Abstract
Molecular spintroinic device based on a singlemolecule magnet is one of the ultimate goals of semiconductor nanofabrication technologies. It is thus necessary to understand the electron transport properties of a singlemolecule magnet junction. Here we study the negative differential conductance and superPoissonian shot noise properties of electron transport through a singlemolecule magnet weakly coupled to two electrodes with either one or both of them being ferromagnetic. We predict that the negative differential conductance and superPoissonian shot noise, which can be tuned by a gate voltage, depend sensitively on the spin polarization of the source and drain electrodes. In particular, the shot noise in the negative differential conductance region can be enhanced or decreased originating from the different formation mechanisms of negative differential conductance. The effective competition between fast and slow transport channels is responsible for the observed negative differential conductance and superPoissonian shot noise. In addition, we further discuss the skewness and kurtosis properties of transport current in the superPoissonian shot noise regions. Our findings suggest a tunable negative differential conductance molecular device and the predicted properties of highorder current cumulants are very interesting for a better understanding of electron transport through singlemolecule magnet junctions.
Introduction
Electronic transport through a singlemolecule magnet (SMM) has been intensively studied both experimentally^{1,2,3,4,5,6,7,8,9,10,11} and theoretically^{12,13,14,15,16,17,18,19,20,21,22,23,24,25,26} due to its applications in molecular spintronics^{27}, but these investigations were focused mainly on the differential conductance or average current. Although the shot noise of electron transport through a SMM has not yet been observed experimentally, new techniques based on carbon nanotubes have been proposed for its possible realization^{28}. Recently, the current noise properties of electron transport through a SMM have been attracting much theoretical research interests^{29,30,31,32,33,34,35,36} due to they can provide a deeper insight into the nature of transport mechanisms that cannot be obtained by measuring the differential conductance or average current^{37,38}. For example, the superPoissonian shot noise can be used to reveal the information about the internal level structure of the SMM, the leftright asymmetry of the SMMelectrode coupling^{32,33} and the angle between the applied magnetic field and the SMM's easy axis^{34}; and distinguish the two types of different nonequilibrium dynamics mechanisms, namely, the quantum tunneling of magnetization process and the thermally excited spin relaxation^{35}. In particular, the frequencyresolved shot noise spectrum of artificial SMM, e.g., a CdTe quantum dot doped with a single S = 5/2 Mn spin, can allow one to separately extract the hole and Mn spin relaxation times via the Dicke effect^{36}.
Among these observed or predicted characteristics, the negative differential conductance (NDC) is especially concerned due to the SMM's potential applications in a new generation of moleculebased memory devices and logic circuits. On the other hand, the shot noise is usually the subPoissonian statistics in noninteracting fermion systems originating from the Pauli exclusion principle. Thus, the superPoissonian shot noise is another important characteristic of transport current. Here, the socalled Fano factor, which is used to characterize the shot noise and defined as the ratio of zerofrequency shot noise and the full Poisson noise, is smaller than one for subPoissonian shot noise and exceeds one for superPoissonian shot noise. According to the definition of Fano factor, the superPoissonian shot noise, namely, the superPoissonian distribution of electron counts has a width that is broader than its mean, whereas for a Poissonian distribution the width and the mean have the same value. For the SMM weakly coupled to two normal metal electrodes, the NDC formation mechanism originates essentially from the nonequilibrium electron occupation of the system eigenstates entering bias voltage window^{14,32,39}, namely, the increased current magnitudes of the new opened transport channels do not compensate the decreased current magnitude(s) of the already opened transport channel(s) and the shot noise in this NDC region is obviously enhanced even up to a superPoissonian shot noise value. In particular, the occurrence of superPoissonian shot noise depends on the effective competition between different transport channels, thus, the SMM's internal level structure and the leftright asymmetry of the SMMelectrode coupling, which can tune the SMM transport channels, have an important influence on the superPoissonian shot noise properties^{32,33,34}. Whereas for the SMM weakly coupled to two electrodes with either one or both of them being ferromagnetic, the spin polarization of the source and drain electrodes play an important role in the forming speed of the correlated SMM eigenstates involved in the electron tunneling processes and thus have a remarkable influence on the transport channels entering bias voltage window^{21,31,40}. Consequently, the spin polarization of the source and drain electrodes will have an significant impact on the NDC and superPoissonian shot noise properties of this SMM system. However, the influences of the spin polarization of the source and drain electrodes on the NDC and superPoissonian shot noise in the SMM system have not yet been revealed.
The goal of this report is thus to study the influences of the spin polarization of the source and drain electrodes and the applied gate voltage on the NDC and superPoissonian shot noise in a SMM weakly coupled to two electrodes with either one or both of them being ferromagnetic and discuss the underlying mechanisms of the observed NDC and superPoissonian shot noise. It is demonstrated that the gatevoltagecontrolled NDC and superPoissonian shot noise depend sensitively on the spin polarization of the source and drain electrodes. In particular, whether the shot noise in the NDC region being enhanced or not is associated with the formation mechanism of the NDC. Moreover, the skewness and kurtosis in the superPoissonian shot noise regions show the crossovers from a large positive (negative) to a large negative (positive) values, which also depend on the spin polarization of the source and drain electrodes. These observed characteristics are very interesting for a better understanding of electron transport through singlemolecule magnet junctions and will allow for experimental tests in the near future.
