H O S T E D BY Available online at www.sciencedirect.

com

ScienceDirect
Journal of Magnesium and Alloys 3 (2015) 16e22
www.elsevier.com/journals/journal-of-magnesium-and-alloys/2213-9567

Full length article

Investigation on mechanical properties and creep behavior of stir cast

AZ91-SiCp composites
Abhilash Viswanath a, H. Dieringa b, K.K. Ajith Kumar a, U.T.S. Pillai a,*, B.C. Pai a
a
CSIR e National Institute for Interdisciplinary Science and Technology, Material Science and Technology Division, Thiruvananthapuram 695019, India
b
MagIC e Magnesium Innovation Centre, Helmholtz-Zentrum Geesthacht, Max-Planck-Strasse 1, D-21502 Geesthacht, Germany
Received 1 September 2014; revised 1 January 2015; accepted 8 January 2015
Available online 6 March 2015

Abstract

The room temperature mechanical properties and high temperature creep behavior of AZ91 alloy reinforced with SiCp synthesized via stir
casting have been evaluated. The mechanical properties showed improvement with respect to the amount of reinforcement content. The creep
testing of the composites carried out at a temperature of 175
C under constant stress of 80, 100 and 120 MPa reveals different creep char-
acteristics depending upon the reinforcement content and the applied load. The true stress exponents of different composites calculated from
minimum creep rate indicate the possible mechanisms of creep deformation.
Copyright 2015, National Engineering Research Center for Magnesium Alloys of China, Chongqing University. Production and hosting by
Elsevier B.V. All rights reserved.

Keywords: Composites; Optical microscopy; Mechanical properties; Creep characteristics

1. Introduction prominent ones. Infiltration processes (pressure less and

pressure assisted) are capable of synthesizing composites with
Being the lightest among the structural materials, magne- higher volume fraction of reinforcements; still, their applica-
sium based alloys have become a competent part in areas tions are limited by the preform size [3]. Fine grained com-
where weight reduction is among primary concerns [1]. Even posites with near-uniform distribution of reinforcements along
though, structural components made of Mg alloys are plenty, with highest possible reinforcement content can be produced
load bearing Mg components, especially operating at higher via powder metallurgy route [4]. However, the restricted size
temperature (>150
C) are rare. Concerns over elevated tem- of the composites and higher cost involved are the inherent
perature applications can be circumvented by developing Mg disadvantages of this process. Even though, interfacial reac-
metal matrix composites (Mg MMC's) with reinforcements tion is a major concern in stir casting, the process is consid-
which are thermally stable. Also, these reinforcements ered as the most adaptable and economical route, due to its
strengthen the matrix by imparting better mechanical and simplicity, flexibility, low processing cost and high production
tribological properties [2]. rate for synthesizing discontinuously reinforced composites
Among the various viable routes for the synthesis of Mg [5].
MMC's, pressure assisted infiltration (squeeze casting), pres- As far as the thermal stability of the reinforcements are
sure less infiltration, powder metallurgy and stir casting are the concerned, ceramic particles are the most desirable due to high
level of hardness, strength and melting point. Among ceramic
reinforcements used in Mg MMC, silicon carbide particles
* Corresponding author. Tel.: þ91 471 2515236; fax: þ91 471 2491712. (SiCp) are the most popular due to high stability and wetta-
E-mail address: utspillai@rediff.com (U.T.S. Pillai). bility in Mg matrix [2]. Furthermore, SiCp reinforced Mg
Peer review under responsibility of National Engineering Research Center
for Magnesium Alloys of China, Chongqing University.
MMC's are reported with improved strength [6] and wear

http://dx.doi.org/10.1016/j.jma.2015.01.001.
2213-9567/Copyright 2015, National Engineering Research Center for Magnesium Alloys of China, Chongqing University. Production and hosting by Elsevier B.V. All rights reserved.
 A. Viswanath et al. / Journal of Magnesium and Alloys 3 (2015) 16e22 17

