HAL Id: jpa-00229582
https://hal.archives-ouvertes.fr/jpa-00229582
Submitted on 1 Jan 1989
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
(CVD) ON OXIDE AND NONOXIDE CERAMIC CUTTING TOOLS
A. Layyous, R. Wertheim
To cite this version:
A. Layyous, R. Wertheim. LOW PRESSURE CHEMICAL VAPOR DEPOSITION (CVD) ON OX-
IDE AND NONOXIDE CERAMIC CUTTING TOOLS. Journal de Physique Colloques, 1989, 50
(C5), pp.C5-423-C5-432. �10.1051/jphyscol:1989552�. �jpa-00229582�
JOURNAL DE PHYSIQUE
Colloque C5, supplgment au n05, Tome 50, mai 1989
LOW PRESSURE CHEMICAL VAPOR DEPOSITION (CVD) ON OXIDE AND NONOXIDE CERAMIC CUTTING TOOLS
A. LAYYOUS and R. WERTHEIM
Iscar Ltd.,PO Box 34, IL-Nahariya 22100, Israel
Des outils de coupe en A1203 + TIC, nilrure de silicium, cabure cimente et zircone stabilisee ont ete revetus par dep8t chimique en phase vapeur [CVD] de multicouches a base de diverses combinaisons de TIC, TiCN, TiC et A1203.
L'adhesion des couches au substrat et la structure des couches ont 6te etudiees par microscopie optique, microscopie Bleclronique a balayage [MEB) et speclroscopie AUGER. Ceci a pemis d'analyser I'interaction chimique du substrat et de TIN A 1000°C. Les outils de coupe ont ete testes en coupe sur acier afin d'etudier la duree de vie, I'usure en depouille et la craterisation. Un echantillon non revttu a ete compare aux echantillons revBtus multicouches. L'adhesion de TIN sur les ceramiques oxyde et non-oxyde s'avere bonne. Par rapport a la ceramique A1203 + Tic non revttue, le materiau revBtu de TiN
+
Tic + TiN presente environ 20% de moins d'usure en depouille et le materiau revetu de TIN + TIC + A1203 jusqu'a 30% de moins. Le revBtement de substrats oxyde et non-oxyde conduit a des usures en depouille b4s similaires. Le meme revgtement TIN + Tic + A1203 par exemple, sur substrat cart>ure WC + Tic + TaC + NbC + Co conduit a une craterisation plus importante due a la p8netration du revetement dans le substrat.Abstract
-
Cutting tools made of A1203+TiC, silicon nitride, carbide, and stabi- lized ZrO2 were coated by chemical vapor deposition (CVD) with a multilayer of TIN, TiCN, TIC and At203 in different combinations. The adhesion of the coated layers to the substrate, and the structure of the layers were investigated by optical microscopy, scanning electron microscopy (SEM) and Auger spec- troscopy. This made it possible to analyze the chemical interaction between the substrate and the TIN at 1000oC. The cutting tools were tested in steel mach- ining in order to investigate tool life, flank wear and cratering. The uncoated specimen was compared to the multilayered, coated specimens. The adhesion of TiN on oxide and nonoxide ceramics showed good results. In comparison to the uncoated ceramic A1203+TiC, the TiN+TiC+TiN coated material had approximately 20% less flank wear, and the TiN+TiC+ A1203 coated material showed up to 30% less wear. Coating of oxide and nonoxide substrate material resulted in very similar flank wear values. When comparing the same coating, TiN+TiC+ A12O3, for example, on a carbide substrate WC+TiC+TaC+NbC+Co, the latter Showed much larger crater wear depth due to, and directly after, the penetration of the coated layer into the substrate.1
-
INTRODUCTIONCoated cutting materials account for ever more market share in turning, drilling and milling operations. The early coatings, dating from the 1970s, consisted of a single Tic layer known for high hardness. In the 1980s, other coating layers and combinations of several layers were developed and introduced (1). Figure 1 shows the various
Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1989552
families of materials according to their properties and applications. Today high speed steel (HSS) and carbide are available both coated and uncoated. Ceramics and cermets (2) are used in the uncoated state; the PCDs or CBNs (3) are used for very special purposes and workpiece materials.
