Web viewmicrostructure . and wear properties. The result revealed that the oil soak and water soak increase as the particle size of the agro-waste materials increases; but the hardness,

DEVELOPMENT AND CHARACTERIZATION OF ASBESTOS FREE BRAKE PAD USING COCONUT SHELL AS A BASE MATERIAL

A.D. Omah, U. U. Orji, I.C. Ezema, P.O. Offor, V.I. Onwuegbuchulam

Department of Metallurgical and Materials Engineering, University of Nigeria Nsukka

ABSTRACTAutomotive brake pad was developed using coconut shell particles as an alternative to asbestos brake pad which has been found to be dangerous to health due to its carcinogenic nature. The coconut shell particles were screened to particle sizes of 300μm and 7100μm. The resin was blended with the agro-waste materials in appropriate ratio. The properties of the brake pads examined include: physical, mechanical, thermal, microstructure and wear properties. The result revealed that the oil soak and water soak increase as the particle size of the agro-waste materials increases; but the hardness, impact strength and wear rate decreases as the particle size increases. The samples containing smaller particle size (i.e. 300μm) gave the better properties in all. The results obtained showed that coconut shell base brake pad can effectively replace asbestos and palm kernel base brake pad in terms of higher thermal efficiency, water absorption rate and low wear rate.

(Keywords: Coconut shell, agro-waste, XRD, TGA, DTA)

1.0 INTRODUCTIONBrake pads generally consist of asbestos fibers embedded in polymeric matrix along with several other ingredients. [1]. Asbestos was widely used in pads for its heat resistance. Inspite of its good properties asbestos is being withdrawn from all application where there is possibility of man consuming or inhaling its dust, because of its carcinogenic nature. Due to this health risk, it is necessary to use alternative material for making non-carcinogenic brake pad [4]. Due to carcinogenic nature of asbestos, a mixture of alternative fibers such as mineral fibers, cellulose, aramid, PAN, chopped glass, steel and copper fibers has now replaced the asbestos in the production of brake pads [4].Brake materials have additional requirements, like resistance to corrosion, light weight, long life, low noise, stable friction, low wear rate, and acceptable cost versus performance. There are two common types of friction brakes: drum/shoe brakes and disk/pad brakes. The design of the brakes affects heat flow, reliability and noise characteristics [2] Existing uses of asbestos are still permitted, while new applications and uses (of asbestos) are banned [5]. Most of the brake pad industry is gradually moving away from asbestos brake pad. The need to develop a new material for asbestos replacement as

friction material and yet maintaining the same mechanical properties, still remains a bone of contention. However, several researches have been carried out in the area of development of asbestos-free brake pads. The use of banana peels, palm kernel shell (PKS) etc. has been investigated [5, 6]. Researches all over the world are focusing on ways of utilizing either industrial or agricultural wastes as a source of raw materials in the industry. These wastes utilization will not only be economical, but may also result in foreign exchange earnings and environmental control. The purpose of this study therefore is to develop new asbestos-free brake pads using agricultural waste like coconut shell. Moreso, coconut shell is readily available and very cheap to obtain.

1.1 Coconut Shell as Brake Lining Ingredient.

The term coconut can refer to the entire coconut palm. It is derived from the Portuguese and Spanish word “coco” meaning “head” or “skull”. Coconut shell is the hard stony endocarps that surround the coconut, and the shells come in shapes and sizes; they are light and naturally sized as shown in Fig. 1 below.

1

JMME. Vol. 10 September 2015 ISSN: 2006-1919pp. 1-8

Coconut shell is recovered as a by-product in coconut production. After extracting the coconut from its shell, the residual nuts are further mechanically crushed to extract the seeds or nuts. The cracked shells are called coconut shell, a virgin biomass with a high calorific value of 5,500cal/kg, compared to that of palm kernel shell which is 3,800kcal/kg-ASTMD 5865. Large quantity of coconut shell is generated annually and only some fractions of it are used for fuel and other domestic and industrial applications, and also in the production of activated carbon. The unused coconut shells are dumped around the processing mill, constituting environmental and economic liability for the mill and the nation at large.

Fig. 1. Cracked coconut shell.

2.0 MATERIALS AND METHODS2.1 EquipmentThe materials and equipment used during the course of this work are: milling machine (Techno Gx160), Drilling machine (THRIGE-TITAN, MOTOR 3N50HZ/DS5002EX, NO 1.6-531699, MADE IN DENMARK), Brinell hardness testing machine (Type DVRB-M, serial no 0109, made in England), Impact testing machine (EXT94064/6705CE), Compressive testing machine (model no 317E-FA, S/N 0100005), Wear resistance

testing machine (No 4970203, Struers Copenhagen/Denmark), Other equipment include grinding/filing machine, electric oven, beam balance, micrometer screw gauge, hand knife, cylindrical dish, hot plate, engine oil (SAE 20/50), water, brake pad mould, digital weighing balance, sieve, honsfield tensometer, scanning electron microscope(SEM), polyester resin and coconut shell with compositional analysis and photo as shown in Table 1 and Figure 2.

Table 1. Compositional analysis of coconut shell

Element Weight% Atomic%

C K 60.99 71.51O K 29.69 26.14Fe L 9.32 2.35

Totals 100.00

2.2 Choice 0f MaterialAlthough, other agro-products similar to coconut shell, like palm kernel shell and cashew nut shell exist, we choose coconut shell because it is one agro-waste material that is found in abundance in most parts of Nigeria. However, Cameroon coconut shell was used because it is harder than Nigerian coconut shell and the shell contains less fibrous material than Nigerian coconut shell. The results obtained from the characterization of the thermo-mechanical performance of coconut shell particles led to the use of coconut shell particles as a candidate material for the replacement of asbestos in break pad production. The characterization of coconut shell particles was investigated through X-ray diffractometer (XRD), and thermogravimetric analysis (TGA/DTA).XRD: X-ray diffractometer (XRD) analysis of the coconut particles was carried out to determine the various element and phase distributions in the particles. (See fig. 4.) The test was carried out on a Philips X-ray diffractometer. The X-ray diffractograms were taken using CuKα radiation at scan speed of 3°/min. The sample was rotated at precisely one – half of the angular speed of the receiving slit, so that a contact angle

2 J. Metallurgy and Materials Engineering, Vol. 10 (2015)

between the incident and reflected beam was maintained. The receiving slit was maintained in front of the counter tube arm, and behind it is usually fitted a scatter slit to ensure that the counter received radiation only from the portion of the specimen illuminated by the primary beam. The intensity diffracted at the various angles was recorded automatically on a chart and the appropriate (θ) and (d) values were obtained. TGA/DTA: Thermal decomposition was observed in terms of global mass loss by using a TA Instrument TGA Q50 thermogravimetric analyzer. This apparatus detects the mass loss with a resolution of 0.1 as a function of temperature. The sample was placed in an open sample pan of 6.4 mm diameter and 3.2 mm deep. The temperature change was controlled from room temperature (30±5°C) to 900°C at a heating rate of 10°C/min. The sampling segment was set at 0.5 second per point. A high purity Argon was continuously passed into the furnace at a flow rate of 60 ml/min at room temperature and atmospheric pressure. This was done to purge the furnace for 30 min to establish an inert environment in order to prevent any unwanted oxidative decomposition. The TG and DTA curves that were obtained from TGA runs were carefully smoothed at a smoothing region width of 0.2°C by using least squares smoothing method, and analyzed by using Universal Analysis 2000 software from TA Instruments.

3.0 METHODS3.1 Raw Material Preparation. A certain quantity of coconut shell was obtained, cleaned and sun dried to remove the fibrous materials from the shell and other extraneous materials.The shell was ground in a conventional milling machine (Techno Gx160) at Ikpa market, Nru Nsukka, Enugu State, Nigeria and the product was graded using sets of 300µm and 710µm standard sieve (Endeccot Ltd, London);

Fig. 2. Pulverized coconut shell

3.2 Development of Experimental Brake Pad

Production of brake pad consists a series of unit operations including mixing, casting, grinding and finishing [10].

MixingThe constituent ingredients, such as resin, pulverized coconut shell, Aluminum oxide, carbon black, graphite, accelerator, catalyst etc. with their respective quantities in grams as shown in table2, were blended until a homogenous mixture was obtained.

Table 2. The constituent ingredient for the formulation of coconut brake pad.S/No Constituent Ingredient Quantities (g) 1 Coconut Shell 240 2 Resin 120 3 Graphite 32 4 Carbon Black 8 5 Aluminum oxide 10 6 Accelerator 2 7 Catalyst 1

CastingBack plates made of iron were inserted in each of the brake pad moulds also made of iron, the mixture was then deposited on the back plates as shown in fig. 3 below and allowed to cure. The brake pads produced were then taken to a filing/grinding machine where the brake pads produced were filed until both surfaces became smooth and desired thickness obtained.

Development of Asbestos-free Brake Pad Using Coconut Shell as Base 3


Fig. 3. Coconut Shell-Based Brake Pad

4.0 DESCRIPTION OF COCONUT SHELL-BASED BRAKE PAD

The disc brake pad is of 0.7cm thickness, and curved radius of 6.9cm and length of 12cm bounded on the back plate of length 13.4cm made of iron. The back plate has a curve radius of 6.7cm, also at the back of the back plate. There are slots that help to locate the pad in correct position in relation to the disc rotor.

5.0 PHYSICO-THERMAL PROPERTIES OF PULVERIZED COCONUT SHELL

Relative density of pulverized coconut shell.The density of the coconut particles was determined by measuring the mass and volume of the sample. The sample was weighed accurately in air using a laboratory balance and then suspended in water. The weight of the sample when suspended in water was determined, the volume of the sample was determined from the effect of displacement by water (Archimedean principle). The density of the sample was then estimated from the equation below:

Density =

massvolume (1)

Thermal Conductivity of Pulverized Coconut Shell. The thermal conductivity of the samples was calculated based on the Fourier’s law of heat transfer which is given as:

K=

ΔQΔt

× LA×ΔT (2)

Where k = Thermal conductivity, ΔQ = the quantity of heat transmitted, Δt = Time, L = Thickness, A = Surface area, ΔT = Temperature difference.

Hardness Survey TestThe resistance of the composite to indentation was studied using a Rockwell hardness testing machine (model-DVRB-M-0109). Based on ASTM, specifications, a 10mm diameter still ball was used and the load W, applied was kept stable at 250kg and

their Rockwell hardness numbers (HRC) were obtained.

Swell Growth AnalysisDimensional stability of the composite when subjected to changes in temperature and humidity was quantified by measuring its percentage swell growth (Khan et al., 2006; Ayruluris, 2008). The percentages of the swell were computed as:

Sp=(Tc−Tbo)100 %Tbo

Where Sp= percentage swell growth, Tbo = thickness of the sample before being introduced into an oven, Tc = thickness of the samples withdrawn from the oven

Thermal Gravimetric AnalysisThe change in weight of the frictional composite with increase in temperature was used to evaluate the thermal integrity of the brake pads.

Wear Test and characteristics of Frictional Composite The wear rate of the friction composite was evaluated; the test specimen was held against rotating disc covered with 220grit rough paper and was given a rotating speed of 600rpm for duration of 3 minutes interval for each experiment run.

6.0 RESULTS AND DISCUSSIONThe XRD pattern (Figure 4) obtained revealed that, the major diffraction peaks were: 12.94, 17.33, 25.84, 34.75, 40.46, 58.68, 65.74, 74.30 and 78.22 and their inter-planar distance were: 7.95, 5.94, 4.00, 3.00, 2.59, 1.83, 1.65, 1.48 and 1.42Å. Phases at these peaks were: Poly (4-octamethylenyloxybenzoic acid methylsiloxane trans-4-ethoxy-4'-stilbazole), Silicon Oxide, Chromium Oxide, 5-Cholesten-3$GB-ol 3-propionate, Magnetite, Silicon Oxide, Cristobalite $GB, syn and each of these phases have a score of 53, 26, 28, 18, 11, 10 and 25, (Table 1 and Figure 3). From this analysis, it is seen that the coconut particles have similar characteristics with other agro-waste particles that can be used as a replacement for asbestos.

(3)

Development of Asbestos-free Brake Pad Using Coconut Shell as Base

A mineralogical analysis carried out by the X-ray diffractometer revealed that the coconut particle contains each of these elements = "H, C, N, O, Al, Si, S, Cr, Fe" and none of these elements = "He, Li, Be, B, F, Ne, Na, Mg, P, Cl, Ar, K, Ca, Sc, Ti, V, Mn, Co, Ni, Cu, Zn, Ga, Ge, As, Se, Br, Kr, Rb, Sr, Y, Zr, Nb, Mo, Tc, Ru, Rh, Pd, Ag, Cd, In, Sn, Sb, Te, I, Xe, Cs, Ba, La, Ce, Pr, Nd, Pm, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, Hf, Ta, W, Re, Os, Ir, Pt, Au, Hg, Tl, Pb, Bi, Po, At, Rn, Fr, Ra, Ac, Th, Pa, U, Np, Pu, Am, Cm, Bk, Cf, Es, Fm, Md, No, Lr, D, T". This implies that with the absence of all these other elements, the coconut particles do not contain harmful radioactive elements that could be cancerous like the conventional asbestos. The TGA and the DTA curves showed three stages each: the TGA curve shows three weight loss steps, while the DTA also shows three stages of thermal decomposition. The temperature of destruction (Tdes) of the coconut particle was determined from DTA curves. DTA data were recorded on “Derivatograph OD 102”, at heating rate of 200C/min in argon. The result of the DTA/TGA scan of the coconut particle is shown in Figure 5.

For the TGA curve, the first step of weight loss occurred between 300C and 2500C. There was an initial weight loss at this stage which was due to the vaporization of water from the coconut particle. The degradation of the particle commenced precisely after 2500C; there was a sharp decrease in the

thermal stability of the particle between 2700C and 3900C. This is the second phase. The thermal stability after this temperature maintained a gradual decrease i.e. from 4000C until final degradation occurred at precisely 8900C. This is the third stage. The DTA showed three endothermic peaks which are 850C, 2900C and maximum thermal decomposition of 3500C respectively. It was observed from the DTA curve that 4000C - 5500C is the temperature range of maximum decomposition/destruction of the particle i.e. the total burning/degradation of the residual particle took place in the above temperature interval. The vapourization of water took place at the first endothermic peak, the burning of cellulose and volatile matter took place at the second endothermic peak, while at the last peak, the maximum thermal decomposition occurred. Thus, the endothermic effects on the coconut particles are seen in three progressions: dehydrogenation, evaporation of some cellulosic materials and the thermal decomposition of the material. The overall effect is seen in the decreased mass of the sample.

Position [°2Theta]

10 20 30 40 50 60 70 80 90

Counts

0

500

1000

1500

[ C10

.41 H

15.57

N0.2

9 O2.1

6 Si ]n

C30 H

50 O

2

Si O2

Si O2

; C30

H50

O2;

Fe3 O

4

Cr O

Si O2

Si O2

Fe3 O

4; Si

O2

Si O2

; Si O

2

Clay powder-5-ver3.raw

Fig. 4. XRD Spectrum of Coconut particles.

It can be seen that coconut shell particles have high heat resistance which is one of the good properties that led to the use of asbestos for break pad production.

Table3: Physicothermal properties of coconut based break pad.

In comparison, the lower thermal conductivity of 300µm particle size of coconut shell implies that the particles are closely packed and it will reduce the amount of heat that will be transmitted into the brake pad when subjected to a high temperature.High thermal conductivity is however undesirable in brake pads. Thus Graphite powder known to be a good lubricant [7] has been added to improve coconut shell particle for brake pad applications.

Fig. 6. Oil absorbtion capacity

Fig. 7. Water absorbtion capacity

Thermal integrity The mass loss of coconut shell (300µm and 710µm particle size) based friction composites is shown in fig. 8. In both materials, the mass losses were comparably low up to about 300ºC. The result in fig. 8 shows that there is no significant difference in weight loss of the two composites over the range of temperature used. However, the 710µm particle size brake pad showed more weight loss w.r.t. temperature than that of 300µm particle size brake pad which implies that it will perform with lower efficiency than the 300µm particle size brake pad in high temperature applications. Typical automotive brake pad materials however are rarely subjected to temperature range higher than 389ºC [9]. Therefore, it is believed that the coconut shell based brake pad will not degenerate under practical application of temperature and time duration.

Fig. 5. DTA/TGA Pattern of coconut particles.


PARTICLE SIZE(µm )

RelativeDensity

ThermalConductivity

PercentageSwells Growth

300µm(g) 1.45 5.71 w/mk 0.11%710µm(g) 1.94 12.05 w/mk 0.42

Fig. 8.Weight loss of friction composite due to increase in temperature

Table 4. Mechanical Properties of the Composite Material.S/N PARTICLE

SIZE(µm )IMPACT STRENGTH (J)

HARDNESS HR(30kgs)

COMPRESSIVESTRENGTH(N/ mm²)

1 300 3.0 714.5 2.252 710 2.0 694.4 0.23

Table 5. Wear resistance test.S/N MATERIALS(µm) WEIGHT

LOSS(g)WEAR INDEX

1 300 0.9 1.52 710 1.3 2.2

7.0 CONCLUSION 1. Coconut shell can be effectively used in

brake pad formulation when properly combined with other addictives.

2. Coconut shell based brake pad was found to be thermally resilient enough not to decompose at typical braking temperatures and duration.

3. Coconut shell based brake pad will require more lubricant for enhanced performance.

4. Coconut shell based brake pad exhibits lower wear rate without degrading the surface of the pad. Also, 300µm particle size of coconut shell shows greater toughness, hardness and compressive strength but has a lower wear rate when compared to 710µm particle size of coconut shell.

5. The lower thermal conductivity of 300µm particle size of coconut shell implies that the particles are closely packed and it will reduce the amount of heat that will be transmitted into the brake pad when subjected to a high temperature.

REFERENCES1. Aigbodion. V., Akande. U. and Hasssn, S. B.

Asuke, F and Agunsoye, J. O (2010), Development of Asbestos- Free Brake Pad Using Bagasse. Journal of Trbology in Industry, Vol. 32. No.1, 2010.

2. Aigbodion. V.S. and J.O.Agunsoye (2010); Bagasse(Sugarcane waste): Non-Asbestos Free Brake Pad Materials, LAP Lambert Academic Publishing, Germany, ISBN 978- 3-8433-8194-9.

3. Chapman T. R., D. E. Niesz, R. T. Fox, and T. Fawcett (1999) “Wear-resistant Aluminum-Boron

cermet’s for automotive Brake Applications,” Wear, 236, pp. 81-87.

4. Dagwa, I.M. and Ibhadode, A.O.A. (2005). Design and Manufacture of Experimental Brake Pad Test Rig Nigerian Journal of Engineering Research and Development , Basade Publishing Press Ondo, Nigeria, Vol. 4, No. 3. 15-24.

5. Blau, P. J (2001). Compositions, Functions and Testing of Friction Brake Materials and their Additives. Being a report by Oak Ridge National Laboratory for U.S Dept. of Energy. www.Ornl.gov/-webworks/cppr/y2001/rpt/112956.pdf

6. Anderson .A. E. (1992) “Friction and Wear of Automotive Brakes,” in ASM Handbook, Friction Lubrication and Wear Technology, Volume 18, ASM International, Materials Park, Ohio, pp. 569-577.

Development of Asbestos-free Brake Pad Using Coconut Shell as Base 7

7. Gurunath P.V. and Bijwe J., (2009). Potential Exploration of Novel Green Resins As Binders for Non-Asbestos-Organic (NAO) Friction Composites in Severe Operating Condition. Wear Journal. Volume 267, Issues 5-8: 789-796.

8 Nicholson .G. (1995) Facts About Friction, P&W Price Enterprises, Inc., Croydon, PA.

9. Mohanty Smart, Chugh Y.P. Development of fly ash – based automotive brake lining. J Tribol Int. 2007; 40(7): 1217-24.

10. Gurunath PV, Bijwe J. “Potential exploration of novel green resins as binder for non-asbestos-organic (NAO) friction composites in severe operating condition.” Wear J 2009: 267(5-8): 789-96.


SYNTHESIS OF BIO- ETHANOL/GASOLINE BLENDS AND ITS EFFECT ON THE PERFORMANCE OF A SPARK IGNITION ENGINE

**Obi, A. I., **Amaghionyeodiwe, C. A., *Ogbeifun, S. and *Dauda, M

*Department of Mechanical Engineering, Ahmadu Bello University, Zaria**Department of Mechanical Engineering, Michael Okpara University of Agriculture, Umudike

ABSTRACTBio-ethanol produced from pineapple and pawpaw peels through biochemical reaction called fermentation by Aspergillus niger and Saccharomyces cerevisiac was investigated for its suitability in running petrol engines. Gasoline blended with 10-50% of the extracted ethanol at intervals of 10% was compared with pure gasoline. Some physical and chemical properties of viscosity, calorific value, specific gravity, octane number and flash point of the bio-ethanol gasoline blends were determined. The performance characteristics and exhaust emission test were also conducted using Petter Paiw single cylinder spark ignition engine and IMR 1400 gas analyzer on the blended fuel samples. It was found that higher fuel consumption and reduced volumetric efficiency were observed with increase in the blended fuel compared with the reference fuel and a marginal increase in brake thermal efficiency and reduction in the exhaust gases of CO, NOx, SOx and HC.

1.0 INTRODUCTIONThe combustion of petroleum products causes environmental pollution and the emission of greenhouse gases generally believed to be responsible for global warming. The Internal Combustion (IC) Engine (a common automobile engine used in Nigeria and other developing countries) is an engine that converts chemical energy in a fuel into mechanical energy, usually made available on a rotating output shaft. ICEs generate undesirable emissions during the combustion process. The emissions discharged into the surroundings pollute the atmosphere and cause the following problems: global warming, acid rain, smog, odors, respiratory and other health hazards. The causes of these emissions are non-stoichiometric combustion, dissociation of nitrogen and impurities in the fuel and air, releasing dangerous gases like: unburnt hydrocarbons (HC), oxides of carbon (COx), oxides of nitrogen (NOx) oxides of sulphur (SOx) and solid carbon particulates into the atmosphere. The trend now therefore is to reduce over dependence on petroleum products in order to cut down the effects of global warming. Other possible advantages of phasing out petroleum products are the presence of sources available from renewable resources which do not require the

use of heavy equipment for extraction, long distance transportation and distribution.

Ethanol blended with gasoline and diesel in various fractions is already in use in many countries. An E10 blend represents 10% bio-ethanol in 90% gasoline while an E20 blend represents 20% bio-ethanol in 80% gasoline. Research has shown that alcohol based fuel is of high quality, low cost and has exceptional engine performance (RFA, 2001). Ethanol blended fuels account for approximately 30% of all automotive fuels sold in the U.S. and ethanol acts as antifreeze in the engine during winter (RFA, 2001). This high quality and high-octane fuel is capable of reducing air pollution and improving automobile performance (Wei-Dong et al., 2002; Al-Hasan, 2003). Road transportation is the dominant means of moving goods and services in developing countries like Nigeria and the demand for road transportation fuel will continue to rise due to increasing population, urbanization, industrialization and socialization, the result being increased environmental problems/ pollution. In 1980, Nigeria emitted 18.9 million metric tons of carbon and since then, carbon emission has been on the increase with 23.5 million metric tons and 27.7 million metric tons emitted in 2001 and 1996 respectively (DPR Nigeria, 2005), hence the


9

necessity to source friendly alternatives. Part of the objectives of this research work includes the extraction of bio-ethanol from pineapple and pawpaw peels, production of E10, E20, E30, E40 and E50 bio-ethanol blends, determination of some Physical and Chemical properties of the produced blends, performance evaluation of the bio-ethanol blends produced on a Spark Ignition (SI) engine and exhaust emission analysis of E10, E20, E30, E40 and E50 blends driven SI engine.

2.0 MATERIALS AND METHOD2.1 Procedure for Extraction of Bio-ethanol from Pineapple and Pawpaw PeelsRipe pineapple and pawpaw obtained from the local Samaru market, were purchased and the peels were collected and taken to the Microbiology Laboratory of Ahmadu Bello University Zaria, where the peels were washed and their outer coat removed, cut into smaller pieces using a knife and then blended with sterile distilled water using electric blender, and stored in a refrigerator prior to use.

2.1.1 Microorganism CulturingPure culture strain of Aspergilus niger and Saccharomyces cerevisiae were isolated and used throughout this study. These organisms were maintained as direct stocks culture from which inoculates were prepared. Fungal spieces of A. niger and S.cerevesiae were originally isolated from soil samples and palm wine respectively, the slant cultures were sub-cultured and grown on potatoes dextrose agar (PDA) in petri dished according to manufacturer’s specification, and sterilized at 1210C for 15mins, Samples were then incubated at room temperature for 5 days.

2.1.2 Preparation of Growth MediumThe growth medium used for preparing the Aspergillus niger inoculum (obtained from garden soil) consisted of 150g of the blended pineapple and pawpaw peels (substrate), peptone, 0.1%; malt extract, 0.1% (w/v), yeast extract, 0.2% (w/v), calcium carbonate 0.2% (w/v), and ferrous sulphate respectively. Saccharomyces cerevisiae (obtained from ripped pineapple peels)

growth medium was prepared using yeast-malt broth at pH 5.5 (Abouzeid and Reddy 2006).

