tool wear in titanium machining
TRANSCRIPT
Tool Wear in Titanium Machining
P. D. Hartung and 6. M. Kramer, Massachusetts Institute of Technology - Submitted by B. F. von Turkovich (1)
S uinmarv
alloys has been shown to be fundamentally different than that in the machining of steel and nickel-based alloys. It is suggested that tool wear is greatly reduced when adhesion occurs between the tool and the chip, preventing relative sliding at the tool/chip interface. This adhesion is promoted by chemical reaction at the interface.
The thickness of the reaction layer is determined by the balance between the diffusion flux of tool constituents through the layer and the removal of tool constituents through chemical dissolution at the interface between the reaction layer and the titanium chip. In this way, for given cutting conditions, a characteristic thickness of the reaction layer is maintained and tool wear is limited by the rate of dissolution of the reaction layer into the titanium.
The existence of a stable reaction layer of Tic on diamond and PIC-based tools (the two most wear- resistant tool materials) has been demonstrated, and the estimated diffusion flux correlates well With the observed wear rate.
- The mechanism controlling the crater wear of cutting tool materials in the machining of titanium
Introduction
Although quite a bit of work has been done in the last thirty years in an effort to find the limits of application of HSS and cemented WC tools in the mach- ining of titanium and its alloys, comparatively little work has been done to determine the potential of other possible tool materials in this application.
Our calculations indicate that essentially no potential tool materials exist that are sufficiently chemically stable with respect to titanium to exhibit low wear rates by virtue of their low chemical solubi- lities in titanium. Instead, to maximize wear resis- tance, relative sliding between the tool and the chip must be eliminated. Under these conditions a very thin, adherent "boundary layer" of chip material forms on the tool face, and wear is limited by the diffusion rate of tool constituents through this layer.
the rate of material removal due to the physical motion of the chip is so great that the contribution of chemical diffusion to the flux of tool material from the interface becomes negligible. This is the case in the machining of steel and nickel alloys at normal speeds. [1,2]
commonly employed to machine titanium react with the workpiece material to form titanium carbide. This reaction layer has high deformation resistance at cut- ting temperatures and adheres strongly to both the tool and the chip. The only tool material which was found to be both more wear resistant and more deforma- tion resistant than WC is polycrystalline diamond. Since diamond is essentially pure carbon, the forma- tion of a Tic layer at the interface is also promoted in this case.
The purpose of the present study is to determine the wear mechanism in titanium machining by comparing the predictions from various theoretical models of tool wear with experimental results. It may be seen that the experimental results are consistent with the mechanism that is described above.
Exper iment a 1 Result?
Turning tests were carried out on Ti 6Al-4V with a conventional c2 grade (Carboloy 820) and C3 grade (Kennametal K68) of cemented carbide. An SNG432 geo- metry was employed with d lead angle of IS degrees. Details of the experimental procedures that were employed may be found in Reference 131. At cutting speeds from 200 sfpm (61.0 m/min) to 400 sfpm (122 m/min, crater wear limits the tool life. Flank wear is stable and does not contribute to tool failure until crater wear weakens the edge and plastic defor- mation of the cutting edge causes acceleration of the wear at the flank. Figure 1 shows the rake face of a K68 insert after machining Ti 6Al-4V at 200 sfpm (61.0 m/min) for 10 minutes. The crater depth in the center of the crater is approximately 23 microns, and the flank wear, while uneven, reached a maximum of .008 in. (0.2 m m ) . At speeds in the range of 400 sfpm (122 m/min) to 2000 sfpm (610 m/min), plastic deforma- tion of the cutting edge becomes pronounced, and is responsible for tool failure. A plot of crater wear rate as a function of cutting speed for the two dif- ferent grades of carbide is shown in Figure 2.