Diamond, PCD. CBN
HIP AI2O3
Coated Carbide
Lower 4 Feed )Higher
Transverse Rupture Strength
Fig. 1
-
Properties and Application of Different Coated and Uncoated Cutting Materials Oxide and nonoxide ceramics are useful at high cutting speeds; but due to their chemical and mechanical properties, their industrial application is limited. In comparison with coated carbide, the use of a coated oxide ceramic substrate may improve resistance to plastic deformation and excess wear when the coating layer is worn off. (1 $4,5,6)In the present investigation an attempt was made to compare different coated substrates including coated ceramic materials. The different coating layers and their adherence to the substrate depend on the layer sequence, thickness and properties. Based on the results, the physical phenomena including wear behavior, chemical reaction and composition, and industrial application could be considered. Substrate materials were alumina with Tic, silicon nitride, zirconium-oxide-based material and
tungsten-carbide-based inserts.
JEST PROCFDURE AND PROPFRTIFS
A combination of machining tests and chemical investigations were carried out and analysed. Different cutting materials based on the compositions of A1203+TiC, Si3N4, stabilized Zr02 and cemented carbide were used, all of them shaped to the standard SNGN 120304T geometry. Ceramic materials were produced by hot pressing; the carbide by pressureless sintering.
Coating was done by chemical vapor deposition (CVD) which took place in a tempera- ture range of 900-1 0500C. The very thin coating layer has a lower friction value which provides less wear and longer tool life. Furthermore, the coating layer has a lower diffusion rate and resists abrasion. The sequence of the layers is built up in the order of TIN, Tic and A1203 coatings.
The chemical reactions which result in the above layering sequences are as follows:
1. TiClq(g)
+
CH4(g) 3- Tic (s)+
4 HCI(g) H22. TiClq(g)
+
112 N2 (g) ->T~N (s)+
4 HCI (g) H2-
{ AICI3(g) +3C02(g)
+
3H2(g) +3~0(g)T+6~Cl(g)T H2Purity of the gas was high (ca. 99.99%) and the deposition of the coated layers was achieved in a hot wall reactor in which the inserts were placed on passive graphite.
Placement in the reactor and coating conditions were similar for all samples.
Thickness and structure of the coating layers are shown in Table 1 according to the produced results of the coating sequence and the different substrates. The number prefixes under "Coating" indicate layer thickness in microns. The coating facility guarantees controlled gas flow, temperature and pressure.
Table I : Tested Samples including Substrate and Coating Layers Sample No. Substrate Coating (Thickness and Layer)
M 1 A1203 +Tic
M 2 A1203 +Tic 2 TiN-5TiC-1 TIN
M 3 A1203 +Tic 2 TiN-0.5 A1203-1 TiN-4TiC-2A1203 M 4 P05-P30 carbide - 2 TiN-0.5 A1203-1 TiNdTiC-2A1203 M 5 H.P.P ss-Si3N4 2 TiN-5TiC-1TiN
M 6 Stabilized Zr02 5 Tic-1 TiN
Control of flow rate is important for fine microstructure of TIN and Tic and amorphous structure of the coated layer A12O3. Production of the A1203 layer is more difficult than that of carbides and nitrides because of the dissolution of chlorides and activation energy for the production of the material. Furthermore, the wetting ability or the adhesive strength between aluminium oxide and carbides or nitrides is lower than between carbides and nitrides.
In the coated samples Si3N4 and ZrO2, the results of chemical reaction and adhesion between TIN as first layer and substrate were investigated. During the investigation the influence of the different coatings on performance and wear behaviour was tested in turning on carbon steel SAE 1040 with 0.45% C. Machining conditions were as follows: cutting speed vc=400m/min, feed f=O.l Gmmlrev and depth of cut ap=l .5mm.
