Giant Magneto Resistance Through a Single Molecule

8/3/2019 Giant Magneto Resistance Through a Single Molecule

1/5

Giant magnetoresistance through a single moleculeStefan Schmaus1,2, Alexei Bagrets2,3, Yasmine Nahas1,2, Toyo K. Yamada1,4, Annika Bork1,

Martin Bowen5, Eric Beaurepaire5, Ferdinand Evers3,6 and Wulf Wulfhekel1,2*

Magnetoresistance is a change in the resistance of a material system caused by an applied magnetic field. Giantmagnetoresistance occurs in structures containing ferromagnetic contacts separated by a metallic non-magnetic spacer,and is now the basis of read heads for hard drives and for new forms of random access memory. Using an insulator (forexample, a molecular thin film) rather than a metal as the spacer gives rise to tunnelling magnetoresistance, whichtypically produces a larger change in resistance for a given magnetic field strength, but also yields higher resistances,which are a disadvantage for real device operation. Here, we demonstrate giant magnetoresistance across a single, non-magnetic hydrogen phthalocyanine molecule contacted by the ferromagnetic tip of a scanning tunnelling microscope. Wemeasure the magnetoresistance to be 60% and the conductance to be 0.26G

0, where G

0is the quantum of conductance.

Theoretical analysis identifies spin-dependent hybridization of molecular and electrode orbitals as the cause of thelarge magnetoresistance.

Spin-polarized transport through a single molecule has attractedinterest, because it combines the goals of downward size scalingof electronic components (to lower power consumption while

increasing speed and density) and increasing computational band-width by manipulating spin in addition to charge. The concept ofspin electronics1,2 was pioneered with the discovery that the resist-ance of two ferromagnetic layers separated by a thin non-magneticspacer may be driven by an external magnetic field or the flow ofcharge. In the case of a thin metallic spacer, this effect is termedgiant magnetoresistance (GMR)3,4, and for ultrathin insulatinglayers it is called tunnelling magnetoresistance (TMR)5,6. Acrossthin organic semiconducting layers, transport may occur via tunnel-ing or sequential tunneling, that is, diffusive hopping7,8, demonstrat-ing a magnetoresistance thanks to the low values of spinorbitcoupling, despite the presence of defects in these imperfectlymastered semiconductors9,10.

Spin electronics has led to new generations of read heads in harddisk drives and random access memory, but further progress indownsizing spintronic devices has been challenging. Indeed,although TMR values may dwarf those of their GMR counterpartsthrough a judicious selection of ferromagnets11 and inorganic12 ororganic13,14 tunnel barriers, downsizing such devices can lead tohigh resistances8 and shot noise, which are incompatible withhigh-frequency applications15. Separately, as the device size isreduced, the GMR bias output shrinks below the bias drop of theleads, which in turn requires one to artificially lower the currentdensities to achieve a high magnetic sensitivity16.

A promising approach to extreme downsizing has been to study

electronic single-molecule devices such as diodes17

and transis-tors1820. Turning now to molecular spintronics, recent studieshave described the spintronic impact of chemisorption of mol-ecules21 onto ferromagnets. This highly conductive interface22 candrive spintronic properties8,21, due in part to ferromagnetic couplingbetween the transition metal site of the molecule and the ferromag-netic surface21,23,24.

Here, we present a combined experimental and theoretical studyon spin-polarized charge transport across a single molecule in theGMR regime that takes advantage of these novel molecular spintro-nic properties. The junction, built with the help of a scanning tun-nelling microscope (STM), consists of two ferromagnetic electrodesbridged by a hydrogen phthalocyanine (H2PcC32H18N8) mol-ecule. Our study shows that, due to charge transfer towards the mol-ecule and the hybridization of molecular orbitals near the Fermilevel with bulk electronic states of the electrodes, transport acrossthe molecule is nearly resonant in the minority channel with anultralow areal resistance product. Surprisingly, the molecule cansustain a large current density with substantial spin-polarization,thus validating the concept of nanoscale spintronic devices.Indeed, despite the absence of transition metal sites on the moleculethat could contribute to spin filtering, we observe a high differentialGMR value of60% across the molecular junction, with an arealresistance product of only 70 mVmm2. Our ab initio theoreticalanalysis explicitly identifies this effect as a strong, generic, spin-dependent hybridization mechanism that can also dominatethe magnetoresistance in the TMR regime, as was reportedphenomenologically8.

