Single-Molecule Nanomagnets

Single-Molecule Nanomagnets

arXiv:1001.4194v2 [cond-mat.mes-hall] 4 Aug 2010

Jonathan R. Friedman
Department of Physics, Amherst College, Amherst, MA 01002, USA

Myriam P. Sarachik
Department of Physics, City College of New York-CUNY, New York, NY 10031, USA

Key Words Single-molecule Magnets, Quantum Tunneling of Magnetization, Quantum Coherence, Berry Phase, Quantum Computation, Spintronics

Abstract Single molecule magnets straddle the classical and quantum mechanical worlds, displaying many fascinating phenomena. They may have important technological applications in information storage and quantum computation. We review the physical properties of two prototypical molecular nanomagnets, Mn12 -acetate and Fe8 : each behaves as a rigid, spin-10 object, and exhibits tunneling between up and down directions. As temperature is lowered, the spin reversal process evolves from thermal activation to pure quantum tunneling. At low temperatures, magnetic avalanches occur in which the magnetization of an entire sample rapidly reverses. We discuss the important role that symmetry-breaking elds play in driving tunneling and in producing Berry-phase interference. Recent experimental advances indicate that quantum coherence can be maintained on time scales sucient to allow a meaningful number of quantum computing operations to be performed. Eorts are underway to create monolayers and to address

.. 2009 and manipulate individual molecules.

1056-8700/97/0610-00

CONTENTS
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Macroscopic Quantum Tunneling of Magnetization . . . . . . . . . . . . . . . . . .

Berry Phase in Molecular Magnets . . . . . . . . . . . . . . . . . . . . . . . . . . .

12

Quantum Coherence; Quantum Computation . . . . . . . . . . . . . . . . . . . . .

15

Addressing and manipulating individual molecules . . . . . . . . . . . . . . . . . .

19

Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

21

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

22

Single-Molecule Nanomagnets

Introduction

True to its name, a single-molecule magnet (SMM) is a molecule that behaves as an individual nanomagnet. Because of their small size and precise characterizability, molecular nanomagnets exhibit many fascinating quantum phenomena, such as macroscopic quantum tunneling of magnetization and Berry-phase interference. They straddle the quantum mechanical and classical worlds, residing in a middle ground that is of abiding interest to physicists. In addition, SMMs may nd application in high-density magnetic storage or as qubits, the processing elements in quantum computers. In this article, we survey some of the remarkable phenomena exhibited by SMMs and discuss progress towards future applications. In Section 2, we review the basic structure and properties of SMMs with reference to the two prototypical molecules, Mn12 -acetate and Fe8 , and provide a brief history. In Section 3, we discuss the reversal of the magnetic moment by quantum tunneling; the crossover from classical spin reversal to pure quantum tunneling; the symmetry-breaking elds that drive tunneling; and the abrupt reversal of the magnetic moment of an entire crystalline sample in the form of a magnetic avalanche. The experimental observation of geometric-phase (Berry-phase) interference is described in Section 4. In Section 5, we discuss recent developments that show that quantum coherence can be maintained in SMMs on time scales sucient to allow a signicant number of qubit operations to be performed. These exciting results make SMMs serious contenders for use in quantum information technologies. In Section 6, we discuss recent experimental eorts to create single layers of SMMs on surfaces and to measure transport through individual molecules.

Friedman & Sarachik

Background

The spin of a SMM ranges from a few to many times that of an electron; the corresponding magnetization of the individual magnets is minuscule. The molecules readily crystallize so that a typical sample contains 1015 or more identical magnetic clusters in (nearly) identical crystalline environments. At the same time, the SMMs are relatively far apart so that the magnetic exchange between them is small and they interact only very weakly with each other. To a very good approximation, a crystalline sample thus behaves at low temperatures as an ensemble of well-characterized, identical, non-interacting nanoscale magnets. Although the symmetry, the magnitude of spin anisotropy, as well as the hyperne elds, dipolar interactions and other properties, vary substantially from one SMM to another, most exhibit the same overall behavior. The central features can be understood with reference to the prototypical SMMs Mn12 -ac and Fe8 shown in Figure 1. First synthesized by Lis in 1980 (1), Mn12 O12 (CH3 COO)16 (H2 O)4 (referred to hereafter as Mn12 -ac) received little attention until its unusually large molecular magnetic moment (2) and magnetic bistability (3) were recognized. Early measurements established a number of important features: a large S = 10 spin, rigid at low temperatures; a large negative magnetocrystalline anisotropy with a barrier U 70 K (2, 4), resulting in a characteristic relaxation time that obeys an Arrhenius law and magnetic hysteresis below a blocking temperature TB 3 K (3, 5, 6, 7, 8, 9, 10). As shown in Figure 1(a), the magnetic core of Mn12 -ac has four Mn4+ (S = 3/2) ions in a central tetrahedron surrounded by eight Mn3+ (S = 2) ions. The ions are coupled by superexchange through oxygen bridges with the net result that the

Single-Molecule Nanomagnets

four inner and eight outer ions point in opposite directions, yielding a total spin S = 10. The magnetic core is surrounded by acetate ligands, which serve to isolate each core from its neighbors and the molecules crystallize into a body-centered tetragonal lattice. While there are very weak exchange interactions between molecules, the exchange between ions within the magnetic core is very strong, resulting in a rigid spin10 object that has no internal degrees of freedom at low temperatures. As illustrated by Figure 2(a), the spins energy can be modeled as a double-well potential, where one well corresponds to the spin pointing up and the other to the spin pointing down. A strong uniaxial anisotropy barrier of the order of 70 K yields doubly degenerate ground states in zero eld. The spin has a set of levels (shown in the gure) corresponding to dierent projections, m = 10, 9, . . . , 9, 10, of the total spin along the easy axis of the molecule (corresponding to the c-axis of the crystal). The magnetic cluster [(tacn)6 Fe8 O2 (OH)12 ]Br8 (referred to as Fe8 ) is shown in Figure 1(b); here (tacn) is the organic ligand 1,4,7-triazacyclononane. Like Mn12 , Fe8 has a spin ground state S = 10, which arises from competing antiferromagnetic interactions between the eight S = 5/2 Fe spins. Modeled by a double-well potential like that of Mn12 (see Figure 2), the spin dynamics are quite similar. However, as shown by the schematic diagrams of Figure 1(c-e), while Mn12 -ac has an easy axis and an essentially isotropic hard plane, Fe8 has three inequivalent axes. The fact that Fe8 is biaxial rather than uniaxial has interesting consequences that are presented in Section 4.