2.1.3 Preparation of Inocula and Fermentation Procedure

Aspergillus niger inoculum was prepared in 250cm3 cotton- plugged conical flask containing 100cm3 of different substrates growth media. The flasks were sterilized and inoculated with 0.11 (OD) Aspergillus niger spores. Each of the flasks was incubated on an environment- controlled incubator shaker (Model 3527-1/34) shaker with agitation rate of 300rpm at 300 C for five days. Saccaharomyces Cerevisiae inoculum was prepared in the same way as the Aspergillus niger inoculum except that yeast malt broth was used. The growth medium was inoculated with 0.08 (OD) yeast cells and incubated for 24hours. The fermentation medium used for ethanol production was identical to the growth medium. Ethanol fermentation was carried out in 1000cm3

conical flasks each containing 300cm3 of medium. The medium was sterilized and inoculated with 5% (v/v) growth media containing Aspergillus niger and Saccaharomyces Cerevisiae and incubated on a shaker with an agitation rate of 300rpm at 300C for seven days.

2.2 Bio-ethanol Blends PreparationBio-ethanol generally has some lower physical and chemical properties compared to gasoline. Hence, for smooth and efficient performance of engines, the bio-ethanol produced was blended with gasoline in different ratios. The bio-ethanol blends were prepared using direct blending method and the blends produced include: E10, E20, E30, E40 and E50 which represent 10% bio-ethanol in 90% gasoline, 20% bio-ethanol in 80% gasoline, 30% bio-ethanol in 70% gasoline, 40% bio-ethanol in 60% gasoline and 50% bio- ethanol in 50% gasoline respectively. This was done by mixing 10ml, 20ml, 30ml, 40ml and 50ml of the bio-ethanol produced with 90ml, 80ml, 70ml, 60ml and 50ml of gasoline respectively in a transparent bottle.

2.3 Determination of the Physical and


Chemical Properties of the Blended FuelThe properties of the bio-ethanol blends tested include: Calorific value, Viscosity, Specific Gravity, Octane number, Flash point and Density.

2.3.1 The Calorific ValueThe calorific values of the bio-ethanol and gasoline blends were measured using a bomb calorimeter. A known amount of fuel was placed in the crucible. The crucible was then placed over a ring and a fine magnesium wire touching the fuel sample was stretched across the electrodes. The lid was tightly screwed and the bomb was filled with oxygen up to 25 atmosphere pressure. The initial temperature was recorded. The electrode was then connected to a 6 V battery and the circuit was completed. As soon as the circuit was completed and current was switched on, the fuel in the crucible burnt with the evolution of heat. Heat liberated by burning of the fuel increases the temperature of water and the maximum temperature attained was recorded.

2.3.2 The ViscosityThe viscosity of the samples was determined using a glass capillary kinematic viscometer at 40 0C (Sivaramakrishnan and Ravikumar 2011). The viscometer was tightly clamped on a retort stand. 100g of each sample was collected into a Pyrex beaker and was gradually heated to a temperature above 40⁰C. The sample was then transferred into the viscometer through the larger opening of the capillary tube and the fluid was allowed to cool until a temperature of 40⁰C was reached. Thereafter, suction was applied to the other end of the capillary tube to draw the fluid to the mark on the upper meniscus level of the capillary tube.The fluid was allowed to run freely to the lower meniscus mark in the capillary tube. The efflux time for the fluid to flow from the upper meniscus mark to the lower meniscus mark was determined with the aid of a stopwatch. The test was triplicated for each sample and the kinematic viscosity was calculated using equation (2.1).

Kinematic Viscosity = kt (2.1)

Where k = Constant of the viscometer expressed in mm2/s2

t = Flow time in seconds of the liquid.

2.3.3 Test for Specific GravityThe specific gravity of the samples was measured at room temperature using a Fisher brand hydrometer (size 0.795-0.910, accuracy 0.001). The measurement was performed according to the method adopted by (Coronado et al., 2009).

2.3.4 Test for Octane numberThe octane number of the blended fuels was determined to ensure that the produced bio-ethanol and gasoline blends have octane numbers complying with octane numbers of motor fuel, as calculated by research method (ASTM D2699) and motor method (ASTM D2700) standards. This was achieved using OCTANE-IM portable octane meter used in refineries to control quality of fuel components and their blends, fuel checking during transportation, storage and consumption.

2.3.5 Test for Flash PointFlash points of the samples were determined using the setup apparatus comprising an electric heater, beaker and thermometer. 80ml of each sample was introduced into a transparent Pyrex beaker placed on an electric heater. The beaker was fitted with the thermometer, clamped on a retort stand. Heat was applied gradually by turning the knob of the electric heater until the observed movement of the particles increased. A flame was constantly brought near the surface of the beaker until “a catch and disappearing” of flame on the surface of the hot liquid occurred. The temperature was noted as the flash point.

2.3.6 DensityThe densities of the blends produced were determined at ambient temperature (28⁰C). A density bottle of mass 50ml was weighed on the analytical balance and the initial weight of the bottle was noted. The samples were then put in the density bottle, the spillage was cleaned and dried, and the bottle was weighed on the analytical balance. The

Synthesis of Bio-ethanol/Gasoline Blends for Spark Ignition Engine 11 11

w2 – w1 vD = (2.2)

WN 5000BP = (2.4)

BP LANK

BMEP = (2.5)

process was repeated twice and the average value was determined as the result. Density of each sample was then computed using equation (2.2) below.

Where D = Density in g/cm³W₂ = Weight of bottle and sample (g)W₁ = Weight of bottle only (g)V = Volume of the Liquid (cm³)

2.4 Experimental Setup of Peter Paiw Spark Ignition Engine for Performance Evaluation of the Produced Fuel Blends

The engine performance and exhaust emissions tests were carried out on a Peters PAIW carburetted single cylinder four-stroke research SI engine. This was done by connecting a four stroke single cylinder petrol engine to the electric dynamometer with the help of coupling. Tachometer for speed reading rpm, U tube manometer, air filter, fuel measuring tube and gas analyzer were arranged as shown in Figure 2.1.

Figure 2.1: Experimental setup of Petters Engine

2.4.1 Determination of Engine Torque and Brake Power

Torque is the engine’s ability to do work. The torque applied by the engine on the dynamometer, T was determined using equation (2.3).

T = W x R (2.3)where,W= LoadR= Torque arm lengthBrake power is the useful power at the

output shaft. For different engine loads. brake power increases with the increase of engine speed due to increase of engine friction. The brake power PB, delivered by the engine and absorbed by the dynamometer was determined using equation (2.4).

where,W = Load readingN = Speed

2.4.2 Determination of Brake Mean Effective Pressure

Mean effective pressure (Pm) is that hypothetical constant pressure which is assumed to be acting on the piston during its expansion stroke producing the same work output as that from the actual cycle. The brake mean effective pressure (BMEP) of an engine is the average (mean) pressure which, if imposed on the pistons uniformly from the top to the bottom of each power stroke, would produce the measure (brake) power output. BMEP is calculated from equation (2.5).

where, Bp = Brake PowerL = Length of Stroke A = Area of the piston (m2)N = Speed in R.P.M and K = Number of Cylinders

2.4.3 Determination of Volumetric Efficiency

Volumetric efficiency (VE) of an engine is the actual amount of air the engine ingests compared to the theoretical maximum. VE of the SI engine was determined from equations (2.6) – (2.10).

(V E) = V a+V f

V s (2.6)

where:

V a = Volume of air = M a R T a

P (2.7)

M a= Mass of air

M a= 0.866 √ P hT a

(2.8)

h = Manometer reading = H Sin θ


θ = 150

R = Gas Constant = 287

V f = Volume of sample

Rate of consumption (2.9)

V S = V C Nn (2.10)V C = Cylinder Swept VolumeNn = Number of cylinders

2.5 NHA- 506EN Automotive Emission Analyzer

This equipment was used to investigate emission products directly from the combustion chamber. The NHA- 506EN gas analyzer is a combustion gas analyzer designed to work under strict adherence to the operating manual and within stipulated temperature and was used for this work. It measures and calculates the Flue Gas Temperature, Excess Air, Carbon dioxide (CO2), Carbon monoxide (CO) (Corrected to 0% O2), Nitrous oxide (NOx) (Corrected to 0% O2), Combustion efficiency and Heat loss.

3.0 RESULTS AND DISCUSSION3.1 Result of Physical and Chemical

Properties test of the Produced Fuel Blends

Results of the calorific value, viscosity, specific gravity, Octane number, Flash point and density of the bio-ethanol blends and the unblended gasoline sample (designated E0) obtained are tabulated in Table 3.1 belowFrom Table 3.1, the calorific values of the blends decrease with an increase of bioethanol in the blends except for E0 and E10 which have the same calorific value; this is because, with 10% of bio-ethanol in the blend, the quantity of bio-ethanol in the blend is not enough to alter the calorific value of the blend. From the result of the viscosity, the unblended gasoline gave a viscosity of 0.61mm2/ s at 400C. There was increase in viscosity with corresponding increase in bioethanol percentage in the blends. Hence, E10 has a viscosity of 0.65mm2/s, E20 has 0.68mm2/s, E30 has 0.69mm2/s, E40 has 0.70mm2/s and E50 has 0.72mm2/s. The specific gravity of the blends increased from 0.7840 for E0 to 0.7952 with E50. This result affirms that bioethanol is heavier than gasoline. This explains why when a mixture of bio-ethanol gasoline is allowed to settle and viewed inside a transparent container; ethanol is seen to settle at the middle sandwiching gasoline on top and water at the bottom.

BlendsCalorific Value KJ/Kg

Viscosity@ 40 (mm2/ s)

Specific gravity (Kg/L)

Octane number FlashPoint (oC)

Density (Kg/m3)

E0 42932 0.61 0.7840 90 28.7 784.0 E10 42932 0.65 0.7852 92 29.1 785.2 E20 41510 0.68 0.7896 93 29.2 789.6

E30 40341 0.69 0.7907 94 29.4 790.7 E40 39450 0.70 0.7924 96 29.6 792.4 E50 37872 0.72 0.7952 97 29.8 795.2

Table 3.1 Calorific value, viscosity, specific gravity, octane number, flash point and density of blends of gasoline and bioethanol produced from pineapple and pawpaw peels.

800 1000 1200 1400 1600 1800 20000

2

4

6

8

10

12

14

E0%E10%E20%E30%E40%E50%

Speed (rpm)

Bre

ak

po

wer

(K

W)

Synthesis of Bio-ethanol/Gasoline Blends for Spark Ignition Engine 13

Figure 3.1: Graph of Break Power against Speed

Figure 3.2: Graph of Torque against Speed

Figure 3.3: Graph of Brake Mean Effective Pressure against Speed

The immiscibility of the blends with higher bio-ethanol fractions seems to be responsible for the unsteady engine production noticed during the tests. The octane number increases from 90 for E0 (100% gasoline) to 97 for E50. Octane number of a fuel indicates its ability to resist pre-ignition and burn evenly. The flash point increases from 28.7 for E0 to 29.8 for E50. The flash point indicates the temperature at which the fuel can vaporize to produce an ignitable mixture with air.The flash point gives some indication on the flammability of the liquid showing that with the addition of ethanol fractions to gasoline the burning characteristics of the mixture reduce. It is important to note that the higher the flash point of a fuel the more difficult it is to start the engine. For example, the flash

point of E50 (29.80C) means it will be difficult to start an engine. Hence, as much as we desire to reduce the harmful emissions associated with combustion of gasoline, other factors such as flash point, octane number, etc should be considered in choosing the optimal blending ratio for the most efficient and safe engine operation.

3.2 Result of Performance Test of the Produced Blends on SI EngineThe brake power, torque, brake mean effective pressure and volumetric efficiency of the various blends were plotted against engine speed and the results presented in Figures 3.1 to 3.4 below.Figure 3.1 shows that brake power increases with increase in engine speed. This is due to increase of engine friction. At the highest speed of 1800rpm, E50 fuel blend developed

800 1000 1200 1400 1600 1800 20000

10

20

30

40

50

60

70

80

90

E0%E10%E20%E30%E40%E50%

Speed (rpm)

To

rqu

e (N

m)

800 1000 1200 1400 1600 1800 20000

2

4

6

8

10

12

E0%E10%E20%E30%E40%E50%

Speed (rpm)

Brea

k m

ea

n e

ffecti

ve p

ress

ure (

ba

r)


the highest power followed by E40, E30, E20 and E10 respectively while, gasoline (E0) developed the lowest brake work as shown in (Fig 3.1). Due to bio-ethanol addition to gasoline, the engine operation enabled leaning effects and improves engine combustion and more power is produced thereby, increasing engine performance as explained by (Al- Hasan, 2003).Figure 3.2 shows decrease of torque as speed increases. The maximum decreasing percentage in torque occurs for gasoline. This is the result of addition of alternative fuel which caused the octane number to rise. This effect is particularly important as it improves the anti-knock quality of the engine (Kumbhar and Patil, 2013).From Figure 3.3, the BMEP of the various blends are seen to increase with a corresponding increase in engine speed until the engine speed gets to 1400rpm where the BMEP starts decreasing. BMEP is 10.01% at 1200rpm for E30 while, minimum value of the BMEP is about 7.34% for E50 as against 6.25% compared to gasoline. At, low engine speeds the higher heating value of gasoline is responsible for high BMEP. This is in agreement with the result obtained by (Nyachaka, 2013).

Finally, Figure 3.4 shows that the value of volumetric efficiency increases with increase in the bio-ethanol content of the blends due to the lower heating value of the ethanol. This reduces the charge temperature of the intake manifold as confirmed by (Al- Hassan, 2003). However, the volumetric efficiency starts to decrease when bio-ethanol content is more than 30%.

3.3 Results of Exhaust Emission of SI Engine for the Various Fuel BlendsThe exhaust emission for each of the blends when tested on a SI engine (running at 1200rpm) were determined and the result shown in figures 3.5 and 3.6From figure 3.5, pure gasoline (E0) emits less CO2 compared to the blended fuel samples. Hence, the amount of CO2 emitted increases as the concentration of ethanol in the samples increases. This is because of the additional oxygen content of ethanol. Also as combustion becomes more efficient, CO gets converted into CO2 (Kumbhar and Patil, 2013). Consequently, CO emission continuously decreased with increase in the ethanol concentration of the blends (Mugal et al., 2012).

800 1000 1200 1400 1600 1800 20000

5

10

15

20

25

30

35

E0E10E20E30E40E50

Speed (rpm)

Vo

lum

etr

ic E

ffici

en

cy (

%)

Figure 3.4: Graph of Volumetric efficiency against Speed


Figure 3.5: Bar Chart of Exhaust emission against Bio-ethanol blends

Figure 3.6: Bar Chart of Exhaust emission against Bio-ethanol blends

The effect of bio-ethanol percentage in the HC emission is shown in (fig 3.6). In all the blends there is decrease in the HC emissions when compared to sole gasoline as a result of hydrocarbon oxidation. This is in agreement with (Laminu et al., 2014), that reduction of the carbon atoms concentration in the blended fuel, high molecular diffusivity and high flammability limits improve mixing process and hence there is better combustion.Generally, NOx formation is increased when an engine runs at its most efficient/hottest part of the cycle. However, it can be seen from figure 3.6 that addition of ethanol to gasoline decreases NOx emission. This is due to lower heating value of the blends compared to pure gasoline. This confirms the investigation made by (Ananda and Saranana, 2010), that lower heating value

decreases the combustion heat energy and lowers the combustion temperature in the cylinder. Hence, E50 blend releases the least amount of NOx (150ppm).

4.0 CONCLUSION AND RECOMMENDATIONS

4.1 ConclusionPineapple and Pawpaw peels were obtained washed and their outer coats removed and cut in small pieces using a knife. Ethanol was extracted from the peels through a biochemical reaction called fermentation. The bio-ethanol produced was blended with gasoline in different ratios to produce E10, E20, E30, E40 and E50 blends which represent 10% bio-ethanol in 90% gasoline, 20% bio-ethanol in 80% gasoline, 30% bio-ethanol in 70% gasoline, 40% bio-ethanol in

E0% E10% E20% E30% E40% E50%0

1

2

3

4

5

6

7

8

CO2(% vol)CO(% vol)

Blend Ratio

Ex

hau

st e

mis

sion

(%

vol)

E0% E10% E20% E30% E40% E50%0

100

200

300

400

500

600

HC (ppm)NOx (ppm)

Blend Ratio

Exh

aust

em

issi

on

(p

pm

)


60% gasoline and 50% bio- ethanol in 50% gasoline respectively.Some physical and chemical properties of the blends were determined and compared with those of pure gasoline. The samples were further subjected to engine test to determine their performance on a four stroke, single cylinder Peter Paiw SI engine. Also CO2, CO, HC and NOX exhaust emissions for each of the blends when tested on the SI engine (running at 1200rpm) were determined and compared with emissions from the engine when run with pure gasoline. The findings of this study show that pineapple peels and pawpaw peels are potential substrates which can be exploited in industries for bio- ethanol production on a commercial scale as they are cheap and more importantly renewable.

4.2 RecommendationsBased on the study the following recommendations are made:

Proper waste management committee should be set up which will religiously collect pineapple and pawpaw peels from consumers, sellers, farmers and fruit juice producers to store the spoilage and post-harvest losses and peelings for bioethanol production.

Further research is recommended to vary the different bioethanol substrates from pineapple and pawpaw separately on different engine parameters.

REFERENCESAbouzeid M.M and Reddy A, (2006) “Direct fermentation of potato starch to Ethanol by Co-Culture of Aspergillus niger and saccharomyces cerevisiac”. Journal of Applied Microbiology. Vol 52 Pp 1055- 1059.

Al – Hasan M, (2003) “Effect of ethanol – unleaned gasoline blends on engine performance and exhaust emission” Energy conversion and management vol.44,pp. 1547-1561.

Ananda C.S and Saravanan C.G (2010) “Emission reduction on ethanol gasoline blends using oxygenated additives” International Journal of Applied Engineering research vol 5.

Coronado M. Yaun W, Wang D and Dowell F.E (2009). “Predicting the concentration and specific gravity of biodiesel-diesel blends using near-infrared spectroscopy. Applied Eng. Agric., 25: 217-221.

Department of petroleum resources DPR, Nigeria 2005.Kumbhar V.S and Patil S.R (2013) “Experimental investigation of the effect of ethanol gasoline blends on performance and emission characteristics of the S.I engine” International conference on advance research in mechanical Engineering ISBN 978-93-82208-93-8.

Laminu S.K, David O.O, Ibrahim A.S, Jeremiah M and Zainab A.K (2014) “The impact of gasoline and synthesized ethanol blends on the emission of a spark Ignition engine”, World Journal of Engineering 11(4) 391-396.Mughal H.U, Bhutta M.M.A, Athar M, Shahid E.M and Ehsan M.S (2012) ‘The Alternative fuels for four stroke compression ignition engines performance analysis. J Transaction of Mech. Eng. 36-155-164.

Nyachaka C.J, Yawas D.S and Pam G.Y (2013) “ Production and performance evaluation of bioethanol fuel from groundnut shell waste”. American Journal of engineering Research (AJER) Vol.2,Issue -12, Pp 303-312

RFA (2001) Renewable fuels Association. http://www.ethanolrfa.org/exchange/2 consulted January 2001.

Sivaramakrishnan K., Ravikumar P (2011) “ Determination of higher heating value of biodiesels.” International Journal of Engineering Science and Technology (IJEST) 3(11)

Wei-Dong H, Rong- Hong C, Tsung-lin W, Ta-Hui L (2002). Engine performance and engine performance and pollution emission of an SI engine using ethanol – gasoline blended fuels. Atmos. Environ., 36(3): 403-410


A NEW COMBINATORIAL APPROACH TO PROPERTY-DRIVEN DESIGN OF Ti ALLOYS FOR BIOMEDICAL APPLICATIONS

Paul S. Nnamchi, Boniface A. Okorie, and Camillus S. Obayi

Department of Metallurgical and Materials Engineering,University of Nigeria, Nsukka, Nigeria.

Tel.: +2347064278906; +447928921079.E-mail addresses: [email protected]; [email protected].

ABSTRACTTi and Ti based alloys containing nontoxic elements, such as Nb, Mo, Ta, and Zr have been proposed as prospective candidates for biomedical applications, due to their excellent corrosion resistance, biocompatibility, high strength, toughness and wear resistance. However, stress shielding has become a major setback to their application. When elastically soft bone tissue (~E=20-40GPa) is replaced by a stiffer implant, the implant takes over a considerable amount of the load, shielding the surrounding parts of the skeleton. Reducing the physiological loads on the bone induces re-sorption processes that give rise to a drop in bone density, mineralization state and strength. Stress shielding can finally promote contact loosening, implant failure, or debris-induced infections. For this, the aim of this study is to develop a new biomaterial to use in the load transfer implant field. An ab initio theoretical calculation was used to couple elastic properties from homogenised multiphase elastic parameters for the design of new Ti-Mo-Nb-Zr alloys with bone matching modulus. The agreement between the predictions and detailed experimental characterization, sheds light on the decisive influence of the multi-phase character of the polycrystalline composites on their structural and mechanical properties. An attempt is also made to highlight the influence of heat treatment and cold work in enhancing the modulus of the alloys. The study shows that the novel combinatorial approach can be highly beneficial as it may lead to something of a breakthrough with respect to reducing the Young’s modulus of metallic biomaterials, which is pertinent to preventing stress shielding and bone resorption in orthopedic implants.

Keywords: Ti based alloys; Young’s modulus, phase stability; ab initio; elastic modulus; elastic constant; biomedical implant application.

1.0 INTRODUCTION Health care has become one of the front burner research fields of this century owing to the dramatic increase in the number of people affected by various diseases. Health care costs and the urgent requirement for biomaterials have placed enormous pressure on government funding agencies and researchers to develop cost effective, appropriate biomaterials to treat various diseases and to regenerate or replace dysfunctional tissues or organs [1]. Thus, various funding organization have allocated considerable funding for the development of the next generation of metallic and associated biomaterials.Amongst metallic biomaterials, titanium is generally considered as one of the most biocompatible and corrosion resistant metals available for biomedical or clinical applications. From 1951, when Leventhal [2] first published an article on the orthopedic application of this metal until now, there has

been a thrust towards new developments in the manufacturing and use of this metal and its varieties of alloys in orthopedic applications and other clinical related activities. Beginning with commercially pure Ti, and then to Ti-6Al-4V alloy, today’s state of the art β-Ti alloys have been added to the list. There is a range of materials and devices based on titanium, which are available for a variety of medical applications. These alloys are known to possess many attractive properties such as excellent corrosion resistance in biological environment, superior biocompatibility, high specific strength and wear resistance and provide adequate mechanical properties when compared to other metallic biomaterials [3.4]. However, stress shielding effect due to high elastic stiffness comparable to that of human bone has become a major setback. When elastically soft bone tissue is replaced by a stiffer implant, the implant takes over a considerable


18

mailto:[email protected]


amount of the load, shielding the surrounding parts of the skeleton. Reducing the physiological loads on the bone induces re-sorption processes that give rise to a drop in bone density, mineralization state and strength [25]. Stress shielding can finally promote contact loosening, implant failure, or debris-induced infections. Therefore, it is of great importance that the elastic mismatch between the bone replacement material and existing bone be minimized in metallic biomaterials intended for orthopaedic application.

Over the past few years, significant effort has been devoted to develop bone matching Ti-based alloys, and a few of them have already been implemented into biomedical applications [5]. They usually contain multiple non-toxic alloying elements, such as Mo, Nb, Ta, and Zr, which are preferred as β stabilizers or/and which play the key role in decreasing the Young’s modulus [6-7]. To design these multi-component alloys, several methods have been proposed like d-electrons concept [9] and Mo-equivalence method [10]. However, it is difficult to determine the optimum alloy composition with sufficiently metastable β phase and enough low modulus which is generally achieved by multiple alloying of simple binary alloys. To design novel Ti-based alloys with desirable properties, it is necessary to develop a theoretical framework in selecting suitable alloying elements.

In recent years, computational material design based on first principles quantum mechanical methods have emerged as an important cost effective tool to design material properties [11,12]. Biocompatible Ti alloys have been investigated experimentally, and by theoretical studies, but such efforts notwithstanding, a review of the literature indicates that multicomponent Ti-Mo alloys have received considerably little attention [13,14], even though the Ti-Mo system has proved to be a good substitute for developing absolutely safe Ni-free biomedical Ti alloys due to their being non-toxic and non-allergic elements. Nevertheless, the Young’s moduli of the Ti-Mo alloy systems reported so far are not low enough [15-17], and there has been

little research effort into the Young’s moduli of multicomponent metastable β-phase Ti-Mo alloys. Moreover, a central theoretical framework of low modulus determining factors in alloys remains elusive and our understanding is not nearly as sharp as it should be in the area. A relatively good understanding could be obtained by density functional theory, which has been proved to be a useful tool to reveal natural properties of many compounds from atomic scale considerations [18, 19].

Generally speaking, if the content of β phase stabilizing elements (e.g. Mo, Nb and Zr) doesn’t exceed a critical amount in the Ti-X binary alloys, three kinds of metastable phase transformations can occur depending on process techniques. The occurrence of β phase would depend on the competition amongst different phases. Therefore, in titanium alloy design for low elastic modulus, a consideration of phase homogenization is important as it can yield results close to experimental results. In this paper, the authors have attempted to calculate by ab initio method the integral elastic response of multi-phase polycrystalline aggregates based on multiple scattering in the Ti-6Mo-xNb-xZr alloys, with respect to thermodynamic phase stability, and on structural and mechanical properties, including Young’s modulus.

