full papers thailand metallurgy conference

The 3rd Thailand Metallurgy Conference (TMETC 3)

Full Papers

“Metallurgical Research for Thailand Development”

26 – 27 October 2009

Century Park Hotel, Bangkok, Thailand

Department of Materials Engineering, Kasetsart University

Iron and Steel Institute of Thailand

National Metal and Materials Technology Center

Sahaviriya Steel Industries Public Co., Ltd

Thai Parkerizing Co., Ltd

Boon Rawd Brewery Co., Ltd.

Thai Nippon Steel Engineering & Construction Corp., Ltd.

Advance Pinnacle Technologies Pte Ltd.

DKSH Ltd.

Council of Engineers, Thailand

Organized by

Sponsored by

Conference Chairman

Asst. Prof. Wisit Locharoenrat Department of Materials Engineering, Kasetsart University

Advisory Committee

Mr. Wikrom Vajragupta Iron and Steel Institute of Thailand

Assoc. Prof. Dr. Weerasak Udomkichdacha National Metal and Materials Technology Center

Assoc. Prof. Dr. Paritud Bhandhubanyong National Science and Technology Development Agency

Assoc. Prof. Dr. Chatchai Somsiri Thainox Stainless Pcl.

Technical Committee

Faculty of Engineering, Kasetsart University

Asst. Prof. Wisit Locharoenrat Dr. Parinya Chakartnarodom

Dr. Ampika Bansiddhi Dr. Ratchatee Techapiesancharoenkij

Dr. Aphichart Rodchanarowan Mr. Thanawat Meesak

Dr.‐Ing. Patiphan Juijerm

National Metal and Materials Technology Center

Dr. Ekkarut Viyanit Dr. Julathep Kajornchaiyakul

Dr. Ruangdaj Tongsri Dr. Kritsada Prapakorn

Organizing Committees

Faculty of Engineering, Chulalongkorn University

Assoc. Prof. Dr. Gobboon Lothongkum Dr. Tachai Luangvaranunt

Assoc. Prof. Dr. Prasonk Sricharoenchai Dr. Seksak Asavavisithchai

Asso. Prof. Charkorn Jarupisitthorn Asst. Prof. Dr. Patama Visuttipitukul

Asst. Dr. Sawai Danchaivijit Dr. Panyawat Wangyao

Asst. Prof. Dr. Ekasit Nisaratanaporn Dr. Boonrat Lohwongwatana

Mr. Suvanchai Pongsugitwat

Metallurgy and Materials Science Research Institute, Chulalongkorn University

Ms. Kanokwan Saengkiettiyut Dr. Yuttanant Boonyongmaneerat

Dr. Nutthita Chuankrerkkul

Faculty of Engineering, King Mongkut’s University of Technology Thonburi

Assoc. Prof. Dr. Chaowalit Limmaneevichitr Dr. Pongsak Tuengsook

Assoc. Prof. Dr Bovornchok Poopat Dr.‐Ing. Paiboon Choungthong

Asst. Prof. Dr Sombun Charoeuvilaisiri Mr. Noppadol Kumanuvong

School of Energy, Environment and Materials, King Mongkut’s University of Technology Thonburi

Asst. Prof. Dr. Siriporn Rojananan Dr. Tippaban Palathai

Dr. Preecha Termsuksawad

Faculty of Engineering, King Mongkut's University of Technology North Bangkok

Asst. Prof. Dr. Somrerk Chandra‐Ambhorn Dr. Nattapong Sornsuwit

Asst. Prof. Dr. Witthaya Eidhed

School of Metallurgical Engineering, Institute of Engineering, Suranaree University of Technology

Dr. Narong Akkarapattanagoon Dr. Rattana Borisuttikul

Dr. Usanee Kitkamthorn Dr. Sakhob Kumkoa

Faculty of Science, Chiangmai University

Assoc. Prof. Dr. Torranin Chairuangsri

Faculty of Engineering, Prince of Songkla University

Asst. Prof. Dr. Thawatchai Plookphol Asst. Prof. Dr. Jessada Wannasin

The 3rd Thailand Metallurgy Conference

AF‐02

Effect of welding processes on the microstructure and hardness properties of weld metal on low alloy steel

AISI 4340

S. Limna, P. Muangjunburee

Department of Mining and Materials Engineering, Faculty of Engineering, Prince of

Songkla University, Hatyai, Songkla, 90112 Thailand

Tel: +6674 287323 Fax: +6674 212897 Email: [email protected]

Abstract

In this work, the microstructure and hardness properties of weld metal on low alloy steel

AISI 4340 have been investigated using shielded metal arc welding, metal active gas

welding and flux cored arc welding processes which the composition of filler wire nearly

the same alloying elements. The samples were mutipass welding. The microstructure and

hardness properties in weld metal were investigated by using optical microscope and micro

hardness tester. The microstructure of weld metal all three processes consisted of acicular

ferrite, grain boundary ferrite, polygonal ferrite and sideplate ferrite. However, the

microstructure of weld metal fabricated using flux cored arc welding process indicates

higher volume fraction of acicular ferrite than metal active gas welding and shielded metal

arc welding process. The results have revealed that the hardness of the weld metal

fabricated using flux cored arc welding process is greater than the weld metal fabricated

using shielded metal arc welding and metal active gas welding process.

Keywords : AISI 4340; Welding; Acicular ferrite; Polygonal ferrite

1. Introduction

Low alloy steel AISI 4340 are used in heavy duties engineering application for a long time.

This is a widely used low alloy steel that offers an advantageous due to high hardness, high

strength and excellent toughness [1-3]. As the machine part-members age, degenerate and

may fail in service or be declared unfit for further service on the basis of inspection and

remaining life assessments. Therefore, the machine part-members are repaired by welding

processes. However, welding can change the microstructure. Thus mechanical properties

were degraded in weld metal and heat affected zone (HAZ). Shielded metal arc welding

(SMAW), Metal Active Gas welding (MAG) and Flux cored arc welding (FCAW) process


AF‐02

are widely used in welding repair of machine part-members [4,5]. MAG and FCAW are a

semi or fully automatic arc welding process in which the electrode is continuously fed to

the weld area. Solid wire used in MAG but FCAW used flux cord wire that has the flux

material in the core of the tube [4]. On the other hand, SMAW is a manual process.

Automatic welding processes are favored over manual processes for the fabrication of

welded joints for number of reasons. Among these are increasing of productivity, lower

cost and a better control of geometry. However, FCAW process has became more popular

due to higher deposition rate and a better weld quality as compared to SMAW process [5].

This paper presents an investigation of microstructure and hardness of weld and base metal

change after welding by SMAW, MAG and FCAW processes.

2. Experimental

The base metal used in this investigation was the commercial AISI 4340 steel. Single bevel

butt joints were prepared to fabricate the weld. The samples were multi-pass welded by

Shielded Metal arc welding (SMAW), Metal Active Gas arc welding (MAG) and Flux

cored wire arc welding (FCAW). Electrodes and process parameters used to fabricate the

weld are given in table 1. Welding completion were post-weld heat treatment at 550๐C for

1 hour. The chemical composition of base metal and weld metal is shown in table 2. Cross

section samples were cut from the all weld samples. The samples were ground surface until

1200 grits. After that, samples were polished to a 1 µ alumina finish. The weld metal

microstructure was revealed by etching with a freshly prepared 2% natal solution. The

microstructure analysis of the weld metals were studied using a light optical microscope.

Vicker’s microhardness testing machine was used to measure the weld metal and base

metal.

3. Results

Microstructures

The main aim of this investigation was to understand the microstructure of welded sample

with different processes such as SMAW, MAG and FCAW processes. The typical

microstructures of base metal and weld metal are presented in Fig.1. The microstructure

feature of the base metal shows tempered bainite (Fig.1a). In general, the microstructures

of weld metal obtained from all processes consisted of acicular ferrite, polygonal ferrite

and sideplate ferrite. Volume fraction of microstructures is presented in Fig.3. An optical


AF‐02

microstructures of weld metals fabricated using SMAW, MAG and FCAW processes

present predominantly acicular ferrite. However, the microstructure of weld metal

fabricated using FCAW process indicates higher volume fraction of acicular ferrite than

MAG and SMAW processes.

Table 1 Welding conditions

Parameter Unit SMAW MAG FCAW

Electrode Types (AWS) - E11018-G H

4R

ER110S-G E110T5-K4H4

Preheat temperature ๐C 350 350 350

Electrode baking temperature ๐C for 1 hr. 350 - -

Mixer gas flow rate l/min - 12 12

Filler diameter mm. 4 1.2 1.2

Current A 145 230 230

Voltage V 26 25 25

Welding speed mm/min 160 300 300

Heat input KJ/mm 1.41 1.15 1.15

Table 2 Chemical composition of base metal and weld metals

Type of materials C Mn Si P S Mo Ni Cr

Base metal 0.39 0.74 0.19 0.024 0.019 0.23 1.72 0.8

SMAW 0.05 1.5 0.4 - - 0.5 2.0 0.4

MAG 0.08 1.4 0.6 - - 0.4 2.5 0.30

FCAW 0.05 1.40 0.005 0.015 0.50 2.40 2.4 0.50

Hardness

Vicker’s hardness testing machine was used to measure the weld metal and the base metal

hardness and the values are presented in Fig.2. From this figure, the hardness distribution

of the weld’s cross section was clearly found to be different among process. The hardness

of the base metal is approximate 290 HV. Weld metals fabricated using SMAW, MAG and

FCAW processes exhibit 250, 230 and 275 HV, respectively. The hardness value of

samples fabricated by FCAW revealed higher hardness in the area of weld metal than

samples fabricated by MAG and FCAW processes.


AF‐02

Fig. 1. Optical microstructures of base metal and weld metal: (a) Base metal;

(b) SMAW; (c) MAG; (d) FCAW

Fig. 2. Vicker’s hardness distribution of base metal (left), weld metal (middle) and base

metal (right) of the samples were fabricated by SMAW, MAG and FCAW

a b

c d


AF‐02

Fig.3. Volume fraction of microstructures in different processes.

4. Discussion

It is a common practice to correlate the various weld metal properties with heat input. The

difference in the ferrite morphology in low alloy steel welds is due to the difference in heat

input. The formation of acicular ferrite is controlled by weld heat input. Thus if the heat

input is higher the content of the acicular ferrite will be very less and vice versa. On the

other hand higher heat input will enhance the formation of coarse pro-eutectoid ferrite or

polygonal ferrite in the weld metal region. Weld cooling rate plays the decisive role in

determining weld microstructure in high strength steels. The general effect of increasing

the cooling rate is to lower transformation temperatures. When cooled at sufficiently low

rates, the microstructure predominantly tends to become polygonal ferrite. In the present

investigation, heat input of 1.41 KJ/mm. was fabricated by SMAW and 1.15 KJ/mm. was

fabricated by MAG and FCAW. The microstructure of the weld metal region consisted of

acicular ferrite, grain boundary ferrite, polygonal ferrite and sideplate ferrite. An acicular

ferrite microstructure has the potential of combining high strength and high toughness.

Acicular ferrite is formed in the interior of the original austenite grains by direct nucleation

from the inclusions resulting in a randomly oriented short ferrite needles with a basket

weave features. It has been accepted that polygonal ferrite is bad for weld metal toughness

because it offers little resistance to cleavage crack propagation. In addition, the results

confirm that the hardness in weld metal fabricated using FCAW process higher than weld

metal fabricated using MAG and SMAW process. Therefore, the welding process has a

significant in the weld metal microstructure. This has a direct influence in weld metal

hardness.


AF‐02

5.Conclusions

The microstructure and hardness properties of weld metal on low alloy steel AISI 4340

fabricated using SMAW, MAG and FCAW processes have been investigated and the

conclusions are as follow:

1. The microstructure constituents such as acicular ferrite, polygonal ferrite and

sideplate ferrite are influenced by heat input.

2. The hardness of weld metal is significantly depending on microstructure.

3. FCAW process indicates both microstructure and hardness better than SMAW and

MAG processes.

6. Acknowledgements

The author gratefully acknowledge the financial support from the electric Generating

Authority of Thailand (EGAT). Thanks are also to the Department of Mining and Materials

Engineering, Prince of Songkla University, Hat- Yai, Thailand for providing equipment

and facilities.

7. References

[1] P.Muangjunburee. 2007. Improvement of Metallurgical and Mechanical Properties of

Welding Surfacing on High Strength Steel AISI 4340 by Various Preheating Temperatures.

Proceedings of international Conference the Frontiers of Technolog: 321-324.

[2] P.Muangjunburee. 2007. Improvement of metallurgical and Mechanical properties of

welding surfacing on high strength steel AISI 4340 by Post-weld heat treatment. The First

South-East Asia IIW Congress: 273-277.

[3] Woei Shyan Lee and Tzay Tian Su. 1999. Mechanical properties and microstructural

features of AISI 4340 high-strength alloy steel under quenched and tempered conditions.

Journal of Materials Processing Technology 87: 198-206.

[4] G. Magudeeswaran, V. Balasubramanian and G. Madhusdhan Reddy. 2008. Effect of

welding processes and consumables on high cycle fatigue life of high strength, quenched

and tempered steel joints. Journal of Materials and Design.

[5] T.Lant, D.L. Robsinson, B.Spafford and J.Storesund. 2004. Review of weld repair

procedures for low alloy steels designed to minimize the risk of future cracking.

International Journal of Pressure Vessels and Piping 78: 813-818.


AF‐05

An investigation of microstructural change of low alloy steel AISI 4150 by Seebeck coefficient

T. Samrana, P. Termsuksawadb*

aSchool of Metallurgical Engineering, Institute of Engineering,

Suranaree University of Technology, Nakhon Ratchasima, 30000 Thailand bDivision of Materials Technology, School of Energy Environment and Materials,

King Mongkut's University of Technology Thonburi, Bangkok 10140 Thailand,

Tel: 0-2470-8643, Fax: 0-2427-9062

Email: [email protected]

Abstract

Low alloyed steel, whose hardness can be increased by heat treatment, have been widely used

in various applications. After heat treatment, it is normally destructively characterized and

tested by many approaches such as microstructure characterization by optical microscope and

hardness testing. It is useful to develop a non-destructive method to characterize its properties

and microstructures. The Seebeck effect is a phenomenon in which the electrical potential

gradient develops due to temperature difference. The magnitude of the Seebeck effect is

demonstrated by the Seebeck coefficient, which can be altered by electronic properties or

microstructure changes. The materials in this study were cylindrical carbon steels AISI 4150

with diameter of 1.3 cm and length of 3 cm. The specimens were heat-treated at 900 oC for 1

hour, and then cooled to room temperature in furnace and in various mediums: air, oil and

water. In addition one of the samples was cooled in salt bath at 350 oC for 1 hour before water

cooled. An x-ray diffractometry (XRD) and optical microscopy (OM) were used to

characterize their crystal structures and microstructures, respectively. The Seebeck coefficient

was measured relative to that of copper. The result indicated that Seebeck coefficient increases

with hardness, which is controlled by microstructure. In conclusion, the Seebeck coefficient

measurement could be possibly applied to study microstructure of low alloyed steels.

Keywords: Seebeck coefficient, low alloyed steel, heat treatment, XRD, microstructure


AF‐05

Introduction

Low alloy steels are steels with additions of alloying elements such as nickel, chromium and

molybdenum. These alloying elements increase hardenability of the steels; as a result,

mechanical properties of these steels can be improved by heat treatment. Because these steels

possess good mechanical properties, they have been widely used in many applications such as

automobile parts, structural steel parts, pipelines, etc. After heat treatment, microstructure of

the steels is altered and their mechanical properties such as hardness and strength were

improved. Typically, mechanical properties of heat treated steels such as hardness and strength

are examined by some destructive tests. Therefore, it is useful to develop a nondestructive

technique to predict these properties. One of the candidates is Seebeck coefficient or

thermoelectric power measurement which measure amount of induced voltage developed by

temperature difference, TVS

TAB ΔΔ

=→Δ 0

lim [1]. Seebeck coefficient or thermoelectric power is

contributed by two components: diffusion and phonon-drag thermoelectric power. Phonon-

drag thermoelectric power is very small and can be negligible at room temperature or above.

Diffusion thermoelectric power is a function of electrical conductivity and effective mass [2].

Effective mass is defined as curvature of electronic structure at the Fermi level [2]. From this

definition, effective mass, m*, is calculated by 2

22*

dkEdm = , where 2 and 2

2

dkEd are Plank’s

constant, and curvature of electronic structure at Fermi level, respectively. Both electrical

conductivity and effective mass are function of microstructure and electronic structure, as a

result, Seebeck coefficient depends on these structures as well. It should be noted that sign of

Seebeck coefficient depends on types of carriers [3]. If a carrier is an electron, Seebeck

coefficient will be negative. In contrast, positive Seebeck coefficient is found when hole is a

carrier. The magnitude of Seebeck coefficient depends on effective mass and difficulty of

carrier transport. Effects of microstructure of carbon steel on Seebeck coefficient were studied

by various research groups [4-7]. Effect of annealing on thermoelectric power of low carbon

steel containing 460 ppm aluminium and 74 ppm nitrogen was investigates by Brami et al [4].

In this study, thermoelectric power increased with amounts of AlN and carbon precipitation.

For ultra low carbon steel, Seebeck coefficient was found to be decreased with increasing

defect concentration or amount of dissolved element in the matrix, and be increased with the


AF‐05

amount of precipitation [5, 6]. The increasing of Seebeck coefficient due to amount of

precipitates was also found in martensitic stainless steel when Caballero et al. studied the

effect of carbide precipitation on Seebeck coefficient of heat treated stainless steel [7]. In

addition, Caballero et al. indicated that Seebeck coefficient can be increased with grain size of

austenite due to decreasing of grain boundary concentration. From literatures; therefore, it is

possible to study effect of heat treatment on microstructure and mechanical properties of low

alloy steel via Seebeck coefficient. This study aims to investigate this relationship in order to

further develop this concept as non-destructive testing for heat treated steel.

Experimental procedure

Low alloy steel grade AISI 4150, whose diameter is 1.3 cm. with the length of 3 cm, was used

in this study. The sample composition was analyzed by emission spectroscopy as shown in

table 1. From this table, main alloying elements in this steel are chromium and molybdenum,

which increases hardenability of the steel.

Table 1: composition of sample, analyzed by emission spectroscopy (wt.%)

C Mn P S Cr Mo Si

AISI 4150 (std.) 0.48-0.53 0.75-1.00 < 0.035 < 0.04 0.75-1.2 0.15-0.25 0.15-0.3

Sample 0.489 0.789 0.021 0.002 0.851 0.177 0.192

The samples were annealed at approximately 900 oC for 1 hr. and subsequently cooled in

different media: furnace cool, air cool, water cool and oil cooled. In addition, one of the

samples was cooled in salt bath at 350 oC for 1 hr and then cooled in water. Three

observations for each treatment were conducted. Seebeck coefficients, relative to Seebeck

coefficient of copper, of each sample were measured after heat treatment. The configuration of

the Seebeck coefficient apparatus was demonstrated in figure 1. The absolute Seebeck

coefficient was calculated by the equation:

Cua STVS +

ΔΔ

= (1)

where aS , CuS , VΔ and TΔ are absolute Seebeck coefficient (μV/K), Seebeck coefficient of

copper (μV/K), induced voltage difference (V) and temperature difference, respectively. The

temperature at the cold side is about 26 oC and temperature difference between hot and cold


AF‐05

sides in this experiment was set at 4 oC. Seebeck coefficient of copper at 300 K is 1.83 μV/K

[1].

Figure 1: Diagram of Seebeck coefficient measurement apparatus

Next, hardness and microstructures of the samples were investigated by hardness test

Rockwell scale C with loading of 150 kgf and optical microscope. The crystal structures of

each sample were also examined by D-8 Bruker x-ray diffractometer using Cu-Kα as x-ray

source, step width of 0.02 degree and step time of 0.04 s.

Results and discussion

Hardness and Seebeck coefficient

Harnesses and Seebeck coefficients of as-received samples and heat-treated low alloy steels

after quenching with different media were shown in figure 2. Negative Seebeck coefficient

pointed out that electron is carrier responsible for thermoelectric power of the samples.

