Available online at www.sciencedirect.

com

Acta Materialia 59 (2011) 7662–7672

www.elsevier.com/locate/actamat

Unexpected influence of Mn addition on the creep properties

of a cast Mg–2Al–2Ca (mass%) alloy
T. Homma a,⇑, S. Nakawaki a, K. Oh-ishi a, K. Hono b, S. Kamado a
a
Department of Mechanical Engineering, Nagaoka University of Technology, Nagaoka 940-2188, Japan
b
National Institute for Materials Science, 1-2-1 Sengen, Tsukuba 305-0047, Japan

Received 5 April 2011; received in revised form 24 August 2011; accepted 30 August 2011
Available online 15 October 2011

Abstract

The addition of 0.3% Mn to a Mg–2Al–2Ca (mass%) alloy improves the minimum creep rate by one order of magnitude compared
with the Mn-free alloy. Network-shaped Mg2Ca (C14 structure) and (Mg,Al)2Ca (C36 structure) compounds coexist in the interdendritic
regions, while plate-shaped Al2Ca (C15 structure) phases are dispersed in the matrix. Ordered Guinier–Preston (GP) zones form dynam-
ically on the basal plane of the a-Mg matrix during creep deformation. The number density of the fine GP zones in the Mn-containing
alloy was found to be higher than that in the Mn-free alloy.
Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.

Keywords: Magnesium alloy; Creep; Guinier–Preston zone; Transmission electron microscopy; Atom probe

1. Introduction systems are ACM522 [6] and MRI153 [7], both of which
contain Al and Ca. Suzuki et al. observed the microstruc-
The interest in Mg alloys as structural materials is ture of die cast Mg–Al–Ca alloys by transmission electron
increasing due to their low specific weight; however, one microscopy (TEM) [8,9] and reported that Laves phases
weak point is their poor creep properties. Thus, various such as Mg2Ca (C14 structure) and (Mg,Al)2Ca (C36 struc-
heat-resistant Mg alloys have been developed, some of ture) formed. A Mg17Al12 (A12 structure) phase also
which are utilized in automotive components such as the appeared, depending on the Al and Ca contents and tem-
powertrain and engine covers [1]. Although rare earth ele- perature. The Mg17Al12 phase usually adversely affected
ments are commonly added to heat-resistant Mg alloys the high temperature mechanical properties.
such as WE54 and QE22 [2], their cost performance is inad- von Bush et al. reported that Ca addition to alloy
equate for automotive applications. Hence, the develop- AZ91D, giving MRI153, dramatically improved the creep
ment of new types of wrought Mg alloys using only properties at 150 °C and 50 MPa [7]. Similar results were
ubiquitous elements such as Al and Zn is desirable for obtained by Zhu et al. [10] for alloy MRI153, who found
automobile components. that the minimum creep rate measured at 175 °C and
It is known that Ca addition is effective in improving the 50 MPa was 2.5  109 s1. The improvement in creep
elevated temperature properties of Mg–Al-based alloy sys- resistance was attributed to the formation of a eutectic
tems [3,4]. This is because intermetallic compounds appear (Mg,Al)2Ca phase with a network microstructure.
at grain boundaries (GB), suppressing GB sliding at high Jing et al. reported that high temperature creep proper-
temperatures [4,5]. Representative commercial alloy ties can be improved by dispersion of the Al2Ca phase [11],
the structure of which was identified by Suzuki et al. [9] as a
⇑ Corresponding author. Tel.: +81 258 47 9760; fax: +81 258 47 9770. C15 structure, one of the Laves phases. Although the sam-
E-mail address: thonma@mech.nagaokaut.ac.jp (T. Homma). ple reported by Jing et al. is in fact an extruded AJC421

1359-6454/$36.00 Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.actamat.2011.08.049
 T. Homma et al. / Acta Materialia 59 (2011) 7662–7672 7663