Friedman & Sarachik

Macroscopic Quantum Tunneling of Magnetization

A number of enigmatic features emerged in the mid-1990s that suggested the possibility of spin reversal by quantum mechanical tunneling in Mn12 -ac. In 1995, Barbara et al. (6) observed an increase in the magnetic relaxation time (H ) when a longitudinal magnetic eld was increased from zero to 0.2 T, above which the relaxation decreased. This is counterintuitive and puzzling, since the application of a eld lowers the barrier for spin reversal so that the relaxation time is expected to decrease monotonically as a function of eld. Barbara et al. (6) suggested that the faster relaxation at zero eld could be due to the coincidence of the level schemes of the two wells. At about the same time, Novak and Sessoli (7) reported relaxation minima at H = 0 and 0.3 T; they speculated that this might be due to thermally assisted tunneling between excited states in a double-well potential. Paulsen and Park (8) reported magnetic avalanches (rapid, complete magnetization reversals) that occurred most often at a specic eld of 1 T. These experiments all found enigmatic behavior at particular values of the magnetic eld, suggesting the possibility of quantum tunneling. The observation by Friedman et al. (11) of macroscopic quantum tunneling of the magnetization in Mn12 -ac in 1996 and its conrmation shortly thereafter by Hern andez et al. (12) and Thomas et al. (13) is widely recognized as a major breakthrough in spin physics (14). A series of steps were discovered in the hysteresis loops in Mn12 -ac below the blocking temperature of 3 K; typical curves are shown in Figure 3. Figure 3(a) shows the magnetization M as a function of magnetic eld Hz applied along the easy axis; the derivative, dM/dHz , which reects the magnetic relaxation rate, is plotted as a function of Hz in Figure 3(b). As shown in Figure 2, Hz tilts the potential, causing levels in the right (left)

Single-Molecule Nanomagnets

well to move down (up). Figure 2(d) shows the eld dependence of the spins energy levels. Levels in opposite wells align at certain values of magnetic eld (dashed lines in Figure 2(d)), allowing the spin to reverse by resonant tunneling. The steps observed in the hysteresis loops at nearly equal intervals of magnetic eld are due to enhanced relaxation of the magnetization at the resonant elds when levels on opposite sides of the anisotropy barrier coincide in energy. This magnetization tunneling phenomenon has now been seen in hundreds of SMMs as well as in some high-spin rare-earth ions (15, 16). SMMs generally owe their simplicity to the fact that they behave as single, rigid spins at suciently low temperatures. The eective spin Hamiltonian can be written as:
2 4 H DSz gz B Sz Hz ASz +H,

(1)

where the rst term gives rise to the anisotropy barrier; the second term is the Zeeman energy that splits the spin-up and spin-down states in a magnetic eld, thereby lifting the degeneracy of the two potential wells; the third is the nexthighest-order term in longitudinal anisotropy; and the last term, H , contains all symmetry-breaking operators that do not commute with Sz . Note that in the absence of H , Sz is a conserved quantity and no tunneling would be allowed. For Mn12 -ac, D = 0.548K, gz = 1.94, and A = 1.173 103 K. The main features of the behavior of Mn12 -ac, a particularly simple and highly symmetric SMM, can be understood by considering only the rst and second terms in Eq. 1; the rst creates the double-well potential and the second provides the tilt of the potential in a magnetic eld. The steps observed in Figure 3 occur at resonant elds HN = N D/gz B at which levels in opposite wells (Figure 2) align. The elds at which the steps occur are consistent with the independently

Friedman & Sarachik

measured values of D and gz (17); the steps are labeled sequentially by integers starting with N = 0 for H = 0. At temperatures below 0.5 K, tunneling proceeds predominantly from the ground state of the metastable well: as shown in Figure 3, the hysteresis loops are essentially identical for T = 0.3 K and T = 0.5 K. As the temperature increases the hysteresis loops become narrower and magnetic relaxation occurs at lower magnetic elds corresponding to smaller values of N : the spins are activated to higher levels from which they can more easily tunnel across the barrier. This thermally assisted tunneling process can be described using a master-equation approach in which transitions between levels occur by tunneling or by the absorption or emission of phonons (18,19,20,21,22,23,24,25). Above the blocking temperature TB (a phenomenological parameter that depends on the time scale of the measurement), sucient thermal energy is available for the magnetization to quickly achieve equilibrium and no hysteresis is observed. This two-term Hamiltonian provides a good description of many other SMMs as well. Although they are small, additional terms in the Hamiltonian are responsible for important details that have provided many new insights. We will introduce these additional terms one by one, and discuss their consequences.
2 , the For the two-term Hamiltonian with a simple quadratic anisotropy, DSz

level crossings corresponding to each resonance occur pairwise at the same value of magnetic eld (Figure 2(d)). That is, every level in the left well simultaneously
4 term, determined by crosses a level in the right well at one value of eld. A Sz

electron spin resonance (ESR) (17) and inelastic neutron scattering (26), removes this coincidence and introduces detailed structure in each step. The magnetic eld corresponding to spin tunneling from level m of the metastable

Single-Molecule Nanomagnets

well to m in the stable well can be easily calculated for the three-term Hamilto2 g S H AS 4 , yielding: nian H = DSz z B z z z

Hm ,m =

A D(m + m) 1 + (m 2 + m2 ) , gz B D

(2)

where (m + m) is the step number N . While the rst term in the brackets gives resonant magnetic elds that are integer multiples of D/gz B independently of the pair (m , m), as illustrated in Figure 2(b and d), the small correction A/D = 2.1 103 causes dierent pairs of levels to cross at dierent values of the magnetic eld within each step, as shown in Figure 2(c and e). This eect becomes increasingly important for larger step numbers (high magnetic eld), as shown by the dashed lines in Figure 2(e). This feature provides an interesting form of spectroscopy that allows a determination of which energy levels are responsible for tunneling. As the temperature is reduced, the relaxation evolves from thermally assisted tunneling (i.e., thermal activation to a higher level from which tunneling proceeds) to tunneling from the lowest m = 10 level of the metastable well. The crossover between these regimes is shown in Figure 4. It is abrupt rather than gradual, suggesting a (discontinuous) rst-order rather than a (continuous) second-order transition (27, 28, 29). We note that a true rst-order, discontinuous transition occurs only in the limit of innite spin; the transition in Mn12 -ac, where the spin S = 10 is large but nite, is thus abrupt rather than discontinuous. In order for tunneling to occur, the Hamiltonian must include terms that do not commute with Sz , which we have collectively labeled H :
2 2 4 4 H = E (Sx Sy ) gB Hx Sx + (C/2)(S+ + S ) + ...

(3)

The rst term on the right-hand side is a second-order transverse anisotropy that

10

Friedman & Sarachik

is present in many low-symmetry SMMs; the second term includes hyperne, dipolar, and possibly other internal transverse elds as well as an externally applied transverse eld; the third term is the fourth-order transverse anisotropy. The source of tunneling in Mn12 -ac was the subject of intense debate and investigation for a number of years. In a perfect crystal, the lowest transverse
4 + anisotropy term allowed by the tetragonal symmetry of Mn12 -ac is (C/2)(S+ 4 ). This imposes a selection rule in which m can change only by integer multiS

ples of 4, m = 4i, i = 0, 1, . . ., allowing only every fourth step for ground-state tunneling. For thermally assisted tunneling, this selection rule prohibits every other step (30). By contrast, all steps are observed with no clear dierences in amplitude between them (see Figure 3). Dipolar elds and hyperne interactions, which would allow all steps on an equal footing, are known to be too weak to cause the rapid tunneling rates observed experimentally (30). Through a series of theoretical (31, 32) and experimental (33, 34, 35, 36, 37) steps, the source of tunneling has been traced to isomer disorder (38) in Mn12 -ac. Specically, variation in the hydrogen bonding of Mn12 molecules with neighboring acetic acid molecules leads to a distribution of quadratic (rhombic) transverse anisotropy. This introduces a locally varying second-order anisotropy superposed on the global tetragonal symmetry of the crystal and induces tunneling through the rst term in H , Eq. 3. The symmetry of such a term permits tunneling at all even-numbered steps. In addition, the isomer disorder produces a distribution of tilts (within 1.7 ) of the molecular easy axes with respect to the global uniaxial direction, the crystals c axis. When a eld is applied along this axis, the tilt distribution gives rise to a distribution of transverse elds that drives tunneling by virtue of the term linear in Hx (in Eq. 3) and allows all steps, both even and

Single-Molecule Nanomagnets odd.