Results
Singlemolecule magnet junction
The SMM junction consists of a SMM weakly coupled to two electrodes, see Fig. 1. The SMM is characterized by the lowest unoccupied nondegenerate molecular orbital (LUMO), the phenomenological giant spin and the uniaxial anisotropy. The SMM Hamiltonian is thus described by
Here, the first two terms depict the LUMO, and U are respectively the electron number operator and the Coulomb repulsion between two electrons in the LUMO, with being the electron creation (annihilation) operators with spin σ and energy ε_{d} (which can be tuned by a gate voltage V_{g}) in the LUMO. The third term describes the exchange coupling between the conduction electron spin in the LUMO and the SMM spin , with being the vector of Pauli matrices. The forth term stands for the anisotropy energy of the SMM whose easyaxis is Zaxis (K_{2} > 0). The last term denotes Zeeman splitting. For simplicity, we assume an external magnetic field is applied along the easy axis of the SMM.
The relaxation in the two electrodes is assumed to be sufficiently fast so that their electron distributions can be described by equilibrium Fermi functions. The two electrodes are thus modeled as noninteracting Fermi gases and the corresponding Hamiltonians read
where is the electron creation (annihilation) operators with energy ε_{α}_{kσ}, momentum k and spin s in α (α = L, R) electrode and the index s = + (−) denotes the majority (minority) spin states with the density of states . The electrode polarization is characterized by the orientation of the polarization vector p_{α} and its magnitude is defined as . Here, the polarization vectors p_{L} (left electrode) and p_{R} (right electrode) are parallel to the spin quantization Z axis and spinup ↑ and spindown ↓ are respectively defined to be the majority spin and minority spin of the ferromagnet. The tunneling between the SMM and the two electrodes are thus described by
Here, for the ferromagnetic electrode case, the electronic tunneling rates depend on the conductionelectron spin, namely, and , where the tunneling amplitudes t_{α} and the density of the state are assumed to be independent of wave vector and energy and ; while for the normalmetal electrode case, p_{α} = 0, thus, .
In addition, we assume that the bias voltage is symmetrically entirely dropped at the SMMelectrode tunnel junctions, i.e., μ_{L} = −μ_{R} = V_{b}/2, which implies that the levels of the SMM are independent of the applied bias voltage and choose meV as the unit of energy. In the Coulomb blockade regime, the occurrence or absence of superPoissonian shot noise is related to the sequential tunneling gap that being the energy difference between the ground state of charge N and the first excited state of charge N − 1 and the “vertical” energy gap between the ground state of charge N and the first excited state of the same charge^{41}. In the present work, we only study the electron transport above the sequential tunneling threshold, namely, . In this bias voltage region, the conduction electrons have sufficient energy to overcome the Coulomb blockade and tunnel sequentially through the SMM. It should be noted that the transport current in the Coulomb blockade regime is exponentially suppressed and the cotunneling tunneling process is dominant in the electron transport, thus, the normalized shot noise will deviate from the present results when taking cotunneling into account. In order to better discuss the occurrence mechanisms of the NDC and the superPoissonian shot noise, the spin of the SMM (e.g., the cyanidebridged trinuclear SMM with an S = 2 ground states^{42}) is chosen as S = 2. The other parameters of the SMM are chosen as: ε_{d} = 0.2, U = 0.1, J = 0.1, K_{2} = 0.04, B = 0.08, Γ_{L} = Γ_{R} = Γ = 0.002 and k_{B}T = 0.02.
We first study numerically the effects of the spin polarization of the two electrodes and the applied gate voltage on the NDC and superPoissonian shot noise in the three different electrode configurations (see Fig. 1), namely, (i) the ferromagnetic lead (Source)  SMM  normalmetal lead (Drain) (i.e., the FSMMN system), (ii) the normalmetal lead (Source)  SMM  ferromagnetic lead (Drain) (i.e., the NSMMF system), (iii) the ferromagnetic lead (Source)  SMM  ferromagnetic lead (Drain) (i.e., the FSMMF system).
The ferromagnetic lead (Source)  SMM  normalmetal lead (Drain)
For the FSMMN system considered here, the conduction electron will tunnel into the SMM from the ferromagnetic lead and then tunnel out of the SMM onto the normalmetal lead. The strengths of tunneling coupling of the SMM with two electrodes can be expressed as , and . Since only the energy eigenvalues of singlyoccupied and doublyoccupied eigenstates and depend on the gate voltage V_{g}, the transition between the singly and doublyoccupied eigenstates, or between the empty and singlyoccupied eigenstates first entering bias voltage window can be tuned by the gate voltage^{14}. For example, for a relatively small or negative gate voltage, the transition from the singly to emptyoccupied eigenstates first takes place; while for a large enough gate voltage that from the double to singlyoccupied eigenstates first occurs.