According to Vedenikov [8], Seebeck coefficient of pure iron at 300 K is approximately +12

μV/K. However, Seebeck coefficient of steel is perturbed by element in solid solution,

microstructure, dislocation and precipitates [5, 6]. Among these contributions, contribution

from solute atom is the greatest because solute atoms act as new diffusion centers for electron

[5]. The contribution from solute atom to Seebeck coefficient, iSΔ , obeys the linear law as

shown by:


AF‐05

∑=Δ iii CKS (1)

where Ki and Ci are the specific thermoelectric power per weight percent of solute element i

(μV/(K-wt%)) and amount of solute element i (wt.%), respectively. The value of Ki depends

on various factors such as chemical composition, texture, grain size, etc [6]. For example the

KC varies from -20 to -52 μV/K-wt% depending on carbon contents [5-7]. The higher the

carbon content, the lower is the KC value. KCr and KMn were reported as -0.30 and -3 μV/K-

wt%, respectively. Beside contribution from solute element, the contribution from dislocation

also leads to negative Seebeck coefficient [5]. The sign of change of Seebeck coefficient due

to contribution from microstructure relies on type of phase transformation. For example,

amount of retain austenite in martensitic stainless steel leads to positive Seebeck coefficient

with specific thermoelectric power constant of +0.087 μV/K-wt% [7]. It can be seen that the

magnitude of specific thermoelectric power due to microstructure is less than those of solute

atom in the order of magnitude; therefore, the negative Seebeck coefficient of steel is

expected.

Figure 2: Hardness and Seebeck coefficients of as-received sample and heat treated low alloy

steels after quenching in different media


AF‐05

Figure 2 also demonstrates dependent of Seebeck coefficient on quenching media or cooling

rate. When cooling rate is high, e.g. quenching in oil or water, magnitude of Seebeck

coefficient decreases and hardness increases. This phenomenon may be used to establish

relationship between Seebeck coefficient and hardness. It is well known that for fast cooling

rate carbon atoms do not have enough time to diffuse out of the austenite to form equilibrium

microstructure of pearlite. Consequently, depending on cooling rate, non-equilibrium

structures such as bainite or martensite will form and hardness of quenched sample increases.

The effect of cooling rate on Seebeck coefficient may be explained by crystal structure and

microstructure as discussed later.

XRD-result

Effect of cooling rate on crystal structure can be seen by x-ray diffraction pattern

demonstrated in figures 3 and 4. From these figures, crystal structures of quenched samples

can be sorted into two groups: 1) samples quenched at low and moderate cooling rate and 2)

samples quenched at high cooling rate. Figure 3 demonstrated that crystal structure of sample

with low cooling rate is body center cubic with diffracted planes: (110), (200) and (211). In

addition, (200) and (211) peaks tend to disappear when cooling rate is high. When considering

at (110) peak (figure 4), (110) peaks of water-quenched and oil quenched samples are shifted

from those of the other samples. In addition, they are broader than those of samples quenched

at low and moderate cooling rate. The shift of the peak indicates that crystal structures of the

oil and water quenched sample are different from the other samples and the broader peaks

indicates occurrence of lattice distortion during fast cooling. As shown in figure 2, magnitude

of Seebeck coefficients of oil cooled and water cooled samples are lower than those of other

samples. The reduction of the magnitude of Seebeck coefficient may be due to phase change

and lattice distortion. The distortion reduces electron movement; consequently, magnitude of

Seebeck coefficient decreases. This explanation can also be applied when the Seebeck

coefficients and x-ray diffraction patterns of only oil quenched and water quenched samples

are compared. However, rather than chemical composition, crystal structure is not only a

factor affecting Seebeck coefficient. To explain variation of Seebeck coefficients of as-

received, furnace cooled, air quenched samples and of sample quenched in salt bath,

microstructure analysis is needed.


AF‐05

Figure 3: X-ray diffraction pattern of as received sample and quenched samples


AF‐05

Figure 4: [110] peak of as received sample and quenched samples

Microstructure

Microstructure of each sample was shown in figure 5. From this figure, microstructures of oil

quenched and water quench sample are martensite with some ferrite. It is well known that

dislocation density of sample quenched with high cooling rate is very high. As a result,

hardness of these samples is high. The dislocation not only increases hardness, it also impedes

electron transport. Therefore, rather than the effect of structure distortion, magnitudes of

Seebeck coefficients of oil quenched and water quenched samples are reduced by the existing

dislocation.


AF‐05

(a) (b)

(c) (d)

(e) (f)

Figure 5: Microstructures of samples: a) as received, and heat-treated sample with different

quenching media: b) furnace, c) air, d) saltbath, e) oil and f) water at magnification of 500x


AF‐05

Figure 5 also points out the presence of ferrite and pearlite in samples cooled in furnace. For

as-received sample and samples cooled in air and salt bath, microstructure of these samples

consists of ferrite and cementite. Unlike microstructure of furnace cooled sample, the ferrite

and cementite of these samples are not lies in lamellar order as illustrated in figure 6. The

nonlamellar array of ferrite and cementite is classified as bainite [9]. In addition, cementite in

as-received, air cooled and salt bath cooled samples disperses all over microstructure.

Consequently, hardness of these samples is higher than that of furnace cooled sample.

Although microstructure of furnace cooled sample is different from those of as-received and

air quenched sample, their Seebeck coefficients are not significantly different. The reason may

be because the phases (ferrite and cementite) present in these samples are the same. However,

this is not conclusive and more investigation is needed. Seebeck coefficient data and XRD

result also point out that although crystal structures of as-received sample and of samples

quenched in air and salt bath are the same, magnitude of Seebeck coefficient of sample

quenched in salt bath is higher than those of the other samples. The higher magnitude of

Seebeck coefficient may be due to larger grain size. Theoretically, grain boundary behaves as

an obstacle for electrical transport. The increasing of magnitude of Seebeck coefficient due to

increasing grain size was also found by Caballero et al [7].


AF‐05

(a)

(b)

Figure 6: Microstructures of samples cooled in

a) furnace cooled and b) salt bath at magnification of 1000x


AF‐05

Conclusions

Seebeck coefficients and hardness of quenched low alloy steels, AISI 4150, were studied.

Although Seebeck coefficient depends on crystal structure and microstructure, hardness of

samples cannot be directly related to Seebeck coefficient. Hardness of these steels can be

explained directly by their microstructures. In contrast, explanation of Seebeck coefficient by

microstructure is more complicate. Magnitude of Seebeck coefficient of quenched samples is

influenced by phases present in sample and grain size.

References

(1) Blatt F.J., Schroeder P.A., Foiles C.L. and Greig D.L., 1976, Thermoelectric Power of

Metals, Plenum Press, New York

(2) N.F. Mott and Jones H., 1936, The Theory of the Properties of Metals and Alloys,

Dover Inc., New York, 310

(3) Kasap, S., 1996, Thermoelectric Effects in Metals: Thermocouples [Online], Available

: http://www.materials.usask.ca/samples/Thermoelectric-Seebeck.pdf [September 9,

2009]

(4) Brahmi, A. and Borrelly, R., 1997, “Study of Aluminium Nitride Precipitation in Pure

Fe-Al-N Alloy by Thermoelectric Power Measurements”, Acta Materialia, 45, 1889-

1897

(5) Lavaire, N., Merlin, J. and Sardoy, V, 2001, “Study of Ageing in Strained Ultra and

Extra Low Carbon Steels by Thermoelectric Power Measurement”, Scripta Materialia,

44, 553-559.

(6) Massardier, V., Lavaire, N., Soler, M. and Merlin, J., 2004, “Comparison of the

Evaluation of the Carbon Content in Solid Solution in Extra-mild Steels by

Thermoelectric Power and by Internal Friction”, Scripta Materialia, 50, 1435-1439.

(7) Caballero, F.G., Capdevila, C., Alvarez, L.F. and García de Andrés, C., 2004,

“Thermoelectric power studies on a martensitic stainless steel”, Scripta Materialia, 50,

1061-1066.

(8) Vedernikov M.V., 1969, Adv. Physics., 18, 337

(9) Krauss G., 1990, Steel: Heat Treatment and Processing Principles, ASM international,

Ohio, U.S., pp.78


BF‐04

Corrosion Assessment of Carbon Steel in Thailand by Atmospheric Corrosion Monitoring (ACM) Sensors

Wanida Pongsaksawada, Ekkarut Viyanita, Sikharin Sorachota,

and Tadashi Shinoharab

aNational Metal and Materials Technology Center (MTEC), Pathumthani, Thailand,

Tel.: 0-2564-6500 Fax: 0-2564-6338 Email: [email protected] bNational Institute for Materials Science, Ibaraki, JAPAN

Tel: +81-298-59-2604 Fax: +81-298-59-2601

Abstract

Atmospheric corrosion of metal depends on material compositions, weather condition (dry,

dew, and rain period), temperature, relative humidity, and airborne sea salt of specific

location. General testing procedure to obtain the corrosion rate is by actual exposure test of

the specimen panels based on time interval plan. In Japan, atmospheric corrosion

monitoring (ACM) sensor, made of an iron-silver galvanic couple, has been developed and

used to sense the corrosivity in terms of galvanic current. Under some atmospheric

conditions, these data can be converted to time of wetness and related to the corrosion rate

of carbon steel. With ACM sensors, it is possible to monitor the corrosion rate in a shorter

time than the exposure test. To apply the ACM sensors in Thailand, it is necessary to

evaluate the effectiveness and correlation between the actual corrosion rate and the sensor

output. In this research during June 2007 – May 2009, we performed exposure tests of

carbon steel (JIS SS400) along with ACM sensors under outdoor and sheltered conditions

at three locations: (1) Rama VI Road, Bangkok (2) Suvarnabhumi International Airport,

Samutprakarn and (3) Royal Thai Navy Dockyard, Chonburi, representing urban, airport,

and marine environments, respectively. Weather data were obtained from temperature,

relative humidity, and ACM sensors. To estimate the corrosion rate, weight loss

measurements were carried out on specimens exposed for 1 month period over 2 years.

Average monthly weight loss ranks from high to low as marine, airport, and urban

environments. The relationship between outdoor corrosion rate and ACM output is found

to be linear on a log-log scale at airport and urban test stations during March 2008 – May

2009.


BF‐04

Keywords Atmospheric Corrosion, ACM sensor, Carbon steel

1. Introduction

Atmospheric corrosion of metal is governed by chemical composition of thin film

electrolyte on the metal surface which is dependent on air pollutants, humidity, and

temperature. Corrosion scientists in several countries have been carried out exposure tests

to investigate the effects of the environment on corrosion rates (Pourbiax (1982), and the

corrosion resistance of different materials (Chen et al. (2005), De La Fuente et al. (2007),

Singh et al. (2008), Sun et al. (2009), Tahara et al. (2005), and Veleva et al. (2009)). The

actual field tests usually take 10-20 years for an evaluation period. To accelerate the

experimental study, simulated wet-dry cyclic tests have been performed for qualitative

observation (Han et al. (2007) and Katayama et al. (2005)).

Electrochemical measurement such as AC impedance monitoring sensor has been

incorporated into the atmospheric corrosion tests by Nishikata et al. (2005), Shitanda et al.

(2007), Wall et al. (2005) to enhance the understanding of corrosion process and monitor

quantitative parameters as a function of environmental factors. Another electrochemical

measurement by atmospheric corrosion monitoring (ACM) sensor relates galvanic current

with corrosion rate. The impedance and ACM sensors have been applied to monitor the

corrosion in industrial plants and infrastructure. In Japan, ACM sensor, made of Fe-Ag

galvanic couple, has been developed and used to monitor the corrosivity of various

atmospheric conditions in the work of Motoda et al. (1994) and Shinohara et al. (2006).

Linear relationship between outdoor corrosion rate and sensor galvanic current output was

found at severe marine and rural/marine environments in Japan (Shinohara et al. (2006)).

In Thailand, atmospheric corrosion tests had been conducted on organic-coated

carbon steel by Bhamornsut et al. (2003), zinc by Phantor et al. (2003), and stainless steel

by Daopiset et al. (2008). This present research is the first to apply the ACM sensor in

atmospheric corrosion study of structural steel in Thailand. The exposure tests of the test

panels as well as the ACM sensors were carried out from June 2007 – May 2009 at three

different environmental conditions. Weight losses and sensor outputs were evaluated.

2. Experimental procedures

Exposure test stations were selected for this field study. The details at each site are

described in Table 1.


BF‐04

Table 1

Structural steel plates (JIS SS400) were cut into rectangular coupons with

dimension of 150mm x 70mm x 6mm. Blue oxide scales were removed by HCl acid,

sandblasting, and mechanical polishing. The initial weights of the samples were recorded.

Exposure tests were carried out in open-air (outdoor) and under shelter (indoor) conditions

for 1 and 12 months. The tests were repeated for 24-month period. An ACM type

corrosion sensor was installed on each test rack and connected to a data logger (Syrinx

Inc.). Picture of a test station is illustrated in Fig. 1. The ACM sensors were replaced

every month. Temperature and humidity sensors were installed under a cover at each

location and connected to the data logger. Electrical current (Q), temperature (T), and

relative humidity (RH) were recorded in a memory card every 10 minutes. After the test,

specimen panels and data were collected for analyses. Two specimens were cleaned

according to ASTM G01 to remove corrosion products. The average weight loss was

determined. Monthly results were related to the sensor data to evaluate correlation with

ACM sensor. Annual results were fitted to a multiple linear regression model as a function

of environmental parameters.

Figure 1

3. Results and discussion

Short – term exposure test

Monthly results from June 2007 – November 2008 were reported in the previous

work (Pongsaksawad et al. (2009)). With additional data from December 2008 – May 2009,

the average monthly weight losses over two years are summarized in Table 2. Corrosivity

ranks from high to low as marine, airport, and urban atmosphere or in the increasing

distance from the sea shore as expected. The sheltered environments are typically less

corrosive than open air condition as seen by smaller magnitude of average corrosion losses.

The corroded sheltered specimens were influenced only by dew condensation, temperature,

relative humidity, sea salt and air pollutants, whereas the specimens exposed outdoor were

influenced by rain fall as well. However, during some months in rainy season, the sheltered

samples were more severely corroded than outdoor samples due to rain wash affect that

removes corrosive species from the metal surface.

Table 2


BF‐04

Correlation between corrosion rate and sensor output

Corrosion rates of one – month exposure test were plotted as a function of the

ACM sensors output to evaluate their relationships. For sheltered condition, the corrosion

rates were related to the daily average electricity (Q). The best correlation (R = 0.7220)

was observed at urban site during March 2008 – April 2009 as shown in Fig. 2(a) as:

log CRurban [mmpy] = 0.165 log Q [C/day] – 0.658 (1)

No correlation was found at marine and airport test sites.

In case of outdoor environment, the current during rain period (Qrain) is much

higher than dew period (Qdew). Thus the effective sensor output (Qeff) is defined as Qeff

=Qdew + �Qrain, where �is 0.2 [Shinohara 2004]. As shown in Fig. 2(b), the relationship

between corrosion rate and effective sensor output for urban site has a strong positive

correlation (R = 0.7113) during March 2009 – May 2009 and follows the expression:

log CRairport, urban [mmpy] = 0.183 log Qeff [C/day] – 1.056 (2)

No correlation is observed for marine exposure sites. In the atmospheric corrosion study

with this Fe-Ag type ACM sensor in Japan (Shinohara et al. (2006)), the ACM sensors

could be used to estimate the atmospheric corrosion rate in severe marine and rural/marine

conditions, but not in the mild marine atmosphere. Thailand has less temperature

fluctuation and longer time of wetness, which may require another type of ACM sensor for

marine environment. Further study by using a long life ACM sensor is under consideration.

Figure 2

Multiple linear regression model

The conventional method to predict the corrosion rate is by finding an empirical

relationship with the active environmental parameters such as in the atmospheric corrosion

study of Vietnam by Hong Lien et al. (2009). The simplest model is a multiple linear

function. Generally, one –year exposure tests are conducted and repeated to obtain reliable

sampling data. In this study, two sets of one – year exposure tests were carried out at each

test station during June 2007 to May 2009. The average corrosion rates of each phase and

other environmental parameters are reported in Table 3.

Table 3

Based on our one – year exposure test data shown in Table 3, the best correlation suggests

that the outdoor corrosion rate (CR) is a function of temperature, relative humidity, and

total rain time as:

CR [g/ m2 / y] = 446.9 - 11.850 T [°C ] + 0.535 RH [%] + 0.028 Train [h/ y] (3)


BF‐04

Temperature has a negative affect on corrosion rate. Higher temperature causes the water

droplet on the specimen surface to evaporate; thereby, the corrosion rate is reduced. Both

relative humidity and total rain time have positive affects on corrosion rate due to

increasing time of wetness. Time of rain slightly contributes to corrosion because it also

washes away the corrosive residues. The calculated corrosion rates (Eq. 3) were plotted

against the actual values as shown in Fig. 3 with R = 0.9733. However, other dependent

variables such as SO2 and Cl- ions were not taken into account since they were not

monitored during the exposure period.

Figure 3

Comparing the two correlation methods discussed above, the ACM sensor is

applicable for corrosion prediction at airport and urban environments. With the use of

ACM sensor, corrosion rate can be monitored in real time without the need to conduct a

long-term field test. For marine site, the multi-variable model can be applied.

4. Conclusions

(1) The atmospheric corrosion of structural steel decreases with increasing distance

from the coast (marine > airport > urban).

(2) In Thailand, the atmospheric corrosion of structural steel under sheltered

environment is generally less corrosive than that under outdoor environment. The

rain wash affect is not a major contribution.

(3) The corrosion losses can be estimated by the ACM sensor output at airport and

urban test station.

(4) The outdoor corrosion losses at marine, airport, and urban atmosphere increases

with decreasing temperature, increasing relative humidity, and increasing time of

rain.

5. Acknowledgement

The authors gratefully acknowledge the financial support from the National Metal and

Materials Technology Center (MTEC), Thailand and the technical support from the

National Institute for Materials Science (NIMS), Japan.

6. References

Bhamornsut, C., L. Chotimongkol, R. Nakkuntod, S. Suphonlai, T. Kodama, and H.

Tanabe, Atmospheric


BF‐04

Degradation of organic coatings in Thailand, Proc. of Japan Society of Corrosion

Engineers

Conference, November 16-21, 2003, Japan.

Chen, Y.Y., H.J. Tzeng, L.I. Wei, L.H. Wang, J.C. Oung, H.C. Shih, Corrosion resistance

and mechanical

properties of low-alloy steels under atmospheric conditions, Corrosion Science, 47,

(2005) 1001-

1021.

Daopiset, S., P. Wanaosod, T. T. Xuan Hang, and T. Anh Truc, Atmospheric corrosion of

stainless steels 304

and 316 with different surface finishes, Proc. of the 5th Thailand Materials Science

and Technology

Conference, September 16-19, 2008, Thailand.

De La Fuente, D., J.G. Castano, M. Morcillo, Long-term atmospheric corrosion of zinc,

Corrosion Science,

49, (2007) 1420–1436.

Han, W., G. Yu, Z. Wang, J. Wang, Characterization of initial atmospheric corrosion

carbon steels by field

exposure and laboratory simulation, Corrosion Science, 49, (2007) 2920–2935.

Hong Lien, L. T., P. Thi San and H. Lam Hong, Atmospheric corrosion of carbon steel in

Vietnam: The

relationship between corrosion rate and environmental parameters and the

classification of

atmospheric corrosivity of carbon steel, Proc. of Japan Society of Corrosion

Engineers Conference, May 22-24, 2009, Japan, A305.

Katayama, H., K. Noda, H. Masuda, M. Nagasawa, M. Itagaki, K. Watanabe, Corrosion

simulation of

carbon steels in atmospheric environment, Corrosion Science 47 (2005) 2599–2606.

Motoda, S., Y. Suzuki, T. Shinohara, Y. Kojima, S. Tsujikawa, W. Oshikawa, S. Itomura,

T.Fukushima and

S.Izumo, Zairyo-to-Kankyo, 43, (1994), 550.

Nishikata, A., F. Suzuki, T. Tsuru, Corrosion monitoring of nickel-containing steels in

marine atmospheric

environment, Corrosion Science, 47, (2005) 2578–2588.


BF‐04

Panther, B. C., M. A. Hooper, G. P. Ayers, I. Cole, W. Limpaseni, W. Somboon, F.

Veersai, W. Veersai,

Atmospheric depositions and corrosion Impacts in Bangkok, Proc.of the 2nd

Regional Conference

on Energy Technology Towards a Clean Environment, February 12-14, 2003,

Phuket, Thailand, Vol.

2, pp. 675-683.