To determine the relative wear rates of the vari- ous potential tool materials in the machining of Ti 6A1-4V, turning tests were performed at 200 sfpm (61.0 m/min) using various coated and uncoated carbides, polycrystalline cubic boron nitride and polycrystal- line diamond [ 3 ] . The average crater wear rates of the various tool materials in the turning of Ti 6Al-4V at 200 sfpm (61.0 m/min) are shown in Table 1.
Once relative sliding occurs at the interface,
The tungsten carbide-based compositions that are
Table 1 - Averaqe Crater Wear Rates OF Various Tool Materials i n the Turning of Ti 6A1-4V st 200 SFPV (61 .0 Thin)
Tool Material !#gar
A 1 2 0 3 (Carboloy 030) HfO2-coated A1203 ZrC-coated Kennametal K7H(C8) HIC-coated Kennametal K68(C2) Cementea Tic (XLO XLRB) TiB2-coated K7H Cast TiCN (Teledyne S D - 3 ) CEN (BorazonR) Tic-coated WC (Carboloy 523) HfC-coated WC (Surftech H-25X) HfN-coated WC (Teledyne HN+) TiN-coated WC (Teledyne TC+) Kennametal K7H ( C 8 ) Surftech H-25X ( C 3 ) Surftech H-2S (C2) Kennametal K68 (C2) Carboloy 8 2 0 (C21 Diamond (SynditeR) Diamond CompaxR)
Rate (u rn /min )
790. 560. 56. 52. 4 3 . 39 3 3 . 30. 30. 22. 11. 11.
a. s 8.0 4.5 2 . 5 2.2 1.4 1.3
Figure 1 - Crater Wear after Machininq Ti 6A1-4V for 10 minutes at 200 sfprn ((51.0 rn/min)
Cuttrnq Speed (m/min)
o i n o 2 0 0 700 400 500
Cuttinq Speed iftimin)
Figure 2 - Wear Rate as a Function of Cutting Speed for two Different Grades of Tungsten Carbide
Annals of the ClRP Vol. 31/1/1982 75
-4 J
- -_- + + + 2 0 0 0 - w: t inti ?Qi?t 0 f 'r I t3 n : ..rl
+++ ( 1 9 3 9 ' K '
+ 1500 -
+ + 1 0 0 0 - +
+ + 500 - , , , , , , , , , , , , , , , , , , , ,
0 5 0 0 1000 1500 2000 cutting Soeed Iftlrnin)
Figure 3 - Area-weighted Average Cutting Temperature as a Function of Cutting Speed for a Kennametal K68 Insert. Feed rate, .005 in/rev (.125mm/rev)
Temperature MeaSUKementS
Tool-chip interfacial temperatures were measured using the chip-tool thermocouple technique. Figure 3 is a plot of measured cutting temperature vs. cutting speed for a K68 insert in the speed range from 50 to 2000 sfpm (15.2 to 610 m/mm). As can be seen in the plot, the indicated cutting temperature asymptotically approaches the meltiny point of titanium at high cut- ting speeds. The temperature indicated by the chip- tool thermocouple technique is the average tempera- ture. It is expected that the maximum cutting temper- ature at the center of the crater will be higher than the average cutting temperature [4]. Based on the results shown in Figure 3, the analyses of tool wear were carried out for temperatures in the range of 1300 - 1500°~, the estimated maximum cutting temperature. and HfC coated tools indicated that there is very little difference in cutting temperatures between coated tools and their uncoated substrates. The aver- age cuttinq temperature for the HfC coated K68 (C3 grade) inserts, 101O0K, was 60°K lower than that of the uncoated inserts, 1070°K, at 200 sfpm (61.0 m/min). The presence oE the coating on the tool surface is expected to have a small influence on the thermal EMF generated by the chip-tool thermocouple. and may account for much of this difference.
Theoretical Analysis
Since crater wear is the dominant form of tool wear in the machining of titanium alloys, OUT first assumption was that the tool wear might be explained by the chemical dissolution of the tool material in titanium. This mechanism controls the wear of carbide tool materials in the machining of steel and nickel- based alloys [ l , 21.