Figures 2(a) and 2(b), 2(e) and 2(f) show the average current and shot noise as a function of the bias voltage for different gate voltages V_{g} with p_{L} = 0.3 and p_{L} = 0.9. For a large enough spin polarization of source electrode p_{L}, the superPoissonian shot noise is observed when the transition from the doubly and singlyoccupied eigenstates first participates in the electron transport with the bias voltage increasing, see the short dashed, short dashdotted and thick dashed lines in Fig. 2(f), whereas for the QD system the superPoissonian noise dose not appear^{43}. This characteristic can be understood in terms of the effective competition between fast and slow transport channels^{32,33,34,41,44,45,46,47,48,49,50} and the forming speed of the new correlated eigenstates^{49}. The current magnitudes of the SMM transport channels can be expressed as^{14,32,34}
where C_{n − 1,m ± 1/2〉,n,m〉} = 〈n − 1, m ± 1/2d_{σ} n, m〉^{2} is a constant which related to the two SMM eigenstates but independent of the applied bias voltage and P_{n,m〉} is the occupation probability of the SMM eigenstate n, m〉. Here, the Fermi function changes very slowly with increasing bias voltage above the sequential tunneling threshold, namely, , thus, . The current magnitude of the SMM transport channel is thus mainly determined by the occupation probability P_{n,m〉} and .
In order to give a qualitative explanation for the underlying mechanism of the observed superPoissonian shot noise, we plot the occupation probabilities of the SMM eigenstates as a function of bias voltage for p_{L} = 0.9 and V_{g} = 0.6 in Fig. 3. With increasing bias voltage, the transport channel begins to participate in the electron transport. When the bias voltage increases up to about 0.6 meV, the new transport channel enters the bias voltage window. In this situation, the conduction electron can tunnel out SMM via the two transport channels and . For the FSMMN system, the electron tunneling between the SMM and the drain electrode (normalmetal lead) is independent of the conduction electron spin, thus the tunneling process mainly relies on the forming speed of the new doublyoccupied eigenstate 2, 2〉. In the case of , a new doublyoccupied eigenstate 2, 2〉 can be quickly formed when the spinup electron tunnels out of the SMM; whereas for the case of the spindown electron tunneling out of the SMM, the forming of the corresponding new doublyoccupied eigenstate 2, 2〉 takes a relatively longer time. Thus, for a large enough p_{L}, the fast transport channel can be modulated by the correlated slow channel , which leads to the bunching effect of the conduction electrons being formed and is responsible for the formation of the superPoissonian noise. When V_{b} > 0.9 meV, the transport channels and enter the bias volatge window, so that the two successive electron tunneling processes and can be formed. Consequently, the formed active competition between the fastandslow transport channels is suppressed even destroyed with the current magnitudes of the two new transport channels increasing, which leads to the superPoissonian shot noise being decreased and even to the subPoissonian.
The normalmetal lead (Source)  SMM  ferromagnetic lead (Drain)
In the NSMMF system, the strengths of tunnel coupling between the SMM and the two electrodes are described by , , . It is demonstrated that the NDC is observed for a small enough or negative gate voltage and a relatively large spin polarization of drain electrode p_{R}, see the solid and dashed lines in Figs. 4(a) and 4(e), especially for a large enough spin polarization p_{R} a strong NDC takes place, see the solid and dashed lines in Fig. 4(e). Moreover, the shot noise can be significantly enhanced and reaches up to a superPoissonian value when the magnitude of the total current begins to decrease, but the superPoissonian value in the NDC region is then decreased with further increasing the bias voltage, see the solid and dashed lines in Figs. 4(b) and 4(f). While for a large enough gate voltage, the peaks of superPoissonian shot noise are observed for a relatively large spin polarization p_{R}, see the short dashdotted and thick dashed lines in Figs. 4(b) and 4(f). The observed NDC and superPoissonian shot noise characteristics can also be attributed to the mechanism of the fastandslow transport channels. Here, we take the V_{g} = −0.1 and V_{g} = 0.6 cases with p_{R} = 0.9 as examples to illustrate these characteristics.
For the V_{g} = −0.1 case, the transition from singlyoccupied to empty eigenstates first participates in the electron transport with increasing the bias voltage, see Figs. 5(a) and 5(b). When the bias voltage is larger than 0.33 meV, the SMM has a small probability of forming the emptyoccupied eigenstate 0, −2〉, see the thick solid line in Fig. 5(a). If the spindown electron tunnels into the SMM, the singlyoccupied eigenstate 1, −5/2〉 can be formed. In this case, for a large enough spin polarization p_{R}, namely, , the spindown electron will remain for a relatively long time in the SMM, so that the electron tunneling processes via the fast transport channels and can be blocked and the conduction electrons appear the bunching effect. On the other hand, the current magnitude of the formed fast transport channel begins to decrease with increasing the bias voltage up to about 4.25 meV, while that of the two new opened transport channels and increase. Since the occupation probabilities of the eigenstates 1, −3/2〉^{−} and 1, −5/2〉, and P_{1,−5/2〉} are much smaller that P_{1,5/2〉}, see Fig. 5(b), thus, for the case the decreased current magnitude of transport channel is much larger than the increased current magnitudes of transport channels and . Thus, a strong NDC is observed, see the solid line in Fig. 4(e). Moreover, the active competition between the fast channel of current decreasing and the slow channels of current increasing can also obviously enhance the shot noise. Consequently, the shot noise is significantly enhanced by the above two mechanisms and reaches up to a very large value of superPoissonian shot noise before the occupation probabilities and P_{1,−5/2〉} are larger than P_{1,5/2〉} and P_{0,−2〉} is larger than P_{0,2〉}. With the bias voltage further increasing, the value of superPoissonian shot noise is decreased quickly but still remains the superPoissonian distribution. This originates from the fact that the transport channels and can form a new effective competition between the fast and slow transport channels. When the occupation probability P_{2,2〉} is larger than P_{0,2〉} (), the active competition between the transport channels and is destroyed by the new transport channel due to the electron tunneling process via the transport channel can occur, which is responsible for the superPoissonian shot noise being decreased to the subPoissonian distribution.