2.0 CALCULATION METHODOur calculations were carried out using

the Cambridge Serial Total Energy Package code (CASTEP) [22, 23). In this code, the Kohn-Sham [31] single electron equation model was used to calculate the fundamental eigenvalue [25]. To minimise the basis set, Perdew, Burke and Ernzerhof (PBE) formalism [32] was used to describe the real electronic functions while the CASTEP code was used for calculations due to its computational efficiency. The pseudo-potential approximation was presented by replacing the nucleus and the core electrons by an effective potential [25].The exchange correlation is treated using the generalised-gradient approximation (GGA) of Perdew, Burke and Ernzerhof (PBE) formalism [32], because of its accuracy in describing the bulk


properties of many materials, and because it is flexible enough to describe the random distribution of the impurity atoms in the Ti lattice [28]. The PBE functional is designed to reproduce closely the PW91 results [26], but the PBE formulation is more robust to guarantee a high level of convergence, cut-off energy of 500 eV and smearing of 0.1 eV electron levels.

For each structure, tests were carried out using different k-point meshes to ensure absolute convergence of the total energy to within a range better than 1.0 meV/atom. Through this study, we calculated the total energies as a function of volume while optimising unit cell-external degree(s) of freedom (i.e. the unit-cell shape) and unit cell-internal degree(s) of freedom (i.e. Wyckoff positions) as permitted by the space-group symmetry of the crystal structure. Such structural optimisations were iterated until the Hellman–Feynman forces were less than 4 meV/Å in magnitude, ensuring a convergence of the energy with respect to the structural degrees of freedom to better than 1 meV/atom ( 0.1 kJ/mol). In addition, all calculations were performed using the ‘‘accurate’’ setting within CASTEP code to avoid wrap-around errors. With the chosen plane-wave cut-off and k-point sampling, the reported formation energies are estimated to be converged to a precision better than 2 meV/atom.

In addition to the optimization, we performed a random distribution of the chemical species over the sites of the structures by replacing Ti atoms with Mo, Zr or Nb atoms, but this had no effect on the system. Nb and Zr atoms are added to binary Ti-6Mo supercell consisting of sixty four atoms to attain an average of 0, 6, 6, 5 and 4 at. % of Nb and 6, 0, 2, 3 and 4at. % of Zr, respectively, (See Fig. 1 and 2).

The stoichiometry was chosen for many reasons. (i) In our previous work on binary Ti–Mo-based alloys, we have seen that some important elastic responses (such as Young’s modulus) are sensitive functions of composition, requiring Mo contents lower than a critical level of about 6-8 at. %, thus requiring composition of around the least stable β phase in the Ti-Mo alloy system [5].

(ii) Further, it has been shown earlier that the composition of the least stable β phase alloy, βc, correlates in some ways with the emergence of many unique properties, such as non-linear superelasticity and very low work hardening rate [28]. (iii) From the view point of memory and time cost, a traceable size of the stoichiometry was chosen. (iv) In line with previous work elsewhere, the stoichiometry was held constant to provide an atom-atom comparison of the relative potency of different substitutional solutes to enhance β stability; and finally (v) holding the symmetry and size of the supercell constant promotes maximum cancellation of numerical errors.

2.1 Calculation details2.1.1 Thermodynamic phase stabilityMaterial design for biomedical application or implant usage is essentially a multi-criteria optimisation constrained by (1) biocompatibility of the alloy elements, and (ii) possession of elastic modulus that is as bone matching as possible. For this perspective, a detailed insight of how alloy composition affects the stability of β-phase is needed to tailor the Young’s modulus. In the present work, we studied the effect of Zr and Nb micro-additions (biocompatible elements) on the relative stability of the Ti-6Mo alloy (i.e. initial least stable β composition) by ab initio calculation. The formation energy of an alloy has been described by several authors as one of the most important quantities that can be used to reveal thermodynamic stability and understand metallurgical trends in the properties of an alloy.

In the present work, the formation energies of the (omega) ω, (hexagonal) α, (orthorhombic) αand (bcc) β phases for the Ti-6Mo-xNb-xZr alloys were calculated according to Eqn.1 below.

EFM( (Tix Mo6+U ))=

Etotbulk ( (Tix Mo6+U ))

∑i

N

−μ( (Tix Mo6+U )) ¿ (1) Here, U represents one or combined atoms of Zr and Nb; N is the total number of atoms per supercell; Etot

bulk (Ti x Mo6+U ) is the first

Property-Driven Design of Ti Alloys for Biomedical Applications 20


principle calculated total energies of the respective alloys; µ is the chemical potential of the alloys in corresponding bulk phase, and is the alloy composition. In the above definition (Eqn.1), the alloy is thermodynamically stable forEF

0 <0. Therefore, from the stand point of thermodynamics, lower formation energy implies better stability of a particular crystal structure over another.

Similarly, the cohesive energyEc, of a single metal compound, defined as the work needed to decompose the crystal into single free atoms [23] and expressed according to Eqn. (2) can be used to measure or complement the structural stability determined by the formation energy.

Ec=1 /n+m+…. ( Etot−nEatomA −nEatom

B +…. ) (2)

Here, nEatomA −nEatom

B are the energy per atom of A and B compounds in an isolated state.

2.1.2 Calculation of polycrystalline elastic parameters

In our previous work on binary Ti–Mo-based alloys [5], we have seen that binary Ti-6Mo alloy exhibits two structures consisting of cubic and orthorhombic phase fractions. Therefore, the subset of the supercells or cubic and orthorhombic symmetries consisting of three (C11, C12, C44) and nine (C11, C12, C13, C22, C23, C33, C44, C55, C66) elastic constants, respectively, were calculated by employing the methodology of the integral elastic response of multi-phase polycrystalline aggregates as explained in [29, 30], which was originally applied by Zeller and Dederichs [31] to determine elastic properties of single phase polycrystals with cubic symmetry. The concept was extended to determine the multiphase composites of hexagonal α phase and bcc β- phase in binary Ti-Nb and Ti-Mo alloys by M. Friak et al [34]. Here, we apply this criterion to calculate (a) the elastic properties of homogenized components of orthorhombic martensitic α´´ phase and the bcc β-phase and (b) the volume fractions in multicomponent Ti-6Mo –xNb-xZr alloys. For materials with orthorhombic symmetry, Eqns. (21) and (22)

in [34] comes to Eqns. (3) and (4), respectively:

15 τ 44=a−b+ β (2d−2 c−e )+3γ (d−c+e )+β ∆❑

1−αβ−9 γ ( K v−~B0 )+β ( β+2 γ ) (c−d )−2 eβγ−1

3 β2 ∆❑+3( C44−~μO

1−2 k (C44−~μO )+

C55−~μO

1−2 k (C55−~μO )+

C66−~μO

1−2 β (C 66−~μO ) )(3)

τ11+2 τ12=9 (K v−

~B0 )+2 β (d−c+e )+3 β2 ∆❑

3 [1−αβ−9γ ( KV−~B0 )+ β ( β+2 γ ) (c−d )−2eβγ−1 β2 ∆❑

3 ](4)

Here, G and B replace μ¿ and B¿ in the equations for β, , and ∆❑. As soon as G and B have been determined, the homogenized polycrystalline Young’s modulus (E) and Poisson’s ratio (ν) can be obtained using standard elastic relationships: The homogenised polycrystalline Young’s modulus is calculated using:

E= 9 BG3 B+G (5)

and the homogenized polycrystalline Poisson’s ratio using:

ν= 3 B−2G3 (2 B+G) (6)

2.2 Experimental verification MethodsIn order to compare the predictions with

experimental data, Ti-6Al-4V and the new multicomponent Ti-Mo alloys, namely, Ti-6Mo-6Nb, Ti-6Mo-6Zr, Ti-6Mo-6Nb-2Zr, Ti-6Mo-5Nb-3Zr, Ti-6Mo-4Nb-4Zr, (all in at. %) were melted, cast and homogenized at T = 1200 ◦C. Characterization was done with X-ray Bragg diffraction method in conjunction with EDXRF analyses. Differential thermal analysis (DTA) was used to identify the changes in the transition temperature, while, the Young’s modulus of the samples was measured by ultrasonic method on an ultrasonic velocity gauge, Olympus 62, UK. A normal incident probe (model M110, 5MHz) and a shear probe (model V221, 5MHz) were used for the measurement of normal and shear velocities of the wave, respectively. The density of the samples was measured on an automatic density meter; for at least five times. The relationships between ultrasonic velocity and the elastic properties

Ti Zr

MoNb

of materials are given below [24-29]. Young’s modulus (E) is expressed as:

E=ρ V S

2 (3V L2−4 V S

2 )V L

2−V S2

(7) Shear modulus (G) is the ratio of

shearing stress τ to shearing strain γ within the proportional limit of a material and is expressed as:

G= ρV S2 (8)

Poisson’s ratio υ is the ratio of transverse contraction strain to longitudinal extension strain in the direction of stretching force, and is expressed as:

¿( 12 )(V L

2 −2V S2)

V L2−V S

2 (9)

In the above set of equations, Vl and Vs are the ultrasonic longitudinal and shear wave velocities, respectively, and ρ is the density of the material.

3. RESULTS AND DISCUSSION3.1. Structure information and

thermodynamic phase stability: theoryIn all the six variants in Table 1, Ti-Mo

alloys are generally characterized as disordered in the ground state and a random distribution of the chemical species (namely Ti, Mo, Nb and Zr) over the sites of the β (bcc), ω (omega), α (hexagonal) or α´´ (orthorhombic) lattices is assumed. The structure of the alloys is approximated by a large supercell whose sites are occupied such as to minimize the total energy of the system (See Fig. 1 and 2).

Fig. 1: The bcc unit cell model of Ti-6Mo-xNb-xZr used in the calculations of elastic coefficients. Here Ti and Mo atoms are shown in grey and deep green, while Zr and Nb are represented in red

and lighter green respectively.

Table 1: Theoretically predicted structural parameters and thermodynamic properties of the multicomponent Ti-6Mo-x-Nb-XZr alloys


Formation Energy, EFm (MeV) of the phases a,(Å) (Cal/Exp.)

Ec(MeV)Compounds Im3m( )β P6/mmm ( )α Cmcm ( ´´)α p3m1( )ωTi-6Mo -48 -28 -41 -35 3.24/3.29 -6.15Ti-6Mo-6Zr -55.3 -35.3 -45.3 -45.1 3.2/3.31 -5.38Ti-6Mo-6Nb -84.1 -44.1 -67.1 -47.1 3.11/3.20 -12.6Ti-6Mo-6Nb-2Zr -72.7 -22.7 -56.7 -42.1 3.19/3.21 -10.2Ti-6Mo-5Nb-3Zr -69.1 -27.1 -57.1 -42.7 3.23/3.26 -8.61Ti-6Mo-4Nb-4Zr -48.9 -28.8 -48.9 -38.9 3.24/3.29 -8.1

From the view point of thermodynamic stability, for an alloy to be thermodynamically stable, its formation energy must be negative. We apply this criterion to study the energetic influence of alloying element on the structural stability. The calculated results of the effect of Nb and Zr micro-additions on the formation energy EFm (meV/atoms,), cohesive energy(E c) (meV/atom) and lattice constant, a (Å) for the various phases are listed in Table 1. The lattice constants of the structures are calculated by geometrical optimization. As seen in Table 1. the alloying elements considered here are good β stabilizers, as the values of the formation energies EFm for β(bcc) phase are more negative than those of ω, α and α´´ phases. From this perspective, we can infer that the synergic additions of Nb and Zr elements in the alloy destabilize the ω and α phases (their formation energy increases with the composition x) but stabilize the α´´ and β-phase (their formation energy decreases with composition x). It is notable that orthorhombic α´´ phase and bcc β phase exhibited similar stability trend with close values of formation energies,EFm,

however, the values for the β (bcc)-phase is lower than that of the α´´ phase. This implies that the ternary and quaternary addition of Nb, Zr or their synergic micro-alloying additions are energetically favourable to

stabilising the β-phase with low composition (<6at. %) in the Ti-Mo alloy.

Similarly, from the values of the cohesive energy (Ec ¿ shown in Table 1, the β-stability of the alloys can be ranked in the order: Ti-6Mo-6Nb> Ti-6Mo-6Nb-2Zr> Ti-6Mo-5Nb-3Zr> Ti-6Mo-4Nb-4Zr> Ti-6Mo-6Zr. The present results are consistent with other first principles calculations [44, 45, 46] as well as previous experiment phenomena [13]. Some of this trend was confirmed experimentally, as explained below

In view of the practical use of the alloys for implant application in human body, T= 310K is a suitable reference for the study of relative stability of the alloys. In order to determine the thermodynamic properties and stability of phases at this temperature, the entropy effects were considered. These can be decomposed for solids into a configurational (mixing) and vibrational contribution [37]. In this study we use a rough estimate of the temperature dependence and neglect the vibrational contribution which is computationally difficult to access. The remaining contribution, the configurational entropy, is

Fig. 2: Layer slab super cell for (110) surface energy and elastic matrix


calculated in the ideal mixing approximation. This approximation becomes exact for alloys where the formation energy depends only on

the concentration, not on the local atomic configuration, which was of no effect.

Fig. 3: Theoretical formation energies of Ti-6Mo-xNb-xZr alloys (where x= 0, 6, 6, 5 and 4 at. % for Nb and x= 6, 0, 4, 3 and 2 at. % for Zr) as a function of x concentration at T=3310K. The lines are for eye guidance.

Table 2: The chemical compositions of the Ti-6Mo-Nb-Zr alloys studied by EDXRF (at a resolution near 150 eV FWHM (all in atomic %).

As shown below, the approximation is well justified in the case of the Ti alloys studied here. The ideal mixing entropy is given by: Sconf (x)=K B [ x .∈( x )+(1−x ) .∈(1−x ) ](10) where x is the alloy composition of a multicomponent alloy Tix Mo6 Nbx Zr x and K B is the Boltzmann constant. The temperature dependent free energy can then be calculated according to:F r (x ,T )=⟨ EF (Ti−6 Mox Nb x Zr )⟩−T Sconf .

(11) where the averaged formation energy of Ti−6 Mox Nbx Zr alloys is obtained from the formation energies of alloys with different

local atomic configuration, but same concentration by averaging using the Boltzmann statistics at the reference temperature. The temperature-dependent free energy of formation allows us to determine the thermodynamic stability of an alloy at a given temperature. The corresponding results are shown in Fig. 3. Although the effect of temperature is less pronounced, evidently the finite temperature significantly reduced the formation energy, which implies that the stability of the alloy is somewhat entropy driven. Comparing the (EFm¿ curves for the different crystal structures, we can infer that (bcc) β-type phase is the most stable phase, while the

metals and alloy structures

Formation energy per atom

(eV/atom

)

-300

-200

-100

0

100

200

Formation energy per atom

(eV/atom

)

-300

-200

-100

0

100

200a"


Materials Mo Nb Zr C Cu Si O H TiTi-6Mo6Zr 5.98 5.93 5.96 0.018 0.005 0.011 0.045 0.0036 Bal.Ti-6Mo-6Zr 5.98 5.93 5.96 0.018 0.005 0.011 0.045 0.0036 Bal.Ti-6Mo-6Nb 5.97 5.96 2.97 0.010 0.003 0.011 0.045 0.0036 Bal.Ti-6Mo-5Nb-3Zr

5.98 4.99 2.97 0.02 0.002 0.012 0.044 0.0036 Bal.

Ti-6Mo-4Nb-4Zr

5.97 3.96 3.98 0.001 0.003 0.010 0.045 0.0036 Bal.

Ti-6Mo-5Nb-3Zr-ST

Ti-6Mo-6Nb-ST

Ti-6Mo-6Nb-2Zr-ST

Ti-6Mo-4Nb-4Zr-ST

(a)

Ti-6Mo-4Nb-4Zr-SW

Ti-6Mo-6Nb-2Zr-SW

Ti-6Mo-6Nb-SW

Ti-6Mo-5Nb-3Zr-SW

(α´´) orthorhombic-type phase is likely a metastable phase at the reference temperature. Therefore, the elastic modulus of the alloys

was modelled as a polycrystalline aggregate of α´´/β composite in the alloy matrices.

The experimental verification of the trend obtained from the theoretical formation energy is presented in the preceding section.

3.2 Comparison with experimental data: Composition and Phase analyses

Chemical analyses (EDXRF) were performed in many different areas (bulk and surface), and results show that the actual chemical composition of the alloys is close to nominal values (See Table 2), agreeing with ASTM F-67. As can be seen, all the

alloys are within ~±1% variance when the experimental and nominal values were compared. The chemical composition of the alloys was homogeneous and no expressive differences were found between the bulk and surface of the samples, which indicates a good homogenisation of the studied alloys. In order to elucidate the compositional sensitivity of alloy elements in the Ti-Mo alloy system, an XRD analysis was carried out at room temperature on solution treated (designated as ST) and swaged plus annealed (SW) specimens.

Fig. 4: XRD profiles of Ti-6Mo-xNb-xZr alloys (where x= 0, 6, 6, 5 and 4 at. % for Nb and x=6, 0, 4, 3 and 2 at. % for Zr) subjected to (a) solution treatment (ST) and (b) Swaged and(SW)


The XRD results for ST and SW samples scanned from 30 to 80 degrees in diffraction angle (2θ) are presented in Fig. 4(a) and (b) respectively. Fig. 4(a) provides unambiguous evidence for the presence of the β phase in all alloys after solution treatment at 1073K for 1hour followed by quenching into water, thus confirming the sensitivity of the initial binary Ti-6Mo alloy to the micro-additions (Nb and Zr). Strong β peaks associated with the (002), (200), (220) and (211) diffraction planes are evident in all the samples. When comparing the results of the different alloy samples, the diffraction peaks of orthorhombic α´´ were detected in Ti-6Mo-5Nb-3Zr-ST and Ti-6Mo-4Nb-4Zr-ST specimens, as indicated by the splitting of the single α´ (1010) peak into three α´´ peaks. These α´´ peaks are of a much weaker intensity, or unnoticed, in the other alloys. This is consistent with the theoretical data; besides the Ti-6Mo-6Zr alloy consisting of no Nb, Ti-6Mo-4Nb-4Zr and Ti-6Mo-5Nb-3Zr alloys have the least negative formation energy and cohesive energy values. According to various references [20,21,22], the amount of α´´ martensite phase in Ti alloys is conditioned by the β-stabiliser content and by grain size. When Nb is low, there exists an orthorhombic α´´ phase, but when Nb is high, the α´´ phase peaks become weaker and more diffuse. In line with this finding, α´´ (020), α´´ (111) and α´´ (021), α´´ (022), α´´ (131) and α´´ (221) peaks exist in the Ti-6Mo-4Nb-4Zr-ST sample. A weaker and more diffused α´´ (021) peak appeared in the Ti-6Mo-5Nb-3Zr –ST sample, while just a single β phase exists in the Ti-6Mo-6Nb-ST and Ti-6Mo-6Nb-2Zr-ST alloys (consisting of high Nb contents). Clearly, with the Nb content increasing, the β phase becomes stable. The diffraction peaks of the α´ or ω phases have not been detected, although the intensity of ω phase is not always high enough to provide evidence of the presence of ω phase when the ω phase has quite a small size or volume fraction. However, the results are in good agreement with the theoretical data. Interestingly, upon SW at 673K for 20 minutes, the α´´ vanishes in the XRD profile of all the samples as shown in Fig.4 (b).

Strong β peaks associated with the (002), (200), (220) and (211) diffraction planes are evident in all the SW specimens without any trace of a metastable phase. This suggests that α´´ phase is formed before it transformed back to β phase upon annealing.

β-transus for pure Ti is reported at 882 °C and the addition of β stabilizing element reduces this value. It is known that, under the Ti-Mo binary system, Mo or any other β stabilizing elements lowers the β transus and promotes the formation of the β phase, while also suppressing other metastable alloys , such as ω, α and α´´ at room temperature. Therefore, DTA was used to characterise the β transus of the ST specimens from 0 to 1200°C during heating and cooling, and the results are shown in Fig. 5. Obviously, it can be seen that it was only in low or none Nb containing alloys (Ti-6Mo-6Zr and Ti-6Mo-4Nb-4Zr) that exothermic peaks were found at 861K (588°C). When comparing the results of the different alloy samples, it is clearly evident that with the Nb content increasing, the β phase becomes stable, lowering β transus temperature below room temperature, thus suggesting that the β to α´´ martensitic transformation can be effectively retarded or even suppressed by specific alloying and subsequent heat treatment. This result is consistent with the result obtained from the theoretical data.

3.2. Comparison with experimental data : Mechanical property

As mentioned before, when elastically soft bone tissue is replaced by a stiffer implant, the implant takes over a considerable amount of the load, shielding the surrounding parts of the skeleton. Reducing the physiological loads on the bone induces re-sorption processes that give rise to a drop in bone density, mineralization state and strength [25]. Stress shielding can finally promote contact loosening, implant failure, or debris-induced infections. Therefore, it is of great importance that the elastic mismatch between the bone replacement material and existing bone be minimized in metallic biomaterials intended for orthopaedic application.


In this study, the elastic response for each condition was evaluated by ultrasonic technique to measure the change in the Young’s modulus of the alloys, and compared with theoretical data. The test was performed for at least five times before taking an average value. 3.2.1 Elastic modulus: Theory and

ExperimentTaking into account the structure symmetry and the volumetric fraction of phase

Table 3: Theoretically predicted polycrystalline integral elastic parameters and phase composition of multicomponent Ti-Mo composites with selected Nb and Zr concentrations

(of actually cast samples) together with the experimental data of Ti, Nb, Zr, Mo and other multicomponent Ti-6Mo alloy.at zero pressure.


Material

Elastic properties (GPa) and phase composition together with the experimental data.

V βTheory V β

expBβ / α ´´ G β /α ´ ´

Eβ /α´Theory E´ [ST ]

exp . E´ [SW ]exp .

υβ /α ´´ G/B

αTi 0 0 111.3 39.4 131.6 - - 0.33 0.35Mo 1 1 120.3 22.5 113.3 - - 0.37 0.6Nb 1 1 136.3 58.6 150.8 - - 0.31 0.43Zr 0 0 103.2 60 93.0 - - 0.36 0.58Ti-6Mo 0.29 0.31 101.2 41 102.1 96.5 117 0.24 0.41Ti-6Mo-6Zr 0.18 0.22 190.73 49.2 70.2 65.2 71.3 0.26 0.42Ti-6Mo-6Nb 0.78 0.73 100 42.18 90.6 77.7 81.4 0.18 0.26Ti-6Mo-6Nb-2Zr 0.67 0.61 114.83 40.46 76.5 58.7 73.5 0.19 0.30Ti-6Mo-5Nb-3Zr 0.58 0.53 118.2 46.12 59.1 61.3 76.1 0.20 0.32Ti-6Mo-4Nb-4Zr 0.42 0.39 104.1 49.6 38.3 41.02 54.0 0.20 0.35

components, the polycrystalline elastic modulus was computed by homogenizing the elastic properties of the polycrystalline α´´/ β composite using Eqns. 3-9 to determine the

Young’s modulus(~E )¿, Poisson’s ratio(υ )¿ , Bulk modulus (B) and Shear modulus (G).

The theoretically predicted elastic material properties compared with experimental data obtained in this study are summarized in Table 3 and visualized in Fig.6. Fig. 6 shows the Young’s moduli of the multicomponent Ti-6Mo –XNb-xZr alloys subjected to solution treatment (ST) and swaging (SW) (process described earlier). All the alloys subjected to solution treatment exhibit considerable low Young’s moduli of <75GPa which is much less when compared with those of SUS 316L, CP-Ti and Ti64 ELI [7]. In the latter case, the values are generally high. For the ST samples, the trend depicted in the figure shows that as the Nb increases, the Young’s modulus increases. The result indicates elastic moduli values of 60.2GPa, 74.7GPa, 68.7GPa, 61.3GPa, and 41.02GPa for Ti-6Mo-6Zr,Ti-6Mo-6Nb,Ti-6Mo-6Nb-2Zr,Ti-6Mo-5Nb-3Zr and Ti-6Mo-4Nb-4Zr

alloy samples, respectively, with an estimated error of ±3%, which represents ~ 41% reduction in the Young’s modulus of Ti-6Mo alloy(See Table 3). The latter indicates 71.3

GPa, 81.4 GPa, 73.5GPa, 77.1 GPa, and 54.0GPa for Ti-6Mo-6Zr, Ti-6Mo-6Nb, Ti-6Mo-6Nb-2Zr, Ti-6Mo-5Nb-3Zr and Ti-6Mo-4Nb-4Zr alloy samples, respectively (for the SW samples).

It is well known that phase(s) as the main constituent of microstructure has a significant effect on the mechanical properties of Ti alloys [31] and that is likely to increase or decrease the Young’s modulus. Therefore the presence of a single β-phase in the SW samples (Fig.4b) is the main factor in the change in Young’s modulus among the designed alloys.