Pongsaksawad, W., S. Sorachot, J. Troset, E. Viyanit, and T. Shinohara, Applying

atmospheric corrosion

monitoring sensor for carbon steel under various exposure test sites in Thailand ,

Proc. of Japan

Society of Corrosion Engineers Conference, May 22-24, 2009, Japan, A304.

Pourbaix, M., The Linear Bilogarithmic Law for atmospheric corrosion, Wiley, New York,

(1982), 107.

Shinohara, T., A. Tahara and Y. Hosoya, Datasheets of Atmospheric Corrosion behaviors of

low alloyed steels

with corrosivities at exposure test sites, Proc. of the 3rd International Conference

on Advanced

Structural Steels, Aug 22-24, 2006, Korea.

Shitanda, I., A. Okumura, M. Itagaki, K. Watanabe, Y. Asano, Screen-printed atmospheric

corrosion

monitoring sensor based on electrochemical impedance spectroscopy, Sensors and

Actuators, B 139,

(2009) 292–297.

Singh, D.D.N., S. Yadav, J. K. Saha, Corrosion of low carbon steel in atmospheric

environments of different

chloride content, Corrosion Science, 50, (2008) 93–110.

Sun, S., Q. Zheng, D. Li, J. Wen, Long-term atmospheric corrosion behaviour of

aluminium alloys 2024 and

7075 in urban, coastal and industrial environments, Corrosion Science, 51, (2009)

719–727.

Tahara, A., T. Shinohara, Influence of the alloy element on corrosion morphology of the

low alloy steels


BF‐04

exposed to the atmospheric environments, Corrosion Science, 47, (2005) 2589–

2598.

Veleva, L., M. Acosta, E. Meraz, Atmospheric corrosion of zinc induced by runoff,

Corrosion Science, 51,

(2009) 2055–2062.

Wall, F.D., M.A. Martinez, N.A. Missert, R.G. Copeland, A.C. Kilgo, Characterizing

corrosion behavior

under atmospheric conditions using electrochemical techniques, Corrosion Science,

47, (2005) 17-32.


BF‐04

Table 1. Locations of exposure test sties

Environment Location Description

Marine Sattahip Navy Dockyard, Chonburi On the ground facing the

Gulf of Thailand

Airport Suvarnabhumi International Airport, Samutprakarn

On the ground nearby the

runway and industrial

district

Urban National Science and Technology Development Agency,

Bangkok

On the roof top of a 7‐ story

building influenced by

heavy traffic


BF‐04

Table 2. Average monthly weight losses of outdoor and sheltered conditions

Weight Loss (g/ m2) Location Phase

sheltered outdoor

June 07 – May 08 46.476 57.786 Marine

June 08 – May 09 44.535 55.614

June 07 – May 08 39.452 56.333 Airport

June 08 – May 09 37.793 48.472

June 07 – May 08 28.280 38.286 Urban

June 08 – May 09 30.867 46.567


BF‐04

Table 3. Corrosion rate of one- year exposure test and environmental parameters:

temperature (T), relative humidity (RH), and time of rain (Train).

Site Phase CR [g/ m2 /

y] T [°C] RH [%] Train [h/ y]

Marine June 07 – May 08 137.381 29.187 66.010 269.833

June 08 – May 09 167.857 28.305 55.079 1089.667

Airport June 07 – May 08 n/a 29.302 58.140 975.833

June 08 – May 09 165.238 28.761 48.594 802.167

Urban June 07 – May 08 110.238 31.7668 55.807 524.167

June 08 – May 09 99.048 32.490 40.281 428.667


BF‐04

Figure 1 ACM sensor and structural steel coupon on an outdoor test rack

ACM sensor


BF‐04

(a) Urban sheltered environment (b) Airport and urban outdoor environment

Figure 2 Linear correlations between monthly corrosion rate and effective sensor output

were found at (a) urban sheltered condition (March 2008 – April 2009) and (b) airport and

urban outdoor condition (March 2008 – May 2009).


BF‐04

R2 = 0.9474R = 0.9733

0

50

100

150

200

250

0 50 100 150 200 250

Actual CR [g/ m2/ y]

Cal

cula

ted

CR

[g/ m

2 / y]

Figure 3 Calculated values compared to the actual values


BF‐06

Pickling Behavior of AISI 304 Stainless Steel in Sulfuric and Hydrochloric Acid Solutions

W. Homjaboka, S. Permpoonb, G. Lothongkuma

aDepartment of Metallurgical Engineering, Faculty of Engineering, Chulalongkorn University,

Patumwan, Bangkok 10330, Thailand

Email: [email protected], [email protected] bThainox Stainless Public Company Limited, 324 Moo 8, Highway no. 3191 Road, Tambol

Mabkha, Nikom Pattana, Rayong 21180, Thailand

Abstract

Oxide scales are formed on AISI 304 stainless steel surface during high temperature

processing as well as a Cr-depleted layer, which grows between the oxide scale and base

metal. Pickling is an important process that includes mechanical and chemical operations, used

to remove oxide scales, Cr-depleted layer and to recover the surface passivity. The multi-step

pickling is commonly used because of its higher efficiency than a single step pickling. In this

study, the multi-step pickling of AISI 304 stainless steel in HCl solution was investigated

instead of H2SO4 solution for the first step of pickling. HF+HNO3 mixed acid is traditionally

used in the second step. The pickling mechanism of HCl and H2SO4 was discussed based on

weight loss and the pickled surface qualities. It was found that pickling efficiency in the first

step directly affects the surface qualities of the final pickled sample. HCl solution showed

much lower pickling efficiency than H2SO4 solution. This resulted in high remaining oxide

scale and intergranular attack at the Cr-depleted layer, which cannot be completely removed in

the second pickling step. Increasing of HCl concentration and electrolytic current were not

enough to improve its pickling efficiency. The addition of small amount H2O2, which is a

strong oxidizing agent, significantly improves the pickling efficiency of HCl. A smooth

surface without any oxide scale and free of intergranular attack can be obtained.

Keywords: Pickling; Hydrochloric acid; Scale; AISI 304 Stainless steel


BF‐06

1. Introduction

Acid pickling is an important step for production of cold rolled stainless steel plate. It is aimed

to remove the oxide scale as well as a Cr-depleted layer growing between the oxide scale and

the base material. Oxide scale and Cr-depleted layer are formed during high temperature

processing. Removing oxide scale processes consist of mechanical descaling and pickling. In

mechanical descaling, scale breaker and shot blasting were used to break up the oxide scale.

This results in easily penetration of pickling solution into oxide scale and enhances the

pickling efficiency [1-5]. Multi-step pickling is used for pickling process because it has higher

efficiency and better surface quality than single step [6-7]. In the first step, electrolytic was

used for increase pickling efficiency [8]. In this step, the mechanism is that the solution

penetrates into metal Cr-depleted layer and the oxide scale is undercut and removed [6]. The

acid type and concentration has strong influence on surface finish quality. In the second step,

HNO3+HF has become widely accepted and used for removal remaining oxide scale and

passivation [8]. The sequence at which the pickling steps are used influences the surface finish

significantly. H2SO4 is a cheap acid and has a good pickling efficiency, which can be

improved by using with electrolytic, so that, H2SO4 with electrolytic is general used for the

first step. However, H2SO4 pickling causes black smut forming. Even though black smut can

be removed by HNO3+HF in the next step, but the surface finish has high roughness and

intergranular attack. In this study, the multi-step pickling behavior of AISI 304 austenitic

stainless steel in HCl solution was experimented for replacing H2SO4 solution in the first step

and the HNO3+HF mixed acid solution was used traditionally in the second step. HCl pickling

has a uniform dissolution behavior with no intergranular attack [8-10]. Results were discussed

based on weight loss and surface finish of the pickled samples.

2. Experimental

2.1. Material

AISI 304 austenitic stainless steel strips were hot-rolled downs to a thickness of 3 mm. The

chemical composition of this material is listed in Table 1. After mechanical descaling process,

test samples of 25x50x3 mm were cut. Then, only unexposed area was painted with EPIGEN

XD005 (acid-resistant at high temperature), and clean with acetone and ethanol. The test

samples were finally dried with air and kept in a desiccator before experiment.


BF‐06

Table.1 Chemical composition (wt.%) of AISI 304 stainless steel used in this study, analyzed

by OES Element Content Element Content

C 0.04 Si 0.342Cr 18.1 P 0.029Ni 8.03 S 0.001Mn 1.076 Fe Balanced

2.2. Pickling

To prepare the HCl, H2SO4, HF and HNO3 electrolytes, analytical grade was used. Purity

50%H2O2 was used in this study. During pickling, temperature was controlled constantly in a

water bath with constant stirring. After pickling, the samples were rinsed with tap water and

brush for removal any reaction products. The pickling conditions were acid concentration of

1.0, 2.0, 3.0, 4.0, 5.0 and 6.0 M at 60°C or 85 °C depending on the purposed tests.

2.3. Characterization

The surface finish was characterized with roughness profiler (Telescan 150) for surface

roughness. Optical microscope (OM) at 200X and scanning electron microscopy (SEM) at

3000X were used for remaining oxide level analysis. Fig.1 showed the evaluation of

remaining oxide on sample surface after the in-house standard.

Fig.1Remaining oxide evaluation after the in house standard on 6 areas observation on test

sample surface at 200X.


BF‐06


HCl solution was investigated instead of H2SO4 solution for the first step of pickling.

HF+HNO3 mixed acid solution was still traditionally used in the second step. The traditional

and studied conditions for this experiment were showed in Fig. 2. The total weight loss

resulting from those multi-step pickling conditions (Fig. 2) was shown in Fig. 3. The weight

loss of H2SO4 condition was high but some oxide scale remains on pickled surface in level 2

as shown in Fig. 4a. Pickling by H2SO4 solution with electrolytic followed by HNO3+HF

solution increased the weight loss and allowed achieving a surface finish free of any oxide

scale as shown in Fig. 3 and 4b.

Fig.2 Multi-step pickling of AISI 304 stainless steel between the traditional and studied

conditions.

Fig.3 Total weight loss of multi-step pickling of AISI 304 stainless steel in H2SO4 at 85°C or

HCl at 85°C followed by HNO3+HF at 45°C.


BF‐06

In case of HCl pickling instead of H2SO4 pickling, it showed that HCl had lower pickling

efficiency than H2SO4 (Fig. 3) and much oxide scale remained (Fig. 4c). Increasing HCl

concentration and HCl pickling with electrolytic did not result in increasing the pickling

efficiency to be higher than H2SO4 pickling efficiency. The surface finish of HCl pickling had

rougher surface and more intergranular attack than H2SO4 pickling as shown in Fig. 4d. The

result was not the same as report by L.-F. Li and J.-P. Celis [9], which said that uniform

dissolution and no intergranular attack were observed by HCl pickling.

Fig.4 SEM surface characterization of AISI 304 stainless steel after multi-step pickling

To understand mechanism of pickling by both HCl and H2SO4 in the first pickling step, which

has a significant effect on the final surface finish after HNO3+HF pickling, the step by step of

Roughness (Rq) = 3.29 μm Remaining oxide level 2




a b

c

4.0 M H2SO4; 85°C followed by HNO3+HF; 45 °C

4.0 M H2SO4 (Electrolytic); 85°C followed by HNO3+HF; 45 °C

4.0 M HCl; 85°C followed by HNO3+HF; 45 °C

4.0 M HCl (Electrolytic); 85°C followed by HNO3+HF; 45 °C d


BF‐06

weight loss was analyzed as shown in Fig. 5 and the surface was characterized by SEM as

shown in Fig. 6. HCl pickling had much lower weight loss than H2SO4 pickling and both

samples surface were covered with oxide scale (Figs. 6a and 6b). HCl pickling had smooth

surface compared with H2SO4 pickling. After the second pickling step with HNO3+HF, HCl

pickling had higher weight loss than H2SO4 pickling and the intergranular attack became more

pronounced on surface finish (Fig. 6d).

Fig.5 Step by step weight loss of AISI 304 stainless steel after pickling in 4.0 M H2SO4 at

85°C or 4.0 M HCl at 85°C followed by HNO3+HF at 45°C


BF‐06

Fig.6 SEM surface characterization of AISI 304 stainless steel after multi-step pickling with

conditions same as in Fig. 5

According to the previous results (Fig. 5 and 6) and discussion, the evolution of surface during

multi-step pickling in H2SO4 and HCl solutions followed by HNO3+HF can be described as in

Fig. 7a and 7b, respectively. The original metal surface consists of oxide scale, Cr-depleted

layer and base metal. On H2SO4 pickling in the first step, H2SO4 transports into oxide scale.

Then, the Cr-depleted layer is attacked or dissoluted. Finally, the oxide scale is removed by





a b

c 4.0 M H2SO4; 85°C followed by HNO3+HF; 45 °C


d

4.0 M H2SO4; 85°C 4.0 M HCL; 85°C


BF‐06

undercutting. Most oxide scale but only some Cr-depleted layer is removed. The surface is

rough because H2SO4 pickling behavior is non-uniform dissolution. The next step pickling by

the selective dissolution of HNO3+HF, intergranular attack appears. Remaining oxide scale

and Cr-depleted layer are almost removed. The final surface finish is completely free of oxide

scale. The evolution of surface finish after pickling in HCl followed by a pickling in

HNO3+HF is showed in Fig. 7b. The same mechanism as H2SO4 is obtained. However, HCl

has lower pickling efficiency than H2SO4. Most of all oxide scale and Cr-depleted layer still

remain. The observed surface is smooth because HCl pickling behavior is uniform dissolution.

By HNO3+HF pickling in the second step, intergranular attack appears because of a selective

dissolution on remaining Cr-depleted layer.

a. H2SO4; 85°C b. HCl; 85°C

Initial Surface

First step

Second step

Fig.7 The multi-step pickling mechanism models of intergranular attack.

According to the mechanism, the most important finding is that the surface finish obtained

from multi-step pickling is greatly affected by the pickling efficiency of the first step. Multi-

step pickling will successively allow achieving a smooth surface finish free of any oxide scale,

when a high enough pickling efficiency with uniform dissolution in the first step is available.

From the result, increasing of HCl concentration and electrolytic currents were not enough to

Oxide scale Cr-depleted layer

Intergranular attack

Base Metal Base Metal

Oxide scale



Intergranular attack


BF‐06

improve its pickling efficiency to be more than the H2SO4 efficiency. The addition of H2O2,

which is a strong oxidizing agent, possibly improved the pickling efficiency of HCl. The

temperature for this study must be fixed at 60°C because H2O2 decomposes at temperature

over than 60°C.

.

Fig.8 Step by step weight loss of AISI 304 stainless steel by pickling with HCl at 60°C or

HCl+H2O2 at 60°C or H2SO4 (Electrolytic) at 85°C followed by HNO3+HF at 45°C.

Addition of H2O2 to improve pickling efficiency of HCl in the first step resulted in increasing

weight loss and having an affect on the second step pickling by HNO3+HF by decreasing

weight loss, as shown in Fig. 8. It also reduced intergranular attack and delivered smooth

surface finish as shown in Fig. 9. Multi-step pickling was successive at 10g/L H2O2 added to

HCl solution. It allowed achieving a higher pickling efficiency than H2SO4 efficiency, smooth

surface finish free of oxide scale, and no intergranular attack


BF‐06

Fig.9 SEM surface characterization of AISI 304 stainless steel after pickling in H2SO4, HCl,

HCl + H2O2 solutions followed by HNO3+HF at 45°C

4. Conclusions

The multi-step pickling of AISI 304 stainless steel in HCl solution as the first step followed

by HNO3+HF as the second step was investigated. The mechanism models of pickling by HCl

or H2SO4 in the first step were proposed. The following conclusions can be drawn from this

study.

1. HCl solution has lower pickling efficiency than H2SO4 solution.





a b

c

4.0 M H2SO4 (Electrolic); 85°C followed by HNO3+HF; 45 °C


4.0 M HCl+10g/L H2O2; 60°C followed by HNO3+HF; 45 °C

4.0 M HCl+10g/L H2O2; 60°C followed by HNO3+HF; 45 °C

d


BF‐06

2. HCl solution can not completely remove Cr-depleted layer and oxide scale.

3. H2O2 addition can improve pickling efficiency of HCl solution. The addition of 10g/L

H2O2 is enough to deliver the smooth surface without any oxide scale and free of intergranular

attack after HNO3+HF pickling.

5. Acknowledgement

The authors would like to thank the Research and Development Center of Thainox

Stainless Public Company Limited for test samples, discussion and analysis equipment. The

thanks also go to the Thailand Research Fund (TRF) and the Office of Small and Medium

Enterprises Promotion (OSMEP) for the research fund.

6. References

[1] Lacombe, B. Baroux and G. Beranger, Stainless Steel, 1st edition., Les Editions de

Physique Les Ulis, France, 1993.

[2] Stainless Steel, ASM Specialty Handbook, 1996.

[3] Mars G. Fontana. Corrosion Engineering, 3rd edition, McGraw-Hill International Editions,

Materials Science and Engineering Series, Singapore, 1987.

[4] Denny A. Jones. Principles and Prevention of Corrosion, 2nd edition, Prentice Hall

International, Inc, Singapore, 1997.

[5] Suwaree Ratanamongkolthaworn, Effects of sulfuric acid concentration, temperature,

ferrous and ferric ion contents on pickling behavior of AISI 304 stainless steel, Master Thesis

in Metallurgical Engineering, Chulalongkorn University, 2007.

[6] L.-F. Li, P. Caenen, M. Daerden, D. Vaes, G. Meers, C. Dhondt, and J.-P. Celis,

Mechanism of Single and Multiple Step Pickling of 304 Stainless Steel in Acid Electrolytes,

Corrosion Science, Volume 47, 2005, pp. 1307- 1324.

[7] L.-F. Li and J.-P. Celis, Intergranular corrosion of 304 stainless steel pickled in acidic

electrolytes. Scripta Materialia, Volume 51, Issue 10, 2004, pp. 949-953.

[8] L.-F. Li, Pickling of Austenitic Stainless Steels. Internal review report Alz-Arcelor France,

2002.

[9] L.-F. Li and J.-P. Celis, Effect of hydrochloric acid on pickling of hot-rolled 304 stainless

steel in iron chloride-based electrolytes, Corrosion Science, Volume 50, 2008, pp. 804-810.


BF‐06

[10] L.-F. Li, Pickling and re-pickling of stainless steel with UGCO and UG3P+H2SO4

electrolytes. Internal review report Alz-Arcelor France, 2002.


CF‐03

The Effect of Welding Speed and Tool Pin Profile on Metallurgical and Mechanical Properties of Joining of Semi-Solid (SSM) Aluminium Alloy A356 by Friction

Stir Welding Process (FSW)

Thongchai Kruepue a and Prapas Muangjunbureeb

a, b Department of Mining and Materials Engineering, Faculty of Engineering, Prince of

Songkla University, Hatyai, Songkla, 90112 Thailand

Tel: 074 287323 Fax: 074 212897 E-mail : [email protected]

Abstract

The effect of joining parameters and tool pin profile on microstructure and mechanical

properties of semi-solid metal A356 joints produced by friction stir welding was

investigated. In this work, the joints were made by using a fixed rotating speed of 1,750

rpm with varying welding speed of 80, 120 and 160 mm/min. In addition, Two different

types of tool pins, cylindrical and square pin, were applied. The Scanning electron

microscope (SEM) reveals fine microstructure and uniform dispersion of Si (Silicon)

particles obtained from cylindrical pin than that of square pin. Transverse and longitudinal

tensile strengths obtained from cylindrical pin are greater than square pin. Furthermore, the

joint made from 1,750 rpm, 160 mm/min with cylindrical pin shows highest strength.

Key words : Semi-solid metal A356, Welding parameter, Thermo-mechanical affected

zone, Stir zone

1. Introduction

There are two types of semi-solid forming technology at the present. Rheo casting is

one of them. It involves the preparation of semi-solid metal (SSM) slurry from liquid alloys

and casting the slurry into a die for component manufacturing. In this work, semi solid metal

was obtained from a new Rheo casting technique called Gas Induced semi-solid (GISS) [1]. It

was clear that the joint between cast Al alloy has increasingly expanded in the usage of

casting component in automotive such as suspension, driveline and engine parts.