11

We end this section with a brief description of the process of spin reversal by magnetic avalanches in molecular magnets. As rst reported by Paulsen and Park (8) in Mn12 -ac, crystals of molecular magnets often exhibit an abrupt and complete reversal of the magnetization from one direction to the other. Poorly understood until recently, these avalanches were attributed to a thermal runaway in which heat is released that further accelerates the magnetic relaxation. In addition to releasing thermal energy, molecular crystals emit bursts of radiation during magnetic avalanches (39, 40, 41). Once considered events to be avoided, as they interfere with a detailed study of the stepwise process of magnetization reversal, magnetic avalanches became the focus of attention and renewed interest stimulated by the theoretical suggestion that the radiation emitted during an avalanche is in the form of coherent (Dicke) superradiance (42). Although the issue of coherence of the radiation has yet to be resolved, recent studies have claried the nature of the avalanche process itself. In particular, local timeresolved measurements using micron-sized Hall sensors have shown that a magnetic avalanche spreads as a narrow interface that propagates through the crystal at a constant velocity that is roughly two orders of magnitude smaller than the speed of sound (43). This process, illustrated schematically in Figure 5, is closely analogous to chemical combustion - the propagation of a ame front through a ammable chemical substance, referred to as chemical deagration. Interestingly, there is clear evidence of the quantum-mechanical nature of the spin-reversal process during an avalanche. Magnetic avalanches have been studied in detail by time-resolved measurements of the local magnetization (44) and by measurements of bulk magnetization during avalanches ignited by surface acoustic

12

Friedman & Sarachik

waves (45, 46), as well as theoretically by Garanin and Chudnovsky (47). As shown in Figure 6(a), the avalanche speed is enhanced at the resonant values of magnetic eld at which tunneling occurs (44, 45, 46); Figure 6(b) shows that at the same resonant elds there are pronounced dips (44) in the temperature required to ignite an avalanche.

Berry Phase in Molecular Magnets

While we have so far restricted our discussion to the Mn12 -ac molecule, many other SMMs exhibit similar behavior. As discussed above, the Fe8 molecule resembles Mn12 : it has a spin of 10 (48), a substantial anisotropy barrier (22 K) and shows resonant tunneling steps in the hysteresis loops (49). As shown schematically in Figure 1(e), Fe8 has three inequivalent directions, providing a hard x axis and a medium y axis within the hard plane. In this case,
2 S 2 ). The presence of it is essential to retain the rst term in Eq. 3, namely, E (Sx y

this term indicates that in zero magnetic eld the spin has two preferred tunneling paths that pass through the y and y directions, respectively, as illustrated by the red and purple curves in Figure 7(a). This leads to a remarkable interference eect. Geometric (or Berry) phase is a fascinating phenomenon in both classical and quantum physics in which a system adiabatically following a closed path in some parameter space acquires a non-trivial phase change. A familiar example of a geometric phase is the Aharanov-Bohm eect in which a charged particle whose path encircles a region of magnetic ux acquires a phase proportional to the ux. When a spins orientation traverses a closed path, it acquires a geometric phase proportional to the solid angle enclosed by that path: (1 cos)d, where and

Single-Molecule Nanomagnets

13

are the polar and azimuthal angles describing the position of the spins vector on the unit sphere. A biaxial spin, like Fe8 , has two least-action tunneling paths for spin reversal. Each path acquires a dierent geometric phase and these paths will therefore interfere. In 1993, well before the discovery of tunneling in SMMs, Garg (50,51) predicted that a magnetic eld could be used to modulate the geometric-phase interference. While a eld oriented along the hard (x) axis preserves the symmetry between the two tunneling paths they both maintain the same amplitude it changes the geometric phase dierence between the paths, altering the interference. The tunnel splitting is modulated by the factor cos(S ), where is the solid angle circumscribed by the two paths, as illustrated by the shaded regions in Figure 7(a) and 7(B). As the magnitude of the eld is increased, decreases and whenever S = (2n + 1) /2 for integer n, the interference is completely destructive, causing the tunnel splitting to vanish. The predicted eld interval between zeros is

H =

2 gB

2E (E + D).

(4)

This interference eect was rst discovered experimentally in Fe8 in 1999 by Wernsdorfer and Sessoli (52). They used a Landau-Zener tunneling method, in which a longitudinal eld sweeps the system rapidly through a tunneling resonance, to determine the tunnel splitting for that resonance. The results, plotted in Figure 7(c), show clear oscillations for the N = 0 resonance when the eld is applied along the hard axis ( = 0). The tunnel splitting drops by nearly an order of magnitude at regular eld intervals of 0.41 T. The eect becomes less pronounced as the eld in the hard plane is tilted away from the hard axis. Qual-

14

Friedman & Sarachik

itative agreement with theory is obtained, and there is quantitative agreement when one includes an additional, fourth-order transverse anisotropy term in the spin Hamiltonian for Fe8 (the third term in Eq. 3), as shown in Figure 7(d). Gargs original calculation considered the case of a purely transverse eld (i.e. N =0). As Figure 7(e) illustrates, the interference was also observed at other resonances (N = 1 and 2). This unexpected nding prompted much theoretical work to understand the observations (55, 54, 53). Another interesting feature of the data is illustrated in Figure 7(e), which shows that the odd resonance (N = 1) is out of phase with the even resonances (N = 0, 2). That is, when the tunnel splitting of an even resonance has a maximum, the tunnel splitting of the odd resonance has a minimum and vice-versa. This so-called parity eect is a manifestation of a selection rule imposed by the secondorder transverse anisotropy for Fe8 , namely, that odd resonances are forbidden for an integer spin. The fact that the tunneling does not go to zero exactly is likely due to dipolar interactions between spins that produce local transverse elds, an eect that has been demonstrated experimentally (56). Since the work of Wernsdorfer and Sessoli, the geometric-phase eect has been observed in other SMMs. Most of these have an eective Hamiltonian like that of Fe8 with a single hard-axis direction (57, 58). Geometric-phase interference has recently been observed in a Mn12 variant (Mn12 -tBuAc) that has a fourth-order transverse anisotropy of the form
C 4 2 (S+ 4 ), in which the system has four + S

hard-axis directions (x and y see Figure1(d)) (59), as had been predicted theoretically (60, 61). Another interference eect has been reported in systems that behave as exchange-coupled dimers of SMMs (62,63), where it is the eective exchange interaction that is modulated by an applied eld, but a full theoretical

Single-Molecule Nanomagnets

15

understanding of the eect has not yet been found. Geometric-phase interference has recently been observed in an antiferromagnetic SMM (64), a manifestation of interference between N eel-vector tunneling paths (65). There has also been recent theoretical work that predicts that uniaxial stress applied the along the hard axis of a four-fold symmetric SMM, like Mn12 -tBuAc, will produce a geometric-phase eect in the absence of a magnetic eld (66).