As evidenced in Table 3 and visualized in Fig. 6, our theoretical result agrees well with

2D Graph 1

Designed Materias

Ti-6Mo Ti-6Mo-6Zr

Ti-6Mo-6Nb

Ti-6Mo-6Nb-2Zr

Ti-6Mo-5Nb-3Zr

Ti-6Mo-4Nb-4Zr

Young's M

odulus [GP

a)

0

20

40

60

80

100

120

140

Young's M

odulus [GP

a)

0

20

40

60

80

100

120

140

TheoryST -ExperimentSW-Experiment

Fig.6: Theoretical calculated Young’s modulus, E of Ti-6Mo-xNb-xZr (x= 0, 6, 6, 5 and 4 at. % for Nb and x= 6, 0, 4, 3 and 2 at. % for Zr) compared with data from experimental result.


the experimental (ST) values within an error margin of ~±2%. It can be recalled that both the theory (Fig. 3) and experiments (XRD and DTA analyses) in Figs. 4 (a) and (b), and 5 indicate that the two alloys that exhibit the lowest Young’s moduli (Ti-6Mo-5Nb-3Zr and Ti-6Mo-4Nb-4Zr) had α″ orthorhombic martensite phase in their matrix. Thus, their low E can be ascribed to the presence of the α´´ plates, which presents a modulus about half of that of the β phase [30]. The small discrepancies in the results could reflect differences in temperature, minor experimental errors or the approximations inherent in DFT-GGA [18, 30]. Therefore, it is apparent that the technique of coupling the elastic properties by homogenising multiphase elastic parameters may lead to something of a breakthrough with respect to reducing the Young’s modulus of metallic biomaterials.

Poisson’s ratio (ν) provides more fundamental information about the characteristics of the bonding forces than any other elastic constants [38]. The 0.25 and 0.5 are the lower and the upper limits for central force solids, respectively [39]. Our theoretical result of the Poisson’s ratio of the designed alloys is in the range of 0.18 and 0.22, which shows that bonding forces are non-central forces and the lower the Nb, the more the directional bonding in the multicomponent Ti-6Mo-xNb-xZr alloys studied. Moreover Poisson’s ratio shows relation with the iconicity of compounds [38]. It has a small value (ν=0.1) for covalent materials and has a typical value of (ν=0.25) for ionic materials.

In our case, the Poisson’s ratio shows that they mainly exhibit covalent characteristics. The addition of Nb and Zr alloy elements led to low Poisson’s ratio, and improved bonding covalency.

Similarly, whether or not a material demonstrates brittle or ductile behaviour is of central importance as it has a direct influence on its applications range. From this perspective, Pugh [27] proposed the ratio of bulk to shear modulus of materials, B/G ratio as an indication of brittle and ductile behaviour: low G means low resistance to shear, hence ductility while low B means low resistance fracture, hence brittleness. The critical value which separates ductile and brittle materials has been evaluated to be equal to 1.75. If B/G>1.75, a material behaves in a ductile manner, and vice versa, if B/G <1.75, a material demonstrates brittleness. From the theoretical data shown in Table 3, all the designed Ti-6Mo-xNb-xZr alloys exhibited good ductile behaviour. Therefore, B/G can describe the brittle/ductile properties of materials, and understanding these differences is essential in analysing the mechanical properties of materials with different slip characteristics.

3.2.2 Hardness The result of Vickers hardness measurement of the multicomponent Ti-6Mo–XNb-xZr alloys subjected to solution treatment (ST) and swaging (SW)plus anneal are listed in Table 4 and they are very similar to those of other biomedical Ti alloys with Nb and Zr additions.

Materials Hardness -ST [HV] Hardness -SW [HV] Yield Strength [σ0.2 (MPa)]

CP –Ti[22] 297.76 - 170-485Ti-6Al-4V[ 22] 325.87 - 825-869ASTMF75[ ](Co-Cr-Mo) 354.9 - 448Ti-10Mo [ 22] 325.3 - 412Ti-15Mo [ 22 ] 342.8 - 544Ti-20Mo [22 ] 358 - 428Ti-6Mo [this work] 298 346 483Ti-6Mo-6Zr [this work] 301 341 -Ti-6Mo-6Nb [this work] 324.13 363 -Ti-6Mo-6Nb-2Zr [this work] 326.98 360 -Ti-6Mo-5Nb-3Zr [this work] 322.7 347 -Ti-6Mo-4Nb-4Zr [this work] 305 345 -

Table 4: Mechanical properties of designed multicomponent Ti-Mo alloys and some typical metallic biomaterials


For example, Vickers Hardness Values of Ti-10Mo-xNb alloys(x=3, 7, 10) were between 396 and 441 HV [13], while those for alloys in Ti-8Mo-xZr system were between 400 and 500HV [7]. For the SW specimens, it is notable that alloys that exhibit α″ orthorhombic martensitic phase (Ti-6Mo-6Zr-ST and Ti-6Mo-4Nb-4Zr-ST) possess significantly lower hardness values (≈301-305 HV), while the hardness of the specimens consisting of single β phase (i.e., Ti-6Mo-6Nb-ST and Ti-6Mo-6Nb-2Zr-ST alloys) are 324.13HV and 326.98 HV, respectively. The highest value compares very well with the hardness 325.87 HV of solution-treated Ti-6-4 alloys (i.e Ti-6Al-4V alloys) that have been accepted by the bio-implant research community as a material of choice. Generally, this can be attributed to two factors: the first is the higher solid solution effect, precipitation hardening exhibited by β phase due to increasing Nb concentration. The other can be linked to the grain size exhibited by the samples.

As envisaged, the hardness values of the SW specimens are higher than those of the ST specimens (See Table 4). This change in hardness of the SW alloys is attributed to the increase in strength during the deformation process. This promotes the development of small grain size, small grain boundaries and high dislocation density. According to the Hall-Petch principle, this should result in high strength and hardness values in the SW designed multicomponent alloys. The hardness value for the specimens are 346HV, 341HV, 363HV, 360HV, 347 HV, 345 HV for Ti-6Mo,Ti-6Mo-6Zr , Ti-6Mo-6Nb,Ti-6Mo-6Nb-2Zr , Ti-6Mo-5Nb-3Zr and Ti-6Mo-4Nb-4Zr, respectively.

4.0 CONCLUSIONS A multidisciplinary approach was used to predict the polycrystalline phase stability, elastic constants and electronic properties of five dual-phase multicomponent Ti-6Mo alloys. The approach combines a thermodynamic analysis with a self-

consistent homogenization scheme that can describe phases with differing crystal structures. Thermodynamics provides the composition and volume fraction of the various phases, while homogenization estimates polycrystalline elastic constants from single crystal ones. All of the input values for this multi-scale approach originate from ab initio calculations, making this approach a strong tool in a theory-guiding materials-design strategy.

In this study, the thermodynamic analysis predicts the volumetric fraction of bcc-β phase composition in the alloys and the results is in good agreement with experimentally measured data. The composite is predicted to consist of mainly cubic β-phase with fairly high Nb content (~ 6at. %). Despite the fact that our theoretical thermodynamic analysis overestimates the volume fractions of β phase compared with those experimentally found, the predicted compositional trend is qualitatively correct.

The resulting Young’s modulus of polycrystalline α/β Ti-Mo alloys decreased as the volume fraction of α´´phase (or Nb content) decreased. Theoretically, a complete suppression of the presence of α´´phase would result in a reduction of the Young’s modulus to about 38 GPa, which is predicted in the case of dual α´´ and β-Ti-6Mo phase containing 4at. % Nb and 4at. % Zr, respectively. While the predicted modulus values are generally lower than that experimentally observed, the compositional trends are predicted correctly. From an alloy design perspective, we can conclude that in order to achieve maximum softness, the amount of the orthorhombic α´´-phase should be optimized via keeping the amount of Nb low enough to ensure thermodynamic stability of the α´´ in β-Ti-Mo.

REFERENCES 1. M. Long, H.J.Rack, Biomaterials 19 (1998) 1621–1639.

2. M. Niinomi, Mater. Sci. & Eng A 243 (1998) 231–236.


3. S. Ankem S, CA. Greene Mater Sci Eng A 263(2) (1999), 127. doi: 10.1016/S0921-5093(98)01170-8

4.. M. Niinomi, Sci Technol Adv Mater 4 (2003), 445. doi:10.1016/ j.stam.2003.09.002

5. P.S.Nnamchi, I Todd and M.W. Rainforth, PhD Thesis, Department of Materials Science and Eng., University of Sheffield, Uk, (2014).

6. Y. Okazaki, S. Rao, Y. Ito, T. Tateishi, Biomaterials 19 (1998) 1197–1215.

7. K. Wang Mater Sci Eng A 213 (1996) 1–2:134. doi:10.1016/ 0921-5093(96)10243-4

8. J. Black and G. Hasting: Handbook of Biomaterials Properties, Chapman and Hall, London, (1998), 167–98.

9. ASM Handbook, vol. 2, Properties and Selection: Nonferrous Alloys and Special-Purpose Materials, Joseph R. Davis, ed., ASM INTERNATIONAL, Materials Park, OH, (1990), 3–15.

10. YL Hao, Li SJ, Sun SY, Zheng CY, Hu QM, Yang R (2005) Appl Phys Lett 87(9):091903. doi:10.1063/1.2037192

11. M. Abdel-Hady, K. Hinoshita, M. Morinaga, Scripta Mater 55-5, (2006), 477. doi:10.1016/j.scriptamat.2006.04.022

12. D. Eylon, RR. Boyer, DA. Koss, Beta titanium alloys in the 1990s. Minerals, Metals & Materials Society (1993), New York

13. C. Zhang, H. Tian, C. Hao, J. Zhao, Q. Wang, E. Liu and C. Dong, J Mater Sci 48 (2013) 3138–3146: DOI 10.1007/s10853-012-7091.

14. D. Raabe, B. Sander, KM. Fria, D. Ma and J. Neugebauer, Acta Mater 55-13 (2007), 4475. doi:10.1016/j.actamat.2007.04.024.

15. H. Ikehata, N. Nagasako, T. Furuta, A. Fukumoto, K. Miwa,T. Saito Phys. Rev.B 70-17, (2004),174113.doi:10.1103/PhysRevB.70.174113.

16. Ho W. F., Ju C. P. and Lin J. H., Biomaterials 20 (1999), 2115-22.

17. L. Trentani, F. Pelillo, F.C. Pavesi, L. Ceciliani, G. Cetta, A. Forlino, Biomaterials 23 (2002) 2863–2869.

18. D.M. Gordin, T. Gloriant, G. Texier, I. Thibon, D. Ansel, J.L. Duval, M.D. Nagel, J, Mater. Sci.: Materials in Medicine 15 (2004) 885–891.

19. S. Nag, R. Banerjee, H.L. Fraser, Mater. Sci.& Eng. C 25 (2005) 357–362.

20. K. Rajamallu, M. K. Niranjan and S.R.Dey, Mater. Sci. &Eng.C 50(2015) 52–58

21. I.J. Polmear, Light Alloys From Traditional Alloys to Nanocrystals, 4th ed., Elsevier, (2006), 301.

22. S.J.Clarck, M.D.Segal, C.J.Pickard, P.J.Hasnip, M.J.Probert, K.Refsen, M.C.Payne. Z.Krystallogr. 220(2005) 567.

23. M.D.Segal, P.J.D.Lindan, M.J.Probert, C.J.Pickard, M.C.Payne. J. Phys: Condens Matter 14 (2002) 2717.

24. W.Kohn, L.J.Sham. Phys. Rev. A 140(1965) 1133.

25. M.C.Payne., M.P.Teta, D.C.Allen, T.A.Arias, J.D.J.Joannopoalos. Rev. Mod. Phys. 64(1990) 1045.

26. J.P.Perdew, J.A. Chevary, S.H.Vosko, K.A.Jackson, M.R.Pederson, D.J.Snigh, C. Fiolhais. Phys. Rev.B. 46(1992) 6671.

27. H.J.Monkhorst, J.D.Pack. Phys. Rev.B. 13 (1970) 5188.

28. M. Abdel-Hady, K. Hinoshita, M. Morinaga, Scripta Mater 57 (2007), 1000-1003.

29. O. Eriksson, Encyclopaedia of Materials: Science and Technology, Elsevier, Amsterdam, (2003), 1–11.

30. B.V. Beznosikov, J. Struct. Chem. 44 (2003) 885.

31. R. Zeller, P.H. Dederichs. Phys. Stat. Solid. B 55 (1973) 831–842.

32. T.R. Middya, A.N. Basu. J. Appl. Phys. 59(1986), 2368–2375.

33. T.R. Middya, M.Paul, A.N. Basu, J. Appl. Phys. 59(1986), 2376–2381.

34. D. Raabe, B. Sander, M. Friak, D. Ma, J. Neugebauer. Acta Mater. 55, (2007) 4475–87.

35. Ho W. F, Ju C. P. and Lin J. H., Biomaterials 20 (1999), 2115-22.

36. W.A. Counts, W.A, M. Friak, D. Raabe, J. Neugebauer. Acta Mater. 57(2009), 69–76.

37. W.A. Counts, W.A, M. Friak, D. Raabe, J. Neugebauer. Adv. Eng. Mat. 12(2010), 572–76.

38. J.F. Nye, Propriétés Physiques des Matériaux, Dunod, 1961.

39. M. Born and K. Huang, Dynamical Theory of Crystal Lattices (Boxford, Clarendon, 1954), 336–8.

40. V. Milman, M.C. Warren, J. Phys.: Condens. Matter 13 (2001) 241.

41 J. Wang, S. Yip, S.R. Phillpot, D. Wolf, Phys. Rev. Lett. 71 (1993) 4182.


42. B.V. Beznosikov, J. Struct. Chem. 44 (2003) 885.

43 S.F. Pugh, Philos. Mag. 45 (1954) 823.

44. J. H. Dai, X. Wu, Y. Song, and R. Yang. J. App. Phy. 112,(2012) 123718.

45 A. D. Becke and K. E. Edgecombe, J. Chem. Phys. 92, 5397 (1990).

46. A. Savin, R. Nesper, S. Wengert, and T. F. F€assler, Angew. Chem. Int. Ed. Engl. 36, 1808 (1997).

47. R. F. W. Bader, Atoms in Molecules: A Quantum Theory (Oxford University Press, New York, 1990).


YIELD RESPONSE OF LEAD TO MULTI-FACTORIAL OPERATIONAL INFLUENCES DURING PROCESSING OF GALENA CONCENTRATE IN

HYDROFLUORIC ACID SOLUTION

C. I. Nwoye*, A. O. Ekete, J.C. Nwobodo and G. C. Egwu1Department of Metallurgical and Materials Engineering, Nnamdi Azikiwe University, Awka, Nigeria

* [email protected]

ABSTRACTThis paper presents the response of lead yield to multi-factorial influence during hydroprocessing of galena concentrate in hydrofluoric acid solution. X-ray diffractometer (XRD) analysis of the as-mined galena was carried out to ascertain the various compounds present. Various quantities of the galena were leached at different initial pH, resulting in different final leaching solution pH. Some selected residues were examined using Scanning Electron Microscopy (SEM) to ascertain nature of phase distribution at the different pH values. Lead yield response analysis was carried out using a derived and validated model. The response coefficient of the lead yield to multi-factorial influence was evaluated to ascertain the viability and reliability of the highlighted dependence. Results of the investigation revealed that while increasing the initial leaching solution pH, lead yield decreased correspondingly as a result of decreasing H+ (ore attacking specie) concentration per unit mass of the ore. Increase in the initial solution pH resulted to increase in the final pH due to consumption of H+ during the leaching process. It was observed that the final leaching solution pH was slightly lower (more acidic) than the initial pH due to dissolution of little sulphur from the galena during the leaching process. The validity of the model; ln ξ = 4.6 (ϑ /₰) – 0.0198ε + 0.0016ɤ was rooted on the core model expression ln ξ - 0.0016ɤ = 4.6 (ϑ /₰) - 0.0198ε where both sides of the expression are correspondingly approximately equal. Regression model generated results showed trend of data point distribution similar to those from experiment and derived model. Standard errors incurred in predicting of lead yield for each value of the variable factors: initial and final leaching solution pH as obtained from experiment, derived model & regression model were 0.3782, 2.7651 & 3.3945 x 10-5 % and 0.4074, 2.5612 & 0.2329% respectively. Furthermore the correlation between lead yield and initial & final solution pH as obtained from experiment, derived model and regression model were all > 0.86. The maximum deviation of model-predicted lead yield from the experimental results was less than 5%. This translated into over 95% operational confidence and response level for the derived model as well as over 0.95 response coefficient of lead yield to the collective operational contributions of the influencing factors.

Keywords: Lead Yield Response, Multi-Factorial Influence, Hydro-processing, Galena, Hydrofluoric Acid.

1.0 INTRODUCTION Presently, the wide applicability of lead for the manufacture of some electronic components and series of industrial alloys has prompted the need for intensive research and development geared towards improving methods of extracting lead from its natural ores. The possibilities of achieving bacterial assisted extraction of lead from its natural ores have been evaluated and found amply rewarding. In the past, the conventional roasting or hydrometallurgical process has been the basic method of lead extraction from galena. The hydrometallurgical route to lead extraction significantly eliminates atmospheric pollution due to production of SO2. Environment friendliness of this process stems on the dissolution of produced gases (in

the leaching solution) as the process progresses.

Studies [1] have been carried out on the non-oxidative leaching of galena with hydrochloric acid in the presence of a metallic chloride. In the course of the research, a succession of metallic chlorides with cations of different valencies were used in an attempt to generalize the leaching behavior of these solutions. The reaction order for leaching galena, in terms of the mean ionic activity of HC1 was 3/2 over a wide range of concentration. Results of the investigation clearly revealed that the addition of soluble chlorides to a HC1 solution increased the leaching rate of the



galena by augmenting the mean ionic activity of the acid. It was observed that the only activity that must figure in the kinetic equation is that of the HC1. The activation energy (58.5 kJ/mole) was found to be independent of the chloride used to increase the activity of the hydrochloric acid; inline with the Arrhenius prefactor.

An alternative method for treating galena involves production of sufficiently soluble lead salts, from which very pure metal can be obtained by electrowinning [1]. Non-oxidative leaching of galena with HCl in the presence of chlorides has been given much consideration, since it permits the direct conversion of lead sulfide into chloride as shown in the equation [1]: PbS + 2HCl = PbCl2 + H2S (1)The accelerating effect of chlorides in the HCl solutions used for leaching galena is well known [2-7] and has been employed in processes on a pilot-plant scaleThe kinetics and mechanisms of dissolution of the major base metal sulphide minerals, pyrite, chalcopyrite, galena and sphalerite in acidic (chloride) media have been investigated [8]. The redox potential was being monitored as minerals were ground in air, then dissolved in air-equilibrated solutions at pH 2.5. Solution samples were analysed by ICP-AES and HPLC, and surfaces of residual sulphides analysed using XPS. Results generated from the research indicated that the rates of dissolution of chalcopyrite, galena and sphalerite in the presence of pyrite were respectively as 18, 31 and 1.5 times more rapid than in single-mineral experiments. In the case of galena, the experimental data suggested extensive release of Pb ions and development of a sulphur-rich surface as galvanically-promoted dissolution progresses.

Successful attempt has been made [9] to leach galena concentrate using ferric chloride brine. The results of the investigation reveal several advantages of ferric chloride over the reagents as a leaching media which includes that it exhibits substantially faster dissolution rates for most sulphides, it is regenerated

easily by chlorination of ferrous chloride leaching by-products, and it has greater potential for the treatment of complex sulphides [9]. Further studies [10] on the ferric chloride brine leaching of galena concentrate have been carried with the view to investigating the thermodynamics and kinetics of the process. It was discovered [10] that under the leaching condition of their work, the distribution of the various metal chloro complexes is relatively insensitive to the extent of PbS dissolution [10]. Evaluation of the kinetics of Cl2-O2 leaching of galena flotation concentrate has been carried out [11]. The results of this investigation indicate that the rate of gas transfer can be enhanced by increasing the partial pressure of the gas and by using vigorous agitation to increase the surface area of the liquid-gas interface. Studies [12, 13] have shown that the final pH of the leaching solution depend on the leaching time, initial pH for the leaching solution and the leaching temperature. A model was derived [14] for predictive analysis of the concentration of dissolved lead in relation to the initial and final solution pH during leaching of galena in butanoic acid. The model shows Pb = Antilog [ exp ((γ/α)0.7407 )] (2)

that the concentration of dissolved lead during the leaching process is dependent on the values of the initial and final leaching solution pH. The validity of the model was rooted in the core expression (LogPb)N =e(γ/α)

where both sides of the expression were correspondingly approximately almost equal. The maximum deviation of the model-predicted concentrations of dissolved lead from the corresponding experimental values is less than 7% which is quite within the acceptable deviation limit of experimental results. The present work aims at analyzing the yield response lead to multi-factorial influence during hydro-processing of galena in hydrofluoric acid solution. An empirical model will be derived, validated and used for the predictive analysis.


2.0 MATERIALS AND METHODSGalena concentrate of weight (5g) and average grain size 150µm was placed in a cylindrical flask containing leaching solution of hydrofluoric acid of initial pH 3.32. The leaching process was allowed for 2hrs at a temperature of 250C. At the end of the leaching process, the final solution pH was measured and the solution filtered. The filtrate was analyzed to determine the concentration of extracted lead. The leaching process was repeated with initial leaching

solution pH values 3.4, 3.66, 3.8, 3.92 and 4.15, and all other process conditions kept constant. The extracted lead concentration was also determined and final leaching solution pH correspondingly measured for each initial leaching solution used. Scanning electron microscopic (SEM) examination was carried out on some residues generated from the leaching process. The essence was to underscore the nature of distributed phases within the reacted galena particles.

(a) (b) (c)

Figs. 1: (a) Lead-zinc ore (as crushed) (b) Lead-zinc ore (pulverized) (c) Galena concentrate sieved to 150μm for the leaching process.

Position [°2Theta]

10 20 30 40 50 60 70 80 90

Counts

0

500

1000

Cu64

O

Pb S;

Cu64

O

Pb S;

Zn O

; Cu6

4 O

Zn O

; Cu6

4 OZn

O; C

u64 O

Cu64

O Pb S;

Cu64

O

Zn S;

Zn O

; Cu6

4 O

Pb S;

Cu64

O

Pb S;

Cu64

O

Zn S;

Cu64

O

Pb S;

Cu64

O Zn S;

Cu64

O

Pb S;

Cu64

O

d2_13_1337-ver2.raw

Fig. 2: XRD Analysis of the lead-zinc ore used.

(a) (b) (c)

Yield Response of Lead to Multi-factorial Operational Influences 35

3.0 RESULTS AND DISCUSSIONVariation of lead yield with initial solution pH Table 1 clearly indicates that on varying the initial leaching solution pH at constant leaching time, temperature and ore mass-input, the lead yield concentration decreases with increase in the initial leaching solution pH all through the initial pH range (3.32- 4.15) used. It was also observed that the final leaching solution pHs increase as the lead yield decreased. Increasing the initial

solution pH implies decreasing the concentration of hydrogen ion (H+) which is the principal attacking specie on the galena. And so during the leaching process, the final solution pH increased because H+ were further reduced as they attack the galena particles. Furthermore, decreasing lead yield occurred as a result of decreasing H+

concentration per unit mass of the ore as the initial leaching solution pH was being increased.

(a) (b)

(c) (d)

(e) (f)

Fig.3: SEM of residues (a), (b), (c), (d), (e) and (f) from leaching process at initial leaching solutions pH: 3.32, 3.40, 3.66, 3.80, 3.92 and 4.15 respectively. 2000x


Table 1 also showed that the final solution pH increase as a result of increase in the initial solution pH. It was observed that the final leaching solution pH was slightly lower (more acidic) than the initial pH due to dissolution of little sulphur from the galena during the leaching process.

Table 1: Variation of extracted lead concentration with initial and final leaching

solution pH

SEM results obtained from the leaching process residues (as shown in Fig. 3) indicate presence of different phases for different residues. This was well attributed to different levels of physico-chemical interaction between the acid and galena at different initial solution pH. The SEM analysis of these residues at initial leaching solution pHs: 3.32, 3.40, 3.66, 3.80, 3.92 and 4.15 showed whitish substance sparsely distributed on the lead structures (Figs. 3 (a) - (f)). This was suspected to be precipitates of lead fluoride produced during the leaching process. Fig. 3(e) shows patches of lustre appearance of lead which suggests non-reaction of the spots with the acid.