Conventional fusion welding of SSM aluminum die casting alloys is generally difficult due to

the formation of blowholes in weld. In addition, the microstructure is also altered. Therefore,

a new welding method is required to overcome theses problems. In recent year, friction stir


CF‐03

welding (FSW) was developed as a solid state joining process in which materials are joined

by the frictional heat as shown in Fig.1 (Thomas : TWI). This process is effective for the

welding of aluminum alloys. However, only a limited number of studies have been carried

out on SSM cast aluminum alloys. The aim of this work is to evaluate the effect of joining

parameters on the microstructure and mechanical properties of the welded SSM A356 alloys

in as cast condition.

Fig.1 Showing the friction stir welding

2. Experimental

The material used in this study was SSM (Semi-Solid Metal) A356 Al alloy 100 mm in

length, 50 mm in width and 4 mm in thickness. The chemical composition is listed in Table 1. In

this study, the two different tool pin profiles as shown in Fig. 2, a tool with a cylindrical pin and a

tool with a square pins, were applied. The length of the pin was 3.2 mm, and the diameter of the pin

was 5 mm. The shoulder of the tool was 20 mm/min. The friction stir welding (FSW) has many

welding parameters, such as tool rotating speed, welding speed and the angle of the tool. In this

work, only the welding speed was changed from 80, 120 and 160 mm/min. Other parameters were

fixed at 1,750 rpm of tool rotating speed and 30 of tool angle. The welding tool was rotated in the

clockwise direction and specimens, which were tightly fixed at the backing plate, were traveled.

The test pieces were cut in the cross-section direction, ground, polished and etched, with Keller’s

reagent. Some of the necessary photographs were taken by optical microscopy (OM), scanning


CF‐03

electron microscope (SEM) with energy dispersive x-ray analysis (EDX) examinations. The Vickers

hardness profile of the weld zone was measured on a cross-section and perpendicular to the welding

direction using a Vickers indenter with a 100 gf load for 10 s and 0.6 mm distance from welding

center. The tensile test was carried out at room temperature using an Instron-type testing machine

with cross-head speed of 1.67x10-2 mm s-1. The shapes and location of the specimens for test are

shown in Fig.3. Two kinds of tensile test specimens were prepared from the welded specimens. One

is transverse to the weld zone and the other is longitudinal to the weld zone.

Table 1 chemical composition of SSM A356 Al alloy (wt.%)

Metal Si Fe Cu Mn Mg Zn Ti Cr Ni Al

A356 7.74 0.57 0.05 0.06 0.32 0.01 0.05 0.02 0.01 Bal.

Cylindrical Square

Fig.2 showing two different tool pin profiles

Fig.3 Locations of the test specimens (A) Discard, (B) Microstructure,

(C) Tensile test and (D) Microstructure and Hardness test


CF‐03


3.1 Effect of the temperature of friction stir welding

The geometry of the tools pin affects the heat generation and the flow of the plastic

material. The temperature results of FSW joints are shown in Fig. 4 (a), . )b ( Zone 1 is the

heat generation increases from the downforce about 28 s., zone 2 is the heat generation

decreases from the stop downforce about 20 s. and zone 3 is the heat generation increases

from the welding speed. It demonstrate, that the welding temperatures during FSW decrease

in the high welding speed. On the top surface, the welding temperatures are almost the same

for two tool types. However, the temperature for three welding speeds of cylindrical pin is

higher than that of the square pin. Therefore, for the cylindrical pin, the frictional area

between the tool pin and the welding material is higher than that of the square pin [7].

(a) The temperature of cylindrical pin (b) The temperature of quare pin

Fig.4 showing the temperature generation results of the friction stir welding

3.2 Effect of the pin geometry on the weld surface appearance of the FSW

Fig.5 shows the surface appearance of the friction stir welded sample obtained from

cylindrical and square tool pins with various welding speeds at 1750 rpm. The top surface of

the joints indicate smooth surface particularly for the higher welding speeds. However, the

welding flash appears at the retreating side of the weld zone where the direction of the tool

rotation moves oppositely to the travel direction for every condition.


CF‐03

(a) 80 mm/min (d) 80 mm/min

(b) 120 mm/min (e) 120 mm/min

(c) 160 mm/min (f) 160 mm/min

Fig. 5 showing the photos of weld surface appearance

3.3 Effect of the tools pin geometry on the macro cross-section of the FSW

Fig. 6 illustrates the macro cross-section photos of the welded joints. Free-defect

joint can be obtained using two different tool pin profiles. There were no voids, cracks or

other weld defects, just as shown in Fig. 6 (a)-(f). An elliptical stir zone with an onion ring

structure was generated for the cylindrical pin. There was a macroscopically visible banded

structure for the square pin. However, shaped band structure appeared to dominate the

advancing side without appearing on the retreating side.

(a) 80 mm/min (d) 80 mm/min

(b) 120 mm/min (e) 120 mm/min

(c) 160 mm/min (f) 160 mm/min

10 mm 10 mm

10 mm 10 mm

10 mm 10 mm

1 mm 1 mm

1 mm 1 mm

1 mm 1 mm

R BM SZ

A TMAZ TMAZ

R BM

A TMAZ TMAZ SZ


CF‐03

Fig. 6 Macro cross-section of the welded joint, (SZ) stir zone,

(TMAZ) thermal-mechanical affected zone, (R) retreating, (A) advancing

3.4 Microstructure of joint

(a) Base metal of SSM A356 Al alloy

(b) R-TMAZ of cylindrical pin (c) A-TMAZ of cylindrical pin


CF‐03

(d) R-TMAZ of square pin (e) A-TMAZ of square pin

(f) SZ of cylindrical pin (g) SZ of square pin

Fig.7 Optical microstructure of the welded joint, (SZ) Stir zone,

(TMAZ) Thermal-mechanical affected zone, (R) Retreating, (A) Advancing

3.4.1 Optical microstructure of FSW

The spheroidal grain microstructure of the BM from Fig. 7 (a) is composed of primary

α phase (white region) and Al-Si eutectic structure (black region). The TMAZ of cylindical

pin and the square pin from Fig. 7 are formed besides the SZ, which are divided into the (b,

d) R-TMAZ and the (c, e) A-TMAZ is depending on the different microstructures at each

zone. The sharp transition between the BM and the SZ is observed in the retreating side. The

compression similar grain structures and a narrow range of deformed structures are observed

at the R-TMAZ. The slightly elongated grain structures or tention similar and a wider range

of deformed structures are observed at the A-TMAZ. The microstructure of the SZ is very

different from that of the BM. The spheroidal grain structure disappeared and finer Si

particles are dispersed over the whole stir zone. There are no voids, cracks or other welded

defects can be observed.

3.4.2 SEM microstructure of FSW

The microstructure of the BM from Fig. 8 (a) is composed of primary α phase and Si

particles structure (Elongated plate), are distributed partially in the primary α phase and


CF‐03

formed eutectic structure. The TMAZ of cylindical pin and the square pin from Fig. 8 are

formed besides the SZ, which are divided into the (b, d) R-TMAZ and the (c, e) A-TMAZ

which depend on the different microstructures. The smaller Si particles structures are

observed for the R-TMAZ and A-TMAZ of the cylindrical pin. However, the finer Si

particles are homogeneously dispersed in the SZ of the cylindrica pin than the square pin and

the plate-like particles disappear. The plate-like Si particles may be broken into slightly finer

particles by the stirring of the welding tool [6, 11]. This is due to the cylindrical pin generates

higher friction than that of the square pin. More heat input can improve the flow of the plastic

material. The material transports from the advanced side to the retreated side, and goes

around the pin, back to the advanced side.

(a) Base metal of SSM A356 Al alloy

(b) R-TMAZ of cylindrical pin (c) A-TMAZ of cylindrical pin

(d) R-TMAZ of square pin (e) A-TMAZ of square pin


CF‐03

(f) SZ of cylindrical pin (g) SZ of square pin

Fig.8 shows the SEM microstructure of the welded joint, (SZ) Stir zone,

(TMAZ) Thermal-mechanical affected zone, (R) Retreating, (A) Advancing

3.5 Effect of the pin geometry on the tensile strength

Table 2 Tensile test results from transverse direction

Tensile test (MPa)

Welding speed

(mm/min) Cylindrical

Failuer

locatio

n

Square

Failuer

locatio

n

80 176.22 Base 143.23 Weld

120 190.85 Base 171.73 Weld

160 193.89 Base 173.61 Base

Fig. 9 Showing the tensile strength of the transverse direction with various welding speeds.


CF‐03

3.5.1 The transverse tensile strength of the FSW

Table 2 and Fig. 9 Show the transverse tensile strength of the friction stir welding.

The tensile strength of the joints increases with the welding speed increases for two different

tools pin. The highest tensile strength of the joints was obtained from the cylindrical pin.

However, for each welding speed, the cylindrical pin indicates higher transverse tensile

strength. This is because the finer Si particles are homogeneously dispersed in the SZ of the

cylindrical pin than that of the square pin.

3.5.2 The longitudinal tensile strength of the FSW

Table 3 and Fig. 10 Show the longitudinal tensile strength of the friction stir welding.

The tensile strength of the joints increases with the welding speed increases for two different

tools pin. The highest tensile strength of the joints was obtained from the cylindrical pin.

However, for each welding speed, the cylindrical pin indicates higher transverse tensile

strength. This is because the finer Si particles are homogeneously dispersed in the SZ of the

cylindrical pin than that of the square pin. In comparison, the longitudinal tensile strength are

higher than transverse tensile strength in each parameter for both tool pins.

Table 3 Tensile test results from longitudinal direction

Tensile test (MPa) Welding speed

(mm/min) Cylindrical Square

80 172.57 170.67

120 190.79 179.83

160 222.23 193.67

Fig. 10 Showing the tensile strength of the longitudinal direction with various welding

speeds.


CF‐03

4. Conclusions

(1) The geometry of the tools pin affects the heat generation. It demonstrates that the

welding temperatures during FSW decrease in the high welding speed. The temperatures for

three welding speeds of cylindrical pin are higher than the square pin.

(2) Free-defect joint can be obtained using two different tool pin profiles. There were

no voids, cracks or other weld defects.

(3) The finer Si particles are homogeneously dispersed in the SZ of the cylindrical

pin than that of the square pin.

(4) The transverse and longitudinal tensile strengths of the joints increase with the

welding speed increases for two different tools pin. The highest tensile strength of the joints

was obtained from 160 mm/min welding speed of the cylindrical pin.

Acknowledgments

This work was financially supperted by TRF. In addition, the authors would like to

thank Department of Mining and Materials Engineering, Faculty of Engineering, Prince of

Songkla University, Hatyai, Songkla, Thailand.


CF‐03

References

[1] J. Wannasin “Development of a Novel Semi-Solid Metal Processing Technique for

Aluminium Casting Applications”

[2] R. Akhter. et al., (2006) “Effect of pre/post T6 heat treatment on the mechanical

properties of laser

welded SSM cast A356 aluminium alloy”, vols 116-117, PP.173- 176

[3] Yeong-Maw Hwang, (2007) “Experimental study on temperature distributions within the

workpiece

during friction stir welding of aluminum alloys”

[4] Yan-hua, et al., (2005) “The influence of pin geometry on bonding and mechanical

properties in friction stir weld 2014 Al alloy ”

[5] K. Kumar (2007) “The role of friction stir welding tool on material flow and weld

formation” A 485 (2008) 367–374

[6] W.B. Lee et al., (2003) “The improvement of mechanical properties of friction-stir-

welded A356 Al alloy” Material Science and Engineering A356 (2003) pp. 154-

159

[7] Hidetoshi Fujii et al., (2005) “Effect of tool shape on mechanical properties and

microstructure of friction stir welded aluminum alloys”

[8] Z.Y. Ma, et al., (2006) “Effect of friction stir processing on the microstructure of cast

A356 aluminum”

[9] Y.G. Kim. et al., (2006) “Effect of Welding parameter on Microstructure in stir zone of

FSW joints of Aluminum die casting alloy” Material Science and Engineering

A 415 (2006) 250-254

[10] K. Elangovan et al., ) 2007 ( “Influences of tool pin profile and welding speed on the

formation of friction stir processing zone in AA2219 aluminium alloy”

[11] M.L. Santella . et al., ) 2005 ( “Effects of friction stir processing on mechanical properties

of the cast aluminum alloys A319 and A356” Scripta Material 53 (2005) 201-

206


CF‐04

Rheocasting of aluminum alloys by the Gas Induced Semi-Solid (GISS) process

R. Burapa, S. Janudom, R. Canyook, J. Wannasin*

Department of Mining and Materials Engineering, Faculty of Engineering, Prince of

Songkla University, Hat Yai, Songkhla, 90112, Thailand

*Corresponding author, e-mail: [email protected]

Abstract

A new semi-solid metal processing technique has been developed to produce semi-solid

slurry with more effectiveness and efficiency at lower costs for the rheocasting process.

This technique is called the Gas Induced Semi-Solid (GISS) process. The GISS process

has been successfully used in laboratory settings to process several aluminum alloys,

including cast and wrought alloys. However, to develop this technique for commercial

applications in industrial settings, it is important to determine the optimized processing

conditions in order to control the resulting slurry temperature and microstructure of the

alloys. In this work, the effects of rheocasting temperatures and rheocasting times on the

resulting slurry temperature and microstructure of A356 aluminum alloy were investigated.

The results indicate the suitable conditions of the GISS process are longer rheocasting time

and lower rheocasting temperature. These conditions result in the formation of fine and

uniform globular structure of the primary α-Al phase.

Keywords: Aluminum alloy; Semi-solid metal; Gas Induced Semi-Solid process;

Rheocasting; Globular structure

1. Introduction

Aluminum alloys have been widely used in many applications such as electronic and

automotive components. These aluminum components are mainly produced by die casting.

In a conventional die casting process, molten metal is injected into a die cavity resulting in

turbulent flow and entrapment of air inside the casting parts. The consequences are oxide

films and porosity defects, which cause several quality issues and lower the mechanical

properties of the components.(1) One way to improve these problems is to apply the semi-

solid metal (SSM) forming technology. SSM forming is a forming process of metal in the


CF‐04

semi-solid state. In addition, the metal is modified during solidification to have non-

dendritic or globular grain structure. One technique for SSM forming is to create semi-

solid slurry directly from the melt and then to form the slurry into parts. This forming

technique is called rheocasting.(2) In recent years, there are various rheocasting processes

that have been developed to produce semi-solid slurries. These processes include the New

Rheocasting (NRCTM) process by UBE Machineries, Inc. (Japan),(3) the Advanced Semi-

Solid Casting Technology by Honda (Japan),(4) the Semi-Solid Rheocasting (SSRTM) by

IdraPrince Inc. (USA),(5) the Sub Liquidus Casting (SLCTM) by THT Presses, Inc. (USA),(6)

the Swirl Enthalpy Equilibration Device (SEED) by Alcan (Canada)(7), and the Rheo-

Diecasting (RDC) process by Brunel University (England).(8) Although several processes

are successfully used in the industry, they are still quite complex and have high capital

costs. As a result, a simple and efficient rheocasting process which offers lower costs for

producing semi-solid slurry is needed.(9) Recently, a new rheocasting process has been

developed at the Department of Mining and Materials Engineering, Prince of Songkla

University, Thailand. This process is called the Gas Induced Semi-Solid (GISS).(9) This

process uses the principle of applying a combination of localized heat extraction between a

cold rod and the molten metal with vigorous convection during the initial stages of

solidification to produce non-dendritic or globular grain structure.(2) The GISS process can

be applied with a large number of cast and wrought alloys such as A356, ADC10, ADC12,

2024, 6061 and 7075. To utilize the process effectively and efficiently, it is important to

understand the effects of the key processing parameters. In this study, the effects of the

rheocasting temperatures (the liquid metal temperatures before starting the introduction of

gas bubbles) and the rheocasting times (the time to introduce gas bubbles) on the resulting

slurry temperature and microstructure of A356 alloy were investigated.

2. Materials and Experimental Procedures

The aluminum alloy used in this study was a commercial cast aluminum alloy A356, which

has a wide solidification range with the solidus and liquidus temperature of 557°C and

613°C, respectively. The chemical composition of this alloy is listed in Table 1.

Table 1. Chemical composition (wt%) of the aluminum A356 alloy used in this study.

Si Fe Cu Mn Mg Zn Ti Al

6.9 0.42 0.05 0.04 0.42 0.01 0.10 Bal.


CF‐04

The GISS process is illustrated schematically in Figure 1. Figure 2 shows the prototype of

the GISS slurry maker used in this study. The machine consists of a graphite diffuser, a

thermocouple, a system to control the inert gas flow rate, a system to control the air

cooling and a central control unit.

Figure 1. Schematic of the GISS process.(2)

Figure 2. The prototype of the GISS slurry maker used in this study.

For all the experiments, the temperature of the graphite diffuser was kept at 40°C, the

diffuser surface area per the melt volume (S/V ratio) was set at 0.35, and the gas flow rate

was controlled at 4 liters/minute. In the experiments, the aluminum alloy was first melted

in a graphite crucible using an electric resistance furnace. The molten metal was fluxed at

710°C before the experiments. Then, about 500 grams of the melt was ladled out of the

crucible using a stainless steel cup coated with a ceramic coating. Subsequently, a

thermocouple was inserted near the middle of the melt to record the temperature data

during the experiments. When the melt cooled down to the set rheocasting temperature, a

porous graphite diffuser was immersed and fine nitrogen gas bubbles were introduced into

the melt. Then, the graphite diffuser was removed from the semi-solid slurry, and the

slurry was allowed to cool in air until the temperature reached 580 °C (about 45% solid

Inert gas bubbles

Inert gas

Flow meter

Graphite diffuser

Crucible

Molten metal

Thermocouple


CF‐04

Middle

fraction). The semi-solid metal was removed from the cup and quenched in water. Samples

were cut from the same position from the quenched semi-solid metals. Figure 3 shows the

schematic location of the samples. The samples were then prepared by a standard grinding

and polishing procedure, and were then etched with the Keller’s reagent. The

microstructure of the samples was observed and analyzed using an optical microscope. In

this study, the experimental conditions investigated include the rheocasting temperatures

and rheocasting times of 620, 635, and 650 °C, and 5, 12, 20 seconds, respectively.

Figure 3. Schematic of the samples and the position of the micrographs.

3. Results and Discussion

Representative cooling curves and the procedure to determine the slurry temperature after

the GISS process is shown in Figure 4. For example, the graphite diffuser was immersed at

the rheocasting temperature of 650 °C and with the introduction of nitrogen gas bubbles

for the rheocasting time of 20 s. When the bubbling was stopped and with a few seconds of

delay, the slurry temperature was determined from the curve. Following this analysis, the

slurry temperatures for different rheocasting temperatures and rheocasting times were

acquired and summarized in Figure 5.

570

580

590

600

610

620

630

640

650

660

0 20 40 60 80 100 120 140 160 180 200 220 240 260

Time (s)

Tem

pera

ture

(C)

Figure 4. Representative cooling curves of A356 alloy and the procedure to determine the

semi-solid slurry temperature.

Rheocasting time = 20 seconds

Start immersion of graphite diffuser at 650°C

Check temperature of semi-solid slurry

Stop immersion of graphite diffuser.

The semi-solid slurry was quenched in water at 580°C


CF‐04

602

604

606

608

610

612

614

616

618

0 5 10 15 20 25

Rheocasting time (s)

Slu

rry

tem

pera

ture

(C)

Rheocasting temp. at 620 CRheocasting temp. at 635 CRheocasting temp. at 650 C

Figure 5. The effects of rheocasting temperatures and rheocasting times on the slurry

temperature of A356 aluminum alloy.

Then, the results from the various slurry temperatures were converted to solid fraction (fs)

data. The Scheil’s equation was used to estimate the solid fraction.(2) For A356 alloy, the

calculation assumed a binary alloy, linear liquidus and solidus lines and the partition

coefficient (k) equals 0.13. Figure 6 shows the solid fraction of A356 slurry under a

combination of rheocasting temperatures and rheocasting times.

02468

101214161820

0 5 10 15 20 25

Rheocasting time (s)

Solid

frac

tion

(%)

Rheocasting temp. at 620 CRheocasting temp. at 635 CRheocasting temp. at 650 C

Figure 6. The effects of rheocasting temperatures and rheocasting times on the solid

fraction of aluminum A356 alloy.