(Mg–4Al–2Sr–1Ca–0.3Mn, mass%) alloy, the creep prop- of 50 MPa. No heat treatment was applied to the samples
erties were improved by Al2Ca phases internally in the prior to the creep test.
grains and Mg2Ca compounds at the GB. On the other The microstructures were observed using an Olympus
hand, Suzuki et al. applied an aging treatment to a die cast BX60M optical microscope and a JEOL JSM-7000F field
AXJ530 (Mg–5Al–3Ca–0.15Sr, mass%) alloy and found emission scanning electron microscope (FE-SEM)
that the creep properties were improved by fine dispersion equipped with EDX operated at 15 keV. Thin foils for
of the Al2Ca phase [12]. These facts suggest that the creep the TEM observations were prepared by low energy beam
properties can be improved not only by the GB com- ion thinning. Conventional TEM observations were made
pounds, but also the fine and intragranular precipitates. using a JEOL JEM-2010 microscope operated at
It is well known that Mn addition to Mg alloys can 200 keV. High resolution transmission electron microscopy
improve the corrosion resistance, particularly in Al-con- (HRTEM) micrographs were recorded parallel to the
taining Mg alloys [2,13], and therefore Mn is usually ½1 0 1 0a direction using a JEOL JEM-2100F microscope
added to commercial Mg alloys. On accidental addition operated at 200 keV. The specimen thickness and the struc-
of Mn to a Mg–Al–Ca alloy we also found an improve- tures of the dispersed precipitates were measured using a
ment in creep properties. Hence, the aim of this paper convergent beam electron diffraction (CBED) technique.
was to understand the effect of 0.3% Mn addition to a Details of the methods can be found elsewhere [14].
cast Mg–2Al–2Ca alloy system on the creep properties High angle annular dark field scanning transmission
based on microstructural observations by optical micros- electron microscopy (HAADF-STEM) observations were
copy (OM), energy-dispersive X-ray spectroscopy (EDX), carried out using the JEOL JEM-2100F microscope oper-
analytical TEM and three-dimensional atom probing ated at 200 keV. An objective lens with a spherical aberra-
(3-DAP). The novel effect of Mn addition on the creep tion coefficient of 0.5 mm was used to obtain a point to
properties and the effect of new precipitation strengthen- point resolution of about 0.19 nm. The HAADF-STEM
ing have been clarified. images were observed using an angular dark field detector
(GIF Tridiem). EDX analyses were performed with a
2. Experimental procedures JEOL JED-2300T. The spot size used was 1.5 nm, and 10
points were measured near the edge of the TEM specimens
Two types of alloys were investigated in this study. The in order to ignore any absorption or fluorescence of the X-
alloy compositions are listed in Table 1 in both mass and rays, the thin foil criterion [15].
mol per cent. In this investigation Al, Ca and Mn are Specimens for atom probe analyses were prepared by
denoted by A, X and M, respectively, following the ASTM mechanical grinding and micro-electropolishing. These
designation. Melting was conducted at 720 °C using Mg of specimens were observed with a field ion microscope, and
99.95% purity, then the Al and Al–10Mn master alloy were then analyzed by 3-DAP using a locally built machine
remelted in an electric furnace in a mixed gas atmosphere equipped with a CAMECA fast delay line detector. Field
of SF6 and CO2, and, finally, Ca was added to the melt. evaporation was assisted by femtosecond laser pulses
After stirring for about 1 min with the aid of Ar bubbling (Amplitude Systems s-pulse) under the following condi-
the melt was poured into a book mold held at 250–300 °C. tions: wavelength 343 nm (ultraviolet wavelength); pulse
The obtained cast dimensions were 150  120  40 mm. power 0.4 lJ; pulse frequency of 100 kHz [16,17]. The mea-
Creep test specimens and samples for microstructural sured temperature was 30 K.
observations were cut from the sample.
Flat specimens with a gauge length of 50 mm and 3. Results
reduced cross-sectional dimension of 10  5 mm were used
for the tensile creep test using an A&D CP-6L-500 machine 3.1. Creep curves
at temperatures of 175 °C and 200 °C and an applied stress
Fig. 1a and b shows creep curves measured at 175 °C
and 200 °C, respectively. The AX22 alloy shows the typical
three regions in the creep curves at both 175 °C and 200 °C:
(i) primary creep; (ii) steady-state creep; (iii) tertiary creep
Table 1 [18]. However, in the Mn-containing AXM220 alloy the
Chemical compositions in mass% and mol.%.
three creep regions are not clearly distinguished. In addi-
Composition tion, the AXM alloy shows a substantially improved (slow)
Mg Al Ca Mn creep rate. It is interesting to note that there is no report on
AX22 improved creep properties due to the addition of Mn,
mass% Balance 2.5 1.8 although most of commercial Mg alloys contain Mn in
mol.% Balance 2.3 1.1 order to improve their corrosion resistance. The minimum
AXM220 creep rates ð_eÞ at 175 °C are 5.6  109 and 2.8  1010 s1
mass% Balance 2.2 2.0 0.3 for the AX22 and AXM220 alloys, respectively, while those
mol.% Balance 2.0 1.2 0.1
at 200 °C are 5.0  108 and 1.0  109 s1 for the AX22
7664 T. Homma et al. / Acta Materialia 59 (2011) 7662–7672

