously.

Acknowledgements

This research is supported from the NSF GOALI/CollaborativeResearch Award #: 0928823. Part of this research is also supportedfrom the USF/GM Research Collaboration Grant.

The authors would like to thank to Richard A. Waldo, NicholasIrish, and Curtis A. Wong at the GM R&D Center.

References

Adler, D.P., Hii, W.W-S., Michalek, D.J., Sutherland, J.W., 2006. Examining the roleof cutting fluids in machining and efforts to address associated environmen-tal/health concerns. Mach. Sci. Technol. 10 (1), 23–58.

Ager, J.W., Drory, M.D., 1993. Quantitative measurement of residual biaxial stressby Raman spectroscopy in diamond grown on a Ti alloy by chemical vapordeposition. Phys. Rev. B. 48, 2601–2607.

Bouzakis, K.-D., Asimakopoulos, A., Skordaris, G., Pavlidou, E., Erkens, G., 2007. Theinclined impact test: a novel method for the quantification of the adhesionproperties of PVD films. Wear 262, 1471–1478.

Buijnsters, J.G., Shankar, P., van Enckevort, W.J.P., Schermer, J.J., ter Meulen, J.J., 2005.Adhesion analysis of polycrystalline diamond films on molybdenum by meansof scratch, indentation and sand abrasion testing. Thin Solid Films 474, 186–196.

Chou, Ch.-Ch., Lee, J.-W., Chen, Y.-I., 2008. Tribological and mechanical properties ofHFCVD diamond-coated WC–Co substrates with different Cr interlayers. SurfaceCoat. Technol. 203, 704–708.

Chou, Y.K., Thompson, R.G., Kumar, A., 2010. CVD-diamond technologies for drydrilling applications. Thin Solid Films 518, 7487–7491.

Cristofanilli, G., Polini, R., Barletta, M., 2010. HF-CVD of diamond coatings onto flu-idized bed (FB) treated CrN interlayers. Thin Solid Films 519, 1594–1599.

Delanoë, A., Bacia, M., Pauty, E., Lay, S., Allibert, C.H., 2004. Cr-rich layer at the WC/Cointerface in Cr-doped WC–Co cermets: segregation or metastable carbide? J.Cryst. Growth 270, 219–227.

Drory, D.M., Bogy, D.B., Donley, M.S., Field, J.E., 1995. Mechanical Behavior of Dia-mond and Other Forms of Carbon. Materials Research Society, Pittsburgh, PA.

Gangopadhyay, S., Acharya, R., Chattopadhyay, A.K., Sargade, V.G., 2010. Effect ofcutting speed and surface chemistry of cutting tools on the formation of BULor BUE and surface quality of the generated surface in dry turning of AA6005aluminum alloy. Mach. Sci. Technol. 14, 208–223.

Glozman, O., Halperin, G., Etsion, I., Berner, A., Shectman, D., Lee, G.H., Hoffman, A.,1999. Study of the wear behavior and adhesion of diamond films deposited onsteel substrates by use of a Cr–N interlayer. Diamond Relat. Mater. 8, 859–864,doi:10.1016/S0925-9635(98)00321-5.

Haubner, R., Kalss, W., 2010. Diamond deposition on hardmetalsubstrates—comparison of substrate pre-treatments and industrial applications.Int. J. Refract. Met. Hard Mater. 28, 475–483.

Hu, J., Chou, Y.K., Thompson, R.G., 2007a. Nanocrystalline diamond coating tools formachining high-strength Al alloys. J. Refract. Met. Hard Mater. 26, 135–144.

Proces

H

K

L

M

M

O

P

P

P

P

the deposition of adherent diamond coatings on cemented tungsten carbide

H. Gomez et al. / Journal of Materials

u, J., Chou, Y.K., Qin, F., Thompson, R.G., Burgess, J., Street, S., 2007b. Characteriza-tions of nano-crystalline diamond coating cutting tools. Surface Coat. Technol.202, 1113–1117.

öpf, A., Feistritzer, S., Udier, K., 2006. Diamond coated cutting tools for machiningof non-ferrous metals and fibre reinforced polymers. Int. J. Refract. Met. HardMater. 24, 354–359.

i, Y.S., Hirose, A., 2007. The effects of substrate compositions on adhesion of dia-mond films deposited on Fe-base alloys. Surf. Coat. Technol. 202, 280–287.

allik, A.K., Binu, S.R., Satapathy, L.N., Narayana, C., Seikh, M.M., Shivashankar, S.A.,Biswas, S.K., 2010. Effect of substrate roughness on growth of diamond by hotfilament CVD. Bull. Mater. Sci. 33, 251–255.

allika, K., Komanduri, R., 1999. Diamond coatings on cemented tungsten carbidetools by low-pressure microwave CVD. Wear 224, 245–266.

kada, K., Taniuchi, T., Tanase, T., 1998. Advances in Powder Metallurgy & ParticulateMaterials. MPIF, Princeton, p. 1.

olini, R., 2006. Chemically vapour deposited diamond coatings on cemented tung-sten carbides: substrate pretreatments, adhesion and cutting performance. ThinSolid Films 515, 4–13.

olini, R., Barletta, M., 2008. On the use of CrN/Cr and CrN interlayers in hot filamentchemical vapour deposition (HF-CVD) of diamond films onto WC–Co substrates.

Diamond Relat. Mater. 17, 325–335.

olini, R., Barletta, M., 2010. Wear resistance of nano- and micro-crystalline diamondcoatings onto WC–Co with Cr/CrN interlayers. Thin Solid Films 519, 1629–1635.

oon, C.Y., Bhushan, B., 1995. Comparison of surface roughness measurements bystylus profiler, AFM and non-contact optical profiler. Wear 190, 76–88.

sing Technology 212 (2012) 523– 533 533

Roy, P., Sarangi, S.K., Ghosh, A., Chattopadhyay, A.K., 2009. Machinability studyof pure aluminum and Al–12% Si alloys against uncoated and coated carbideinserts. Int. J. Refract. Met. Hard Mater. 27, 535–544.

Sarangi, S.K., Chattopadhyay, A., Chattopadhyay, A.K., 2008. Effect of pretreatment,seeding and interlayer on nucleation and growth of HFCVD diamond films oncemented carbide tools. Int. J. Refract. Met. Hard Mater. 26, 220–231.

Sutherland, K., 2008. Machinery and processing: managing cutting fluids used inmetal working. Filtration Separation 45, 20–23.

Uhlmann, E., Koenig, J., 2009. CVD diamond coatings on geometrically complexcutting tools. CIRP Annals Manuf. Technol. 58, 65–68.

Volinsky, A.A., Moody, N.R., Gerberich, W.W., 2002. Interfacial toughness measure-ments for thin films on substrates. Acta Mater. 50, 441–466.

Warcholinski, B., Gilewicz, A., 2009. Tribological properties of CrNx coatings. JAMME37, 498–504.

Xiao, X., Sheldon, B.W., Konca, E., Lev, L.C., Lukitsch, M.J., 2009. The failuremechanism of chromium as the interlayer to enhance the adhesion of nanocrys-talline diamond coatings on cemented carbide. Diamond Relat. Mater. 18,1114–1117.

Xu, Z., Lev, L., Lukitsch, M., Kumar, A., 2007. Effects of surface pretreatments on

substrates. Diamond Relat. Mater. 16, 461–466.Zhang, Z.M., He, X.C., Shen, H.S., Sun, F.H., Chen, M., Wan, Y.Z., 2000. Pre-treatment

for diamond coatings on free-shape WC–Co tools. Diamond Relat. Mater. 9,1749–1752.