Machining time was 16 minutes and both crater KT and flank wear VB were measured.
Machining conditions were constant during all tests.
The ceramic inserts MI, M2 and M3 were compared to the standard coated carbide grade M4. Wear behaviour and wear structure were checked by scanning electron microscopy (SEM) and optical microscopy. To characterize the chemical reactions between the TIN layer and the different substrates, Auger spectroscopy was used.
3
-
TFST RFSULTSThe CVD coating process in the heated reactor took place in a partial pressure atmos- phere of hydrogen or nitrogen. The hydrogen reduced surface impurities and oxides
and improved adhesion to the substrate of the first coating layer. In our investigation the TIN layer is always referred to as the first coating layer.
The results of the coated carbide material in sample M4 (Figure 2) are discussed, based on chemical analysis of the elements found in the coating layers and in the substrate. The patterns described in the diagram were measured by Auger spec- troscopy. The diagram shows the multilayer coating including two layers of A1203; the first TIN layer separates the substrate from the A1203.
Subslrate
- -
Coating layers Measur~ng A dlrectlcnDepth ~n rneasurlng direction
Fig. 2
-
Element Concentration of the Carbide Multilayer Coated Insert Measured with Auger Spectrosopya. The high cobalt content observed on the carbide surface does not prevent adhesion of the TIN layer to the substrate.
b. The TIN layer reacts as a complete boundary against diffusion of cobalt into the A1203 layer. The efficiency of the diffusion boundary depends on layer thickness and cobalt cohtent, especially at an elevated temperature of 1000oC. It is known that cobalt is subjected to grain-boundary diffusion when coated with Tic as a first coating layer. When using TIN as a first layer the effective layer thickness with a diffusion boundary is about 2 ~ m .
c. When coating with Tic as the first layer on a carbide substrate, a chemical reaction takes place between the free titanium atoms (from the dissolution of TiC14) and the carbon from the substrate. The Tic layer produced in the chemical reaction partially by decarbonization from the substrate surface causes phase transformation and the appearance of an eta phase. (7)
co +
WC- I C O ~ ~ ~ C {c06w6cThe phase transformation depends on the carbon content on the carbide surface. The eta phase is very brittle and should be avoided in carbide, especially that used for milling inserts. The use of a TIN layer is a much more effective diffusion boundary compared to a Tic layer.
The chemical stability of the coated carbide inserts in a temperature of 1000oC is com- pared to the stabilized Zr02 samples at the same temperature and in the same active gases. Test results are shown in Figure 3.
Kinetic Energy EV Kinetic Energy EV
Fig. 3
-
Concentration Profiles of Zr02 Substrate Coated with Tic-TiNFrom the Auger spectroscopy, the optical microscopy and the machining results, the following remarks and conclusions may be made:
a. In the upper diagram of Figure 3, two continuous layers can be observed, one of (Zr,Ti)x(O,C)y and the second of Tix (O,C)y. The ZrOz reacts with gas and especi- ally with the Ti atoms in the atmosphere of TiC14, hydrogen and methane to produce the new phase (Zr,Ti)x(O,C)y as shown in Interface laver II. The two layers adhere to the substrate and to the Tic layer and have a thickness of 1 to 2 microns. The surface of Zr02 reacts chemically and results in combined oxides (red) in which the origin of the oxygen is the Zr02. It might be that the new phase (Zr, Ti)x(O,C)y located near the substrate has a low percentage of carbon.
b. A chemical composition near the Tic layer is a Ti(O,C)y combination in which oxygen is the more dominant element than carbon. This indicates that the ZrO2 is chemically unstable under these temperatures and conditions or that Jnterface laver II acts as a boundary layer.