The GISS process utilizes the cold graphite diffuser and the introduction of fine nitrogen

gas bubbles to decrease the temperature of the melt below its liquidus temperature. The

Liquidus temperature = 613°C


CF‐04

rheocasting temperature and rheocasting time affect the slurry temperature and,

consequently, the solid fraction with the relationships shown in Figures 5 and 6,

respectively. To create more solid phase in the melt, the rheocasting temperature is

lowered and the rheocasting time is increased. The obtained data give the important

processing information about the required rheocasting times to achieve a certain amount of

solid fraction in the melt with different starting rheocasting temperatures. For example, to

achieve about 5% solid fraction in the melt, the rheocasting times should be about 7, 9, and

15 seconds for the rheocasting temperatures of 620, 635, and 650 °C, respectively. The

microstructure of aluminum A356 alloy solidified without the application of the GISS

process showing coarse dendritic structure is given in Figure 7. The white phase in Figure

7 is primary α-Al phase and the dark continuous matrix is the quenched eutectic phase. The

typical microstructures of A356 alloy produced by the GISS process under the rheocasting

time of 5 seconds for various rheocasting temperatures are shown in Figure 8. The

experimental results show that the primary α-Al phase varied with the solidification

conditions from coarse dendritic, to rosette-like, and to globular structure.

Figure 7. Microstructure of aluminum A356 alloy solidified under normal conditions

showing coarse dendritic microstructure.

The experimental results for the case of rheocasting time of 5 seconds show that with the

rheocasting temperature of 650 °C, the primary α-Al phase has rosette-like morphology, as

shown in Figure 8(a). With the rheocasting temperatures of 635 °C and 620 °C, as shown

in Figures 8(b) and 8(c), respectively, the primary α-Al phase consists of mostly globular

and some rosette-like grains. Figure 9 shows the experimental results under different

rheocasting times at the same rheocasting temperature of 620 °C. With the rheocasting

time of 12 seconds, most of the primary α-Al phase appears globular with some rosette-

like structure, Figure 9(a). When the rheocasting time increases to 20 s, the morphology of


CF‐04

primary α-Al phase is mainly globular with a uniform distribution in the structure, Figure

9(b).

Figure 8. Microstructures of A356 alloy produced by the GISS process under the

rheocasting time of 5 seconds for different rheocasting temperatures: (a) 650 °C;

(b) 635 °C; and (c) 620 °C.

Figure 9. Microstructures of A356 alloy produced by the GISS process under the

rheocasting temperature of 620 °C for different rheocasting times: (a) 12 s and (b) 20 s.

These results suggest that the primary α-Al phase tends to be fine globular structure with a

uniform distribution in the eutectic phase when the rheocasting temperature is decreased

and rheocasting time is increased. The results may be explained by the dendrite

fragmentation mechanism.(10) The non-dendritic structure is developed from a large

number of initial dendrite fragments going through the ripening mechanism which results

in globular grain structure. For the GISS process, the results obtained in this study suggest

that a combination of localized heat extraction with the introduction of fine nitrogen gas

bubbles through the graphite diffuser to create the vigorous convection during immersion

of the graphite diffuser in the molten metal causes dendrite arms to break off from the

mother dendrites. This process helps to generate secondary nuclei particles, which then can

grow to form a non-dendritic or globular structure within a few seconds. With the lower

(a) (b) (c)

(a) (b)


CF‐04

rheocasting temperature, the mother dendrites will be finer making it easier and faster for

the dendrite arms to be detached. In addition, with longer rheocasting times, more dendrite

arms will be detached and the longer ripening time will lead to more globular structure.

4. Conclusions

1. This study gives important information for processing about the required

rheocasting times to achieve a certain amount of solid fraction in the melt with different

starting rheocasting temperatures.

2. Fine and uniform globular structures for aluminum A356 alloy were obtained

when the rheocasting temperature was low and the rheocasting time was long.

3. The GISS process can be used to produce semi-solid slurries effectively and

efficiently when the proper processing conditions are selected.

5. Acknowledgements

The authors would like to thank the Reverse Brain Drain Project (RBD), the

National Science and Technology Development Agency (NSTDA) for funding this

research project. In addition, we would like to thank Mr. Thiensak Chucheep and

Innovative Metal Technology (IMT) Team for helping with the experiments.

6. References

1. de Figueredo A, Ed. 2001. Science and Technology of Semi-Solid Metal Processing.

The North American Die Casting Association, U.S.A.

2. Wannasin, J. and Thanabumrungkul, S. 2008. Development of a semi-solid metal

processing technique for aluminium casting applications. Songklanakarin J. Sci.

Technol., 30(2): 215-220.

3. Kaufmann, H., Wabusseg, H. and Uggowitzer, P.J. 2000. Metallurgical and Processing

Aspects of the NRC Semi-Solid Casting Technology. Aluminum, 76(1-2): 70-75.

4. Kuroki, K., Suenaga T., Tanikawa, H., Masaki, T., Suzuki, A., Umemoto, T. and

Yamazaki, M. 2004. Establishment of a Manufacturing Technology for the High

Strength Aluminum Cylinder Block in Diesel Engines Applying a Rheocasting

Process. Proceedings of the 8th International Conference on Semi-Solid Processing

of Alloys and Composites, Limassol, Cyprus.


CF‐04

5. Yurko, J., Martinez, A. and Flemings, M. 2003. The Use of Semi-Solid Rheocasting

(SSRTM) for Automotive Casting. SAE 2003 World Congress & Exhibition,

Detroit, Michigan, U.S.A.

6. Jorstad, J., Thieman, M. and Kamm, R. 2004. Fundamental Requirements for Slurry

Generation in the Sub Liquidus Casting Process and the Economics of SLCTM

Processing. Proceedings of the 8th International Conference on Semi-Solid

Processing of Alloys and Composites, Limassol, Cyprus.

7. Douter, D., Hay, G. and Wales, P. 2004. SEED: A New Process for Semi-Solid

Forming. Canadian Metallurgical Quarterly, 43(2): 265-272.

8. Fan, Z., Fang, X. and Ji, S. 2005. Microstructure and Mechanical Properties of Rheo-

Diecast (RDC) Aluminium Alloys. Mater. Sci. Eng. A, 412(1-2): 298-306.

9. Wannasin, J., Junudom, S., Rattanochaikul, T. and Flemings, M.C. 2008. Development

of the Gas Induced Semi-Solid Metal Process for Aluminum Die Casting

Application, Solid State Phenomena Vols. 141-143: 97-102.

10. Wannasin, J., Martinez, R.A. and Flemings, M.C. 2006. Grain refinement of an

aluminum alloy by introducing gas bubbles during solidification. Scripta

Materialia. 55: 115-118.


CF‐09

Effects of replacing binder with powder space holder on properties of metal injection moulded foam

U. Koikula , A. Manonukulb,*, S. Suranuntchaia

aKing Mongkut’s University of Technology Thonburi, Pracha-U-Thit Rd., Thungkru,

Bangkok, 10140. bNational Metal and Materials Technology Center, 114 Thailand Science Park, Paholyothin

Rd., Klong Luang, Pathumthani, 12120.

*E-mail: [email protected] (Corresponding author)

Abstract

Metal foam can be produced using metal injection moulding with powder space holder. In

this study, the effects of volume fraction of spacer holder on the foam properties were

studied. Spherical poly (methyl methacrylate) (PMMA) particles were used as a space

holder material. The 30% volume fraction of stainless steel 316L powder was mixed with

varied volume fractions of binder and PMMA. There were five volume fractions of

PMMA, which are 30% to 50% with an increment of 5%. The results shown that the

volume fraction of PMMA affected the properties. As the volume fraction of PMMA

increased, the number of pore increased but the sintered density and the mechanical

properties decreased.

Keywords: metal foam; metal injection moulding; powder space holder

1. Introduction

The interest in metal foam has significantly increased due to their extended applications,

for example, automotive parts, filters, cushions, insulators and biomedical implants

(Degischer and Kriszt, 2002). Currently, there are different manufacturing methods for

metal foams. The conventional process is the gas injection method, where gas bubbles are

injected into a liquid metal. The liquid metal is transferred using a conveyer belt to

solidify. This process is very effective in continuously producing large size foams, but it is

difficult to control the process to obtain a uniform structure. Another process is the

deposition method, which starts from the ionic state of metal and deposits a polymeric

foam preform with open cells. Similar to the deposition method, the investment casting


CF‐09

method also uses a polymeric foam perform. In the investment casting method, the

polymeric foam preform is dipped into graphite slurry or coated with a thin layer by metal

vaporisation. Subsequently, the polymeric foam preform is removed by thermal treatment

(Ashby et al., 2000). These two processes can produce a complex shape part, that can be

fabricated by preforming the polymeric foam but both methods are expensive (Gibson and

Ashby, 1988).

Metal injection moulding (MIM) is a manufacturing process combining the

traditional powder metallurgy process and plastic injection moulding (German, 1997). It is

capable of producing small parts with complex shape in a mass production scale. Hence,

MIM using powder space holder (MIM-PSH) has been developed for producing complex

metal foam part (Gülsoy and German, 2008). The MIM with powder space holder for

producing foam is similar to conventional MIM as shown in Fig. 1. There are four main

steps, which are mixing, injection, debinding and sintering. In the first step mixing,

metallic powder, binder and powder space holder are homogeneously blended together.

The mixture is then granulated and injected to obtain “green” parts. The green part is

heated to remove binder and powder space holder. This step is the debinding step and

“brown” parts with the structure of foam are obtained after debinding. Brown parts are

then sintered at a high temperature to obtain a metal foam. It is noted that MIM-PSH can

produce both open-cell and close-cell foams. Poly (methyl methacrylate) (PMMA) is a

common powder space holder. PMMA can be easily decomposed in the debinding stage

and the metal foam with uniform foam structure can be manufactured by MIM-PSH

(Gülsoy and German, 2008). In addition, MIM-PSH can be cost-effective for microsized,

highly complex porous shape with high dimensional accuracy (Williams, 2007). As a

result, PMMA was used in this work as the powder space holder. Nishiyabu et al. (2008)

studied the propertied of 316L foam produced by MIM with 30 and 60% volume fraction,

and 10 and 40 µm average size of PMMA. The effect of powder space holder shape was

also studied (Jiang et al., 2005). Spherical and strip carbaminde particles were used as the

powder space holder. Previous works only studied two volume fractions of powder spacer.

Therefore, this work systematically investigated the effect of volume fraction of PMMA

(powder space holder) on the properties of metal foam produced by MIM-PSH.

2. Experiment procedures


CF‐09

In this work, the water–atomised stainless steel 316L powder (PF-20F) provided by Atmix

Co. Ltd., Japan, was used. The powder has the average size of 10.5 µm. PMMA was

supplied by Sunjin Chemical Co. Ltd., Korea. PMMA was used as the powder space holder

in this work and had a particle size of 84.7 µm. Figure 2 shows the scanning

Figure 1: Schematic representation of metal injection moulding using powder space holder

technique.

electron microscopy (SEM) images of the 316L powder and PMMA particle. The 316L

powder is rounded, while the PMMA particle is spherical. The binder in this experiment

comprised of three components: paraffin wax as a plasticiser, high density polyethylene as a

backbone polymer and stearic as a acid surfactant. The binder reduces the viscosity of the

feedstock and facilitates injection moulding. The backbone polymer provides the essential

strength of the green parts. The surfactant strengthens the adhesion between binder and

powder and weakens the agglomeration of the powder (Huang and Hsu, 2009).

(a) (b)

Figure 2: SEM micrographs of (a) 316L powder and (b) PMMA particle.


CF‐09

Stainless steel 316L powder, PMMA and a polyacetal-based binder were mixed

together in five batches with different volume fractions as shown in Table 1. The solid

loading of metal powder was kept constant at 30% volume fraction. The volume fraction of

PMMA and binder were correspondingly varied with the constant combined volume fraction

of 70%. The volume fraction of PMMA was increased from 30% to 50% in an increment of

5%, while the volume fraction of binder was decreased from 40% to 20% in an increment of

5%. Thus, the experiment was designed to replace binder with more PMMA, while the solid

loading of metal powder was constant.

This mixture was injected into tensile-test-specimen shape. Green parts were

thermally debound at 450 °C for 1 hour in air and sintered at 1100 °C for 2 hours in an argon

atmosphere. The densities of the green and sintered parts were measured. The sample were

cut, mounted, grinded with silica papers and polished with diamond solution for the

observation of microstructures using the optical microscopy. Hardness in HR15W scale and

tensile tests were tested and reported.

Table 1. Fraction by volume of each component: PMMA, binder and metal powder.


Processing parameters

Most processing parameters for mixing and injection moulding were kept constant apart from

the mixing and injection temperatures, which were varied with the volume fraction of

PMMA. Figure 3 shows the variation of mixing and injection temperatures. The mixing and

Volume fraction of

PMMA

(% vol)

Volume fraction of

binder

(% vol)

Volume fraction of

metal powder

(% vol)

30 40 30

35 35 30

40 30 30

45 25 30

50 20 30


CF‐09

injection temperatures increased as the volume fraction of PMMA increased and the volume

fraction of binder decreased. It is noted that the result for the 50% volume fraction of PMMA

cannot be shown because it was not possible to inject the feedstock with 50% PMMA. From

Table 1, a feedstock with 50% volume fraction of PMMA contained 30% by volume of metal

powder and 20% by volume of binder. This means that during injection, there was only 20%

liquid phase during injection (binder) and 80% solid phase during injection (PMMA and

metal powder). This resulted in the high viscosity of feedstock and it was not possible to

inject this feedstock with 50% volume fraction of PMMA regardless of the injection

condition. As the volume fraction of PMMA increased, the viscosity of feedstock was higher

and it was more difficult for the feedstock to flow. As a result, the higher mixing and

injection temperatures were required to increase flow ability (Supati et al. 2000).

Figure 3: Mixing and injection temperatures as a function of the volume fraction of PMMA.

Density and microstructure

Figure 4 shows the green and sintered densities as a function of volume fraction of PMMA.

The sintered density is higher than the green density for all percentages of PMMA showing

consolidation during sintering. The sample had similar green density because the volume

fraction of metal powder was kept constant. The average green density of all specimens was

3.05 g.cm-3. The sintered density decreased with increasing volume fraction of PMMA. The

sintered density of 3.58 g.cm-3 was observed in the 30% PMMA volume fraction, which is

the highest sintered density. On the other hand, the sintered specimen with 45% PMMA

volume fraction has the lowest sintered density of 3.49 g.cm-3.


CF‐09

Figure 4: Green and sintered densities as a function of the volume fraction of PMMA.

The microstructures of the sintered 316L stainless steel specimens with four different

volume fractions of PMMA are shown in Fig. 5. The number of pores depended on the

PMMA contents. For all microstructures, the pores were introduced by the burnout of

PMMA. All such pores retained the spherical shape of powder space holder and distributed

homogeneously in the 316L stainless steel matrix. The number of pores increased with the

increased addition of PMMA. The microstructure of 45% PMMA volume fraction exhibited a

large number of pores distributed thoroughly inside the specimen as shown in Fig. 5 (d).

There were more pores distributed in the microstructure of 45% PMMA volume fraction than

the other volume fractions.

Mechanical properties

Sintered metal foam was subjected to hardness and tensile tests. The hardness of metal foam

specimens was tested using Rockwell W (HR15W). The hardness of sintered parts varied

with the volume fraction of PMMA as shown in Fig. 6. The result shows that as the volume

fraction of PMMA increased from 30% to 45%, the hardness decreased from 32-21 HR15W.

The error was also displayed. The error increased as the volume fraction of PMMA increased.

It is noticed that the 45% volume fraction of PMMA had the largest error and this volume

fraction had a large number of pores. The microstructures of the specimens showed the

number of pores increased with increasing volume fraction of PMMA. Therefore, the

hardness values decreased as the number of pores in the specimens increased.


CF‐09

(a) 30% vol PMMA (b) 35% vol PMMA

(c) 40% vol PMMA (d) 45% vol PMMA

Figure 5: Optical microstructure of sintered parts cross section as a function of the volume

fraction of PMMA.

Figure 6: Hardness of sintered parts as a function of the volume fraction of PMMA


CF‐09

Figure 7 shows the tensile strength and the elongation of sintered parts as a function

of the volume fraction of PMMA. The tensile strength and elongation were dependent on the

volume fraction of PMMA. As the volume fraction of PMMA increased, both tensile strength

and elongation decreased. The 30% volume fraction of PMMA had the highest tensile

strength of 125 MPa with 14% of elongation. The lowest values were obtained for the 45%

volume fraction of the PMMA, which had the lowest tensile strength of 97 MPa and the

elongation of 10%. The results are similar to the hardness results. The mechanical properties

decreased with increased in porosity or the volume fraction of PMMA increased.

Figure 7: Tensile strength and elongation of sintered parts as a function of the volume

fraction of PMMA

4. Conclusions

Stainless steel 316L foams can be produced by applying a powder space holder method to a

metal injection moulding process. The material used for space holding is a spherical PMMA

particle. Five different volume fractions of PMMA were varied to replace the binder and the

volume fraction of metal powder was kept constant. The experimental results show that the

sintered density was higher than the corresponding green density. The sintered density

decreased when the volume fraction of PMMA increased. The microstructure showed that the

number of pore depended on the fraction of PMMA. The spherical shape pores were

homogeneously dispersed in the 316L stainless steel matrix. The number of pores increased

with increasing volume fraction of PMMA. Tensile strength, elongation and hardness

decreased as the volume fraction of PMMA increased.


CF‐09

5. References

Ashby, M. F., Evans, A., Fleck, N.A. Gibson, L. J., Hutchinson, J. W. and Wadley, H. N. G.

2000.Metal Foam: A Design Guide. Butterworth-Heinemann. Boston.

Degischer, H. P. and Kriszt, B. 2002. Handbook of Cellular Metals. Wiley. Weinheim.

German

R.M. and Bose, A. 1997. Injection Molding of Metals and Ceramics. MPIF. New Jersy.

Gibson, L. J. and Ashby, M. F. 1988. Cellular Solids Structure & Properties. PERGAMON

PRESS. Oxford.

Gülsoy, H.O. and German, R.M. 2008. Production of micro-porous austenitic stainless steel

by powder injection molding. Scripta Materialia. 58: 295-298.

Huang, M. Y. and Hsu, H. C. 2009. Effect of backbone on properties of 316L stainless steel

MIM compact. Sci. Forum. 209: 981-984.

Nishiyabu, K., Matsuzaki, S. and Tanaka, S. 2007. Net-Shape Manufacturing of Micro

Porous Metal Components by Powder Injection Molding. Materials Science Forum.

534-536: 981- 984.

Supati, R., Loh N.H., Khor, K. A. and Tor, S. B. 2000. Mixing and Characterization of

Feedstock for Powder Injection Molding. Materials Letters. 46: 109-114.

Williams, B., 2007. Powder injection moulding in the medical and dental sectors. Powder

Injection Moulding International. 1: 12-19.


CF‐10

Effect of temperature and pressure on the densification of titanium silicide compound

P. Thapnuya, S. Larpkiattaworna, T. Luangvaranuntb, J. Ikeuchia

aThailand Institute of Scienctific and Technological Research 35 Moo3 Technothani

klong5 klongluang pathumthani bChulalongkorn University Rama4 Road Pathumwan, Bangkok

Tel: 02-5779274 Fax: 02 -5774160-1 E-mail: [email protected]

Abstract

Titanium silicide compound was synthesized from the mixture of titanium and silicon

powder with atomic ratios of 60:40 and 70:30. The powders were mixed by ball milling

and pressed by different methods: by using uniaxial pressing at 64 MPa, cold isostatic

pressing (CIP) at 200 MPa, hot forging (HF) at 648 MPa, and hot pressing (HP) at 24 MPa.

The samples were then sintered at 1300 oC or 1600 oC for 2 h. soaking time in argon

atmosphere. The sintered samples were characterized for phases constitution using X-ray

diffraction (XRD). Ti5Si3 was observed as main compound from both 60:40 and 70:30

mixtures. Archimedes’ method and scanning electron microscope (SEM) was used to

measured density and investigate microstructure of sintered samples. It was found that the

sample prepared from the 70:30 mixture has higher density than that sample of 60:40

mixture for all applied pressures. The density of samples prepared from the mixture of

70:30 and 60:40 sintered at 1300oC are in the range of 53-60% and 42-55%, respectively.

It was found that densities of all samples sintered at 1300 oC were not much different when

applied higher forming pressure by using CIP and HF. This means that pores are created

during sintering. However the microstructure of samples formed by CIP and HF showed

some big pore inside the sample body while sample formed at lower pressure by uniaxial

press showed a more uniform pore size. By increasing the sintering temperature to 1600

oC, the density of uniaxial pressed sample was increased to 85% and pore size get smaller

than the one sintered at 1300 oC. On the other hand, by applying a lower pressure during

sintering the sample at 1600 oC by hot pressing, this can produce high density sample of

99% with a few amount of small closed pore.

