Machinability of Gray Cast Irons - AFS - Modern Castings - WebOnly0305

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 17

AFS Transactions 02-159 (Page 1 of 17 )

Machinability of Gray Cast Iron


R. D. Griffin
H.J. Li
E. Eleftheriou
C.E. Bates
University of Alabama, Birmingham, Alabama

Copyright 2002 American Foundry Society

ABSTRACT

The University of Alabama at Birmingham, in cooperation with the American Foundry Society, with companies across North
America and with support from the US Department of Energy, is conducting a project to develop an understanding of the
factors that control the machinability of gray and ductile irons. Machining tests were developed using a lathe and carbide
tools. Evaluation of the machinability of three continuous cast pearlitic gray irons of the same strength and hardness was
performed. The results showed that the wear rates of these irons varied by about five times even at the same strength level.
Since these results indicate that tool wear rates are not only determined by the strength and hardness of the irons that are
being machined, the goal of this paper is to determine the factors that contribute to these differences in machinability. The
higher wear rates were associated with decreased free carbon in the irons. Higher wear rates were also associated with higher
hard inclusion concentrations even though the total concentration of inclusions was very low.

INTRODUCTION

Many factors can influence tool life when machining iron. These include metallurgical conditions such as graphite size and
distribution, composition, ferrite/pearlite ratio, cooling rate from the eutectic through the eutectoid temperatures, and the
presence of either endogenous or exogenous inclusions.

Several factors that influence machinability are schematically illustrated in Figure 1. This figure represents a tool advancing
through a metal part containing a variety of graphite flakes and abrasive macro-and micro-inclusions that might include
oxides, carbides, nitrides, sand, and other materials. The advancing tool creates a compression zone below and ahead of the
tool rake and flank faces. The flow characteristics of the material is a function of the metal modulus, strength, work-
hardening coefficient, chip-forming characteristics, and metal ductility.

Rake face friction energy

Shear zone
Shear
energy

Macroinclusions
Microinclusions Tool

Compression zone Flank face


Friction energy

Workpiece

Fig. 1 Schematic of a Tool Advancing Through A Metal Part


AFS Transactions 02-159 (Page 2 of 17 )

Plastic deformation produced by the advancing tool in the workpiece generates heat that must be dissipated either through the
chip, workpiece, or the tool. The metal being removed also impinges on the tool rake face of the tool and produces frictional
heat. Under some circumstances, the heat and abrasion cause craters to develop on the tool rake face.

Several phases can be present in iron, and their volume and distribution have significant effects on tool wear. Massive
carbides formed during solidification are hard and can obviously degrade the machining characteristics by chipping or
breaking tool tips.

Some of the carbon dissolved in austenite during eutectic solidification must diffuse from the austenite and migrate to
graphite flakes or nodules as the metal cools to the eutectoid temperature. The presence of elements that inhibit carbon
diffusion reduces the rate of carbon transfer and produces austenite supersaturated with carbon. High cooling rates from the
eutectic to the eutectoid temperature may not provide enough time for the carbon to diffuse to the graphite. Supersaturated
austenite then decomposes in the eutectoid range to produce higher volumes of abrasive (micro)carbides in the pearlite
(Kovaks).

Inoculant additions and solidification rates are also important. These two factors have significant effects on both the eutectic
cell structure and the graphite size and distribution, which in turn affect the carbon diffusion distance and the chip forming
characteristics of the metal. The carbon must be able, in the time available as the iron cools from the eutectic to the eutectoid
temperature, to diffuse from the austenite and attach itself to the graphite flakes or nodules. Larger distances between
graphite flakes and nodules require more time for carbon diffusion to the graphite.

The graphite distribution also affects the mechanical strain that must be overcome at the tool tip and the chip forming
characteristics of the metal. The volume and distribution of the graphite may also affect the friction characteristics of the iron
in contact with the rake and flank faces of the cutting tool. The friction characteristics affect the amount of heat produced
during the shearing ahead of the tool tip and that, in turn, affects the tool temperature. Higher tool temperatures generally
cause faster tool wear.

Molding and metal handling practices can introduce oxides into the metal that abrade, wear, and chip cutting tools. Sand
grains picked up from the mold and incorporated into the metal or adhering to the surface of castings degrade machinability
because of their abrasiveness.