Sample characterizationSTM2527 and break-junction28,29 techniques have generally provento be powerful and versatile tools with which to study the transportproperties of single-molecule junctions. In this work, we use magneticelectrodes to measure the magnetoresistance of single molecules usingspin-polarized STM (Sp-STM). Experiments were carried out in a

home-built STM instrument working in ultra-high vacuum at 4 K(ref. 30) on individual H2Pc molecules sandwiched betweencobalt-coated tungsten tips and ferromagnetic cobalt nano-islandson Cu(111) single crystals (see Supplementary Information).

Figure 1a presents the STM topography of the sample. The cobaltnano-islands exhibit a spontaneous out-of-plane magnetization dueto a strong interfacial anisotropy31,32. With cobalt-coated tips

1Physikalisches Institut, Karlsruhe Institute of Technology (KIT), 76128 Karlsruhe, Germany, 2DFG-Center for Functional Nanostructures, Karlsruhe Instituteof Technology (KIT), 76128 Karlsruhe, Germany, 3Institute of Nanotechnology, Karlsruhe Institute of Technology (KIT), 76128 Karlsruhe, Germany,4Graduate School of Advanced Integration Science, Chiba University, Chiba 263-8522, Japan, 5Institut de Physique et Chimie des Materiaux de Strasbourg,UMR 7504 UdS-CNRS, 67034 Strasbourg Cedex 2, France, 6Institut Fur Theorie der Kondensierten Materie, Karlsruhe Institute of Technology (KIT),D-76128 Karlsruhe, Germany. *e-mail: [email protected]

ARTICLESPUBLISHED ONLINE: 20 FEBRUARY 2011 | DOI: 10.1038/NNANO.2011.11

NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology 1
mailto:[email protected]://www.nature.com/doifinder/10.1038/nnano.2011.11http://www.nature.com/naturenanotechnologyhttp://www.nature.com/naturenanotechnologyhttp://www.nature.com/doifinder/10.1038/nnano.2011.11mailto:[email protected]


2/5

(10 monolayers) showing an out-of-plane magnetization, we can usethe sensitivity of the Sp-STM technique to the spin-polarizeddensity of states33 to determine the orientation (parallel or antipar-allel) of the magnetization of individual islands relative to that of thetip. Figure 1b shows typical differential conductance (dI/dV) curvesmeasured on top of cobalt islands of parallel and antiparallel orien-tation. Particularly large differences in the spectra are found at2350 meV, which corresponds to the surface state of cobalt34. Byrecording maps of the differential conductance at this bias voltage

(Fig. 1a), we can thus identify the spin orientation of the cobaltislands and determine the differential TMR. The latter is definedas the difference in differential conductance (dI/dV) divided bythe smaller differential conductance of the tunnel junction formedby the tip and cobalt island for ensembles of junctions with paralleland antiparallel oriented cobalt islands. As can be seen in Fig. 1c,the differential TMR for such cobalt/vacuum/cobalt junctions isstrongly energy-dependent, as reported previously32. Remarkably,the differential TMR reaches only5% at low bias voltage.

Spin-dependent conductance across a single moleculeThe H2Pc molecules that were also deposited adsorb lying flat on thecobalt surface24, and are therefore easily recognized in Fig. 1 by theirfour aromatic isoindole (BzPy) side groups35. Although the localspin-polarization in the tunnelling regime has been imaged with