resistance [7] than base alloys. However, investigations into specimens were tested at 175
C under constant stresses of 80,
high temperature creep properties of SiCp reinforced com- 100 and 120 MPa using an ATS lever arm system.
posites with Mg matrix synthesized via stir casting are rare.
The present investigation aims to reveal the room temper- 3. Results and discussion
ature mechanical properties and high temperature creep
properties of AZ91/SiCp composite synthesized via stir cast- 3.1. Microstructural observations and interfacial
ing. The creep testing of the composites has been performed characteristics
by applying compressive loads as the stress applied in
compressive manner represents real situations as in case of The XRD analysis done for phase identification revealed
bolting of gear housing or parts of engine [8]. Stress exponents the presence of the Mg matrix, the inherent intermetallic phase
calculated from minimum creep rate reveal the mechanism of Mg17Al12 and the reinforcement SiC. The traces of oxide layer
deformation involved. MgO along with Mg2Si intermetallic phase are also detected.
The details of the analysis have been published elsewhere [10].
2. Experimental Fig. 1 depicts the optical micrographs of AZ91 base alloy
and AZ91-SiCp composites with 5, 10, 15, 20 and 25 wt%
2.1. Composite preparation SiCp content. A typical two phase microstructure of AZ91
alloy with a-Mg phase and b-Mg17Al12 phase along and
About 1200 g of AZ91 (Mg-9.3Al-0.8Zn-0.18Mn) was adjacent to the grain boundaries is shown in Fig. 1(a). Near-
used per casting as the matrix for preparing the composites. uniform distribution of SiCp can be noted in microstructure
The alloy was melted in a steel crucible using a resistance of the composites depicted in Fig. 1 (bef). Moreover, the
heating furnace under protective atmosphere of argon gas. At reinforcements are found to be dispersed in the matrix with
the melt temperature of 750
C, stirring was carried out using very few agglomerations. Generally, the increase in particle
a steel impeller rotating at 750 rpm. The SiCp having size reinforcement content tends to increase the agglomeration in
~23 mm was preheated to 600
C and was added into the composites. This is attributed to the finer size of the re-
corner of the vortex during stirring. The amount of SiCp cor- inforcements and variations in parameters like stirring time,
responding to 5, 10, 15, 20 and 25% by weight of the matrix stirring speed, stirring temperature, etc [11]. Moreover,
took about 10, 20, 30, 40 and 50 min respectively for the noticeable presence of MgO in the form of white patches or
addition. The stirring was continued for 10 min after complete that of Mg2Si in needle like or dots cannot be evidenced in the
addition of SiCp to ensure the mixing of reinforcement into the microstructure as of XRD results. This nullifies the effect of
matrix. Subsequently, the melt was poured into a 300
C such phases on formation of microstructure.
preheated mild steel mould to obtain the castings. Moreover, the grain size is reduced from 65 microns for the
base alloy to 34 microns for 5 wt% SiCp containing com-
2.2. Microstructure posites and then to 22 microns for composite containing 10 wt
% SiCp. However, the grain size reduction is comparitively
The specimens for microscopy, obtained from the same negligible for higher reinforcement containing composites i.e.
area for all the castings to have comparable cases, were sub- 21 microns for 25 wt% SiCp.
jected to standard polishing using SiC abrasive papers and The possible refinement mechanism behind reduction in
selvyte cloth with diamond paste. The polished specimens grain size of the matrix due to the presence of SiCp is het-
were etched with acetic picral etchant (5 ml acetic erogeneous nucleation. The reinforcing SiCp which can act as
acid þ 6 g picric acid þ 10 ml H2O þ 100 ml ethanol). Leica the heterogeneous nucleation site may get captured by
DMRX optical microscope was used for the microstructural growing Mg crystals and finally stay within the Mg grains in
analysis. Microstructures with higher magnification were the composites. Hence, the presence of SiCp within the grain
incurred using a scanning electron microscope (JEOL, JSM boundaries of the composite is justified. Moreover, certain
35C) with an accelerating voltage of 15e30 keV. crystallographic orientation relationships between the particle
and matrix and the restricted growth of Mg grains by SiCp
2.3. Mechanical properties during solidification can further contribute to the reduction in
grain size [11,12] which is evidenced by the presence of the
Samples prepared according to ASTM E8 standard [9] were reinforcements around the grains as clusters.
tested for room temperature tensile properties using a computer Interfacial reactions are of extreme importance while
controlled INSTRON 8801 universal testing machine set with a considering the mechanical and creep properties of the com-
cross head speed of 2 mm/min. The fractured surfaces of the posite as the load transfer between the reinforcements and the
tensile tested specimens were also examined under SEM. matrix is governed by the interface between them. The SEM
microstructure of AZ91-20SiCp composite shown in Fig. 2a
2.4. Creep characteristics reveals good distribution of the particles. The high magnifi-
cation SEM micrograph in Fig. 2b shows the intimate bonding
Cylindrical specimens with length 15 mm and 6 mm of SiCp with Mg matrix which nullifies the occurance of any
diameter were used for compression creep tests. The significant chemical reaction at the Mg-SiCp interfaces.
18 A. Viswanath et al. / Journal of Magnesium and Alloys 3 (2015) 16e22