The Casting Engineering Laboratory at the University of Alabama at Birmingham, in conjunction with the American
Foundry Society and a number of industrial participants, are determining the factors that affect the machinability of cast
irons. Previously, the group has focused on determining the root causes of changes in machinability as evaluated by drilling
experiments (Bates). These machinability evaluations were performed by measuring the wear rates of high-speed steel drills
on test castings that were produced in commercial foundries. During the past two years, the group has expanded its
capabilities to include turning experiments using more wear resistant tool materials. The machinability of many commercial
castings can be evaluated with the new method. This paper presents the results of one such study on the machinability of
continuous cast gray iron. A companion paper presents similar results for a series of continuous cast ductile irons.
Continuous cast iron has the advantage of reduced porosity and inclusions so that catastrophic tool breakage can more easily
be avoided.

EXPERIMENTAL PROCEDURES

CASTING PRODUCTION
Three gray irons were produced for the study including a 67 mm diameter cylindrical bar designated G1 and two 77 mm
diameter cylindrical bars designated G2 and G3, respectively. The composition of the irons is given in Table 1.

To produce the bar stock, molten iron was placed in a holding crucible and pulled horizontally through a water-cooled
graphite die located near the crucible base. This process gives a solid outer rim and a molten core as the iron moves through
the die leading to a finer graphite and higher ferrite content in the rim than in the center of the bar. The irons were inoculated
with proprietary alloy containing manganese and zirconium. The irons were air-cooled, notched and broken off into six and
twelve feet lengths prior to shipment to UAB.

LATHE MACHINABILITY EVALUATION


Three-quarter inch by three-quarter inch square, un-coated fine-grained carbide (SPG 631) inserts were used to perform the
tool life (wear) experiments. The inserts were 0.1900 inches thick with an 18 degree taper. Before being used, each insert
was examined for cutting edge uniformity. The presence of a chipped edge or other anomalous feature caused the insert to be
AFS Transactions 02-159 (Page 3 of 17 )

discarded. The insert was placed on a gauge block with parallel sides such that three of the corners overhung the gauge block
and the fourth corner overhung a center hole on the block. This prevented any edge build-up on a cutting edge from affecting
the wear pattern determination.

Table 1. Composition of Experimental Gray Irons (Weight Percent)

Iron G1 G2 G3
CE 3.87 3.77 3.80
C 3.26 3.18 3.19
Si 2.330 2.290 2.310
P 0.056 0.026 0.055
S 0.048 0.052 0.043
Mn 0.690 0.360 0.700
Ni 0.054 0.050 0.054
Cu 0.060 0.040 0.050
Sn 0.045 0.011 0.045
Cr 0.042 0.038 0.043
Ti 0.007 0.008 0.008
V 0.007 0.004 0.008
Mo 0.008 0.007 0.008

The iron was sectioned into seven inch lengths and mounted in the hydraulic chuck in the lathe. Tool wear was initiated
using face cuts on the lengths. The flank of the tool showed progressively more wear as more metal was machined. The area
and shape of the wear land on each insert edge was recorded periodically by removing the tool from the CNC lathe and
digitally recording the wear pattern. The rate of flank wear was related to the machinability of the material by plotting the
average flank wear as a function of the volume of material removed.

Each insert was used until it had an average wear of 0.015 inches across the cutting edge or until it had been used to remove
at least 70 cubic inches of metal. Care was taken during handling to prevent inserts from chipping or shattering. Gross
fractures caused by aggressive cuts did not produce a natural wear pattern and the associated measurements were rejected.

The insert wear rates obtained with each material were determined using the following procedure:

1. The insert was examined under an optical comparator to ensure there were no chipped edges.

2. The insert was positioned on the gauge block so the cutting edge was parallel to the x-axis of the optical
comparator. The origin of the cutting edge was considered to be the extreme point on the cutting edge located at the
intersection of the horizontal and vertical axis. This location was established by rotating the cross-hair of the
comparator Y axis until it was aligned with the clearance plane of the edge. This rotation of the cross-hair was
temporary and necessary for establishing the origin. As the wear profile was traced, the cross-hair was rotated back
to a normal orientation. This procedure established a reference point and prevented edge wear at the tip of the insert
from affecting the measured flank wear area.