3.1 Model FormulationExperimental data generated from this research work were used for the model formulation. Computational analysis of the data shown in Table 1, gave rise to Table 2 which indicate that; ln ξ - Sɤ ≈ K (ϑ /₰) - Nε (3) Introducing the values of K, S, and N into equation (3) reduces it to; ln ξ- 0.0016ɤ = 4.6 (ϑ /₰) – 0.0198ε (4)

ln ξ = 4.6 (ϑ /₰) – 0.0198ε + 0.0016ɤ (5)

Taking the exponential of both sides of equation (5) reduces it to;

ξ = e[4.6 (ϑ /₰) – 0.0198ε + 0.0016ɤ] (6)

Where(ϑ) = Initial pH of leaching solution(₰) = Final pH of leaching solution(ζ ) = Conc. of extracted lead (mg/kg) (ε) = Leaching time ( hrs) (ɤ) = Leaching temperature (0C)

K = 4.6, S = 0.0016, and N = 0.0198. These are empirical constant (determined using C-NIKBRAN [15]

4.0 BOUNDARY AND INITIAL CONDITION

Galena concentrate was placed in cylindrical flask 30cm high containing leaching solution of hydrofluoric acid. The leaching solution is non flowing (stationary). Before the start of the leaching process, the flask was assumed to be initially free of attached bacteria and other micro organism. Initially, the effect of oxygen on the process was assumed to be atmospheric. In all cases, weight of lead used was 5g. The initial and final pH ranges of leaching solutions used are 3.32- 4.15 and 3.11-3.98 respectively. Leaching time of 2 hrs (120 minutes) and constant leaching temperature of 25oC were used for all samples. Hydrofluoric acid concentration at 0.27mol/litre and average ore grain size;150µm were also used. The leaching process boundary conditions include: atmospheric level of oxygen (considering that the cylinder was open at the top) at both the top and bottom of the ore particles in the gas and liquid phases respectively. A zero gradient was assumed for the liquid scalar at the bottom of the particles and for the gas phase at the top of the particles. The sides of the particles were assumed to be symmetries.


(ϑ) (₰) (ζ )3.323.403.663.803.924.15

3.11 3.21 3.48 3.61 3.71 3.98

130.64 130.28 128.76 127.72 127.50 127.05

4.1 Model Validation

Table 2: Variation of ln ξ- 0.0016ɤ with 4.6 (ϑ /₰) - 0.0198ε

Equation (6) is the derived model. The validity of the model is strongly rooted on equation (4) where both sides of the equation are correspondingly approximately equal. Table 2 also agrees with equation (4) following the values of ln ξ- 0.0016ɤ and 4.6(ϑ /₰) - 0.0198ε evaluated from the experimental results in Table 1. Furthermore, the derived model was validated by comparing the lead yield predicted by the model and that obtained from the experiment. This was done using the 4th Degree Model Validity Test Techniques (4th DMVTT); statistical graphical, computational and deviational analysis.

R2 = 0.9506

90

100

110

120

130

140

150

160

3.2 3.4 3.6 3.8 4 4.2

Initial solution pH

Con

c. o

f ext

ract

ed le

ad (m

g/kg

)

Fig.4: Coefficient of determination between lead yield concentration and initial leaching solution pH as

obtained from experiment

R2 = 0.7468

90

100

110

120

130

140

150

160

3.2 3.4 3.6 3.8 4 4.2

Initial solution pH

Con

c. o

f ext

ract

ed le

ad (m

g/kg

)

Fig.5: Coefficient of determination between lead yield concentration and initial leaching solution pH as

obtained from derived model

R2 = 0.9425

90

100

110

120

130

140

150

160

3 3.2 3.4 3.6 3.8 4

Final solution pH

Con

c. o

f ext

ract

ed le

ad (m

g/kg

)

Fig.6: Coefficient of determination between lead yield concentration and final leaching solution pH as

obtained from experiment

R2 = 0.7832

90

100

110

120

130

140

150

160

3 3.2 3.4 3.6 3.8 4

Final solution pH

Con

c. o

f ext

ract

ed le

ad (m

g/kg

)

Fig. 7: Coefficient of determination betweenlead yield concentration and final leaching solution pH as

obtained from derived model


lnξ- 0.0016ɤ 4.6(ϑ /₰) - 0.0198ε 4.8324 4.8297 4.8180 4.8098 4.8081 4.8046

4.8709 4.8327 4.7982 4.8024 4.8208 4.7568

4.2 Statistical Analysis Standard Error (STEYX)The standard errors incurred in predicting lead yield for each value of the initial & final leaching solution pH considered as obtained from experiment and derived model were 0.3782 and 2.7651 & 0.4074 and 2.5612 % respectively. The standard error was evaluated using Microsoft Excel version 2003.

Correlation (CORREL)The correlation coefficient between lead yield and initial & final leaching solution pH were evaluated from the results of the derived model and experiment, considering the coefficient of determination R2 from Figs. 4-7. The evaluation was done using Microsoft Excel version 2003.

R = √R2 (7)The evaluated correlations are shown in Tables 3 and 4. These evaluated results indicate that the derived model predictions are significantly reliable and hence valid considering its proximate agreement with results from actual experiment.

Table 3: Comparison of the correlations evaluated from derived model predicted and ExD results based on initial solution pH

Table 4:Comparison of the correlation evaluated from derived model-predicted ExD based on final solution pH

4.3 Graphical Analysis Comparative graphical analysis of Figs. 8 and 9 show very close alignment of the curves from the experimental (ExD) and model-predicted (MoD) lead yields.

90

100

110

120

130

140

150

160

3.2 3.4 3.6 3.8 4 4.2

Initial solution pH

Con

c. o

f ext

ract

ed P

b (m

g/kg

)

ExDMoD

Fig.8: Comparison of lead yield concentrations

(relative to initial leaching solution pH) as obtained from experiment and derived model

90

100

110

120

130

140

150

160

3 3.2 3.4 3.6 3.8 4

Final solution pH

Con

c. o

f ext

ract

ed P

b (m

g/kg

)

ExDMoD

Fig.9: Comparison of lead yield concentrations

(relative to final leaching solution pH) as obtained from experiment and derived model.

Furthermore, the degree of alignment of these curves is indicative of the proximate agreement between both experimental and model-predicted lead yields.

90

100

110

120

130

140

150

160

3.2 3.4 3.6 3.8 4 4.2

Initial solution pH

Con

c. o

f ext

ract

ed le

ad (m

g/kg

)

ExDMoDReG

. Fig.10: Comparison of lead yield concentrations (relative to initial leaching solution as obtained from experiment, derived model and regression model)


Analysis Based on initial solution Ph ExD D-Model

CORREL 0.9750 0.8642

Analysis Based on final solution pH ExD D-Model

CORREL 0.9708 0.8850

4.4 Comparison of derived model with standard model

The validity of the derived model was also verified through application of the regression model (Reg) (Least Square Method using Excel version 2003) in predicting the trend of the experimental results. Comparative analysis of Figs. 10 and 11 shows very close alignment of curves and areas covered by lead yield, which precisely translated into significantly similar trend of data point’s distribution for experimental (ExD), derived model (MoD) and regression model-predicted (ReG) results of lead yield. Also, the calculated correlations (from Figs. 10 and 11) between lead yield and initial & final leaching solution pH for results obtained from regression model gave 1.0000 & 0.9906 respectively. These values are in proximate agreement with both experimental and derived model-predicted results. The standard errors incurred in predicting lead yield for each value of the initial & final solution leaching considered as obtained from regression model were 3.3945 x 10-5 and 0.2329% respectively.

90

100

110

120

130

140

150

160

3 3.2 3.4 3.6 3.8 4

Final solution pH

Con

c. o

f ext

ract

ed le

ad (m

g/kg

)

ExDMoDReG

Fig.11: Comparison of lead yield concentrations (relative to final leaching solution pH) as obtained

from experiment, derived model and regression model.

4.5 Deviational Analysis The deviation Dv, of model-predicted lead yield from the corresponding experimental result was given by Dv = ζ MoD – ζExD x 100 (8) ζExD

where ζExD and ζMoD are extracted lead concentration from experiment and derived model respectively.Critical analysis of the lead yield obtained from experiment and derived model shows low deviations on the part of the model-predicted values relative to values obtained from the experiment. This is attributed to the fact that the surface properties of galena and the physico-chemical interactions between the galena and the leaching solution which played vital roles during the leaching process were not considered during the model formulation. This necessitated the introduction of correction factor, to bring the model-predicted extracted lead concentration to those of the corresponding experimental values.

110

115

120

125

130

135

140

3.32 3.4 3.66 3.8 3.92 4.15

Initial solution pH

Con

c. o

f ext

ract

ed le

ad

(mg/

kg)

-6-5-4-3-2-1012345

Dev

iatio

n (%

)MoD Deviation

Fig.12: Variation of deviation with lead yield concentration (relative to the initial leaching solution

pH)

110

115

120

125

130

135

140

3.11 3.21 3.48 3.61 3.71 3.98

Final solution pH

Con

c. o

f ext

ract

ed le

ad

(mg/

kg)

-6-5-4-3-2-1012345

Dev

iatio

n (%

)

MoDDeviation

Fig.13: Variation of deviation with lead yield concentration (relative to final leaching solution pH)

Deviational analysis from Figs. 12 and 13 indicates that the precise maximum deviation of model-predicted lead yield from the experimental results is 4.7%. This translates into over 95% operational confidence and


response level for the derived model as well as over 0.95 response coefficient of lead yield to the collective operational contributions of the influencing factors.Consideration of equation (8) and critical analysis of Figs. 12 and 13 shows that the least and highest magnitudes of deviation of the model-predicted lead yield (from the corresponding experimental values) are + 0.26 and – 4.7. Figs. 5, 7, 12 and 13 indicates that these deviations correspond to lead yields: 130.621and 121.0738 mg/kg; initial leaching solution pH: 3.4 and 4.15 as well as final leaching solution pH: 3.21 and 3.98 respectively.

Correction factor, Cf to the model-predicted results is given by

Cf = - ζMoD – ζExD x 100 (9) ζExD

Critical analysis of Figs. 12, 13 and Table 5 indicates that the evaluated correction factors are negative of the deviation as shown in equations (8) and (9). The correction factor took care of the negligence of operational contributions of the surface properties of the galena and the physico-chemical interactions between the galena and the leaching solution which actually played vital role during the leaching process. The model predicted results deviated from those of the experiment because these contributions were not considered during the model formulation. Introduction of the corresponding values of Cf from equation (9) into the model gives exactly the corresponding experimental values of lead yield. Table 5: Correction factor to model-predicted

lead yield

Table 5 also

shows that the least and highest correction factor (to the model-predicted lead yield) are – 0.26

and + 4.7 %. Since correction factor is the negative of deviation as shown in equations (8) and (9), Table 5, Figs. 12 and 13 indicate that these highlighted correction factors correspond to lead yields: 130.621and 121.0738 mg/kg; initial leaching solution pH: 3.4 and 4.15 as well as final leaching solution pH: 3.21 and 3.98 respectively.It is very pertinent to state that the deviation of model predicted results from that of the experiment is just the magnitude of the value. The associated sign preceding the value signifies that the deviation is a deficit (negative sign) or surplus (positive sign).

5.0 CONCLUSIONThe yield response of lead to multi-factorial influence during hydroprocessing of galena concentrate in hydrofluoric acid solution was analyzed using a derived and validated model. Increase in the initial solution pH resulted to increase in the final pH. The final solution pH was slightly lower (more acidic) than the initial pH due to dissolution of little sulphur from the galena during the leaching process. The validity of the model was rooted on the core model expression ln ξ - 0.0016ɤ = 4.6 (ϑ /₰) - 0.0198ε where both sides of the expression are correspondingly approximately equal. Standard errors incurred in predicting of lead yield for each value of the variable factors: initial and final leaching solution pH as obtained from experiment, derived model & regression model were 0.3782, 2.7651 & 3.3945 x 10-5

% and 0.4074, 2.5612 & 0.2329% respectively. The maximum deviation of model-predicted lead yield from the experimental results was less than 5%. This translated into over 95% operational confidence and response level for the derived model as well as over 0.95 response coefficient of lead yield to the collective operational contributions of the influencing factors.

REFERENCES



Initial pH Final pH Cf (%) 3.32 3.40 3.66 3.80 3.92 4.15

3.11 3.21 3.48 3.61 3.71 3.98

- 3.88 - 0.26 +1.99 +0.78 - 1.24 +4.70

1. Nunez, C., Espiell, F and Garcia-Zayas, J. (1990). Kinetics of Galena Leaching in Hydrochloric Acid-Chloride Solutions. Metallurgical Transactions B 21b:11.

2. Muir, D. M., Gale, D. C., Parker, A. J., Giles D. E. (1976). Proc. Australas. Inst. Min. Metall., 259:23-35.

3. Tarabaev, S. I., Budon, V. I., Mateeva, K. T., and Milyutina, V. A. (1959). Chem. Abstr., 53: 6941b.

4. Esmirlov, V. V., Taraskin, D. A., Chemezov, N. S., Burov, G. D., and Tseft A. L. (1969). Chem. Abstr., 70:39845v.

5. Cambazoglu, M. and Ozkol, S.(1980). Complex Sulphide Ores, M. J. Jones, London, pp. 7-11.

6. Majima, H., Awakura, Y., and Misaki, N. (1981). Metall. Trans. B, 12B:645-649.

7. Majima, H., and Awakura, Y. (1979). Xlll Int. Mineral Processing Congress, Warsaw Polish Scientific Publishers, Warsaw, pp. 665-689.

8. Abraitis, P. K., Pattrick R. A. D., Kelsall, G. H and Vaughan, D. J. (2004). Acid Leaching and Dissolution of Major Sulphide Ore Minerals: Processes and Galvanic Effects in Complex Systems. Mineralogical Magazine. 68(2):343-351.DOI:10.1180/0026461046820191

9. Dutrizac, J.E. (1986). The Dissolution of Galena in Ferric Chloride Media. Metallurgical Transactions B , 17(1): 5-17.

10. Seon-Hyo, K. K., Henein, H., and Warren, G.W. (1986). An Investigation of the Thermodynamics and Kinetics of the Ferric Chloride Brine Leaching of Galena Concentrate. Metallurgical Transaction B, 17(1):29-39.

11. Dix, R. B., and Hendrix, J. L. (1986). Kinetics of Cl2-O2 Leaching of Lead–zinc Flotation Concentrates. University of Nevada Reno, 89557.

12. Nwoye, C.I. (2008). Bioleaching Studies of Ishiagu Galena. Ph.D Thesis, Metallurgical and Materials Engineering Department. Federal University of Technology, Owerri, Nigeria.

13. Pinches, A. (1975). Bacterial Leaching of an Arsenic Bearing Sulphide Concentrate. The Institute of Mining and Metallurgy, England, 34. [14]Nwoye, C. I. and Mbuka, I. E. (2010). Model for Predictive Analysis of the Concentration of Dissolved Lead in relation to the Initial and Final Solution pH during Leaching of Galena in Butanoic Acid. Journal of Academia Arena, 2(6), 54-61.

15. C. I. Nwoye, (2008). Data Analytical Memory; C-NIKBRAN .

EXPERIMENTAL CORRELATION BETWEEN HEAT TREATMENT PARAMETER AND CORROSION BEHAVIOUR OF 800 SERIES Ni-Fe

SUPERALLOY IN HCl SOLUTION

C.C. Nwogbu1, B.A. Okorie1,2, and V. S. Aigbodion2

1Department of Metallurgical and Materials Engineering, Enugu State University of Science and Technology, Enugu, Nigeria.

2Department of Metallurgical and Materials Engineering, University of Nigeria, Nsukka.

ABSTRACTExperimental correlation between heat treatment parameter and corrosion behaviour of 800Ni-Fe superalloy in HCl solution has been investigated. The sample was produced by spark plasma sintering (SPS) and was heat treated at temperatures of 800 and 1000oC and time of 1 and 2hours at each temperature . Electrochemical studies were carried out in HCl medium, using potentiodynamic anodic polarization technique. From the results the heat treated alloy showed better corrosion resistance than the untreated alloy. Heat treatment temperature and time have great influences on the corrosion behaviour and morphology of the alloy. The sample heat treated at 800oC for 1hour gave a high protection efficiency of 75.33% . It has been established that heat treatment can be used to improve the corrosion resistance of the alloy.

Keywords: Ni-Fe alloy, Microstructure, Corrosion and Heat treatment

1.0 INTRODUCTION Nickel-iron base alloys appear to be a promising alternative to austenitic stainless steels because of their better corrosion resistance, thermal conductivity and mechanical properties. Using these alloys, complex processes and waste streams can be handled safely due to their high corrosion resistance. [1,2]. These alloys possess higher tolerance for alloying elements in solid solution than stainless steels and other iron base alloys[3]. The good metallurgical stability of the nickel base alloys make them a better alternative to stainless steel[4,5]. Owing to the excellent mechanical and physical properties, nickel base superalloys are extensively employed in nuclear power plants[6]. The 800 nickel base solid solution strengthened superalloy with major elemental composition as Ni-Fe-Cr; is a non-magnetic, corrosion and oxidation resistant alloy. The major elements nickel and chromium, provide good resistance to oxidizing environments. In nuclear power plants. Alloy 800 is used as steam generator tubes [7]. The selection of this material is

because of its good mechanical strength, thermal conductivity, high formability and corrosion resistance. The high temperature performance of the Ni-base superalloy Superni-75 has been evaluated under cyclic conditions for 1,000 h in real service environment of the waste incinerator based upon medical waste as fuel. The performance has been characterized via surface morphology, phase composition and element concentration using the combined techniques of XRD, SEM/EDX, BSEI and EPMA. Initially, due to chlorine-based corrosion attack on the Superni-75 alloy, there was inner penetration of the corrosive species. However, with the growth of a thin Cr2O3interface layer along the scale/surface boundary, the performance of the alloy improved against the attack by the flue gases in the real service conditions of the medical waste incinerator. Boiler tubes made of Superni-75 were estimated to have an erosion-corrosion rate of about 65 mils/year. Paula Rojas [8] reported on the electrochemical behaviour and corrosion


43

https://www.researchgate.net/publication/229333445_Experimental_correlation_between_varying_silicon_carbide_and_hardness_values_in_heat-treated_Al-Si-FeSiC_particulate_composites?ev=prf_pub



resistance of glassy Fe68.6-Ni28.2-Mn3.2 (at%) specimens which were studied in different concentrations of HCl solutions. The results indicated that the corrosion rate increased with increasing concentration of the HCl solutions. Electrochemical impedance spectroscopy results were analysed by fitting the experimental data to an equivalent circuit using the ZSim Demo program, and suitable equivalent circuit models were determined. The results obtained from the impedance and polarization measurements are in good agreement. The thermodynamic parameters were evaluated for the corrosion process and discussed. In order to further research in this novel area for better service condition, the present work has been undertaken.

2.0 MATERIALS AND METHOD A spark plasma sintering machine (model SPS 10-3), manufactured by Thermal Technologies LLC, was used to produce the alloy samples. Specimens of diameter 100 mm, were produced using dies and punches of graphite. The samples were produced at a temperature of 1150°C temperature and a pressure of 5MPa with heating and cooling rate of 10°C/min . The thermocouple inserted into the bottom punch was used to measure the temperature. All the samples were produced in a closed furnace where 10-2 torr vacuum was maintained throughout the experiment. A standard Ni-Fe base superalloy(Incoloy 800) with composition shown in Table 1a below was used.

A carbolite furnace was used for the heat treatment of the samples. The samples were placed inside the furnace and the following heat treatment programme was used. i. Hold at 1000oC for 1 and 2 hours; then

rapidly cool by quenching in water. ii. Hold at 800oC for 1 and 2 hours; then air

cool, X-ray diffraction (XRD) analysis, with Cu-Kα radiation,

was conducted using a PANalytical X'Pert PRO. The XRD was operated at 45kV voltage and 40mA current.

The 2θ angles between 1°and 90° were scanned and analyzed using the Bragg law. A

Rietveld refinement software, TOPASTM, was used for quantitative analysis. A TESCAN Scanning Electron Microscope was used in the research. The polished a n d e t c h e d samples were firmly held in the sample holder using a double-sided carbon tape before putting them inside the sample chamber. The SEM was operated at an accelerating voltage of 20 kV.

Electrochemical measurements were carried out using an Autolab Potentiostat with the General Purpose Electrochemical Software package. The samples were cold mounted with epoxy leaving a working area of 0.785 cm2. The working surface was ground with grinding papers from 600 to 1000 grit, and then cleaned with distilled water and ethanol. A conventional three electrode cell, consisting of Ag/AgCl, Platinum and samples was used as: reference, counter and working electrodes respectively. The medium used for the electrochemical measurement was 0.5MHCl. The measurement was carried out at room temperature. The potentiodynamic potential scan was fixed from -1.5 V to +1.5 V with a scan rate of 0.012 V/s.

3.0 RESULTS AND DISCUSSION Figure 1a shows the XRD pattern of the untreated alloy, while Figures 1b-1e display the XRD patterns of the heat treated samples. From the XRD spectrum of the alloy, it is observed clearly that there is presence of Ni,Fe (Awaruite), Al0.3Fe3Si0.7(Aluminum Iron Silicon), and Cr7C3(Carbon Chromium) phases. After heat treatment it is observed that new phases appeared such as: FeSi(Fersilicite,syn [NR]), Manganese Carbide, Chromium Iron Carbide, Manganese Silicon Carbide. The presence of

Table 1a: Composition of the superalloy Incoloy 800 used


Ni Fe Cr Mo Mn Si C Al Ti Cu Other

32.5 44.5 21 - 0.8 0.5 0.05 0.4 0.4 0.4 -

( Ni,Fe )( Awaruite) phase is common to all the samples; this is expected because the

specimen is a Ni-Fe base superalloy. By comparing the XRD of the control(Figure 1a) with the heat treated samples(Figures 1b-1e), one can observe that there is a great

change in the spectrum which resulted in more diffraction peaks, and a larger quantity

of hard carbides with smaller inter-particle distance. In the heat treated samples, it is clear that the various phases formed after the heat treatment depend on the heat treatment condition as shown in Table 1.

Position [°2Theta]

10 20 30 40 50 60 70 80

Counts

0

400

1600

3600

Cr7 C

3

( Ni ,

Fe );

Ni3 (

Al , T

i )Al0

.3 Fe

3 Si0.

7; Cr

7 C3

( Ni ,

Fe );

Cr7 C

3; Ni

3 ( Al

, Ti )

Al0.3

Fe3 S

i0.7;

Cr7 C

3

( Ni ,

Fe );

Ni3 (

Al , T

i )

Cr7 C

3; Ni

3 ( Al

, Ti )

Cr7 C

3

SAMPLE 800.ASC

Figure 1a: XRD spectrum of the untreated alloy

Position [°2Theta]

10 20 30 40 50 60 70 80

Counts

0

400

1600

3600

6400

Al0.5

Fe0.5

; Mn2

2.6 Si

5.4 C4

Mn22

.6 Si5

.4 C4

Mn22

.6 Si5

.4 C4

( Ni ,

Fe );

Mn5 C

2; Al0

.5 Fe

0.5; M

n22.6

Si5.4

C4

Mn5 C

2; Fe

Si2;

C0.12

Fe0.7

9 Si0.

09; M

n22.6

Si5.4

C4

Ni2.6

7 Ti1.

33

Ni2.6

7 Ti1.

33Mn

Si; C

0.12 F

e0.79

Si0.0

9

SAMPLE_1.ASC

Figure 1b: XRD spectrum of the alloy heat treated at HT10001hr

Position [°2Theta]

10 20 30 40 50 60 70 80

Counts

0

400

1600

3600

6400

Al Ni

2 Ti; A

l0.96

Ni1.

04; A

l0.4 F

e0.6

Cr22

C6

( Ni ,

Fe );

Al Ni

2 Ti; A

l0.4 F

e0.6;

Cr22

C6

Fe7 C

3

Al0.4

Fe0.6

; Cr22

C6

Cr22

C6 Al Ni

2 Ti

Al0.96

Ni1.

