Optimization of Mechanical Properties in A356 Via Simulation And

Home , Gauge (instrument), Heat treating, Pattern (casting), Riser (casting), Sprue (manufacturing)

OPTIMIZATION OF MECHANICAL PROPERTIES IN A356 VIA

SIMULATION AND PERMANENT MOLD TEST-BARS

CHIA-JUNG CHEN

Submitted in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

Dissertation Adviser: Professor David Schwam

Department of Materials Science and Engineering

CASE WESTERN RESERVE UNIVERSITY

January, 2014 CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of

Chia-Jung Chen

Doctor of Philosophy candidate for the degree *.

John Lewandowski

David Schwam

Gerhard Welsch

Malcolm Cooke

11/15/2013 (date)

*We also certify that written approval has been obtained for any proprietary material contained therein. TABLE OF CONTENTS

Content Page

TABLE OF CONTENTS [1]

LIST OF TABLES [5]

LIST OF FIGURES [6]

ACKNOWLEDGEMENTS [12]

ABSTRACT [13]

1. Chapter 1: Aluminum Cast Alloys 1

1.1. Introduction 1

1.1.1 3xx Alloy 2

1.2. Intrinsic effects on mechanical properties of 3xx alloys (element effects) 3

1.2.1. Grain Refining of 3xx alloys 3

1.2.2. Factors Effect on Grain Size 4

1.2.3. Microstructural Modification 6

1.2.3.1. Modification of A356 alloy 7

1.2.3.2. Modification of 319 alloy 8

[1] Content Page

1.2.4. Heat Treatment of A356 Alloy 9

1.3. Extrinsic effects on mechanical properties of 3xx alloy (porosity and 14 inclusion effects)

1.3.1. Effect of Porosity on Mechanical Properties of A356 14

1.3.1.1. Hydrogen in Molten Aluminum Alloys (gas porosity) 17

1.3.1.2. Shrinkage Porosity 20

1.3.1.3. Element effects on shrinkage porosity 20

1.3.1.4.The Removal of gas porosity 21

1.3.2. Effect of Inclusions on Mechanical Properties 22

1.3.2.1. Measurement of inclusion content 22

1.3.2.2. Methods of Inclusion Removal from Molten Aluminum 25

1.4 Quality Index for Aluminum Alloy Castings 28

2. Literature Review of Effects on Mechanical Properties of A356 and 319 28 Alloys 2.1. Effect of Microstructural Modification and Heat Treatment on 32 Mechanical properties 2.2. Effect of SDAS (Grain Size) on Mechanical Properties 34

2.3 Effect of Porosity on Mechanical Properties 36

2.4. Inclusion Effect on Mechanical Properties 38

3. Experimental Procedure 40

[2] Content Page

3.1. Melt Preparation 40

3.2. Mold Preparation 42

3.3. Filter Preparation 48

3.4. Reduced Pressure Test 49

3.5. PoDFA 50

3.6. Experimental Cooling Rate 50

3.7. Tensile Test 54

3.8. Fatigue Test 55

3.9. Magma Simulation 57

4. Results and Discussions 59

4.0. Hypotheses and Objectives 59

4.1. Characterization of Defects 59

4.1.1. Reduced Pressure Test 59

4.2. Improvement in Mechanical Properties of A356 Alloy 61

4.2.1. Gating System Effect on Mechanical Properties 62

4.2.1.1. Sprue Size Effect 62

4.2.1.2. Effect of the Knfie Ingate 72

4.2.2. The Effect of Coating the Gage Section 79

[3] Content Page

4.2.3. Filtration Effect 82

4.2.4. Best Mechanical Properties of A356 Alloy from Each Mold 83

4.2.5. Quality Index 88

4.3. Improvement in Mechanical Properties of 319 Alloy 91

4.3.1. Effect of Mg and Mn addition in 319Alloy 91

4.3.2. Effect of the Knife Ingate 92

4.3.3. Filtration Effect 95

4.4. Improving Fatigue Properties with HIP 97

5. Conclusions 100

6. References 103

[4] LIST OF TABLES

Table Page

1.1. Series of Cast and Wrought Aluminum Alloys 1

1.2. Composition of A356 and 319 Alloys 2

1.3. Heat Treating Temper Code 11

1.4. Typical Inclusions in Aluminum Cast Alloys 16

1.5. Summary of trends in porosity distribution and level 21

2.1. Gas Levels, Solidification Rates and Tensile Properties of CA-B135-T6 38 Sand Cast Step Castings

Table Page

2.2. Inclusion/oxide - Mechanical Properties Relationships for A356.2 Alloy 39

4.1. Reduced pressure test relationship to Alspek reading of A356 alloy 60

4.2. Compositions of A356 Alloy in This Study 61

4.3. Compositions of 319 Alloy in This Study 91

[5] LIST OF FIGURES

Figures Page

1.1. Phase diagram of aluminum and silicon alloy 3

1.2. Hall-Petch relationship diagram 4

1.3. Grain refiner effect on undercooling 5

1.4. Ti content effect on average grain size 5

1.5. Polarized light micrographs showing the effect of B addition on the A356 grain structure 6

1.6. Typical microstructure of as-cast A356 alloy 7

1.7. Eutectic Si structure observed in 319 alloy 8

1.8. SEM microstructures observed in the 319 type aluminum alloys 9 after T6 heat treatment

1.9. Artificial aging of aluminum alloys 12

1.10. SEM micrograph of A356 alloy solutionized in conventional furnace for 6 hours at 1000oF and aged at 340oF 13

1.11. BFTEM image and [001] SAD pattern from the 319 alloy 13 1.12. Typical shrinkage pore surrounded by dendrites and eutectic phase 15

1.13. Typical gas pore 15

1.14. Effect of temperature on solubility of hydrogen in pure aluminum 17

1.15. A schematic figure of reduced pressure test apparatus 19

1.16. Cross section of RPT samples 21

1.17. Schemetic picture of degassing device 22

[6]

Figures Page

1.18. PoDFA (Porous disc filtration apparatus) 23

1.19. Typical micrograph of a sectioned PoDFA sample 24

1.20. Higher magnification on cake section 24

1.21. Effect of time and temperature on oxidation of aluminum 26

1.22. Schematic figure of a fluxing system 26

1.23. Modes of inclusions capture 27

1.24. Yield strength-elongation relationship of A356/A357 alloys 29

1.25. Schematic illustration of the quality index, QT, as the ratio of the experimentally observed elongation as a fraction of the maximum possible 30

elongation QT = eF/eF(max) at the given level of yield strength, σY.

1.26. The nomogram with iso- QT lines for A356/A357 alloy castings shows four distinct regions 31

2.1. UTS and elongation as a function of solution time 33

2.2. Ultimate tensile strength and % elongation of Mg-free and Mg-containing 34 samples

2.3. The average UTS and elongation of cast A356-T6 aluminum alloy curves 35

2.4. S/N curves of E319 with low SDAS and 70μm SDAS 36

2.5. Ultimate tensile and yield strengths versus gas content and per cent voids for 37 remelted A356 alloy 3.1. Electric resistance furnace 40

3.2. Recycled A356 aluminum 41

3.3. Degassing treatment 41

[7] Figures Page

3.4. Fluxing and skimming 42

3.5. ALSPEK hydrogen system 42 3.6. Electrical heater 43

3.7. Oxy-hydrogen flam 43

3.8. Standard Sthal mold 44

3.9. Stahl mold with large round shape sprue (Stahl HS Mold) 44

3.10. Case mold version.1 45

3.11. Case mold version.2 46

3.12. Case mold version.3 46

3.13. Schematic of the cast bars from Case mold version.3 and version.4 47

3.14. Grinding of the knife ingate 47

3.15. In-mold filters 48

3.16. Top view of the in-crucible filter box (black) immersed in Al 48 3.17. RPT test equipment 49

3.18. PoDFA equipments 50

3.19. (a) Dimensions of the step casting and (b) Illustration of the set-up for 51 cooling rate measurement

3.20. Respective cooling rates 52

3.21. Typical cooling rates for various casting processes 53

3.22. Sthal and Case mold version.1 and step mold test bars 54

[8] Figure Page

4.1. PoDFA results for five different conditions 61

4.2. The fracture surface of Case mold v1 test bar 63 4.3. The fracture surface of a Stahl mold test bar 64 4.4. Reduction in melt front velocity of Case Mold v.2 65

4.5. Standard Stahl mold and Stahl HS mold 66

4.6. Filling time in Stahl mold and Stahl HS mold 67

4.7. Melt front velocity of Stahl mold and Stahl HS mold 68

4.8. The effect of sprue size on mechanical properties of A356 alloy on Stahl 69 mold

4.9. Sprue size of Case mold v.2 and v.3 70

4.10. Effect of sprue size on mechanical properties of A356 alloy on Case mold 71

4.11. Macrograph of Stahl mold and Stahl HS mold 72 4.12. Case mold v.3 and v.4 73

4.13. Solidification time of Case mold v.3 and v.4 74

4.14. Micro porosity of Case mold v.3 and v.4 74

4.15. SDAS of Case mold v.3 and v.4 75

4.16. The effect of the knife gate on mechanical properties of A356 alloy 76

4.17. Fracture surface of Stahl mold and Case mold v.4 77 4.18. Shrinkage porosity in Stahl mold test bars 78

4.19. Perfectly broken of Case mold V4 test bars 78

4.20. Test bar after removal of the knife ingate 79

[9] Figure Page

4.22. The micro-porosity at gage section of Case mold V4 with two kinds of 81 coating

4.23. Applying a coating on the gage section effects on the fracture surfaces of 82 tensile test bars of A356 alloy in different molds

4.24. In-mold filter effect on mechanical properties of A356 alloy 83

4.25. Schematic of the step mold and step casting 84

4.26. Micro-porosity prediction in three different molds 85

4.27. SDAS prediction for three different molds 85

4.28. Mechanical properties of samples from the 2” Step Casting with and w/o 86 HIP

4.29. Best mechanical properties for test bars cast in different molds (T6 87 condition)

4.30. Mechanical properties of 356-T6 alloy as a function of iron content and 89 aging time.

4.31. Data from this study in QT lines for A356 alloy casting 90

4.32. SEM images of low and high Mn content 319 alloy 92

4.33. Fracture surface of Stahl mold and Case mold v.4 test bar 92

4.34. The effect of Knife ingate on alloy 319 in the as-cast condition 93

4.35. Effect of knife ingate on properties of 319 alloy in T6 condition 94

4.36. Effect of knife ingate on alloy 319 in T7 condition 94

4.37. Effect of in-mold filter on alloy 319 in the as-cast condition 95

4.38. Effect of in-mold filter on alloy 319 in T6 and T7 conditions 96

[10] Figure Page

4.40. Fatigue properties of Step Mold 1” sample with and without HIP 98

4.41. Dimensions of fatigue test bar 99

[11] Acknowledgements

I would like to thank Professor Schwam for lending his insight and experience to this project. I am proud to have been given the opportunity to work with him. Also, I would especially like to thank Dr. David Neff and Dr. Xuejun Zhu for all of his time and effort while being an invaluable help thorough the length of this work. My thanks also to our foundry staff, Rich Tomazin and Rich Miller for all of their hard work and help along the way.

I would like to thank the American Foundry Society whose funding made this work possible.