3. The coordinates of the datum were recorded so the same position could be found in successive measurements.

4. The amount of tool wear was measured as progressively larger volumes of metal were removed from the
experimental castings. The frequency of the measurements depended on the tool wear rate being experienced.

5. The inserts were catalogued and archived for future examination.

Turning experiments were performed using at least four speeds on each material, and triplicate experiments were performed
at each speed. All experiments were performed at the same feed per revolution (0.030 inch). The CNC lathe was
programmed to keep a constant surface speed as face cuts were made across the castings.

The insert wear land area was digitized from the original edge (which may have been worn off by successive passes) to a
length of 0.042 inches (engagement length corresponding to a depth of cut of 0.03 inches and an angle of approach of 45
degrees) using an optical comparator.
AFS Transactions 02-159 (Page 4 of 17 )

A schematic illustration of a worn insert is shown in Figure 2. Flank wear areas and lengths were measured, and the flank
area divided by the flank wear length provided the “average” flank wear.

Depth of Cut

Average Wear
Maximum
Wear Height

Fig 2. Schematic of Worn Tool


A representative tool wear curve for one of the irons is illustrated in Figure 3. The solid line drawn through each data set
represents the best linear fit to the insert wear data, and the dashed lines on either side of the solid line represent one standard
deviation in tool wear. Tool wear data obtained under specific conditions is summarized in the boxes on each of the graphs.
The data include the volume of metal machined away before the tool reached a specific amount of average wear, usually 6.5
mils, and the least squares linear slope of the wear curve. The data obtained with each insert edge is plotted with a different
symbol so the wear on any insert can be followed if desired.

CURVE F IT OF TURNING T OOL W E AR F OR GRAY IRON, G2


AT 6 5 0 S F M
0.025
A vg. V olume Removed @
6.5 Mils of W ear (in^3) =
6.28

Least S quares Linear


0.020
S lope
(Mils/100 in^3) = 89.6

Least S quares Linear


Intercept
0.015
Flank W ear (in)

W ear (in) = 0.000986

E dge 1

0.010 E dge 2

E dge 3

A vg. V ol. removed


0.005 at E nd
P ooled Linear Fit

One S td. D ev. of


Linear Fit
0.000
0 10 20 30 40 50
V o lu m e o f M etal R em o ved (in ^3)

Fig 3. Representative Tool Wear Curve for Gray Iron


AFS Transactions 02-159 (Page 5 of 17 )

SPECIMEN CHARACTERIZATION
One inch thick sections were removed from each casting for evaluation of the microstructure and certain physical and
chemical properties. Samples for microstructural evaluation were removed at ½ radius to avoid rim effects. The samples
were etched with Stead’s Reagent (2 grams CuCl2-2H2O, 8 grams MgCl2-6H2O 4 ml HCl 100 ml grain alcohol) and the
eutectic cell size and count were determined manually on 25X magnification optical microscope images. The cell count was
determined by counting the number of cells within each image while the cell size was determined using a line intercept count.

The volume percent graphite (Vv), ratio of surface area of graphite to volume of material (SV), mean spacing between
graphite flakes, and graphite flake thickness were made. Procedures for mounting and examining the specimens used for
microstructure have been previously published. (Griffin et.al.)

Scanning electron microscope (SEM) analysis using energy dispersive x-ray spectroscopy (EDS) was used to identify the
hard inclusions on nital etched samples. Quantitative measurements of the volume percent and number density of hard
inclusions including alloy carbo-nitrides, massive carbides and steadite, and complex carbides were performed on samples
that were etched with nital (4% nitric acid in methanol). The measurements were performed on one hundred images
collected at 500X magnification on an optical microscope. Measurements of the area of the inclusions were performed using
an image analysis system. The area of each alloy carbo-nitride found was estimated by measuring the diameter and
calculating a corresponding area. Each steadite and eutectic carbide identified was traced and the area determined with the
image analysis system.

Chemical compositions and combined carbon measurements were also performed on the irons. Mechanical properties
measured included tensile properties, compression properties, Brinell hardness measurements, and pearlite microhardness.
The foundry that produced the irons provided tensile results. Pearlite Vickers microhardness was measured on polished and
Nital etched samples using a diamond indenter with a 50g load. Images of the indents were acquired at 500X magnification
and measured using an image analysis system. Twenty measurements were made on each specimen but only the top ten
microhardness measurements were used for calculating average pearlite microhardness to attempt to avoid contributions from
graphite.

