Inelastic Neutron Scattering Study of Mn_12-Acetate
Extracted source text ยท arXiv cond-mat/9809133
\documentstyle[preprint,aps]{revtex}
\begin{document}
\draft
\title{Inelastic Neutron Scattering Study of Mn$_{12}$-Acetate}
\author{Yicheng Zhong and M. P. Sarachik}
\address{Physics Department, City College of the City University of
New York, New York, New York 10031}
\author{Jonathan R. Friedman}
\address{Department of Physics and Astronomy, The State University of New York
at Stony Brook, Stony Brook, NY 11794}
\author{R. A. Robinson, T. M. Kelley, H. Nakotte\cite{hn} and A. C. Christianson}
\address{Manuel Lujan Jr. Neutron Science Center, Los Alamos National
Laboratory, Los Alamos, NM 87545}
\author{F. Trouw}
\address{Intense Pulsed Neutron Source, Argonne National Laboratory, Argonne,
IL 60439}
\author{S. M. J. Aubin and D. N. Hendrickson}
\address{Department of Chemistry and Biochemistry, University of California
at San Diego, La Jolla, CA 92093}
\date{\today}
\maketitle
\begin{abstract}
We report zero-field inelastic neutron scattering experiments on a
deuterated powder sample of Mn$_{12}$-Acetate consisting of a large number of
nominally identical spin-$10$ magnetic clusters. Our resolution enables us
to see a series of peaks corresponding to transitions between the
anisotropy levels within the spin-$10$ manifold. A fit to the spin Hamiltonian
$H = -D S_z^2-\mu_B{\bf B}\cdot {\bf g}\cdot {\bf S}-AS_z^4+C(S_+^4+S_-^4)$ yields an anisotropy
constant $D =(0.54\pm 0.02)$ K and a fourth-order diagonal anisotropy coefficient
$A=(1.2\pm 0.1)\times 10^{-3}$ K (the other terms being negligible).
Performed in the absence of a magnetic field,
our experiments do not involve the $g$-values as fitting parameters, thereby
yielding particularly reliable values of $D$ and $A$.
\end{abstract}
\pacs{PACS numbers: 78.70.N, 29.30.H, 36.40.Cg, 75.45.+j}
\noindent {\bf INTRODUCTION}
High-spin molecular magnets provide a unique laboratory for the study of
Quantum Tunneling of Magnetization (QTM). To date, the most intensively
studied system of this type is Mn$_{12}$-Acetate,
[Mn$_{12}$O$_{12}$(CH$_3$COO)$_{16}$(H$_2$O)$_4$]$\cdot 2$CH$_3$COOH$\cdot4$H$_2$O
(hereafter referred to as Mn$_{12}$-Ac).
First synthesized by Lis\cite{lis}, it consists of Avogadro's number of
weakly interacting\cite{weakint}, chemically identical Mn$_{12}$-Ac
molecules residing on a body-centered tetragonal
lattice. The magnetic core of each molecule contains four Mn$^{4+}$ ($S=3/2$)
and eight Mn$^{3+}$ ($S=2$) ions which form an $S=10$ ground state at low
temperatures\cite{spin10}. A strong magnetocrystalline anisotropy results
in a double-well potential with each molecule's $(2S+1)=21$ states yielding two
degenerate ground states $m=\pm 10$, and a set of doubly degenerate excited
states $m=\pm 9$,$\pm 8$....(except for $m=0$)\cite{spin10,double}. Below the
blocking temperature of $\approx 3$ K, a remarkable series of steps were found
in the hysteresis loops of oriented-powder samples at regular intervals of
magnetic field, steps which were interpreted as a manifestation of
QTM\cite{friedman}. Experimental confirmation of these steps was
provided soon thereafter in studies of single crystals\cite{thomas}.
Other experimental evidence supports this interpretation\cite{fominaya,luis1}
but there is no general agreement on the mechanism responsible for the QTM in
Mn$_{12}$-Ac. Up to fourth-order terms, the spin Hamiltonian of the
system can be written as:
\begin{equation}
H= -DS_z^2-\mu_B{\bf B}\cdot {\bf g}\cdot {\bf S}-AS_z^4+C(S_+^4+S_-^4)=H_0+H'
\label{equation}
\end{equation}
where $D$ is the anisotropy constant, the second term represents the Zeeman
energy, and the remaining are higher-order terms in the crystalline anisotropy.