If the mechanics of chip flow are assumed to be similar for different tool materials cutting under identical conditions, the relative wear rates of the tool materials may be taken as the ratio of the products of the solubility and the molar volume for each material.
Cutting temperature measurements taken on Tie2
( 1 ) Vwear 1 - '1'1 Relative Wear Rate = - - - 'wear 2 '2'2
where:
vwear = wear rate of tool material i
Vi
Ci
= the molar volume of tool material i
= the equilibrium concentration (solubility) of tool material i in the work material at the cutting temperature
The molar volumes of many tool materials are available in the literature, and are easily estimated from molecular weight and density information or from lattice parameter measurements.
lities of tool materials in titanium are described in [31, and estimated solubilities are given in Table 2. The values in Table 2 may be taken as estimates of the maximum possible solubilities, based on chemical pro- perties alone. While estimates based on chemical pro- perties may be quite accurate when the predicted solubilities are small, physical and geometric effects become significant when large concentrations of solute atoms must be taken into solution. It has been as-
The methods employed for estimating the solubi-
sumed that the solubility of the tool material in titanium will not be greater than the solubility of i t s least soluble component. That is, if tungsten carbide is being employed as a tool material, it may be assumed that the solubility of WC will be no qreater than the solubility of C, since the solubility ot tungsten in titanium i s much greater than that of carbon.
Table 2 - Estimated Solubilities of Tool Yaterials in Titanium at Various Temperatures
Solubility (nole 9 ) Tool .\later ia 1 1300°K 140O0K 1500°K
HfC NbC Sic TaC Ti C vc WC ZrC BN HfN Si3N4 TiN
1.27 23.70
16.03 7.75
4.23
10.92
48.58 *
1.41 20.73
14.32 7.75
* 4.42
12.80
48.48
*
1.53 18.45
13.02 7.75 * * 4.58
13.35
48.17 A1203 HfO2 7.85 14.67 29.46 Zr02 10.47 21.58 39.50 La203 0.033 0.076 0.16 Ti82 * t
t *
*Chemical reaction occurs.
Table 3 lists the solubilities of the tool con- stituents in titanium, obtained from phase diagrams [5, 6, 7, 81. Comparison of Table 3 with Table 2 reveals that all of the tool materials except for HfN, HfO and La2O have predicted solubilities in titanium thaz are grea2er than that of at least one of their constituents. In all of these cases, the solubility of the tool material may be approximated by the solu- bility of the least soluble component divided by the number of atoms of that component per molecule of tool material.
Table 3 - Reported Solubilities of Tool Constituents in Titanium at Various Temperatures
Solubility (atomic % ) Tool Constituent 1300'K 1400'K 1500DK Ref.
A 1 B C Hf La N Nb 0 Si Ta Ti V W Zr
15.0 1.0 0 . 6
1.0 23.6
34.0 2.2
*
*
100.
*
19.0 28.0 [51 1.0 1.0 [6 1 0.6 0.6 [ 7 1 * [5 i 1.0 1.0 I S 1
23.5 23.2 171 [ E l
34.0 34.0 171 3.0 4.0 i5i * 171
[71 [51 [51
100. 100.
*These constituents are soluble over a wide range of temperatuLes and concentrations
It may be seen that predictions based on the solubility argument do not agree well with test results. In particular the analysis does not explain the high wear rates of HfC and ZrC coated tools rela- tive to that of WC (see Table 1). Figure 4 shows the rake face of an HfC-coated, cemented WC substrate next to an identical, uncoated, Kennametal K68 (C3) sub- strate, after both tools have turned Ti 6A1-4V for 30 sec. at 2GO sfpm (61.0 m/min). The uncoated tool has an adherent layer of titanium over the entire crater and shows less than 1.5 microns of crater wear. On the other hand, the HfC coated insert shows between 28 and 34 microns of wear on the rake face. Even after the 6 micron thick HfC coating was completely pene- trated, the coated tools still continued to wear at a much greater rate than the uncoated tools. The fact that the coated tool wore at about 20 times the rate of the identical uncoated tool can not be due to the different solubilities of the two materials. It appears as if the wear mechanism in the machining of titanium is fundamentally different from that in the machining of steel and nickel alloys.
