Processing and mechanical properties of pure mg and in situ aln reinforced mg 5al composite 4

4.4 b and c can be observed in the 10h- and 20h-MMed composite samples which are in larger grain size regime.. As shown in Table 4.1, after 10h of milling, grain size of the MMed powders

in nanometer regime [1-7] Severe plastic deformation and MA/MM processes can be employed to refine the matrix grains [8,9]

MM is one of the most effective processes for dispersing ex-situ nanoparticles more uniformly in metal matrix [5,6,10-14] and inducing in-situ nanoparticles in the composites during milling A better bonding between metal matrix and in-situ formed nanoparticles which are clean, ultrafine and thermally stable renders the excellent mechanical properties

Inherent deficiencies such as low stiffness, high wear rate, and high chemical reactivity, loss of mechanical strength at high temperature and creep resistance restrict the

industrial applications of Mg and its alloys [15] By adding micro and nanosized ceramic particles in Mg matrix, these drawbacks can be overcome [3, 6] In this study,

Mg nanocomposites with 1 wt% of in-situ AlN formed reinforcement were synthesized for milling durations up to 40h and their physical and mechanical properties were accessed The contribution of texture developed during extrusion to tensile deformation was examined by means of pole figure measurements For comparative study, pure Mg samples were also synthesized and tested using identical parameters used for composite samples

4.2 Experimental

Mg chips/turnings (Drehspaene) (Acros Organics) and Al powder (Alfa Aesar, -325 mesh) of 99.5% and 99% purity respectively were used as starting materials AlN composite powder was synthesized in-situ by MM of Al powder and pyrazine for 100h

as described in Chapter 3 The nominal composition of the composite is 1wt%AlN (Mg-5Al-1AlN) 35g of composite mixture together with 0.5 to 3 wt% of stearic acid, CH3(CH2)16COOH, and hardened carbon steel balls were loaded into 500

Mg-5wt%Al-ml stainless steel vial in a 99.9% pure argon atmosphere in an AMBRUAN glove box The weight ratio of Mg chip to ball is 1:20 A Retsch PM400 Planetary Ball Mill was employed for MM at 300 rpm Each batch of powder was mechanically milled for different durations of 0 (as-blended), 10, 20, 30 and 40 hours at room temperature 0h-MMed sample was obtained by blending the composite mixture at low rotational speed

of 100 rpm for 1h Same milling conditions were applied to MM of pure Mg chips Mg-5Al-1AlN and pure Mg samples are designated hereafter as xxh-MMed composite sample and xxh-MMed Mg sample respectively, where xx is milling hours while the as-blended powder mixture or as-received Mg chips are indicated as 0h

After milling, a small quantity of powder was withdrawn for the examination of structural changes by means of an X-ray diffractometer (XRD) The milled powders were cold-compacted using 35mm diameter metal die at sixty tons of uniaxial compaction pressure The green compacts were sintered in a tubular furnace under argon gas flow for 2 hours at 500ºC The sintered billets were then hot-extruded at an extrusion ratio of 25:1 to cylindrical rods of 7mm diameter

The grain size of the as-received Mg chips and as-blended extruded specimens was measured using optical microscope and the microstructure of as-milled specimens were characterized using Jeol 2010F TEM The extruded rods were machined into cylindrical tensile specimens with a gauge diameter of 5mm and a gauge length of 25

mm according to ASTM E8M-96 standard Uniaxial tensile test was conducted at room temperature using an automated Instron 8501 servo hydraulic testing machine at controlled strain rates of 3.33x10-4 s-1 The deformation was monitored using a 25-mm clip-on extensometer

Resistivity measurement was carried out by Jandel Multi Height Four-Point Probe Stand with Keithley K6200 DC current source and Keithley K2182 nanovoltmeter The bulk resistivity  was obtained from the equation:

where s is the spacing of the probe in cm, I the test current in ampere and V the

measured voltage in volt

Setaram TMA 92-16.18 was employed to investigate the nature of thermal expansion

of the samples by means of coefficient of thermal expansion (CTE) Thermal behavior

of the bulk sample was further investigated by heating the samples in differential scanning calorimeter DSC-2910 to 700°C at 10°C/min To calculate the specific heat capacity, thermal analysis was carried out using DSC from 323K to 453K at a constant

heating rate of 20 K/min in argon atmosphere Specific heat capacity C p,sample was obtained from equation 4.2 [16]

sapphire p reference sapphire

reference sample sample

sapphire sample

y

y m

displacement of the sample Δy sample-reference and the sapphire standard Δy sapphire-reference

are the difference between the distances from the reference baseline in the thermal plot The standard specific heat capacity of sapphire was obtained from the thermodynamic data [17] and is expressed as