Finally, I would like to thank my family and friends who provided support throughout my pursuit of a Ph.D. degree.

[12] Optimization Of Mechanical Properties In A356 Via Simulation

And Permanent Mold Test-Bars

CHIA-JUNG CHEN

Abatract

Using as-cast test bars is a quick and convenient method of determining as-cast

metal quality in the foundry--although such results are only representative of the

section of a casting solidifying at the same rate as the test-bar. Unfortunately the

current standard test-bar mold suffers from shrinkage porosity which detracts from

best properties. In this work computer simulation has been utilized to predict and

design an improved permanent mold test bar mold. A356 and 319 alloys have been

melted and treated with best metal cleaning practices (degassing, filtration) in order to

assess the effect of clean metal as a baseline of this research prior to microstructural

enhancements (modification, grain refinement, SDAS).

The results show that with a knife-ingate in the re-designed test-bar mold, better

as-cast properties can be obtained on a more consistent basis throughout a varying

mold temperature range. With best in-furnace clean metal practice and virgin ingot,

applying filters in the test bar mold have minimal effect, but show that filtration is

beneficial if melting recycled metal.

In the standard heat treated T6 condition, the new test-bar mold delivers superior mechanical properties as measured by the Quality Index method.

[13] Chapter 1: Aluminum Cast Alloys

1.1 Introduction

In industrial applications, aluminum alloys can be divided into two groups. One is cast aluminum alloys, the other is wrought aluminum alloys. Cast aluminum alloys are used to cast final products; wrought aluminum alloys are cast into ingots or billets, then formed by rolling, extrusion or forging into final products. Table.1 lists the main designations of cast and wrought aluminum alloys. The designation provides key information on the alloy. For instance, 1xx and 1xxx series means the aluminum content exceed 99%; 2xx and 2xxx series means that copper is the primary alloying element in these alloys.

Table.1 Designation of cast and wrought aluminum alloys

Cast Prime Composition Wrought Prime Composition

1xx 99% Al 1xxx 99% Al

2xx Cu 2xxx Cu

3xx Si, Cu, Mg 3xxx Mn

4xx Si 4xxx Si

5xx Mg 5xxx Mg

6xxx Mg, Si

7xx Zn 7xxx Mg, Zn

8xx Li 8xxx Li

1.1.1 3xx Alloy

3xx series alloys are widely used in the aluminum casting industry. Silicon is the primary alloying element in this alloy, added to increase fluidity and strength.

A356 and 319 alloys were used in this study and their compositions are listed in

Table.2. A356 is easy to cast in complex shape products. Silicon can also improve strength of aluminum alloy. Besides silicon, A356 alloy contains 0.3% magnesium which can precipitate Mg2Si with silicon to become a strengthening phase by heat treatment;

319 alloy also has good castability; it contains about 3% copper and precipitates Al2Cu as a strengthening phase upon heat treatment.

Table.2 Composition of A356 and 319 alloys

Alloy/Element Si Mg Cu Mn Fe

A356 7.0% 0.3% 0.2% 0.1% 0.2%

319 6.5% 0.1% 3.5% 0.1% 0.6%

Fig.1 shows the phase diagram of aluminum-silicon alloys. Generally, silicon content in aluminum alloy will not exceed the eutectic point (silicon content 12.6%) to avoid formation of a primary silicon phase. The primary silicon phase is very hard, therefore will reduce machinability of aluminum alloys.

Fig.1 Phase diagram of aluminum-silicon alloys

1.2 Intrinsic Effects on Mechanical Properties of 3xx Alloys (element effects)

1.2.1 Grain Refining of 3xx Alloy

Generally, the strength of cast metals is related to grain size via the Hall-Petch equation

-1/2 σ = σ0 + kD Equation.1

Where σ0 is the yield stress; σi is the friction stress, or overall resistance of the crystal to

dislocation movement; k is the locking parameter, or hardening contribution of grain

boundaries. Hall-Petch strengthening is a method of strengthening materials by reducing

their average grain size. It is based on blocking the motion of dislocations at grain

boundaries, thus decreasing the number of dislocations present in grains. As the

dislocations motion from grain to grain becomes more difficult, the material would

become stronger. Fig.2 shows the Hall-Petch relationship. In this study, the grain size of

A356 alloy is generally larger than 10nm, therefore we will not consider the inverse of

Hall-Patch relationship.

Fig.2 Hall-Petch relationship diagram

1.2.2 Factors Affecting Grain Size

Adding grain refiners is commonly practiced in cast aluminum alloys. The mechanism of grain refining is to produce a heterogeneous solid phase which can make grain nucleation easier in the melt. Undercooling occurs often during the solidification

process. Generally, a large undercooling temperature will lead to larger grain size.

Adding grain refiner can eliminate undercooling, resulting in grain refinement.

Conventional grain refiners, different compounds of Ti, such as TiB2 and TiAl3 are used as grain refiners in this study. Fig.3 shows the effect of the grain refiner on undercooling. The more the titanium and boron present in A356 alloy, the less the undercooling. Fig.4 shows the effect of titanium content on average grain size. Over 800 ppm of titanium can lead to a grain size of 600 µm. Fig.5 shows the decrease in grain size with boron increase in A356 alloy [1].

Fig. 3 Grain refiner effect on undercooling [1]

Fig.4 Effect of Ti content on average grain size [1] 5

Fig.5 Polarized light micrographs showing the effect of B addition on the A356 grain structure: (a) without addition, (b) 63 ppm, (c) 254 ppm, and (d) 1265 ppm [1]

1.2.3 Microstructural Modification

Modification is commonly practiced to enhance mechanical properties of cast aluminum alloys. The purposes of modification is to change the needle shape phase of silicon into a spherical shape or change the needle shape secondary phase into a different component phase with a different shape. It is easy to image why brittle, needle-shaped Si particles present in a cast alloy can act as stress concentration sites and cause it to break in a brittle manner. In the case of A356 alloy, we are trying to modify the silicon needle phase into a spherical shape; in 319 alloys, we are modifying the needle silicon phase into a spherical morphology and the needle β phase into Chinese script-shape α phase.

1.2.3.1 Modification in A356 alloy

Fig.6 shows the typical microstructure of an A356 alloy. As shown in Fig.6 (a), the eutectic silicon phase is needle-shape like. In a tensile test, the fracture initiates readily at these silicon particles, because these silicon needles are very brittle. In order to improve properties, sodium or strontium are used as modifiers in A356 alloys. Generally, 0.01% sodium or 0.02% strontium is used. After modification, the silicon needles become powder-like, as shown in Fig.6 (b). By modification, the mechanical properties are improved significantly, especially the elongation [2].

Fig.6 Typical microstructure of as-cast A356 alloy [2]

1.2.3.2 Modification in 319 alloys

S. El-Hadad et al. [3] discussed the effect of Bi addition on the microstructure of

319 alloy. Fig.7 shows an eutectic silicon structure modified by strontium in a low magnesium-content 319 alloy. Starting from a well modified eutectic silicon region in

Fig.7(a), additions of bismuth coarsen the silicon structure with bismuth increases from 0 ppm to 1555 ppm, Fig.7(B); when bismuth is increased from 1555 ppm to 6060 ppm, the silicon structure will be refined, Fig.7(C). Similar modification is observed with Ca additions to 319 alloy.

Fig.7 Eutectic Si structure observed in 319 alloy with (a) 0 ppm,(b) 1555 ppm, and (c) 6060 ppm Bi additions [3].

J. Y. Hwang et al. [4] investigated the effect of Mn addition on the microstructure of

319 alloy. It is known that the Mn/Fe ratio larger than 5 can modify the β-Al5FeSi phase effectively [5]. Fig.8(a) shows a microstructure of 319 alloy with only 0.02% manganese addition after T6 heat treatment. Plate-like β phases (Al5FeSi) are presented in this figure.

It is well known that the intermetallic β-Al5FeSi phase that forms during the solidification of type 319 alloys is detrimental to the mechanical properties because it is

both brittle and exists as thin plates [6]. For 0.35% Mn addition in the 319 alloy, Fig.8(b), the amount of the plate-like β phases is reduced and Chinese script α phases

(Al15(Fe,Mn)3Si2) are present. When Mn is increased to 0.65% and 0.85%, β-Al5FeSi phases almost disappear and the α-Al15(Fe,Mn)3Si2 phases become larger, as illustrated in

Fig.8(c) and Fig.8(d).

Fig.8 SEM microstructures observed in the 319 type aluminum alloys after T6 heat treatment. (a) 0.02%Mn, (b) 0.3%Mn, (c) 0.65%Mn, and (d) 0.85%Mn sample showing the α and β intermetallic phases [6].

1.2.4 Heat Treatment of A356 Alloy

Precipitation hardening heat treatment is the most useful process to strengthen aluminum alloy. Aluminum cast alloys, 2xx, 3xx, 7xx series cast alloys and 2xxx, 6xxx,

7xxx series wrought alloys can be strengthened by precipitation hardening heat treatment.

Table.3 shows 10 codes used to designate heat treatment conditions. The most commonly used is T6 condition which can obtain the highest strength in aluminum alloys. T6 temper generally can be divided into three steps:

Solution treatment is the first step in the precipitation hardening process. The alloy is heated above the solvus temperature and soaked there until a homogeneous solid solution is formed. The precipitates (θ or β) and segregation are re-dissolved into the matrix.

Quenching is the second step in which the aluminum alloy is rapidly cooled from α phase to form a supersaturated solid solution. This solid solution is not an equilibrium phase.

Artificial aging is the third step applied to precipitate supersaturated elements into the aluminum alloy matrix to strengthen it. The aluminum is keep under a tight temperature window (around 320~340°F) for a few hours. For example, the aging temperature for the

A356 alloy is 320°F and the aging time is 4~6 hours to obtain optimum strength. Fig.9 shows the aging time-strength relationship for aluminum-copper alloy. The sequence of particle precipitation with time in aluminum-copper alloy is GP-Zone→ θ” phase→ θ’ phase→ θ phase. As this figure illustrates, the best condition is between θ” and θ’ precipitated in the alloy When the GP-Zone is too small and θ is too big they have less ability to block dislocations than θ” and θ’ [7]. 10

Table.3 Heat treating temper codes

Code Content

Cooled from an elevated temperature shaping process and naturally T1 aged to a substantially stable condition

Cooled from an elevated temperature shaping process, cold worked, T2 and naturally aged to a substantially stable condition

Solution heat treated, cold worked, and naturally aged to a T3 substantially stable condition

Solution heat treated, and naturally aged to a substantially stable T4 condition

Cooled from an elevated temperature shaping process then artificially T5 aged

T6 Solution heat treated then artificially aged

T7 Solution heat treated then and overaged/stabilized

T8 Solution heat treated, cold worked, then artificially aged

T9 Solution heat treated, artificially aged, then cold worked

Cooled from an elevated temperature shaping process, cold worked, T10 then artificially aged

Fig.9 Artificial aging of aluminum alloys [7]

Fig.10 shows a high magnification picture under SEM. Needle-like Mg2Si phases are present in the A356 alloy matrix. The maximum mechanical properties are controlled by the heat treatment condition to produce uniform needle-like Mg2Si phases in the alloy [8].