Quantum Coherence; Quantum Computation

Of the many potential applications for SMMs, one of the most interesting is the possibility that they could be used as qubits, the processing elements of quantum computers. Quantum computers exploit uniquely quantum properties like superpositions of states and entanglement. They can, in principle, solve certain problems, like factoring large numbers, more eciently than could be done with any known algorithm for a classical computer (67). The basic processing elements of quantum computers are qubits, which like classical bits, can be put into logical |0 > and |1 > states. Unlike classical bits, however, they can also be put into superposition states, i.e. a|0 > + b|1 >. Qubits can also be entangled with one another where the state of an individual qubit is ill-dened. Many physical systems have been proposed as possible qubits, from microscopic systems like trapped ions (68) and nuclear spins (67) to macroscopic ones like quantum dots (69) and superconducting devices (70). In order to build a practical quantum computer, there are two broad criteria that must be fullled: 1) The coherence of superposition and entangled states must be maintained for periods of time long enough to complete a calculation without appreciable errors, and 2) the qubits must be individually controlled and manipulated within

16

Friedman & Sarachik

a large-scale architecture. Microscopic systems more readily fulll the rst criterion, as we routinely describe their behavior using quantum mechanics. However, it is extremely challenging to manipulate individual atomic-sized objects and to integrate them into an architecture of myriads of qubits. Macroscopic systems easily fulll the second criterion since we can fabricate many qubits on a chip and address them with individual wires, but, concomitantly, their quantum mechanical behavior is easily destroyed through their stronger interactions with the environment, a process generically described as decoherence. SMMs as qubits may oer the best of both worlds. With magnetic moments an order of magnitude larger than the moment of an electron, they may be easier to manipulate than atomic-sized objects. Yet their quantum behavior may mirror atomic-scale objects more than macroscopic ones. SMMs have many important advantages as potential qubits. Many properties (e.g. barrier height, tunneling rate, interaction with environmental degrees of freedom) can be chemically engineered. Magnetic elds can be used to tune the barrier height and, in particular, the tunnel splittings. Moreover, microwave elds can be used to manipulate the quantum state of the spin and create superposition states. Interest in using SMMs in quantum computing was galvanized by theoretical work by Leuenberger and Loss in 2001 (71) that showed how Grovers quantum search algorithm could be implemented within a single, high-spin SMM using an elaborate superposition state (but no entanglement). Realizing that proposal is still well beyond current technology but activity in the eld remains high. We now turn to a discussion of eorts to measure and exploit coherent quantum phenomena in SMMs.

Single-Molecule Nanomagnets

17

Several groups have been studying the eects of microwave radiation on SMMs with the intent of observing coherent phenomena, such as Rabi oscillations or spin-echo eects. Early work showed photon-assisted tunneling eects and measured excited-state lifetimes (72, 73, 74, 75, 76, 77, 78, 24, 79, 80). A major advance was made in 2007 by Ardavan et al. (81), who observed spin-echo in Cr7 Ni and Cr7 Mn SMMs (as above, these names are shorthand for more complicated chemical formulas that, importantly, include many hydrogen ions). These are variants of Cr8 wheels, which are antiferromagnetic with S=0. Substituting Ni or Mn for one Cr (as well as a compensating cation) results in a SMM with S=1/2 or 1, re2 = I , the identity operator) spectively. The former has no zero-eld anisotropy (z

and the orientation of the molecules with respect to the applied eld is therefore irrelevant, except for a small anisotropy of the g factor. This allowed Ardavan et al. to dilute the molecules in a solvent to the point that intermolecular dipole interactions were negligible. Using standard ESR techniques, they observed spin echos and measured excited-state lifetimes, T1 , as high as 1 ms and dephasing times, T2 , of 100 - 1000 ns at low temperatures, as shown in Figure 8. Interestingly, nearly identical results were obtained for the S=1 Cr7 Mn SMM (red open squares), despite the addition of anisotropy. A substantial component of the decoherence in these systems derives from hyperne elds, as illustrated by the blue, lled circles in Figure 8, which shows T2 values for a deuterated sample of Cr7 Ni. The reduction in hyperne elds by the substitution of deuterium for hydrogen raises the coherence time to 3 s at low temperature. SMMs based on polyoxometalates have been suggested as qubit candidates because they lack nuclear spins (82). Antiferromagnetic rings remain interesting low-spin systems and potential qubits.

18

Friedman & Sarachik

Recent experiments have shown that pairs of Cr7 Ni wheels can be exchangecoupled (83) and there has been a theoretical proposal for turning a linear chain of rings into a scalable quantum processor (84). Coherent phenomena have now been observed in a few other SMMs: V15 (85), Fe4 (86) and Fe8 (87). In the rst two, this was achieved by diluting the molecules in solvent, a method similar to that used by Ardavan et al. For Fe8 , Takahashi et al. (87) used a large magnetic eld and low temperatures to polarize the molecules within a single crystal so that each molecule experiences nearly the same local dipolar eld. Also of note is recent work on Mn2+ ions diluted in a MgO nonmagnetic matrix (88). These single ions have a relatively large spin (S=5/2) and show coherent phenomena similar to SMMs. The qubit state of an SMM need not be the spin state. SMMs that lack inversion symmetry (89) can have denite chirality states. These states can be used as eective qubit states that can couple to electric elds through spin-orbit eects. Experimental evidence for chirality states was recently reported for a SMM triangle of three Dy ions (90). Entanglement between qubits is essential for making a universal quantum computer. Entanglement may exist within the molecules, but if each SMM is to act as a single qubit, one is interested in entanglement between SMMs. Passive entanglement (in which the energy eigenstates happen to be entangled states) has been observed in dimers of SMMs through magnetization (91), spectroscopy (92) and specic heat measurements (83). As of this writing, controlled entanglement the creation of a well dened entangled state on demand through a controlled gate operation has not yet been achieved. There have been several theoretical proposals for creating entanglement between SMM qubits: using radiation

Single-Molecule Nanomagnets

19

pulses (84), by injection of a linker spin from an STM tip (93), or by coupling to rf electric elds in a cavity (89). While these methods have promise for testing fundamental physics and proof-of-principle demonstrations of entanglement, they are probably unworkable for a large-scale quantum computer.