I’iqure 4 - Comparison of Crater Wear on Lncoatcd and HfC-coater! Tools after Turninq Ti 6A1-4V for 30 seconds at 200 sfpm (61.0 m/nin),.005 ln/rev (.125 mp/rev) feed and .050 in (1.25 n?) de;th.
Upper Bound Calculation of the Diffusion Flux
It is possible that the diffusion of the tool constituent materials within the chip controls the wear. Cook and Nayak 191 have modelled the cutting process with the assumption that no deformation occurs in the chip as it slides across the tool-chip inter- face. This model yields an upper bound to the diffus- ion rate and, therefore. provides an upper bound for the wear rate of tool materials titanium due to diffusion.
The predicted wear rate in crater, from this model, is:
I D ) + ‘wear = -KC .ti
the wedr rate material
in the-machining of
the center of the
of the tool
the ratio of molar volumes of the tool material and the chip material
the equilibrium concentration of the tool material in the chip
the diffusion coefficient of the slowest diffusing tool consti- tuent in the chip
the time it takes for the criip to move from the edge of the tool to the center of the crater
The time required €or the chip to slide from the tool edge to the center of the crater was estimated from chip thickness measuremen s and was found to be equal to approximately 3.2x10-! secogds. The molar volume of titanium is 10.64 cm /mole 1101.
tool materials, calculated at 1400°K using Equation 2, are contained in Table 4 . The equilibrium concen- trations were obtained from Tables 2 and 3 , and were assumed to be equal to the solubility of the least
The predicted wear rates ot various potential
Table 4 - Predicted Wear Rates From the Upper Bound Diffusion Model (at 1400°K)
Tool Ratio of Diffusion Predicted Molar Coefficient+ s ~ ~ ~ , ” ’ I ~ l ~ : y Wear Rate
Material Volumes* (cm2/sec) (um/min)
Diamond 0.321 2.28~101: 0 . 6 NbC 1.266 3.66~10 0.6 Tic 1.147 4 . 8 2 ~ 1 0 1 ~ 0.6 vc 1.025 2.92x10-’ 0.6 WC 1.165 1 . 0 5 ~ 1 0 - ~ ~ 0.6 ZrC 1.472 1.40~10 0.6 TiN 1.080 4 . 8 2 ~ 1 0 1 ~ 23.5 Zr02 2.068 1.4OxlO-: 17.0 TiB2 1.492 4.82~10 0.5
176. 27.7 28.8 20.0
63.0 9.48
1063. 2508.
53.2
*Ratio of the molar volume of the tool material to the molar volume of titanium 111,131.
?Diffusion coefficient of the slowest-diffusing constituent [11,121. Diffusion data was not available for those systems that are not included in the table.
soluble componont in cases where the solubility is greater than the solubility of the least soluble con- ponent..
the actual measured wear rates (Table 1 ) reveals that this model not only gives a gross overestimate of wear rate, as may be expected f o r an upper oound analysis, but also does not rank the materials properly in order of their wear resistance.
Interfacial Conditions Control the Wear
Comparison of predicted results [Table 4) with
Because of its low wear rate relative to all Other tool materials tested, and the lack of informa- tion concerning it in the literature, the wear of polycrystalline diamond tools was carefully monitored in machining Ti 6A1-4V at 200 sfpm (61.0 m/min), to determine what properties of diamond might be respons- ible for the low wear rate.
Profile traces across the rake face of the tool show no apparent wear until a cutting time of 4 min. In the micrograph, an adherent layer of titanium appears to cover the surface (Figure 5 ) . After 15 minutes of cutting, three zones of accelerated wear are apparent on the rake face of the tool which show over 20 microns of wear, but still there are regions with essentially no wear.