In the case of 319 alloy, it is known that needle like θ’-Al2Cu phase is the strengthening phase after T6 heat treatment [9]. Fig.11(a) shows a TEM image of Mg- free 319 alloy after T6 heat treatment. Needle like θ’-Al2Cu phases are present uniformly along [001] direction. Fig.11(b) shows Mg-containing alloy after the T6 heat treatment.

The additional Mg elements produce Q phases after T6 heat treatment, showing as black spots in the image. Arnberg et al. [10] reported that the composition of the Q phase was approximately Al4Cu2Mg8Si7. The Q phases could increase tensile strength dramatically after T6 heat treatment as will be discuss in Chapter 3.

Fig.10 SEM micrograph of A356 alloy solutionized in a conventional furnace for 6 hours at 1000oF and aged at 340oF [8]

Fig.11(a) BFTEM image and [001] SAD pattern from the Mg-free alloy (b) BFTEM image and [001] SAD pattern from the Mg-containing alloy after the T6 heat treatment[10].

1.3 Extrinsic Effects on Mechanical Properties of 3xx alloy (porosity and inclusion effects)

1.3.1 Effect of Porosity on Mechanical Properties of A356

Metal cleanliness is considered as a key factor in the performance of the next generation of metal castings. Reduction of porosity and inclusions produced during casting can significantly increase the mechanical properties of cast products.

Gas porosity and shrinkage porosity are two of the most common defects in cast aluminum alloys. Gas porosity is formed when gas becomes entrapped in the molten metal. This type of porosity is usually spherical in appearance. It can be avoided through proper degassing of the molten metal, proper gating design and good pouring practices.

Shrinkage porosity is caused by the volumetric contraction which taken place when molten aluminum solidifies. The shape of shrinkage porosity is irregular and usually interdendric. Increasing the cooling rate can reduce the size, and form a more uniformly dispersed shrinkage porosity; this will improve the fatigue properties of the cast aluminum. Fig.12 and Fig.13 show typical shrinkage and gas porosity respectively [11].

Fig.12 Typical shrinkage pore surrounded by dendrites and eutectic phase [11]

Fig.13 Typical gas pore [11] 15

Inclusions may comprise dross, oxides, salts or sludge. Typical inclusions present in aluminum alloys are shown in Table.4 [12]. Generally, most inclusions exhibit complex structures, are hard and brittle. In most cases, inclusions larger than 10 microns to 20 microns may have an adverse effect on mechanical properties. If properly practiced, fluxing and filtration can remove inclusions from the molten metal.

Table.4 Typical inclusions in aluminum cast alloys [12]

1.3.1.1 Hydrogen in Molten Aluminum Alloys (gas porosity)

Hydrogen is the only gas which is soluble in aluminum alloy melts. The solubility of hydrogen in aluminum alloy melts is depended on temperature. Fig.13 shows the effect of temperature on solubility of hydrogen in pure aluminum [13]. When the temperature exceeds 1220F (melting point), the hydrogen solubility increases dramatically. Hydrogen gas is produced from the dissociation of water vapor in the atmosphere. The hydrogen gas problem is therefore more severe in hot and humid environments.

During solidification, as the casting cools down, the solubility of hydrogen in aluminum drops and the gas is rejected from the metal, forming porosity in the aluminum casting. This type of porosity is also known as gas porosity. The detrimental effect of gas porosity on mechanical properties (tensile strength, elongation) of aluminum alloys is well documented.

Fig.14 Effect of temperature on solubility of hydrogen in pure aluminum [13]

Predeterminations of gas content

Reduced Pressure Test

The Reduced Pressure Test (RPT) is one of the most commonly used methods to assess the hydrogen level in the aluminum alloys. In RPT, also known as the vacuum density test (VDT) a sample is usually taken from molten aluminum (between 100 and 200 grams) with a metal or ceramic crucible, and allowed to solidify under vacuum. Fig.15 shows a typical RPT apparatus. As the sample solidifies under vacuum, hydrogen bubbles nucleate because of the reduced pressure. Fig.16 shows typical voids with respect to each density and hydrogen level. Since porosity is formed at lower pressure, the volume of these voids is larger than the voids formed under atmosphere for the same gas content. By this method, it is relatively simple to detect and quantify the hydrogen level from a small sample even if the hydrogen level is relatively low.

Calculation of Gas Content

The volume of gas porosity in the sample can be calculated by comparing the density of the measured sample with the theoretical density of the same aluminum alloy [14].

Let D=Density of sample D=mass/volume

D0=Standard density of the sample (gas free)

Mair=Mass of sample in air

Mwater=Mass of sample in water

Volume of sample =Mair-Mwater

Volume of Porosity=Mair/(D-D0)= Mair/[( Mair/ Mair-Mwater)-D0]

Fig.15 Schematic figure of the reduced pressure test apparatus [14]

Fig.16 Cross section of RPT samples [14]

1.3.1.2 Shrinkage Porosity

Shrinkage porosity is usually found at under-riser and centerline locations in the castings. The obvious reason is such last-to-solidify zone is usually surrounded by completely solidified metal that blocks additional feeding with molten metal. As the last- to-solidify metal cools down, the volume contraction cannot be compensated without extra feeding, leading to porosity.

Shrinkage porosity is also detected frequently near the junction of a thick and thin section and at the center of fairly thick cast parts. In this case it is the difference in cooling rates that drives the formation of shrinkage porosity by preventing adequate feeding to the last-to-solidify regions.

1.3.1.3 Elemental effects on Shrinkage Porosity

A. Knuutinen et al. investigated Porosity formation in aluminum alloy A356 modified with Ba, Ca, Y and Yb. All additions increased the porosity level compared to the unmodified alloy and it increased with increased addition level. Additions of Ca and

Y caused porosity to become increasingly concentrated in the hot spot. Additions of Ba and Yb resulted in small, round, dispersed porosity. These results indicate a strong correlation between porosity amount and distribution and the eutectic solidification mode

[15],[16].

Table.5 Summary of trends in porosity distribution and level, eutectic solidification modes and proposed eutectic solidification mode in the hot spot of the casting [16]

1.3.1.4 Removal of Gas Porosity

Degassing

As it is well-established that hydrogen produces gas porosity, degassing of molten aluminum has become an essential step in molten metal treatment for attaining high quality castings.

Degassing systems are based on injecting an inert gas (nitrogen or argon) into the bottom of the melt. When the nitrogen or argon gas is injected into the melt, the differential in hydrogen pressure will cause hydrogen start to diffusing into the ascending gas bubbles. If we use a lance to inject inert gas, the degassing process is as efficient. A better result is obtained with a rotary impeller degasser, because relatively fine and

evenly spread bubbles are produced. A schematic picture of degassing device is shown in

Fig.17 [17].

Fig.17 schemetic picture of degassing device [17]

1.3.2 Effect of Inclusions on Mechanical Properties

1.3.2.1 Measurement of Inclusion Content

PoDFA

PoDFA (Porous disc filtration apparatus) identifies the inclusions and measures their concentration in the melt for each type of inclusion. The PoDFA method is based on filtration of a predetermined quantity of liquid aluminum through a very fine high- porosity disc under vacuum condition. Due to the pressure differential, the liquid

aluminum can easily penetrate the filter. However, any inclusions present in the molten metal are retained as a “cake” on top of the filter. Fig.18 (a) shows the PoDFA equipment and Fig.18 (b) shows the schematic picture of the PoDFA. In Fig.18 (b), the left image shows the aluminum melt before filtration; the right image shows it after filtration. After filtration, the inclusion will remain on top of the filter disc. The inclusion concentration can be determined by sectioning and image analysis of the filter disc and the “cake”.

Fig.19 shows a typical micrograph of the inclusions “cake” on the filter disc. The trapped inclusions are clearly visible at the bottom of the micrograph [18]. The inclusion concentration is calculated by the following equation:

…...Equation.2

Fig.18 PoDFA (Porous disc filtration apparatus) [18]

Fig.19 Typical micrograph of a sectioned PoDFA sample [18]

Fig.20 higher magnification on cake section [18]

1.3.2.2 Methods of Inclusion Removal from Molten Aluminum

Fluxing

Metal cleanliness can be improved by proper use of flux. Aluminum oxide (Al2O3) particles and films form rapidly on the surface of the molten aluminum during melting and holding. Fluxes can be used to remove these oxides. Although the density of oxides is higher than liquid aluminum, they tend to accumulate at or near the melt surface because of surface tension effects and adsorbed gases. Fig.21 shows that oxidation markedly increases with the temperature. Drossing fluxes are the most economical. Dross may contain more than 80% metal suspended in less than 20% oxide; Treated with drossing fluxes dross changes its appearance from wet to dry as its metallic aluminum content decreases. These exothermic fluxes release oxygen and generate heat by combusting a portion of the metallic aluminum in the dross. Drossing fluxes are added either by weight at about 0.2-1% of the metal charged or by melt surface area at about

0.5lb/ft2. The amount of flux needed depends on the cleanliness of the charge materials and how much dross is already present [19].

Fig.22 shows a modern method of flux delivery called flux injection. The traditional method of spreading flux onto the melt surface is not as efficient since the flux stays predominantly on the surface rather than mixing with the melt. Flux injection can overcome this limitation by delivering the powdered flux underneath the melt surface.

Typical flux injection equipment involves a dry powder feeder that mixes powdered flux into an inert gas stream carrying it through a lance immersed in the melt. Therefore, flux injection not only removes inclusions but also performs degassing [20].

Fig.21 Effect of time and temperature on oxidation of aluminum [19]

Fig.22 Schematic figure of a fluxing system [20]

Filtration

Foam filters made of ceramics are an effective way to remove inclusions from aluminum alloys. Alumina (Al2O3) particles and films are the most common and readily formed inclusion in molten aluminum. Al2O3 is formed due to contact of molten aluminum with air if during casting. If the metal pouring velocity is relatively high, more turbulence will be produced trapping air in the molten metal. Excessive turbulence will generally lead to formation of more Al2O3. Fig.23 shows the melt flow pattern through a filter. The originally turbulent flow before the filter becomes laminar flow after passing through the filter and being slowed down [21].

Fig.23 modes of inclusions capture [21]

1.4 Quality Index for Aluminum Alloy Castings

Drouzy, Jacob and Richard indicated that a linear relationship between ultimate tensile stress (UTS) and log(eF) in the underaged condition up to the peak strength [22].

They introduced the quality index, Q, for underaged and peak-aged alloys as:

Q = UTS + c log(eF) Equation.2

Where c is a constant (=150 MPa). The authors also indicated some processes can affect yield strength and the quality index with minor modifications to chemical composition, heat treatment and solidification rate.

Another quality index, QT, was proposed by M. Tiryakioglu and J. Campbell based on the concept of ductility potential for cast aluminum alloys [23-26]. The authors used extensive data taken from hundreds of tensile tests excised from aerospace and premium castings (Fig.24) to determine trends in maximum values of elongation versus yield strength for A356/A357 alloys. The maximum elongation (ductility potential) of cast aluminum alloys, eF(max) can be estimated by using equation.3:

Equation.3

-1 Where β0 and β1 are coefficients, the values are 36 and 0.064(MPa ) respectively for

A356/A357 alloys.

These coefficients were determined manually so that the lines could follow the trend of maximum points. This index, QT, is found by:

Equation.4

The concept of QT is presented in Fig.25. This quality index is easy to use and provides a realistic estimate of the current casting quality level.