Fig. 1. Optical micrographs of AZ91 alloy with a) 0, b) 5, c) 10, d) 15 and e) 20 and f) 25 wt-% SiCp content.

Moreover, the sharp and clean (precipitate and reaction stating that SiC wetted well with Mg and Mg is a better host
free) interface also depicts good interfacial integrity of SiCp for SiCp embedment than Al. Moreover, the presence of Al in
with the matrix. Studies conducted on the feasibility of the AZ91 alloy has tremendous influence in improving the
compo-cast AZ91/SiC composites had similar conclusion wettability of SiCp reinforcements [13].

Fig. 2. SEM micrographs of a) AZ91-20SiCp composite and b) Showing clean interface.

 A. Viswanath et al. / Journal of Magnesium and Alloys 3 (2015) 16e22 19

3.2. Mechanical properties and tensile fractography carbide is very high (3900 MPa) compared to the base alloy
value of 310 MPa. The significant increase in compressive
The variation in yield strength (YS) and ultimate tensile strength observed in the composite can also be attributed to
strength (UTS) of AZ91 alloy with respect to the various SiCp partial closure of the fine microscopic cracks during
content is depicted as histogram in Fig. 3. The YS of the compression loading. Similar observation is also reported for
composites during tension shows a gradual increase with SiCp reinforced AZ92 Mg alloy composites [15].
respect to the increment in amount of reinforcements. Studies The tensile fracture surfaces of the composites (Fig. 5 a and
pertaining to the tensile strength of Mg matrix composites b) reveal both ductile and brittle mechanisms. The pockets like
have revealed that within a certain range, both yield strength regions surrounding the fractured and deciphered SiCp are
and elastic modulus of the composites increase linearly with termed as “tear ridges”. These ridges are indicative of ductile
increase in amount of reinforcement. The key strengthening failure and so are the pockets of dimples of varying size and
mechanisms in Mg matrix composites are particle strength- shape in the fracture surface. The mechanical properties of the
ening, work hardening, load transfer and grain refinement of soft, ductile and elastically deforming AZ91 alloy matrix gets
the matrix alloy by the reinforcement [2]. The finely dispersed augmented by the hard, brittle and essentially plastically
SiCp strengthen the Mg matrix by blocking the dislocation deforming reinforcing SiCp. Moreover, the development of a
motion within the composites. Also, the smaller sized grains resultant triaxial stress state in the matrix aids in limiting the
have improved strength within the elastic region and facilitate flow stress of the composite [16]. Also, brittle fracture region
enhanced tensile properties for the composites at room tem- increases with increase in SiCp when compared to the ductile
perature [14]. In the present composite system, the grain size region.
reduction for AZ91-5SiCp and AZ91-10SiCp is more pre-
dominant compared with AZ91 alloy with 15, 20 and 25 wt% 3.3. Creep behavior
SiCp. The variation in YS also depicts similar characteristics,
with increment being high for 5 and 10 SiCp containing The increase in yield strength, ultimate compressive
composites as of composites with higher amount of reinforc- strength and the hardness is the motivation behind evaluation
ing phase. of creep properties. Mg alloys have an inherent disadvantage
Moreover it is evident from the histogram that the of relatively low creep resistance. The grain boundary slide
improvement in strength during tension is limited to the elastic and dislocation slip at both the basal and the non-basal planes
region, as the UTS of the composites do not show considerable results in high creep rate of Mg alloys [17]. Thus, incorpora-
variation from that of base alloy. The residual stress produced tion of hard particles which can impede the grain boundry
during the deformation process, failure of agglomerated sub- slide and the dislocation slip in the soft matrix can improve
micron SiCp and the presence little amount of porosity are high temperature creep properties of Mg alloys.
the possible reasons behind constant levels of UTS of the Fig. 6 shows exemplarily creep curves of tests performed
composites. with AZ91-25SiCp composite at a temperature of 175
C with
The increase in ultimate tensile strength (UCS) of the stress levels of 80, 100 and 120 MPa. The variation in creep
composites (Fig. 4) is attributed to the presence of hard rein- deformation of the composite as a function of time for
forcement particle which restrict the flow of dislocations in the different applied loads depicted in Fig. 6a symbolizes a typical
matrix. Moreover, the compressive strength of the silicon creep curve. The creep deformation increases with increase in