Addressing and manipulating individual molecules

If SMMs are to be useful qubits, they need to be individually addressed and controlled. Other possible applications, such as using SMMs for magnetic memory or in spintronic devices, similarly require individual addressability. Eorts to achieve this are multidisciplinary, lying at the interface of the elds of supramolecular chemistry, molecular electronics, spintronics and quantum control. In this section, we briey review experimental eorts to create single layers of SMMs on surfaces that can be individually addressed and to measure individual SMMs through transport techniques. Since large anisotropy lies at the heart of most SMM behavior, it is essential that when depositing the molecules on a surface, this property be preserved and that the molecules have a well-controlled orientation. In addition, as evidenced by the eects of solvent disorder discussed above, it is also important that any symmetry breaking interactions be minimized. Given these constraints, it is not surprising that progress in depositing both chemically and magnetically intact SMMs on surfaces has been slow. Progress in this area has been outlined in recent review articles (94, 95, 96); we highlight a few salient points here and refer the reader to those reviews for more details. Many eorts have focused on the deposition of Mn12 -ac and its variants. Techniques include vapor deposition and chemical self-assembly. While some of these

20

Friedman & Sarachik

techniques result in monolayers or thin lms of molecules, there is evidence that in many cases the molecules are not all intact, e.g. small fragments are found by STM imaging and Mn2+ ions are detected while the unperturbed compound contains only Mn3+ and Mn4+ ions. Moreover, many of the key magnetic characteristics, such as hysteresis and tunneling, are absent in these monolayers. Some progress has recently been made by Voss et al. who reported deposition of largely intact Mn12 molecules using a ligand exchange process in which both the Au surface and the molecules are prefunctionalized (97). Using STM, they found a bandgap for the molecules consistent with that predicted by ab initio calculations (98). There has been progress in attaching other SMMs to surfaces, such as Cr7 Ni rings (99) and Fe4 tetrahedrons (100, 101). The latter has proved to be very promising: it is the only SMM to remain magnetically intact, showing both hysteresis and evidence of tunneling after grafting to the surface, as indicated by X-ray magnetic circular dichroism and muon spin-rotation techniques. These techniques indicate that Fe4 molecules on surfaces exhibit slow relaxation at low temperatures, although with a lower barrier than that found in bulk crystals (100, 101). In another notable recent result, the Fe6 -POM molecule has recently been bonded to the surface of single-wall carbon nanotubes and magnetic measurements indicate that the molecules exhibit hysteresis and magnetization tunneling (102). There have been a few attempts to incorporate an individual SMM in a transistorlike device in which the molecule is attached to two leads and the current-voltage characteristic is measured (103,104,105). An example is shown in Figure 9, which shows the dierential conductance of a Mn12 molecule attached to two Au leads

Single-Molecule Nanomagnets

21

as a function of bias voltage and an external gate voltage capacitively coupled to the molecule (103). Complex behavior is found, including regions of negative dierential conductance (see Figure 9(b)). The results have been interpreted in terms of the high-spin states of Mn12 with the interaction of electrons added to the molecule during conduction. It should be noted that neither in this nor similar studies by other groups has hysteresis been found in the transport properties as a function of applied magnetic eld. This may be due to damage to the molecules similar to that found when they are deposited on surfaces. It may also be the result of strong coupling to the leads that distorts the molecule or otherwise breaks its symmetry. The experimental results on transport through individual SMMs have stimulated a great deal of theoretical work (106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 120). These include proposals for observing a novel form of the Kondo eect (107, 113, 114, 115), a geometric-phase modulation of transport (109, 113, 114, 120), spin ltering (116) by the SMM, and current-induced switching of the SMM spin (111, 117). Observing any of these eects may be challenging in light of the diculty in making electrical contacts to SMMs without signicantly perturbing them. However, recent results are encouraging (100, 101, 102) and some groups are working on other, less invasive techniques for measuring the properties of individual SMMs (121).

Summary

From an unheralded beginning in 1980, when the rst molecular magnet was synthesized, activity in the eld of SMMs has grown rapidly and now involves

22

Friedman & Sarachik

an unusually broad range of disciplines including physics, chemistry, material science, nanoscience and nanotechnology, spintronics, and quantum information. As new techniques (such as spin polarized STM, scanning magnetometry, magnetic circular dichroism) become available, they are being brought to bear to probe fundamental questions and to investigate the potential of SMMs for various applications. Work is proceeding along many dierent fronts. Chemists have expended a great deal of eort in a quest to nd SMMs with larger anisotropy barriers to enable operation at higher temperatures; a new barrier-height record of 86.4 K has recently been set in a S = 12 [Mn6 ] complex (122), toppling the record held by Mn12 ac and its variants. Much progress has been made depositing molecules on surfaces and measuring transport through individual SMMs with the goal of using them for both quantum and classical information processing and storage. A particularly attractive feature of SMMs is that they are amenable to engineering by chemical and self-assembly techniques to control and design desirable properties of the molecules and the interactions between them. These characteristics include spin, barrier height, hyperne elds, dipole and exchange interactions and spin-phonon coupling. In the space allotted to this review, it is impossible to cover all the interesting aspects of the eld. For further information, the reader is referred to references (123, 124, 125, 126).

Acknowledgments

JRF acknowledges the support of the National Science Foundation under grant no. DMR-0449516; MPS acknowledges the support of the National Science Foundation under grant DMR-0451605

Single-Molecule Nanomagnets Literature Cited 1. Lis T. 1980. Acta Cryst. B 36:2042-2046

23

2. Caneschi A, Gatteschi D, Sessoli R, Barra AL, Brunel LC, Guillot M. 1991. J. Am. Chem. Soc. 113:5873-5874 3. Sessoli R, Gatteschi D, Caneschi A, Novak MA. 1993. Nature (London) 365:141-143 4. Sessoli R, Tsai H-L, Schake AR, Wang S, Vincent JB, Folting K, Gatteschi D, Christou G, Hendrickson DN. 1993. J. Am. Chem. Soc. 115:1804-1816 5. Sessoli R. 1995. Mol. Cryst. Liq. Cryst. 273A:145-157 6. Barbara B, Wernsdorfer W, Sampaio LC, Park J-G, Paulsen C, Novak MA, Ferre R, Mailly D, Sessoli R, Caneschi A, Hasselbach K, Benoit A, Thomas L. 1995. J. Magn. Magn. Mater. 140-144:1825-1828 7. Novak MA, Sessoli R. 1995. In Quantum Tunneling of MagnetizationQTM94, ed. Gunther L, Barbara B, pp. 171-188. Dortrecht: Kluwer 8. Paulsen C, Park J-G. 1995. In Quantum Tunneling of MagnetizationQTM94, ed. Gunther L, Barbara B, pp. 189-207. Dortrecht: Kluwer 9. Novak MA, Sessoli R, Caneschi A, Gatteschi D. 1995. J. Magn. Magn. Mater. 146:211-213 10. Paulsen C, Park J-G, Barbara B, Sessoli R, Caneschi A. 1995. J. Magn. Magn. Mater. 140-144:379-380 11. Friedman JR, Sarachik MP, Tejada J, Ziolo R. 1996. Phys. Rev. Lett. 76:38303833 12. Hern andez JM, Zhang XX, Luis F, Bartolom e J, Tejada J, Ziolo R. 1996. Europhys. Lett. 35:301-6 13. Thomas L, Lionti F, Ballou R, Gatteschi D, Sessoli R, Barabara B. 1996.

