Malkin
Malkin
Malkin
Experimental and theoretical studies of the dynamic magnetization in swept magnetic fields of the
orthorhombic SrY2 O4 single-crystals doped with the Dy3+ Kramers ions (0.01 and 0.5 at.%) with
natural abundances of even and odd Dy isotopes are presented. Impurity ions substitute for Y3+ ions
at two nonequivalent crystallographic sites with the same local Cs symmetry but strongly different
crystal fields. Well pronounced double-loop hysteresis is observed at temperatures 2, 4, 5 and 6 K for
sweeping rates of 5 and 1 mT/s. The microscopic model of spectral, magnetic and kinetic properties
of Dy3+ ions is developed based on the results of EPR, site selective optical spectra and magnetic
relaxation measurements. The derived approach to the dynamic magnetization in the sweeping field
based on the numerical solution of generalized master equations with time-dependent transition
probabilities induced by the electron-phonon interaction, quantum tunneling and cross-relaxation
allowed us to reproduce successfully the evolution of the hysteresis loop shape with temperature,
sweeping rate and concentration of paramagnetic ions.
tions of paramagnetic ions accounting for the electron- the result of long spin-phonon relaxation times of the
phonon interaction, cross-relaxation and Landau-Zener- quasi-Ising-type Dy2 ions caused by the large energy gap
Stückelberg (LZS) non-adiabatic quantum tunneling [20– mentioned above.
22]. The measured low-temperature dynamic magnetiza- Magnetization measurements were carried out using
tion and magnetic field dependencies of relaxation rates the vibrating sample magnetometer (VSM) with the op-
are presented in Section II, while Section III describes tion of the PPMS-9 universal system (Quantum Design,
modeling of the observed hysteresis loops. The results of USA). Magnetization (M ) was measured as a function of
spectroscopic studies (electron paramagnetic resonance a magnetic field Bb applied along the b-axis. In order to
(EPR) and optical spectra) are presented in Supplemen- assess the irreversible nature of magnetization, measure-
tal Material [23] containing necessary references [24–29]. ments were carried out in increasing and then decreasing
The article ends with the Conclusions. field with a constant speed v = dB/dt. The equilibrium
magnetization was measured at a reduced number of field
values with setting the target field, then leaving a sam-
II. EXPERIMENTAL DETAILS AND RESULTS ple to relax (typically for 5 minutes) and then taking the
magnetization measurement.
In the present study, the magnetic and spectral char- Figure 1 shows the dynamic magnetization M (Bb )
acteristics of impurity Dy3+ ions in Sr(Y1−x Dyx )2 O4 measured on the single crystal SrY2 O4 :Dy3+ (0.01 at.%)
(x = 10−4 and x = 5 · 10−3 ) single crystals grown by the in a swept magnetic field applied along the b-axis as well
floating zone method [23, 30] were measured by means of as the equilibrium one at temperatures 2, 4, 5 and 6 K
the low temperature EPR and site-selective laser spec- (Bmax = 0.6 T, v = 5 and 1 mT/s).
troscopies. The analysis of the registered spectra based In the magnetic fields oriented along the crystallo-
on the calculations of crystal-field (CF) parameters in graphic axes, four Y1 sites, as well as four Y2 sites, in
the framework of the semi-phenomenological exchange the unit cell are magnetically equivalent. According to
charge model (ECM) [31, 32] allowed us i) to associate the results of EPR measurements [23]), the correspond-
the observed EPR signals with exact quantum transi- ing g-factors of the Dy3+ ions are gbb (Dy1) = 2.7 and
tions between the sublevels of the CF ground state dou- gbb (Dy2) = 19.28, so, the Dy2 ions provide the dominant
blets and ii) to assign spectral lines in selectively excited contribution to the measured magnetization.