)4

635.474(

)10742.0(6)10757.1(2)10087.12(2547.104

5 0

2 6 2

6 3

x T

x

C p sapphire

(4.3)

4.3 Results and discussion

4.3.1 Mass structure investigation by XRD

X-ray diffractometer was employed to perform structural investigation on MMed powders and extruded specimens in the transverse direction XRD spectra of the MMed composite powders and the extruded composite samples are shown in Figs 4.1(a) and (b) respectively In Fig 4.1(a), all Al peaks from the as-blended specimen

disappeared in the MMed powder due to solid solution of Al with Mg resulting in the formation of Al12Mg17 A new phase of MgAl2O4 was detected in the MMed powder after 10 and 20h During milling, MgO and Al2O3 oxide layers on as-received Mg chips and Al power surfaces fractured into very fine particles to favor solid state reaction for the formation of MgAl2O4 according to the following reaction [18]

Very weak AlN peaks were observed in all MMed specimens suggesting the complete immiscibility of AlN in Mg With increasing milling duration, broadening of XRD peaks and declining in peak intensity are observed due to the reduction in grain size and introduction of microstrain during milling

In Fig 4.1(b), intensities of Mg (100) and (110) peaks for the extruded specimens increase with milling duration, confirming the formation of deformation texture Texture occurs due to the deformation-induced crystallographic plane rotating preferentially along the extrusion direction as a result of extrusion at high extrusion ratio Contamination from process control agent, stearic acid and powder handling atmosphere can also be observed from the weak MgO (200) peak and MgH2 (211) peak, especially at milling durations of 20h and longer Al12Mg17 peaks disappeared in all the as-milled specimens It might be due to very fine in size and very minimal in quantity to produce a visible diffraction peak It is also highly possible that Al reacted with the excess nitrogen molecules from AlN composite powder to form AlN This will be confirmed by thermal analysis of the composite samples by DSC in section 4.3.3 Weak MgAl2O4 peaks emerged in all as-milled samples as a result of solid state reaction during sintering between MgO and Al2O3, both of which are inherited from the surfaces of the as-received Mg chips and Al powders

From X-ray diffraction patterns of pure Mg in Figs 4.2 (a) and (b), no apparent contamination from stearic acid or milling and handling atmosphere was observed in as-milled powders and as-extruded samples It might be due to negligible amount of contamination or the particles of contamination by-products were too fine to produce prominent diffraction peaks The intensity of Mg (100) peak was exceptionally high in the as-extruded specimens and that of basal plane (002) peak becomes lower with milling duration It implies that extrusion induced deformation textures exist in the extruded samples

Fig 4.3(a) shows the microstructure of the as-received Mg chips In Fig 4.3(b), Al12Mg17 decorated along the Mg matrix grain boundaries and AlN along the Mg chip boundaries in the 0h-MMed extruded composite sample Deformation due to cold compaction and hot extrusion was not high enough to inject AlN particles into the grain boundaries From Fig 4.3(c), no apparent grain elongation could be observed along the extrusion direction Fig 4.3(d) shows the clean grain boundaries of pure Mg samples

(a) (b)

Figure 4.3 Optical micrograph of (a) as-received Mg chip, (b) 0h-MMed composite sample in cross-sectional area, (c) 0h-MMed composite sample in longitudinal direction and (d) 0h-MMed Mg sample in longitudinal direction

At higher magnification, entangled dislocation pile-ups within the grain were observed from TEM image as shown in Fig 4.4 (a) During milling, mechanically cold-worked powders resulted in generation of dislocations, multiplication and congealing that produced nanosized grains [19] The grains were highly strained and contained numerous defects When the grain is extremely small, the formation of new nanocrystals via dislocation movement stops because of inability of individual grain to support more than one dislocation [20] As such, some limited dislocations (Figs 4.4 (b) and (c)) can be observed in the 10h- and 20h-MMed composite samples which are

in larger grain size regime However, for 30h- and 40h-MMed composite samples, no

d100=2.75Å

indication of the presence of dislocation can be observed in Figs 4.4 (d) and (e) Fig.4.4 (f) shows a HRTEM image of lattice pattern from 30h-MMed composite sample It reveals that the particle is AlN polycrystal with an inter-planer spacing of 2.75 Å corresponding to the (100) plane of AlN crystal 40h-MMed composite showed the smallest grain size of 33 nm compared to the coarse and ultrafine grain-sized samples In Fig 4.5, HRTEM investigation reveals the estimated thickness of grain boundary was about 1 nm and it appears to be free of contamination or particles with disordered phases

Figure 4.5 HRTEM observation of grain boundary marked with dotted line

As shown in Table 4.1, after 10h of milling, grain size of the MMed powders was significantly reduced from 24 m to 44 and 41 nm in the composite sample and pure