As for the V_{g} = 0.6 case, the transport channel , which is a slow transport channel for the case, first participates in the electron transport, see Figs. 5(c) and 5(d). With the bias voltage increasing up to about 0.4 meV, the fast electron tunneling process via the transport channel takes place, thus, the effective competition between fast and slow transport channels can form and the shot noise is rapidly enhanced and reaches up to a relatively large superPoissonian value. However, the new transport channels and can be quickly opened with the bias voltage further increasing, then the fast transport channel will be weakened when a spindown electron tunnels out the SMM through the transition from the eigenstates 2, −2〉 to 1, −3/2〉^{−}, so that the formed effective competition between fast and slow transport channels is suppressed and even destroyed. Moreover, when the transport channel does not participate in the quantum transport originating from the occupation probabilities P_{2,2〉} and P_{1,5/2〉} being approaching zero, the two transport channels and can not form a new effective competition between fast and slow transport channels due to a relatively fast electron tunneling process via can take place. Consequently, the superPoissonian shot noise is decreased quickly to a subPoissonian value and displays a sharp peak.
The ferromagnetic lead (Source)  SMM  ferromagnetic lead (Drain)
We now consider the FSMMF system, the strengths of the spindependent SMMelectrode coupling are characterized by and , here we set p_{L} = p_{R} = p. For a small enough or negative gate voltage and relatively large spin polarization of the source and drain electrodes p, an obvious NDC is observed but weaker than that in the NSMMF system, especially for a large enough spin polarization p, see the solid and dashed lines in Figs. 4(a) and 6(a) and 4(e) and 6(e). While for a relatively large gate voltage, such as V_{g} ≥ 0.4 meV, a weak NDC can be observed for a large enough spin polarization p, but that in the NSMMF system does not occur. Interestingly, for a small enough or negative gate voltage, the shot noise in the NDC region is dramatically enhanced and reaches up to a superPoissonian value, see the solid and dashed lines in Figs. 6(b) and 6(f); whereas for a large enough gate voltage the formed superPoissonian shot noise in the NDC region is decreased, see the short dashed, short dashdotted and thick dashed lines in Fig. 6(f). This characteristic depends on the formation mechanism of the NDC, which is illustrated by the examples of V_{g} = −0.1 and V_{g} = 0.6 with p = 0.9.
For a negative gate voltage V_{g} = −0.1, the fast transport channel first enters the bias voltage window. When the bias voltage increases up to about 0.48 meV, the new spinup electron tunneling processes, namely, , , , and the spindowm electron tunneling processes, namely, , , , , begin to participate in the quantum transport, see Figs. 7(a) and 7(b). This leads to the current magnitude of the fast transport channel decrease, but the increased current magnitudes of the new opened transport channels are too small to compensate the decreased current magnitude of . Thus, a NDC region can form, in which the corresponding shot noise is rapidly enhanced by the active competition between the fast channel of current decreasing and the slow channels of current increasing and reaches up to a large superPoissonian value, see the solid line in Fig. 6(f). With further increasing the bias voltage, the formed active competition between the fast channel of current decreasing and the slow channels of current increasing is weakened and even disappears, but the effective competition between the spinup and spindown electron tunneling processes is still valid due to and , thus, the value of the formed superPoissonian begins to continually decrease but still remains superPoissonian distribution. When the bias voltage increases up to 0.8 meV, the current magnitudes of the transport channels originating from the transitions between the double and singlyoccupied eigenstates are already larger than that of the some transport channels originating from the transitions between the singly and emptyoccupied eigenstates, for example, . In this case, the formed effective competition between the fast and slow transport channels is suppressed and finally destroyed due to these transport channels via the transitions from the double to singlyoccupied eigenstates entering the bias voltage. Consequently, the superPoissonian shot noise is decreased quickly up to a subPoissonian value, see the solid line in Fig. 6(f).
Compared with the V_{g} = −0.1 case, for V_{g} = 0.6 the transport channel first participates in the quantum transport. When the bias voltage increases up to about 0.48 meV, the spinup transport channels , , , , and the spindown transport channels , , , can be opened, while the current magnitude of the transport channel begin to decrease, see Figs. 7(c) and 7(d). For the case, the decreased current magnitude of the spindown transport channel is smaller than the increased current magnitudes of the new opened transport channels, thus, the NDC does not appear. Whereas the active competition between the fast channel of current decreasing and the slow channels of current increasing in a relatively small bias voltage range can form but soon be destroyed, so that the shot noise is significantly enhanced up to a very large superPoissonian value, then this value begins to decrease but still remains superPoissonian distribution due to the effective competition between the spinup and spindown electron tunneling processes being still valid, see the thick dashed line in Fig. 6(f). In particular, it is interesting note that the current magnitudes of the transport channels and increase with further increasing the bias voltage, while the current magnitudes of the other transport channels , , , , , , and decrease. This feature leads to the occurrence of a weak NDC. In this NDC bias voltage range, however, the superPoissonian shot noise value continually decreases, see the thick dashed line in Fig. 6(f). When the transport channels originating from the transitions from the singly to emptyoccupied eigenstates enter the bias voltage, the physical mechanism of decreasing superPoissonian shot noise is the same as the V_{g} = −0.1 case, namely, the formed effective competition between the spinup and spindown electron tunneling processes is weakened even destroyed by these current increased transport channels. This is responsible for the superPoissonian shot noise being decreased to a subPoissonian value.