Brinell hardness measurements were made with a 10 mm diameter steel ball indenter and 3000 kg load on one inch thick
sections that had been ground plane parallel prior to indentation. An automated image analysis system was used to measure
the Brinell indentation diameters.

STATISTICAL ANALYSIS
Statistical analysis was performed to determine if there were significant differences between measurements on the materials.
A confidence level of 95% was the cut-off used for statistical significance. The data was also statistically analyzed for
factors that might influence machinability. While the overall machinability data that is presented is an average of the results
from three tools, the statistical analysis used the results for each tool. The analyses are based on the volume of material
removed at 650 surface feet per minute (sfm). Only linear and exponential fits between the factors and volume of material
removed were considered and only variables that had a p statistic at the 95% confidence level (p<0.05) or better were
considered to be a non-random event. In addition to the probability that a variable has a significant contribution to the
response being measured, the R2 statistic was also considered. The R2 statistic gives a measure of how much of the variation
in the dependent variable can be explained by the independent variable. Values of R2 less than about 25% are considered to
have a weak effect; values ranging from about 25% to 55% have a moderate effect; and values above 55% are considered to
have a strong effect.

RESULTS AND DISCUSSION

MACHINABILITY EVALUATION
The average volume removed of material for 6.5 mils of tool wear and the average linear wear slopes are listed in Table 2.
For each material, the average and standard deviation for three tools at each speed is presented. The slopes for the upper
speeds evaluated were quite steep which can lead to exaggeration of the differences in wear rates between tools that are
actually wearing at approximately the same rate. For this reason, the remainder of this paper will focus on the average
volume of material removed at 6.5 mils of wear because this measurement is less prone to error. A higher volume of material
removed at 6.5 mils of wear is an indication of better machinability.

A graph of the volume removed at 6.5 mils of wear versus turning speed is illustrated in Figure 4 with error bars representing
+ one standard deviation in the measurements. The volume removed decreased with increased turning rates and converged at
the speeds greater than about 650 sfm. There was no statistically significant difference in tool life as a function of either
AFS Transactions 02-159 (Page 6 of 17 )

speed or material above this speed. Higher cutting speeds produced heat at the tool tip at a higher rate and thermal effects
reduced the machinability and overwhelmed any differences in the irons.

Table 2. Machinability Results from Experimental Gray Irons

Turning Speed (sfm) Machinability Parameter G1 G2 G3


650 Linear Wear Slope (mils/100 in3) 105 (33) 91 (6) 137 (11)
800 Linear Wear Slope (mils/100 in3) 167 (49) 250 (166) 322 (192)
1200 Linear Wear Slope (mils/100 in3) 822 (283) 769 (237) 1099 (328)
1600 Linear Wear Slope (mils/100 in3) 878 (77) 586 (35) 1341 (375)
650 Volume Removed at 6.5 mils of Wear (in3)4.50 (0.74)6.35 (0.72)1.74 (0.80)
800 Volume Removed at 6.5 mils of Wear (in3)2.22 (0.40)1.76 (0.71)1.42 (0.61)
1200 Volume Removed at 6.5 mils of Wear (in3)0.90 (0.22)0.54 (0.53)0.55 (0.13)
1600 Volume Removed at 6.5 mils of Wear (in3)0.63 (0.14)1.07 (0.24)0.51 (0.18)

Note: Numbers in Parenthesis are Standard Deviations

Volume Removed at 6.5 mils of Wear vs. Turning Speed

8
Volume Removed at 6.5 mils of Wear (in3)

G1
4 G2
G3

0
500 700 900 1100 1300 1500 1700
Turning Speed (sfm)

Fig 4. Effect of Turning Speed on the Volume of Material Removed at 6.5 mils of Wear

Analysis showed a statistically significant difference in the machinability in the three irons at the 95 % confidence level at
650 sfm. The highest volume removed was found in the iron designated G2 with an average volume removed of 6.4 in3
followed by G1 with an average of 4.5 in3 and finally G3with an average of 1.7 in3.

MECHANICAL PROPERTY EVALUATION


There was little variation in the strength and hardness of the irons as illustrated in Table 3. The tensile strength in the iron
ranged from about 35 to 38 ksi. No correlation between the tool wear rate and the tensile strength was found.