Figure 5 - Wear on the Rake Face of a Polycrystalline Diamond Tool after T u r n l n g Ti 6A1-4V at 200 sfpm (61 C61.0 m/min), .005 in/rev (.125 mm/rev) and .05O In (1.25 mm) depth of cut..
The “scalloped” appearance of the rake face wear on the diamond tools is not unique to these tools, but can be seen quite clearly on the CBN and the coated carbide tools. This effect is unusual and has not been discussed in the metal cutting literature. Uncoated cemented tungsten carbide tools do not show this discontinuous, scalloped form of wear when mach- ining Ti 6A1-4V, but crater in a more conventional manner. Figure 6 shows the surface of cemented WC and polycrystalline CBN tools after etching the titanium from the surfaces with hydrofluoric acid for 20 min. The surface of the WC tool appears to be chemically polished whereas the surface of the CBN tool is rough and appears to be abraded.
It is necessary to identify a mechanism by which rapid wear can occur in one region of the crater area while essentally no wear is present in an adjacent region. Different regions of the rake area must be sublect to different interfacial conditions during the cut.
titanium adheres to the tool and no relative sliding occurs at the tool-chip interface. A boundary layer of titanium forms at the interface, and the relative motion between the tool and chip is generated inter- nally by shear within the titanium chip material. This “dead” boundary layer quickly becomes Saturated with tool constituents, limiting the mass transport of tool constituents from the tool surface. This would explain why essentially no wear occurs in certain regions of the crater area of some tools.
PK bably tne most significant example is BOKaZOn8 (CBN), which has essentially no wear in one region of its crater area, and a 3G micron deep pit in another (Figure 6 ) . Tests on Inconel 718 with CBN tooling have shown that the wear of is very sensitive to interfacial conditions, with a wear rate at 135 sfpm (41.1 m/min) five times greater than that at 600 sfpm (183 m/min) 1141.
It is possible that, under certain conditions,
77
Fiqure 7 - Enerqy Disoersive X-Ray Analysis of a Polycrystall ine Diamond Tool after Yachininq Ti 6A1-4V for 20 minutes.
Polycrystalline Cubic Boron Nitride
Figure 6 - Comparison of the Wear of Cemented Tunqsten Carbide and Polycrystalline Cubic Boron Nitride in the Turning of Ti 6A1-4V at 200 sfpm (61.0 m/min), .005 in/rev (.125 mm/rev) feed and .050 in (1.25 mm) depth. Cuttinc; Times: WC, 10 minutes: CBN, 0.5 minutes. Adherent titanium has been removed by etching in HF.
Analysis of the Tool-Chip Interfacial Conditions
It is not possible to see the tooll'chip interface while the cut is in process, so it is necessary to look for signs of different boundary conditlons on the tool surface after the cut is made.
the tools look like at high magnification, to obtain x-ray spectra of the elements present in these speci- fic areas, and to obtain mappings of the location of the specific elements of interest which have been shown to be present on the tool, a scanning electron microscope (SEM) with an energy dispersive x-ray anal- yser (EDS) was used.
20 minutes at 61.0 m/min (200 sfpm) is shown before (Figure 7 ) and after (Figure 8 ) etching in hydro- fluoric acid (HF) for 20 minutes. Hydrofluoric acid reacts violently with metallic titanium and slowly with titanium carbide and was, therefore, used to remove any unbound titanium that was present on the surface. Figure 7 shows a backscatter image (top left) of the rake face and its associated titanium map, as well as a secondary image of both the rake and flank surfaces at the tool tip. The scalloped wear pattern is very clear in these pictures, with the areas of least wear corresponding to the narrower por- tions oE the crater area. Titanium covers the worn areas of the tool. After etching, the tool was anal- yzed on an AMR SEM with a Tracor Northern TN-2000 EDS (Figure 8 ) . Secondary images show the topography of the tool after the unbound titanium has been removed. X-ray analysis of the crater area shows that titanium is present in some form in the worn portion of the tool after etching. The elemental map shows the greatest concentration of titanium in the upper right hand corner of the image area, in a region of rela- tively low wear.
