Homogeneous limit of Cd(1-x)Mn(x)GeAs(2) alloy: electrical and magnetic properties

0.31 MB·

by  L. Kilanski

Homogeneous limit of Cd Mn GeAs alloy: electrical and magnetic
1-x x 2
properties
L. Kilanski,1,a) M. G´orska,1 E. Dynowska,1 A. Podg´orni,1 A. Avdonin,1 W. Dobrowolski,1 I. V. Fedorchenko,2
and S. F. Marenkin2
1)Institute of Physics, Polish Academy of Sciences, Al. Lotnikow 32/46, 02-668 Warsaw,
Poland
2)Kurnakov Institute of General and Inorganic Chemistry RAS, 119991 Moscow,
Russia
(Dated: 10 April 2014)
4
1 We present the studies of structural, electrical, and magnetic properties of bulk Cd Mn GeAs crystals
1-x x 2
0
withlowMn content, x, varyingfrom0to 0.037. The studiedsampleshaveexcellentcrystallographicquality
2
indicatedbythepresenceofdiffractionpatternsneverbeforeobservedexperimentallyforthiscompound. The

r electricaltransportin our samples is dominated by thermal activation of conducting holes from the impurity
p
statestothevalencebandwithactivationenergyofabout200meV.Thedefectstatesactingasionicscattering
A
centers with concentration in the range from 6 to 15×1017 cm−3 are observed. The effective Mn content in

9 our samples, x¯ , determined from fit of the susceptibility data to the Curie-Weiss law, is very close to the
θ
averagechemicalcontent,x. Itindicatesthatthe Mnionsaredistributedrandomly,substitutingtheCdsites

] in the host CdGeAs lattice. We observe a negative Curie-Weiss temperature, |θ|≤3.1 K, increasing as a
i 2
c function of x. This indicates the significance of the short-range interactions between the Mn ions.
s
-
l PACS numbers: 72.80.Ga,75.30.Hx, 75.30.Et, 75.50.Pp
r
t Keywords: semimagnetic-semiconductors;magnetic-impurity-interactions,exchange-interactions
m
.
t
a I. INTRODUCTION inducing magnetoresistive effects with either negative
m values up to -50% for Zn Mn GeAs at T =1.4 K
1-x x 2
- ThestudiesonsemiconductorsbasedonMn-alloyedII- (Refs.10 and 12) or positive values for Cd1-xMnxGeAs2
d VI materials1,2 haveledtodevelopmentofa newclassof with 3 at. % of Mn at T =1.4K (Ref. 11). However, in
n
o semiconductors called diluted magnetic semiconductors order to understand the complex magnetic properties of
c (DMS). The discovery of carrier mediated ferromag- ferromagnetic semiconductor systems, it is necessary to
[ netism in p-type IV-VI semiconductors (Ref.3) followed study low paramagnetic ion alloying regime, where the

with development of III-V group compounds (Ref. 4) aggregationof magnetic impurities does not occur.
2
v resulted in considerable development of these materials In the present paper we focused on the
7 in view of their possible application in spintronics. In Cd1-xMnxGeAs2 crystals without any signatures of
3 DMS it is possible to study and independently tune MnAs clusters, observed before in the samples with
0 and control their electronic and magnetic properties. x>0.05.11 The homogeneous Mn distribution in the
4 Apart from classical III-V, IV-VI, and II-VI DMS, in Cd1-xMnxGeAs2 lattice can be accomplished by lim-
1. which the Curie-Weiss temperature, θ, does not exceed iting the Mn content, x, to 0.037. The thermally
0 200K, there are several other groups of prospective and activated transport was observed in our samples with
4 intensively studied materials.5 impurity states with activation energy Ea=200 meV.
1 Complex diluted magnetic semiconductors, such Our samples showed paramagnetism with the effective
v: as II-IV-V chalcopyrite materials, are perceived as Mn-concentration reduced with respect to the average
2
i prospective candidates for being used in spintronics.6,7 Mn content, x.
X
Room temperature ferromagnetism in Zn Mn GeAs
1-x x 2
r
a and Cd1-xMnxGeAs2 alloys with θ as high as 367 K
for Zn1-xMnxGeAs2 with 3 at.% of Mn8 was found II. STRUCTURAL CHARACTERIZATION
(via direct observation of NMR spectra characteristic
of MnAs hyperfine structure) to be related to the
For purpose of this research bulk Cd Mn GeAs
presence of MnAs clusters.9–11 The combination of 1-x x 2
crystals with low Mn content, x, varying from 0 up to
the semiconductor matrix and metallic clusters called
0.037, were prepared by using the vertical Bridgman
nanocomposite can be prospective from the point of method.13 Our samples were synthesized from stoichio-
view of applicable magnetotransport effects. The MnAs
metricratiosofCdAs ,Ge,andMnofatleast5Npurity.