Sp-STM for phthalocyanine molecules with 3d metal centres24,36,we here focus on GMR transport measurements across non-mag-netic H2Pc molecules in the contact regime. To observe the tran-sition from the TMR regime to the GMR regime across themolecule, we alter the junction geometry and achieve tip-to-mol-ecule contact by positioning the STM tip above the aromatic sidegroups of a H2Pc molecule, opening the feedback loop and decreas-ing the tip-to-molecule distance by a small offset (1 s21) whilemeasuring the tunnelling current2527,37,38. We first observe an expo-nential increase of the conductance due to the decreasing tunnellingbarrier width (Fig. 2a). Below a certain tip-to-surface separation(typically 34 ), the conductance abruptly increases and thereafterdepends only weakly on the distance. As reported regarding alkane-dithiol molecular wires26, this underscores the jump-to-contact of a

molecular segment. For H2Pc molecules, the transition between theflat adsorption and the contact geometry has been ascribed to the20 meV vibrational bending mode of the aromatic side groups ofthe H2Pc molecule. Note that when voltages higher than thisvibrational mode are applied in the contact geometry, the vibrationcan be excited by the current and the molecular contact becomesunstable (see Supplementary Information).

The conductance after the jump-to-contact encompasses twocontributions: the current across the molecule Gmol and a leakage

current Gtun due to direct tunnelling between the tip and sample26

(Fig. 2c). To obtain a reasonable estimate of the molecular conduc-tance Gmol, we therefore subtract the conductance Gtun measuredbefore the jump from the conductance Gcont measured after the jump, that is, Gmol Gcont2Gtun. Because measurements on twoparallel and antiparallel oriented cobalt islands are required,using identical tip conditions, to extract the GMR correctly, werecord the traces from two such islands in the same topographicscan. This method notably eliminates possible magnetostrictioneffects between the two ferromagnetic electrodes, because neitherthe tip nor the island magnetization need be switched by an externalmagnetic field.

In the particular measurement at 10 mV in Fig. 2a, we find a con-ductance ofGP 0.26G0 for the parallel and GAP 0.19G0 for theantiparallel magnetic configuration in units of G0 2e

2/h. These

rather large conductance values underscore a strong hybridizationbetween the electronic levels of the metallic electrodes and thoseof the molecule21. This is in sharp contrast to the orders of magni-tude lower values reported for almost insulating molecular TMRdevices8,13 and single-molecule devices based on C60 (ref. 39).

Each such measurement was repeated several hundred times onthe four side groups of about 10 molecules adsorbed onto the twoparallel and antiparallel oriented islands, leading to a distributionof the conductances across these parallel-type and antiparallel-typejunctions as depicted in Fig. 2d. Compared to previous results26,the width of the resulting conductance distribution is relativelynarrow. The broadening and possible structural elements of thesedistributions could reflect variation in the contact and binding geo-metries, but also fluctuations in the spin-polarized density of states

1.0 0.5 0.0 0.5 1.0

Sample bias (V)

1.0 0.5 0.0 0.5

Parallel

ab

c

Parallel

Antiparallel

Antiparallel

1.0

dl/dV

(nAV

1)

60

40

20

0

TMR(%)

100

150

50

0

50

Sample bias (V)

4.5(nAV1)

13.5(nAV1)

5nm

Figure 1 | H2

Pc molecules adsorbed on cobalt islands with different out-of-plane magnetic orientations. a, Topographic image of H2Pc molecules adsorbed

onto two cobalt islands on the Cu(111) surface. Colour code: measured d I/dVat 2310 mV. The two island species can be distinguished by the magnetization

parallel (in yellow) and antiparallel (in red) to the tip magnetization. b, Typical dI/dVspectra taken on parallel and antiparallel oriented cobalt islands (markedby red and blue crosses in a) clearly reveal spin-polarized density of states below the Fermi edge. c, Energy dependence of the optimistic TMR ratio

calculated from the dI/dVspectra. The highest value is measured at approximately 2350 meV, and is used to distinguish between the magnetic orientation

of the islands.