Fig. 3. Histogram showing variation in yield strength (YS) and ultimate tensile Fig. 4. Variation in ultimate compressive strength (UCS) of the AZ91 alloy
strength (UTS) of the composites with respect to SiCp. with increase in reinforcement content.
20 A. Viswanath et al. / Journal of Magnesium and Alloys 3 (2015) 16e22

Fig. 5. SEM image of the fractured surface of the tensile tested specimens of a) AZ91/10 SiCp and b) AZ91/20 SiCp composites.

stress and associated time decreases as the load elevates. conditions. It is clearly visible that minimum creep rate is
Hence, increase in applied stress reduces the test duration. reached in the first 10% of deformation and with increase in
The creep rate of the composite is plotted against time for applied load, the deformation at which the minimum creep
varying stress levels in Fig. 6b. The curves have similar rate occurs, increases.
characteristics with those from Fig. 6a in terms of variation of Creep deformation is the representation of the time
creep rate and time with respect to applied load. Moreover, dependant plastic flow occurring in crystalline and non-
reduction in minimum creep rate can also be evidenced with crystalline materials under an applied load which need not
decreasing the applied load. The creep curves in Fig. 6c are the to be above the elastic limit of the material. Moreover, it is
diagrammatical representation of the creep rate of the com- well documented that the classical creep curve of materials
posite against its creep deformation under different loading consists of three distinct stages namely primary stage in which

Fig. 6. Creep results of tests performed with AZ91 þ 25SiC at 175
C. a) Creep deformation over time, b) Creep rate over time and c) Creep rate over creep
deformation from tests at applied stresses of 80, 100 and 120 MPa.
 A. Viswanath et al. / Journal of Magnesium and Alloys 3 (2015) 16e22 21