2
clusters, however, interact with the conducting carriers
The sample synthesis was done in vacuum sealed (to a
pressure of about 10−2 Pa) graphitized quartz ampoules
with the CdGeAs seed. The growth was performed at
2
thetemperatureofabout950±0.5K.Thegrowthwasfin-
a)Electronicmail: [email protected] ished with rapidcooling (about 5-10K/s)downto room
2
temperature in order to prevent segregationof magnetic
impurities and increase the uniformity of the as-grown
crystals. A more detailed description of the growth pro-
cedure can be found in Refs.14 and 15.
Cd Mn GeAs
The chemical composition of the as-grown samples 0.963 0.037 2
was determined using the energy dispersive x-ray flu-
5
orescence method (EDXRF). Measurements were done 10
at room temperature with the use of the Tracor X-ray
y
Spectrace 5000 Spectrometer equipped with Si(Li) de- sit
tector. The as-grown Cd1-xMnxGeAs2 ingots were cut en
t
into1mm-thickslicesperpendiculartothegrowthdirec- n
I 4
tion. The measured Mn molar fraction, x, determined 10
with maximum relative uncertainty of EDXRF method
not exceeding 10%, is found to be almost constant as
a function of the ingot length. The EDXRF data in-
dicate that all our samples have a proper stoichiometry
3
of the compound equal to 1:1:2. The EDXRF spectra 10
show no evidence of the fluorescent lines coming from 20 40 60 80 100 120
unwanted dopants in the studied alloy. This means that 2 [deg]
FIG. 1. Diffraction patterns observed for the exemplary
unintended impurity concentrations in the alloy are less
than 1015 cm−3. The values of the chemical composi- Cd1-xMnxGeAs2 sample with x=0.037.
tion, x, together with the estimated experimental errors
are gathered in Table I. The EDXRF analysis indicates
TABLEI.Preliminaryresultsofabasicstructuralcharacter-
that our samples have chemical composition, x, chang-
ization including the chemical composition, x, and the chal-
ing from 0 for pure CdGeAs2 crystal up to 0.037. The copyritestructure lattice parameters, a and c, respectively.
presentsamplesareofamajorimprovementwithrespect
x±∆x a±∆a (˚A) c±∆c (˚A)
toouroldercrystals(seeRef.11)due togoodstoichiom-
0 5.9452±0.0002 11.2153±0.0003
etry. 0.004±0.001 5.9425±0.0002 11.2207±0.0006
Thestructuralqualityofoursampleswasstudiedwith 0.013±0.001 5.9439±0.0002 11.2156±0.0006
the use of the high resolution x-ray diffraction method 0.024±0.002 5.9436±0.0002 11.2116±0.0004
(HRXRD).Weusedmultipurposediffractometer(X’Pert 0.037±0.003 5.9417±0.0003 11.2082±0.0006
PRO MPD, Panalytical configured for Bragg-Brentano
diffractiongeometry)equipped with a stripdetector and
an incident-beam Johansson monochromator. The Cu data for this compound18: a=5.942˚A and c=11.224˚A.
K x-ray radiation with wavelength equal to 1.5406˚A
α1 Nearly monotonic, decreasing a(x) and c(x) dependen-
was used. This instrument allows obtaining diffrac-
cies suggest that the Mn ions substitute the Cd ions in
tion patterns with an excellent resolution and counting
thecrystallattice,sincethetetrahedralradiusoftheMn
statistics. The obtained HRXRD patterns for both pure ion is smaller than that of Cd.19 The a(x) and c(x) de-
CdGeAs crystalandMn-alloyedsamples showexcellent
2 pendencies are fitted with the linear function of x with
crystalline quality of our samples (see Fig.1). The crys-
good agreement between the experimental data and the
tal structure of our samples is improved with respect to
fitted lines. The lattice parameters change as a function
the old results.14,17 The HRXRD patterns obtained for
of x with a slope similar to that presented in Ref. 17.
alloursampleswereanalyzedwiththeuseofRietveldre-
A more detailed description of the HRXRD data can be
finementmethod. Thedataanalysisindicatesthatallour
found in Ref.20.
samplescrystallizedintetragonalchalcopyritestructure.