Fig. 1. Creep curves obtained at (a) 175 °C and (b) 200 °C for 400 h for the AX22 and AXM220 alloys under an applied stress of 50 MPa. (c) Creep curves
measured at 175 °C for up to 2051 h are also given. Smoothing was utilized to prepare the creep curves.

and AXM220 alloys, respectively. e_ obtained for the particles in the AXM220 sample. Since the Al and Ca con-
AXM220 alloy is approximately one order of magnitude tents in the AX22 and AXM220 alloys are almost identical,
lower than that for the AX22 alloy. the types of network structure compounds in these two
samples must be same. Fig. 4 shows bright field (BF)
3.2. Grain boundary structures in the as-cast specimens images and selected-area electron diffraction (SAED) pat-
terns taken from the compounds. The SAED patterns indi-
In order to understand the effect of Mn addition on the cate that there are two phases, Mg2Ca and (Mg,Al)2Ca.
creep properties, low magnification microstructures were This result is consistent with that reported by Suzuki
observed by OM, as shown in Fig. 2. For both the AX22 et al. [9] at a ratio of Al to Ca of 1:1.
and AXM220 alloys network structured compounds can
be seen in the interdendritic regions. As a result the GB 3.3. Intragranular precipitates in the as-cast samples
are covered by these compounds. EDX maps taken from
the same sample indicate that Al and Ca are enriched in The BF TEM images of intragranular precipitates are
these compounds (Fig. 3). Mn is detected in some of the shown in Fig. 5, which were taken near B ¼ ½1 0 
1 0a .

Fig. 2. OM images observed from the as-cast specimens.

 T. Homma et al. / Acta Materialia 59 (2011) 7662–7672 7665

Fig. 3. EDX maps obtained from Mg, Al, Ca and Mn. SEI, secondary electron image.

Plate-like or disk-shaped precipitates lying on the basal the planar precipitates were estimated to be 1.7  1019 and
plane are uniformly dispersed. Based on foil thickness mea- 1.8  1019 m3 in the AX22 and AXM220 alloys, respec-
surements using a CBED technique the number densities of tively. Note that the number densities in the two specimens
7666 T. Homma et al. / Acta Materialia 59 (2011) 7662–7672

Fig. 4. BF TEM images obtained from network structure compounds as shown in Figs. 2 and 3.

Fig. 5. BF TEM images observed in the interior regions of the grains. The samples are as-cast and crept specimens.

are almost identical, regardless of Mn addition. The struc- by solid and white circles, while additional spots can also
ture of the plate-like precipitates was examined by a micro- be seen in the MBD patterns, as indicated by open and dot-
beam diffraction (MBD) technique. The results are shown ted circles. The diffraction patterns are consistent with an
in Fig. 6. Diffracted spots from the matrix are indicated Al2Ca phase with a C15 structure (a = 0.804 nm, space
 T. Homma et al. / Acta Materialia 59 (2011) 7662–7672 7667