The Brinell hardness ranged from 210 in iron G2 to 214 and 215 in irons G1 and G3, respectively. The pearlite
microhardness mirrored the Brinell hardness. The pearlite microhardness was lowest in sample G2 at 353 VHN and about
the 363 and 365 in the G1 and G3 irons. There was no statistically significant difference in the Brinell or pearlite
microhardness of the three irons.
AFS Transactions 02-159 (Page 7 of 17 )

Table 3. Mechanical Property Results from Experimental Gray Irons

Measurement Sample Number of Average Std. Dev.


Measurements
Ultimate Tensile G1 1 34,734 ---
Strength (psi) G2 1 37,830 ---
G3 1 36,656 ---
Brinell Hardness G1 5 214 6
G2 5 210 4
G3 5 215 8
Pearlite G1 Top 10 average of 363 14
Microhardness 20 measurements
(VHN) G2 Top 10 average of 353 10
20 measurements
G3 Top 10 average of 364 12
20 measurements

CHEMISTRY EVALUATION
A number of different parameters can be used to describe the amount and distribution of carbon in iron. These include the
carbon equivalent, the weight percent carbon, the weight percent free carbon, and the weight percent combined carbon. A
summary of the carbon equivalent, total carbon, the weight percent free carbon and the weight percent combined carbon is
presented in Table 4. The weight percent total carbon and combined carbon were measured and the free carbon was obtained
by taking the difference between the two.

There was some variation in the carbon equivalent and carbons in the three irons investigated. The range in carbon
equivalent was 0.17% and carbon varied by a total of 0.08%. The carbon equivalent was lowest in the iron G2 at 3.77%
followed by 3.80% in iron G3 and was highest at 3.87% in iron G1. The carbon concentration followed the ranking of the
carbon equivalent with the lowest amount of carbon in G2 at 3.18%, closely followed by G3 at 3.19% and the highest in G1
at 3.26%. No correlation between carbon equivalent or carbon concentration and the volume of metal removed at 6.5 mils of
wear with a turning speed of 650 sfm was found.

The distribution of the carbon varied significantly between the three irons. The amount of free carbon was the lowest in iron
G3 at 2.51 wt% while the combined carbon was highest at 0.68%. Intermediate free carbon and combined carbon
concentrations were found in iron G1 at 2.62% and 0.64%, respectively. The highest free carbon concentration was found in
iron G2 at 2.89%. This iron also had the lowest concentration of combined carbon with a value of only 0.29%.

Table 4. Carbon Equivalent, Carbon, Free Carbon and Combined Carbon in Experimental Gray Irons

Iron G1 G2 G3
Carbon Equivalent (wt%) 3.87 3.77 3.80
Carbon (wt%) 3.26 3.18 3.19
Free Carbon (wt%) 2.62 2.89 2.51
Combined Carbon (wt%) 0.64 0.29 0.68

Higher volumes of metal removed correlated with higher free carbons and lower combined carbons as illustrated in Figures 5
and 6. A strong correlation was found between the volume removed and the free carbon with a p value of 0.0012 and an R2
statistic of about 80% for a linear fit. The results were slightly weaker for the correlation between the combined carbon and
the volume removed which had a p value of 0.0075 and R2 of about 66% for a linear fit. These results indicate that the form
that the carbon takes in the iron can affect machinability.
AFS Transactions 02-159 (Page 8 of 17 )

Volume of Metal Removed at 650 sfm Vs. Free Carbon

Volume Removed at 6.5 mils of Wear (in^3)


7

R2 = 0.7975
4
p=-0.0012

0
2.45 2.5 2.55 2.6 2.65 2.7 2.75 2.8 2.85 2.9 2.95
Free Carbon (wt. %)
Raw Data Average Linear (Raw Data)

Fig 5. Effect of Free Carbon Concentration on the Volume of Metal Removed at 6.5 mils of Wear at 650 sfm

Volume of Metal Removed at 650 sfm Vs. Combined Carbon

8
Volume Removed at 6.5 mils of Wear (in^3)

5
R2 = 0.6631
p=0.0075
4

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Combined Carbon (wt. %)

Raw Data Average Linear (Raw Data)

Fig 6. Effect of Combined Carbon Concentration on the Volume of Metal Removed at 6.5 mils of Wear at 650 sfm