optical spectra to transitions between well-defined CF All registered dependencies M (Bb ) in the swept fields
sublevels of the ground and several excited multiplets (Fig. 1) demonstrate well pronounced double-loop hys-
of the impurity Dy3+ ions at Y1 (Dy1) and Y2 (Dy2) teresis. Loop areas reduce monotonously with decreasing
lattice sites. As a result, we obtained the total sets of a sweeping rate and increasing a temperature. We note a
self-consistent single-ion spectral and magnetic parame- change of the loops shape from quasi-rectangular upturns
ters (two sets of 15 CF parameters (Table S3 in [23]), and downturns close to zero values of the applied field Bb
g-factors (Table S1 in [23]) and hyperfine coupling con- to a rounded spindle-type one between 5 and 6 K for de-
stants), the energy level patterns (Table S2 in [23]) and creasing sweeping rates in the range 5−1 mT/s. A similar
the corresponding electronic and electron-nuclear wave profile of the hysteresis loops was found in the dynamic
functions in external magnetic fields. This has served a magnetization of the SrY2 O4 :Dy single-crystal sample
base for a detailed modeling and interpretation of the with much higher dysprosium concentration (0.5 at.%,
measured dependencies of the dynamic magnetization on Fig. 2).
temperature, sweeping rate of the magnetic field and con- Relaxation rates of the magnetization in SrY2 O4 :Dy3+
centration of Dy3+ ions. single-crystals with different concentrations of impurity
It is important to underline here some specific differ- ions (0.01 and 0.5 at.%) at a given temperature T and
ences between the single-ion properties of the Dy1 and fixed magnetic field B0 along the b-axis were measured
Dy2 centers in SrY2 O4 :Dy crystals. Possessing the same using fast switching of the magnetic field from the initial
Cs point symmetry, due to differences in the location value Bin in the equilibrium state of the sample to the
of nearby oxygen ions (see Fig. S1 in [23]), these cen- B0 value (B0 < Bin or B0 > Bin ). The magnetization
ters reveal different magnetic anisotropies (of easy-axis evolution was tracked by periodic measurements (typi-
and easy-plane type at Y1 and Y2 sites, respectively) cally every second) after the target field B0 had been
and strongly different energy gaps (energy barriers in set. Each magnetization was in fact an average over the
the relaxation processes) between the first excited and integration time of the PPMS lock-in detector.
the ground CF doublets, E2 − E1 = 68 K and 300 K The measured magnetization evolution during the
for Dy1 and Dy2, respectively. The maximum principal thermalization processes at the temperatures 2 and 4 K
value of the g-tensor of the ground doublet of Dy2 ions gives evidence for the two magnetic subsystems with fast
equals g2 = 19.28 (g1 and g3 are less than 0.1) with the and slow relaxion rates and strongly different contribu-
principal direction in the (ab)-plane slightly tilted (∼ ±9◦ tions to the total magnetization. We identify these two
for magnetically nonequivalent ions) from the b-axis con- subsystems with the Dy3+ ions at Y1 and Y2 sites. In
trary to Dy1 ions with the maximum g-factor g3 = 13.6 particular, the measured time dependencies of the mag-
along the c-axis. The discovered magnetic hysteresis is netization after setting the field B0 , shown in Fig. 3 for
3
9 9
(a) (b)
8 8
7 7
6 6
5 5
T=6 K
)
3+
4 4
/Dy
T=5 K
B
3 3
mMagnetic moment (
2 2
1 1
T=4 K
0 0
-1 -1
-2 -2
-3 -3
T=2 K
-4 -4
|dB/dt| = 5 mT/s |dB/dt| = 1 mT/s
-5 -5
-0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6
FIG. 1. Dynamic magnetization of the SrY2 O4 :Dy (0.01 at.%) single-crystal in magnetic fields B k b for the sweeping rate of
(a) 5 and (b) 1 mT/s at 2, 4, 5 and 6 K. Red and black lines represent the results of measurements and modeling. respectively.
Dashed lines show the equilibrium magnetization. Magnetization curves for T = 4, 5 and 6 K are shifted upward by 3, 5 and
6.5 µB /Dy3+ , respectively; µB is the Bohr magneton.