TheRietveldfitsallowedprecisecalculationofthelattice
parameters for the samples with different chemical com-
III. MAGNETOTRANSPORT DATA
position, x. The results ofour calculationsarepresented
in Table I. The lattice parameters are Mn composition
In order to obtain fundamental information about
dependent. The a(x) andc(x) dependencies aredecreas-
the electricaltransportin Cd Mn GeAs crystals with
ing functions of the Mn content for most of our crystals 1-x x 2
varyingchemicalcompositionweperformedthetempera-
with exception of the sample with x=0.004. However,
turedependentHalleffectmeasurements. TheHalleffect
the volume of the unit-cell V is a monotonic, decreasing
measurementsweredone usinga standardsixcontactdc
function of x. It indicates that in the case of the two
currenttechnique inthe presence ofstaticmagnetic field
samples with the smallest Mn content in our series, the
of induction not exceeding B=13 T. Prior to magneto-
deformationofthelatticeconstantoccurredalongthedif-
transport measurements the samples were cut into Hall
ferent crystallographic axes. The lattice parameters ob-
bars with dimensions of about 1×1×8 mm3, chemically
tainedbyusforpureCdGeAs areclosetotheliterature
2 cleaned and etched. The contacts to the samples were
3
16
10 200
Cd1-xMnxGeAs2 6 1017 cm-3
180
x = 0
Ea 200 meV
x = 0.024 160
-3 m]1015 x = 0.037 s)]140 9 1017 cm-3
V
c
[ (120
1 2 /
- R)H 14 [cm100 1.5 1018 cm-3
(e10 80
60
x = 0.024
40
13 x = 0.037
10 20
x = 0
0
3 4 5 6 7 8 9 100 150 200 250 300
-1
1000/T [K ] T [K]
FIG. 2. The Hall carrier concentration p=(eRH)−1 as a FIG. 3. The Hall carrier mobility as a function of the tem-
function of the inverse of the temperature for the selected perature for selected Cd1-xMnxGeAs2 samples with different
Cd1-xMnxGeAs2 samples with different chemical composi- chemicalcompositions. Themarkersrepresenttheexperimen-
tions(symbols). Thelinesshowfitstothethermallyactivated tal data and the lines are obtained form fits to Eq. 1. The
carrier transport model. fitted concentrations, NI, are shown next to thelines.
tionσ =epµ,whereσ istheconductivitytensorcom-
made using gold wire and indium solder. The measure- xx xx
ponent parallelto the currentdirection, measuredin the
ments were done over a broad temperature range from
absence of the external magnetic field, and e is the el-
T =1.4K up to 300K.
ementary charge. The carrier mobility as a function of
We measured a series of isothermal magnetic field
the temperature, calculated for each studied sample, is
dependencies of the off-diagonal resistivity component
presentedfor selected Cd Mn GeAs crystalswith dif-
ρ (B). The obtained ρ (B) dependencies are linear 1-x x 2
xy xy ferentchemicalcompositionsinFig.3. Thedataindicate
over the studied temperature range. No anomalous Hall
thatinalloursamplesaboveT ≈180Kthephononscat-
effect was observed in our samples. As a result of the
tering is responsible for the decrease of the carrier mo-
ρ (B) measurementswecalculatedthe temperature de-
xy bility with increasing the temperature. The maximum
pendence of the Hall carrier concentration p for all our
value of the carrier mobility for our samples is not pro-
Cd Mn GeAs samples presented in Fig. 2. Our sam-
1-x x 2 portionaltotheaverageMncontent,x. Itindicates,that
ples became highly resistive at temperatures lower than
Mn ions are not directly responsible for the carrier scat-
100K andthe Hallconstantbecame unmeasurable. The
tering in our samples. This conclusion is in agreement
Hall constant for all our samples points into the p-type
with the data gathered in Fig. 2. Below T ≈180 K a
conductivity with relatively low carrier concentration,
p<1016 cm−3, at room temperature. The Hall effect decreaseofthemobilitywithdecreasingthetemperature
is connected with impurity scattering. The µ(T) depen-
results show that the carrier transport in our samples is
dence at temperatures between 100K and 170K can be
thermally activated in the valence band at temperatures
fairly well fitted to the T3/2 dependence, indicating that
higher than 150 K. The data gathered in Fig. 2 allowed
theconcentrationofscatteringcentersistemperaturein-
ustoestimatetheactivationenergyofbandcarriers,E .
a dependent due to compensation.