04; A

l0.4 F

e0.6

SAMPLE_2.ASC

Figure 1c: XRD spectrum of the alloy heat treated at HT10002hrs

Correlation between Heat Treatment Parameter and Corrosion Behaviour 45 41

Position [°2Theta]

10 20 30 40 50 60 70 80

Counts

0

400

1600

3600

Fe Si

; Ni74

Si26

Al Fe

; Ni74

Si26

Fe Si

; Fe3

C; Al

2 Ti

C0.12

Fe0.7

9 Si0.

09; F

e3 C;

Cr3 N

i2; Al

2 Ti

Ni74

Si26

Al2 Ti SAMPLE_3.ASC

Figure 1d: XRD spectrum of the alloy heat treated at HT8001hr

Position [°2Theta]

10 20 30 40 50 60 70 80

Counts

0

400

1600

3600 Fe Si

; Mn1

5 C4

Al Cr

Fe2;

Fe N

i3; N

i2 Si;

Mn1

5 C4;

( Cr ,

Fe )7

C3Fe

Si; M

n15 C

4Ni

2 Si

Fe N

i3; M

n15 C

4Ni

2 Si; M

n15 C

4

SAMPLE_4.ASC

Figure 1e: XRD spectrum of the alloy heat treated at HT8002hrs

Control HT10001HR HT10002HR HT8001HR HT8002HR( Ni , Fe )(Awaruite) ( Ni , Fe )

( Awaruite)( Ni , Fe )( Awaruite) ( Ni , Fe )( Awaruite) Al Cr Fe2

Al0.3 Fe3 Si0.7(Aluminum Iron Silicon)

Mn5 C2(Manganese Carbide)

Al Ni2 Ti(Aluminum Nickel Titanium)

Fe Si(Fersilicite, syn [NR])

Fe Si(Fersilicite, syn [NR])

Cr7 C3(Carbon Chromium)

Al0.5 Fe0.5(Aluminum Iron)

Al0.96 Ni1.04(Aluminum Nickel)

C0.12 Fe0.79 Si0.09(Carbon Iron Silicon)

Fe Ni3(Awaruite)

Ni3 ( Al , Ti )( "Udimet 500")

Mn Si(Manganese Silicon)

Fe7 C3(Iron Carbide) Fe3 C(cementite) Ni2 Si(Nickel Silicon)


( Cr , Fe )7 C3(Chromium Iron Carbide)

Al0.4Fe0.6(Aluminum Iron)

Cr3Ni2(Chromium Nickel)

Mn15C4(Manganese Carbide)

Ni2.67 Ti1.33(Nickel Titanium)

Cr22 C6(Carbon Chromium)

Al2 Ti(Aluminum Titanium)

( Cr , Fe )7 C3(Chromium Iron Carbide)

Mn22.6 Si5.4 C4(Manganese Silicon Carbide)

Ni74 Si26(Nickel Silicon)

Table 1: Identified Patterns List

From the Figures 2a-2e, one can observe a great difference between the morphology of the untreated alloy(see Figure 2a) and those of the heat treated samples (see Figures 2b-2e). Samples heat treated for 1hr have a more refined structure. The dark spots show the presence of some of the new phases developed during heat treatment. The electrochemical potential of the alloy was investigated using HCl solution. Table 2 presents the corrosion data including the corrosion rate, while Fig. 3(a) and 3(b) present results of the polarization tests. From the results obtained in Table 2 and Figure 3, the corrosion rate of the samples generally decreases after heat treatment. The untreated sample has the higher corrosion rate. This was attributed to the high anodic potential reached by the sample. Meanwhile, as the sample was heat treated there was decrease in corrosion rate. This may be attributed to the formation of a hard thin film, which may have retarded the ingress of Cl-1

ions. As the heat treatment time decreased from 2 to 1hr the corrosion rate of the alloy decreased. Heat treatment contributes immensely to the corrosion behavior of the alloy (see Figure 3) e.g corrosion rates of 0.471, 0.244, 0.358, 0.116 and 0.318mm/year were obtained for the untreated, heat treated at 1000oC for 1, 2hrs and 800oC for 1, 2hrs respectively. The various hard phases formed after heat treatment are the major factor responsible for the improvement in the corrosion behavior of this alloy

Correlation between Heat Treatment Parameter and Corrosion Behaviour 47

Table 3a: Upper and lower levels used with their response

Figure 3a: Variation of the open circuit potential with time

Figure 3b: Potentiodynamic polarization curve

Table 2: Electrochemical corrosion data

Figure 2: SEM image of a) Untreated alloy b)HT10001hr c) HT10002hrs d)HT8001hr e)HT8002hrs


N SAMPLES Icorr (A/cm2) ba (v/dec) LPR Rp (Ωcm2) -Ecorr (V) CR (mm/yr)

Untreated 4.398E-5 0.330 159.1 -0.30 0.4710HT10001hr 2.275E-5 0.042 121.2 -0.304 0.2437HT10002hrs 3.349E-5 0.046 102.1 -0.289 0.3586HT8001hr 1.085E-5 0.043 180.2 -0.322 0.1162HT8002hrs 2.977E-5 0.72 80.84 -0.356 0.3180

The heat treated sample may act as a cathode, which prevents conductance of ions but allows electronic conductance to some extent. It should also be emphasized that the electrical potential in all the samples decreases in the direction from the anode to the cathode (positive ions move towards the cathode, negative ions towards the anode)(see Figure 3a). The surface of the anode in reality is much smaller than that of the

S/No Temperatures(A) oC Time(B) hrs Corrosion rate(mm/yr)1 800 2 0.31882 1000 2 0.35863 1000 1 0.24374 800 1 0.1162

cathode, i.e. the electrolytic current density is much higher near the anode. It is worthy to note in this present work that heat treatment

temperature and time play a vital role in the electrochemical potential of all the samples. The optimum condition that led to the higher corrosion resistance is the sample heat treated at HT8001hr. This optimum condition gave a protection efficiency of 75.33%. Two factors and levels of factorial design experiment were used to study the influences of heat treatment temperature and time on the corrosion behaviour of the materials. Table 3a, shows the upper and lower levels of each variable with their response values.Two factors and levels of factorial design of experiment were used to study the influences of heat treatment temperature and time on the corrosion behaviour of the materials. Table 3a, shows the upper and lower levels of each variable with their response values.Figure 4, shows the estimated response surface for the samples. It is observed that the corrosion behavior is highly influenced by temperature and time. The corrosion rate increases with increase in heat treatment temperature and time. For example as the temperature increases from 800 to 1000oC and time from 1 to 2hours the corrosion rate increased rapidly (see Figure 4). Equation 1 shows the dependence of corrosion rate on temperature and time.Corrosion rate= -0.048750+1.82500E-004* Temperature+0.11160* Time (1)

DESIGN-EXPERT Plot

Corrosion rate

X = A: TemperatureY = B: Time

Design Points

B- 1.000B+ 2.000

B: TimeInteraction Graph

A: Temperature

Corro

sion r

ate

800.00 850.00 900.00 950.00 1000.00

0.183173

0.233036

0.2829

0.332764

0.382627

Figure 4a: Interaction curve for the corrosion behaviour

DESIGN-EXPERT Plot

Corrosion rateX = A: TemperatureY = B: Time

0.20885

0.245875

0.2829

0.319925

0.35695

Cor

rosio

n rate

800.00

850.00

900.00

950.00

1000.00

1.00

1.25

1.50

1.75

2.00

A: Temperature B: Time

Figure 4b: 3-D plot for the corrosion behaviour From equation 1, it can be clearly seen that the coefficients associated with temperature and time are positive. It indicates that as the temperature rises from 800 to 1000oC, the corrosion rate rises by 1.82500E-004. Also as the time increases from 1 to 2hours the

Table 3b: ANOVA for Selected Factorial Model

Correlation between Heat Treatment Parameter and Corrosion Behaviour 49

Hardness valuesSource Sum of Squares DF Mean square Fvalue Pvalue

Model 0.014 2 6.893E-003 633.00 0.0281A 1.332E-003 1 1.332E-003 122.34 0.0574B 0.012 1 0.012 1143.67 0.0188Residual 1.089E-005 1 1.089E-005CorTotal 0.014 3

corrosion rate rises by 0.11160. This further supports the earlier observation in Figure 4a and 4b. Analysis of variance (ANOVA) was used to determine the design parameters significantly influencing the corrosion rate. Table 3b shows the results of ANOVA at 95% confidence level (significance level of α=0.05). The last column of Table 3b shows the contribution (P) of each parameter to the response, indicating the degree of influence on the results.The Model F-value of 633.00 implies that the model is significant. There is only a 2.81% chance that a "Model F-Value" this large could occur due to noise.Values of "Prob > F" less than 0.0500 indicate that the model terms are significant. In this case B(time) is a significant model term.

The "Pred R-Squared" of 0.9874 is in reasonable agreement with the "Adj R-Squared" of 0.9976.

Std. Dev. 3.300E-003 R-Squared 0.9992Mean 0.28 Adj R-Squared 0.9976C.V. 1.17 Pred R-Squared 0.9874

PRESS 1.742E-004 Adeq Precision 51.822

4.0 CONCLUSIONSFrom the results and discussion above the following conclusions can be made:1. The heat treated alloy showed better

corrosion resistance than the untreated alloy

2. Heat treatment temperature and time have great influences on the corrosion behaviour and morphology of the alloy.

3. The sample heat treated at 800oC for 1hour gave a high protection efficiency of 75.33% .

4. It has been established that heat treatment can be used in improving the corrosion resistance of the alloy.

REFERENCES1. G.B. Viswanathan, P.M. Sarosi, M.F. Henry,

D.D. Whitis, W.W. Milligan, M.J. MILLS- a ‘‘Investigation of Creep Deformation Mechanisms at Intermediate Temperatures in René 88 DT.’’ Acta Mater, 53, pp. 3041-3057, 2005

2. D. Locq, A. Walder, M. Marty,P. Caron ‘‘Development of New PM Superalloys for High Temperature Applications.’’ EUROMAT, Intermetallics and Superalloys Vol. 10, WILEY-VCH Verlag Gmbh,Weinheim, Germany (D.G. Morris et al.,eds), pp. 52-57, 2000

3. S.T. Wlodek, M. Kelly, D.Alden -The Structure of N18. Superalloys 1992, TMS, Warrendale, PA, U.S.A. (S.D. Antolovich et al., eds), pp. 467-476, 1992.


4. M. Soucail, M. Marty, H. Octor-‘‘Development’’ of Coarse Grain Structures

in a Powder Metallurgy Nickel Base Superalloy N18. Scripta Mater, 34, 4, pp.519-525, 1996

5. Khadijah M. Emran, ‘’Effects of concentration and temperature on the corrosion properties of the Fe–Ni–Mn alloy in HCl solutions,’’ Res Chem Intermed (2015) 41:3583–3596

6. W. A. Wesley and H. R. Copson, “Effect of NonCondensable Gases on Corrosion of Nickel in Steam Condensate,” TRANS. ELECTROCHEM. SOC.,May, 1949

7. Harminder Singh T. S. Sidhu ‘‘High Temperature Corrosion Behavior of Ni-based Superalloy Superni-75 in the Real Service Environment of Medical Waste Incinerator,’’ Oxid Met (2013) 80:651–668

8. Paula Rojas, Rosa Vera, Carola Martínez María Villarroel., ‘‘Effect of the Powder

Metallurgy Manufacture Process on the Electrochemical Behaviour of Copper,’’ Nickel and Copper Nickel Alloys in Hydrochloric Acid’’, Int. J. Electrochem. Sci., 11 (2016) 4701 – 4711

EFFECT OF PARTICLE SIZE ON THE MECHANICAL AND PHYSICAL PROPERTIES OF COW BONE PARTICLE REINFORCED POLYESTER

COMPOSITE

I.C. Ezema1*, A.D Omah1, V.S Aigbodion1, Y. Suleman1, O. D Kalu1 & C.D Nebenu1

a) Department of Metallurgical & Materials Engineering,University of Nigeria, Nsukka

*Corresponding Author: e-mail: ike chukwu. ezema@ unn.edu.ng ; +234-8148320961

ABSTRACTThis research work studied the effect of different particle sizes of cow bone particles on some physical properties and mechanical properties of cow bone particulate filled polyester composite. The cow bone particles were sun dried for several days, ground and sieved to different micro sizes and used to reinforce unsaturated polyester resin. Hand lay-up technique was used in the composite manufacture. It was observed that the particle size of 300µm gave the best tensile modulus and best compressive strength while the particle size of 180µm gave the best tensile strength. The water absorption of the composite increased with increase in particle size while the density of the composite increased as a result of the reinforcement. The developed composite has good compressive strength and can be useful in application with minimum requirement on strength and light weight such as floor and wall tiles.

Key words: Cow bone particles, Particle size, Bio-composite, Mechanical Properties, Water absorption

1.0 INTRODUCTIONPolymeric composites are being used extensively in various applications from domestic house hold items to industrial, automobile and aerospace applications due to their high strength to low weight ratios. [1]. Pure polymeric materials do not have such wide application in engineering because of their low mechanical properties. These mechanical properties can be greatly improved by using techniques from nature

where most special structural materials are indeed a mixture of different materials which indicate composites with enhanced properties.

Polymers are often combined with fillers and/or fibers to improve their mechanical and/or physical properties. The fillers usually consist of wood flour, china clay, quartz powder or other powdered minerals. The filler is incorporated not only to improve the physical property of the composite but

JMME. Vol. 10 September 2015pp. 51-56


sometimes to reduce the polymer content and hence the cost. The costs of most commonly used composite fillers are high such as talc, calcium carbonate, kaolin, silica, and carbon black. Most of them are also non biodegradable.

The problem in most developing countries is that unused natural biomaterials are being treated as a waste instead of being used as an industrial input. Animal bones are of typical example and constitute a big environmental problem (environmental pollution) and so its productive use will rid the environment of these wastes. Indeed environmental, economic/cost issues are the basis for utilization of agricultural bio-waste products in the industry, and this has been the focus of considerable research in recent times [2-3].

In recent times, there has been increased legislation and awareness about our having a clean environment through clean energy and production of biodegradable materials. Cow bones seen in various abattoirs are dehydrated waste products only used for animal feed supplements in most developing countries. It presents a serious environmental problem if not properly disposed. Cows are killed every day and as such the volume of this cow bone generated daily is on the increase as only a limited quantity is used by animal feed mills. Its use in composite manufacture will introduce a new dimension to its applications and will not only help clean up our environment but will find value addition and good economic returns for cow owners/butchers.Recently, extensive research work has been

carried out on other biodegradable reinforcements such as natural fibers as replacement for synthetic fibers in polymer reinforced composites [4-8]. Many have also worked on the use of various particulates with outstanding results [9-14].

This paper is an extension of the search for biodegradable materials for use as polymer composite reinforcements for various applications. The focus is to investigate the effect of cow bone particle size on the physical and mechanical properties of an unsaturated polyester resin composite in order to determine its usefulness or otherwise in engineering applications. The very approach of drying and grinding them for use as composite reinforcement adopted in this research has not only provided a sure solution to environmental problems but has opened a door way to its value addition for various domestic and industrial applications.

2.0 MATERIALS AND METHODS2.1 MaterialsThe raw cow bones were obtained from Nsukka Ikpa/Ogige market abattoir. Polyester resin and Methyl Ethyl Ketone Peroxide (MEKP) were obtained from Ndidiamaka Chemicals Enugu, while Cobalt 2-ethylhexanoate and mould release agent were obtained from Manweb Nigeria Ltd Lagos. The metal moulds were supplied by ICE-JEB

Technical Services. Acetone mould cleansing agent was obtained from Lavans Chemicals Company Ltd, Nsukka.

(a)

(b)


Plate1:( a) Raw cow bone dumped as waste at Ikpa market Nsukka.(b) washed and sun-dried cow bone (c)

ground and sieved cow bone particles

2.2 Methods2.2.1 Conversion of Cow Bones into

Particulate Fillers The cow bones were collected from the abattoir at Ikpa/Ogige market, Nsukka, Nigeria. They were washed with detergent and a local spurge in a rotating bath to remove the oily part and traces of dirty contaminants. The washed cow bone was rinsed with clean water and sprayed on a mat at room temperature for 24 hours to remove water. They were sun dried in an open air hot sun environment for 5 days. Bone crusher was used to crush the bone to fine particles. Sieving was done using different sets of sieves having mesh sizes of 180µm, 300µm and above 300µm. After sieving, the different sizes were used as reinforcement of the matrix.

2.2.2 Manufacture of the CompositeFlat metal moulds with 4.5mm wall

thickness were prepared and hand lay–up method was used in making the composite. Here, the unsaturated polyester resin was measured into a beaker and the cow bone particles were added and then stirred vigorously until even dispersion was achieved. Addition of about 1.5% weight of catalyst and 1% accelerator were added and stirred for some time before casting the sample in the mould. Samples were made for unreinforced polyester resin and for180µm,

300µm and >300µm particle sizes at volume fraction of 40% each.

Plate2: Moulds for producing (a) Tensile Specimens (b) Compressive and impact specimens

Plate3: Sample of the developed composite

(c)

(c)

(b)

(a)

Properties of Cow-bone particle reinforced Polyester Composite 53

Various test specimens were cut according to ASTM standards for polymer composites: ASTM638 for tensile Test; ASTM D695 for compression Test and ASTM D370 for impact Test.

2.2.3 TestingThe tests carried out on the produced

samples are: Physical Properties Test (Density and Water Absorption test) and Mechanical Properties Test (Tensile, Compressive and Impact test).

Density TestingThe density of the samples was determined by first measuring the dimensions of the samples to calculate the volume and measuring on a digital weighing balance the mass of each sample. The density (ρ ) of each sample was obtained using the following equation

ρ = Mass

Volume (1)

Water Absorption Test: The water absorption was determined by first weighing each sample using a digital weighing balance to obtain the initial weight w1, Thereafter the samples were immersed in water for 24 hours, removed, dried with towel and then allowed to dry in an open air for 30 minutes at room temperature and the weight (W2) recorded. The percentage of water absorption was calculated according ASTM D570] as follows.

% Weight gained =

W 2−W 1

W 1x100

(2)where W2 is the wet weight and W1 is the dry weight of the samples

Tensile TestingHounsfield Tensometer was used for the tensile test. The samples were cut to standard (ASTM D638) and each sample was subjected to test by loading to its maximum load carrying capacity, after which the tensile

strength was calculated using the standard formula.

Compressive Test The compressive Strength was also tested using the Hounsfield Tensometer. The samples were cut to standard (ASTM D695-96). They were also subjected to their highest compressive load carrying capacity and the compressive strength was calculated using standard formula.

Impact Testing The Impact testing was carried out on the samples using the Charpy Impact testing machine to determine the impact energy. The specimens were cut according to standard (ASTM A370). Each sample was then placed on the machine, and the pendulum was allowed to hit the specimen when it swings under gravity. The Impact energy was obtained by reading energy loss of the pendulum as a result of hitting the sample, directly from the machine. The impact strength was calculated from the relation.

Gc=U/A (J/m 2) (3)

3.0 RESULTS AND DISCUSSIONSThe results of the tests conducted on the control sample (ctrl) and the developed composites with 180µm, 300m and above 400m air dried cow bone particle sizes are presented as follows

3. 1 Density of Composite


Fig 1.1: Effect of particle size on the density of sun dried cow bone- polyester composite.

From Figure 1.1, it is deduced that the density of polyester was increased as a result of the particles added to reinforce it. It was also observed that the smaller the particle size the higher the density such that 180µm developed composite is denser than others. This property is not favorable for applications where weight to strength ratio is of much consequence.

3.2 Water Absorption

Fig 1.2: Effect of particle size on the water absorption

of cow bone- polyester composite.

From Figure 1.2 it is observed that the higher the particle sizes of the cow bone particles, the higher the percentage of water absorption of the composite. This is because increase in the size of the particles increases the pore spaces for water percolation.

3.3 Tensile PropertiesThe result of the tensile strength and tensile modulus are presented in Figure 1.3a and Figure 1.3b respectively.

Fig 1.3: Effect of particle size on the (a) Tensile strength (b) Young’s Modulus of cow bone- polyester

composite.

From Figure 1.3, it is observed that the tensile strength decreased with increasing particle sizes of the cow bone particles. This is due to the fact that the finer particles develop a larger interfacial area with the matrix, thereby providing for more efficient load transfer to the particles. On the hand, there was a remarkable increase in the tensile modulus at 300µm particle size while other sizes of 180µm and 400µm did not show any improvement which is an indication of under reinforcement and over reinforcement respectively. This proves that the optimum reinforcement particle size is 300µm when stiffness is of primary consideration.

(b)

(a)

Properties of Cow-bone particle reinforced Polyester Composite

3.4 Compressive Strength

Fig 1.4: Effect of particle size on the compressive strength of cow bone- polyester composite.

From Figure 1.4, it is observed that variation in particle sizes of the cow bone particles does not show any linear relationship in their compressive strength. The sample with cow bone particle size of 300µm possesses the highest compressive strength.

3.5 Impact Strength

Fig 1.5: Effect of particle size on the impact strength of cow bone-polyester composite.

From Figure 1.5 presenting the Impact test results, there is no linear relationship between the impact strength of the samples and the particle size. The results show that impact strength decreased as the particle size increased: however there was a further increase of strength with particle size 400µm.

4.0 CONCLUSIONThe results of this investigation show the possibility of using cow bone particles as reinforcement in composites production in areas where compressive strength is a major requirement. In this research, the density increased with the reinforcements but decreased with increase in particle size. The particles size of 300m gave the best combination of tensile strength, tensile modulus and compressive strength but had the least impact strength. The water absorption of the composite increased with increase in particle size due to bigger pore spaces for water percolation as particles size increased. Due to good compressive strength, the cow bone particle /polyester composite can be a good replacement for high cost ceramic tiles for walls and floors.

REFERENCES1 Satyanarayana, K.G., Pai, B.C, Sukumaran, K.

and Pillai, S.G.K.: Fabrication and Properties of Lingocellulosic Fiber–Incorporated Polyester Composites; Handbook o f Ceramics a nd Composite: (Synthesis and Properties), Marcel Dekker Inc. New York, Vol 1. pp. 339 – 386. (1990)

2 Kaith, B.S. and Chauhan, A. A.: Synthesis, Characterization and Evaluation of the Transformations in Hibiscus Sabdariffa – Graft – Poly (Butyl Acryl ate): E-Journal o f Chemistry . Vol 5. No 21, pp. 980-986. (2008)

3 Danesh, M.A, Tabari H.Z, Reza H, Noradin N and Shams M.: Investigation of Morphological and Thermal Properties of Waste Newsprint/recycled Polypropylene Nanoclay Composite: BioResources Vol 7, No1, pp 936-945. (2012)

4 Maleque, M.A. and Behl, F.Y: Mechanical Properties Study of Pseudo – Stem Banana Reinforced Epoxy Composite. The Arabian Journal for Science and Engineering, Vol. 32, No. 2B, pp 359 – 364.(2006)

5 Zurale, M. M. & Bhide, S. J.: Properties of Fillers and Reinforcing Fibres, Mechanics of Composite Materials 34, 463-472. (1998)

6 Julson, J. L.; Subbaroa, G.; Stokke, D. D.; Gieselman, H. H. & Muthukumarappan, K.: Mechanical Properties of Bio Renewable Fibre/Plastic Composite, Journal of applied polymer science, 93, 2484-2493. ((2004)

7 Amer, A.A., Azza, E., Malash, G.F. and Nahla, A.T. (2007); Extensive Characterization of Raw Barley Straw and Study the Effect of Steam Pretreatment: Journal of Applied Science Research. Vol 3. No11, pp. 1336-1342

8 Hayes, B. S. & Seferis, J. C.: Modification of Thermosetting Resins and Composite through Preformed Polymer Particles, A Review” Polymer Composite 22, 451-467. (2001),

9 D’Almeida, J. R. M. & Manfredini, B. H. P: Hardness Evaluation of Epoxy Resin Filled with Mineral waste, Journal of Applied Polymer Science 84, 2178-2184, (1987).

10 Aigbodion V. S., Hassan S. B., & Oghenevweta E. J.: Potential of Maize Stalk Ash as Reinforcement in Polyester Composites, Journal of Minerals Characterization & Engineering, Vol.11, No.4, pp.543-557, (2012)

11 Hee-Soo Kima, Sumin Kima, Hyun-Joong Kima and Han-Seung Yang,. Thermal properties of bio-flour-filled polyolefin composites with different compatibilizing agent type and content, Thermochimica Acta, 45(1): 181-188, (2006)

12 Justin R. Barone, F. Walter Schmidt, Christina F.E. Liebner:. Compounding and molding of polyethylene composites reinforced with keratin feather fiber, Composites Science and Technology, 65: 683-692.( 2005)

13 Vignesh J., C.M. Selvam. Experimental Evaluation of Wood dust Particulate Reinforced Polymer Composites, IRACST – Engineering Science and Technology: An International Journal (ESTIJ), ISSN: 2250-3498 Vol.5, No.4,(2015).

14 S. Shuhadah, M. Supri and H. Kamaruddin: Thermal analysis, water absorption and morphology properties of eggshell powder filled low density polyethylene composites, in: Proceeding of MUCET 2008, UniMAP, Kangar, Perlis, pp. 15–16. (2008)


BIODIESEL PRODUCTION USING WASTE COOKING OIL

Amaghionyeodiwe, C. A., Ikechukwu I., Igwe, J., Obi, A. I., Ikechukwu, R. & Ukakwu, P

Department of Mechanical Engineering, Michael Okpara University, Umudike

ABSTRACTIn this study, a domestic blender was used as the reactor to synthesize biodiesel from Used Cooking Oil (UCO). The transesterification process involved 1 litre of UCO, 250g of methanol, 6.5g of NaOH at a reaction temperature of 65⁰C and reaction time of 20 minutes and the process was catalysed by alkali base. The process was triplicated and average results evaluated. A high yield of 93.71% of UCO methyl ester (UCOME) was obtained after it was washed and dried. The UCOME produced was blended with petroleum in the following percentage by volume of the UCOME; 5%, 10%, 20%, and 30% corresponding to B5, B10, B20, and B30 respectively. The fuel properties of UCOME and its blends were measured and found to satisfy both ASTMD6751 and EN14214 Standards for biodiesel fuel, and other reported works from different authors. From the biodiesel blends obtained, B30 and B20 proved to be more efficient compared to other blends due to their satisfactory properties such as their densities (0.847 and 0.846) g/cm respectively ³ and flash point of 145 C and 136 C respectively.⁰ ⁰

1.0 INTRODUCTIONBiodiesel fuel has become attractive throughout the world due to increasing knowledge of fossil fuel depletion and concerns of environmental protection. It is a renewable fuel, non-toxic and does not spoil water quality. It is also biodegradable. Used cooking oil is readily available and could be employed in the production of biodiesel rather than discarding it. It can also be used directly in most diesel engines without requiring extensive engine modifications (Nada Elsolh,2011). Used cooking oil is one of the economical sources for biodiesel production. However, the products formed during frying, can affect the transesterification reaction and the biodiesel properties. Nevertheless the production of biodiesel from waste vegetable oil offers economic, environmental and waste management benefits. Producing biodiesel from used frying oil is environmentally beneficial, since it provides a cleaner way for disposing these products; including valuable cuts in CO2. Its other advantages include reduction of sulphate, hydrocarbon emissions and particulate matter. It runs a diesel engine just as petroleum-based diesel would. The increase in population and improved economic activities the world over have led to huge increase in energy demand. The major part of all energy consumed worldwide comes

from fossil source. However, these sources are limited. The increasing depletion of world petroleum reserves, uncertainties concerning availability of petroleum source and its increasing environmental effects have inspired the need for a cleaner renewable energy source which should be environmentally friendly. Nigeria’s transport sector mainly depends on fossil fuels. This is particularly worrying because fossil fuels are the chief culprit implicated in the environmental issue of climate change phenomenon commonly referred to as global warming. This problem has resulted in intense search for alternative renewable energy considered to supplant the dwindling conventional Transportation fuels in Nigeria. Fatty acid methyl ester (FAME) commonly referred to as Biodiesel shows great potential as substitutes for petroleum diesel. (Anitha, 2012).Biodiesel, an alternative diesel fuel is made from renewable biological sources such as vegetable oils, animal fat, or used cooking oils. This fuel is biodegradable and non-toxic, and has low emission profiles as compared to petroleum diesel. It is synthesized from direct transesterification of vegetable oils. It is a renewable fuel comprising of mono-alkyl esters of long chain fatty acids, derived from a renewable lipid feedstock (Sharma and Singh, 2008). It can also reduce greenhouse gas effect and does not contribute to global


57

warming. Biodiesel has become a substitute for petroleum diesel because of its similar properties to the traditional fossil diesel with little or no engine modification (Ma and Hamna, 1999, Oliveira et al, 2006). It is essentially free of sulphur, making it a cleaner burning fuel than petroleum diesel with less emission of carbon monoxide (CO) and unburnt hydrocarbons. It can be used in its pure form B100 or can be blended with petroleum diesel. Technically, biodiesel is better used as a blend fuel in order to improve its performance.