Statistical analysis was also performed on the other elements that varied between the three irons and these elements included
silicon, manganese, copper and tin. The silicon concentration varied from 2.29 to 2.33 wt. % while the copper concentration
varied from 0.04 to 0.06 wt. %. Neither element showed a significant correlation with the wear rates in the irons.
AFS Transactions 02-159 (Page 9 of 17 )

Statistically significant correlations of higher wear rates with higher manganese and tin concentrations were found in the
irons as illustrated in Figures 7 and 8. The manganese ranged from 0.36 to 0.7 wt. % and the p value was 0.0133 and the R2

Volume of Metal Removed at 650 sfm Vs. Wt. % Manganese

8
Volume Removed at 6.5 mils of Wear (in^3)

5
R2 = 0.6074

4 p = 0.0133

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Manganese (wt. %)

Raw Data Average Linear (Raw Data)

Fig 7. Effect of Manganese Concentration on the Volume of Metal Removed at 6.5 mils of Wear at 650 sfm

Volume of Metal Removed at 650 sfm Vs. Wt. % Tin

8
Volume Removed at 6.5 mils of Wear (in^3)

5
R2 = 0.5852
p=0.0163
4

0
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05
Tin (wt. %)

Raw Data Average Linear (Raw Data)

Fig 8. Effect of Tin Concentration on the Volume of Metal Removed at 6.5 mils of Wear at 650 sfm
AFS Transactions 02-159 (Page 10 of 17 )

concentration was 60% for a linear fit. The p value for the tin was 0.0163 and the R2 statistic was 59% with a range of 0.01
to 0.045 wt %.

MICROSTRUCTURAL CHARACTERIZATION
The microstructure of the samples consisted primarily of Type A graphite in a fully pearlitic matrix. Occasional small areas
of type D graphite and extremely low steadite concentrations were also found in the irons. Iron G2 had a few areas with trace
ferrite, and this iron was the most machinable iron of the three. Representative images of the samples are shown in Figures 9
through 11.

Fig. 9. Representative Microstructure of Iron G1, Nital Etch, 100X

Fig 10. Representative Microstructure of Iron G2, Nital Etch, 100X


AFS Transactions 02-159 (Page 11 of 17 )

Fig 11. Representative Microstructure of Iron G2, Nital Etch, 100X

Eutectic cell measurement results are presented in Table 5. There was no statistically significant variation in the cell count or
size between the three irons. The average cell count was between 7.1 and 8.1/mm2 and the average cell size ranged from 0.31
to 0.34 mm.

Table 5. Eutectic Cell Count and Size Measurements in Experimental Gray Irons

Measurement Sample Number of Average Std. Dev.


Measurements
Eutectic Cell G1 10 7.3 1.7
Count G2 3 7.1 2.4
(number/mm2) G3 7 8.1 1.1
Eutectic Cell Size G1 10 0.33 0.04
(mm) G2 3 0.34 0.03
G3 7 0.31 0.03

The results of the quantitative measurements of the graphite structure in the three irons are summarized in Table 6. The
range in the graphite results was limited although significant differences were found between the highest and lowest values in
all the measurements. The iron designated G1 had the highest volume percent graphite at 8.6% followed by G3 at 7.9% and
then G2 at 7.6%. No correlation between the volume percent graphite in these irons and the wear rate of the tools was
found.

The ratio in the surface area of graphite to the volume of material (SV) was 108.7 1/mm in sample G1, 89.3 1/mm in sample
G2, and 96.6 1/mm in sample G3. At the same volume percent graphite, a lower SV would be expected for an iron that was
primarily type A with large regularly shaped flakes while smaller or more irregularly shaped flakes would result in a higher
SV value. No correlation between machinability and graphite SV was seen in the irons.

Mean graphite spacing ranged from 34.4 µm in the iron G1 to 42.6 µm in iron G2. Iron G3 had an intermediate value of
42.6 µm. Mean flake thickness only ranged from 3.2 to 3.5 µm. The machinability values did not correlate at a statistically
significant level with either the mean graphite spacing or the flake thickness.
AFS Transactions 02-159 (Page 12 of 17 )

Table 6. Graphite Shape and Distribution Data in Experimental Gray Irons

Measurement Sample Number of Fields Average Std. Dev.