The calculated values of E are almost constant as a
a The Hallcarriermobility inp-type GaAs canbe fitted
function of the Mn-content, x, and are equal to about
with the use of empirical two-band model in the low-
200 meV. Similar activation energies for both pure and
field approximation, describing the hole conductivity in
Mn-alloyed samples together with the fact that Mn is
valence and impurity bands. The Hall effect scattering
isovalent in chalcopyrite structure indicates that the Mn
factors are assumedto be equalto unity for all our sam-
ions are not directly related to the presence of impurity
ples. Theempiricalmobility,asdescribedinRef.22,may
states. It is then evident that other defect types, pos-
be fitted to the expression
sibly antisite defects21 that are the main defect type in
CdGeAs2, because they have lower formation energies 1 1 1
= + , (1)
than Cd or Ge vacancy defects, are present with moder- µ A·T3/2·N−1 B·T−2.3
ate concentration in our samples and are responsible for I
the observed impurity states. where N is the concentration of scattering centers, A
I
The Hall carrier mobility, µ, is defined from the equa- andB areconstantsdescribingtheionizedimpurityscat-
4
tering and the combined scattering due to phonons, re-
spectively. For our purpose we took the A and B values
from GaAs, a binary equivalent ofternary CdGeAs . At
2
low temperatures, where ionic impurity scattering pro- x = 0.004
cesses are of a major importance the A value is equal to x = 0.013
2.5×1020(cm·V·s)−1, while at high temperatures, where u]1.5 x = 0.024
m
phononscatteringis the mainscatteringmechanism,the e x = 0.037
/
value ofB equals2.5×103.22 We performedfitting ofthe 6 g
experimental µ(T) data (symbols in Fig.3) to the Eq.1 101.0
with A=2.5×1020 (cm·V·s)−1, B=2.5×103, and NI as -1 ] [
the only fitting parameter. As a resultof the fitting pro- )
cedure the NI values were estimated for each sample. Re(0.5
The concentration of the ionic scattering centers, NI, [
changes from 6×1017 cm−3 to 1.5×1018 cm−3, for the
sample with x=0.024 and 0, respectively. The obtained
0.0
defectconcentrationvaluesforthesampleswithdifferent
xarenotproportionaltothe amountofMn. Itindicates
0 20 40 60 80 100 120 140 160
thattheMnincorporationintheCdGeAs doesnotpro-
2 T [K]
duce ionic scattering centers. FIG. 4. The inverse of the real part of the dynamic
The magnetoresistance measurements were also done magnetic susceptibility measured (symbols) for the selected
in parallel with the Hall effect measurements. All our Cd1-xMnxGeAs2 samples with different chemical composi-
samples have high resistance and an absence of free car- tion. ThelinesrepresenttheoreticalcurvesfittedtotheEq.2
riers at temperatures lower than 100 K. At T >100 K .
the obtained magnetoresistance curves for all our sam-
ples indicate the presence of positive magnetoresistance
of the form
with relative amplitudes not exceeding 2%. The positive
C
magnetoresistancescaleswiththesquareofthemagnetic χ(T)= +χ , (2)
dia
field indicating that this effect is the classical effect due T −θ
to the orbitalcarriermovementinthe presenceofanap-
where
plied magnetic field.
N g2µ2S(S+1)x¯
C = 0 B θ. (3)
3k
B
Here C is the Curie constant, χ =−2.5×10−7 emu/g
IV. MAGNETIC PROPERTIES dia
is the diamagnetic contribution to the magnetic sus-
ceptibility originating from the host lattice (the value
Magnetic properties of our Cd Mn GeAs samples wasdeterminedfromourmagnetizationmeasurementsof
1-x x 2
were studied with the use of LakeShore magnetome- CdGeAs ), N is the number of cation sites per gram, g
2 0
ter/susceptometer system. This instrument allows mea- istheg-factorofthemagneticion(forMng=2),S=5/2
surements of both static and dynamic magnetic proper- isthespin-magneticmomentumoftheMnions,µ isthe
B
tiesofsolids. Thetemperaturedependentdynamicmag- Bohr magneton, k is the Boltzmann constant, and x¯
B θ
netic susceptibility as well as the magnetic field depen- is the effective magnetically-active Mn content.
dencies were measured. The experimental data gathered in Fig. 4 were fit-
Thedynamicmagneticsusceptibility,χ,wasmeasured ted to the Eq. 2 assuming the constant value of the
over the temperature range from 4.3 up to 200 K. The diamagnetic contribution to the magnetic susceptibil-
sample was put into the alternating magnetic field with ity χ =−2.5×10−7 emu/g estimated above for the
dia
frequencyf equalto625HzandamplitudeH =10Oe. pure CdGeAs sample. We fitted the experimental
AC 2
As a result, we obtained the temperature dependencies (Re(χ ))−1(T)curves with two fitting parameters: the
AC
of both real and imaginary parts of the ac susceptibility. Curie-Weiss temperature θ and the Curie constant C.