24 Nature 383:145-147 14. Ziemelis K. 2008. Nat. Phys. 4:S19

Friedman & Sarachik

15. Barbara B, Giraud R, Wernsdorfer W, Mailly D, Lejay P, Tkachuk A, Suzuki H. 2004. J. Magn. Magn. Mat. 272-276:1024 16. Ishikawa N, Sugita M, Wernsdorfer W. 2005. Angew. Chem. 117:2991-5 17. Barra A-L, Gatteschi D, Sessoli R. 1997, Phys. Rev. B 56:8192-8198 18. Villain J, Hartmann-Boutron F, Sessoli R, Rettori A. 1994. Europhys. Lett. 27:159-64 19. Garanin DA, Chudnovsky EM. 1997. Phys. Rev. B 56:11102-18 20. Fort A, Rettori A, Villain J, Gatteschi D, Sessoli R. 1998. Phys. Rev. Lett. 80:612-5 21. Luis F, Bartolom J, Fernndez JF. 1998. Phys. Rev. B 57:505-13 22. Leuenberger MN, Loss D. 1999. Europhys. Lett. 46:692-8 23. Chudnovsky EM. 2004. Phys. Rev. Lett. 92:120405 24. Bal M, Friedman JR, Chen W, Tuominen MT, Beedle CC, Rumberger EM, Hendrickson DN. 2008. EPL 82:17005 25. Garanin DA. 2008. arXiv:0805.0391 26. Mirebeau I, Hennion M, Casalta H, Andres H, Gudel HU, Irodova AV, Caneschi A. 1999. Phys. Rev. Lett. 83:628-631 27. Kent AD, Zhong YC, Bokacheva L, Ruiz D, Hendrickson DN, Sarachik MP. 2000. Europhys. Lett. 49:521-527 28. Bokacheva L, Kent AD, Walters MA. 2000. Phys. Rev. Lett. 85:4803-4806 29. Mertes KM, Zhong Y, Sarachik MP, Paltiel Y, Shtrikman H, Zeldov E, Rumberger EM, Hendrickson DN, Christou G. 2001. Europhys. Lett. 55:874-879 30. Hern andez JM, Zhang XX, Luis F, Tejada J, Friedman JR, Sarachik MP,

Single-Molecule Nanomagnets Ziolo R. 1997. Phys. Rev. B 55: 5858-5865

25

31. Chudnovsky EM, Garanin DA. 2001. Phys. Rev. Lett. 87:187203; 2002. Phys. Rev. B 65:094423 32. Park K, Baruah T, Bernstein N, Pederson MR. 2004. Phys. Rev. B 69:144426 33. Mertes KM, Suzuki Y, Sarachik MP, Paltiel Y, Shtrikman H, Zeldov E, Rumberger EM, Hendrickson DN, Christou G. 2001. Phys. Rev. Lett. 87:227205 34. del Barco E, Kent AD, Rumberger EM, Hendrickson DN, Christou G. 2003. Phys. Rev. Lett. 91:047203 35. del Barco E, Kent AD, Hill S, North JM, Dalal NS, Rumberger EM, Hendrickson DN, Chakov N, Christou G. 2005. J. Low Temp. Phys. 119-174 36. Hill H, Edwards RS, Jones SI, Dalal NS, North JM. 2003. Phys. Rev. Lett. 90:217204 37. Takahashi S, Edwards RS, North JM, Hill S, Dalal NS. 2004. Phys. Rev. B 70:094429 38. Cornia A, Sessoli R, Sorace L, Gatteschi D, Barra AL, Daiguebonne C. 2002. Phys. Rev. Lett. 89:257201 39. Tejada J, Chudnovsky EM, Hern andez JM, Amig o R. 2004. Appl. Phys. Lett. 84:2373-2375 40. Hern andez-M nguez A, Jordi M, Amig o R, Garc a-Santiago A, Hernandez JM, Tejada J. 2005. Europhys. Lett. 69:270-276 41. Shar O, Keren A. 2009. Phys. Rev. B 79:180404 42. Chudnovsky EM, Garanin DA. 2002. Phys. Rev. Lett. 89:157201 43. Suzuki Y, Sarachik MP, Chudnovsky EM, McHugh S, Gonzalez-Rubio R, Avraham N, Myasoedov Y, Zeldov E, Shtrikman H, Chakov NE, Christou G. 2005. Phys. Rev. Lett. 95:147201

26

Friedman & Sarachik

44. McHugh S, Jaafar R, Sarachik MP, Myasoedov Y, Finkler A, Shtrikman H, Zeldov E, Bagai R, Christou, G. 2007. Phys. Rev. B 76:172410 45. Hern andez-M nguez A, Hernandez JM, Maci` aa F, Garcia-Santiago A, Tejada J, Santos PV. 2005. Phys. Rev. Lett. 95:217205 46. Hern andez-M nguez A, Maci` a F, Hernandez JM, Tejada J, Santos PV. 2008. J. Magn. Magn. Mater. 320:1457-1463 47. Garanin DA, Chudnovsky EM. 2007. Phys. Rev. B 76:054410 48. Barra A-L, Debrunner P, Gatteschi D, Schulz CE, Sessoli R. 1996. Europhys. Lett. 35:133-8 49. Sangregorio C, Ohm T, Paulsen C, Sessoli R, Gatteschi D. 1997. Phys. Rev. Lett. 78:4645-8 50. Garg A. 1993. Europhys. Lett. 22:205-210 51. Garg A. 1995. Phys. Rev. B 51:15161-15169 52. Wernsdorfer W, Sessoli R. 1999. Science 284:133-135 53. Garg A. 2003. In Exploring the Quantum/Classical Frontier: Recent Advances in Macroscopic Quantum Phenomena, ed. JR Friedman, S Han, pp. 219-49. Hauppauge, NY: Nova Science. And references therein. 54. Chen Z-D. 2003. Europhys. Lett. 62:726 55. Bruno P. 2006. Phys. Rev. Lett. 96:117208 56. Wernsdorfer W, Sessoli R, Caneschi A, Gatteschi D, Cornia A, Mailly D. 2000. J. Appl. Phys. 87:5481-6 57. Wernsdorfer W, Chakov NE, Christou G. 2005. Phys. Rev. Lett. 95:037203 58. Wernsdorfer W, Soler M, Christou G, Hendrickson DN. 2002. J. Appl. Phys. 91:7164-6 59. da Silva Neto EH, Friedman JR, Lampropoulos C, Christou G, Avraham N,

Single-Molecule Nanomagnets Myasoedov Y, Zeldov E. to be published 60. Kim G-H. 2002. J. Appl. Phys. 91:3289-93 61. Park C-S, Garg A. 2002. Phys. Rev. B 65:064411

27

62. Ramsey CM, del Barco E, Hill S, Shah SJ, Beedle CC, Hendrickson DN. 2008. Nat. Phys. 4:277-81 63. Wernsdorfer W, Stamatatos TC, Christou G. 2008. Phys. Rev. Lett. 101:237204 64. Waldmann O, Stamatatos TC, Christou G, Gudel HU, Sheikin I, Mutka H. 2009. Phys. Rev. Lett. 102:157202 65. Chiolero A, Loss D. 1998. Phys. Rev. Lett. 80:169-172 66. Foss-Feig MS, Friedman JR. 2009. Europhys. Lett. 86:27002 67. Nielsen MA, Chuang IL. 2000. Quantum Computation and Quantum Information. Cambridge: Cambridge University Press. 676 pp. 68. Blatt R, Wineland D. 2008. Nature 453:1008-1015 69. Loss D, DiVincenzo DP. 1998. Phys. Rev. A 57:120-126 70. Clarke J, Wilhelm FK. 2008. Nature 453:1031-1042 71. Leuenberger MN, Loss D. 2001. Nature 410:789-793 72. Sorace L, Wernsdorfer W, Thirion C, Barra AL, Pacchioni M, Mailly D, Barbara B. 2003. Phys. Rev. B 68:220407 73. del Barco E, Kent AD, Yang EC, Hendrickson DN. 2004. Phys. Rev. Lett. 93:157202 74. Bal B, Friedman JR, Suzuki Y, Mertes KM, Rumberger EM, Hendrickson DN, Myasoedov Y, Shtrikman H, Avraham N, Zeldov E. 2004. Phys. Rev. B 70:100408(R) 75. Petukhov K, Wernsdorfer W, Barra AL, Mosser V. 2005. Phys. Rev. B