4 h i
|dB/dt| = 5 mT/s
M (t) = M1 (B0 ) + M2 (B0 )(1 − e−t/τ ) /2, (1)
B
m
2
where τ is the relaxation time, M1 (B0 ) and M2 (B0 ) the
Magnetic moment (
4
3
10
0.3 T
/Dy)
0.25 T
(s)
B
3
m
0.2 T
t
Magnetic moment (
Relaxation time
1
0.15 T 2
10
0.1 T 2
0.07 T
1
1
10
0
0.1 0.15 0.2 0.25 0.3
0 500 1000 1500 2000 2500
Time (s)
Magnetic field B
0
(T)
FIG. 3. Registered (solid lines) relaxation to the equilibrium FIG. 4. The relaxation time dependencies on the applied
magnetization in SrY2 O4 :Dy (0.01 at.%) sample after a fast magnetic field at T = 2 K in SrY2 O4 :Dy crystals with Dy
setting of the external field B0 along the b-axis at T = 2 K; for concentrations of 0.01 at.% (1) and 0.5 at.% (2). Line (2) is
each curve, a value of B0 is shown in the figure. Red dashed a guide for an eye.
lines represent the fits of the data with single exponential slow
evolution (1) of the magnetization of Dy2 ions.
Here Mα.λ are components of the magnetic moment op-
erators of Dy3+ ions, and ρ(λ, t) is the single-ion density
sites. This fact proves the existence of an additional re- matrix of an impurity ion at the site Y1 (λ = Dy1) or Y2
laxation mechanism effective in the region of weak mag- (λ = Dy2) satisfying the generalized master equation [32]
netic fields. In the considered range of temperatures, with the time dependent relaxation terms. To model the
Orbach–type relaxation is ineffective because of a large, dynamic magnetization, we use the secular approxima-
300 K, gap between the ground and the first excited CF tion that is adapted for the considered system; the non-
sublevels of the ground multipet of the Dy3+ ions at Y2 diagonal elements of ρ(λ, t) in the basis of eigenfunctions
sites. However, the observed abrupt change of the shape of the single-ion Hamiltonian Hλ (t) are neglected. The
of the hysteresis loops at the elevated temperatures of diagonal elements ρnn (λ, t) = ρn (λ, t) which determine
5 − 6 K as compared with the loops at 2 − 4 K (see populations of single-ion electronic (or electron-nuclear)
Fig. 1) evidences for the two-phonon Raman relaxation states with energies En (λ, t) satisfy the equations of mo-
processes with the strong temperature dependence of the tion
relaxation rate.
∂ρn
Quantum tunneling and cross-relaxation processes af- =
fect the dynamic magnetization only in the relatively nar- X∂t
row range of sweeping magnetic field −0.02 T < Bb (t) < [ρk Wk→n (t) − ρn Wn→k (t) + Γkn (t)(ρk − ρn )] +
+0.02 T comparable to widths of hyperfine structures of k(k6=n)
the ground Kramers doublets of odd dysprosium isotopes X
CR
Wn←p,l←k (t)(ρp ρk − ρl ρn ), (3)
at Y2 sites.
kpl(k6=l,p6=n)
9
(a) TABLE I. Magnetic fields Bb (p, q) (in units of 10−4 T) at
Dy)
T=2K the crossing points of the hyperfine sublevels (p and q) of the
164
6
ground state doublet of 163 Dy2 ions (I) and 161 Dy2 ions (II)
/
B
3
Magnetic momernt (
9 (b)
1 0 28.2 57.6 88.1 119.8 152.6
164
T = 4 K
6 2 -28.2 0 29.3 59.9 91.6 124.4
/
B
m
-6
-9
transitions of magnetic moments affecting the magneti-
zation involve at least three energy levels) or of two dif-
-0.9 -0.6 -0.3 0.0 0.3 0.6 0.9 ferent magnetic subsystems with overlapping transitions
Magnetic field B
b
(T) in their responses to external time-dependent perturba-
tions. Thus, the CR processes affect the dynamic magne-
FIG. 5. Computed dynamic magnetization of even dyspro- tization of the two subsystems of odd isotopes of quasi-
sium isotopes at Y2 sites promoted by the electron-phonon Ising Dy2 ions with approximately equal energy gaps be-
interaction at temperatures (a) 2 K and (b) 4 K for sweeping tween crossings of different pairs of hyperfine sublevels
rates of 5 mT/s (1) and 1 mT/s (2). Dashed lines correspond of the ground doublet at some fixed magnetic fields (see
to the equilibrium magnetization. Fig. 6) and of the equilibrium and nonequilibrium sub-
systems of even isotopes at Y1 and Y2 sites, respectively,
nearby the zero value of the sweeping field. The CR ef-
tional relaxation channels, namely, the cross-relaxation fects were considered qualitatively in Refs. [43, 44].
and the quantum tunneling, which are effective at low CR
The probability Wn←p,l←k for the simultaneous tran-
magnetic fields in the region of crossing hyperfine or Zee- sitions k → l of ion 1, and p → n of ion 2 can be written
man sublevels of the ground state Kramers doublet of as follows [45, 46]:
Dy3+ ions at Y2 sites.