Mg sample respectively However, longer milling did not produce further grain refinement The crystalline sizes are respectively 32, 26 and 22 nm after 20, 30 and 40h-MM in composite sample In pure Mg samples, average crystalline sizes after 20,

30 and 40h-MM are 31, 28 and 25 nm respectively

Grain boundary

Table 4.1 Grain sizes (in nm) of powders (P) and extruded samples (E) after different milling durations





cos05.1

d

where B s is the broadening due to reduction in crystallite size,  the wavelength of

X-ray, d the crystallite size and  the diffraction angle

Grain sizes of the extruded composite samples were estimated from 50 grains from TEM images taken at different locations and different samples The Scion Image software was employed to calculate the grain size Grain size distribution of the as-milled composite samples is illustrated by the histograms in Fig 4.6 Average grain sizes were estimated to be 116, 86, 42 and 33 nm for the 10, 20, 30 and 40h-MMed composite samples respectively Narrower grain size distribution between 20 to 70 nm was observed after 40h-MM

10h Average = 116 nm

0.1 0.2

20h Average = 86 nm

0.1 0.2

40h Average = 33 nm

(d) Figure 4.6 Grain size distributions of MMed composite samples after (a) 10h, (b) 20h, (c) 30h and (d) 40h milling

Crystallite sizes of the as-milled extruded composite and pure Mg samples were in the range of 116-33 nm and 183-127 nm respectively It is clear that retardation of grain growth was effective in composite specimens by second phase particles such as AlN and in-situ formed Al12Mg17, MgH2, MgAl2O4, and MgO during sintering and extrusion Such stable grain structure with no grain growth up to 550°C has been reported in commercially pure Mg with mean particle diameter of 20m reinforced with 1 vol % of nanoscaled alumina particles (the mean diameter 12 nm) [21, 22]

4.3.2 Electrical resistivity

The electrical resistivity () of nc materials with increased volume fraction of atoms at the grain boundary is expected to be higher than both coarse grained polycrystalline

metals and amorphous alloys At a constant temperature,  increases with decreasing grain size The magnitude of electrical resistivity in composite can be tailored by changing the volume fraction of electrically conducting component The total resistivity of a crystalline metallic specimen is the sum of the resistivity due to thermal agitation of the metal ions of the lattice and the resistivity due to the presence of imperfections in the crystal The resistivity of a metal results from the scattering of conduction electrons Lattice vibrations scatter electrons because the vibrations distort the crystal Imperfections such as impurity atoms, interstitials, dislocations and grain boundaries scatter conduction electrons because the electrostatic potential in their immediate vicinity differs from that of the perfect crystal

Table 4.2 Electrical resistivity of composite (c) and pure Mg samples (Mg) milled for different milling durations

up to 20h-MM followed by a decrease after 30 and 40h of MM This might indicate

the lesser processing flaws such as porosity, internal cracks, etc and the diminishing dislocation activities in 30 and 40h-MMed samples with higher degree of grain refinement as evident in Fig 4.4 (d) and (e) It is noteworthy that the difference in electrical resistivity between composite and pure Mg becomes lesser with increasing milling time exhibiting 182, 257, 191, 88 and 25% higher in composite samples after 0,

From Table 4.3 and Fig 4.8, the same trend of α c and α Mg is observed for both

composite and pure Mg samples Milling increases the α value with grain refinement

accompanying the increasing grain boundary volume after 10h-MM As the CTE of grain boundary is larger than that of crystalline state (2.5 – 5 times larger) [29], more grain (or phase) boundaries in the samples with higher milling durations would increase the thermal expansion The greatly enhanced thermal expansion coefficient,

on the other hand, reveals the ultrafine structures in the sample

Table 4.3 Coefficient of thermal expansion (CTE) of composite (α c) and pure Mg

samples (α Mg) milled for different milling durations

Figure 4.8 Coefficient of thermal expansion (CTE) of composite (α c) and pure Mg

samples (α Mg) milled for different milling durations

It is interesting to note that except for the 20h-MMed sample, α increased up to 30h

MM followed by a drop in 40h-MMed samples in both material systems The homogeneous distribution of reinforcement particles and in-situ formed second phase particles hinder the lattice expansion and grain growth, and consequently reduce the thermal expansion of the composite samples compared to those of pure Mg samples

Thermal behavior of the bulk sample was further investigated by heating the samples

in DSC to 700°C at 10°C/min DSC traces of the composite and pure Mg samples MMed for different milling durations are as shown in Fig 4.9 A sharp endothermic peak appears corresponding to melting which is taken as the onset temperature of the peak It is clear that the longer the milling hours, the higher is the melting temperature

(T m) of the composite samples The positions of the endothermic peaks are respectively

shifted from 586°C for as-blended sample to 622, 625, 645 and 647°C for the 10, 20,