Sample
M5
which is a nonoxide ceramic based on silicon nitride produced by hot pressing has a TiN+TiC+TiN sequence of coated layers.a. The phenomena of chemical reaction between the different gases on the sur- face of the substrate was not observed on the Si3N4 substrate with its solid solution beta phase structure. No new phases could be identified. This means that the sub- strate materials are chemically and thermodynamically stable in the presence of the used gases up to the tested temperature (1 000oC).
b. The adhesive strength of the first TiN layer on the substrate was checked by mechanical testing. The use of this material for steel cutting proved that the coating did not peel off despite heavy machining conditions.
c. Coating of TIN and Tic on the Si3N4 substrate improves the performance of these inserts when cutting steel due to the coating layer itself acting as a diffusion barrier between substrate and steel. From application and experiments, it is well known that the beta phase Si3N4 materials are successfully used mainly for the machining of cast iron. It is not successful in machining steel due to high wear caused by the diffusion of nitrogen and silicon into the heated chips of steel and the chemical reaction resulting in silicon-ferrum compositions.
Comparison between the uncoated oxide ceramic (A1203+TiC) and the same material but with different coatings was done with samples M I , M2 and M3. Structure and adhesion of the layers are similar to those obtained on the Si3N4 substrate. The oxide ceramic substrate was chemically and thermodynamically stable. The investigation mainly-included machining tests described as follows.
Referring to the stabilized Zr02 substrate, it is proven that this material can be used as a cutting material for steel but not for cast iron. The reason is the fact that at high cutting speeds, the temperature in the contact zone can reach up to 1000 OC; under these conditions in a carbon-reached atmosphere, the Zr02 will be quickly deoxidized, resulting in microcracks on the surface and total failure of the material. Coating of this material with TIN and T i c or A1203 reduces its sensitivity to deoxida- tion and other surface-failure phenomena.
4
-
JVlECHANfCAL TEST RESULTSMechanical tests compared the uncoated ceramic material A1203+TiC to the same material with different coating, and to a standard coated carbide grade. From the com- parison of
MI,
M2 and M3 (Table I ) , the influence of the coating on oxide ceramic can be analysed. From the comparison of types M3 and M4, the influence of different sub- strates (carbide and oxide) with the same coating layer sequence is investigated.Figure 4 shows crater depth KT as a function of machining time t. Crater depth KT for uncoated ceramic M1 and coated ceramics M2 and M3 showed a very linear, similar pattern. It reached an 0.35mm depth after 16 minutes machining time. Thus, in this case the additional A1203 layer did not influence crater development. The aluminum oxide which serves for thermal protection does not serve as an additional wear- resistant layer. Crater depth on the carbide coated substrate is very small, up to approximately 8 minutes; after 16 minutes it deepens very quickly to 1.2mm. The phenomena can be explained by the sudden penetration of the coated layers and the exposure of the substrate to higher wear due to the higher cutting speeds. The coating layers have different properties and, therefore, different applications. The Tic has the higher resistance against mechanical wear and therefore has improved tool life. When machining cast iron, the A1203 layer is very important because of the thermal bound- ary protecting the substrate, and the nonhornogeneity of the cast iron, causing higher wear and higher temperatures in the crater area.
Mach~nlng Condlllons Wolkplece Maler~al SAE 1045
Insert SNGN 120308
Cuulng Speed Vc.400 m/mln 1-0 16 mmlrev Depth of Cut ap.15 mm 0 2
16 mln 20
Machining Time t
Fig. 4
-
Crater Depth of Coated and Uncoated Ceramic Materials Compared with a Coated Carbide GradeFigures 5, 6 and 7 show the wear geometry of the cutting edge, including top rake and clearance faces of the different tested materials.