Fig. 6. BF TEM images and MBD patterns of the plate-like precipitates. The white and solid circles are diffracted spots from the matrix, while the open
and dotted circles are diffracted spots from the C15 structure.

group F d   C15//
3 m), and {1 1 1}C15//(0 0 0 1)a and [0 1 1] again determined based on the foil thickness measure-
½0 
1 1 0a orientation relationship between the C15 structure ments, and were found to be 1.3  1019 and
and the a-Mg matrix, as reported previously [12]. 1.9  1019 m3 for AX22 and AXM220, respectively. Thus
In order to understand the chemistry of the Al2Ca phase the number densities do not change significantly before and
chemical analysis by 3-DAP was carried out. The 3-DAP after steady-state creep. Moreover, the number density is
elemental maps obtained from the as-cast AX22 and not altered irrespective of the addition of Mn.
AXM220 samples are shown in Fig. 7. As can clearly be In spite of the fact that the number density suggests that
seen, segregation of Al and Ca into the planar precipitates the Al2Ca phase does not influence improvements in the
is confirmed in both the samples. The ratio of Al to Ca is creep properties due to the addition of Mn, some fine con-
almost 2:1, indicating the formation of an Al2Ca phase trast newly appears in the matrix between the planar Al2Ca
(Fig. 7a and b). In addition, partitioning of Mn to the phases, as indicated by the white arrowheads in Fig. 8. The
Al2Ca phase can also be seen in the AXM220 sample. fine contrast can only be seen when the g vector of  2 1 1 0 is
Although Mn partitions into the planar Al2Ca phase the excited using the two beam condition. Large planar Al2Ca
number densities do not alter significantly, irrespective of phase precipitates are apparent. The plates are on the basal
Mn addition, according to the TEM analysis results. plane. The average diameters of the fine precipitates mea-
sured perpendicular to [0 0 0 1]a are 19.4 and 14.3 nm in
3.4. Intragranular precipitates in the crept specimens AX22 and AXM220, respectively. Thus, the diameter of
the fine precipitates decreases on Mn addition. Moreover,
The microstructures during creep showed subtle differ- strain contrast can be seen around the fine, planar precip-
ences when observed as conventional BF images, as indi- itates. It should be noted that streaks along the c-axis of
cated in Fig. 5. Specimens of the AX22 and AXM220 the hcp structure and at 1=3f1 1 2 0g and 2=3f1 1  2 0g
alloys were crept to the steady-state for 2051 h at 175 °C appear, as shown by the white arrows, in both the speci-
(Fig. 1c). In both alloys the Al2Ca phase precipitated on mens due to the presence of the fine, plate-like precipitates.
the basal plane of the matrix. The number densities were These facts suggest that ordered Guinier–Preston (GP)
7668 T. Homma et al. / Acta Materialia 59 (2011) 7662–7672

Fig. 7. 3-DAP elemental maps of Mg, Al, Ca and Mn in the as-cast specimens.