We now study the skewness and kurtosis properties of the transport current in the superPoissonian shot noise bias voltage regions. It is well known that the skewness and kurtosis (both its magnitude and sign) characterize, respectively, the asymmetry of and the peakedness of the probability distribution around the average transferredelectron number during a time interval t, thus that provide further information for the counting statistics beyond the shot noise. In the NSMMF system with a given small enough or negative gate voltage, for a relatively large p_{R}, the skewness shows a crossover from a large negative to a relatively small positive values, while the kurtosis shows a crossover from a large positive to a relatively small negative values, see the solid and dashed lines in Figs. 4(c) and 4(d); whereas for a large enough p_{R}, the transition of the skewness from a large negative to a large positive values takes place and forms a Fanolike resonance, see the solid, dashed and dotted lines in Fig. 4(g), while the transitions of the kurtosis from a large positive to a large negative values and then from the large negative to a large positive values take place and form the double Fanolike resonances, see the solid, dashed and dotted lines in Fig. 4(h). In contrast with a small enough or negative gate voltage, for a large enough gate voltage, the skewness and kurtosis for a relatively large p_{R} show, respectively, the crossovers from a large positive to a relatively small negative values and from a small positive to a relatively large negative values, see the short dashdotted and thick dashed lines in Figs. 4(c) and 4(d); while for a large enough p_{R}, the skewness and kurtosis show, respectively, the crossovers from a small positive to a relatively large negative values and from a small negative large to a relatively large positive values, see the short dashdotted and thick dashed lines in Figs. 4(g) and 4(h), but the variations in the magnitudes of the skewness and kurtosis are much smaller than that for a small enough or negative gate voltage, see Figs. 4(g) and 4(h). As for the FSMMF system with a given relatively large p, the skewness for a small enough or negative gate voltage shows a large negative value, see the solid, dashed and dotted lines in Figs. 6(c) and 6(g), whereas for a large enough gate voltage that shows a large positive value, see the short dashed, short dashdotted and thick dashed lines in Figs. 6(c) and 6(g); while the kurtosis shows the doublecrossover from a large positive to a relatively small negative values and then from the negative to a large positive values but the latter has a remarkable variation in the magnitude of the kurtosis, see Figs. 6(d) and 6(h). Moreover, we found that the magnitudes of the skewness and kurtosis are more sensitive to the active competition between the fast channels of current decreasing and the corresponding slow channels of current increasing than the shot noise, see the short dashed, short dashdotted and thick dashed lines in Figs. 2, 4 and 6.
Discussion
Now, we discuss the feasibility of the experimental tests of the current noise properties of the SMM junctions. The differential conductance of the Mn_{12}^{1,2,8}, Fe_{4}^{5,10} and TbPc_{2}^{7,9,11} SMM junctions has been experimentally demonstrated. On the other hand, the highorder current cumulants of electron transport through a semiconductor quantum dot have been realized experimentally^{51,52}, especially the fifteenorder cumulants can be extracted from the highquality realtime singleelectron measurements^{52}. Consequently, in principle, the quantum transport of individual electrons through a SMM can be detected using a quantum point contact^{51,52} or a carbon nanotube fieldeffect transistor^{28} acting as a charge sensor. This provides a possible opportunity to experimentally test the predicted negative differential conductance, superPoissonian shot noise and the highorder current cumulants of the SMM junctions considered here. Moreover, it should be noted that the shot noise in a SMM junction depends on the internal level structure of the SMM^{32,33,34}, the leftright asymmetry of the SMMelectrode coupling^{32,33,34} and the spin polarization of the source and drain electrodes. However, in the present SMM junctions fabricated by the breakjunction and electromigration techniques^{1,2,5,10}, the angle of the easy axis of the SMM with respect to the polarization vectors of the source and drain electrodes^{5,10} and the leftright asymmetry of the SMMelectrode coupling varies from sample to sample. Thus, the shot noise properties of the SMM junction varies from sample to sample.
In summary, we have studied the the NDC and superPoissonian shot noise properties of electron transport through a SMM weakly coupled to two electrodes with either one or both of them being ferromagnetic and analyzed the skewness and kurtosis properties of the transport current in the superPoissonian shot noise regions. It is demonstrated that the occurrences of the NDC and superPoissonian shot noise depend sensitively on the spin polarization of the soure and drain electrodes and the applied gate voltage. For the FSMMN system, when the transition from the double to singlyoccupied eigenstates first enters the bias voltage window, which corresponds to a large enough gate voltage, the superPoissonian shot noise is observed for a large enough spin polarization of left electrode p_{L}. As for the NSMMF system, the NDC and superPoissonian shot noise can be observed for a relatively large spin polarization of right electrode p_{R} and a small enough or negative gate volatge, especially for a large enough p_{R} a strong NDC and a very large value of the superPoissonian shot noise appear and the shot noise in the NDC region is first enhanced up to a superPoissonian value and then is decreased but still remains superPoissonian distribution for a large enough p_{R}; while for a large enough gate voltage and a relatively large p_{R} the superPoissonian shot noise is only observed. Compared with the NSMMF system, for the FSMMF system a relatively weak NDC and a large superPoissonian shot noise bias voltage range are observed; whereas the formed superPoissonian shot noise in the NDC region is continually decreased for a large enough gate voltage and spin polarization of left and right electrodes p. Furthermore, the transitions of the skewness and kurtosis from a large positive (negative) to a large negative (positive) values are also observed, which can provide a deeper and better understanding of electron transport through singlemolecule magnet junctions. The observed NDC and superPoissonian shot noise in the SMM system can be qualitatively attributed to the effective competition between the fast and slow transport channels and the NDC properties suggest a gatevoltagecontrolled NDC molecular device.