ThemeasurementsperformedonallourCd Mn GeAs The fitted curves are presented together with the exper-
1-x x 2
crystals indicated a vanishing imaginary part of the ac imental data in Fig. 4. As we can clearly see the mag-
magnetic susceptibility. The temperature dependence of neticsusceptibilityofoursamplescanbeverywellrepro-
theinverseoftherealpartoftheacmagneticsusceptibil- duced with the use of the Curie-Weiss law defined with
ityfortheselectedCd Mn GeAs samplesispresented Eq.2. The fitting parameters, θ and C, are gathered for
1-x x 2
in Fig. 4. Our results clearly indicate the paramagnetic all our samples in Table II. The Curie-Weiss tempera-
behaviorofthe(Re(χ ))−1(T)foralloursamples. The tures, determined for all our samples are negative with
AC
temperature dependence of the inverse of the magnetic the values increasing as a function of the Mn content, x.
susceptibility, (Re(χ ))−1(T), was fitted over the tem- This indicates the significance of the short-range inter-
AC
peraturerange20−160KtotheCurie-Weissexpression actions between the Mn ions. The Curie constant values
5
zinc-blende II-VI DMS with Mn24. Larson et al. sug-
TABLEII. Susceptibility parameters for Cd1-xMnxGeAs2. gested that in the superexchange interaction the anion
x C (10−5) x¯θ θ [K] J/kB [K] is of more importance than the nonmagnetic cation.26
[emu·K/g] Therefore,itwouldbealsointerestingtocompareourre-
0.004 7.0±0.2 0.005±0.001 -0.08±0.02 -0.46 sults with magnetic properties ofcompounds with group
0.013 9.4±0.2 0.007±0.001 -0.41±0.03 -1.69
V anion, e.g., III-V DMS. This is not trivial, since in
0.024 28±2 0.021±0.002 -2.4±0.2 -3.26
Mn-dopedIII-VDMSthedivalentmanganesesubstitutes
0.037 66±3 0.050±0.005 -3.1±0.3 -1.77
the trivalent group III cation. That results in high hole
concentration, usually above 1021 cm−3. In materials
with such high free carrier concentration we observe a
obtained for our samples, gathered in Table II, are an
carrier induced ferromagnetism due to the RKKY ex-
increasingfunctionoftheMncontent. Itindicatesanin-
change interaction and ferromagnetic Curie-Weiss tem-
creaseintheamountofmagneticallyactiveMnionswith
peratures of an order of 100 K (for a review see e.g.
increasingxinoursamples. ThevaluesoftheCuriecon-
Ref. 5). To find III-V DMS with magnetic properties
stant can be used to calculate the effective Mn content,
like in our material we must consider only very dilute
x¯ , using Eq. 3. The estimated x¯ values are gathered
θ θ III-V DMS in which the free carrier concentration does
in Table II. The discrepancy between the x and x¯ val-
θ not exceed 1019 cm−3. In In Mn As with free car-
ues is almost zero (within the accuracy of the estima- 1-x x
rier concentration below 1019 cm−3 J/k equals -1.6 K
tion of both quantities) for the sample with the smallest B
(Ref. 27) and in Ga Mn N with free carrier concen-
x=0.004. For higher x the difference between x and x¯ 1-x x
θ tration below 1018 cm−3 J/k equals -1.9 K(Ref. 28).
increases. It is a signature that the distribution of the B
These values are very similar to the values of J/k ob-
Mn ions in the Cd Mn GeAs lattice becomes imper- B
1-x x 2 servedhere. Ourresultsindicate,thatthesuperexchange
fect. Therearetwoexplanationsofthisdifference: (i)the
interactioninchalcopyriteDMSdepends stronglyonthe
charge state of a fraction of Mn ions might be different
type of the chemical bond, as was suggested for other
than Mn2+, reducing its total magnetic momentum and
DMS experimentally24,25 and theoretically.26
(ii) antiferromagnetic Mn pairing leading to turning off
The static magnetic properties of Cd Mn GeAs
their magnetic moment. However, it should be empha- 1-x x 2
samples were studied with the use of the Weiss extrac-
sized,thatnosignaturesofthepresenceoflargeMnAsor
tion method employed to the LakeShore 7229 suscep-
other clusters are present in our samples. For the high-
tometer/magnetometersystem. Thisinstrumentallowed
est Mn content, x=0.037, the x¯ value is higher than
θ us to study the magnetic moment at different temper-
x. The difference (taking into account the uncertainties
atures from 4.3 up to 200 K and in the presence of
in x and x¯ estimation) might originate from clustering
θ an external magnetic field with maximum induction not
and/or underestimation of the experimental errors.