Fig.24 Yield strength-elongation relationship of A356/A357 alloys [26]

Fig.25 Schematic illustration of the quality index, QT, as the ratio of the experimentally observed elongation as a fraction of the maximum possible elongation QT = eF/eF(max) at the given level of yield strength, σY [26].

Fig.26 shows a nomogram for QT with four distinct regions. It was developed and based on the experience of the authors. When tensile data are in Region 1 (QT = 0 to

0.25), which corresponds to eF = 5% in T6 condition. The authors mention that “Old oxides” from remelts and/or ingot should be first removed. Region 2 represents melts that are free from major old oxides. In this region, quality can be improved by further improvement, such as design of a gating system or chilling of the casting. Improvement efforts in Region 3 involve a perfect filling system (perhaps by computational modeling) and paying attention to all details of melt preparation or even using vacuum casting method. Region 4 is not achievable with current casting technology.

Fig.26 The nomogram with iso- QT lines for A356/A357 alloy castings shows four distinct regions [26].

Chapter 2 - Literature Review on Mechanical Properties of A356 and 319 Alloys

2.1 Effect of Microstructural Modification and Heat Treatment on Mechanical

Properties

S. Shivkumar et al. investigated the heat treatment of A356 alloy [27]. This study looked at how strontium modifies the microstructure and applied the T6 heat treatment to increase the mechanical properties of sand and ASTM B-108 permanent mold catings.

For sand casting, the results showed that 0.02 % strontium could increase the UTS from 23.9 to 24.4ksi (164.6 to 168.1 MPa) and the elongation from 4.25% to 6.12% in the as-cast condition. After 1000oF for 800 min solution treatment and 310F for 4h artificial aging, the UTS for modified and unmodified samples are both 39.3ksi (271 MPa), but the elongation increased from 5.5% to 8.4%. Strontium didn’t show much effect on the UTS but increased the elongation significantly.

On the other hand, for the ASTM B-108 permanent mold test bars, the results showed that 0.02 % strontium could increase the UTS from 28.7 to 29.6ksi (197.7 to 204 MPa) and the elongation from 7.42% to 9.45% in the as-cast condition. After 1000oF x 800 min solution treatment and 310oF for 4h artificial aging, the UTS for modified and unmodified samples were both around 49ksi (337 MPa), and the elongation increased from 8.38% to 8.66%. Therefore, strontium modification didn’t show much effect for permanent mold test bars.

Fig.1 shows the solution time effect on UTS and elongation of sand and ASTM B-108 mold.

Fig.1 UTS and elongation as a function of solution time [27]

J. Y. Hwang et al. investigated the effect of Mg on the structure and properties of type 319 aluminum casting alloys [28]. The test bar mold used in this study was not mentioned. After casting, the samples were solution-treated at 910oF for 8 h. The subsequent T6 aging treatment was carried out at 380oF for 8 h. The results showed that the UTS increased from 21.7 to 29ksi (150 to 200MPa) and the elongation decreased from 0.75% to 0.45% for Mg-free test bars. On the other hand, the UTS increased from

28.8 to 46.4ksi (198 to 320MPa) and the elongation decreased from 0.9% to 0.1% for

Mg-containing (0.45%Mg) test bars.

Fig.2 Ultimate tensile strength and % elongation of Mg-free and Mg-containing samples in the as-cast (blue) and the T6 (red) conditions [28].

2.2 Effect of SDAS (Grain Size) on Mechanical Properties

Guang Ran et al. investigated the effect of hot isostatic pressing (HIPing) on the microstructure and tensile properties of and unmodified A356-T6 cast aluminum alloy

[29]. The A356 alloy was cast into a plate and then machined into test bars according to

ASTM E8M. The solution treatment was 1000oF for 5 h and artificial aging was 320oF for 4 h. In this study, the authors tried to compare different SDAS with UTS and elongation. The results showed that the UTS decreased from 36 to 32.9 ksi (248 to

227MPa) and the elongation decreased from 0.5% to 0.35% as SDAS increased from 82 to 96 micron for non-HIPed samples. On the other hand, the UTS decreased from 35.4 to

34.5 ksi (246 to 238MPa) and the elongation decreased from 0.5% to 0.35% as SDAS increased from 82 to 96 micron for HIPed samples.

Fig.3 The average UTS and elongation of cast A356-T6 aluminum alloy curves [29]

X. Zhu et al. investigated effects of microstructure and temperature on fatigue behavior of E319-T7 cast aluminum alloy in very long life cycles [30]. In this study, they discussed how tensile strength and fatigue property changed with SDAS. The E319 alloy was subjected T7 heat treatment that solution at 925oF for 8 h and aging at 500oF for 4 h.

30μm SDAS (low SDAS) and 70μm SDAS (medium SDAS) samples showed 42ksi and

26ksi UTS respectively. Fig.4 shows S/N curves of 30μm SDAS (low SDAS) and 70μm

SDAS (medium SDAS) fatigue samples. Obviously, low SDAS E319 alloy has much higher fatigue life than medium SDAS E319 alloy.

Fig.4 S/N curves of E319 with low SDAS and 70μm SDAS [30]

2.3 Effect of Porosity on Mechanical Properties

R. K. Owens et al. investigated that effect of nitrogen and vacuum degassing on properties of a cast aluminum-silicon-magnesium alloy [31]. Fig.5 shows the ultimate tensile and yield stress of remelted 356 aluminum alloy decreases as porosity increases.

The ultimate tensile and yield stress are 41.5ksi and 35ksi respectively with no void content. As porosity content increases to 2%, the ultimate tensile and yield stress decrease to 35ksi and 28ksi respectively.

Fig.5 Ultimate tensile and yield strengths versus gas content and per cent voids for remelted A356 alloy [31]

B. Chamberlain et al. also investigated the effect of gas content on the tensile properties on cast aluminum alloys [32]. The aluminum alloy used in this study is CA-

B135 (0.72% Si and 0.35%Mg), similar to 356 type alloys. At same solidification rates

(0.28-0.32), the ultimate tensile, yield stress and elongation are 41.8-44.2ksi, 33.2-

34.85ksi and 6.5-7.5% respectively at 0.15 ml H2/100g gas levels. As the gas level increases to 0.36 ml H2/100g, the ultimate tensile, yield stress and elongation decrease to

38.3-41.1ksi, 30.42-33.86ksi and 3-3.5% respectively.

Table.1: Gas levels, solidification rates and tensile properties of CA-B135-T6 sand cast step castings [32]

2.4 The Effect of Inclusions on Mechanical Properties

L. Liu and F. H. Samuel investigated the effect of inclusions on the tensile properties of A356.2 cast aluminum alloys [33]. Table.3 lists the total inclusions, harmful inclusions and oxide films as measured by the PoDFA technique, and the corresponding tensile properties measured for similar melt conditions. It is evident from this table that the oxide films have a far more deleterious effect on the mechanical properties. In Table.3, Serial no.1 shows the highest mechanical properties since it includes no oxide films. Therefore, oxide films were assessed as the most harmful inclusions in this alloy. Moreover, by comparing Serial no.10 with Serial no.26, Serial no.10 contains 3.78 mm2kg-1 total harmful inclusions and Serial no.26 contains 1.12 mm2kg-1 total harmful inclusions, yet the mechanical properties of Serial no.10 are much higher than Serial no.26. This provides further evidence that oxide films are relatively more harmful to mechanical properties of A356 alloy than particulate inclusions.

Table.2 Inclusion/oxide-mechanical properties relationships for A356.2 alloy [33]

Chapter 3: Experimental Procedure

3.1 Melt Preparation

The A356 metal was melted in a Lindberg electric resistance melting furnace with a

1,000 lb crucible shown in Fig.1. The melt temperature was held at 1350±10˚F.

In this study, virgin and recycled alloy were used. Sometimes the text refers to virgin metal as “clean” and recycled metal as “dirty” metal. Fig.2 shows how we prepared the dirty melt. The recycled parts were from ingates and runner sections of commercially cast products.

Degassing was performed by bubbling argon through a steel lance for 30 minutes

(Fig.3). Flux was used in the melt to remove inclusions (Fig.4). The hydrogen level was controlled to 0.1 ml/100g Al as minimum by an Alspek system (Fig.5). No modification and grain refining was carried out in this part of the study.

Fig.1 Electric resistance furnace 40

Fig.2 Recycled A356 aluminum

Fig.3 Degassing treatment

Fig.4 Fluxing and skimming

Fig.5ALSPEK hydrogen system

3.2 Mold Preparation

Three kinds of molds, a standard Stahl permanent mold (type ASTM B-108), a HS

Stahl mold with an enlarged sprue cross-section, and a modified Stahl mold designated as

“Case mold” (filter mold) were used in this study. A detailed description of the modifications in the Stahl mold design, leading to the present Case mold is provided later in this section. The Stahl mold and Case filter mold were pre-heated to 625˚F. We used

three methods to pre-heat the molds: an electrical resistance heater (Fig.6); an oxy- hydrogen flame torch (Fig.7) and built-in electrical calrods in the SH-Stahl mold.

Fig.6 Electrical heater

Fig.7 Oxy-hydrogen flame

Fig.8 shows the standard Stahl mold. Fig.9 shows the HS Stahl mold with the larger and round cross-section sprue replacing the thin, rectangular cross-section of the standard

Stahl mold,. This larger cross-section increases feeding speed and prevents premature chilling and solidification of the molten metal chill in the gage sections.

Fig.8 Standard Sthal mold

Fig.9 Stahl mold with large round shape sprue (Stahl HS Mold)

Fig.10~13 show the Case mold version 1 to version 4. Fig.10 shows the sub-insert with filter prints for use of filters designated to remove impurities. The sprue of the Case mold version1was deepened to 0.4” from 0.2” to ensure molten metal can pass through the filter. In retrospective, this change was unnecessary as the wider cross-section of the filter print allowed unrestricted flow through the filter. Fig.11 shows the Case mold version 2. The diameter of the entrance to the riser was enlarged from 0.5” to 0.75 to reduce the front velocity to below 20 in/sec. Fig.12 shows the Case mold version 3. The sprue was narrowed back from 0.2”, uniformly deep in each side to an inclined shape

(0.1” at the top and 0.06” at the bottom on each side). This change provides a slower, stable filling pattern, controls the filling velocity and reduces the air aspirated. Fig.13 shows the schematic figure of test bars made with the Case mold version 3 and version 4.

The key difference is in version 4 we introduced a knife ingate between the runner and the test bar. The knife ingate can be removed carefully with a grinder (Fig.14)

Fig.10 Case mold version.1 45

Fig.11 Case mold version.2

Fig.12 Case mold version.3

Fig.13 Schematic of the cast bars from Case mold version.3 and version.4.

Fig.14 Grinding of the knife ingate

3.3 Filter Preparation

In this study, we used two types of filters: in-mold filters in the Case mold and a large, in-crucible filter. Fig.15 shows three sizes of in-mold filters used. From left to right are 30ppi, 20ppi, and 10ppi respectively. Fig.16 shows the in-crucible filter.