Methods
The SMMelectrode coupling is assumed to be sufficiently weak, so that the sequential tunneling is dominant. The transitions are well described by the quantum master equation of a reduced density matrix spanned by the eigenstates of the SMM. Under the second order Born approximation and Markov approximation, the particlenumberresolved quantum master equation for the reduced density matrix is given by^{53,54}
with
where , , and (f_{α} is the Fermi function of the electrode α). Liouvillian superoperator is defined as . ρ^{(n)} (t) is the reduced density matrix of the SMM conditioned by the electron numbers arriving at the right electrode up to time t. In order to calculate the first four cumulants, one can define . According to the definition of the cumulant generating function , we evidently have e^{−F(χ)} = Tr[S(χ, t)], where the trace is over the eigenstates of the SMM. Since Eq. (6) has the following form
S (χ, t) satisfies
In the low frequency limit, the counting time is much longer than the time of electron tunneling through the SMM. In this case, F (χ) can be expressed as^{55,56,57,58}
where λ_{1} (χ) is the eigenvalue of L_{χ} which goes to zero for χ → 0. According to the definition of the cumulants, one can express λ_{1} (χ) as
Here, the first four cumulants C_{k} are directly related to the transport characteristics. For example, the firstorder cumulant (the peak position of the distribution of transferredelectron number) gives the average current 〈I〉 = eC_{1}/t. The zerofrequency shot noise is related to the secondorder cumulant (the peakwidth of the distribution) . The thirdorder cumulant and fourorder cumulant characterize, respectively, the skewness and kurtosis of the distribution. Here, . In general, the shot noise, skewness and kurtosis are represented by the Fano factors F_{2} = C_{2}/C_{1}, F_{3} = C_{3}/C_{1} and F_{4} = C_{4}/C_{1}, respectively. Moreover, the Fano factors F_{2} < 1, F_{2} = 1 and F_{2} > 1 describe, respectively, the subPoissonian shot noise, the Poisson noise and the superPoissonian shot noise.
The low order cumulants can be calculated by the Rayleigh–Schrödinger perturbation theory in the counting parameter χ, which was developed in Refs. 57, 59, 60. In order to calculate the first four current cumulants we expand L_{χ} to four order in χ
Along the lines of Refs. 57, 59, 60, we define the two projectors and Q = Q^{2} = 1 − P, obeying the relations PL_{0} = L_{0}P = 0 and QL_{0} = L_{0}Q = L_{0}. Here, 0〉〉 being the steady state ρ^{stat} is the right eigenvector of L_{0}, namely, L_{0}0〉〉 = 0 and is the corresponding left eigenvector. In view of L_{0} being singular, we also introduce the pseudoinverse according to , which is welldefined due to the inversion being performed only in the subspace spanned by Q. After a careful calculation, λ_{1} (χ) is given by
From Eqs. (11) and (13) we can identify the first four current cumulants:
The above four equations are the starting point of the numerical calculation.
References
Heersche, H. B. et al. Electron transport through single Mn12 molecular magnets. Phys. Rev. Lett. 96, 206801 (2006).
Jo, M. H. et al. Signatures of molecular magnetism in singlemolecule transport spectroscopy. Nano Lett. 6, 2014–2020 (2006).
Grose, J. E. et al. Tunnelling spectra of individual magnetic endofullerene molecules. Nat. Mater. 7, 884–889 (2008).
Loth, S. et al. Controlling the state of quantum spins with electric currents. Nat. Phys. 6, 340–344 (2010).
Zyazin, A. S. et al. Electric field controlled magnetic anisotropy in a single molecule. Nano Lett. 10, 3307–3311 (2010).
Roch, N. et al. Cotunneling through a magnetic singlemolecule transistor based on N@C60 . Phys. Rev. B 83, 081407(R) (2011).
Komeda, T. et al. Observation and electric current control of a local spin in a singlemolecule magnet. Nat. Commun. 2, 217 (2011).
Kahle, S. et al. The quantum magnetism of individual manganese12acetate molecular magnets anchored at surfaces. Nano Lett. 12, 518–521 (2012).
Vincent, R., Klyatskaya, S., Ruben, M., Wernsdorfer, W. & Balestro, F. Electronic readout of a single nuclear spin using a molecular spin transistor. Nature 488, 357–360 (2012).
Burzurí, E., Zyazin, A. S., Cornia, A. & van der Zant, H. S. J. Direct observation of magnetic anisotropy in an individual Fe4 singlemolecule magnet. Phys. Rev. Lett. 109, 147203 (2012).
Thiele, S. et al. Electrically driven nuclear spin resonance in singlemolecule magnets. Science 334, 1135–1138 (2014).
Romeike, C., Wegewijs, M. R., Hofstetter, W. & Schoeller, H. Quantumtunnelinginduced Kondo effect in single molecular magnets. Phys. Rev. Lett. 96, 196601 (2006).