Gudenko et al. studied the electron paramagnetic res- exceeding B=9 T. Our studies covered the isothermal
measurements of the magnetization M as a function
onance (EPR)in Cd Mn GeAs with x=0.06.23 They
1-x x 2 of the magnetic field B done at few selected tempera-
reported two Curie-Weiss temperatures: θ =255K and
1 tures T <200K. The magnetization as a function of the
θ =−3 K. The Curie-Weiss temperatures for our sam-
p2les are close to θ . Gudenko et al. attributed the θ to magnetic field measured at T =4.5 K for the selected
2 2 Cd Mn GeAs crystalswith differentchemicalcompo-
the centers where Mn ions substitute the Cd ions. We 1-x x 2
sition,x,ispresentedinFig.5. TheM(B)measurements
suppose that this is the case in all our samples.
showed the lack of measurable magnetic irreversibility
Since our experimental results show that θ≪T, we
of our samples. The magnetization results obtained at
canuseθ to estimatethe exchangeinteraction,whichwe
T >5K are typical of paramagnetic materials.
haveassumedto be the nearest-neighborexchange,from
Because the M(B) curves at low temperatures carry
the relation
mostinformationabout the paramagneticmedia and for
3θ clarity of presentation only the results obtained at the
J/k = , (4)
B
x¯ S(S+1)z lowest measured temperature, T =4.3 K, are presented
θ
and analyzed.
wherezisthenumberofnearest-neighborcationsites,12
The M(B) curve for the nonmagnetic CdGeAs sam-
forthechalcopyritestructure. (Werememberthatinour 2
ple showed a diamagnetic, negative slope. The diamag-
chalcopyrite compound x refers to one kind of cations,
netic susceptibility for CdGeAs crystal, estimated from
i.e., half of the total cation number.) the slope of the M(B) curve, χ2 =−2.5×10−7 emu/g.
In the last column of Table II we present the values dia
This result is in agreementwith the diamagnetic suscep-
of the exchange parameter, J/k , for all our samples.
B tibility value determined from the ac susceptibility data.
The errors,estimatedfromthe scatterinthe experimen-
Themagnetizationforthesampleswithx≤0.013reaches
tal data and from the uncertainties in the fitting, are
saturation in moderate magnetic fields, B<9T, used in
about ±30% for J/k . The values of J/k are nega-
B B our experiments. However, for x>0.013 the magneti-
tive and relatively small, characteristic for the antifer-
zation curves do not show saturation. Therefore, the
romagnetic superexchange via an anion.24,25 They are
M(B) curves had to be fitted with the modified Bril-
of the same order but somewhat smaller than those for
6
similartotheCurie-Weisstemperature,θ,withintheac-
curacyofestimationofbothquantities. We estimate the
error in T to be about 20%. The only exception oc-
0
2.5
curs for the sample with x=0.037; that might originate
T = 4.5 K
from the increasing influence of Mn-pairs or short-range
0.037
clustering in the magnetism of this compound. The sat-
2.0
urationmagnetization, M , estimated for our samples is
S
an increasingfunction of x. The M values were used in
] 0.024 S
u/g1.5 order to calculate the amount of magnetically active Mn
m ions, x¯ , using Eq.6. For the samples with the low Mn
m
e
M [ content, x≤0.013, the calculated x¯m values (Table III)
1.0 areveryclosetoxandx¯θ. Forhigherx,however,wehave
x≈x¯ ≤x¯ for x=0.024 and x≤x¯ ≈x¯ for x=0.037.