Fig.15 In-mold filters

Fig. 16 Top view of the in-crucible filter box (black) immersed in Al

3.4 Reduced Pressure Test

The solubility of gas in molten aluminum decreases with temperature. During solidification of castings, excessive gas is rejected from the melt and forms undesirable porosity. The reduced pressure test (RPT) is a method used widely in the aluminum casting industry to monitor the quantity of gas in molten metal. This project also utilized the RPT to monitor and quantify gas presence on porosity of aluminum in the molten aluminum. Buttons were sampled from the melt and solidified under reduced pressure, one before degassing and the other after degassing in each experimental session. The samples were then sectioned to reveal gas porosity. The density measurement by weighing RPT samples in air and water shown in Fig. 17(b) is discussed in detail in

Chapter 2.

Fig.17 RPT test equipment (a) vacuum pump/chamber (b) balance

3.5 PoDFA

To determine the inclusion content in the melt, two PoDFA samples were also cast in each section of experiment, one before degassing and the other was after degassing.

Fig.18 PoDFA equipment

3.6 Experimental Cooling Rate

A356 and 319 alloys were cast into an instrumented vertical permanent step mold.

The dimensions of the step casting are shown in Fig.19. The mold was coated with Hill

& Griffith ConcoteTM Mag 669, and pre-heated to 400oF using an electrical die heater. In order to measure the local cooling rates, thermocouples were embedded in the center of each step as shown in Fig.19.

Fig.19: (a) Dimensions of the step casting and (b) Illustration of the set-up for cooling rate measurement.

As expected, the highest cooling rate is obtained in the thinnest 0.5” step and gradually decreases toward the heaviest 3” step as depicted in Fig.20. The values of the measured cooling rates are alloy specific, as the thermal properties of the alloy play an important role in the transfer of the heat from the solidifying molten metal to the mold.

However, the cooling rate is affected by many other parameters. The size and configuration of the casting and the location within the casting, can make a significant difference. The metal at the center of a casting will generally solidify at a slower rate that the center. The flow of the incoming metal into the mold can also make a difference, as it will heat the mold more in some areas than in others.

Fig.20: Respective cooling rates.

The initial temperature of the mold is an important factor. The higher the initial temperature, the slower the cooling rate of the casting, as the heat transfer is a function of the temperature difference between the casting and the mold. Preheating of molds is practiced to ensure good filling of the thin sections. However, excessive mold temperatures lower the mechanical properties of the casting and can slow down cycle time.

Another factor that determines cooling rate is the heat transfer coefficient at the interface between the mold and the casting. This coefficient can be controlled to some extent by application of coatings on the surface of the mold. Insulating and/or conductive mold coatings (washes) are applied on specific parts of the mold to generate directional solidification patterns that promote a defect free casting. In general, different casting processes are associated with certain ranges of cooling rates. Sand and investment casting utilize insulating media, and are at the low end of the solidification rate range.

Die casting employs metal molds and is at the high end of the solidification range. Most other processes fall between these. Fig.21 is a qualitative map of cooling rates across metal casting processes. As described earlier, many additional factors play a role in determining the local cooling rate during solidification.

Fig.21: Typical cooling rates for various casting processes

The cooling rate determines the microstructure which in turn drives the mechanical properties. Experimental measurement of the cooling rate under production circumstances is often times difficult. A common practice is to use instead the inter- dendritic arm spacing that is inversely proportional to the cooling rate. An advantage of this method is the ability to examine the local inter-dendritic arm spacing in the area of interest. In the present study, the inter-dendritic arm spacing has been used as an indicator of cooling rate. Computer simulation predictions of inter-dendritic arm spacing are reported and compared to the experimental results.

3.7 Tensile Test

The diameter of the gauge section in the Stahl, Case and step mold test bars is 0.5”.

Fig.22 shows test bars cast with the standard Stahl mold, the Case mold and the step mold.

The test bars were tested primarily in the as-cast condition at room temperature at a strain rate of 10-3s-1.

Fig.22(a) Stahl mold test bars Fig.22(b) Case mold version.1 test bars

Fig.22(c) Step mold sample Fig.22(d) Step mold 2” machined test bars

3.8 Fatigue Test

The fatigue test unit is an MTS 810 and ran at 125 MPa fully reversed sinusoidal loading for testing 0.25” material. An alignment fixture is used to obtain less than 15 micro strain prior to testing. The software is MTS Multipurpose elite. Fig.23 shows the image of MTS 810.

Fig.23(a) Fatigue testing machine

Fig.23(b) Fatigue testing machine grip 56

3.9 Magma Simulation

Computer aided simulation analysis was used to model the process-microstructure relationship for both Stahl Mold and the modified design. The Finite Volume Method

(FVM) based commercial software MagmaSoft [35] was used and a 3D simulation model was created. The model was meshed with about 1,500,000 non-body-fit structured hexagon elements. The edge length of the smallest element was chosen to be 0.3mm so as to ensure at least five layers of elements in thin wall sections (such as the knife ingate).

The mold material was H13 tooling steel, with the thermal conductivity of 17Wm

-1K-1 (10 Btu/ft×hr×℉) at room temperature and 25Wm-1K-1(14 Btu/ft×hr×℉) at elevated temperature (>204℃, 400℉). The cast alloy was aluminum alloy A356, with solidus temperature of 550℃ (1022℉) and liquidus temperature 616℃ (1141℉) [36].

Two kinds of coatings, Graphite coating and Dycote 8 coating, thermally conductive and insulating, respectively, were used in the study. According to the previous experimental observations [37], the following assumptions were made for the interface heat transfer coefficient (HTC). The HTC was defined to be casting surface temperature dependent. The HTC of graphite coated interface was 1500Wm-2K-1(264

Btu/ft2×hr×℉) for temperatures below 550℃ (1022℉), and 3000Wm-2K-1(528 Btu/ft2

×hr×℉) for temperatures above 616℃ (1141℉). The HTC of Dycote 8 coated interface was 750Wm-2K-1(132 Btu/ft2×hr×℉) for temperatures below 550℃ (1022℉), and

1521Wm-2K-1(268 Btu/ft2×hr×℉) for temperatures above 616℃ (1141℉). The HTC 57

was assumed to be linearly dependent on temperatures between 550℃ (1022℉) and 616

℃ (1141℉).

The casting processes, comprising filling and solidification, were simulated. The filling process was simulated by filling pressure control, whereby the molten metal was assumed to fall from 50.8mm (2 inch) above the mold and generate 1.2kPa fill-in pressure. Initial mold temperature was set to be uniform at T0 (to be specified later).

The pouring temperature of molten A356 was chosen at 705℃ (1300℉), which was within the normal pouring temperature range of 680-705 ℃ (1250-1300℉ [38]. The mold was cooled in air with ambient temperature of 20℃ (68℉). The cast part was solidified in the mold for one minute, and then cooled in ambient air for 5 minutes.

The microstructural features of cast test bars, such as the secondary dendrite arm space (SDAS) and shrinkage induced micro-porosity were predicted based on the heat transfer simulation. Equations developed by Carlson and Beckermann [36] from Niyama criterion [39] were adopted to compute the micro-porosity and SDAS as a function of cooling rate and thermal gradient. The accuracy of the simulation was examined by comparing it to the experimental results.

Chapter 4 - Results and Discussion

4.0 Hypotheses and Objectives

The hypothesis of this study is systematic application of molten metal treatment, optimized strengthening mechanisms and post-casting treatments should produce an improvement in mechanical properties of A356 and 319. These are the most common among cast aluminum alloys in automotive and industrial applications. A reliable and reproducible method for measuring mechanical properties is essential in order to determine the effect of the various processing parameters. The Stahl permanent mold was selected, as it is commonly used to test melt quality in commercial casting operations.

However, shrinkage porosity was frequently encountered in the fracture surfaces of the test bars. This porosity has an overriding impact on mechanical properties and had to be eliminated. An effort to redesign of Stahl mold by a combination of computer simulation and experimental verification was undertaken. The subsequent objective of the study was to determine the effect of processing variables on the mechanical properties of two mainstream alloys, 356 and 319, employing the redesigned Stahl mold.

4.1 Characterization of Defects

4.1.1 Reduced Pressure Test

Table.1 shows the relationship between the results of the reduced pressure and the

Alspek hydrogen concentration readings of A356 alloy. The pressure test results were used to determine the specific gravity of each sample and then calculate the percentage of porosity. The results did not so much depend on the condition of the melt (clean or dirty) but rather on temperature and relative humidity. The percent porosity determined with the

reduced pressure test seems to be proportional to the Alspek readings. Fig.1 shows the

PoDFA result related to table.1 as a bar graph. The total inclusion content of clean and dirty metal before degassing and fluxing were 0.6 and 2 mm2/Kg. After degassing and fluxing, the spinel and Mg-oxide were mostly removed, but somehow induced Al-carbide.

A possible source is the degassing system that introduces argon gas bubbles with a spinning graphite shaft and impeller. Finally, we used an in-situ filter in the melt that reduced the total inclusion content to only 0.002mm2/Kg Mg-oxide. The efficient reduction in inclusions with the in-melt filter is another reason why we decided to introduce filters in the test bar mold.

Table.1 Reduced pressure test relationship to Alspek reading of A356 alloy

Porosity in RPT Specific Gravity H2 Content Sample Degas Sample in RPT Sample (ml/100g) (%) (g/cm3)

Clean No 8.05 2.47 0.31

Clean Yes 2.11 2.63 0.09

Dirty No 2.89 2.61 0.16

Dirty Yes 1.87 2.64 0.09

Dirty + In-melt Yes 4.34 2.56 0.19 filter

Fig.1: PoDFA results for five different conditions

4.2 Methods for Improvement in Mechanical Properties of A356 alloy

The composition of the 356 aluminum alloy is shown in table 2. One objective of this study is to improve the melt quality of this alloy so as to obtain maximum mechanical properties in the casting. However, if shrinkage porosity is present, it may reduce the mechanical properties, even in the absence of inclusions. Therefore, oxides, shrinkage porosity and inclusions have to all be eliminated or reduced to very low levels for good mechanical properties to be obtained. Shrinkage porosity is determined primarily by the solidification pattern of the casting. This in turn is determined by the gating design and feeding of the casting. These are the main topics discussed in this chapter.

Table 2: Composition of A356 alloy in this study

Alloy/Element Si% Mg% Fe% Sr% Ti%

A356 6.975 0.357 0.090 0.09 0.088

4.2.1 Effect of the Gating System on Mechanical Properties

4.2.1.1 Effect of Sprue Size

An early concern in implementing filters in the test bar permanent mold, was the filter would slow down the flow of the molten metal, causing premature solidification, before the cavity is filled. Therefore, the thickness of the sprue in the Case mold (Case mold v1 and v2) was increased to 0.4”. However, after increasing the dimension of the sprue, an undesirable increase in oxides was noticed, as illustrated in Fig.2. The fracture of test bars cast without degassing and without a filter had 10%~20% of the surface covered by a black film; in contrast, another kind of film (white) covers on the fracture surface of the Stahl mold test bars (Fig.3).

EDX results shown in Fig.2 indicate the black film has much higher oxygen content then the oxide-free fracture surface. Fig.2 also includes an SEM micrograph of the oxide film. It clearly shows the dendritic structure typical around folded oxide films. The white film in the SEM micrograph (Fig.3) shows a clear dendritic structure yet didn’t contain higher oxygen compared to the dimpled region. We therefore concluded the white region is not an oxide film but rather a shrinkage pore.