Volume Percent G1 30 8.6 2.0
Graphite G2 30 7.6 1.3
G3 30 7.9 1.6
SV Graphite G1 30 108.7 15.7
(1/mm) G2 30 89.3 16.0
G3 30 96.6 18.2
Mean Spacing G1 30 34.4 5.8
Between Graphite G2 30 42.6 7.4
Flakes (µm) G3 30 39.6 8.1
Mean Flake G1 30 3.2 0.5
Thickness (µm) G2 30 3.5 0.8
G3 30 3.3 0.6

Only two types of hard particles were found in the irons and these included steadite and an alloy particle that was most likely
a carbide. A representative optical microscope image of the steadite is illustrated in Figure 12. A representative optical
microscope image of an alloy carbide is illustrated in Figures 13. They appeared pale orange in color and typically had a
blocky shape. An EDS spectrum from one of the alloy carbides is shown in Figure 14. The inclusions were rich in
zirconium with lesser amounts of titanium also detected. Conclusive identification of the particles as a carbide was not
possible because EDS is insensitive to carbon. The zirconium in the irons may have originated from the inoculant as the
irons were inoculated with a proprietary alloy that containing zirconium.

Fig 12. Representative Optical Microscope Image of Steadite in Iron G3, Nital Etch, 700X
AFS Transactions 02-159 (Page 13 of 17 )

Fig 13. Representative Optical Microscope Image of Alloy Carbides in Iron G3, Nital Etch, 700X
Intensity

Energy
Fig 14. Representative EDS Spectrum from an Alloy Carbide in Iron G3

The number density and volume percent of inclusions in the irons is summarized in Table 7. In all three irons, the
concentration of hard inclusions was extremely low and would be rated at the “trace” concentration or less by simple visual
rating. However, careful measurement revealed that there were significant differences in the hard particle concentrations. In
all the irons, number density of the alloy carbides was higher than the number density of steadite but the volume percent of
steadite was higher than the volume percent of alloy carbides.

The total of the number density and volume percent of all the hard particles was calculated and is summarized in Table 7.
Statistical analysis showed a significant difference in the inclusion concentrations in the irons at the 95% confidence level.
The highest hard particle content as measured by both number density and volume percent was found in iron G3 with a
number density of 72/mm2 and a volume percent of 0.245%. Iron G1 had the next highest inclusion concentration with a
AFS Transactions 02-159 (Page 14 of 17 )
2
number density of 56/mm and a volume percent of 0.169%. The lowest concentration of hard particles was found in iron G2
with a number density of 11/mm2 and a volume percent was 0.029%. A strong correlation between machinability and the
number density and volume percent of hard particles in the casting was found. The volume of material that could be removed
before 6.5 mils of wear decreased with increasing hard inclusion concentrations as illustrated in Figures 15 and 16. For the
correlation between volume removed and the number density of hard particles, the p value was 0.0016 and the R2 statistic
was 78%. The correlation was even stronger with the volume percent of hard inclusions as the p value was 0.0006 and the R2
statistic increased to 83%. These results indicate that changes in the concentration of hard inclusions in the irons can lead to
reductions in machinability even at very low total concentrations.

Table 7. Number Density and Volume Percent of Hard Inclusions in Experimental Gray Irons

Iron G1 G2 G3

No./mm2 Alloy Carbides 31 (35) 7 (13) 30 (31)

No./mm2 Steadite 25 (27) 4 (10) 42 (44)

No./mm2 Total Hard Inclusions 56 (42) 11 (16) 72 (61)

Volume Percent Alloy Carbides 0.065 (0.123) 0.009 (0.024) 0.032 (0.050)

Volume Percent Steadite 0.104 (0.136) 0.020 (0.060) 0.212 (0.306)

Volume Percent Total Hard Inclusions 0.169 (0.173) 0.029 (0.063) 0.245 (0.316)

Volume of Metal Removed at 650 sfm Vs. Number Density of


Hard Inclusions

8
Volume Removed at 6.5 mils of Wear

5
R2 = 0.7798
(in^3)

4 p = 0.0016

0
0 10 20 30 40 50 60 70 80 90
Number Density of Hard Inclusions (#/mm^2)

Raw Data Average Linear (Raw Data)