28 72:052401

Friedman & Sarachik

76. Bal M, Friedman JR, Rumberger EM, Shah S, Hendrickson DN, Avraham N, Myasoedov Y, Shtrikman H, Zeldov E. 2006. J. Appl. Phys. 99:08D103 77. Petukhov K, Bahr S, Wernsdorfer W, Barra AL, Mosser V. 2007. Phys. Rev. B 75:064408 78. Bahr S, Petukhov K, Mosser V, Wernsdorfer W. 2007. Phys. Rev. Lett.99:147205 79. Bahr S, Petukhov K, Mosser V, Wernsdorfer W. 2008. Phys. Rev. B 77:064404 80. de Loubens G, Garanin DA, Beedle CC, Hendrickson DN, Kent AD. 2008. Europhys. Lett. 83:37006 81. Ardavan A, Rival O, Morton JJL, Blundell SJ, Tyryshkin AM, Timco GA, Winpenny REP. 2007. Phys. Rev. Lett. 98:057201 82. Stamp PCE, Gaita-Arino A. 2009. J. Mater. Chem. 19:1718-1730 83. Timco GA, Carretta S, Troiani F, Tuna F, Pritchard RJ, Muryn CA, McInnes EJL, Ghirri Candini AA, Santini P, Amoretti G, Aronte M, Winpenny REP. 2009. Nat. Nanotechnol. 4:173-178 84. Troiani F, Aronte M, Carretta S, Santini P, Amoretti G. 2005. Phys. Rev. Lett. 94:190501 85. Bertaina S, Gambarelli S, Mitra T, Tsukerblat B, Muller A, Barbara B. 2008. Nature 453:203-206 86. Schlegel C, van Slageren J, Manoli M, Brechin EK, Dressel M. 2008. Phys. Rev. Lett. 101:147203 87. Takahashi S, van Tol J, Beedle CC, Hendrickson DN, Brunel LC, Sherwin MS. 2009. Phys. Rev. Lett. 102:087603 88. Bertaina S, Chen L, Groll N, van Tol J, Dalal NS, Chiorescu I. 2009. Phys.

Single-Molecule Nanomagnets Rev. Lett. 102:050501

29

89. Trif M, Troiani F, Stepanenko D, Loss D. 2008. Phys. Rev. Lett. 101:217201 90. Luzon J, Bernot K, Hewitt IJ, Anson CE, Powell AK, Sessoli R. 2008. Phys. Rev. Lett. 100:247205 91. Wernsdorfer W, Aliaga-Alcade N, Hendrickson DN, Christou G. 2002. Nature 416:406-409 92. Hill S, Edwards RS, Aliaga-Alcalde N, Christou G. 2003. Science 302:10151018 93. Lehmann J, Gaita-Arino A, Coronado E, Loss D. 2007. Nat. Nanotechnol. 2:312-317 94. G omez-Segura J, Veciana J, Ruiz-Molina D. 2007. Chemical

Communications:3699-707 95. Voss S, Burgert M, Fonin M, Groth U, Rudiger U. 2008. Dalton Trans.:499505 96. Moroni R, Buzio R, Chincarini A, Valbusa, d. Mongeot FB, Bogani L, Caneschi A, Sessoli R, Cavigli L, Gurioli M. 2008. J. Mater. Chem. 18:109-115 97. Voss S, Fonin M, Rudiger U, Burgert M, Groth U. 2007. Appl. Phys. Lett. 90:133104 98. Boukhvalov DW, Al-Saqer M, Kurmaev EZ, Moewes A, Galakhov VR, et al. 2007. Phys. Rev. B 75:014419 99. Corradini V, Moro F, Biagi R, De Renzi V, del Pennino U, Bellini V, Carretta S, Santini P, Milway VA, Timco G, Winpenny REP, M. Aronte M. 2009. Phys. Rev. B 79:144419 100. Mannini M, Pineider F, Sainctavit P, Danieli C, E. Otero E, Sciancalepore C, Talarico AM, Arrio M-A, Cornia A, Gatteschi D, Sessoli R. 2009. Nat.

30 Mater. 8:194-197

Friedman & Sarachik

101. Mannini M, Pineider F, Sainctavit P, Joly L, Fraile-Rodr guez A, Arrio MA, Moulin CCD, Wernsdorfer W, Cornia A, Gatteschi D, Sessoli R. 2009. Adv. Mater. 21:167-173 102. Giusti A, Charron G, Mazerat S, Compain J-D, Mialane P, Dolbecq A, Riviere E, Wernsdorfer W, Ngo Biboum R, Keita B, Nadjo L, Filoramo A, Bourgoin J-P, Mallah T. 2009. Angew. Chem. Int. Ed. 48:4949-4952 103. Heersche HB, de Groot Z, Folk JA, van der Zant HSJ, Romeike G, Wegewijs MR, Zobbi L, Barreca D, Tondello E, Cornia A. 2006. Phys. Rev. Lett. 96:206801 104. Jo M-H, Grose JE, Baheti K, Deshmukh MM, Sokol JJ, Rumberger EM, Hendrickson DN, Long JR, H. Park H, D. C. Ralph DC. 2006. Nano Lett. 6:2014-2020 105. Henderson JJ, Ramsey CM, del Barco E, Mishra A, Christou G. 2007. J. Appl. Phys. 101:09E102 106. Leuenberger MN, Mucciolo ER. 2006. Phys. Rev. Lett. 97:126601 107. Romeike C, Wegewijs MR, Hofstetter W, Schoeller H. 2006. Phys. Rev. Lett. 97:206601 108. Barraza-Lopez S, Avery MC, Park K. 2007. Phys. Rev. B 76:224413 109. Gonz alez G, Leuenberger MN. 2007. Phys. Rev. Lett. 98:256804 110. Lehmann J, Loss D. 2007. Phys. Rev. Lett. 98:117203 111. Misiorny M, Barna s J. 2007. Phys. Rev. B 76:054448; Misiorny M, Barna s J. 2007. Phys. Rev. B 75:134425; Misiorny M, Barna s J. 2007. Europhys. Lett. 78:27003; Misiorny M, Barna s J. 2008. Phys. Rev. B 77:172414 112. Romeike C, Wegewijs MR, Ruben M, Wenzel W, Schoeller H. 2007. Phys.