L. Liu and F. H. Samuel investigated the effect of inclusions on the tensile properties of A356.2 aluminum cast alloy. Their results show that oxide films are extremely deleterious to the mechanical properties [33]. R. Fuoco and E. R. Correa investigated the effect of gating design on the quality of aluminum gravity castings. Their results show good gating systems can prevent formation of oxide films [34]. Too large cross-section sprues can cause turbulence, and produce oxide films.. 62

Fig.2: The fracture surface of Case mold v1 test bar

Fig.3: The fracture surface of a Stahl mold test bar

In order to understand how these two kind of porosity are produced, we examined the influence of the diameter between the filter and runner of the Case mold v1. The

Magma [35-39] flow simulation in Fig.4, shows the melt front velocity between the filter and runner was larger than 30 in/sec (Fig.4 (a)). This velocity is considered too fast and could produce turbulence thus create oxide in the metal. Fig.4 (b) shows the Case mold v2, where the thickness between the filter and runner was increased from 0.5 inch to 0.75

inch. The simulation now shows the melt front velocity was reduced to about 17 in/sec.

However, this change still didn’t eliminate the oxide film in the test bars. Therefore, the cause of the oxide film present in the test bars must be the size of the sprue cross-section.

Fig.4: Reduction in melt front velocity of Case Mold v.2

Another experiment was conducted with two Stahl molds with different sprue sizes, to better understand how the gating system influences the formation of the oxide films and resulting mechanical properties. This experiment employed the standard Stahl mold Fig.5

(a) with a thin rectangular sprue,the modified Stahl mold Fig.5 (b) with a larger cross- section, and the Stahl HS mold with a round sprue. By increasing the size of the sprue, the fillng time is reduced from 10 to 5 seconds. However, test bars cast with the Stahl mold with a larger cross-section round sprue still exhibited oxide films on the fracture surface just like Case mold version 1 and 2. 65

Fig.5 (a) Standard Stahl mold Fig.5 (b) Stahl HS mold

To understand the effect of different sprue sizes on the microstructure and mechanical properties additional computer simulations were run. These simulations show the filling time obtained with the standard Stahl mold (Fig.6 (a)) is below one second; the filling time in the modified Stahl HS mold (Fig.6 (b)) is longer. The molten metal flows continuously in the Stahl mold, but is “falling” from the top to the bottom in the Stahl HS mold. As a result, more air is trapped in the runner and the test bars during pouring in the

Stahl HS mold.

Fig.6 (a) Filling time in Stahl mold Fig.6 (b) Filling time in Stahl HS mold

Fig.7 shows the molten metal front velocity as it flows into the runner. Magma simulation predicts the melt front in that part of the Stahl mold to be about 20 in/sec

(Fig.8 (a). For the Stahl HS mold it is about 30 in/sec Fig.8 (b). Therefore, the larger sprue size will create a faster melt front velocity. As discussed before, more air is aspirated by the larger sprue, producing more oxide films in the test bars.

Fig.7: Melt front velocity of Stahl mold and Stahl HS mold

Due to the increased oxides content, the mechanical properties of the Stahl HS mold test bars in as-cast and T6 condition were lower than the Stahl mold test bars. Fig.8 shows that the UTS and elongation of the Stahl and Stahl HS mold test bars in the as-cast condition are 26.55ksi, 5.23% and 23.65ksi, 4.67% respectively. In the T6 condition the

UTS and elongation are 44.36ksi, 7.1% and 41.74ksi, 5.6% respectively.

Fig.8: The effect of sprue size on mechanical properties of A356 alloy on Stahl mold

To further investigate the sprue size effect on filling time, we used Magma to simulate filling time for the 0.4” and 0.1” sprue sizes in the Case mold. In Fig.9 (a), the sprue size is 0.4” and filling time of sprue, runners and test bars were similar at the same height. However, when the sprue size is 0.1” as shown in Fig.9 (b), the filling time in the sprue was much faster than other locations. As before, when the Stahl and Stahl HS molds were compared, the larger sprue size will increase air entrapment in the test bars and reduce mechanical properties.

Fig.9: (a) 0.4” Sprue size (Case mold V2) Fig.9 (b) 0.1” Sprue size (Case mold V3)

Fig. 10 shows a similar result when comparing the Stahl HS mold to the standard

Stahl mold. The test bars cast with the Case Mold v.2 contain lots of oxides. The mechanical properties these test bars were lower than bars cast with the Case Mold v3 test bars. The UTS and elongation of Case Mold v2 and v3 test bars in the as-cast condition were 22.81ksi, 4.38% and 24.32ksi, 5.08% respectively.

Fig.10: Effect of sprue size on mechanical properties of A356 alloy on Case mold

Based on the reported results, we can conclude the Stahl mold and Case Mold v3 with the 0.1 inch sprue thickness eliminate air aspiration effectively. The fracture surface of bars cast with these molds only contain a small shrinkage region at the center (Fig.11 (a)); however, Stahl HS mold and Case Mold v1 and v2 with the larger sprue size will produce undesirable black oxide film at the fracture surface (Fig.11 (b)).

Fig.11: (a) Macrograph of Stahl mold and Fig.11 (b) Macrograph of Stahl HS mold Case mold version 3 and Case mold version 1 and 2

In summary, the small 0.1” thickness of the sprue in the Stahl mold is facilitating longer fill times, with no turbulence and/or air entrapment. Conversely, the oxide film identified on the fracture surface, explain the lower tensile properties measured for test bars with a thick sprue. To correct this problem, the sprue of the Case mold was welded and machined back to the original thickness of 0.1”.

4.2.1.2 Effect of the Knife Ingate

Even with the thinner sprue size, that should eliminate air aspiration, the fracture surface of Stahl mold and Case mold v 3 still exhibits shrinkage defects.

Fig.12 (a) and Fig.12 (b) show the Case mold with and without a 0.07” wide “knife ingate” respectively; Fig.13 (a) and Fig.13 (b) shows the simulation of solidification in the Case mold with and without knife ingate respectively. In Fig.13 (a), the solidification

rate in the gage section is very fast (around 35 sec) and the melt solidifies from the gage section towards the grip sections of the test bars. This solidification pattern is detrimental to mechanical properties because it prevents feeding. Since no more melt can feed into the gage section shrinkage porosity would be produced. Fig.14 (a) shows the prediction of shrinkage porosity in the Case mold without a knife ingate. The shrinkage porosity in the center of the gage section is about 2.8%. Fig.13 (b) shows the melt solidifying from the outer surface to the inner section. This solidification pattern allows molten melt to continue filling the gage section of test bars. Fig.14 (b) predicts the shrinkage porosity of the Case Mold with a knife ingate (v4). The knife ingate is very effective in reducing shrinkage porosity in the gage section of the test bars. The shrinkage porosity predicted by the simulation in a Case Mold with a knife ingate was only 1%.

Fig.12 (a) Case mold without Fig.12 (b) Case mold with knife ingate (v3) knife ingate (v4)

Fig.13 (a) Solidification time of Fig.13 (b) Solidification time of Case mold without knife ingate (v3) Case mold with knife ingate (v4)

Fig.14 (a) micro porosity Fig.14 (b) micro porosity of Case mold without knife of Case mold with knife ingate (v3) ingate (v4)

Fig.15 (a) and (b) show the prediction of SDAS in the Case Mold with and without a knife ingate respectively. The Case Mold with a knife ingate had a slightly larger SDAS than without a knife ingate. The smaller SDAS can increase mechanical properties of both A356 and 319 alloys, but if the porosity is large enough, it will overwhelm other effects even if we only had 3µm SDAS difference.

Fig.15 (a) SDAS of Case Fig.15 (b) SDAS of Case mold without knife ingate (v3) mold with knife ingate (v4)

Despite the improvements obtained with Case Mold v3, the porosity present in the gage section still dominates the mechanical properties. Therefore, based on the simulation results, we replaced the bottom gating with a 0.07” wide knife ingate between the runner and the test bar. This version was designated as Case Mold v4. Fig.16 shows

the mechanical properties of test bars cast with Stahl mold and Case-H mold in the as- cast condition. The UTS of Case Mold v4 is just a little highe, but the elongation is much higher than the Stahl mold in the as-cast and T6 condition (the elongation is increased from around 5.23% to 8.3% at as-cast condition and 7.1% to 12.98% in theT6 condition).

Fig.16 The effect of the knife gate on mechanical properties of A356 alloy

Fig.17 (a) and Fig.17 (b) show the microstructure of test bars cast with the Stahl mold and Case mold v4 in the middle of the gage section. This microstructure provides clear evidence on the ability of the knife ingate to improve elongation dramatically. It is because the knife ingate eliminates most porosity in the gate section of the test bar. There are many shrinkage pores on the fracture surface of a Stahl Mold test bar shown in Fig.17

(a); however, the fracture surface of bars cast with the Case Mold v4 is nearly perfect

(45∘) as shown in Fig.17 (b) (c).

Fig.17 (a) Fracture surface of Stahl Mold Fig.17 (b) Fracture surface of Case mold Test bar version 4 test bar

Fig.17 (c) Stahl mold test bar broke in 90 degree (left) and Case mold test bar broke in 45 degree

Fig. 18 is a SEM image of the fracture surface from a Stahl Mold test bar. Shrinkage is pervasive on the surface shown in Fig.18 (a). Fig.18 (b) shows at higher magnification a shrinkage pore with dendritic structure; the fractography of bars cast with the Case

Mold v4 shows no porosity. Fig.19 (a) and Fig.19 (b) show at different magnification the fracture surface of bars cast with the Case mold v4. All the fractures are porosity free and present a dimpled, ductile surface.

Fig.18: Shrinkage porosity in Stahl mold test bars

Fig.19: Perfectly broken of Case mold V4 test bars 78

Although the knife ingate improves the mechanical properties and microstructure it poses a challenge when trimming the test bars. The knife ingate is at gage section and has to be removed carefully to avoid introducing any defects in the gage section. Fig.20 shows a test bar after removing the knife ingate.

Fig.20: Test bar after removal of the knife ingate

4.2.2 The Effect of Coating the Gage Section of the Mold

When casting test bars, controlling the cooling rate of the gage section during solidification is very important. Coating the gage is therefore widely practiced.

Insulating coatings like Dycote34 are generally used on the runner and the riser. We used this coating in the gage section of Case Mold v4, otherwise the molten metal will not easily flow through this thin section without solidifying prematurely. Coating is a very tedious process, so at the beginning we just applied Dycote34 on the entire parting line including the gage section. Applying Dycote34 on the gage section can however reduce the mechanical properties dramatically. Shrinkage on the fracture surface could be observed as illustrated in Fig.23.

Fig.21 and Fig.22 illustrate the cooling rate and micro-porosity for two coatings applied on Case Mold v4. Faster cooling rate and less micro-porosity were obtained when a graphite coating was applied on the gage section. A possible reason for the effectiveness of the graphite coating in preventing shrinkage is the faster cooling that prevents a hot spot from forming in the gage section

Fig.21: The cooling rate at gage section of Case mold V4 with two kinds of coating

Fig.22: The micro-porosity at gage section of Case mold V4 with two kinds of coating

Fig.23 shows the fracture surface of test bars with graphite and Dycote34 coating on the gage sections in three molds. Fig.23 (a) indicates that using graphite coating on Stahl mold and Case Mold v3 produced a small shrinkage porosity region at the center on the test bar; Fig.23 (b) indicates Stahl mold and Case Mold v3 coated with Dycote34 produced a large shrinkage porosity region in the test bar. It confirms the prediction of the simulation that the hot spot in the gage section will results in shrinkage porosity.