30 and 40h-MMed samples as shown in Fig 4.9 (a) This could be due to the variation

of solid solution of Al in Mg matrix with milling duration From Mg-Al phase diagram,

the amount of solid solution of Al in Mg during milling is calculated using the melting

temperature at various milling durations It is estimated that 2, 1.8, 0.2 and 0 wt% of

Al formed solid solution with Mg after 10, 20, 30 and 40h of milling respectively It

can be observed that Al formed lesser solid solution with milling time leading to the

melting temperature to shift higher and ultimately close to the melting temperature of

pure Mg for 40h-MMed composite sample

645°C 645°C 646°C 646°C

(a) (b) Figure 4.9 DSC traces of (a) composite samples and (b) pure Mg samples at different

milling durations

Although weak Mg17Al12 peaks are detected in XRD patterns of 0h-MMEd sample, it

is noted that the melting of Mg17Al12 is not detected from the thermal traces It might

be due to too little Mg17Al12 present to produce an endothermic peak in the DSC traces

It is also highly possible that the reaction between Al and excess nitrogen atoms from

AlN composite powder forms AlN instead of producing Mg17Al12 The heat released

thus lowering the overall melting temperature by few degrees The melting

temperatures of MMed samples increase near to T m of Mg matrix (650°C) after 40h of

milling indicating no apparent influence of second phases and reinforcement on

melting with longer milling duration It is interesting to note that milling has no

influence on the melting of pure Mg samples exhibiting similar melting temperature of

645-646°C in all samples as shown in Fig 4.9 (b)

0h 40h 30h 10h 20h

(a) (b)

Figure 4.10 C p values of (a) composite and (b) pure Mg samples for different milling

durations

It can be seen from Fig 4.10 that the increase in C p of nanocrystalline samples with

temperature is approximately linear as reported by Lu et al [16] In absolute value of

C p, composite samples show highest value in the 40h-MMed sample followed by 0, 30,

10 and 20h-MMed samples as shown in Fig 4.10(b) It can be observed from Fig

4.10(b) that pure Mg samples show insignificant change in C p values although the

20h-MMed sample exhibits slightly higher C p Generally, C p values of composite samples

are higher than those of pure Mg samples

Specific heat of a material is closely related to its vibrational and configurational

entropy, which is affected significantly by the nearest neighbor configurations, e.g

interatomic potentials The enhancement of C p in nanocrystalline materials might be

attributed to the complicated structures of grain and/or phase boundaries The nature of the property difference needs further theoretical and experimental studies Based on

statistical mechanics and quantum theory [30], C p can be expressed as:

3 4

kT k N C



where N A is Avogadro number, k Boltzmann’s constant, T the absolute temperature, h

Planck’s constant and D Debye frequency (maximum allowable phonon frequency) The specific heat capacity is mainly dependent on the maximum phonon frequency D

as shown in equation 4.6 After initial milling of 10 and 20h, grain refinement with larger volume fraction of disordered grain boundaries resulted in high vibrational densities and higher Debye phonon frequency D, which consequently leads to the

decreasing trend of C p for these milling durations However, a reversed trend is observed for longer milling durations It can be associated with an increase in the configurational and vibrational entropy with longer milling time due to higher interfacial disorder and lattice defects such as dislocations, grain boundaries, vacancies and impurities Lu et al [16] verified experimentally that grain boundaries and lattice defects are mainly responsible for the increase in specific heat

The activation enthalpy ΔH and activation entropy ΔS can be estimated from the C p

data using the following relationship,

H (

Mg Mg-5Al-1AlN

0.13 0.12 0.11

Mg Mg-5Al-1AlN

(a) (b) Figure 4.11 (a) ΔH and (b) ΔS of composite samples and pure Mg samples estimated

from the data of specific heat capacity

It can be seen from Fig 4.11 that both ΔH and ΔS of composite samples are higher

than those of pure Mg samples The AlN reinforcement particles and second phase

particles might hinder the lattice vibration and this causes the higher ΔH and ΔS for the

composite samples For composite samples both ΔH and ΔS values decrease with

milling duration up to 20h and significantly increase after 30 and 40h-MM Different

trend is observed for the pure Mg samples As milling increases, both ΔH and ΔS

increases until 10h-MM However, further milling decreases those values until

30h-MM and slight increases in 40h-30h-MMed samples It is noted that except for the

10h-MMed samples, other 10h-MMed samples show lower values of ΔH and ΔS compared to

those of as-blended specimens

Decrease in ΔH with prolonged milling duration implies that the value of activation

energy barrier Q in the grain boundary diffusivity in 30 and 40h-MMed samples is

lower compared to rest of the samples Due to low activation energy barrier, there will

be higher possibility for an applied stress to overcome the barrier to trigger the grain

boundary diffusion at room temperature