$H_0= -DS_z^2-g_{para}\mu_B B_zS_z-AS_z^4$ includes all terms that commute
with $S_z$ and do not give rise to tunneling;
$H'= -g_{perp}\mu_BB_xS_x+ C(S_+^4+S_-^4)$ represents symmetry-breaking
terms that could
give rise to tunneling, associated with a transverse magnetic field and
transverse fourth-order anisotropy terms. Major efforts
are currently underway to determine the relative importance of magnetic fields
and crystalline anisotropy in
accounting for the relaxation rates observed in
Mn$_{12}$-Ac\cite{garanin,luis2,fort}. Accurate, reliable experimental
determinations of the spin Hamiltonian,
Eq. (1), thus provide crucial information.
EPR measurements performed recently in Mn$_{12}$-Ac have yielded two
different sets of values for the coefficients $D$ and $A$ of Eq. (1). Barra
{\it et al.}\cite{barra} measured high-field EPR spectra at frequencies ranging from
$150$ to $525$ GHz in magnetic fields up to $25$ T on a polycrystalline powder
sample, yielding $g_\parallel =(1.93\pm 0.01)$, $g_\perp=(1.96\pm 0.01)$,
$D =(0.56\pm 0.04)$ K and $A=(1.1\pm 0.1)\times 10^{-3}$ K. Using
high-sensitivity EPR techniques in the frequency range between $35$ and $115$
GHz, Hill {\it et al.}\cite{hill}
studied a submillimeter single crystal; their results imply
$D =0.59$ K and $A=0.88\times10^{-3}$ K\cite{hillcalc} with $g_\parallel$
ranging from 1.97 to 2.08 and $g_\perp=1.9$.
EPR measurements are normally done in a magnetic field and the $g$-values,
generally unknown, are treated as (additional) fitting parameters.
In contrast, neutron scattering experiments are normally
performed in the absence of external magnetic fields, and yield a more
direct determination of the coefficients $A$
and $D$. An inelastic neutron
scattering study by Hennion {\it et al.}\cite{hennion} of
partially deuterated Mn$_{12}$-Ac found
a well-defined peak around $0.3$ THz ($1.24$ meV) which was attributed to
excitations from $m=\pm 10$ to $m=\pm 9$. The peak broadens
on its low energy side as the temperature increases, but these
authors were unable to resolve any detailed structure.
In the present study, we have performed zero-field inelastic neutron
scattering experiments on fully deuterated Mn$_{12}$-Ac. The excitation
spectra were measured
with relatively uniform and high resolution at finite neutron
energy transfer up to $20$ meV, covering the excitation energies
of the spin-10 manifold of Mn$_{12}$-Ac. Since $g$-factors do not
enter the problem in the absence of a magnetic field, this method
allows a more accurate determination of the spin Hamiltonian.
\noindent {\bf EXPERIMENTS AND RESULTS}
A 14-gram deuterated Mn$_{12}$-Ac powder sample was prepared for
the inelastic neutron scattering experiments. The sample was
characterized following the method of
ref. 5 and steps at the same values of magnetic field were
seen in its hysteresis loops. We used the PHAROS chopper
spectrometer\cite{robinson} at the LANSCE spallation neutron source at Los
Alamos, covering energy transfers between $0$ and $20$ meV with
resolutions of $0.4$ and $0.8$ meV FWHM (full-width-at-half-maximum)
at two different incident energies (12 and 20 meV), and temperatures between
$1.4$ and $77$ K. We also used
QENS\cite{bradley}, an inverse geometry crystal analyzer spectrometer at
the Intense Pulsed Neutron Source at Argonne National Laboratory with final
neutron energy of $3.63$ meV at five different temperatures ranging from
$1.4$ K up to $30$ K. Its energy resolution is $\approx 100 \mu$eV FWHM.
Data taken at temperatures of 1.4 K, 10 K, 17 K and 30 K are shown in Fig. 1.