Biodiesel can be produced from several vegetable oils and animal fats with a varied composition in fatty acids. However, biodiesel produced from vegetable oils have more economical setback due to high cost of pure vegetable oils. This has led to high cost of production of biodiesel making it more expensive than petroleum diesel. Because vegetable oils are a source of staple food, their usage for biodiesel production may also result in high cost of food supply, as more land will be needed for the cultivation of oil crops, in order to tackle the challenges between food supply and fuel. One of the means to address the higher price hurdle of biodiesel production is to produce biodiesel from waste fats and oils (Anitha, 2012). Used Cooking Oils (restaurant greases) can be recycled and used to produce biodiesel. This will significantly reduced the dependence on edible vegetable oils for biodiesel production and will as well reduce the cost of production. Used Cooking Oil is readily available from homes, restaurants, bakeries, etc, and is affordable at low or no cost. In addition, it is on record that millions of tons of used cooking oils are disposed annually in manners

that contaminate the environment (Samuel et al, 2013). Therefore, the proper utilization and management of Used Cooking Oils (UCO) will pave way for a conducive environment.

2.0 MATERIALS AND METHODSBiodiesel fuel can be produced through the chemical reaction of transesterification and esterification in which vegetable oil or animal fats are reacted with alcohol (methanol or ethanol) under the influence of a catalyst (acid, alkali base or enzyme). The ratio of oil is an important variable affecting the yield of methyl ester. The stoichiometry of the transesterification reaction requires 3moles of alcohol per mole triglyceride to yield three moles of fatty esters and one mole glycerol. Alkali metal alkoxides are the most effective transesterification catalyst compared to the acidic catalyst. Sodium hydroxide is among the most efficient catalysts used for this purpose, although KOH can also be used. Trans-methylations occur faster in the presence of an alkaline catalyst than those catalysed by the same amount of acidic catalyst as has been reported (Refaat et al., 2008, Mohammed et al, 2012, Samuel et al., 2013, Kumar et al., 2014).Due to high free fatty acid of Used Cooking Oil, it is necessary to carry out titration test of the oil in order to determine the additional catalyst (NaOH) needed for the conversion of triglyceride into methyl esters. The oils that do not contain FFA require 3.5g of NaOH per litre of oil as catalyst. The excess FFA demands additional NaOH for neutralization which is determined by titration test. High FFA content leads to soap formation during the transesterification process.


The materials used in this work include:(a) Used Cooking Oil(b) Sodium Hydroxide(c) Isopropyl Alcohol(d) Distilled Water(e) Laboratory Glass Wares(f) Laboratory Apparatus(g) Domestic Blender(h) Electric Hot Plate

The major feedstock used in this experiment is Used Cooking Oil (UCO), which was donated by Mr Biggs’s (UAC Restaurants Limited), Aba Road, Umuahia, Abia State, Nigeria. The fatty acid composition of the Used Cooking Oil is shown in Table 1. Methanol and Isopropyl alcohol used are analytical grade products of BDH Chemicals LTD, Poole, England, and Sodium hydroxide (pellet) were supplied by the Chemistry Department. The Distilled water was obtained from the Manesty Distillation Plant at the Agricultural and Bio-resources Engineering laboratory, MOUAU.

Table 1: Fatty Acid composition of Used Cooking Oil (From: Obi, 2000)

FATTY ACID COMPOSITION

PERCENTAGE (%)

Lauric, C12:0 3.0Myristic-oleic, C14:1 2.3Palmitic, C16:0 7.3Stearic, C18:0 1.9Oleic, C18:1 12.4Linoleic, C18:2 9.6Linolenic, C18:3 15.4Others (unidentified) 48.2

The Laboratory glassware include 250ml and 500ml beakers, 250ml conical flask, 1000ml conical volumetric flask (Pyrex), graduated cylinder and pipette, 1000cm² volumetric flask, burettes, separating funnel, stirrer, syringe, thermometer and test jar.The blender used was a multi Dry and Wet Master Chef Blender, with a clear gas jar of 1.25 Litres capacity. 6.8KW Ceran Electric Hot Plate, Analytical Weigh Balance, spatula, retort stand, and clamp, litmus paper, filter cloth and funnel.

2.3 Experimental Procedure

The experimental procedure for the production of biodiesel includes:(a) UCO Pre-treatment.(b) Titration Test.(c) Sodium Methoxide production.(d) Alkali base Transesterification Reaction.(e) Settling and Separation.(f) Purification.

2.3.1 Pre-Treatment of the Used Cooking OilThe Used Cooking Oil (UCO) was filtered with a filter cloth to remove debris and other solid materials and then heated up to 30⁰C.

2.3.2 Titration TestOne gram of NaOH was dissolved and made up to the mark in 1000cm³ volumetric flask with distilled water. The oil solution was prepared by mixing oil and isopropanol in the ratio of 1:10 respectively to form the mixture. The pH was determined to be 8.62 using pH meters; then 20ml of isopropyl/oil solution mixture was measured out into a 250cm³ conical flask and the NaOH solution was titrated against the mixture (isopropyl/oil solution). The titration test was repeated three times and the result of the average was determined as the percentage of free fatty acid (FFA) content of the Used Cooking Oil and the extra catalyst needed to neutralize the acid was computed using the expression (1) below:

X + 3.5g NaOH (1)where X = the titrated value (NaOH/distilled water solution) and 3.5g NaOH is the standard catalyst required per 1000ml oil for transesterification reaction.

2.3.3 Sodium Methoxide ProductionSix grams of sodium hydroxide pellet was measured into a 250ml beaker; 200ml of methanol was also measured out in a separate 250ml beaker. The methanol was gradually added to the sodium hydroxide, and the mixture was stirred for 13 minutes until the sodium hydroxide pellets completely dissolved to form sodium methoxide. Equation (2) below shows the formation of sodium methoxide solution. CH3OH + NaOH = CH3ONa +H2O (2)

Biodiesel Production Using Waste Cooking Oil 59

TriglycerideMethanol Methyl EsterGlycerol

Figure 1: Equation for Transesterification Reaction

(3)

2.3.4 Alkali Base Transesterification reaction1000ml of Used Cooking Oil (UCO) was measured into a conical flask and was preheated to the temperature of 65⁰C. The prepared sodium hydroxide solution was poured into the Wet Mill Jar of the blender, and the preheated Used Cooking Oil was carefully added. The Wet Mill lid was

securely tightened and the blender switched ON. The blender was operated at the highest speed for complete agitation of the reagents for 20 minutes before the blender was switched OFF. Figure 1 below represents the transesterification reaction for the conversion of 1mole of triglyceride and 3 moles of ethanol to 3 moles of methyl Ester and 1 mole of Glycerol respectively.

2.3.4 Settling and SeparationThe mixture was poured from the Wet Mill of the blender into a separating funnel clamped on the retort stand and the lid was tightly secured. The mixture was allowed to stand overnight for complete separation and settling of the expected products (biodiesel and glycerol) respectively. The upper layer which is the methyl ester was a pale gold biodiesel while the lower layer of dark brown is the glycerol that settled at the bottom of the separating funnel. The valve separating the funnel was opened and the glycerol was discarded into the beaker.

2.3.5 PurificationThe biodiesel produced was purified firstly, by washing it with warm distilled water. The distilled water used was warmed up to 40⁰C; equal ratio of biodiesel and warm distilled water was mixed in the separating funnel by first pouring the methyl ester, then adding the warm distilled water. The separating funnel was tightly secured and the separation funnel was shaken slowly for proper mixing and to

avoid emulsion of air bubbles that causes delay in separation of biodiesel from the washed impurities. The separating funnel was clamped on a retort stand and allowed to stand until proper separation was observed. The milkfish layer that contains the impurities at the bottom of the separating funnel was carefully drained by opening the tap of the separating funnel leaving the biodiesel that is the upper layer in the separating funnel. This sequence was repeated four times until a clear biodiesel with water was observed. The washed biodiesel was heated to 110⁰C to evaporate off the water and methanol present in the fuel.

Table 2: Preparation of Biodiesel Blends


COMPOSITION OF BLENDS VOLUME OF MIXTUREB5 95ml of petroleum diesel + 5ml of methyl-esterB10 90ml of petroleum diesel + 10ml of methyl-esterB20 80ml of petroleum diesel + 20ml of methyl-esterB30 70ml of petroleum diesel + 30ml of methyl-esterB100 100ml of methyl-ester

2.3.6 Biodiesel Blend PreparationThe biodiesel produced from UCO was blended with petroleum diesel oil using direct blending method. The biodiesel blends were prepared according to the measured percentage shown in Table 2, by placing the mixtures in a transparent bottle.

2.4 Property Test for BiodieselThese properties of the methyl-ester blends produced were tested in accordance with ASTM D6751 and EN14214 standards. They are:

(a) Density(b) Flash and Fire points(c) Viscosity(d) Cloud and Pour Points

2.4.1 Test for DensityThe densities of the samples were determined at ambient temperature (28⁰C) for each blend. A density bottle of mass 50ml was weighed on the analytical balance and the initial weight of the bottle was noted. The samples were put in the density bottle, the spill cleaned and dried, and the bottle weighed on the analytical balance. The densities were measured in triplicates and the average values were recorded as the results. The densities of the samples were computed using appropriate equations.

2.4.2 Test for Flash Point/Fire PointThe Flash point of the sample was determined using the setup comprising a hot – plate, beaker and thermometer. 80ml of the sample was placed in a transparent Pyrex beaker resting on a hot plate as the source of heat and a thermometer was introduced clamped on a retort stand. Heat was applied gradually by turning the knob of the hot plate until the observed movement of the particles increased. A flame was gradually brought close to the surface of the beaker until “a catch and disappearing” flame on the surface of the hot liquid was observed. The temperature at which this happens is noted. Further heating of the samples with gradual moving of the flame closer to the vapour was continued and the temperature at which the sample ignites without the flame disappearing from the surface of the sample was recorded as the Fire Point.

2.4.3 Test for ViscosityThe viscosities of the samples were determined using a glass capillary kinematic viscometer. The viscometer was tightly

clamped on a retort stand. 100g of each sample was collected into a Pyrex beaker and was gradually heated to a temperature above 40 C. The sample was then⁰ transferred into the viscometer through the larger opening of the capillary tube and the fluid was allowed to cool until a temperature of 40 C was reached. Thereafter,⁰ suction was applied to the other end of the capillary tube to draw the fluid to the mark on the upper meniscus level of the capillary tube.The fluid was allowed to run freely to the lower meniscus mark in the capillary tube. The efflux time for the fluid to flow from the upper meniscus mark to the lower meniscus mark was determined with the aid of a stopwatch. The test was triplicated for each sample and the kinematic viscosity was calculated from the formula below:Kinematic Viscosity = ktWhere k = Calibrated value of the viscometer expressed in square millimetres per second square (3.0) t = Flow time in seconds of the liquid.The viscosities of the samples were determined using a simple viscometer setup as shown in figure 2:

Figure 2: Simple Viscosity Setup Apparatus

2.4.5 Test for Cloud and Pour Points


The cloud and pour points of the sample were analysed by reducing their temperature. The samples were introduced into transparent cylindrical glasses and placed in a cooling bath containing crushed ice.The temperature of the samples was monitored as it decreased; the temperature at which the clear sample loses its sharpness due to cloud formation was noted as the sample’s cloud point. The sample was further cooled in order to pour it at every increase in temperature. The sample was inspected at every one-minute interval by tilting it horizontally. The temperature at which the samples cannot pour freely for five seconds was noted as the pour point.

3.0 RESULTS AND DISCUSSION

Table 3: Titration Results

The results from the titration test were determined from the average of the titre values as shown in Table 3.

2.5g of NaOH was added to 3.5g NaOH as the required catalyst, while 6.5g NaOH was used to react with 1litre of UCO and 250g of methanol to convert the free fatty acid (FFA) into methyl – ester known as biodiesel.The alkali base transesterification reaction of the Used Cooking Oil and Sodium Methoxide at a ratio of 4:1 of UCO to methanol as stated previously, yielded 93.71% increase in biodiesel production after 20 minutes. The collected by-product (glycerol) was determined to be 177.23g that could be further processed to produce soap. It was observed also that 142.13g of the total reacting masses accounted as losses which could be due to unreacted methanol, residual catalyst and the emulsion removed at the water washing stage of the production process. The results stated are averages of three different experimental runs. Detailed results for each of the experimental runs are presented in Table 4

3.1 Properties of Biodiesel BlendsThe physical properties of the biodiesel

blends such as its Viscosity, Density, Cloud Point, Pour Point, Flash Point, and Fire Point were determined and compared with the ASTMD06751 Standard and EN 14214 Standard for biodiesel as shown in Table 5. It is seen that the fuel properties of biodiesel

Table 4: Results of Experimental Runs

Table 5 Properties of Biodiesel Blends and Used Cooking Oil


Biodiesel Production Using Waste Cooking Oil

TEST TITRE VALUE (ml)1st Run 2.402nd Run 2.53rd Run 2.54th Run 2.6

Average 2.5

EXPERIMENTAL CONDITIONS 1ST RUN 2ND RUN 3RD

RUNAVERAGE

Reaction Temperature (⁰C) 65 65 65 65Reaction Time (min) 20 20 20 20Used Cooking Oil (UCO) quantity (g) 1000 1000 1000 1000Methanol Quantity (g) 250 250 250 250NaOH (Catalyst) concentration (g) 6.5 6.5 6.5 6.5UCO Biodiesel obtained (g) 925.20 955.80 930.40 937.13Glycerol obtained (g) 188 164.70 179 177.23Losses (g) 143.3 136 147.10 142.13UCO Biodiesel Yield % 92.52 95.58 93.04 93.71

PROPERTIES UCO B100 B30 B20 B10 B5 ASTMD6751

EN 14214

Kinematic Viscosity @ 40 C⁰ 36.94 8.344 5.690 5.309 5.309 5.073 1.9 – 6.0 3.5 – 5.0

Specific Gravity @ 28 C⁰ 0.910 0.882 0.847 0.846 0.840 0.834 0.860 – 0.9 0.860 – 0.9

Flash Point ( C)⁰ 259 210 145 136 132 126 130min 120minFire Point ( C)⁰ 263 238 165 153 143 138 - -Cloud Point ( C)⁰ 7.2 3.6 2.9 2.7 2.6 2.4 Reported ReportedPour Point ( C)⁰ 5.4 2.4 1.6 1.5 1.4 1.2 Reported Reported

blends are in agreement with the ASTMD6751 and EN14214 Standards, and other reports from some authors.

3.2 Flash PointThe flash point of a fuel is the lowest temperature at which the fuel will ignite when exposed to a flame or spark. Biodiesel fuel has a higher flash point over conventional diesel fuel (74⁰C – 80⁰C). The biodiesel produced from UCO (B100) has a flash point of 210⁰C and this satisfies the minimum requirement of ASTMD6751 and EN14214 standard as provided in the table 4.4.

The

flash point of the biodiesel produced is higher than those reported by Samuel et al (139⁰C),

Igbokwe and Nwafor (150 C)⁰ and Alamu et al (167 C⁰ ) as shown in the table 4.4. B20 (123 C⁰ ) of Ibekwe and Nwafor is lower than the minimum requirement as reported by ASTMD6751 (130 C) compared to the⁰ B20 (136 C) of this work which is⁰ within the range of the required standard. However, B5 indicates the lowest flash point 126⁰C which is slightly below the minimum requirement of ASTMD6751 standard and above that of EN14214 standard as shown in Table 5Furthermore, UCO displays a great rise in flash point compared to the biodiesel blends as shown in Figure 3. In addition, the high flash point of UCO biodiesel and biodiesel blends compared to petroleum diesel makes it

safer to handle and for storage.

Table 6: Characterization of experimental Used Cooking Oil Methyl-Ester in comparison with other standards

Figure 3: Flash Point of Biodiesel Blends and Used Cooking Oil

PROPERTIES UCO METHYL ESTER(B100)

B20 IBEKWEAND

NWAFOR

SAMUEL ET AL

ALAMU ET AL

ASTMD6751

METHYL ESTER

METHYL ESTERB100 B20

Kinematic Viscosity @40 C⁰ 8.344 5.309 2.99 2.04 4.72 4.839 1.9 – 6.0

Specific Gravity @ 28 C⁰ 0.882 0.846 0.865 0.846 0.879 0.883 0.860 – 0.90Flash Point ( C)⁰ 210 136 150 123 139 167 130minFire Point ( C)⁰ 238 153 - - - - -Cloud Point ( C)⁰ 3.6 2.7 7.5 3.0 - 6 ReportedPour Point ( C)⁰ 2.4 1.5 0.0 -12 - 2 Reported

3.3 Fire PointThe fire point of a fuel is the lowest temperature at which the vapour produced by that given fuel will continue to burn for at least five seconds after ignition by an open flame. The fire point of the samples were determined and found to be more than 10 C higher than the flash point⁰ as shown in Table 5. The fire point for biodiesel is significantly higher than that of conventional fuel; this is more reason why biodiesel is safer than petroleum diesel. 3.4 Density Density is one of the desirable properties of a fuel; it is an important parameter for diesel fuel injection system. Fuel density affects engine performance, as the density increases the energy content of the fuel also increases. Higher density of fuel results to low volatility and poor atomization of biodiesel during fuel injection in combustion chamber causing incomplete combustion and carbon deposits in combustion chamber.

Figure 4: Density for the Biodiesel Blends and UCO

Figure 5: Viscosity of Biodiesel Blends and Used Cooking Oil


From the results presented in Figure 4, the density recorded for B100 is within the range limit for both ASTMD6751 and EN 14214 standards for biodiesel. From Figure 4, UCO has a density 0.910g/cm³, which is above the range of ASTMD6751 and EN14214 Standards for biodiesel, which when used may result to engine ignition delay. Moreover, the density of Used Cooking Oil B100 (0.882g/cm³) was observed to be in good agreement with that of Samuel et al and Aluma et al. However, the remaining blends (B5, B10, B20, and B30) were slightly lower than the given standards for biodiesel. The B20 of this research work is of the same value (0.846g/cm ) with that of³ Igbokwe and Nwafor (9999), this value 0.846g/cm compared to³ ASTMD6751 and EN14214 (i.e. 0.883 and 0.860 – 0.90)g/cm³ respectively for biodiesel indicates the lower density of petroleum diesel.3.5 ViscosityViscosity is a measure of the internal flow resistance of a liquid. The viscosity of biodiesel should be low for good atomization of fuel spray. This is the primary reason why biodiesel is used as an alternative fuel instead of neat vegetable oil and animal fat that will ultimately lead to engine operational problems when used directly as fuel. The kinematic viscosity of B100 measured at 40⁰C was 8.344mm²/s and was found to be high compared to the acceptable standards. However, the viscosity of the blends (B5, B10, B20, and B30) falls within the ASTMD6751 Standards and slightly above EN 14214 Standard for biodiesel. The viscosities comparison for the produced biodiesel blends and Used Cooking Oil as shown in Figure 5 revealed that UCO at 40⁰C has a viscosity of 39.944mm²/s which is very high compared with that recorded for ASTMD6751 and EN 14214 Standards. This is the major reason why vegetable oil or UCO cannot be used directly in an engine due to its high viscosity.

3.6 Cold Flow Characteristics

The cold flow properties of the UCO methyl ester blends were measured by determining the cloud point and pour point. The cloud point and pour point are important low-temperature fuel parameters. The cloud point as a desirable fuel property indicates the temperature at which the solidification of heavier components of biodiesel result in the formation of cloud of crystals. It is observed from Table 6 that the cloud point of B100 (3.6) is lower than those reported by Alumu et al (6⁰C), Igbokwe and Nwafor (7.5 C). In addition, the B20 of⁰ Igbokwe and Nwafor (3.0 C) is⁰ slightly higher than the B20 (2.7 C)⁰ of this study.While the pour point of a fuel indicates the lowest temperature, at which the wax becomes visible when the fuel is cooled to a certain temperature. It is the lowest temperature at which the fuel can flow. Specifications for cloud point and pour point are not in the biodiesel standards, although ASTMD6751 and EN 14214 requires that the cloud point be reported, this is because each country has different climatic conditions. As presented in Table 6, the UCO methyl ester (B100) has a satisfactory cloud point (3.6⁰C) and pour point (2.4⁰C) compared to Alamu et al, 6⁰C and 2⁰C respectively for cloud point and pour point. However, the cloud point and pour point reported by Igbokwe and Nwafor (0.0 C⁰ and -12 C⁰ ) respectively is significantly lower than the determined cloud and pour points.

4.0 CONCLUSIONBiodiesel was successfully produced from Used Cooking Oil through alkali catalysed transesterification reaction using methanol in the presence of NaOH at a reaction time of 20 minutes. The produced methyl-ester was characterised and from the obtained results, the methyl esters produced can be effectively used in a diesel engine since they fulfil the requirement of ASTMD6715 and EN14214 Standards for biodiesel fuel. The UCO has been proved to be a good feedstock for biodiesel as high yield of 93.71% was obtained and fuel properties were within the standard prescribed by ASTM and EN. The densities of the blends (B30, B20, B10, and B5) are closer


to that of petroleum diesel but they have higher viscosity compared to petroleum diesel.

4.2 RECOMMENDATIONBiodiesel produced from Used Cooking Oil could be used in a diesel engine without engine modification. However, it is advisable to use biodiesel in its blended form. The blends B20 and B30 from this study have satisfactory properties compared to petroleum diesel and therefore should be recommended for use in any diesel engine.In addition, for the purpose of mass production, a better separation technique should be used in order to reduce the separation time such as a commercial centrifugal separating machine.The production of biodiesel should be encouraged in order to increase employment opportunities, reduce over dependency on petroleum diesel, and reduce global warming.

REFERENCES (1) A.A. Refaat, N. K. Attia, H. A. Sibak, S. T. El

Sheltawy, and G. I. (2008). Production optimization and quality assessment of biodiesel from waste vegetable oil, International Journal of Environmental Science and Technology, vol. 5, no. 1, pp. 75–82.

(2) A. Anitha (2012). Transesterification of Used Cooking Oils Catalyzed by CsTPA/SBA15 Catalyst System in Biodiesel production. IACSIT International Journal of Engineering and Technology, Vol. 4, No. 1, (February 2012).

(3) A.Jimoh, A.S, Abdulkareem, A.S. Afolabi, J.O. Odigue and U.C. Odili (2012). Production and Characteristics of Biofuel from Refined Groundnut Oil. Pp 21-28.

(4) Antolin, G., F. V. Tinaut, Y. Briceno, V. Castano, C. Perez, and A. I. Ramirez. (2002). Optimization of biodiesel production by sunflower oil transesterification. Bioresource Technology 83(2):111-114.

(5) Arjun B. Chhetri 1, K. Chris Watts 2 and M. Rafiqul Islam (2008). Waste Cooking Oil as an Alternate Feedstock for Biodiesel Production. Energies 2008, 1, 3-18; DOI: 10.3390/en1010003.

(6) Beatrice Onyinye Ojiego1, O.C. Onyial and F.W. Abulraman. Production of Biodeisel from used Vegetable Oil.

(7) Bello, A. O. Fade-Aluko, S.A Anjorin and T.S. Mongaji (2011). Characteristics and evaluation of

African bush mango nut (Dika nut) (Irvingia gabonensis) oil biodiesel as alternative fuel for diesel engines. Journal of pertroleum Technology and Alternative Fuel vol.2 (9), pp. 176-180, September 2011.

(8) Canakci, M. (2007). The potential of restaurant waste liquids as biodiesel feedstocks. Bioresource Technology, vol.98, No.1,(January, 2007), pp. 183-190, ISSN 0960-8524.