Fig 15. Effect of Number Density of All Hard Inclusions on the Volume of Metal Removed at 6.5 mils of Wear at 650 sfm
AFS Transactions 02-159 (Page 15 of 17 )

Volume of Metal Removed at 650 sfm Vs. Volume Percent of


Hard Inclusions

Volume Removed at 6.5 mils of Wear (in^3) 7

R2 = 0.8334
6
p = 0.006

0
0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350
Volume Percent of Hard Inclusions

Series1 Series2 Linear (Series1)

Fig 16. Effect of Volume Percent of All Hard Inclusions on the Volume of Metal Removed at 6.5 mils at 650 sfm

The inclusion data is consistent with the free carbon/combined carbon data as illustrated in Figure 17. The combined carbon
concentration and the volume percent of hard particles increased together. This data may explain some unexpected results in
the pearlite microhardness. Typically, higher combined carbon concentrations are associated with higher pearlite
microhardness in cast irons. However, in these irons, even with a significant variation in the combined carbon content, no
differences in the pearlite microhardness was detected. It seems probable that some of the differences in the combined
carbon concentrations between the irons may be accounted for by increasing hard inclusion counts as the bulk of these hard
inclusions are carbides in the form of titanium carbo-nitrides, complex alloy carbides and steadite.
AFS Transactions 02-159 (Page 16 of 17 )

Volume Percent Hard Inclusions Vs. Combined Carbon

0.350

0.300
Volume Percent Hard Inclusions

0.250

0.200
R2 = 0.9952
p = 0.0447
0.150

0.100

0.050

0.000
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Combined Carbon

Fig 17. Relationship Between the Combined Carbon Concentration and the Volume Percent Hard Inclusions in
Experimental Gray Irons

SUMMARY AND CONCLUSIONS

The Casting Engineering Laboratory at UAB has developed the capability of testing the machinability of cast iron using
turning experiments and carbide tooling materials. This study was conducted to determine possible contributions to the
variation in machinability in three gray irons. The irons had approximately the same tensile strength, Brinell hardness,
pearlite microhardness and eutectic cell size.

1. The wear rates of the irons converged at turning speeds greater than 650 sfm. Above this speed, no significant
difference in the machinability of the irons was found.

2. At 650 sfm, the volume removed increased by about five times between the “easiest” and “hardest” to machine
irons even though the strength and hardness of the irons was approximately the same. These results indicate that tool
wear rates are not only determined by the strength and hardness of the irons that are being machined.

3. Higher wear rates in the irons was associated with lower free carbon concentrations and higher combined carbon.

4. Higher wear rates in the irons was associated with higher manganese and tin concentrations.

5. Iron G2 had trace ferrite in the sample examined. This iron was also the most machinable iron.

6. No correlation between the wear rates and the differences in the graphite structure was detected in the irons.

7. The total concentration of hard inclusions in the irons was quite small with the most prevalent hard inclusion
being steadite. Higher combined carbon concentrations were associated with the samples with the highest hard
inclusion concentrations.

8. Higher wear rates were found in the irons with the higher concentrations of hard inclusions as measured both by
number density and volume percent.
AFS Transactions 02-159 (Page 17 of 17 )

ACKNOWLEDGMENTS

We are deeply grateful for the company that interrupted their hectic production schedules to produce the castings used in this
study. We also truly appreciate the two companies that provide the combined carbon and chemical analysis for this work.
The contributions from these three companies are invaluable to this project. A special word of thanks also goes to the
American Foundry Society and our Industrial Sponsor's for their financial support.

We are also indebted to the US Department of Energy for support provided under agreement DE-FC07-97ID13555. Our hope
is that this information on the factors that affect machinability in gray iron will be of assistance to companies throughout the
United States in advancing the technology needed to produce machinable thin wall iron castings.

REFERENCES

Bates, C.E. Effects of Inoculation on the Machinability of Gray and Ductile Cast Iron An American Foundrymen's Society
Special Report, July 1998.
Griffin, R.D.; Scarber, P.; Janowski, G.M.; Bates, C.E. Quantitative Characterization of Graphite in Gray Iron, Transactions
of the American Foundrymen's Society V 104 Paper 96-129 P 977-983, 1996
Kovacs, B.V. Pearlite Stabilization in Cast Irons. American Foundrymen's Society Transactions 1981, p. 79-96.

You might also like