zones form in the AX22 and AXM220 alloys. The presence Average Al and Ca compositions of the AX22 alloy mea-
of these ordered GP zones was reported by Hono et al. in sured by EDX spot analyses are given in Table 2. The Al
several Mg-based alloys [19–21]. In addition, it was found and Ca contents of the AX22 alloy are 1.8 and 0.2 mass%,
that ordered GP zones form dynamically during creep respectively, while the Al and Ca contents of the AXM220
deformation, regardless of Mn addition. alloy are 2.0 and 0.45 mass%, respectively. Thus the Ca
The structure of the fine precipitate was further investi- composition in the matrix is increased slightly by Mn
gated by HRTEM imaging, as shown in Fig. 9. The fine GP addition.
zones observed in the AX22 and AXM220 alloys can be
confirmed, indicated by the white arrowheads. It is clear 4. Discussion
that the ordered GP zones possess a monolayer structure
on the basal plane, and the interface between the GP zone This study has shown the detailed microstructural evo-
and matrix is coherent. lution in the alloys AX22 and AXM220 before and after
3-DAP analysis of the AX22 and AXM220 alloys was creep testing, and the effect of Mn addition on the creep
performed after creep deformation. The results are shown properties has been addressed. For this purpose the influ-
in Fig. 10. It is clear that the GP zones are densely distrib- ence of Mn addition on the structure–property relationship
uted in the matrix, and are mainly composed of Al and Ca. has been investigated.
The number densities of the ordered GP zones determined Creep curves obtained at 175 °C and 200 °C (Fig. 1) pro-
from the 3-DAP elemental maps for the AX22 and vide a new insight into the effects of Mn addition in a Mg–
AXM220 alloys are approximately 5.1  1022 and Al–Ca alloy system. At both temperatures the minimum
2.5  1023 m3, respectively. In addition, Mn seems to par- creep rates of the AXM220 alloy were approximately one
tition into the GP zones, suggesting that Mn addition influ- order of magnitude lower than those of the AX22 alloy.
ences the increment in number density. Nevertheless, while the purpose of Mn addition to Mg
alloys is usually to improve the corrosion resistance, a dif-
3.5. Solid solution in the crept specimens ferent effect of Mn addition has been addressed in this
study, significant enhancement of the creep properties. In
The matrix composition after creep deformation was general Mn is added to Al-containing commercial Mg
also confirmed by EDX spot analysis. The measurements alloys, for example ACM522, MRI153, among others,
were carried out by checking the distribution of the Al2Ca but the effect has never been described as far as the authors
phase as well as GP zones in HAADF-STEM images. are aware.
 T. Homma et al. / Acta Materialia 59 (2011) 7662–7672 7669

Fig. 8. BF TEM images observed in the AX22 and AXM220 alloys crept at 175 °C for 2051 h. The incident beam direction is parallel to ½1 0 1 0a with
g¼ 2 1 1 0.

Fig. 9. HRTEM images obtained from ordered GP zones.

Network structure compounds form in the interdendrit- identified the structure as C15 (Al2Ca), in alloy AM50 with
ic regions after casting (Figs. 2 and 3). Luo et al. reported 1.72% added Ca, which is expected to yield effective GB
grains surrounded by a GB eutectic phase with (Mg,Al)2Ca strengthening or to resist plastic flow of the a-Mg grains
in alloy AC53 (Mg–4.5Al–3.0Ca). The intermetallic phase [5,22]. On the other hand, Suzuki et al. carefully analyzed
prevented GB sliding during creep [4]. Terada et al. also the eutectic compounds in alloy AXJ530 at the periphery
reported the presence of a eutectic phase on the GB, and of the a-Mg grains and found that C14 (Mg2Ca) and
7670 T. Homma et al. / Acta Materialia 59 (2011) 7662–7672

Fig. 10. 3-DAP elemental maps of Mg, Al, Ca and Mn in the crept specimens.