A similar result is shown in Fig.23 (c) and Fig.23 (d). The knife ingate by itself can remove most shrinkage porosity in the gage section of test bars; if we apply Dycote34 in the gage section, a small shrinkage porosity region will be produced. Fig.23. (d) confirms the simulation prediction with regard to the shrinkage porosity shifting off the center, towards the knife ingate direction.

Fig.23: Applying a coating on the gage section affects on the fracture surfaces of tensile test bars of A356 alloy in different molds

4.2.3 Effect of Filtration

PoDFA results indicate in-crucible filtration can remove most inclusions from the melt.

The in-mold filter can improve the melt cleanliness during the treatment process. The

45ppi in-mold filter contains 75% porosity, while the in-crucible filter contains only 40% porosity. Fig.24 shows the mechanical properties of the A356 alloy with and without the

45ppi filter in the Case Mold v4 in as-cast and T6 conditions. The properties of test bars are very similar with or without the filter.

Fig.24: In-mold filter effect on mechanical properties of A356 alloy

4.2.4 Best Mechanical Properties of A356 alloy Test Bars Cast with Each Mold

Fig.25 illustrates a Step Mold used to cast Step Castings. The 2” second step was used to machine test bars. The step mold was used to produce almost porosity free casting because it solidifies in a directional manner (from thin part to thick part). The thin part is in contact with the mold wall, therefore solidifies first. In this study, we use the 2” part to make test bars.

Fig.25: Schematic of the step mold and step casting

Fig.26 shows the comparison of micro-porosity among the Stahl mold, Case Mold v4 and Step Mold. The Stahl mold test bar has about 3% micro-porosity at the center of the gage section. The Case Mold v4 test bar and Step Mold 2” section have less than 1% porosity. Fig.27 illustrates the SDAS in the Stahl mold, Case Mold v4 and Step Mold.

The SDAS of test bars cast in the Stahl mold and Case Mold v4 at the center of the gage section is about 24~27 µm; in the step mold 2” section the SDAS is about 35 µm. The prediction of the simulation for the 2” section of the step mold is very close to the cast samples; however, the SDAS at the center of the test bar is only about 20 µm for Stahl mold, 18 µm for Case Mold v4.

Fig.26: Micro-porosity prediction in three different molds

Fig.27: SDAS prediction for three different molds

Hot isostatic pressing (HIP) should be able to remove most of the porosity in the 2” section of the Step Mold. Fig.28 shows the tensile testing results for an A356 alloy with and without HIP in the as-cast and T6 conditions. The HIP improves both UTS and elongation;, the tensile properties of bars machined from the Step Mold are lower than the Stahl and Case-H Mold samples. This may be because the SDAS of the test bars taken from the Step Mold samples is much larger than in the Stahl and Case-H Mold samples.

Fig.28: Mechanical properties of samples from the 2” Step Casting with and w/o HIP

Fig.29 shows the best mechanical properties of A356 alloy cast in different molds.

The squeeze casting mold, Case Mold v4 and Step Casting can be considered porosity free, but the Stahl mold is not. The mechanical properties of squeeze cast and Case Mold v4 bars are very similar. If the SDAS is lower than 20 µm in a porosity-free test bar, we obtain the best mechanical properties of A356 alloy. Because the 2” section of the step casting has a larger, 35 µm SDAS it yields lower mechanical properties than squeeze casting and/or bars cast with the Case Mold v4.

Fig.29: Best mechanical properties for test bars cast in different molds (T6 condition)

4.2.5 Quality Index

It is convenient to use the quality to describe overall casting quality and to compare different sets of data on castings. The quality index accounts for differences in strength levels, which may be caused by differences in Mg content or aging time. The quality index is defined by the formula as mentioned in the literature review:

Q = UTS + 150 log E where Q and UTS are given in MPa and the elongation to fracture, E, is given in percent.

The best mechanical properties obtained so far in our study are superimposed on a plot from a recent in a paper by G.K. Sigworth and T.A. Kuhn [40]. Our result show as a star in Fig.30. The aging time we used in this study was 6 hours (320℉) and the Case-H mold.

43.84Ksi UTS and 13.4% elongation (QI=471). This is a very good result according to this plot because our A356 contains 0.09% Fe. We can say the Case mold v4 is better than standard Stahl mold, because the knife ingate reduces shrinkage porosity in the gage section of the test bars.

Fig.30: Mechanical properties of 356-T6 alloy as a function of iron content and aging time. Castings were aged 2, 6 and 18 hours at 310oF (155oC). (Ken Whaler 10) [40]

To compare with another quality index QT also mentioned in the literature review, three average data points were used from Fig.29 and shown in Fig.31 [26]. The green star represents the Stahl mold test bars (σY. = 215 and QT = 0.33), the blue star stands for the step mold with HIP test bars (σY. = 243 and QT = 0.49) and the red star represents the

Case-H mold test bars (σY. = 245 and QT = 0.66). The data from the Case-H mold test bars could only reach the top of Region 2. The reasons of why literature data could reach

Region 3 are:

1. The authors didn’t show all the dimensions of the test bars. Elongation depends on

the shape an thickness of gauge section. (The authors only mentioned the kind of

mold used such as sand mold, permanent mold or premium casting) 89

2. The data in Region 3 probably used vacuum casting which could produce porosity

free test bars or employed fast cooling rate to obtain extremely fine grain test bars.

3. The authors should have used a single mold to generate the plot, because the three

stars from this study have the same melt preparation and the test bars have the same

size of the gauge section. The only difference is the mold. This plot could be used as

evidence on how the Case-H mold improves the quality of A356 test bars.

Fig.31: Data from this study in QT lines for A356 alloy castings [26]

4.3 Improving Mechanical Properties of Alloy 319

In this study, we used four different 319 alloys: low Mg - low Mn, low Mg – high Mn, high Mg – low Mn, and high Mg – high Mn alloys. The compositions are shown in

Table.3.

Table.3: Composition of Alloy 319 used in the study

Alloy/Element Si% Cu% Mg% Mn% Fe% Sr% Ti% Mn/Fe

319 5.808 3.468 0.072 0.158 0.690 0.18 0.124 0.23

319 6.014 3.281 0.063 0.328 0.637 0.16 0.140 0.51

319 7.081 3.358 0.321 0.089 0.537 0.17 0.135 0.16

319 7.290 2.990 0.357 0.357 0.520 0.17 0.146 0.69

4.3.1 Effect of Mg and Mn Additions to 319 Alloy

Adding Mg and Mn can increase mechanical properties in alloy 319. Fig. 32 shows the Mn modified microstructure of 319. Fig.32 (a) shows many needle-shape secondary phases (β phases - Fe (Al5FeSi)) in low Mn 319 alloy; and Fig.32 (b) shows modified phases (α phases - Al15(Fe,Mn)3Si2) in high Mn 319 alloy.

Fig.32 (a) Low Mn 319 alloy Fig.32 (b) High Mn 319 alloy

4.3.2 Effect of the Knife Ingate

A similar trend for the A356 alloy is shown in Fig.33. The fracture surface shows a shrinkage pore at the center of the Stahl mold test bars; Test bards made with the Case mold v4 do not show defects on the fracture surface.

Fig.33 (a) Fracture surface of Stahl Mold Fig.33 (b) Fracture surface of Case mold Test bar v4 test bar

Fig.34, Fig.35, and Fig.36 show how test bars cast with the Stahl mold compare with those cast with the Case Mold v4 for four kinds of 319 alloys in the as-cast, T6 and T7 condition respectively. Fig.34 indicates Mg increases the UTS; Mn increases both UTS and elongation of 319 alloy. The knife ingate didn’t change the mechanical properties of low Mg – low Mn and high Mg – low Mn 319 alloys, but increased by a small degree the properties of high Mg – high Mn alloy. It is possible that the low Mn 319 alloy is too brittle to benefit from the absence of the shrinkage pores in the test bars. Fig.35 shows a similar results relative to the as-cast condition. Fig.36 demonstrates the knife ingate could increase mechanical properties in the T7 condition especially the elongation.

Fig.34: The effect of Knife ingate on alloy 319 in the as-cast condition

Fig.35: Effect of knife ingate on properties of 319 alloy in T6 condition

Fig.36: Effect of knife ingate on alloy 319 in T7 condition 94

4.3.3 Effect of Filtration

Fig.37 and Fig.38 show the effect of in-mold filtration in Case mold v4 test bars for three kinds of 319 alloys in as-cast, T6 and T7. Fig.37 shows a low Mg – high Mn 319 alloy in the as-cast condition with and without filter. The filter only improves a little on both UTS and elongation. Fig.38 shows the filter doesn’t improve the UTS and reduces by a little elongation in high Mg – high Mn 319 alloy in both T6 and T7 condition.

Fig.37: Effect of in-mold filter on alloy 319 in the as-cast condition

Fig.38: Effect of in-mold filtration on alloy 319 in T6 and T7 conditions

Fig.39 shows the results of A319 alloy with and without HIP at as-cast and in T7 condition. The HIP improves elongation but reduces yield stress and only improves a little the UTS. Like A356 alloy, the tensile properties of the step mold samples are much lower than the Stahl and Case-H mold samples. This may because the SDAS of the step mold samples is much larger than Stahl and Case-H mold samples.

Fig.39: Mechanical properties of Step Mold 2” tensile samples with and without HIP

4.4 Improving Fatigue Properties with HIP

Fig.40 shows the fatigue properties of samples excised from the Step Mold, 1” step sample with and without HIP that corresponds to the reference data with 125 MPa fully reversed sinusoidal loading. The reference curve of the A356 alloy is from the Casting

Technologies Company (CTC). The blue triangles represent A356 alloy without HIP in

T6 condition and the red circles represent A356 with HIP. The no-HIP samples have lower results than the reference curve while the HIP samples show better results than

reference curve.

HIP also improves the fatigue properties of the 319 alloy. The blue stars represent

319 alloy without HIP in T7 condition; the red circles represent it with HIP. The 319 reference curve is from X. Zhu’s study [30]. Evidently, reducing microporosity by HIP improves the fatigue properties dramatically. Fatigue test bars were cut and machined from Step Mold 1” section.

Pores are known to accelerate micro-crack initiation as a result of pore-induced plastic zones which create locally high regions of plastic strain at or very near the specimen surface, adjacent to the pore. The main mechanism by which HIPing improves fatigue life is the reduction in crack initiation due to the lower level of porosity facilitated by hot isostatic pressing.

Fig.40 Fatigue properties of Step Mold 1” sample with and without HIP [30] 98

0.250+/-0.001 1.050 (Ref.) 1.050 (Ref.)

0.500+/-0.001

R 2.000 3.50

Fig.41 Dimensions of fatigue test bar

5. Conclusions

A number of key conclusion were derived from this study and are listed below.