Keywords: Titanium silicide; Cold Isostatic Press; Hot Forge; Hot Press


CF‐10

1. Introduction

Titanium silicide compound such as TiSi3, TiSi2, TiSi, Ti5Si4 and Ti5Si3 can be prepared

from various ratio of titanium and silicon metal. Among these silicide compounds, Ti5Si3 is

known as an intermetallic compound which suitable for high temperature applications due

to the properties of a high melting point (2130° C), moderate density (4.32 g/cm3), high

temperature oxidation resistance, high hardness (11.3 GPa) and high young modulus (225

GPa) [1-2]. Titanium silicide can be prepared by a variety of powder techniques such as

hot pressing, hot isostatic pressing, reactive sintering, mechanical alloying, and thermal or

plasma spraying [3-6]. Due to the limited fracture toughness of Ti5Si3 at room temperature,

most researchers have paid attention to produce multiphase in Ti5Si3 compound by addition

of Al, C, Ni, or Nb [4, 6, 7]. However, Research works on densification of Ti5Si3

dependence on pressure and temperature are limited. In this paper, the variety of pressure

and temperature were applied to prepare Ti5Si3, and then their density and microstructure

were observed.

2. Experiment procedures

Ti powder (99.7% purity, average particle size < 45μm) and Si powder (99.7% purity,

average particle size < 45μm) were mixed in atomic ratio of 60:40 and 70:30 for 20 hr in

Ar gas atmosphere. The mixed powder was compacted into specimens with 2 cm

diameter and 0.5 cm thickness using different pressure of 64 MPa (uniaxial pressing),

200 MPa (cold isostatic pressing), 648 MPa (hot forging) and 24 MPa (hot pressing).

The uniaxial pressed specimens were sintered at temperature varied from 1100 to

1600°C. While other specimens were sinter at 1300 and 1600 °C in argon atmostphere.

The heating rate and soaking time were 15 °C/min and 2 hr, respectively. The sinter

specimens were measured for density by Archemidis method. X-ray diffractometer

(XRD) and Scanning electron microscope (SEM) were used to determine the phase

constitution and microstructure of the sintered specimens respectively.


The sintered Ti: Si specimens of 60:40 and 70:30 were characterized for phase constitution

by XRD as patterns shown in Figure.1 and 2 respectively. The results show that single

phase of Ti5Si3 was formed in the 60:40 specimens sintered at temperature 1100-1500 °C

and at high temperature of 1600 °C, TiC was observed together with Ti5Si3 phase. On the


CF‐10

other hand, TiC can be formed together with Ti5Si3 in 70:30 specimens sintered at

temperature range of 1100-1500 °C and then this TiC phase will transform to Ti3SiC2 at

1600 °C. The TiC and Ti3SiC2 phases in specimens could be from the diffusion of carbon

inside the furnace into specimens during sintering process. Higher content of Ti in the

specimen causes easier forming of TiC at low temperature and TiC will react with some

Ti5Si3 to form Ti3SiC2 at high temperature.

25 30 35 40 45 50 55 60 65 70 75 80

Diffraction angle ( 2θ )

Inte

nsity

Ti5Si3 TiC

11000C

12000C

13000C

15000C

16000C

Figure 1: XRD patterns of the Ti:Si mixture of 60:40 sintering at different temperatures

25 30 35 40 45 50 55 60 65 70 75 80

Diffraction angle ( 2θ )

Inte

nsity

Ti5Si3Ti3SiC2TiC

11000C

12000C

13000C

15000C

16000C

Figure 2: XRD patterns of the Ti:Si mixture of 70:30 sintering at different

temperatures


CF‐10

Table 1 shows density of specimens formed at different pressures after sintering at

1300°C. It was found that increasing the forming pressure could slightly increase the

density of specimen after sintering. The specimen with high Ti content yield higher

density than that of lower Ti content which agree with the SEM micrographs in Figure

3 Moreover, Figure 3 shows that the specimens formed at higher pressure have bigger

pore size than those formed at lower pressure. These phenomena can be explained that

the Si vapor is trapped inside specimens which have been compacted at high pressure

prior to sintering. On the other hand, Si vapor generated from specimens with lower

forming pressure can easily move out during sintering before densification, and that

results in smaller pores.

Table1 Density of Ti:Si mixture of 70:30 and 60:40 formed at various pressure and sintering

at 1300°C.

Bulk density Apparent density

(Bulk

density/Apparent

density) x100 Ti:Si

60:40 70:30 60:40 70:30 60:40 70:30

Uniaxial press

(63.69 MPa) 1.89 2.49 4.37 4.35 43.24 57.24

CIP

( 200 MPa ) 1.79 2.33 4.27 4.36 41.92 53.44

HF ( 648MPa ) 2.38 2.63 4.31 4.40 55.22 59.77

Table2 Density of TiSi mixture of 70:30 sintered at 1600°C

Ti:Si Bulk density Apparent

density

(Bulkdensity/

Apparent

density) x100

%Porosity by

Archimedis

method

Uniaxial press

(63.69 MPa) 3.70 4.31 85.84 14.16

Hot press

(24.24 MPa ) 4.34 4.40 98.63 1. 37


CF‐10

V

Figure.3 SEM micrographs of TiSi mixture of 60:40 and 70:30 formed at pressure (a) 60:40

(uniaxial Press 64 MPa), (b) 70:30 (uniaxial Press 64 MPa), (c) 60:40 (CIP 200 MPa), (d)

70:30 (CIP 200 MPa), (e) 60:40 (HF 648 MPa), (f) 70:30 (HF 648 MPa)

Figure 4: SEM micrographs of TiSi mixture of 70:30 sintered at 1600°C formed by (a)

uniaxial press at 64 MPa and (b) hot press at 24 MPa

According to Table 1 and Table 2, the density of specimen formed at 64 MPa increases

significantly from 57% to 86%, when the sintering temperature is raised up from 1300 °C

to 1600 °C. Furthermore, by simultaneously applying low pressure (24 MPa) and heating

(1600 °C) the density of specimens can be raise up to 99% which is shown in Figure 4.

(a) (b) (c)

(d) (e) (f)

(a) (b)


CF‐10

This means that temperature is more effective on increase in density than pressure;

however applying pressure during sintering is the most effective to get high densification.

4. Conclusions

Ti5Si3 can be synthesized from Ti and Si powders (70:30 and 60:40) sintered at 1100-1600

°C. During sintering Si vapor can generated and form the pores inside specimen, and

retards the densification. Increasing forming pressure is insignificant in densification

during sintering. On the other hand, applying lower pressure during sintering can

remarkably enhance the densification of Ti5Si3 specimens.

5. Acknowledgements

The authors would like to thank Thailand Reasearch Fund (TRF), Nanoshield Ltd. and

Japan International Cooperation Agency (JICA) for the support of this work

6. References

[1] D.P. Riley,C.P. Oliver, E.H. Kisi. In-situ neutron diffraction of titanium silicide, Ti5Si3,

during self-propagating high-temperature synthesis (SHS). Intermetallics 14 (2006) 33-38

[2] Massalski TB .et.al. Binary alloy phase diagrams.2nd ed.Materials Park, OH: ASM

Int.:1990

[3] N.S. Stoloff. Materials Science and Engineering: A 261 (1999), 169-180

[4] A. Calka, A.P Radlinski, R.A.Shanks, and A.P. Pogany.Formation of titanium silicides

by mechanical alloying .10 (1991), 734-737

[5] R.Mitra, Met.Meter.Trans.Microstructure and mecchanical behavior of reaction hot –

pressed titanium silicide and titanium silicide based alloys and composites.A 29A(1998),

1629-1641.

[6] C.L.Yeh, W.H. Chen, and CC. Hsu. Formation of titamium silicides Ti5Si3 and TiSi2 by

self-propagating combustion synthesis. 432 (2007), 90-95

[7] L.Zhang and J.Wu. Ti5Si3 and Ti5Si3–base alloys: Alloying behavior, microstructure

and mechanical property evaluation, Acta matter 46(10) (1998), 3535-3546


CF‐12

Effect of aluminium on sintered properties of Cu-10wt%Sn bearing

V.Chobaomsup, T.Luangvaranunt

Department of Metallurgical Engineering, Faculty of Engineering, Chulalongkorn

University, Phyathai Rd., Bangkok, Thailand 10330

Tel: +66 2218 6947, 089-128-1572 Fax: +66 2218 6942 Email:[email protected]

ABSTRACT

Self-Lubricating bearings are one of the oldest industrial applications of porous powder

metallurgy part, dating back to the mid-1920. They remain the highest part produced by the

P/M industry. The objective of this research was to study effect of sintering time, sintering

temperature and ratio of adding aluminium on sintered properties of Cu-10wt%Sn bearing

that produced from powder metallurgy processing. Various physical and mechanical

properties such as density, percentage of porosity and hardness were tested to clarify the

effect of processing parameters. Phase identification and microstructure were analyzed by

X-Ray Diffractometer and optical microscope respectively. Sintering time in the

experiment was 5, 30, 45, 60 and 90 min, sintering temperature was 830°C and 900°C and

ratio of adding aluminium was 0wt%Al (no adding), 5wt%Al and 10wt%Al. It was found

that the larger the addition of aluminium, the greater was the reduction in density and

hardness in all sintering conditions. However, additional heat treatment after sintering, by

isothermal annealing at 750oC for 1 h and quenching in water, increased the hardness of all

specimens. The best processing condition to obtain high hardness was sintering at 900oC

for 30 to 60 min, followed by isothermal annealing at 750oC and quenching in water.

Keywords: Self-Lubricating Bearing; Cu-10wt%Sn; Powder metallurgy

1. Introduction

Porous parts are divided into two groups, filters and self-lubricating bearing. Glass,

ceramics and metallic materials can be used as the starting materials [1]. Nevertheless

sintered metal powders has the best performance as starting materials, which has high

strength, high thermal resistance, high corrosion resistance, durability and ease to control

porosity and permeability. Self-lubricating bearings are one of the oldest industrial

applications of porous P/M part, dating back to the mid-1920. They remain the highest part


CF‐12

produced by the P/M industry. Metal powders used for porous parts are selected according

to the application. The most commonly used powders include bronze, stainless steel, nickel

and nickel-base alloys, titanium and aluminium. As mentioned above, while operating self-

lubricating bearing, it receives acting force all the times even through it is lubricated.

Consequently, improvement self-lubricating bearing in reducing wear and has less friction

coefficient by dispersion hardening make bearing has longevity.

2. Experimental procedures

Premixed 90-10wt% of copper-tin powder and 99.9% pure aluminium powder were mixed

and blended together in various ratio of added aluminium: 0wt%Al (no adding), 5wt% and

10wt%Al. The mixture was compacted into cylindrical shape (1.1 mm. diameter and 1-1.3

mm. height) under a pressure of 2000 kg. Weight, size of the sample was measure to

calculate density before sintering. Sintering is in a batch type alumina tube furnace

maintained at 830 °C and 900 °C for 5, 30, 45, 60 and 90 min under argon atmosphere.

After sintering samples weight, size were measured. Density and porosity are measure by

Archimedes’ method. The samples were phase identified by using X-ray Diffractometer.

Microstructure was investigated by using optical microscope and scanning electron

microscope.


Results will be discussed in two parts: result from sample sintered at 900 °C and after heat

treatment.

3.1 Sintering temperature at 900 °C

It was found that microstructure of different sintering time sample look similar. Fig.1

shows microstructure of 0wt%Al, 5wt% and 10wt%Al sample sintered at 900 °C for 30

min, (a) – (c) at center of the sample (d) – (f) at edge of the sample. Pores at the center of

the sample are quite round but pores at the edge of the sample are irregular, and with added

aluminium the pores become more irregular.


CF‐12

Fig.1 microstructure of 0wt%Al, 5wt% and 10wt%Al sample sintered at 900°C for 30 min

(a) – (c) at the center of the sample (d) – (f) at the edge of the sample

Fig.2 and 3 demonstrate the density and hardness of 0wt%Al (no adding), 5wt%Al and

10wt%Al samples sintered at 900°C.

0123456789

0 20 40 60 80 100Sintering Time (min)

Bul

k D

ensi

ty (g

/cm

3 )

0% Al 5% Al 10% Al

(a) (b) (c)

(d) (e) (f)


CF‐12

Fig. 2 Effect of sintering time and amount of aluminium on the density of Cu-10wt%Sn

samples sintered at 900 °C.

0

10

20

30

40

50

60

70

0 20 40 60 80 100Sintering Time (min)

Har

dnes

s (H

V 1k

g)0% Al 5% Al 10% Al STD

Fig.3 Effect of sintering time and amount of aluminium on the hardness of Cu-10wt%Sn

samples sintered at 900 °C.

Density of no adding aluminium samples sintered at 900°C for 30 min was the greatest

which was 7.81 g⋅cm-3, with minimum porosity of 9.53% and the maximum hardness was

54.4 HV. Density of 5wt%Al and 10wt%Al adding samples sintered at 900°C for 30 min

was greatest, which were 6.63 and 5.44 g⋅cm-3 respectively. The porosity was 15.53% and

27.96% respectively. However the hardness was maximum at 51.30 and 37.90 HV when

sintered at 900°C for 5 min. Samples without addition of aluminium have higher hardness

than the ones with aluminium for all sintering time. Effect of aluminium on the density of

the samples is as following: the lager addition of aluminium, the greater is the reduction in

density in all sintering time. It is lower than the reference samples from the K.Powder

factory. Therefore porosity increases and hardness reduces when adding larger amount of

aluminium in all sintering time. The same correlation as found in samples sintered at 830

°C. When the samples have less density and more porosity, this can cause stress

concentration at the edge of the pores. Stress concentration is one of the causes that make

the sample have lower hardness. Two possible causes of pore occurrence are vaporization

of aluminium and Kirkendall void. Vaporization of aluminium occurs because aluminium


CF‐12

has higher vapor pressure at high temperature. Vapor pressure of pure aluminium obeys

equation 1. [2]

Eq.1

P is a vapor pressure in Pascal and T is an absolute temperature. Vapor pressure at 900 °C

is 0.17 × 10-3 Pa. Atmospheric pressure is much larger than vapor pressure of pure

aluminium at 900°C therefore vapor pressure effect can be neglected. Weight loss after

sintering of samples sintered at 900°C for different sintering time is shown in Table 1.

Table 1 Weight loss after sintering at 900°C for different sintering time

Lubricant weight plus aluminium weight for 0wt%Al, 5wt%Al and 10wt%Al samples are

0.045 g, 0.375 g and 0.715 g respectively. It can be seen that weight loss of all samples is

less than lubricant weight plus aluminium weight. Therefore the main cause of pore

occurrence in samples is from Kirkendall effect which is diffusion phenomenon of two

species with different diffusion coefficient. This causes void in the samples. In this case

void caused from tin and aluminium diffusing into copper faster than the reverse. Long

sintering time gives tin and aluminium more time to diffuse into copper and larger amount

of pores were created. Fig 4 – 6 show XRD patterns of samples sintered at 900°C for 30,

60 and 90 min.

Sintering time (min) 0 wt%Al (g) 5 wt%Al (g) 10 wt%Al (g)

5 0.040 0.030 0.030

30 0.015 0.055 0.035

45 0.040 0.035 0.045

60 0.025 0.065 0.035

90 0.025 0.050 0.035

( )T

P 16211917.10log −=


CF‐12

20 30 40 50 60 70 80

Fig.4 XRD Pattern of samples sintered at 900°C for 30 min

20 30 40 50 60 70 80


α

AlCu4 β

α α α

α

α

α α α 10wt% Al

5wt% Al

0wt% Al

2 Theta (Deg)

Inte

nsity

(a.u

.)

β

2 Theta (Deg)

Inte

nsity

(a.u

.)

α

AlCu4 β

α α

α

α

α

α α α 10wt% Al

5wt% Al

0wt% Al

β


CF‐12

20 30 40 50 60 70 80


XRD pattern show that the main phase found in samples is solid solution of 10wt% tin in

copper (α phase, JCPDS #44-1477) which can be found in all samples. Furthermore, β

phase (JCPDS #06-0621) which is solid solution of high tin content in copper (25 – 26.5

wt % tin) and intermetallic compound AlCu4 (JCPDS #28-0006) still present in 10wt%Al

samples. Splitted peak also occur in the samples sintered at 900°C for 90 min, this splitted

peak is matched with Cu-5.6Sn JCPDS files (JCPDS #31-0487). Relative amount of phases

is as following: the amount of α phase in 10wt%Al samples sintered at 900°C increases

with sintering time and amount of β phase in 10wt%Al samples sintered at 900°C

decreases with increasing sintering time. While amount of α phase increase with sintering

time, so are the vacancies occurring from diffusion of tin into copper. Therefore density

and porosity trend are as aforementioned.

3.2 Results from heat treatment

Fig 7 shows the microstructure of samples sintered at 900°C for 30 min and heat treated at

750°C for 60 min of 0wt%Al, 5wt% and 10wt%Al sample, (a) – (c) at the center of the

sample, (d) – (f) at the edge of the sample.

2 Theta (Deg)

Inte

nsity

(a.u

.)

α

AlCu4

α α

α

α

α

α α

α AlCu

4 10wt% Al

5wt% Al

0wt% Al

β Cu 5.6Sn

Cu 5.6Sn

Cu 5.6Sn


CF‐12

Fig 7 Microstructure of samples sintered at 900°C for 30 min and heat treated at 750°C for

60 min of 0wt%Al, 5wt% and 10wt%Al sample (a) – (c) at the center of the sample (d) –

(f) at the edge of the sample.

From the microstructures it was found that pores at center and edge of the sample slightly

increase when compare to the ones before heat treatment. Table 2 show the density,

porosity and hardness of 0wt%Al (no adding), 5wt%Al and 10wt%Al samples sintered at

900°C and heat treatment at 750°C.

Table 2 Density, porosity and hardness of 0wt%Al (no adding), 5wt%Al and 10wt%Al

samples sintered at 900°C and heat treatment at 750°C.

Comparing density and porosity before and after heat treatment, it was found that

density after heat treatment slightly lower than before heat treatment so porosity after heat

treatment is slightly higher than before heat treatment. Hardness after heat treatment of

0wt%Al is slightly lower than before heat treatment. However, the sample with added

Al amount (wt%) Bulk Density (g/cm3) % Porosity Hardness (HV 1kg)

0 7.38 15.70 51.00

5 6.43 19.66 69.20

10 5.26 29.97 39.60

(a) (b) (c)

(d) (e) (f)


CF‐12

aluminium, 5wt%Al and 10wt%Al, increase 37% and 25% respectively. Fig.8 show XRD

pattern of sample sintering at 900°C for 30 min and heat treated at 750°C for 60 min

Fig.8 XRD Pattern of samples sintered at 900°C for 30 min and heat treated at 750°C for

60 min

According to XRD pattern of heat treated samples, it was found that the main phase

existing in samples is solid solution of tin in copper (α phase), same as in samples before

heat treatment, and α phase can be found in all samples. Five weight percent samples

contain both α phase and slight amount of β phase and AlCu4. Ten weight percent samples

contain α, β phase and AlCu4. When consider relative amount of phases in heat treated

samples, it was found that for no adding aluminium sample contains only α phase. For

added aluminum sample, 5wt%Al contains β phase, as a new phase which was not found

in the sample before heat treatment. Ten weight percent aluminum sample has no new

phase after heat treatment. Therefore the cause of reduction of hardness of no adding

aluminum sample (0wt%Al) is pores formation and coalescence. Although density of

5wt%Al sample is slightly reduced, but the hard new phase AlCu4 existing in the sample

increases its hardness.

α

AlCu β

α αα

α

α

α α

10wt% Al

5wt% Al

0wt% Al

2 Theta (Deg)

Inte

nsity

β

AlCuAlCuβ

20 30 40 50 60 70 80


CF‐12

4. Conclusions

1. The larger addition of aluminium, the greater reduction in density and hardness in

all sintering conditions and they are lower than the reference samples from the

factory.

2. Porosity increases and hardness is reduced when adding larger amount of

aluminium in all sintering time.

3. The main cause of pore occurrence in samples is from Kirkendall effect.

4. Additional heat treatment at 750°C for 60 min can increase hardness of 5wt%Al

and 10wt%Al samples by 37% and 25% respectively.