ARTICLES NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2011.11

NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology2
http://www.nature.com/doifinder/10.1038/nnano.2011.11http://www.nature.com/naturenanotechnologyhttp://www.nature.com/naturenanotechnologyhttp://www.nature.com/doifinder/10.1038/nnano.2011.11


3/5

on the islands32,40. Gaussian fits were used to determine the averageconductances and the GMR from the measurement statistics. Wefind GP (0.253+0.005)G0 and GAP (0.158+0.005)G0. Thesevalues result in an optimistic GMR ratio at V 10 mV of

GMR=GP GAP

GAP= (61+ 9)%

Surprisingly, this GMR ratio is one order of magnitude largerthan the differential TMR ratio found for direct tunnellingbetween the tip and the cobalt surface.

Transport calculations

To understand what causes this large value of GMR, we performedtransport calculations based on density functional theory (DFT)using the non-equilibrium Greens function (NEGF) formalism andthe TURBOMOLE package41 (see Supplementary Information).The atomic structure of the molecular junction was found by opti-mizing the H2Pc geometry on a Co(111) surface that was modelledusing a 65-atom cluster. Our analysis suggests that H2Pc adsorbspreferentially in the bridge position onto Co(111) (Fig. 3a), due toa binding energy28.17 eV that is larger than the one found ineither the hollow site position (28.06 eV) or the atop site position(27.45 eV), which is consistent with earlier findings24,42.

We have calculated spin-polarized transport in the linearresponse at low bias voltage for two junction geometries, schemati-cally shown in Fig. 3a,b, corresponding to the TMR and GMR

transport regimes (see Supplementary Information). Qualitatively,these calculations (Fig. 3c) reproduce very well our experimentalfindings (Fig. 2a). We confirm the exponentially increasing conduc-tance in the tunnelling regime, Gtun(d)/ e

2bd, for which the dis-

tance d between the two electrodes is still large. The slope isindependent of the relative alignment of electrode magnetizations,and the computational values ofbtheo 1.87 21 (and the workfunction Wtheo 3.24 eV) are in agreement with experiment (b1.9+0.3 21; W 3.2 eV).

Once contact between the tip and the molecule has been estab-lished, the variation ofGcont(d) with the contact distance d is veryweak, just as observed in the experiment. It is, however, still sensitiveto the relative orientation of the magnetization of the electrodes. Wefind that GAP(d) is always much lower than GP(d). To quantitativelycompare with experiment, we consider the theoretical GMR ratio atthe distance d at which the ratio rGtun/Gcont matches the valuer 4 found experimentally. We thus find that GMR 65% and isonly weakly dependent on d.

We now discuss the basic conduction mechanism that under-

scores the large GMR measured. Pc molecules are characterizedby the energetically isolated highest occupied molecular orbital(HOMO) and the nearly doubly degenerate lowest unoccupied mol-ecular orbital (LUMO)43,44. We note that the HOMO* levels corre-sponding to the aromatic group hybridize only very weakly, withalmost no amplitude on the bridging nitrogen (Nb)4. In contrast,the LUMO states are located on two of the four aromatic groups,with a strong hybridization to all nitrogen atoms forming theinner macrocycle (see Supplementary Information). Because thenitrogen bond to cobalt includes states at the Fermi energy EF,(refs 21,45), transport should occur via the (quasi-degenerate)LUMO level. We confirm this fact by examining in Fig. 4a the trans-mission probability per spin direction T,(E) near the Fermi energyacross a parallel oriented junction. We find that, for both spin

8 7.5 7 6.5

Surace-to-tip distance ()

0.1

1

Conductance,

2e

2h

1

Contact

c

a b

Tunnelling

Co(111) surace

Co(111) tip

Contact , parallelContact , antiparallelFlat , parallelFlat , antiparallel

Figure 3 | Ab initio simulations of currentdistance traces and themagnetoresistance effect across a H

2Pc molecule. a,b, Contact geometry

used in the transport calculation: for a H2Pc molecule adsorbed on the

Cobalt island (a), and in simultaneous contact with the tip and the cobalt

surface through a lifting of the aromatic group (b). Cobalt sites, grey;

hydrogen, white; carbon, green; nitrogen, cyan. c, Conductance of H2Pc

sandwiched between two parallel or antiparallel aligned Co(111) surfaces in

the tunnelling (before contact) and ballistic (after contact)

junction geometries.