θ m θ m
Thediscrepancybetweenthevalues,especiallyvisiblefor
0.5 0.013 x=0.037,mightoriginatefromlocalimperfectionsinthe
sample. We have very good agreement between the sus-
0.004 ceptibility and magnetization data. However, it is also
0.0 possible that we underestimated the error in parameter
0 2 4 6 8
B [T] estimation for the sample with x=0.037. It must be
FIG. 5. The magnetization as a function of the applied
emphasized, that our data indicate a rather good distri-
magnetic field measured (symbols) at T=4.5 K for the se-
bution of Mn ions in our samples.
lected Cd1-xMnxGeAs2 samples with different chemical com-
positions. The lines represent the theoretical curves fitted to
theEq.5.
V. SUMMARY
TABLE III. Parameters for the modified Brillouin function
We explored the structural, electrical, and magnetic
fit.
propertiesofbulk Cd Mn GeAs crystalswithlowMn
x MS [emu/g] x¯m T0 [K] content, x, varying fr1o-xm 0xto 0.0327. Our samples have
0.004 0.27±0.02 0.005±0.001 0.061±0.012
0.013 0.43±0.03 0.008±0.002 0.63±0.12 good structural quality, with lattice parameters chang-
0.024 2.0±0.2 0.037±0.005 2.8±0.5 ing as a function of the Mn content, x, according to the
0.037 2.7±0.3 0.050±0.008 2.0±0.4 Vegard rule.
The transport properties of our samples show p-type
conductivity due to impurity states present near valence
louin function in order to properly estimate the satura- band,withactivationenergyofabout200meV.Thecar-
tion magnetization of our samples and the role of short riertransportinoursamplesattemperatureshigherthan
range Mn-pairing. We fitted our experimental data to 150Kisthermallyactivated,whileatT <150Kadegen-
the expression29 erated transport dominates. The carrier mobility shows
the presence of ionic carrier scattering centers with con-
gµ SB centration from 6 to 15×1017cm−3. The scattering cen-
B
M =MSBS +χdiaB, (5) ters are not related to the Mn impurities.
k (T +T )
B 0 ! TheparamagneticCurie-Weissbehaviorindicatesthat
themajorityofMnionsinoursamplesareisolated(ran-
where
domly distributed in the host CdGeAs lattice). The
2
M =x¯ N µ gS. (6) estimates of the active Mn content of our samples, in
S m 0 B
case of most of them, give values similar to those ob-
HereB isthe Brillouinfunctionandx¯ is theeffective, tained with the use of the EDXRF method. It indicates
S m
activeMncontentestimatedfromthe saturationmagne- thatthe majorityofMnions substitute the Cd positions
tization values. The term χ B represents the diamag- in the CdGeAs lattice, and possess Mn2+ charge state
dia 2
netic contribution of the CdGeAs . The experimental with highmagnetic momentum. We provethat the total
2
M(B)curveswerefittedwithtwofittingparameters: M solubility of Mn in Cd Mn GeAs is around x=0.04,
S 1-x x 2
andT . Theresultsofthefittingprocedurearepresented avaluelowerthanforII-VI andratherhighwithrespect
0
inFig.5togetherwiththeexperimentaldata. Aswecan to III-V DMS bulk crystals grown under thermal equi-
see the experimental curves are well fitted with the the- librium conditions. Both II-VI and III-V DMS we may
oretical model. The fitting parameters, obtained during considerasternaryanalogsofCd Mn GeAs . Theneg-
1-x x 2
the fitting procedure, are gathered for all our samples ativeCurie-Weisstemperatures,withvaluesincreasingas
in Table III. The T temperatures have positive values. a function of x are observed indicating an antiferromag-
0
Positive T in Eq. 5 corresponds to negative θ in Eq. 2. netic exchange interaction. The average value of the ex-
0
ThevaluesofT changewiththeMncontent,x,inaway changeparameter,J/k ≈-1.8±1K,isofthesameorder
0 B
7
as those observed in II-VI DMS and in very dilute III-V S.A. Varnavskiy, I.V. Fedorchenko, and S.F. Marenkin, Solid
DMS with carrier concentration below 1019 cm−3. This State Commun. 151,870(2011).
result may be explained by assuming the superexchange 12L. Kilanski, I.V. Fedorchenko, M. G´orska, E. Dynowska,
M. Wo´jcik, B.J. Kowalski, J.R. Anderson, C.R. Rotundu,
via an anion.
S.A. Varnavskiy, W. Dobrowolski1, and S.F. Marenkin, Phys.
Stat. Sol. B 248,1601(2011).
13S.F. Marenkin and V.A. Morozova, Inorg. Mater. 35, 1190
VI. ACKNOWLEDGMENTS (1999).
14S.F.Marenkin,V.M.Novotortsev,K.K.Palkina,S.G.Mikhailov,
V.T.Kalinnikov,Inorganic Mater. + 40,93(2004).