identical (1) before and after the creep test and (2) between
Table 2 the AX22 and AXM220 alloys. This suggests that there
Solute composition measured in the matrix by EDX spot analysis may be another effect by which the creep properties of
(mass%).
the cast AXM220 alloy is improved by Mn addition.
Matrix composition As indicated in Figs. 8 and 9, fine and ordered GP zones
Mg Al Ca Mn dynamically precipitate during creep deformation in both
AX22 Balance 1.8 0.2 – AX22 and AXM220. Indeed, strain fields are visible near
AXM220 Balance 2.0 0.45 0.25 the GP zones, as shown in Fig. 8. The strain fields should
provide an impediment to dislocation motion on the basal
and non-basal planes. It has also been reported that a large
C36 (dihexagonal (Mg,Al)2Ca) structures coexisted [8,9]. number density of precipitates and a coherent interface
Zhu also provided microstructural analyses on the creep between the precipitate and matrix may improve the creep
properties of alloy MRI153 and reported eutectic particles properties [23]. In fact, the number density of the GP zone
of a (Mg,Al)2Ca phase with a C36 structure together with a in the alloy AX22 is 5.1  1022 m3, while that in the
b-(Mg17Al12) phase with a bcc structure [10], thereby alloy AXM220 is 2.5  1023 m3. Moreover, dynamic
improving the creep resistance by obstructing dislocation precipitation during the creep test can significantly increase
annihilation at GB. In the present work it is found that the creep resistance, according to Lumley et al. [24],
C36 and C14 phases coexist in the interdendritic regions, although their alloy was an Al–Cu-based alloy. The bene-
as indicated in Fig. 4, and therefore a similar improvement ficial effect of dynamic precipitation is associated with the
in the creep properties can be expected: the reported Mg– retained solute content within the alloy, and the precipitate
Al–Ca alloy systems were produced by die casting or may act to provide greater obstacles to dislocation climb
squeeze casting and therefore their grain sizes were much and glide.
smaller than permanent mold cast Mg alloys. In the case The ordering can basically give rise to improvements in
of permanent mold casting GB sliding may not be impor- the creep properties [25]. In addition, L12 structure, known
tant in improving the creep properties compared with die as the c0 phase, significantly improves the high temperature
casting and squeeze casting technologies. mechanical properties of Ni-based superalloys due to the
Since the network structure compounds have similar ordering state, although the structure is long-range order
morphologies in both the AX22 and AXM220 samples rather than short-range order [25]. Thus the ordered GP
the compounds at GB are not expected to be obstacles to zones can be expected to improve the creep properties.
improving the creep properties by Mn addition. The Note that the GP zone observed in Al–Cu-based alloys
TEM observations showed that (0 0 0 1) platelets of Al2Ca cannot form at such high temperatures as 175 °C and
phase were dispersed within the grains (Fig. 5). This planar 200 °C due to reversion [26], and thus the dispersion of
Al2Ca phase has been reported by Jing et al. [11] and the ordered GP zones in Mg alloys at relatively high tem-
Suzuki et al. [12] in extruded AJC421 and die cast perature is beneficial in improving the high temperature
AXJ530 alloys, respectively. Jing et al. concluded that the mechanical properties.
fine Al2Ca phase effectively impedes dislocation movement Following the idea of combining precipitation strength-
during creep deformation at 175 °C and 70 MPa, while ening and solid solution strengthening reported by Terbush
Suzuki et al. mentioned that the Al2Ca phase is an effective et al. [27], the effect of solid solution strengthening in our
obstacle to non-basal dislocation glide at 175 °C and alloys has been evaluated, as indicated in Table 2. The 3-
80 MPa. Although the planar Al2Ca phase seems to pre- DAP analysis can also directly provide the matrix compo-
vent dislocation motion even in the AX22 and AXM220 sition, however, the result will be localized. EDX spot
alloys, the measured number densities were almost analysis was utilized to show compositional differences in
 T. Homma et al. / Acta Materialia 59 (2011) 7662–7672 7671

Table 3 On the other hand, Uesugi and Higashi [30] and Yasi
Symbols and definitions used in Eqs. (1) and (2). et al. [31] individually proposed the possibility of a reduc-
Symbol Definition tion in stacking fault energy (SFE) in Mg in the presence
v0 Constant of Ca based on first principles calculations. If the SFE
q Dislocation density decreases then the minimum creep rate shown in Eq. (2)
b Burgers vector will decrease, leading to an improvement in the minimum
M Taylor factor
kb Boltzmann constant
creep rate for dislocation climb. However, the addition of
T Absolute temperature Mn may, in fact, increase the SFE [31]. This would have
G Shear modulus an adverse effect on the minimum creep rate. Hence the
X Atomic volume effect of the Mn solid solution may be negligible.
e
D Diffusion coefficient for solute atoms The reason for the small increase in Ca content in the
r Applied stress
Cs Concentration of solute
matrix on addition of Mn may be that Al and Mn have a
dm Misfit parameter large negative enthalpy of mixing (70 kJ g1) as calcu-
L Distance lated by Miedema et al. [32]. Meanwhile, Mn and Ca have
hc Height a large positive enthalpy of mixing (59 kJ g1). According
Dl Lattice diffusion coefficient for solvent to the proposed models for ordered GP zones [19,20] the
Cj Density of jog on dislocation
csf Stacking fault energy
sites for solute elements may be occupied by Al and Ca
atoms in the case of AX22. If Mn is added to AX22 then
Mn may occupy some of the solute positions in the ordered
GP zone, ejecting Ca, due to the strong interaction between
Table 4
Al and Mn. As a result, the Ca content of the matrix is
Misfit parameters (dm) of solutes [29]. slightly enriched in AXM220 compared with AX22, lead-
Mg Al Ca Mn
ing to the prevention of dislocation motion due to solid
solution strengthening.
AX22 0.16 0.143 0.197 0.112
dm 0.11 0.23 0.3
5. Conclusions