(9) Cetinkaya M. and Karaosmanoglu F., (2004), Optimization of base-catalyzed transesterification reaction of used cooking oil, Energy Fuels 18, 1888-95.

(10) Darnoko, M. P.; Ballesteros, E.; Lopez, F. J.; and Mittelbach, M. (2004). Optimization of alkali-catalyzed transesterification of Brassica Carinata oil for biodiesel production Energy and Fuels 18, [77-83].

(11) Dossat, V.; Combes, D.; Marty, A. Continuous enzymatic transesterification of high oleic sunflower oil in a packed bed reactor: influence of the glycerol production. Enzyme Microbe. Technol. 1999,25, 194-200.

(12) F. Ma and MA. Hanna. (1999). Biodiesel production: a review. Bioresource Technology. 70 (1): 1-15..

(13) Granados, M. L., Poves, M. D. Z, Alonso, D.M., Mariscal, R., Galisteo, F. C., Moreno-Tost, R., Santamaria, J. and Fierro, J. L. G. (2007). Biodiesel from sunflower oil by using activated calcium oxide. Applied catalysis B: Environmental 73 (3-4): 317-326.

(14) Gemma, V.; Mercedes, M.; Jose, A. Optimization of brassica carinata oil Methanolysis for biodiesel production. J. Am. Oil Chem. Soc. 2005, 82, 899-904.

(15) Ginzburg BZ (1993) Liquid fuel (oil) from halophilic algae: a renewable source of non-polluting energy. Renew Energy 3:249252. Dio:10.1016/0960-1481(93)90031-B.

(16) Haas, M.J., Scott, K., Marmer, W. N., and Folia, T.A., (2004). In situ alkaline transesterification: An effective method for the production of fatty acid esters from vegetable oils. JAOCS, 81(1),82.

(17) Helwani US Jr, Othman Mr, Aziz Mr. Fernando WJN and Kim J (2009). Technologies for production of biodiesel focusing on green catalytic techniques: a review. Fuel Processing Technology. 90: 1502-1514.

(18) Hilbert T., Mittelbach M., Schmidt E., (2005). Animal fats perform well in Biodiesel, 5th

International Colloquium on Fuels in Germany held January 12-13.

(19) Hem, C.J.; Joe, T.; Terry, W. Optimization of cottonseed oil ethanolysis to produce biodiesel


high in gossypol content. J. Am. Oil Chem. Soc. 2008, 85, 357-363.

(20) Issariyakul, T.; KulKarni, M. G.; Dalai, A. K.; Bakhshi, N.N. Production of biodiesel from waste fryer grease using mixed methanol/ethanol system. Fuel processing Technology 2007, 88, 429436.

(21) Karmee S. J. and A. Chadha (2005) Preparation of biodiesel from crude oil of Pongamia pinnata. Bio resources Technology 96(4): 1425-1429.

(22) Kemp, W. H. (2006). Biodiesel: Basics and beyond: A Comprehensive guide to production and use for the home and farm. Tamworth, Ont.: Aztext Press.

(23) Korus, R. A., Hoffman, D.S., Bam, N., Peterson, C.L. and DROWN, D.C. (1993). Transesterification process to manufacture ethyl ester of rape oil. In Proceedings, First Biomass Conference of the Americas: Energy, Environment, Agriculture, and Industry. Vol. 11. Ed. National Renewable Energy Laboratory. Nrel, Golden, Colorado, pp. 815-826.

(24) Knothe, G. (2006). Analysing biodiesel: Standards and other methods. Journal of American Oilk Chemists Society, 83(10), 823-833.

(25) Kumar Ved and Kant Pandan (2014). Differential Studies of Alkali Catalysed Production of Biodiesel from Sorghum Oil. Research Journal of Chemical Sciences ISSN 2231-606X Vol. 4(1), January (2014). 66-70.

(26) Janaun, J. and Ellis,N. (2010). Perpestives on biodiesel as a sustainable fuel. Renewable and Sustainable Energy Reviews Vol. 14, No. 4,(May 2010), pp. 1312-1320, ISSN 1364-0321.

(27) Jitputti, J.; Kitiyanan, B., Rangsunvigit, P., Bunyakiat, K., Attanatho, L., and Jenvanitpanjakul, P.(2006). Transesterification of crude palmernel oil and crude coconut oil by different solid catalysts, Chem. Eng. J., 116,61-66.

(28) Karmee S.K., Chadha A. Preparation of biodiesel from crude oil of Pongamia Pinnata . Bioresour Technol 2005; 96:1425-9.

(29) Kerihuel A. Kumar, M. S., Bellettre, j., and Tazerout, M.(2006). Ethanol animal fat emulsions as a diesel engine fuel-Part 1: Formulations and influential Parameters. Fuel, Vol. 85, No. 17-18, (December 2006),pp. 2371-2684, ISSN 0016-2361.

(30) Kose O., Tuter M., and Ayse H. A. (2002), Immobilized Candidaantarticalpase-Catalyze dalcoholysis of cottonseed oil in a solvent-free medium. Bio. Res. Technol. 83, 125D129.

(31) Knothe , G., Van Gerpen, J., Krahl, J. (2005). The Biodiesel Handbook. AOCS Press, Champaign, Illinois.

(32) Lance, D. L. ; Goodfellow, C. L.; Williams, j.; Bunting, W.; Sakata, I.; Yoshida, K.; Taniguchi, S.; Kitano, K. The Impact of Diesel and Biodiesel Fuel Composition on a Euro V HSDI Engine with Advanced DPNR Emissions Control; SAE Technical Paper No. 2009-01-190; SAE International: Washington, DC, USA, 2009.

(33) Leung, D. Y. C.; Guo, Y. Transestrification of meat and used frying oil: Optimization for biodiesel production. Fuel processing Technology 2006, 87, 883-890.

(34) Licht, F.O. (2009). World-Biodiesel Production, World Ethanol and Biofuels Report, (4), 16:365.

(35) Lin, C. Y.; Li, R. J. Engine performance and emission characteristics of Marine fish-oil biodiesel produced from the discarded parts of marine fish. Fuel Process. Technol. 2009, 907, 883-888.

(36) Li L, Du W, Liu D., Wang L, Li Z. (2006). Lipase-catal. Lipase-catalysed trans-esterification of rape seed oils for biodiesel production with novel organic solvent as the reactor medium. J Mol catal B: Enzym 2006, 43:58-62.

(37) Ma, F., Clements, L. d., and Hanna, M. A. (1998). The effects of Catalyst, free fatty acids, and water on trans-esterification of beef tallow, Transaction of the ASAE, Vol. 41, No. 5, pp. 1261-1264, ISSN 0001-2351.

(38) M. Martinez, and J. Aracil (2004). Integrated Biodiesel production: a comparison of different homogeneous catalysts systems, Bio resource Technology, vol. 92, pp. 297-305

(39) Meher L. C. Sagar D. V. Naik S. N. (2006). Technical aspects of biodiesel production by trans esterification: A review. Renew Sustain Energy Rev. 10(3): 248-268. 1314.

(40) Meneghetti, S. M. P., Meneghetti, M. R., Woif , C. R., Silva, E. C., Lima, G.E.S., Coimbra, M.A., Soletti, J. I. , and Carvalho, S.H.V.(2006). Ethanolysis of castor and cotton seed oil: a systematic study using classical catalysts. Journal of the American Oil Chemists Society, Vol. 83, No. 9, pp. 819-822, ISSN 1558-9331.

(41) Miao X, Wu Q (2006) Biodiesel production from heterotrophic microalgal oil. Bio resources Technol 97:841-846. Doi:10.1016/j.biortech. 2005.04.008.

(42) Mohammed I. Jahirul 1, Richard J. Brown 1, Wijitha Senadeera 1, Ian M. O’Hara 2 and Zoran D. Ristovski. (2009). The Use of Artificial Neural Networks for Identifying Sustainable Biodiesel Feedstocks. Energies review. pp. 31, 74-80. ISSN 1996-1073.

(43) Mohammed I. Jahirul 1, Richard J. Brown 1, Wijitha Senadeera 1, Ian M. O’Hara 2 and Zoran D. Ristovski. (2013). The Use of Artificial Neural


Networks for Identifying Sustainable Biodiesel Feedstocks. Energies review. pp. 31, 74-80. ISSN 1996-1073.

(44) Nada Elsolh (2011): Unpublished M.Sc thesis, Kassel and Cairo Universities.

(45) Oliveira J.S., Montalavo R., Dahar L., Suarez P.A.Z., Rubim, J.C. Determination of methyl ester contents in biodiesel blends by FTIR-ATR and FTNIR spectroscopies, Talanta, 69, 1278-1284.

(46) O. J. Alamu, M. A. Waheed and S.O. Jekayinfa (2007). Alkali-catalysed Laboratory Production and Testing of Biodiesel fuel from Nigerian Palm Kernel Oil. Agricultural Engineering International: the CIGR Ejournal. Manuscript Number EE 07

(47) 009. Vol. IX. July, 2007.

(48) P. Chitra, P. venkatachalam and A. Sampathrajan, “Optimization of experimental conditions for biodiesel production from alkali-catalysed trans esterification of jatropha curcus oil’’, Energy Development, Vol. IX, pp. 13-18, 2007.

(49) Pierre CK (2008). Development of a biodiesel industry in Rwanda: Processing and fuel quality control. Processing of the Workshop on biodiesel production and marketing, Kigali, Rwanda.

(50) Ramadhas A.S., Jayaraj S, Muraleedharan C (2004). Use of Vegetable oils as I.C. engine fuels a review: data bank. Rev. Enferm. 29:727742.

(51) Sahoo P. K., Das L. M., Babu MKG, Naik SN (2007). Biodiesel development from high acid value polanga seed oil and performance evaluation in a CI engine. Fuel 86:448-454.

(52) Selmi, B., Thomas, D., 1998. Immobilizedlipse-catalyzed ethanolsis of sunflower oil in solvent-free medium. J. Am. Oil Chem. Soc. 75, 691-695.

(53) Sharma Y.C., Singh B.D. (2007). Development of biodiesel fro karanja –a tree found in rural India, Fuel 87, 17401742.

(54) Sharma Y. C., Singh B. D. (2008). Development of biodiesel fro karanja- a tree found in rural India, Fuel 87, 17401742.

(55) Shah S, Sharma S and Gupta MN (2003). Enzymatic trans esterification for biodiesel production. Indian Journal of Biochemistry and Biophyysics. 40: 392-399.

(56) Solomon Giwa and clement Ogunbona (2014). Sweet almond (Prunus amygdalus “dulcis”) seeds as a potential feedstock for Nigerian Biodiesel Automotive Project. Ambiente and Agua An Interdisciplinary Journal of Applied Science ISSN 1980993X.

(57) Tashtoush GM, Al-Widyan MI, Al-Jarrah MM (2004). Experimental study on evaluation and optimization of conversion of waste animal fat into biodiesel. Energy conversion Management, 45:2697-711.

(58) Tickell, J. (2006). Biodiesel America: How to achieve energy security, Free America from middle-east oil dependence and make money growing fuel. ISBN 9780970722744, Yorkshire press, USA.

(59) Tiwari, A. K., Kumar, A., Raheman, H. (2007). Biodiesel production from Jatropha oil (Jatropha-curcas) with high free fatty acids: An Optimized process. Biomass Bioenergy, 31(8): 569-575.

(60) Van Gerpen, J., Pruszko, R., Clements, D., Shanks, B., Knothe, G. (2006). Building a Successful Biodiesel Business (2nd Edition). Biodiesel Basics ISBN: 0-9786349-0-X.

(61) Veljkovic, V.B., Lakicevic, S.H., Stamenkovic, O.S., Todorovic, Z. B., Lazie, K.L.,2006. Biodiesel production from tobacco (Nicotiana tabacum L.) seed oil with a high content of free fatty acids. Fuel 85, 2671-2675.

(62) Wang, Y., S. Ou, P. Liu and Z. Zhang. (2007). Preparation of biodiesel from waste cooking oil via two-step catalysed Process. Energy Conversion and Management. 48: 184-188.

(63) Watanabe Y, Shimada Y, Sugihara A, Tominaga Y. Enzymatic conversion of waste edible oil to biodiesel fuel in a fixedbed bioreactor. J. Am. Oil Soc. Chem. 2002;78(2):703707.

(64) Xie, W. and Li, Haung (2006). Alumina-supported potassium iodide as a heterogeneous catalyst fo biodiesel production from soybean oil. Journal of molecular catalysis A: Chemical 255(1-2): 1-9.

(65) Xu, H., Miao, X. and Wu, Q. (2006). High quality biodiesel production from of microalga chlorella protothecoides by heterotrophic growth in fermenters. Journal of Biotechnology 126:499-507.

(66) Zhang, Y., Dube, M.A., McLean, D.D. and Kates, M. (2003). Biodiesel production from waste cooking oil: 1. Process design and technological assessment. Bioresource Technology, Vol. 89, No 1, pp 1-16, ISSN 0960-8524.


THE EFFECT OF SLAG ADDITION ON THE MOULDING PROPERTIES OF GREEN SAND

Sanya, Olajide T.a,*, Agunsoye, Johnson O.b, Akinruli, Ifedayo J.a, Oke, Olugbenga O.c

a Department of Glass and Ceramic Technology, Federal Polytechnic Ado Ekiti, Ekiti State

b Department of Metallurgical and Materials Engineering, University of Lagos, Akoka, Lagos

c Department of Mineral Resources Engineering, Federal Polytechnic Ado Ekiti, Ekiti State

ABSTRACTSlags generated by small, medium and large scale foundries across the country are in huge volumes, and constitute severe environmental menace. Each foundry expends part of its revenue for the disposal of these slags, while it still needs to procure green sand for moulding at a cost. This paper explores the possibilities of using particulate admixture slag for foundry green sand production. Sample basic slag was pulverized in a ball mill and classified into a range of particle sizes between 80-300 British Standard (BS). Each grade of particle size was used to prepare blends containing varying percentages by weight of slag ranging from 1–10%. The mechanical properties such as permeability, green strength, moisture content and shatter index were assessed, using standard laboratory practice. The results reveal that the particle size of the added slag has more influence on the moulding properties of the slag-bearing sand. However, when slag addition is above 1.5 percent, all the moulding properties of the green sand (permeability, green strength, moisture content and shatter index) fall outside the accepted ASTM (America Society for Testing and Materials) D2216 and D2434 Standards,. It is recommended that not more than 1.5 percent particulate slag by weight per batch of green sand of 150BS particle size should be added to the moulding sand

Keywords: Slag, moulding properties, permeability, moisture content, hydrophilic 1.0 INTRODUCTION1.1 General

Steel foundry slag is an undesirable by-product formed during the purification or melting of iron or steel scraps in iron melting furnaces. Slag from melting and holding furnaces arises from the oxidation of extraneous materials in the charge such as rust, dirt, coatings and other impurities. Slag can also be formed from erosion and wear of the refractory lining, oxidized ferroalloys and other sources(Andrews, Gikunoo, Ofosu-Mensah, Tofah, & Bansah, 2012)(Katz,2004). Iron and steel slag can be broadly categorized as basic and acidic slags. Excessive slag formation as a result of dirt and impurities contributes greatly to the erosion of furnace lining especially in electric induction furnaces. The severity of the erosion often leads to furnace leak and outright damage of the burst-bar. Also, high nickel content slag corrodes iron and steel in the presence of moisture. Slags containing sulphurous leachate discolour water resulting

in poor drainage condition (Andrews,Gikunoo, Ofosu-Mensah, Tofah, & Bansah,2012). Besides, most foundry operators in Nigeria often incur huge losses on transportation for slag disposal. Apart from the millions of naira spent annually on transportation, the huge piles of slags still lying in some companies constitue environmental pollution (Sanya , Akinruli, &Oke, 2014). However, in some developed nations, furnace slags are widely utilised as raw materials for cement production, concrete manufacturing, soil conditioners, fertilizers and in road construction (Mohammed & Arun, 2012) (Sofilic, Sofilic, & Brnardic,2012).The consumption of slag by cement and construction industries helps in reducing the problem of environmental pollution (Mohammed & Arun , 2012). The attempt to cut down expenses on slag disposal by resource recovery and recycling has not received adequate attention by researchers, particularly in Nigeria. In this present study, an attempt has been made to investigate the


69

possibility of utilising foundry slag as an admixture to produce green sand moulds.

2.0 EXPERIMENTAL PROCEDURE2.1 MaterialsA deliberate effort was made to source for basic foundry slag from some selected foundries within Lagos metropolis. Lump slags of different sizes were collected. 20kg pure silica sand was sourced from a sand mine and used for the preparation of the green sand mould. Other additives for the green mould include bentonite, fine coal dust, boiled cassava starch and water. Samples were taken from batch discharged conveyor belt of 1 ton capacity sand mixer in Nigeria Foundries Limited. 500g samples were put into a measuring cup and covered with damp cotton to avoid loss of moisture through evaporation. American Society for Testing Materials (ASTM) standard test sieves with vibrator were used to grade the starting materials; a calibrated digital weighing balance measuring up to 0.05g accuracy was used for weighing the batch constituents. The Hydraulic Powered Ramming Machine, Universal Strength Testing Machine, Shatter Index Equipment and Permeability Machine used were of the standard prescribed by the Foundry Association of Nigeria.

2.2 Experiment and MethodsPrior to permeability, green strength and shatter index laboratory test, 150g of each sand sample was weighed out on the electronic weighing scale and transferred to the specimen sleeve, with the base already plugged in. Standard ramming method was employed to prepare 100mm height by 50mm diameter cylindrical specimens. Permeability which is determined by measuring the quantity of air that passes through a given sample of sand in a prescribed time and under standard condition was tested by passing air through each rammed sample. Each rammed sand specimen was placed in the mercury cup of the permeability meter. The permeability test procedure adopted was in accordance with American Foundry Society (AFS) Universal Sand Strength machine. This machine consists of a pusher arm and weight arm, both hanging from a pivot bearing at the top of the machine. As the weight arm was pushed up higher, the load increased until the specimen was crushed. Then the compression strength in ton/in2 was read at the magnetic marker end on the graduated scale. A fresh rammed specimen was prepared to determine the shatter index value. The specimen was allowed to fall freely from a height. The shatter index equipment and the lever were pulled such that the sand was removed from the sleeve on to the receiving sieve placed directly below the equipment. Then the lumps of the sand on the sieve were collected and weighed on the weigh balance to determine the weigh X of the sand. The value of the shatter index was computed using Eq. 1.

Shatter index= X150

×100(1)To determine the moisture content of the blend, 50g of sand

specimen was accurately weighed, W1, on the electronic weighing scale and poured in a pan according to ASTM D2216 specification. The timer for the blower of the moisture teller was set for the required time to dry the sand (approximately 5 min) and air at 110℃ was blown over and through the sand. Thereafter, the dried sand was taken out of the oven and re-weighed, W2. The percentage of moisture content in the green sand was determined using Eq. 2.

Moisture content=W 2−W 1

W 1× 100(2)

Standard. For green compression strength test, the test was performed on the slag incorporated sand specimens by using.

3.0 RESULTS AND DISCUSSIONThe properties of the control samples are shown in Table 1. Results from the other experimental samples are presented in the sections below.


3.1 PermeabilityThe permeability results are shown in Figure 1. It is evident from the result that the permeability of the slag bearing moulding sand showed no significant difference with coarser slag addition for mesh sizes of 80BS and 100BS. Conversely, for pulverised slag particulates of 150BS and 300BS, the permeability decreased with increase in percentage of slag added to the blend. This shows that the addition of powdered slag particles reduced the permeability of the blend fairly significantly. The addition of coarser slag aggregates will permit gases to pass out of the mould during pouring. However, inclusion of high quality pulverised slag particulates will hold gases in the mould during pouring resulting casting defects.

3.2 Green StrengthThe green strength of the moulding sand decreased as the percentage of slag addition increased (see Figure 2). The experimental result also shows that green strength for 80 BS mesh size is 61.00KN/m2 and 300BS mesh is 37.00KN/m2 for 10% slag addition.

This implies that the green strength of the slagincorporated moulding sand varies with grain size and percentage weight of the slag. In other words, the finer the slag, the lower the green strength. Thus, addition of finer slag to the green

1 2 3 4 5 6 7 8 9 10 110

20406080

100120140160

80BS100BS150BS300BS

Percent weight of slag addition in the green sand

Perm

erab

ility

(mm

H20)

Table 1 Control test results of Permeability, Green Strength, Moisture Content and Shatter Index

Figure 1 Plot of permeability of moulding sand against percentage weight of slag addition

0 2 4 6 8 10 120

10203040506070

80BS100BS150BS300BS


Gre

en S

tren

gth,

KN

/m2

Figure 2 Plot of Green Strength of moulding sand against percentage weight of slag addition

Effect of Slag Addition on Properties of Green Sand 71

Trial Permeability (mmH2O)

Green Strength (KN/m2)

Moisture Content(%)

Shatter Index(%)

1 72.00 69.00 5.20 65.102 88.00 78.00 5.00 65.503 70.00 73.00 4.40 67.504 80.0 63.00 4.00 63.10

Average 77.50 70.75 4.65 65.30

sand blend would lower the moulding properties in terms of green strength significantly compared to the coarser slag aggregates. Like the permeability results, addition of very fine slag (below 150BS) would make the moulding sand unsuitable for casting purpose. However, backing sand would retain its strength when slags of 80-100BS mesh sizes are added in relatively low percentage (within 1-2.5% range).

0 2 4 6 8 10 120

0.51

1.52

2.53

3.54

4.55

80BS100BS150BS300BS

Percent weight of slag addition to the green sand

Moi

stur

e Co

nten

t %

Figure 3 Plot of Moisture Content of moulding sand against percentage weight of slag addition

0 2 4 6 8 10 120

10

20

30

40

50

60

70

80

80BS100BS150BS330BS


Shatt

er In

dex

, %

Figure 4 Plot of Shatter Index of moulding sand against percentage weight of slag addition

3.3 Moisture ContentThe moisture content data that were collected are shown in Figure 3. From the graph, there is a sharp decline in moisture content on slag addition. The reduction in moisture content readings when pulverised slag particles were added is because of the hydrophilic nature (showing strong affinity to absorb water) of slag grains. Low quantity (not more that 1% of the blend) of slag for all mesh size ranges may be added to green sand to regulate its moisture content.

3.4 Shatter IndexThe values of the shatter index of the slag bearing sand decrease with increase in percentage of slag addition as shown in Fig 4. Addition of low quantity of 80-100BS mesh size of slag would yield an appropriate blend that can effectively resist shock.

4.0 CONCLUSIONS AND RECOMMENDATION

4.1 ConclusionsThe following conclusions may be deduced from this study:

1. The grain size of the added slag has more influence on the moulding properties of the slag-bearing sand.

2. Except for the permeability result, the moulding properties examined (green strength, moisture content and shatter index) decline with increase in slag addition.

3. The permeability of the slag incorporated blend increases as percentage of slag increases for 80-100 BS mesh size; however, the blend permeability decreases as percentage of slag increases for 150-300BS mesh size.


4. The addition of granulated slag reduces the moisture content of the blend. Slag addition is hydrophilic in nature.

5. Slag addition not more than 1.5% of the blend could be used to regulate moisture content of the moulding sand.

4.2 RecommendationIt is recommended that slag to be added to moulding sand should be graded into particle size of 150BS and the moulding sand should not contain more than 1.5 percent by weight per batch.

REFERENCESAndrews, A., Gikunoo, E., Ofosu-Mensah, L., Tofah, H., & Bansah, S. (2012). Chemical and Mineralogical Characterization of Ghanaian Foundry Slags. Journal of Minerals & Materials Characterization & Engineering , 183-192.

Aweda, J., & Jimoh, Y. (2009). Assessment of Properties of Natural Moulding Sands in Ilorin and Ilesha, Nigeria. Journal of Research Information in Civil Engieering , 68-77.

Dimtrova, S. (2000). Metal sorption on blast-furnace slag. Water Reserach , 228-232.Fredericci C., Z. E. (2000). Crystallization mechanism and properties of a blast furnace slag glass. Journal of Non-Crystalline Solids , 64-75.

Effect of Slag Addition on Properties of Green Sand 73

Mohammed, N., & Arun , D. P. (2012). Utilization of Industrial Waste Slag as Aggregate in Concrete Applications by Adopting Taguchi's Approach for Optimiaztion . Open Journal of Civil Engineering , 95-105.

Katz, S. (2004). Slags' Effects on Cast Iron Production. American Foundry Society , 1-13.Nwonye, E., & Odo, J. (2010). Effect of Starch on the Moulding Characteristics of Ngwo Foundry Sand. Annual Nigerian Materials Congress NIMACON 2011, (pp. 107-114). Akure, Nigeria.

Sanya , O., Akinruli, I., & Oke, O. (2014). The Effects of Slag Addition on the Moulding Properties of Green Sand. 13th Annual Nigeria Materials Congress NIMACON 2014. Yaba, Lagos.

Sofilic, T., Sofilic, U., & Brnardic, I. (2012). The Significance of Iron and Steel Slag as By-Product for Utilization in Road Construction. 12th International Foundrymen Conference , (pp. 419-436). Croatia.

Documents

Web viewmicrostructure . and wear properties. The result revealed that the oil soak and water soak increase as the particle size of the agro-waste materials increases; but the hardness,