AU er Electron Spectroscopy for Analysis of Tool
To determine what the worn and unworn areas of
A Compax' tool which has turned Ti 6A1-4V for
* The Auger electron spectroscopy (AES) technique
for chemical analysis of surfaces can be used to unambiguously identify the composition of a solid surface [151. To determine whether titanium carbide was formed in the crater area of the diamond tool of Figure 8 , Auger spectra were taken of worn and unworn regions of the tool and of a Tic control (an unused
X-ray Spectra of the Etched Surface
Figure 8 - Energy Dispersive X-Ray Analysis of the Polycrystalline Diamond Insert of Fiqure 7 after etching in HF for 20 minutes to remove any unbound titanium from the tool surface.
Carboloy Tic coated tool). Comparison of a spectrum from the crater of the diamond tool with the spectrum from a Tic control and an unworn portion of the dia- mond tool reveals a similarity between the carbon peak shapes of the crater area and the Tic control which is not characteristic of the unworn diamond material (Figure 9 ) . Auger results indicate that titanium oxy- carbides are present on the surface of the crater area of the diamond tool, and as the depth below the sur- face of the tool increases, the oxycarbides decrease in quantity and a layer of relatively pure Tic exists to a depth of on the order of 100 nm below the sur- face.
An Auger analysis was performed on a cemented tungsten carbide tool (Kennametal K68, grade C3) which has machined Ti 6A1-4V for 30 seconds. The tool was etched for 20 minutes in HF to remove any metallic titanium from the surface before performing the anal- ysis to see if a reaction layer of titanium carbide was formed in the crater area during the cut. An unworn portion of the tool was used as a WC control. Auger spectra were taken at three successively greater depths below the crater surface of the tool. Near the surface, the Auger spectrum shows the presence of tungsten, titanium, bound carbon and oxygen in the crater area of the cemented WC tool. Very little cobalt is present. The carbon level decreases proportionately with the titanium level as the depth below the original crater surface increases, and then
C
C
seconds to establisn a 100 nm thick layer at 1500'K under static conditlons. This time may be compared with cutting t-sts, io which a 1 0 0 nm thick layer of Tic was found on the surface of the crater after 20 minutes of cutting. It nay oe assumed that, after 20 minutes the reaction layer has established an equill- brium thickness.
Table 5 - Predicted Growth of a Tic Layer on Diamond
Tarabol ic Layer Time (sec) for a Temperature Growth Time layer thickness
(cm2/sec) ( L m ) 5 0 0 i 7508, 10008, ( O R ) Constant* (min) Thickness of:
1300 9.51~10-l~ 1 0.239 2.63 5.92 10.52 20 1.068
1400 5.19xlO-'l 1 0 . 5 5 8 0.48 1.08 1.93
1500 2.26xlO-'O 1 1.165 0.11 0.25 0.44
, I
20 2 . 4 9 7 4
I - 0 , -\I\-.-. __
1 TI 20 5.210 !
_ ' I h
Crater Area of Diamond Tool
"From Reference [201.
Tic Control
3iamond Control
Figure 9 - Comparison of the Auger Spectrum of the Crater Area of a Diamond Tool with the Titanium Carbide and Diamond Control Spectra.
increases to correspond with the carbon Level of the unworn WC material indicating that Tic is forming on the crater surface of the WC-based tools. The presence of oxygen indicates that oxycarbides are also formed during the cut.
the crater of the kC-based tool is an indication that the Tic grains are removed from the surface as a result of binder removal (probably by diffusion), and that the layer of Tic has to be continuously replen- ished by removing carbon from the WC grains below.
Theoretical Implications of the Auger Results
the surface of the diamond tool sugrjests that the wear rate of the tool might be calculated from the rate of diffusion of carbon through the Tic layer.