the matrix, in combination with HAADF-STEM observa- The effect of 0.3% Mn addition on the creep properties
tions. The analyses were carried out on several regions of of the alloy AX22 has been investigated by OM, EDX, 3-
the internal grains. As can clearly be seen, Al and Ca solid DAP, TEM and HRTEM. The novel effect of Mn addition
solution strengthening occurs in both specimens. When Mn on the creep properties and the effect of precipitation
is added to alloy AX22, the Ca composition increases. This strengthening have been clarified. The following conclu-
means that the Ca solid solution is more effective in the sions can be drawn. The minimum creep rates obtained
AXM220 alloy than in the AX22 alloy, despite the fact that at 175 °C and 200 °C in AXM220 were 2.8  1010 and
the difference in Ca concentration is no more than 1.0  109 s1, respectively. These values are approxi-
0.25 mass%. There may be two possibilities for the preven- mately one order of magnitude lower than those for
tion of dislocation motion in the case of solid solution AX22. Network structure compounds containing Al and
hardening: (i) dislocation glide; (ii) dislocation climb. The Ca form in the interdendritic regions after casting. The
effect of solid solution strengthening on dislocation move- structures of the compounds are C14 and C36 in both
ment during creep may be expressed as: AX22 and AXM220. Mn addition does not influence the
e 1
qb k b T r D types of compounds. A plate-like Al2Ca phase with a
e_ glide ¼ v0 ð1Þ C15 structure (a = 0.804 nm, space group F d  3 mÞ is dis-
M GX G b2 C s dm 2
persed within the grains. Mn addition does not influence
1 q XG r Dl  csf 2 the number density of the Al2Ca phase before and after
e_ c lim b ¼ v0 bL Cj ð2Þ
M hc k b T G b2 Gb the creep test.
where the symbols used in the text are summarized in Ta- Ordered GP zones dynamically form on the basal plane
ble 3 [28]. The related misfit parameters in our samples of the a-Mg matrix during creep deformation in both the
are tabulated in Table 4. The misfit parameter is calculated AX22 and AXM220 alloys. The average diameter of the
from: GP zone is decreased by Mn addition. The measured num-
rs  r0 ber densities in AX22 and AXM220 are approximately
dm ¼ ð3Þ 5.1  1022 and 2.5  1023 m3, respectively. They are
r0
coherent with the matrix, and possibly impede dislocation
where rs and r0 are the atomic radii of the solute and sol- motion. The Ca content of the matrix is slightly increased
vent elements, respectively. Thus we can expect that Ca by the addition of Mn. This may lead to solid solution
and Mn are effective elements in reducing the minimum strengthening. As a result, Mn addition can improve the
creep rate for dislocation glide according to Eq. (1). creep properties by a combination of precipitation
7672 T. Homma et al. / Acta Materialia 59 (2011) 7662–7672