strain rate decreases with increase in time, a secondary stage Table 1

where stain rate is reasonably constant and ternary stage where Stress exponents, threshold stresses and true stress exponents of AZ91-SiCp
composites calculated from creep tests at 175
C.
strain rate increases with time [18]. However, in the case of
discontinuously reinforced composites, duration of secondary AZ91 with various wt-% SiCp 0 5 10 15 20 25
stage is relatively shorter. In such cases, the deformation Stress exponent (n) 7.2 7.4 7.7 8.5 8.4 8.9
mechanisms are best described using minimum creep rate Threshold stress (sthr) 24.4 24.7 26.4 30.1 31.4 34.4
True stress exponent (nt) 5.4 5.5 5.6 5.8 5.6 5.7
which can be observed most readily in plots of creep rate
against time [19].
The dependence of minimum creep rate ε_ s on temperature
particles, the ‘n’ value naturally increases with increase in
and applied stress is given by the Norton-equation as:
reinforcement content. However, in the present study, the ‘n’
ADGb  s n value for AZ91-5SiCp and AZ91-10SiCp composites are very
ε_ s ¼ ð1Þ close to that of the base alloy. These alloys have phenomenally
kT G
reduced grain size in comparison with composites with higher
where, A is a material dependent constant, G the shear SiCp content. For the composites with 15, 20 and 25 wt%
modulus, b the Burgers vector, k the Boltzmann constant and n SiCp, the presence of elevated amount of reinforcement
the stress exponent, D is the diffusion coefficient, which is compensate for reduction in grain size, thus improving the
stress exponent.
D ¼ D0 $expðQC =RTÞ : ð2Þ Studies have shown that in comparison with metals, the
D0 is the frequency factor and QC the activation energy for creep behavior of discontinuously reinforced alloys is unusual
creep. in two ways [24,25]. Firstly, the stress dependence of creep
The stress exponent n gives information about rate- rate, as accounted by value of stress exponent ‘n’, is unnatu-
controlling deformation mechanisms during creep (n ¼ 3 rally high. Secondly, the temperature dependence of creep
represents creep controlled by viscous glide [20], n ¼ 5 is for rate, usually measured as activation energy ‘Q’, is much larger
creep controlled by high temperature climb of dislocation due than that for self-diffusion.
to diffusion of the lattice [21], n ¼ 7 symbolizes creep by low The high stress dependence of creep rate is due to the
temperature climb of dislocation due to diffusion of the core presence of threshold stress (sthr), the stress below which no
[22], n ¼ 8 is for lattice diffusion controlled creep deformation creep deformation occurs, and is attributed to the interactions
[23]). of the dislocation with the reinforcement. There are several
The n values are determined by plotting the minimum creep explanations for threshold stress. It is often referred as the
rates in a double logarithmic plot versus stress field as shown Orowan stress, which is the additional stress needed for
in Fig. 7a. The filled squares in the graph represent AZ91 and bowing the dislocations between the reinforcements [26]. The
the empty ones are those of the composites reinforced with sthr is also referred as the additional back stress which is
different amounts of SiCp. The base alloy and the composites required for climbing over an obstacle or as the stress needed
with lower reinforcement content (5 and 10 wt %) evince less for detaching dislocations from obstacles [27,28]. Hence, the
significant difference in minimum creep rates. However, the stress term (s) in eq (1) is to be replaced with the term of
composites containing 15, 20 and 25 wt% SiCp have an effective stress (seff) to avoid the anomalous nature of the
improved creep resistance. The calculated stress exponents n stress exponents. The equation gets modified as:
for the base alloy and composites are given in Table 1.
ADGb seff nt
The values of stress exponent given in Table 1 shows ε_ s ¼ ð3Þ
variation in co-relation with the grain size and volume of re- kT G
inforcements. For composites reinforced with thermally stable

Fig. 7. Minimum creep rates as function of a) applied stress and b) effective stress for tests at 175
C for AZ91 alloy with 0, 5, 10, 15, 20 and 25 wt-% SiCp content.
22 A. Viswanath et al. / Journal of Magnesium and Alloys 3 (2015) 16e22