These conclusions have both fundamental and practical implications for the metal casting industry:

1. Magma computer simulation of the molten metal flow demonstrated reducing sprue size of the Case mold from 0.4” to 0.1” prevents air entrapment in the test bars. The computer simulations also predict significantly reduced micro-porosity in the modified

Case Mold, largely due to the incorporation of the knife-ingate design. This design promotes a laminar, turbulence-free flow of the metal into the mold cavity while providing a more directional heat flow pattern in the gage section of the test specimen.

2. Employment of an in-furnace filter had a very positive effect on reducing inclusions and therefore improving tensile properties in the test-bar. The total inclusion content of clean and dirty metal before degassing and fluxing were 0.6 and 2 mm2/Kg. After degassing and fluxing, the spinel and Mg-oxides were by –and-large removed; using an in-furnace filter reduced the total inclusion content to only 0.002mm2/Kg Mg-oxide.

3. Experiments with 30 ppi in-mold filters had little or no effect on mechanical properties of the test-bars. Applying filtration and degassing best practices enabled the recycled metal to reach nearly the same properties as those obtained from a virgin ingot alloy. In other words, if the molten metal is very clean, additional filtration makes no difference.

4. The knife ingate design improves the UTS, elongation and Q/QT of permanent mold test bars. The Case-H Mold that includes an in-mold heater, could keep the mold temperature at 625℉ and allow the knife ingate to stay open and fill the test bars well. 100

For the traditional Stahl mold design, the average UTS, elongation and quality indexes Q and QT are 41.13ksi, 6.68%, 407Mpa and 0.33 respectively; for the Case mold test bars

(after improving the gating system), the average UTS, elongation, Q and QT are 43.84ksi,

13.4%, 471Mpa and 0.66 respectively.

5. Tensile properties were improved by Hot Isostatic Pressing, especially the elongation which is more sensitive to the porosity than Ultimate Tensile Strength. For the A356 alloy, the elongation was improved from an average of 7.17 to 10.83 in the T6 condition; for the 319 alloy, the elongation was improved from an average of 3.17 to 3.5 in the T7 condition.

6. The fatigue properties of aluminum 356 and 319 are very sensitive to the presence of porosity. The fatigue life could be improved significantly by HIPing of the Step Mold samples. For the A356 alloy at 125 MPa fully reversed sinusoidal loading at 60 Hz, the fatigue life was improved from an average of 4.5 x 105 to 3.5 x 106. For the 319 alloy the fatigue life was improve from an average of 1.2 x 104 to 3.5 x 106.

Because of these improvements, the study has attracted strong interest in the aluminum casting community The Case mold design is currently considered by ASTM as a possible replacement for the traditional ASTM B108 test bar mold.

Future Work

Improved mechanical properties in 0.5” diameter test bars made feasible with the knife ingate design were demonstrated in this study. Similar improvements were obtained by Hot Isostatic Pressing. Both are attributed to a reduction in the porosity level.

101

Quantifying the porosity level in the test bars and step castings, and correlating it to the mechanical properties is recommended. Recent developments in high resolution computer tomography (CT) should enable characterization of micro porosisity.

In this study, we used test bars machined from the step casting for hot isostatic pressing studies. Hipping test bars cast with the Stahl and Case mold to determine the effect of HIPing on the mechanical properties is suggested.

Since fatigue is very sensitive to porosity, a broader study on the effect of HIPing on cast aluminum A356 and 319 alloys should be undertaken.

102

6. References

1. Shahrooz Nafisi and Reza Ghomashchi, “Boron-based refiners: Implications in conventional casting of Al-Si alloys”: Materials Science and Engineering A 452-453 (2007) 445-453

2. X. Jian, T.T. Meek, and Q. Han, “Refinement of eutectic silicon phase of aluminum A356 alloy using high-intensity ultrasonic vibration”: Scripta Materialia 54 (2006) 893-896

3. S. El-Hadad, A.M. Samuel, F.H. Samuel, H.W. Doty, and S. Valtierra, “Effect of Bi and Ca additions on the microstructure of Sr-modified 319 type alloys under variable cooling conditions”: AFS Transactions 2004 Paper 04-008(02).pdf, PAGE 1 of 13

4. J.Y. Hwang, H.W. Doty, and M.J. Kaufman, “The effects of Mn additions on the microstructure and mechanical properties of Al-Si-Cu casting alloys”: Materials Science and Engineering A 488 (2008) 496-504

5. C. Villeneuve, A.M. Samuel, F.H. Samuel, H.W. Doty, and S. Valtierra, “Role of trace elements in enhancing the performance of 319 aluminum foundry alloys”: AFS Transactions 01-023 (1 of 14)

6. L. Lu and A.K. Dahle,: Metallurgical and Materials Transactions A, vol. 36A, (2005) pp. 819-835

7. Robert W. Cahn and Peter Haasen, “Physical metallurgy”: 1996 Elsevier Ltd

8. S.K. Chaudhury, L. Wang, and D. Apelian, “Fluidized bed reactor heat treatment of A356 alloy: microstructure analysis and mechanical properties”: AFS Transactions 2004 paper 04-055(02).pdf, Page 1 of 16

9. E. Rincon, H.F. Lopez, M.M. Cisneros, and H. Mancha, “Temperature effects on the tensile properties of cast and heat treated aluminum alloy A319”: Materials Science and Engineering A 519 (2009) 128-140

10. C. Ravi and C. Wolverton, “First-principles study of crystal structure and stability of Al-Mg-Si-(Cu) precipitates”: Acta Materialia 52 (2004) 4213-4227

11. F.J. Tavitas-Medrano, S. Valtierra, J.E. Gruzleski, F.H. Samuel, and H.W. Doty, ”A TEM study of the aging behavior of 319 type alloys”: AFS Transactions 2008, Paper 08-19(02).pdf Page 1 of 16

12. G. Boudreault, A.M. Samuel, F.H. Samuel, and H.W. Doty, “Microsturctural observations of porosity in A319.2 alloy: effect of mold type/cooling rate”: AFS Transactions 19990446A.pdf, Page 1 of 10 103

13. Diran Apelian, “How clean is the metal you cast? The issue of assessment: a status report

14. L. Parmenter, D. Apelian, and F. Jensen,”Development of a statistically optimized test method for the reduced pressure test”: AFS Transactions 1998, pp. 365-378

15. A. Knuutinen, K. Nogita, S.D. McDonald, and A.K. Dahle, “Porosity formation in aluminum alloy A356 modified with Ba, Ca, Y and Yb”: Journal of Light Metals 1 (2001) 241-249

16. A.N. Khorasani, ”Effect of degassing on aluminum-silicon alloys”: AFS Transactions 1995, pp. 265-269

17. L.A. Godlewiski and J.W. Zindel, “Capability study of a filtration process to predict aluminum melt quality”: AFS Transactions 01-097 ([Page 1 of 11)

18. Metallographic analyses 8 PoDFA samples of January 1, 2006: The ABB Group

19. B.P. Cochran, M.L. Fenyes, and J.L. Jeanneret, “Flux practice in aluminum melting” AFS Transactions 1992, pp. 163-169

20. L.M. Fenyes, C. Nelson, and G. Reinkmann, “Melt quality up front”: Al Conf. 1992, pp. 577-589

21. Bob Braun, “Improving the reliability and consistency of a critical casting using foam filtration to control the mold filling process”: AFS Compiled Literature – Inclusions in Aluminum Casting Alloys, pp. 808-820

22. M. Drouzy, S. Jacob, and M. Richard, "Interpretation of Tensile Results by Means of Quality Index and Probable Yield Strength": AFS International Cast Metals Journal, June 1980, pp. 43-50

23. M. Tiryakioğlu and J. Campbell, “Quality index for aluminum alloy castings”: AFS Transactions 2013, Panel 13-1525

24. M. Tiryakioğlu, J. Campbell and N.D. Alexopoulos, “On the ductility potential of cast Al-Cu-Mg (206) alloys”: Materials Science and Engineering A, v. A506, pp. 23-26, (2009)

25. M. Tiryakioğlu, J. Campbell, “Ductility, structural quality, and fracture toughness of Al-Cu-Ag (A201) alloy castings”: Mater. Sci. Technol., vol. 25, pp. 784-789 (2009)

26. M. Tiryakioğlu, J. Campbell and N.D. Alexopoulos, “On the Ductility of Cast Al- 7%Si-Mg Alloys”, Metallurgical and Materials Transactions A, v. 40A, pp. 1000- 1007, (2009) 104

27. S. Shivkumar, and Ricci S. Steenhoff B Jr, “An experimental study to optimize the heat treatment of A356 alloy”: AFS Transactions 1989, pp. 791-810

28. J.Y. Hwang, R. Banerjee, H.W. Doty, M.J. Kaufman, “The effect of Mg on the structure and properties of type 319 aluminum casting alloys”: Acta Materialia 57 (2009) 1308-1317

29. Guang Ran, Jingen Zhou, and Q.G. Wang, “The effect of hot isostatic pressing on the microstructure and tensile properties of an unmodified A356-T6 cast aluminum alloy”: Journal of Alloys and Compounds 421 (2006) 80-86

30. X. Zhu, A. Shyam, J.W. Jones, H. Mayer, J.V. Lasecki, and J.E. Allison, “Effects of microstructure and temperature on fatigue behavior of E319-T7 cast aluminum alloy in very long life cycles”: International Journal of Fatigue 28 (2006) 1566-1571

31. R.K. Owens, H.W. Aantes, and R.E. Edelman, “Effect of nitrogen and vacuum degassing on properties of a cast aluminum-silicon-magnesium alloy (type 356)”: AFS Transactions 1957, pp. 37-44

32. B. Chamberlain and J. Sulzer, “Gas content and solidification rate effect on tensile properties and soundness of aluminum casting alloys”: AFS Transactions 1964, pp. 63-70

33. L. Liu and F.H. Samuel, “Effect of inclusions on the tensile properties of Al-7% Si- 0.35% Mg (A356.2) aluminum casting alloy”: AFS Compiled Literature – Inclusions in Aluminum Casting Alloys, pp. 383-395

34. R. Fuoco and E.R. Correa, “The effect of gating system design in the quality of aluminum gravity casting”: AFS Transactions, 2003

35. MagmaSoft v4.4, Magma Foundry Technologies, Chicago, IL, USA.

36. Carlson, K.D., and Beckermann, C., "Prediction of Shrinkage Pore Volume Fraction Using a Dimensionless Niyama Criterion," Metall. Mater. Trans. A, Vol. 40A, 2009, pp. 163-175

37. T.G. Kim and Z.H. Lee “Time-varying Heat Transfer Coefficients between Tube- shaped Casting and Metal Mold”, Int. J. Heat Mass Transfer. 1997, vol.40, pp.3513- 25.

38. D. Emadi, L.V. Whiting, M. Sahoo and D. Larouche, “Revisiting the ASTM B108 Test Bar Mold for Quality Control of Permanent Mold Cast Aluminum Alloys”: AFS Transactions, 2004, vol.112, pp. 225-236

39. E. Niyama, T. Uchida, M. Morikawa and S. Saito, “A method of shrinkage prediction 105

and its application to steel casting prctice”: AFS Int. Cast Met. J, 1982, vol. 7(3), pp.52-63.

40. G.K. Sigworth and T.A. Kuhn, “Use of ‘standard’ molds to evaluate metal quality and alloy properties”: AFS Transations 2008,

106