CR 2π
Wn←p,l←k = | < n, l|H12 |p, k > |2 δ(ωpn − ωlk ),
h̄2 C
C. Cross-relaxation (13)
where H12 is the Hamiltonian of interaction between the
Energies of hyperfine sublevels of the ground doublet ions, h̄ωpn is the difference between energies of states
of Dy3+ ions (odd isotopes) at Y2 sites versus the mag- p and n, and h. . . iC denotes a configurational averag-
netic field Bb dependencies obtained by the numerical ing over the distribution of the paramagnetic ions in the
diagonalization of the corresponding single-ion Hamil- crystal lattice.
tonian are presented in Fig. 6. Because of the strong Probabilities of CR processes depend strongly on the
Ising-type magnetic anisotropy, each hyperfine sublevel overlap of excitation energies of interacting ions. Note
can be characterized by the nuclear spin projection (Iy = also that the cross-relaxation within the subsystem of
−5/2, −3/2, . . . , 5/2) on the applied field that has a max- magnetically equivalent two-level ions (even dysprosium
imum weight in the corresponding electron-nuclear wave- isotopes) does not change the total magnetization. Thus,
function. Values of the magnetic fields Bb at the crossings when considering the dynamic magnetization of the mul-
of the hyperfine sublevels are presented in Table I. tisite SrY2 O4 :Dy3+ crystals with strongly different g-
Cross-relaxation (CR) transitions lead to the forma- tensors of impurity ions at Y1 and Y2 sites, we can limit
tion of a nonequilibrium state of an ensemble of magnet- ourselves by taking into account the CR processes sepa-
ically equivalent multi-level paramagnetic ions (flip-flop rately within the subsystems of 161 Dy2 and 163 Dy2 ions
8
of Dy1 and Dy2 ions in the magnetic field Bb (t) cross- 1/2 4
ing the zero value. Taking into account finite widths of 0.08
p
)
-1/2 =3
-1
energy levels and assuming the homogeneous distribu-
Energy (cm
0.04
-3/2 p=2
1/2 q= 3
X
CR λλ′ (λ) (λ′ ) (λ) (λ′ )
Magnetic field Bb (10
-4
T)
gαβγδ (ωpn − ωlk )kαβγδ Jα,np Jβ,lk Jγ,pn Jδ,kl , (14) 0.12
161 -5/2 p=6
αβγδ (b) Dy2
-3/2 p=5
-1/2 p=4
0.08
′ 161 161 163 163
where either λ = λ = Dy1, Dy2, Dy1, Dy2, 1/2 p=3
)
-1
3/2 p=2
0.04
Energy (cm
CR
tration of λ′ -ions per Y lattice site, gαβγδ (ω) is the 5/2 p=1
λλ′ λλ′
CR form-function, kαβγδ are the lattice sums kαβ,γδ = 0.00
5/2 q=6
P
a
s αβ (R )a
λλ ,s δγ
′ (R λλ ,s
′ ) over sites of a Bravais lattice,
-0.04
and 3/2 q=5
! 1/2 q=4
-0.08
1 xλλ′ ,sα xλλ′ ,sβ
aαβ (Rλλ′ ,s ) = 3 δαβ − 3 2 . (15) -5/2 q=1 -3/2 q=2
-1/2 q=3
-4
Here, Rλλ′ ,s is the radius-vector of the site s in the sub- Magnetic field B
b
(10 T)
SrY2 O4 doped with Kramers Dy3+ ions is the second in- The important result of our work is the demonstration
organic dilute rare-earth paramagnet that exhibits a hys- of the strong CR effects on the dynamic magnetization,
teretic behavior of the dynamic magnetization similar to in particular, the renormalization of the LZS incoher-
SMM. ent transition probabilities at anticrossing points in the
electron-nuclear manyfold of states in swept magnetic
The derived approach to the magnetization dynam- fields.