where, seff ¼ s  sthr and nt is the true stress exponent. support received from Mr. K.K. Ravikumar for mechanical
The estimation of values of threshold stress is done here testing, Mr. M.R Chandran for SEM analysis and dedicated
using the method developed by Li and Langdon [29]. Each efforts from Mr. J.S Harikrishnan, Mr. K.A. Ajukumar and
curve through the individual datum points are extrapolated to Mrs. K.S. Deepa, is also highly acknowledged.
lie essentially vertical to a strain rate of 1010 s1, which
corresponds to a deformation rate of approximately 1% in 3 References
years. The predicted stress levels at this low strain rate are very
close to threshold stress. The double logarithmic plots of [1] C. Blawert, N. Hort, K.U. Kainer, Trans. Indian Inst. Met. 57 (2004)
minimum creep rates versus the effective stress used for 397e408.
[2] H.Z. Ye, X.Y. Liu, J. Mater. Sci. 39 (2004) 6153e6171.
calculating true stress exponent is depicted in Fig. 7b. The [3] H. Lianxi, W. Erde, Mater. Sci. Eng. A 278 (2000) 267e271.
calculated sthr values and associated true stress exponents are [4] B.W. Chua, L. Lu, M.O. Lai, Compo. Str. 47 (1999) 595e601.
given in Table 1. [5] J. Hashim, L. Looney, M.S.J. Hashmi, J. Mater. Process. Technol. 92e93
The true stress exponent values of the composites lie in the (1999) 1e7.
range of 5.4 and 5.8. As per literature, this can be attributed to [6] X.J. Wang, X.S. Hu, K.B. Nie, K. Wu, M.Y. Zheng, Trans. Nonfer. Met.
Soc. China 22 (2012) 1912e1917.
the dislocation climbing at higher temperature [21,30]. [7] C.Y.H. Lim, S.C. Lim, M. Gupta, Wear 255 (2003) 629e637.
[8] H. Dieringa, Y. Huang, P. Maier, N. Hort, K.U. Kainer, Mater. Sci. Eng.
4. Conclusions A 410e411 (2005) 85e88.
[9] Standard test methods for tension testing of metallic materials http://
Stir casting is an effective method for fabrication of AZ91- enterprise1.astm.org/DOWNLOAD/E8-04.1203314-1.pdf.
[10] K.K. Ajith Kumar, A. Viswanath, T.P.D. Rajan, U.T.S. Pillai, B.C. Pai,
SiCp composites with reinforcement content up to 25 wt-% Acta Metall. Sin. Engl. Lett. 27 (2014) 295e305.
and the reinforcements get near uniform distribution in the [11] M.C. Gui, J.M. Han, P.Y. Li, J. Mater. Eng. Perform. 12 (2003) 128e134.
matrix with very less agglomeration and have cleaner inter- [12] P. Poddar, V.C. Srivastava, P.K. De, K.L. Sahoo, Mater. Sci. Eng. A
face. The grain size reduction in 5 and 10 SiCp containing 460e461 (2007) 357e364.
composites is high when compared with 15, 20 and 25 SiCp [13] A. Luo, Metall. Mater. Trans. A 26 (1995) 2445e2455.
[14] S. Pal, K.K. Ray, R. Mitra, Mater. Sci. Eng. A 527 (2010) 6831e6837.
containing composites and variation in mechanical and creep [15] M. Jayamathy, S.V. Kailas, K. Kumar, S. Seshan, T.S. Srivatsan, Mater.
attributes are closely related with the same. The yield strength Sci. Eng. A 393 (2005) 27e35.
and ultimate compressive strength of the composites increases [16] K. Manigandan, T.S. Srivatsan, T. Quick, Mater. Sci. Eng. A 534 (2012)
with increase in reinforcement content. But, the ultimate ten- 711e715.
sile strength remains comparable with that of the base alloy. [17] P. Lukac, Z. Trojanova, Z. Drozd, Key Eng. Mater. Dev. Light Metals
121e128 (2000).
The base alloy and the composites with lower reinforcement [18] Z.Y. Ma, S.C. Tjong, Compo. Sci. Technol. 61 (2001) 771e786.
content (5 and 10 wt %) evince less significant difference in [19] F.A. Mohamed, K.T. Park, E.J. Lavernia, Mater. Sci. Eng. A 150 (1992)
minimum creep rates. However, the composites containing 15, 21e35.
20 and 25 wt% SiCp have an improved creep resistance. As the [20] F.A. Mohamed, T.G. Langdon, Acta. Metall. 22 (1974) 779e788.
true stress exponent values of the composites lie in the range [21] O.D. Sherby, P.M. Burke, Prog. Mater. Sci. 13 (1968) 323e390.
[22] S.L. Robinson, O.D. Sherby, Acta. Metall. 17 (1969) 109e125.
of 5.4 and 5.8, the creep can be attributed to the dislocation [23] O. Sherby, R. Klundt, A. Miller, Metall. Trans. A 8 (1977) 843e850.
climbing. [24] H.E. Evans, G. Knowles, Met. Sci. 14 (1980) 262e266.
[25] A.H. Clauer, N. Hansen, Acta. Metall. 32 (1984) 269e278.
Acknowledgment [26] E. Orowan, Dislocations and mechanical properties, The American
Institute of Mining and Metallurgical Engineering Inc, New York, 1954.
[27] E. Arzt, D.S. Wilkinson, Acta. Metall. 34 (1986) 1893e1898.
The authors are grateful to the Aeronautical Research and [28] E. Arzt, M.F. Ashby, Scr. Metall. 16 (1982) 1285e1290.
Development Board (ARDB), New Delhi for the financial [29] Y. Li, T.G. Langdon, Scr. Mater. 36 (1997) 1457e1460.
grant for this work (GAP 214739). Mr. Prasanth Sujayakumar [30] J. Weertman, J. Appl. Phys. 28 (1957) 1185e1189.
has been very helpful throughout the course of the work. The