Romeike, C., Wegewijs, M. R., Hofstetter, W. & Schoeller, H. Kondotransport spectroscopy of single molecule magnets. Phys. Rev. Lett. 97, 206601 (2006).
Timm, C. & Elste, F. Spin amplification, reading and writing in transport through anisotropic magnetic molecules. Phys. Rev. B 73, 235304 (2006).
Elste, F. & Timm, C. Transport through anisotropic magnetic molecules with partially ferromagnetic leads: Spincharge conversion and negative differential conductance. Phys. Rev. B 73, 235305 (2006).
Elste, F. & Timm, C. Cotunneling and nonequilibrium magnetization in magnetic molecular monolayers. Phys. Rev. B 75, 195341 (2007).
Timm, C. Tunneling through magnetic molecules with arbitrary angle between easy axis and magnetic field. Phys. Rev. B 76, 014421 (2007).
Wegewijs, M. R., Romeike, C., Schoeller, H. & Hofstetter, W. Magnetotransport through singlemolecule magnets: Kondopeaks, zerobias dips, molecular symmetry and Berry's phase. New J. Phys. 9, 344 (2007).
González, G. & Leuenberger, M. N. Berryphase blockade in singlemolecule magnets. Phys. Rev. Lett. 98, 256804 (2007).
González, G., Leuenberger, M. N. & Mucciolo, E. R. Kondo effect in singlemolecule magnet transistors. Phys. Rev. B 78, 054445 (2008).
Misiorny, M., Weymann, I. & Barnaś, J. Spin diode behavior in transport through singlemolecule magnets. Europhys. Lett. 89, 18003 (2010).
Misiorny, M., Weymann, I. & Barnaś, J. Influence of magnetic anisotropy on the Kondo effect and spinpolarized transport through magnetic molecules, adatoms and quantum dots. Phys. Rev. B 84, 035445 (2011).
Xie, H., Wang, Q., Chang, B., Jiao, H. J. & Liang, J. Q. Spin current and polarization reversal through a singlemolecule magnet with ferromagnetic electrodes. J. Appl. Phys. 111, 063707 (2012).
Misiorny, M. & Barnaś, J. Effects of transverse magnetic anisotropy on currentinduced spin switching. Phys. Rev. Lett. 111, 046603 (2013).
Xie, H., Wang, Q., Xue, H. B., Jiao, H. J. & Liang, J. Q. Intrinsic spinrelaxation induced negative tunnel magnetoresistance in a singlemolecule magnet. J. Appl. Phys. 113, 213708 (2013).
Wang, R. N., Rodriguez, J. H. & Liu, W. M. Negative magnetoresistance and spin filtering of spincoupled diironoxo clusters. Phys. Rev. B 89, 235414 (2014).
Lapo, B. & Wolfgang, W. Molecular spintronics using singlemolecule magnets. Nat. Mater. 7, 179–186 (2008).
Gruneis, A., Esplandiu, M. J., GarciaSanchez, D. & Bachtold, A. Detecting individual electrons using a carbon nanotube fieldeffect transistor. Nano Lett. 7, 3766–3769 (2007).
Romeike, C., Wegewijs, M. R. & Schoeller, H. Spin Quantum tunneling in single molecular magnets: fingerprints in transport spectroscopy of current and noise. Phys. Rev. Lett. 96, 196805 (2006).
Imura, K. I., Utsumi, Y. & Martin, T. Full counting statistics for transport through a molecular quantum dot magnet: Incoherent tunneling regime. Phys. Rev. B 75, 205341 (2007).
Misiorny, M., Weymann, I. & Barnaś, J. Spin effects in transport through singlemolecule magnets in the sequential and cotunneling regime. Phys. Rev. B 79, 224420 (2009).
Xue, H. B., Nie, Y. H., Li, Z. J. & Liang, J. Q. Tunable electron counting statistics in a singlemolecule magnet. J. Appl. Phys. 108, 033707 (2010).
Xue, H. B., Nie, Y. H., Li, Z. J. & Liang, J. Q. Effect of finite Coulomb interaction on full counting statistics of electronic transport through singlemolecule magnet. Phys. Lett. A 375, 716–725 (2011).
Xue, H. B., Nie, Y. H., Li, Z. J. & Liang, J. Q. Effects of magnetic field and transverse anisotropy on full counting statistics in singlemolecule magnet. J. Appl. Phys. 109, 083706 (2011).
Wang, R. Q., Shen, R., Zhu, S. L., Wang, B. & Xing, D. Y. Inelastic transport detection of spin quantum tunneling and spin relaxation in singlemolecule magnets in the absence of a magnetic field. Phys. Rev. B 85, 165432 (2012).
ContrerasPulido, L. D. & Aguado, R. Shot noise spectrum of artificial singlemolecule magnets: Measuring spin relaxation times via the Dicke effect. Phys. Rev. B 81, 161309 (2010).
Blanter,. Ya, M. & Büttiker, M. Shot noise in mesoscopic conductors. Phys. Rep. 336, 1–166 (2000).
Quantum Noise in Mesoscopic Physics, edited by Yu. V. Nazarov (Kluwer, Dordrecht, 2003).
Farajian, A. A., Esfarjani, K. & Kawazoe, Y. Nonlinear coherent transport through doped nanotube junctions. Phys. Rev. Lett. 82, 5084 (1999).