strengthening by the ordered GP zone and solid solution [10] Zhu SM, Mordike BL, Nie JF. Metall Mater Trans A
strengthening by Ca. 2006;37A:1221–9.
[11] Jing B, Yangshan S, Feng X, Shan X, Jing Q, Weijian T. Scripta
Mater 2006;55:1163–6.
Acknowledgements [12] Suzuki A, Saddock ND, Termush JR, Powell BR, Jones JW, Pollock
TM. Metall Mater Trans A 2008;39A:696–702.
This work was supported in part by Grants-in-Aid for [13] Emley EF. Principles of magnesium technology. New York: Perg-
amon Press; 1966.
Scientific Research 22246094 (2010) and 21360348 (2009) [14] Homma T, Mendis CL, Hono K, Kamado S. Mater Sci Eng A
and the World Premier International Research Center Ini- 2010;527:2356–62.
tiative on Materials Nanoarchitectonics, MEXT, Japan. [15] Williams DB, Carter CB. Transmission electron microscopy. New
Mr. R. Asakawa and Mr. H. Takahashi of Nagaoka Uni- York: Plenum Press; 1996.
versity of Technology, Japan, are gratefully acknowledged [16] Hono K, Okubo T, Chen YM, Kodzuka M, Oh-ishi K, Sepehri-
Amin H, Li F, Kinno T, Tomiya S, Kanitani Y. Ultramicroscopy
for their support. TH also wish to thank Associate Profes- 2011;111:576–83.
sor C.R. Hutchinson of Monash University, Australia, for [17] Oh-ishi K, Mendis CL, Ohkubo T, Hono K. Ultramicroscopy
his help and encouragement. 2011;111:715–8.
[18] Kassner ME, Prez-Prado MT. Prog Mater Sci 2000;45:1–102.
[19] Ping DH, Hono K, Nie JF. Scripta Mater 2003;48:1017–22.
References [20] Oh-ishi K, Watanabe R, Mendis CL, Hono K. Mater Sci Eng A
2009;526:177–84.
[1] Mordike BL, Ebert T. Mater Sci Eng A 2001;302:37–45. [21] Jayaraji J, Mendis CL, Ohkubo T, Oh-ishi K, Hono K. Scripta Mater
[2] Friedrich HE, Mordike BL. Magnesium technology. Berlin: Springer; 2010;63:831–4.
2006. [22] Terada Y, Ishimatsu N, Mori Y, Sato T. Mater Trans 2005;46:145–7.
[3] Ninomiya R, Ojiro T, Kubota K. Acta Metall Mater 1995;43:669–74. [23] Kassner ME. Fundamentals of creep in metals and
[4] Luo AA, Balogh MP, Powell BR. Metall Mater Trans A alloys. Oxford: Elsevier; 2009.
2002;33A:567–74. [24] Lumley RN, Morton AJ, Polmear IJ. Acta Mater 2002;50:3597.
[5] Terada Y, Sota R, Ishimatsu N, Sato T, Ohori K. Metall Mater Trans [25] Nabarro FRN, de Villiers F. Physics of creep and creep resistant
A 2004;35A:3029–32. alloys. London: Taylor & Francis; 1995.
[6] Koike S, Washizu K, Tanaka S, Baba T, Kikawa K. Honda R&D [26] Guinier A. Solid State Phys 1959;9:293–398.
Tech Rev 2001;12:167–73. [27] Terbush JR, Saddock ND, Jones JW, Pollock TM. Metall Mater
[7] von Bush F, Schumann S, Friedrich H, Aghion E, Bronfin B, Trans A 2010;41A:2435–42.
Mordike BL, Bamberger M, Eliezer DM. In: Kaplan HI, editor. Mag [28] Maruyama K, Nakajima E. Materials science of high temperature
tech 2002. Warrendale (PA): TMS; 2002. p. 61–7. strength. Tokyo: Uchida-Rokakuho; 2002.
[8] Suzuki A, Saddock ND, Jones JW, Pollock TM. Scripta Mater [29] Takeuchi T, Inoue A. Mater Trans 2005;46:2817–29.
2004;51:1005–10. [30] Uesugi T, Higashi K. J JILM 2004;54:82–9.
[9] Suzuki A, Saddock ND, Jones JW, Pollock TM. Acta Mater [31] Yasi JA, Hector Jr LG, Trinkle DR. Acta Mater 2010;58:5704–13.
2005;53:2823–34. [32] Miedema AR, de Boer FR, Boom B. CALPHAD 1977;1:341–59.