5. Acknowledgement

This project is financially supported by the Graduated School of Chulalongkorn

University. The authors would like to thank K.POWDER METAL CO., LTD for providing

premixed Cu-10wt%Sn powder.

6. References

[1] ASM Powder Metallurgy Committee. 1993. ASM Handbook Volume 7 Powder

Metallurgy. Fifth Printing. United State of America

[2] David R. Lide (ed), CRC Handbook of Chemistry and Physics, 84th Edition. CRC

Press.Boca Raton, Florida, 2003; Section 4, Properties of the Elements and Inorganic

Compounds; Vapor Pressure of the Metallic Elements

[3] P.Villars, A.Prince & H.Okamoto, ASM Handbook of Ternary Alloy Phases Diagrams

Volume 4.


DF‐03

Feasibility Study of Lard Oil and LPG as Fuels for Aluminum Crucible Furnace

Supakiat Supasin1, Sumpun Chaitep2, Saritporn Vittayapadung3, Lin Lin3

1Graduate Student, Mechanical Engineering Department, Faculty of Engineering,

Chiang Mai University, Chiang Mai, Thailand

Tel: +6686 6570937 Fax: +6653 942062 Email: [email protected] 2PARA Laboratory, Mechanical Engineering Department, Faculty of Engineering,

Chiang Mai University, Chiang Mai, Thailand

Tel: +6653 942005 Fax: +6653 942062 Email: [email protected] 3School of Food & Biological Engineering, Jiangsu University, Zhenjiang, P.R. China

Email: [email protected], [email protected]

Abstract

This research was to study the used of alternative fuel from animal easily found in

Thailand. Generally, diesel oil would be used for metals melting processes. A new package

of aluminum crucible furnace and burner was designed and built, lard oil was selected to

be used in this research. This paper indicates the possibility of using lard oil in

combination with liquefied petroleum gas instead of diesel oil in aluminum melting

process.

The experiment results found that overall thermal efficiency of the system was

equal to 6.47%. Five kilograms of aluminum was melted in 34.02 minutes, with fuel

consumption of lard oil at 0.0015 kg/s, combined with fuel consumption of liquefied

petroleum gas at 0.0013 kg/s and total fuel cost for aluminum melting was 25.1 baht/kg.

While, the overall thermal efficiency of crucible furnace using diesel oil was equal to

6.64% with 33.22 minutes of melting time. The used of diesel oil obtained fuel

consumption of 0.0027 kg/s aluminum melting cost was to 37.5 baht/kg. Finally, a

comparison of aluminum melting cost under different fuel was described using lard oil in

aluminum melting process. It could be concluded that lard oil has all benefit and

appropriated to be used as a main fuel in melting process over the conventional use of

diesel oil.

Keywords: Lard Oil; LPG; Aluminum Crucible Furnace; Overall Thermal Efficiency


DF‐03

1. Introduction

At present, the foundry process is widely used in modern industries. Several specialized

furnaces are used to melt the metal. Furnaces are refractory lined vessels that contain the

material to be melted and provide the energy to melt it. Furnace design is a complex process,

and the design can be optimized based on multiple factors. [1-2] Furnaces in foundries can be

any size and they are designed according to the type of metals that are to be melted. Also,

furnaces must be designed around the fuel being used to produce the desired temperature.

Electricity, fossil oil, gas and drying fuel are normally used in melting process [3-5].

Metal casting is a manufacturing process by which a liquid material is usually poured into a

mold, which contains a hollow cavity of the desired shape, and then allowed to solidify [6].

Melting is performed in a furnace. Virgin material, external scrap, internal scrap, and alloying

elements are used to charge the furnace. Virgin material refers to commercially pure forms of

the primary metal used to form a particular alloy. The solidified part is also known as a

casting, which is ejected or broken out of the mold to complete the process. Casting materials

are usually metals or various cold setting materials that cure after mixing two or more

components together [7-8]. Metal casting is most often used for making complex shapes that

would be otherwise difficult or uneconomical to make by other methods. Generally, cast steel

and nonferrous metal, such as, copper, brass and aluminum would be used in metal casting

process [9].

Aluminum Crucible in foundry industry is using much of energy, especially in its

production process [10]. It is generally known that the prices of diesel oil and gas fuel are

daily increased. Diesel oil was usually used in aluminum melting process. The demand of high

energy consumption was caused of high investment cost [11]. In the foundry process of some

materials was also use a lot of energy consumption [12]. Therefore, it was interesting, if this

process could be decreased the energy consumption.

Due to, Thailand is agricultural country, plenty of animal husbandry is popularly practiced

[13]. Animal fat is obtained from the tissues of mammals in the commercial processes of

rendering or extracting. It consists predominantly of glyceride esters of fatty acids and

contains no additions of free fatty acids. Chemical reaction of animal fat is shown in Figure 1

[14]. In actuality, the animal source is not specified or required to give the origin of

slaughtered animals. However, Thailand can produce animal fat, especially, swine fat or lard

oil which is enough to expect used instead of diesel oil in this research. Figure 2 show animal

slaughtered (swine, cattle and buffalo) for consumption in Thailand (year 2008) [15].


DF‐03

Figure 1. Chemical reaction of animal fat

-

1,000,000

2,000,000

3,000,000

4,000,000

5,000,000

6,000,000

7,000,000

8,000,000

Cattle Buffalo Swine

Num

ber

S laughter livestock Expected S laughter livestock

Figure 2. Animal slaughtered in Thailand 2008.

From overall review, it was found that animal fat, especially, lard oil might be used instead

of fossil fuel such as diesel oil. Consequently, lard oil was used as main fuel in this research

[17]. The main objective of this research was to design and build the burner for aluminum

crucible furnace using Liquefied Petroleum Gas (LPG) combine with lard oil as fuels in

aluminum melting process. This would be the alternative way for energy saving and decreased

the investment cost in foundry industry.

2. Lard Oil as Fuel

The pig abdominal fat is one of low cost product and it can be produced as the lard oil fuel

with quantity ratio of 90% by mass [16]. Lard oil is consisted chiefly of olein that is expressed

from lard and used especially as a lubricant, cutting oil or illuminant. Its specification could

say that similar to diesel oil. Table 1 show the comparative properties results between lard oil

and diesel oil which was experimented in PARA laboratory, Faculty of Engineering, Chiang

Mai University, Chiang Mai, Thailand [18].

Table 1 Property of Lard Oil and Diesel

CH2 - OH

CH - OH

CH2 - OH

+ 3R - C - OHO

CH2 - O

CH - O

CH2 - OO

- C - R

- C - RO

- C - R

O+ 3H2O

[Glycerol] [Fatty Acid] [Fat & Oil] [Water]

CH2 - OH

CH - OH

CH2 - OH

+ 3R - C - OHO

CH2 - O

CH - O

CH2 - OO

- C - R

- C - RO

- C - R

O+ 3H2O

[Glycerol] [Fatty Acid] [Fat & Oil] [Water]


DF‐03

Properties of Fuels Lard

Oil*

Diesel

Oil

Heating Value

(kJ/kg) 41,688 42,500

(40oC) 44.077 2.2 –

5.3 Viscosity

(cSt) (90oC) 11.491 -

Specific Gravity @

25oC 0.85 -

* Lard oil measured from the experiment.

3. Theory

Crucible Furnace system used some thermal theories to calculation as follows [19-21],

3.1 Heat quantity of fuel, fQ

•

The heat quantity of fuel is calculated from the fuel combustion process which was

determined from equation (1).

f fQ m LHV

• •

= × (1)

Where; fQ

• = Fuel heat quantity; kW

fm

• = Fuel mass flow rate; kg/s

LHV = Fuel low heating value; kJ/kg

3.2 Air sensible heat, airQ•

The temperature of fuel fed in the combustion process was equal to outside temperature, in

case of no fuel preheating. Air sensible heat could be calculated follow in equation (2)

( )airair p a ambairQ m C T T

• •

= × × − (2)

Where;

airQ• = Air Sensible heat; kW


DF‐03

airm• = Air mass flow rate; kg/s

airpC = Specific heat capacity (Air); kJ/kgoC

aT = Air Temperature in Burner; oC

ambT = Ambient Temperature; oC

3.3 Heat loss through the furnace wall, wallQ•

The calculation of heat loss through the furnace wall consisting of refractory brick and

steel sheet material was shown in equation (3).

, ,

3 12 1 ( /( / )2 2

w in w outwall

b s

T TQ In r rIn r r

k L k Lπ π

• −=

+

(3)

Where;

wallQ•

= Heat loss through the furnace wall; kW

,w inT = Inside temperature of furnace wall; oC

,w outT = Outside temperature of furnace wall; oC

1r = Radius from center of furnace to inside refractory brick; mm

2r = Radius from center of furnace to inside steel sheet; mm

3r = Radius from center of furnace to outside steel sheet; mm

bk = Thermal Conductivity of refractory brick; W/m-K

sk = Thermal Conductivity of steel sheet; W/m-K

L = Height of furnace; mm

3.4 Exhaust gas heat loss, flueQ

•

While, crucible furnace was operated, aluminum was melted by heat quantity of fuel. The

heat loss of exhaust gas at the top of furnace that was released to the outside from a stack to

environment could be calculated as follow in equation (4)

( ),f p flue flue ambflueQ m C G T T• •

= × × × − (4)

Where;


DF‐03

flueQ• = Heat loss in exhaust gas; kW

fm• = Fuel mass flow rate; kg/s

,p flueC = Specific heat capacity; kJ/kgoC

flueT = Temperature of exhaust gas; oC

ambT = Ambient Temperature; oC

G = An exhaust gas quantity; -

3.5 Heat transfer to aluminum material, mQ

•

The heat quantity would be transferred to aluminum material in the crucible inside the

furnace. This calculation could be determined by equation (5)

,

,

( )60 ( ) 60

A p A melt start Am

melt melt end melt

m C T T m LHQt t t

• ⎡ ⎤−⎡ ⎤ ×= + ⎢ ⎥⎢ ⎥× − ×⎢ ⎥⎣ ⎦ ⎣ ⎦

(5)

Where;

mQ•

= Heat transfer to aluminum; kW

Am = Mass of aluminum in one batch; kg

,p AC = Specific heat capacity of aluminum; kJ/kgoC

meltT = Melting point of aluminum; oC

startT = Initial temperature of aluminum; oC

LH = Latent heat; J/kg

meltt = Initial melting time; min

,melt endt = End of melting time; min

3.6 Thermal efficiency of crucible furnace, furnaceη

Thermal efficiency is a measure of the efficiency of converting a fuel to energy and useful

work. It used for evaluated the performance of aluminum crucible furnace. Thermal efficiency

of aluminum crucible furnace could be followed by equation (6)

mfurnace

input

Q

Qη

•

•= × 100 (6)


DF‐03

Where;

furnaceη = Thermal efficiency; %

mQ•

= Quantity heat product; kW

inputQ• = Quantity heat inlet; kW

4. Materials & Methods

4.1 Design and Construction

This research was to design and construct crucible furnace system for aluminum melting

process. The crucible furnace used materials which easily found in Thailand, it consisted of

Refractory brick No.SK36, Steel sheet St.37, Refractory mortar No.70AM.

In case of burner (See Figure 3), its components included of stainless steel (SUS-304) pipe

4 inch, Lard oil pipe used SUS-304 size 3/8 inch diameter, Venturi size 4 inch diameter and

Heat buffer was modified from refractory brick No.C2.

Figure 3 shows the components of burner which constructed in our laboratory. From figure

3, air inlet at position 1 to 3, in the other hand, LPG was also supplied from position 2 to 3.

Position 3 was the combustion and burner to burn LPG and air inlet. Heat transferred from

position 3 (660oC) to the venturi in position 4 and heat buffer transferred heat to lard oil which

moved inside the SUS-304 stainless pipe.

E x h a u s t o u t l e t7

H e a t e x c h a n g e r6

H e a t b u f f e r5

V e n t u r i a n d l a r d o i l o u t l e t4

L P G b u r n e r3

L P G in l e t2

A i r i n l e t1

D e s c r i p t i o n sN o .





L P G b u r n e r3

L P G in l e t2

A i r i n l e t1






L P G b u r n e r3

L P G in l e t2

A i r i n l e t1






L P G b u r n e r3

L P G in l e t2

A i r i n l e t1


Figure 3. Components of burner

Lard oil viscosity was decreased and heated up while being moved to position 4. It was

then forced feed and entrained with moving air to position 5. Finally, the droplet touched high

temperature heat buffer. The flame oxidation was visible through out behind position 7.


DF‐03

The crucible furnace with burner was constructed in this research as shown in figure 4.

Figure 4. Aluminum crucible furnace using liquefied petroleum gas and lard oil as fuel

4.2 Experimental Methodology

The experiment of aluminum crucible furnace using petroleum gas and lard oil as fuel

system for melting 5 kg/batch of aluminum was tested in PARA Laboratory, Faculty of

Engineering, Chiang Mai University, Chiang Mai, Thailand.

5. Results & Discussions

5.1 Principle Burner Testing

After finished burner construction, it was tested to determine the properties. Burner testing

results were shown in Table 2.

Table 2 Burner tested properties

Mass flow rate (kg/s)

No Air Lard LPG

Output

Temp.

(oC)

Combustion

Efficiency

(%)

1 0.042 0.0010 0.0001 1,027.0 96.76

2 0.050 0.0013 0.0001 1,076.0 74.10

3 0.055 0.0014 0.0001 1,051.0 72.36

4 0.061 0.0014 0.0001 965.3 88.80

5 0.061 0.0012 0.0001 853.4 97.32

The testing results of burner were found that the capacity of combustion burner which was

designed and constructed could be used in aluminum crucible for melting process. Because

burner provided the maximum temperature (1,076oC), was higher than melting point of


DF‐03

aluminum (660oC). Figure 5 was the burner testing.

Figure 5. Burner testing in PARA Laboratory

5.2 Aluminum melting process using lard oil and LPG as fuel

The results of aluminum melting process using lard oil and LPG as fuel exhibited the

temperature inside furnace, air released temperature, crucible temperature and temperature of

aluminum were related in this experiment. It was found that air flow rate representing 0.0302

kg/s, while, lard oil fuel flow rate and LPG fuel flow rate were 0.0015 kg/s and 0.0013 kg/s,

approximately. The experiment was done till aluminum (solid) become aluminum (liquid).

The result was shown in figure 6.

Graph in figure 6 indicated that the trends of temperature lines were continuously increase.

While, all aluminum in crucible was melted, it obtained the highest temperature of aluminum

melting was 962.38oC and took 34.02 minutes melting time. Finally, molten aluminum was

ready to be poured into the molds.

0

200

400

600

800

1000

1200

0 5 10 15 20 25 30 35

Melting Time (Minute)

Tem

pera

ture

(o C)

Temperature in FurnaceTemperature OutletTemperature CrucibleTemperature Aluminum

Figure 6. Aluminum temperatures in crucible furnace using LPG and lard oil as fuel

5.3 Aluminum melting process using diesel oil as fuel

The results of aluminum melting process using diesel oil as fuel exhibited the temperature

inside furnace, air released temperature, crucible temperature and temperature of aluminum

were related in this experiment. It was found that air flow rate representing 0.030 kg/s, while,

diesel oil fuel flow rate was 0.0026 kg/s. The result was shown in figure 7.


DF‐03

0

200

400

600

800

1000

1200

0 5 10 15 20 25 30 35


Tem

pera

ture

(o C)

Temperature in FurnaceTemperature OutletTemperature CrucibleTemperature Aluminum

Figure 7. Aluminum temperatures in crucible furnace using Diesel as fuel

Graph in figure 7 indicated that the trends of temperature lines were continuously increase.

While, all aluminum in crucible was melted, it obtained the highest temperature of aluminum

melting was 1,000oC and took 33.24 minute melting time consumption. Finally, molten

aluminum was ready poured into the molds.

5.4 The comparison of aluminum temperatures between lard oil and Diesel oil in melting

system.

From item 4.2 and 4.3, the aluminum temperatures caparisoned between lard oil and diesel

oil in melting of 4 kg aluminum. Graph in figure 8 showed the aluminum temperature using

different fuels.

The result shown in figure 8 was found that aluminum temperature using diesel oil was

higher than lard oil in melting process, due to, high heating value of diesel oil was higher than

lard oil. However, the results of these 2 fuels used did not much different. Burning

temperature of lard oil as fuel was also enough to melt the aluminum.

0

200

400

600

800

1000

1200

0 5 10 15 20 25 30 35


Tem

pera

ture

(o C)

Furnace Diesel Furnace Lard Oil

Figure 8. The aluminum temperature comparison using different fuels

Therefore, it was proved that lard oil with LPG as fuel in aluminum melting process could

be used instead of diesel oil. This would be one of new choice to select and use alternative


DF‐03

energy which easily found in Thailand, remunerated the fossil fuel.

6. Conclusions

This research was to study the used of alternative fuel from animal easily found in

Thailand. Generally, diesel oil would be used for metals melting processes. The new package

of aluminum crucible furnace and burner were designed and built, lard oil was selected to use

in this research. This paper indicates the possibility of using lard oil in combination with

liquefied petroleum gas instead of diesel oil in aluminum melting process.

The experiment results were found that overall thermal efficiency of the system was equal

to 6.47%. Five kilograms of aluminum was melted in 34.02 minutes, with fuel consumption of

lard oil at 0.0015 kg/s, combined with fuel consumption of liquefied petroleum gas at 0.0013

kg/s and aluminum melting cost was to 25.1 baht/kg. While, the overall thermal efficiency of

crucible furnace using diesel oil was equal to 6.64% with 33.22 minutes of melting time

process. The used of diesel oil obtained fuel consumption of 0.0027 kg/s aluminum melting

cost was to 37.5 baht/kg. Finally, a comparison of aluminum melting cost under different fuel

was described using lard oil in aluminum melting process. It could be concluded that lard oil

has all benefit and appropriated to use as main fuel in melting process over the conventional

use of diesel oil.

7. Acknowledgements

The authors gratefully thank the financial supports from postgraduate thesis fund, Small

Gas Turbine Development Research, Institute for Science and Technology Research and

Development, PARA (Propulsion & Aerodynamics Research & Application) and FAME

(Food & Agricultural Machinery Engineering) Laboratories and the department of Mechanical

Engineering Faculty of Engineering Chiang Mai University, Thailand. And special thanks to

the School of Food & Biological Engineering for this cooperative work.

8. References

[1] B. Cabric, T. Pavlovic and A. Janicijevic, “Regulation of the crystallization in a crucible

Furnace”, Journal of Crystal Growth, Vol. 20, 1998. pp 339-340.

[2] H. Calisto, N. Martins and N. Afgan, “Diagnostic system for boilers and furnaces using

CFD and neural networks”, Expert Systems with Application, Vol. 35, 2008. pp 1780-

1787.

[3] A.R. Khoei, I. Masters and D.T. Gethin, “Design optimization of aluminum recycling


DF‐03

process using Taguchi technique”, Journal of Materials Processing Technology, 127, 2002.

pp. 96-106.

[4] V. Krivandin and B. Mrkov, Metallurgical Furnaces, Mir Publishers. Moscow. Russia,

1980. 509 pp.

[5] Y.J. Zhang, P.V. Barr and T.R. Meadowcroft, “Continuous Scrap Melting In a Short

Rotary Furnace”, Minerals Engineering, Vol. 21, 2008. pp 178-189.

[6] L.J. Yang, “The effect of casting temperature on the properties of squeeze cast aluminum

and zinc alloys”, Journal of Materials Processing Technology, 140, 2003. pp. 391-396.

[7] O. Auchet, P. Riedinger, O. Malasse and C. Iung, “First-principles simplified modeling of

glass furnaces combustion chambers”, Control Engineering Practice, 116, 2004. pp. 1443-

1456.

[8] P.M. Sobrinho, J.A. Carvalho, J.L. Silveira and P.M. Filho, “Analysis of aluminum plates

under heating in electrical and natural gas furnaces”, Energy, Vol. 25, 2000. pp. 975-987.

[9] A.R. Khoei, I. Masters and D.T. Gethin, “Numerical modeling of the rotary furnace in

aluminum recycling processes”, Journal of Materials Processing Technology, 139, 2003.

pp. 567-572.

[10] M. Jackson, M.L. Pantoya and W. Gill, “Characterization of a gas burner to simulate a

propellant flame and evaluate aluminum particle combustion”, Combustion and Flame,

153, 2008. pp. 58-70.

[11] G.L Borman and K.W. Ragland, Combustion Engineering, WCB McGraw-Hill. Int.,

1998. 613 pp.