Kroll has suggested that graphite crucibles are able to withstand the attack of liquid titanium metal by forming a stable layer of titanium carbide on the surface, which prevents direct contact between the melt and the crucible 1161. Adelsberg and Cadoff [ 1 7 ] have shown that the formation of a Tic layer on graph- ite from molten titanium follows a parabolic growth law of the form:
The absence of cobalt in the surface layers of
The existence of a 100 nm thick layer of Tic on
X2 = Kpt ( 3 )
K (Tic) = 0.2 exp (-61800/RT) cm2/sec (4) P where: x = layer thickness, cm
= parabolic growth constant, cm2/sec KP t = carburization time, sec
T = temperature, OK
R = the universal gas constant = 2 cal/mole-°K
In Table 5 , the rate of growth of the Tic layer is predicted at 1300, 1400 and 1500°K under the assumption that the parabolic growth law holds at temperatures under the melting point of titanium. As can be seen in the table, it takes less than 0.5
Figure.10 - Model for the Wear oE Diamond by the sffusion of Carbon through a Tic Reaction Layer
Calculation of the Wear Rate of Diamond i n the - Machining of Titanium
In Figure 10, a simple model for the wear process is shown, where the wear rate is given by:
where: vwear = the wear rate of the tool (cm/sec)
Vt
Ct = the concentration of carbon in the tool material (moles/mole)
Vi
= the molar vo ume of the tool material (cin 3 bole)
= the molar volume of titanium carbide at point i in the titanium arbide reaction layer (cm /mole)
= the concentration of carbon at point i in the titanium carbide reaction layer (moles/mole)
5
Ci
b = the reaction layer/chip boundary
0 = the reaction layer/tool boundary
D = the diffusion coefficient of carbon in the tita ium carbide reaction layer (cm 9 /sec)
t = the thickness of the titanium carbide reaction layer ( c m )
There is a large amount of variation in reported dif- Eusion coeEficients for carbon in Tic 117, 181. The values of [ 1 7 ] were considered the most reliable Since they were measured using a layer growth technique with a similar geometry to our assumed model. These values were used in predicting the wear rate. The reaction layer thickness, t, was assumed to be 100 nm, which was the approximate measured value in the AES study. All concentrations are from phase diagram data [ S ] . Molar volumes of titanium carbide are calculated from the lattice parameter measurements of 161 . At 1400°K, the parameters of interest are:
D = 9.26~10-l~ cm2/sec
79
E. Rudy, "Ternary Phase Equilibrium in Transi- tion Metal- Boron-Carbon-Silicon Svstems. Part
= 3.42 moles/cm' f o r diamond 12.4 moles/cm3 for WC
Vt
Ct 'b = 12.02 cm3/mole
'Jo = 12.21 cm'/mole
Cb
CO
t = 100 nm (from Auger results)
= 1 for both diamond and WC
= .67
= .98
The predicted wear rate at diamond at 1400°K is 7 . 7 7 ~ 1 0 - ~ cm/sec = 0.47 micronshin. ponding predicted wear rates at 1300 and 1500°K are 0.10 and 1.85 microns per minute: agree quite well with test results, which show an actual wear rate of from u.6 to 1.6 microns/minute when machining titanium with diamond at 200 SFPM (61 rn/min).
and 6.7 microns/min at 1300, 1400 and 1500°K res- pectively. This also agrees with the observed wear rate of cemented WC, whicn is from 2.2 to 2.5 microns/ min.
CONCLUSIONS
Most potential tool materials either rapidly dissolve in or chemically react with titanium when they are used to machine titanium alloys. When chemical reac- tion occurs, a reaction layer forms, the thickness of which is determined by the balance between the rate of diffusion of tool material through the layer, and the rate of dissolution of the reaction layer in the work- piece. It is suggested that the wear rate of tool materials which maintain a stable reaction layer is limited by the diffusion rate of tool constituents from the tool-chip interface. At equilibrium, this flux will be equal to the diffusion flux of tool con- stituents through the reaction layer.