Single-Molecule Nanomagnets Rev. B 75:064404

31

113. Wegewijs MR, Romeike C, Schoeller H, Hofstetter W. 2007. New Journal of Physics 9:344 114. Gonz alez G, Leuenberger MN, Mucciolo ER. 2008. Phys. Rev. B 78:054445 115. Roosen D, Wegewijs MR, Hofstetter W. 2008. Phys. Rev. Lett. 100:087201 116. Barraza-Lopez S, Park K, Garc a-Su arez V, Ferrer J. 2009. Phys. Rev. Lett. 102:246801 117. Lu H-Z, Zhou B, Shen S-Q. 2009. Phys. Rev. B 79:174419 118. Misiorny M, Weymann I, Barna s J. 2009. Phys. Rev. B 79:224420 119. Pemmaraju CD, Rungger I, Sanvito S. 2009. Phys. Rev. B 80:104422 120. Chang B, Liang JQ. 2009. Eur. Phys. J. B 69:515-522 121. Bogani L, Wernsdorfer W. 2006. Nat. Mater. 7:179-186 122. Milios CJ, Vinslava A, Wernsdorfer W, Moggach S, Parsons S, et al. 2007. J. Am. Chem. Soc. 129:2754-5 123. Friedman JR. 2003. In Exploring the Quantum/Classical Frontier: Recent Advances in Macroscopic Quantum Phenomena, ed. JR Friedman, S Han, pp. 219-49. Hauppauge, NY: Nova Science 124. Bagai R, Christou G. 2009. Chem. Soc. Rev. 38:1011:1026 125. Gatteschi D, Sessoli R. 2003. Angew. Chem. Int. Ed. 42: 268-297 126. Gatteschi D, Sessoli R, Villain J. 2006. Molecular Nanomagnets. Oxford: Oxford University Press

32

Friedman & Sarachik

Figure 1: (a) Chemical structure of the core of the Mn12 molecule. The four inner spin-down Mn3+ ions (green) each have spin S = 3/2; the eight outer (red) spinup Mn4+ ions each have spin S = 2, yielding a net spin S = 10 for the magnetic cluster; black dots are O bridges and arrows denote spin. Acetate ligands and water molecules have been removed for clarity. (b) Structure of Fe8 . Brown atoms represent Fe, with arrows denoting spin; red circles are O, mauve circles are N, and gray circles are C. Br, H, and ligands are not shown. (c) Spherical polar plots of energy as a function of orientation for a classical spin with uniaxial anisotropy; (d) same as (c) with additional fourth-order transverse anisotropy. (e) Spherical polar plot of energy for a spin with biaxial anisotropy.

Single-Molecule Nanomagnets

33

Figure 2: (a) Double-well potential in the absence of magnetic eld showing spin-up and spin-down levels separated by the anisotropy barrier. Dierent spin projection states |m > are indicated for S = 10. The arrows denote quantum tunneling. (b) Double-well potential for the N=4 step when a magnetic eld is applied along the easy axis. (c) Double-well potential with a small fourth-order
4 ; due to the presence of this term, dierent pairs of term in the Hamiltonian, Sz

energy levels coincide at dierent magnetic elds. (d) Energy of spin projection states |m > versus applied magnetic eld derived from the rst two terms of the Hamiltonian, Eq. 1. The vertical dashed lines indicate that all level pairs cross simultaneously in this simple approximation. (e) Energy levels as a function of magnetic eld when a fourth-order term is included. As denoted by the dashed lines, dierent level pairs are in resonance for dierent magnetic elds within a given step N .

34

Friedman & Sarachik

Figure 3: (a) Magnetization of a Mn12 -ac crystal normalized by its saturation value as a function of magnetic eld applied along the uniaxial c-axis direction at dierent temperatures below the blocking temperature; the magnetic eld was swept at 10 mT/s. (b) The derivative, dM/dHz of the data in part (a) as a function of magnetic eld.

Single-Molecule Nanomagnets

35

Figure 4: (a) dM/dHz as a function of magnetic eld and temperature for tunneling step N = 7. As shown in Figures 2c and 2e, dierent level pairs (m , m) within a given step are in resonance at dierent magnetic elds, so that enhanced relaxation due to quantum tunneling shifts to dierent pairs as the eld and temperature vary. Note the abrupt transfer of weight from thermally-assisted tunneling (left-hand peak) to pure quantum tunneling from the lowest state of the metastable potential well (right-hand peak). (b) The abrupt transfer of weight shown for several steps, as indicated.

36

Friedman & Sarachik

Figure 5: Upper panel: Schematic diagram of magnetic eld lines as spins reverse direction during a magnetic avalanche traveling from top to bottom of a Mn12 crystal. Lower panel: The local magnetization measured as a function of time by an array of micron-sized Hall sensors placed along the surface of the sample. Each peak corresponds to the bunching of the magnetic eld lines as the deagration front travels past a given Hall sensor. The propagation speed for this avalanche is 10 m/s, approximately two orders of magnitude below the speed of sound.

Single-Molecule Nanomagnets

37

Figure 6: (a) The speed of propagation of the magnetic avalanche deagration front is plotted as a function of the eld at which the avalanche is triggered; note the enhancement of propagation velocity at magnetic elds corresponding to quantum tunneling (denoted by vertical dotted lines). From Ref. (45); reproduced with permission. (b) The temperature above which an avalanche is ignited displays clear minima at the same resonant elds (vertical dashed lines).

38

Friedman & Sarachik

Figure 7: (a) Bloch sphere for spin tunneling in zero eld. The blue arrows indicate the ground-state directions of the spin (corresponding to minima of the potential wells in Figure 2(a)). The two least-action (instanton) tunneling paths are indicated by the red and purple curves. The geometric phase that produces the interference is proportional to the solid angle subtended by the surface spanning the two paths (shaded hemisphere). (b) When a eld H is applied along the hard x axis, the two ground-state orientations are tilted toward that axis, altering the tunneling paths (red and purple), the solid angle between them (shaded region), and the interference between the paths. (c) Measured tunnel splittings for Fe8 when the eld is applied along the hard axis ( = 0) and in several other directions in the hard plane, as labeled. (d) Calculated tunnel splitting for the

Single-Molecule Nanomagnets

39

10

A T

10

Spin relaxation times (ns)

10

10

10 4 10

B T

10

10

10

Temperature (K)
Figure 8: Measured values of (a) the spin relaxation time T1 and (b) the spin dephasing time T2 as a function of temperature T for Cr7 Ni (blue open circles) and Cr7 Mn (red squares). The blue lled circles show T2 for a deuterated sample of Cr7 Ni. From Ref. (81); reproduced with permission.

40

Friedman & Sarachik

10 Vb (mV) 0

(a)

-10 -1 20 Vb (mV) 0 -0.5 Vg (V) 0 I (pA) -5 0 0.5

(b)

(c)

-20 0 0.4 Vg (V)

-10 -10

0 10 Vb (mV)

Figure 9: (a) Grayscale plot of the dierential conductance through an individual Mn12 -ac molecule as a function of bias voltage Vb and gate voltage Vg . The arrow indicates a region of complete current suppression. (b) Same as (a) over a dierent range of Vb and Vg . The arrow indicates a 14 meV excitation characteristic of Mn12 molecules used as an indicator of successful incorporation of the SMM into the transistor. (c) I Vb curves for a xed value of Vg (vertical white line in (a)) shows a region of negative dierential conductance. From Ref. (103); reproduced with permission.