In Figure 5 the absence of an A1203 layer caused high temperature changes and, as a result, cracking phenomena. The same substrate with the additional ceramic coating layer in Figure 6 shows less wear on the clearance face, and is much more consistent in size, without cracks. The crater shape itself with the additional VG wear at the end is very similar in both types. The crater wear of the carbide-coated substrate in Figure 6 is much larger and reflects the temperature distribution on the top rake face. The higher heat conductivity of carbide, compared to that of ceramics, is one of the
additional reasons for increased crater wear. Heat transfer from crater sides is greater, temperature is lower and crater width (KB) is smaller than in the center. The constant crater width and depth on the ceramic-coated material reflects the thermal properties and the temperature distribution.
Figure 7 shows the crater and flank wear geometry qf uncoated ceramic inserts. Crater dimensions are similar to those of coated grades, but differ especially on both crater, ends.
The shapes of flank wear on the various samples differ from those of crater wear.
When comparing M2 and M3 types, the influence of diffferent coating layers can be observed. The wear resistance of the ceramics not coated with aluminum oxide (Fig- ure 5) is lower than those coated with aluminum oxide (Figure 6). The least clearance face wear is observed on carbide coated material (Figure 7) and is even smaller than that on the ceramic coating. This might be explained by the fact that more energy is re- quired for the creation of crater wear than for flank wear, and that we still did not pene- trate into the substrate. Both M3 and M4 types coated with A1203 have lower friction and less wear compared to types M I and M2. Flank wear of uncoated ceramic mate- rial (Figure 8 ) is larger but consistent over the entire depth of cut. Different chemical reactions on flank wear phenomena require additional investigation.
Figure 9 shows flank wear of the different materials. Initial flank wear and subsequent wear development of the uncoated ceramic material M I is very high, reaching 0.5mm after 16 minutes. Ceramic and carbide substrates (M3 and M4) coated with a com- bined layer sequence including A1203 showed much lower wear.
Fig. 5 - Wear Geometry Including Flank and Crater Wear of a TiN+TiC+TiN CVD Coated Ceramic Insert
Fig. 6 - Wear Geometry Including Flank and Crater Wear of a TiN+Al203+TiN+TiC+
AI2O3 CVD Coated Carbide Insert
Fig. 7 - Wear Geometry Including Flank and Crater Wear of a Carbide Insert (WC+TiC +NbC+TaC+Co) Coated with TiN+AI203+TiN+TiC+ AI2O3
Fig. 8
-
Flank and Crater Wear Geometries on an Uncoated Ceramic (A1203+TiC) InsertMachining Time t
Fig.
9-
Flank Wear of Coated and Uncoated Ceramic Materials Compared with a Coated Carbide Grade5
-
CONCLUSIONS1. TiN as a first coating layer acts as a diffusion boundary on the surface for cobalt.
2. ZrO2 zirconium oxide has limited chemical stability in gas environments which causes reactions into new phases.
3. Coating with TiN+TiC is possible on any of the investigated substrates.
4. Si3N4+A1203+TiC are chemically and thermodynamically stable; no new phases are produced.
5. Crater wear on coated carbides is much larger than on coated ceramic materials due to the higher thermal conductivity of carbide.
6. Flank wear on uncoated ceramics is larger than that on carbide.
7. Flank wear dimensions and geometries depend on coating layers and are smaller for A1203 materials.
6
-
REFERENCES111 Pastor, H., Refractory and Hard Metal, December (1987) 196.
121 Ettmayer, P. and Koilaska, H., Symposium on Ceramic Cutting Tools, Hagen, West Germany (1 988) 163.
I31 Steinmetz, K., Symposium on Ceramic Cutting Tools, Hagen, West Germany (1 988) 21 5.
141 Gruss, Walter W., Ceramic Bulletin
a,
6 (1988) 993.151 Brandt, G. and Mikus, M., Wear
228
(1 987) 99-1 12.I61 Tonshoff, Hans K. and Bartsch, Sven, Ceramic Bulletin
a,
6 (1988) 1020.I71 Layyous, A., Sixth European Conference on CVD (1987) 58.