Calculations based on this diffusion 1im.ited model accurately predict the wear rates of p l y - crystalline diamond and tungsten carbide. These materials showed experimental evidence of the forma- tion of stable reaction layer, and were the most wear resistant materials tested.
For those materials for which a stable layer forms, it may be possible to reduce tool wear by saturating the titanium workpiece with the fastest diffusing tool constituent, thereby eliminating the concentration gradient of that element within the workpiece, and reducing the driving force for dif- fusion. In addition, alteration of the surface condi- tions of the tool material so as to provide a stable adherent layer of workpiece material on the tool sur- face that limits diffusion should serve to decrease the diffusion rate of tool constituents into the chip and reduce the wear rate. If relative sliding occurs between the tool and the chip, it is expected that the wear rate will be much higher.
ACKNOWLEDGEMENTS
The corres-
These predictions
The wear rate of WC is predicted to be 0.36, 1.7
Many thanks to the Boeing Commercial Airplane Company, Fabrication Division for their support of this work. The opportunity to draw on the experience of BOeing engineers in titanium machining helped a lot.
REFERENCES
[l] B. M. KrameK and N. P. Suh, "Tool Wear by Solu- tion: A Quantitative Understanding," Trans. ASME, Journal of Engineering for Industry, Vol. 102, No. 3, November 1980, pp. 303-309.
[2] P. D. Hartung and B. M. Kramer, 'Theoretical Considerations in the Machininq of Nickel-Based Alloys,' in Cutting TOO^ Materials, F. Gorsler (ed.), ASM, 1981, pp. 57-76.
[31 P. D. Hartung, Tool Wear in Titanium Machininq, S.M. Thesis, Department of Mechanical Engineer- ing, M.I.T., June 1981.
141 E. Bickel, "The Temperature on a Turning Tool," International Research in Production Engineer- '3, ASME, New York, 1963, pp. 89-94.
[71
I81
I 9 1
[I61
1171
V. Compendium of Phase Diagram Data," Technical Report AFML-TH-65-2, Part V, May 1969.
M. Hansen, Constitution of the Binary Alloys, McGraw-Hill, New York, 1958.
R. Elliott, Constitution of the Binary Alloys, First Supplement, McGraw-Hill, New York, 1965.
N. Cook and P. Nayak, "Tne Thermal Mechanics of Tool Wear," Trans. ASME, Journal of Engineering for Industry, Vol. 88, No. 1, 1966, p. 93.
B. M. Kramer, Ph.D. Thesis, Department of Mech- anical Engineering, M.I.T., 1979.
R. C. Weast, ed., CRC Handbook of Chemistry and F'F';sfcs, 60th Edition, CRC Press, Florida, R. J. Wasileski and G. L. Kehl, "Diffusion of Nitrogen and Oxygen In Titanium," Journal of the Institute of Metals, vol. 83, 1954-55, pp. 94- 104.
W. B. Pearson, A Handbook of Lattice Spacings and Structures o f t a l s and Alloys, Volume 2, Pergamon Press, New York, 1967, p. 420.
E. L. Emerson, Wear Performance of Cubic Boron Nitride, S.B. Thesis, M.I.T. Department of Mech- anical Engineering, June 1981.
L. E. Davis et al., Handbook of Auger Electron Spectroscopy, Second Edition, Physical Electron- ics Industries, Inc., Eden Prairie, Minnesota, 1976.
A. D. Mcpuillan and M. K. Mcpuillan, Titanium, Butterworths Scientific Publications, London, 1956, p. 70.
L. M. Adelsberg and L. H. CadofE, "The Reactions of Liquid Titanium and Hafnium with Carbon," Transactions of the Metallurgical Society of AIME, Volume 239, June 1967, pp. 933-935. - Diffusion Data, Vol. 2, No. 3, Diffusion Infor- mation Center, Cleveland, Ohio, December 1968, p. 369.
I51 W. G. Moffatt, The HandDook of Binary Phase Dia- grams, General Electric Company, Schenectady, New York, 1981.
80