The large maximum centered about zero energy is due to elastic scattering. At
$1.4$ K, a single sharp peak is observed at
$1.24$ meV; we attribute this to excitations from spin
states $m=\pm 10$ to $m=\pm 9$. We note that at $1.4$ K, the overwhelming
majority of spins
are in the ground states $m=\pm 10$. As the temperature is raised and some
of the spins are thermally activated to higher energy states, new peaks
develop on the low energy side of the $1.24$ meV peak; we attribute these to
transitions from $m=\pm 9$ to $\pm 8$, $\pm 8$ to $\pm 7$, etc. Transitions
such as those between $m=\pm 9$ and $\pm 7$ are forbidden by neutron scattering
selection rules, $\Delta S=0,\pm 1, \Delta m=0,\pm 1$. Due to the increased
population of higher energy levels at higher
temperature, peaks also appear that are symmetrically placed with respect to $E=0$
on the neutron energy-gain side. No maxima appear above $1.24$ meV up
to $\approx 3$ meV, where further excitations occur that are possibly associated
with transitions between different spin manifolds\cite{hennion}; this
confirms that the peak at $1.24$ meV corresponds to transitions between the
ground and first excited states of the spin-10 manifold. The maxima are
labelled by the index $m$, which denotes the level from which
each excitation occurs; thus, the $1.24$ meV peak is
labeled $10$, the adjacent peak, $9$, and so on. As shown below, the positions
of these peaks contain key information regarding
the spin Hamiltonian of Mn$_{12}$-Ac.
Since there is no externally applied magnetic field in our
experiments, and the Zeeman energy due to the internal magnetic
field of Mn$_{12}$-Ac (estimated to be several hundred Oe\cite{hartmann})
is at least two orders of magnitude smaller than the
anisotropy energy, the term $-\mu_B{\bf B}\cdot {\bf g}\cdot {\bf S}$ in spin Hamiltonian (1) can
be safely neglected. Furthermore, the fourth-order transverse
anisotropy term $C(S_+^4+S_-^4)$ has little effect on the Eigen-energies
of the states with large $|m|$. The energy of the states
probed in our experiments near the bottom of the anisotropy wells can
thus be approximated by $E_m=-Dm^2-Am^4$, and the energy of
excitation from levels $\pm m$ to $\pm (m-1)$ will be
\begin{equation}
\Delta E_m = E_{m-1} - E_m = D(2m-1) + A [m^4-(m-1)^4].
\label{equation}
\end{equation}
In Fig. 2, six excitation energies are plotted as a function
of the index $m$. The deviation from linear dependence clearly
indicates the importance of including a diagonal fourth-order
term. A two-parameter fit to Eq. (2) gives
$D=(4.67\pm 0.18)\times 10^{-2}$ meV=$(0.54\pm 0.02)$ K and
$A=(1.04\pm 0.10)\times 10^{-4}$ meV=$(1.2\pm 0.1)\times 10^{-3}$ K. These values
are very close to the EPR
results obtained by Barra {\it et al.}: $D=(0.56\pm 0.04)$ K,
$A=(1.1\pm 0.1)\times 10^{-3}$ K.
\noindent {\bf DISCUSSION}
For our values of $D$ and $A$,
the full height of the anisotropy barrier
(defined as the energy difference between $m=0$ and $m=10$) is
calculated to be $(66\pm 3)$ K. The ratio $|Am^4/Dm^2|\approx 0.2$ for $m=10$. Since
$A$ and $D$ have the same sign, the level spacings near the bottom
of the anisotropy wells are relatively sparser,and the distribution of
levels near the top of the barrier denser. Due to the presence of a
fourth-order term, the energy levels
will not come into resonance simutaneously for a given field
applied along the anisotropy axis. Since two levels of
different quantum numbers $m$ and $m'$ are degenerate when
$H=-(m+m')[D+A(m^2+ m'^2)]/(g_{para}\mu_B)$, all pairs of states $m'= -m$
come into resonance simultaneously only in zero magnetic field.
Moreover, the spacing between steps in the hysteresis loops will not
be constant. Detailed comparison of the magnetic fields at which
maxima in the relaxation rate occur with the calculated level crossings using the
parameters given above can, in principle, allow a determination of the
specific levels near the top of the barrier that participate in the tunneling.
In summary, we have used zero-field inelastic neutron scattering
to probe the excitation spectrum of Mn$_{12}$-Ac. Our resolution enables us to
observe a series of peaks within the range from $0$ to $1.24$ meV;
we attribute these peaks to transitions within the $S=10$ manifold.
A two-parameter fit yields
values for the anisotropy constant $D$ and the coefficient of the fourth-order
diagonal anisotropy $A$ that are inconsistent with those deduced from EPR
experiments of Hill {\it et al.} and agree well with results of Barra
{\it et al.}\cite{barra}.
Work was supported at City College by NSF grant DMR-9704309 and at the
University of California, San Diego by NSF grant DMR-9729339. Work at Los
Alamos and Argonne National Laboratories was supported by the U.S.
Department of Energy, Office of Basic Energy
Sciences, under contracts W-7405-ENG-36 and W-31-109-ENG-38, respectively.
\begin{references}
\bibitem[*]{hn} Permanent address: Physics Department, New Mexico State
University, Las Cruces, NM 88003.