Scientific work was financed from funds for science in 15S.F.Marenkin,W.A.Morozova,L.I.Ocertianova,W.M.Truhan,
2011-2014, under the project no. N202 166840 granted S.G. Mikhailov, A.W. Molcanov, and G.S. Juriev, Chemical
by the NationalCenter for Science of Poland. This work Technology of Inorganic Substances andMaterials 12,5(2005).
16A.Yu. Mollaev, I.K. Kamilov, R.K. Arslanov, T.R. Ar-
has been supported by the RFBR projects no. 12-03-
slanov, U.Z. Zalibekov, V.M. Novotortsev, S.F. Marenkin, and
31203 and 13-03-00125.
V.M.Trukhan,Appl. Phys. Lett.100,202403(2011).
17V. M. Novotortsev, K. K. Palkina, S. G. Mikhailov, A. V.
1J.KossutandW.Dobrowolski,Handbook ofMagneticMaterials
Molchanov, L. I. Ochertyanova, and S. F. Marenkin, Inorganic
(North-Holland,Amsterdam,1993),pp.231305.
Mater. + 41,439(2005).
2W. Dobrowolski, J. Kossut, and T. Story, Handbook of Mag-
18V.H.Pfister,ActaCrystallogr.11,221(1958).
netic Materials (Elsevier, Amsterdam, 2003), Chaps. II-VI, pp.
19P.Pyykko¨,Phys. Rev. B 85,024115 (2012)
289377.
20E. Dynowska, W. Paszkowicz, L. Kilanski, W. Dobrowolski,
3T.Story,R.R.Galazka,R.B.Frankel,andP.A.Wolff,Phys.Rev.
I.V.Fedorchenko, andS.F.Marenkin,inpreparation.
Lett.56,777(1986).
21Tula R. Paudel and Walter R. L. Lambrecht, Phys. Rev. B 78,
4H. Ohno, A. Shen, F. Matsukura, A. Oiwa, A. Endo, S. Kat-
085214(2008).
sumoto,andY.Iye,Appl. Phys. Lett.69,363(1996).
22J.S.Blakemore,J. Appl. Phys.53,R123(1982)
5T.Dietl,Nat. Mater.9,965(2010).
6S.C.ErwinandI.Zˇuti´c,Nat. Mater.3,410(2004). 23S.V. Gudenko, B.A. Aronzon, and V.A. Ivanov, JETP Letters
82,532(2005).
7S.Picozzi,Nat. Mater.3,349(2004).
24J.Spalek,A.Lewicki,Z.Tarnawski,J.K.Furdyna,R.R.Galazka,
8V. M. Novotortsev, S. F. Marenkin, S. A. Varnavskii, L. I. Ko-
andZ.Obuszko,Phys. Rev. B 33,3407(1986).
roleva, T. A. Kupriyanova, R. Szymczak, L. Kilanski, and B.
Krzymanska,Russ. J. Inorg. Chem.+ 53(1),22(2008). 25M.GorskaandJ.R.Anderson,Phys. Rev. B 38,9120(1988).
9L. Kilanski, M. G´orska, V. Domukhovski, W. Dobrowol- 26B.E.Larson,K.C.Haas,H.Ehrenreich,andA.E.Carlsson,Solid
ski, J.R. Anderson, C.R. Rotundu, S.A. Varniavskii, and State Commun. 56,347(1985).
S.F.Marenkin,Acta Phys. Pol. A114,1151(2008). 27S.vonMolnar,H.Munekata,H.Ohno,andL.L.Chang,J.Magn.
10L. Kilanski, M. G´orska, W. Dobrowolski, E. Dynowska, Magn. Mat.93,356(1991).
M. Wo´jcik, B.J. Kowalski, J.R. Anderson, C.R. Ro- 28M. Zajac, J. Gosk, M. Kaminska, A. Twardowski, T. Szyszko,
tundu, D.K. Maude, S.A. Varnavskiy, I.V. Fedorchenko, and andS.Podsiadlo,Appl. Phys. Lett.79,2432(2001).
S.F.Marenkin,J. Appl. Phys.108,073925 (2010). 29J.A.Gaj,R.Planel,andG.Fishman,SolidStateCommun.29,
11L. Kilanski, W. Dobrowolski, E. Dynowska, M. Wo´jcik, 435(1979).
B.J.Kowalski,N.Nedelko,A.S´lawska-Waniewska,D.K.Maude,