[12] Energy Information Administration (EIA), Official Energy Statistics from the U.S.

Government, 2008. [Online] Available: http:// tonto.eia.doe.gov/oog/diesel.asp.

[13] Department of Livestock Development, Livestock Infrastructure Information, Ministry of

Agriculture and Cooperative of Thailand, 2007. [Online] Available:

http://www.dld.go.th/i /index.html

[14] N. Rattanapanont, Food Chemistry, 1st edition, Odian Store, Bangkok, 2001. 487 pp.

[15] Department of Livestock Development, Animal Statistics Ministry of Agriculture and

Cooperative of Thailand, 2008. [Online] Available:

http://www.dld.go.th/jxbvict/stat_web /index _stat.html

[16] U. Werner, U. Stohr and N. Hees, Biogas plants in animal husbandry, BAU-Biogas

advisory unit, 1999. 153 pp.

[17] S. Supasin and S. Chaitep, “Lard as an Alternative Fuel Replacing Diesel Oil in Crucible

Furnace”, The 2nd Symposium on Engineering and Architecture for the Sustainable


DF‐03

Development in the Greater Mekong Sub-Region, Luang Prabang, Laos, 2008. pp 91-94.

[18] S. Supasin, S. Chaitep and N. Vorayos, “Design of Burner for Aluminum Crucible

Furnace using Liquefied Petroleum Gas (LPG) and Land Oil as Fuels”, The 15th Tri-

University International Joint Seminar and Symposium, Zhenjiang, P.R. China, 2008. pp

194-197.

[19] J. Gosse, Technical Guide to Thermal Processes, Cambridge University: England, 1986.

227 pp.

[20] K.V. Mitzlaff, Engines for Biogas Theory, Modification, Economic, Operation, Federal

Republic of Germany, 1988. 133 pp.

[21] S. Sarannit, Heat Transfer, Technology Promotion Association (Thailand-Japan), 1st

edition, 2002. 467pp.


DF‐05

Development of an aluminum semi-solid extrusion

process

T. Rattanochaikula, S. Janudoma, N. Memongkolb ,J. Wannasina*

aDepartment of Mining and Materials Engineering, Faculty of Engineering, Prince of

Songkla University, Hat Yai, Sonkhla 90112, Thailand. bDepartment of Industrial Engineering, Faculty of Engineering, Prince of Songkla

University, Hat Yai, Sonkhla 90112, Thailand.

*Corresponding Author: [email protected]

Abstract

An aluminum extrusion process is mainly used to fabricate long tubes, beams and rods for

various applications. However, this process has a high production cost due to the high

investment cost of high-pressure machinery. The objective of this work is to develop a new

extrusion process using a semi-solid metal forming technology. In this study, a laboratory

extrusion system was used to fabricate aluminum rods with the diameter of 12 mm. The

semi-solid metal process used in this study is the Gas Induced Semi-Solid (GISS)

technique. To study the feasibility of the GISS extrusion process, the effects of extrusion

parameters such as plunger speed and solid fraction on the extrudability and microstructure

of extruded samples were investigated. The results show that the plunger speed and solid

fraction of the semi-solid metal need to be carefully controlled to produced complete

extruded parts.

Key words: Aluminum alloys; Aluminum extrusion; Semi-Solid Metal; Extrusion process;

Microstructure; Gas Induced Semi-Solid (GISS)

1. Introduction

Extrusion is one of various forming processes that is used to produce long, straight metal

products with constant cross section, such as bars, solid and hollow sections, tubes and

wires [1]. In the process, a billet is heated and forced through a die orifice. The products

from this extrusion process are near net shape and long. However, the extrusion process


DF‐05

requires a high-pressure machine to force the metal in the solid state. In addition, defects

such as surface defect and piping can be present in the products of an extrusion process [1].

Semi-solid rheo-extrusion is a new extrusion process that has several advantages

such as low extrusion force, high fluidity of materials, and low friction force between the

die and the materials [2]. In a rheo-extrusion process, the metal alloy is melted in a furnace

and then extruded at a temperature between the solidus and liquidus temperature of the

metal alloy. The slurry is forced through a die orifice to form a desired part.

Several previous studies have been reported regarding the behaviors of the rheo-

extrusion process. However, no complete research in the rheo-extrusion process has been

published [3-5]. To apply the rheo-extrusion process in the production of commercial parts,

it is important to conduct further studies. This research paper reports a preliminary research

and development work of a new rheo-extrusion process using the Gas Induced Semi-Solid

(GISS) technique. In this study, the effects of the plunger speeds and solid fractions on the

extrudability of an aluminum A356 alloy were investigated.

2. Materials and Experimental Procedure

The raw material used in this work is aluminum A356 alloy. The chemical composition of

the alloy is shown in Table 1.

Table1. Chemical composition of aluminum A356 alloy

Element Si Fe Cu Mn Mg Zn Ti Al

Weight% 6.9 0.42 0.05 0.04 0.42 0.01 0.10 Bal.

Preparation of semi-solid slurry: The aluminum A356 alloy was melted in an electric

furnace at the temperature of about 650°C. Approximately 300 grams of the molten

aluminum was taken from the crucible by a ladle. When the temperature of the molten

aluminum was about 620°C, a graphite diffuser was immersed to induce nitrogen gas for 5

seconds. A semi-solid slurry with the solid fraction of about 10% was then obtained. A

schematic drawing of the GISS technique and the GISS machine is shown in Figure 1.


DF‐05

(a) (b)

Figure 1: (a) Schematic drawing of the GISS technique [6] and (b) the GISS machine to

prepare semi-solid slurry

Rheo-extrusion test: The semi-solid slurry from the GISS machine was then poured into

a shot sleeve with the inner diameter of 40 mm. The shot sleeve was preheated to about

350°C-400°C. Next, the slurry was forced by a plunger at various speeds of 2, 4, and 6

cm/s through a die, a graphite support and a water-cooled tube. The inner diameter of the

die was 12 mm. The schematic drawing of this rheo-extrusion process is shown in Figure

2. The holding time of the slurry in the shot sleeve, 0 second and 5 seconds at each

plunger speed, was also studied in this work. Figure 3 shows the extrusion die and the

laboratory-scale extrusion machine. This machine has a 20-ton capacity with a hydraulic

system.

Figure 2: The schematic drawing of this extrusion process.

Figure 3: The extrusion die and laboratory-scale machine used in this study.


DF‐05

Results analysis methods

The extruded samples were analyzed by three criteria to determine the extrudability. The

methods are briefly described as follow:

Length of the samples:, the length of the samples was measured after the extrusion

test. In this work, the criterion for the required length was 15 cm. shorter samples then the

criterions were rejected.

Surface quality: The surface of the samples was also examined. Samples with

smooth surfaces in all the area would pass the requirement.

Microstructure uniformity: the microstructure of samples was observed using an

optical microscope.

The samples were cut and obtained from two positions as shown in Figure 4. The samples

were then prepared for metallographic analysis using the standard grinding, polishing and

etching procedure. Good extruded parts should have uniform microstructure throughout the

length.

Figure 4: The sampling position.


The representative extruded samples from the experiments are given in Figure 5. The

results show that faster plunger speed and lower holding time yield longer samples. Only

the sample with a low plunger speed of 2 cm/s and a longer holding time of 5 s did not pass

the length criteria. (see Table 2.) The results suggest that conducting the rheo-extrusion

process with a high speed and with the low-solid-fraction slurry (no holding time) gives the

longest length as expected since the slurry can flow easier and faster.

1 inch 1 inch

Sample


DF‐05

(a)

(b)

Figure 5: The samples from GISS extrusion at (a) each plunger speed and 5 seconds of

holding time and (b) each plunger speed and 5 seconds of holding time.

(a) (b)

Figure 6: The surface finished of samples of (a) 4cm/s of plunger speed and holding time is

5 seconds and (b) 4cm/s of plunger speed and no holding time.

Table 2. The semi-solid extrudability of A356

Al-alloy

0s 5s

length Surface length Surface

2 cm/s x x

4 cm/s x

6 cm/s x x

2cm/s

4cm/s

6cm/s

24cm

23cm

14cm

2cm/s

4cm/s

6cm/s35cm

30cm

27cm


DF‐05

However, when the surfaces of the samples were examined, the samples produced by fast

speed and at a low solid fraction have surface defect as shown in Figure 6(b). Only the

samples produced by lower speed (2-4 cm/s) and at a higher solid fraction pass the surface

quality requirement. The fast flow speed of the slurry may cause turbulent flow causing the

surface defect. By increasing the viscosity of the slurry through increasing the solid

fraction, the slurry will have laminar flow at the same flow speed.

From these results, only samples produced by the conditions of 4 cm/s plunger speed and

holding time of 5 seconds pass the requirements of length and surface quality.

(a) Middle

(b) Edge

Figure 7: The representative microstructures of the cross section of the samples

For all the samples, the microstructures at the edge and the middle are similar. Figure 7

shows representative microstructures at the edge and the middle of the samples. The

micrographs show that the solid particles are concentrated near the center of the channel


DF‐05

during the flow. Representative microstructures of the samples at the tip and the base of the

rods at various plunger speeds and holding times are given in Figure 8-9.

Plunger speed is 2 cm/s and no holding time



Figure 8: The microstructure of each sample that no holding time.

At tip of samples At base of samples


DF‐05

Plunger speed is 2 cm/s and 5 seconds of holding time



Figure 9: The microstructure of each sample that have 5 seconds of holding time.

In general, the amount and distribution of the primary α phase in all the samples are quite

uniform. However, the eutectic structures in the samples at the tip and the base of the rods

are different. The eutectic phase at the tip has coarse structure, as shown in Figure 10. Fine

eutectic structure is observed at the base of the rod.

The results show that the metals near the tip have longer solidification time so that the

eutectic structure can grow larger. To improve this, a better cooling system should be

applied in the rheo-extrusion system.

(a) Tip (b) Base

Fig 10: The different eutectic structure at each position.

At tip of samples

Coarse Eutectic Fine Eutectic

At base of samples


DF‐05

4. Conclusions

From this study, the following conclusions can be drawn:

1. The extrusion behavior of an aluminum A356 alloy using the GISS technique is

influenced by the plunger speed and the solid fraction of the slurry in the shot sleeve.

2. The higher solid fraction of the slurry helps reduce the surface defects of extruded

parts.

3. The non-uniformity of the eutectic microstructure is caused by inefficient cooling of

the extruded samples. This problem can be improved by adding a better cooling

system along the die.

4. This preliminary study gives important information for the development at the rho-

extrusion machine using the GISS technique in the future.

5. Acknowledgements

The authors gratefully thank the Department of Mining and Materials Engineering,

Faculty of Engineering, Price of Songkla University for financial support and facilities. We

also thank Miss Rungsinee Canyook for the metallographic preparations and the

Innovative Metal Technology (IMT) team for all the kind supports.

6. References

1. Pearson, C. and Parkins, R. 1960. The extrusion of metal. London : CHAPMAN &

HALL LTD.

2. Gerhard, H. and Reiner, K. 2004. Thixoforming. WILEY-VCH Verlag GmbH & Co.

KGaA.

3. B.S. Lee., D.H. Joo., M.H. Kim. 2005. Extrusion behavior of Al-Cu alloys in the semi-

solid state, Materials Science and Engineering A402 (2005) 170-176

4. Zhang, L.N., Wang, S.Q., Zhu, M.F., Wang, N. and Wang S.D. 2003. The extrusion

behaviour of Zn-20% Al alloy in the semi-solid state, Journal of Materials Processing

Technology Vols. 44: 91-98.


DF‐05

5. Jae-Ho Hwang, Dae-Cheol Ko, Gyu-Sik Min, Byung-Min Kim and Jae-Chan Choi.

Finite element simulation and experiment for extrusion of semi-solid Al 2024,

International Journal of Machine Tools and Manufacture Vols. 10: 1311-1328.

6. Wannasin, J. and Thanabumrungul, S. 2008. Development of a semi-solid metal

processing technique for aluminum casting applications. Songklanakarin J. Sci.

Technol., 30(2): 215-220.

7. Wannasin, J., Junudom, S., Rattanochaikul, T. and Flemings, MC. 2080.

Development of the Gas Induce Semi-Solid Metal Process for Aluminum Die Casting

Applications, Solid State Phenomena Vols. 141-143:97-102.


PF‐21

The Preparation of Silicon Nitride by Silicon Source from Rice Husk Ash

S. Rattanaveeranon1*and D. Bhongsuwan2

1 Department of General Education (Physics), Rajamangala University of Technology

Rattanakosin, Salaya, Phuttamontol, 73170, Thailand, Tel: +662-8894585-7 ext.2920 Fax +662-8894585 ext.2920

2 Materials Science Program, Faculty of Science, Prince of Songkla University, Hat yai, Songkhla 90110, Thailand, Tel: +6674-288396 Fax: +66-74218701

*Corresponding Author E-Mail : [email protected]

Abstract: This study aims at the preparation of silicon nitride powder from rice husk silica

by chemical and thermal degradation. The composition of silica powders were mixed with

carbon powders by ratio 20 : 12 wt% and calcined at temperature 1400°C at the rate of 5

°Cmin-1 under N2 atmosphere of 1 dm3.min-1.The soaking temperature was maintained for

a period of 4, 5 and 6 hours, respectively. The XRD analysis shows the presence of

cristobalite at firing temperature 1400°C and soaking time of 4 hours. The appropriate

temperature for silicon nitride formation is at 1400°C with the soaking time of 6 hours. The

scanning electron micrograph shows the surface morphology of silicon nitride phase

consisting of fibers and/or whiskers.

Introduction

Silicon nitride (Si3N4) has been widely used to fabricate cutting tools and high-temperature

structural applications due to its excellent mechanical, physical and chemical properties.

The Si3N4 ceramics with a tailored microstructure are promising high performance

materials because of such unique properties as light weight, good strain tolerance, damage

tolerance and thermal shock resistance. Traditionally, the Si3N4 ceramics have been used as

hot gas filters, high-temperature separation membranes, and catalyst supports. Recently

porous Si3N4 ceramic is also attractive in electromagnetic wave penetrating materials, as a

strategy to reduce dielectric constant and loss.

Materials and Methods

Starting material, by using the rice husk of 65 g. was reacted by 3 M of hydrochloric acid

concentres and then filtered the sample by plastics grate, then washed the sample with pure


PF‐21

water several times, finally washed the sample with distilled water, dried the sample at 100 °C with the soaking for 2 hours and then brunt the sample at 850 °C with the soaking for 3

hours (increasing temperature rate = 20 °C⋅min-1). It was obtained the pure silica powder

(SiO2) (pure silica over than 95 wt.%,), The sample was prepared by abovementione to mix

with activated carbon (pure 99.99 wt.%, 0.52 μm).The samples were prepared to mix by

ration 20 : 12 wt.% (SiO2 : C).The mixtures were ball-milled for 24 h using ceramic balls.

Absolute ethanol was used as the milling media.The resultant slurry was dried to obtain an

agglomerate-free powder mixture to dry box at 100 °C for at least 3 hours to ensure that the

powders were completely free of alcohol. The dried powders were then sieved to 60-

mesh.The sample was calcined at 1400 °C with varying the period of calcined time for 4, 5

and 6 hours, respectively.X-ray diffractometry (XRD) analysis was conducted to examine

the phases in the obtained silicon nitride ceramics and X-ray fluorescence(XRF) was

analyzed the chemical compounds of rice husk ash. The laser particle sizing analyzer

(LPSA) was measured an average particle. The morphologies of combustion products were

studied by using scanning electron microscopy (SEM).

Results

Table1. shows the analysis by using XRF technique using for analyzing chemical

compounds of rice husk ash (RHA).The main chemical compounds of this product is silica

and there is amorphous structure.(more than 99 wt.%) Furthermore, there is a trace element

such as alumina and calcium oxide of only 0.65. The hydrochloric concentrate of 3 M is

enough to remove all of the impurity in rice husk ash.

Table 1: Show that the chemical compounds of rice husk ash which calcined at 850°C for 3

hours.

Compounds Concentrate (%) Silica (SiO2) 99.35 Alumina (Al2O3) 0.23 Calcium Oxide (CaO) 0.41 Others 0.01

Fig1. Shows the distribution of particle size of silica powder.


PF‐21

Particle size (μm)

10-2 10-1 100 101 102 103 104

Volum

e (%

)

0

1

2

3

4

5

10 %

50 %

100 %

The distribution of particle size of silica powder consists of 3 types. First, the particle

size is less than 10 % which is 1.66 μm. Second, the mean particle size of silica powder is

83.04 μm. Third, the particle size is more than 90 % which is 326.08 μm, respectively.(the

samples were ball milled for 12 hours)

Fig.1 The particle size of silicon oxide was milled for 12 hours.

Fig.2 shows the XRD pattern of silicon nitride calcined at temperature 1400 °C at the rate

of 5 °C/min in nitrogen atmosphere of 1 dm3⋅min-1 soaked for 4 hours. There is no silicon

nitride in the sample but there is only cristobalite-silica structure Fig.2(a). When the

sample was calcined for 5 hours , it began to have silicon nitride and silicon oxide nitride

Fig.2(b). Furthermore, there is unreated silica which betides the silica-cristobalite only.

The sample was calcined for 6 hours, there was only the silicon nitride Fig.2(c) and there

were two forms ; α-silicon nitride and β-silicon nitride mixing in the sample.


PF‐21

(a)

(b)

(c)

Fig.2 The XRD pattern of SiO2+C was calcined in pure nitrogen gas 99.9% with the soaking time for 4 hours (a), 5 hours (b) and 6 hours(c)

Fig.3 The SEM morphology of SiO2+C compound was calcined in pure nitrogen gas

99.9% with the soaking time for 4 hours

2- Cristobalite low 1

2 2

3 1-Si3N4 3-Si2N2O2

1 11

1-Si3N4 1

1 1

1

1


PF‐21

Fig.4 The SEM morphology of SiO2+C compoundswas calcined in pure nitrogen gas 99.9% withthe soaking time for 5 hours.

Fig.5 The SEM morphology of SiO2+C compoundswas calcined in pure nitrogen gas 99.9% withthe soaking time for 6 hours.

Discussion

The hydrochloric concentrate of 3 M is enough to remove all the impurity in rice husk ash.

These impurities were easily leaching by concentrate acid and then washed the sample

with distilled water several times. After that the rice husk was calcined at 850 °C (carbon

was thermal decomposed at 550 °C) Finally, the pure silica rice husk was obtained.

From the XRD pattern, it shows if the silica rice husk ash mixed with carbon at the

ratio of 20 : 12 wt.% which is calcined at 1400 °C for 4 hours, silicon nitride is not

obtained. But there is a form of silicon nitride and silicon oxide nitride impure agglutinated

in the sample. If the sample is calcined for 6 hours, there is only the silicon nitride with

two forms; α-silicon nitride and β-silicon nitride mixing in the sample. The calcined time

of 6 hours was the optimal condition for preparing the silicon nitride.


PF‐21

The SEM morphology Fig.3-5 at 5,000 magnifications, it shows that the surfaces of

calcined sample of 4 hours became porous which is weaving into a network across the

sample but in some parts there were small sphere affiliated within a long stripe. When

increasing the calcined time upto 5 hours, the there was less porous in the material of the

surface then it was conglomerated into a clump and there was a little bit fiber. But if we

increased the calcined time up to 6 hours there was the least porous in the material and

there was the whisker which was orthogonal with the surface of sample Conclusions

The hydrochloric concentrate of 3M can well remove the impurities in rice husk and

results in the purity of product more than 99 %. The silicon nitride was formed when

calcined at temperature of 1400 °C more than 6 hours continuously which results in α-

silicon nitride and β-silicon nitride mixing in the sample. The porosity of samples were

decrease when the calcined temperature was increase.

References

[1] Dianying C., Baolin Z., Hanrui Z. and Wenlan L.

(2003) “Combustion synthesis of network silicon

nitride porous ceramics”, Cera Inter 29 pp. 363–364

[2] Grechikhin L. I. and Golubtsova E. S.(2004) :

‘Silicon-Nitride-Based Nanoceramic Materials’, Inor. Mater., 41, pp. 140-127

[3] Lee J.S., Muna J. H., Hanb B. D.,Kimb H. D., Shin

B.C.,and Kim S.(2004) “Effect of raw-Si particle

size on the properties of sintered reaction-bonded

silicon nitride” Cera Inter., 30 pp. 965–976

full papers thailand metallurgy conference

Documents