\bibitem{lis}T. Lis, Acta Cryst. B {\bf 36}, 2042 (1980).
\bibitem{weakint}R. Sessoli, D. Gatteschi, A. Caneschi, and M. A. Novak,
Nature {\bf 365}, 141 (1993).
\bibitem{spin10}A. Caneschi, D. Gatteschi, R. Sessoli, A. L. Barra, L. C. Brunel, and
M. Guillot, J. Am. Chem. Soc. {\bf 113}, 5873 (1991).
\bibitem{double}R. Sessoli, H.-L. Tsai, A. R. Schake, S. Wang, J. B. Vincent,
K. Folting, D. Gatteschi, G. Christou, and D. N. Hendrickson, J. Am. Chem.
Soc. {\bf 115}, 1804 (1993).
\bibitem{friedman}J. R. Friedman, M. P. Sarachik, J. Tejada, and R. Ziolo, Phys. Rev.
Lett. {\bf 76}, 3830 (1996).
\bibitem{thomas}L. Thomas, F. Lionti, R. Ballou, D. Gatteschi, R. Sessoli,
and B. Barbara, Nature {\bf 383}, 145 (1996).
\bibitem{fominaya}F. Fominaya, J. Villain, P. Gandit, J. Chaussy, and A. Caneschi,
Phys. Rev. Lett. {\bf 79}, 1126 (1997).
\bibitem{luis1}F. Luis, J. Bartolome, J. F. Fernandez, J. Tejada, J. M. Hernandez,
X. X. Zhang, and R. Ziolo, Phys. Rev. B. {\bf 55}, 11448 (1997).
\bibitem{garanin}D. A. Garanin and E. M. Chudnovsky, Phys. Rev. B {\bf 56}, 11102 (1997).
\bibitem{luis2}F. Luis, J. Bartolome, and J. F. Fernandez, Phys. Rev. B {\bf 57}, 505 (1998).
\bibitem{fort}A. Fort, A. Rettori, J. Villain, D. Gatteschi, and R. Sessoli,
Phys. Rev. Lett {\bf 80}, 612(1998).
\bibitem{barra}A. L. Barra, D. Gatteschi, and R. Sessoli, Phys. Rev. B {\bf 56}, 8192 (1997).
\bibitem{hill}S. Hill, J. A. A. J. Perenboom, N. S. Dalal, T. Hathaway, T. Stalcup, and
J. S. Brooks, Phys. Rev. Lett. {\bf 80}, 2453 (1998).
\bibitem{hillcalc}S. Hill {\it et al.}\cite{hill} considered fourth order terms of the form
$[D_{4\parallel}\times S_z^4 + D_{4\perp}\times (S_x^4+S_y^4)]$ and set
$D_{4\parallel}= D_{4\perp}$.
\bibitem{hennion}M. Hennion, L. Pardi, I. Mirebeau, E. Suard, R. Sessoli, and A. Caneschi, Phys.
Rev.B {\bf 56}, 8819 (1997).
\bibitem{robinson}R. A. Robinson, M. Nutter, R. L. Ricketts, E. Larson, J. P. Sandoval,
P. Lysaght, and B. J. Oliver, in proceedings of ICANS-XII, Abingdon, U.K. 24-28 May 1993, Rutherford Appleton Laboratory, Report 94-025, Vol. I, pp. 44-51.
\bibitem{bradley}K. F. Bradley, S-H Chen, T. O. Brun, R. Kleb, W. A. Loomis, and J. M.
Newsam, Nucl. Instrum. Meth. A {\bf 270}, 78 (1988).
\bibitem{hartmann}F. Hartmann-Boutron, P. Politi, and J. Villain, Int. J. Mod.
Phys. {\bf 10}, 2577 (1996).
\bibitem{hernandez}J. M. Hernandez, X. X. Zhang, F. Luis, J. Tejada, J. R. Friedman,
M. P. Sarachik, and R. Ziolo, Phys. Rev. B {\bf 55}, 5858 (1997).
\end{references}
\begin{figure}
\caption{Neutron scattered intensity versus energy at temperatures of
1.4, 10, 17 and 30 K, taken on the QENS spectrometer at Argonne National Laboratory.}
\label{fig1}
\end{figure}
\begin{figure}
\caption{Energies of the peaks of Fig. 1 plotted as a function of index $m$
($m$ denotes initial states for energy loss and final states for energy gain).}
\label{fig2}
\end{figure}
\end{document}