ics involved the comprehensive experimental studies of
spectroscopic, magnetic and kinetic properties of the syn-
thesized single-crystal samples with different concentra- ACKNOWLEDGMENTS
tions of Dy3+ ions, the subsequent analysis of the mea-
sured EPR and site-selective optical spectra and the This work was supported by the Russian Science
magnetic relaxation rates in the framework of the semi- Foundation (project No. 19-12-00244). BZM is grate-
phenomenological crystal field model, and numerical so- ful to O.A. Petrenko for useful discussions and to
lutions of the master equations. This approach enabled M.V. Vanyunin for help in developing a Matlab code for
us to successfully reproduce the hysteresis cycle shapes solving a system of nonlinear equations of motion. The
observed as well as their transformation with tempera- authors are grateful to M.A. Cherosov for his assistance
ture, sweeping field rate and concentration of Dy3+ ions. in magnetization measurements.
[30] B. Malkin, S. Nikitin, I. Mumdzhi, D. Zverev, [41] I. Chiorescu, W. Wernsdorfer, A. Müller, H. Bögge, and
R. Yusupov, I. Gilmutdinov, R. Batulin, B. Gabbasov, B. Barbara, Journal of magnetism and magnetic materi-
A. Kiiamov, D. Adroja, et al., Physical Review B 92, als 221, 103 (2000).
094415 (2015). [42] L. F. Chibotaru et al., Physical Review B 97, 024427
[31] B. Malkin, in Modern Problems in Condensed Matter Sci- (2018).
ences, Vol. 21, edited by A. Kaplyanskii and R. Macfar- [43] R. Giraud, A. Tkachuk, and B. Barbara, Physical review
lane (Elsevier, 1987) pp. 13–50. letters 91, 257204 (2003).
[32] I. Rousochatzakis and M. Luban, Physical Review B 72, [44] W. Wernsdorfer, S. Bhaduri, R. Tiron, D. Hendrick-
134424 (2005). son, and G. Christou, Physical review letters 89, 197201
[33] N. Bloembergen, S. Shapiro, P. Pershan, and A. JO, (2002).
Phys. Rev 114, 445 (1959). [45] S. Bertaina, B. Barbara, R. Giraud, B. Malkin,
[34] N. Prokof’ev and P. Stamp, Reports on Progress in M. Vanuynin, A. Pominov, A. Stolov, and A. Tkachuk,
Physics 63, 669 (2000). Physical Review B 74, 184421 (2006).
[35] A. Abragam and B. Bleaney, Electron paramagnetic res- [46] B. Malkin, M. Vanyunin, B. Barbara, and S. Bertaina,
onance of transition ions (Clarendon P., 1970). Journal of alloys and compounds 451, 473 (2008).
[36] G. Taran, E. Bonet, and W. Wernsdorfer, Journal of Ap- [47] D. Garanin and E. Chudnovsky, Physical Review B 56,
plied Physics 125 (2019). 11102 (1997).
[37] A. Abragam, J. Jacquinot, M. Chapellier, and M. Gold- [48] J. Villain, A. Würger, A. Fort, and A. Rettori, Journal
man, Journal of Physics C: Solid State Physics 5, 2629 de Physique I 7, 1583 (1997).
(1972). [49] M. N. Leuenberger and D. Loss, Physical Review B 61,
[38] E. Chudnovsky, D. Garanin, and R. Schilling, Physical 1286 (2000).
Review B 72, 094426 (2005). [50] N. Andriushin, S. E. Nikitin, G. Ehlers, and
[39] R. Gaillac, P. Pullumbi, and F.-X. Coudert, Journal of A. Podlesnyak, Physical Review B 106, 104427 (2022).
Physics: Condensed Matter 28, 275201 (2016). [51] G. I. Andreev, I. Romanova, S. Korableva, O. Morozov,
[40] I. Chiorescu, W. Wernsdorfer, A. Müller, H. Bögge, and A. Semakin, M. Cherosov, A. Kiiamov, and M. Tagirov,
B. Barbara, Physical Review Letters 84, 3454 (2000). Materials Research Bulletin 156, 112002 (2022).
[52] B. Barbara, Comptes Rendus Physique 6, 934 (2005).