Luo, W., Wang, R. Q., Hu, L. B. & Yang, M. Spindependent negative differential conductance in transport through singlemolecule magnets. Chin. Phys. B 22, 047201 (2013).
Aghassi, J., Thielmann, A., Hettler, M. H. & Schön, G. Shot noise in transport through two coherent strongly coupled quantum dots. Phys. Rev. B 73, 195323 (2006).
Li, D. et al. An S = 2 cyanidebridged trinuclear singlemolecule magnet. Inorg. Chem. 45, 5251–5253 (2006).
Lindebaum, S., Urban, D. & König, J. Spininduced charge correlations in transport through interacting quantum dots with ferromagnetic leads. Phys. Rev. B 79, 245303 (2009).
Safonov, S. S. et al. Enhanced shot Noise in resonant tunneling via interacting localized states. Phys. Rev. Lett. 91, 136801 (2003).
Djuric, I., Dong, B. & Cui, H. L. SuperPoissonian shot noise in the resonant tunneling due to coupling with a localized level. Appl. Phys. Lett. 87, 032105 (2005).
Aghassi, J., Thielmann, A., Hettler, M. H. & Schön, G. Strongly enhanced shot noise in chains of quantum dots. Appl. Phys. Lett. 89, 052101 (2006).
Sánchez, R., Kohler, S., Hänggi, P. & Platero, G. Electron bunching in stacks of coupled quantum dots. Phys. Rev. B 77, 035409 (2008).
Xue, H. B., Zhang, Z. X. & Fei, H. M. Tunable superPoissonian noise and negative differential conductance in two coherent strongly coupled quantum dots. Eur. Phys. J. B 85, 336 (2012).
Zhang, H. W., Xue, H. B. & Nie, Y. H. Full counting statistics of a quantum dot doped with a single magnetic impurity. AIP Advances 3, 102116 (2013).
Xue, H. B. Full counting statistics as a probe of quantum coherence in a sidecoupled double quantum dot system. Annals of Physics (New York) 339, 208 (2013).
Fujisawa, T., Hayashi, T., Tomita, R. & Hirayama Y. Bidirectional counting of single electrons. Science 312, 1634–1636 (2006).
Flindt, C. et al. Universal oscillations in counting statistics. Proc. Natl. Acad. Sci. (USA) 106, 10116–10119 (2009).
Li, X. Q., Luo, J., Yang, Y. G., Cui, P. & Yan, Y. J. Quantum masterequation approach to quantum transport through mesoscopic systems. Phys. Rev. B 71, 205304 (2005).
Wang, S. K., Jiao, H., Li, F., Li, X. Q. & Yan, Y. J. Full counting statistics of transport through twochannel Coulomb blockade systems. Phys. Rev. B 76, 125416 (2007).
Bagrets, D. A. & Nazarov, Yu. V. Full counting statistics of charge transfer in Coulomb blockade systems. Phys. Rev. B 67, 085316 (2003).
Groth, C. W., Michaelis, B. & Beenakker, C. W. J. Counting statistics of coherent population trapping in quantum dots. Phys. Rev. B 74, 125315 (2006).
Flindt, C., Novotny, T. & Jauho, A. P. Full counting statistics of nanoelectromechanical systems. Europhys. Lett. 69, 475 (2005).
Kießlich, G., Samuelsson, P., Wacker, A. & Schöll, E. Counting statistics and decoherence in coupled quantum dots. Phys. Rev. B 73, 033312 (2006).
Flindt, C., Novotny, T., Braggio, A., Sassetti, M. & Jauho, A. P. Counting Statistics of NonMarkovian Quantum Stochastic Processes. Phys. Rev. Lett. 100, 150601 (2008).
Flindt, C., Novotny, T., Braggio, A., & Jauho, A. P. Counting statistics of transport through Coulomb blockade nanostructures: Highorder cumulants and nonMarkovian effects. Phys. Rev. B 82, 155407 (2010).
Acknowledgements
This work was supported by the NKBRSFC under grants Nos. 2011CB921502, 2012CB821305, NSFC under grants Nos. 11204203, 61405138, 11275118, 61227902, 61378017, 11434015, SKLQOQOD under grants No. KF201403, SPRPCAS under grants No. XDB01020300.
Author information
Affiliations
Contributions
H.B.X. conceived the idea and designed the research and performed calculations. J.Q.L. and W.M.L. contributed to the analysis and interpretation of the results and prepared the manuscript.
Ethics declarations
Competing interests
The authors declare no competing financial interests.
Rights and permissions
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder in order to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
About this article
Cite this article
Xue, HB., Liang, JQ. & Liu, WM. Negative differential conductance and superPoissonian shot noise in singlemolecule magnet junctions. Sci Rep 5, 8730 (2015). https://doi.org/10.1038/srep08730
Received:
Accepted:
Published:
Further reading

Direction inversion of spinpolarization current and exceptional tunnel magnetoresistance
AIP Advances (2020)

Noisy defects in the highTc superconductor Bi2Sr2CaCu2O8+x
Nature Communications (2019)

Full counting statistics in a serially coupled double quantum dot system with spin–orbit coupling
Annals of Physics (2018)

DopingInduced Negative Differential Conductance Enhancement in SingleMolecule Junction
Physica E: Lowdimensional Systems and Nanostructures (2018)

Efficient spincurrent injection in singlemolecule magnet junctions
AIP Advances (2018)
Comments
By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.