Surace

Surace

Tip

0.1

a

d

b

c

ParallelAntiparallel

0.00

50 ParallelAntiparallel

40

30

20

Numberofmeasurements

10

00.2

Molecular conductance, Gmol (G0)

0.4

Conductance,

G0

0.05 0.10 0.15Ofset (nm)

0.20

0.01

GtunGcont=Gmol+Gtun

Gmol

Gtun

Surace

Tip

Gtun

Figure 2 | Currentdistance traces and magnetoresistance measured

across single H2

Pc molecules. a, Typical set of conductance-distance curves

measured on top of a H2Pc molecule adsorbed onto parallel and antiparallel

magnetized islands with a constant tunnelling voltage of 10 mV

(G0 (2e2/h)). b, As the tip approaches the molecule, the tunnel barrier

width decreases, so the conductance increases exponentially. c, Below a

certain tip-to-surface separation (typically 34 ) the conductance abruptly

increases as the molecule jumps into contact, and then varies only slightly

upon further reducing the distance. Transport across the contacted molecule

reflects both tip-to-surface tunnelling and conduction across the molecule.

d, Histogram of corrected molecular conductances (381 parallel and 366

antiparallel). A Gaussian fit is used to determine the statistical conductance

in the parallel and antiparallel configurations, and thus the GMR ratio. Error

bars indicate statistical errors in the conductance distribution.

NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2011.11 ARTICLES



4/5

channels, GP near EF is indeed dominated by a peak centred slightlyabove EF indicating transmission through the LUMO level. We finda similar origin of electron transport for an antiparallel orientedjunction (Fig. 4b).

The peak width is determined by the amount of hybridization ofthe molecular LUMO orbital (through NCo binding) with the

states of the cobalt electrodes. Because the surface density of statesof the cobalt electrodes at EF is enhanced for minority spin electronsas compared to majority spin electrons, the LUMO broadeningfound for a parallel oriented junction is stronger in the minoritytransport channel than in the majority channel (Fig. 4a). Thisbroadening asymmetry is, of course, less pronounced for an antipar-allel oriented junction (Fig. 4b). Note that our experiments operateat low voltage, 10 meV, in the regime where the current response tothe bias voltage is still linear. In this regime, the conductanceresembles transmission at the Fermi energy.

The dependence of LUMO broadening on a given spin channeland the relative alignment of electrode magnetizations both, in turn,have a direct impact on the GMR obtained across the molecular junction. Indeed, the conductance across a single level of thequasi-degenerate LUMO generically takes on the BreitWignerform46 (here a factor of 2 accounts for the degeneracy of theLUMO: G 2GsubstrateGtip/((ELUMO2 EF)

2 (Gsubstrate Gtip)2/4).

This expression considers the energy separation between theLUMO and EF, as well as the LUMO broadenings (inverse lifetimes)Gsubstrate and Gtip due to hybridization to the substrate and tip,respectively. Each is, in turn, split into Gmin(maj) depending onthe spin channel considered. Because transport is off-resonant,that is, |ELUMO2 EF| G

min,maj, we have GP 2 (GminGmin

GmajGmaj)/(ELUMO2 EF)2, while GAP 4G

minGmaj/(ELUMO2 EF)2.

Introducing the ratio @ Gmaj/Gmin, we thus find

GMR(Gmin Gmaj)2

2GminGmaj=(1 @)2

2@

This simple formula implies two important rules of thumb for spin-polarized transport off-resonance across a molecule. First, the GMRis insensitive to the precise location of the resonance energy pro-vided that EF lies reasonably within the level broadening. Second,the GMR is mainly indicative of the ratio @of minority and majoritymolecular orbital broadenings due to hybridization. As such, the

GMR may reach larger (here order-of-magnitude) values than itsTMR counterpart without molecules.Several papers have recently described means of tuning this spin-

polarized broadening of molecular orbitals near EF, and the result-ing amplitude/sign of interfacial spin polarization useful for spininjection into organic materials. For instance, Barraud andco-workers invoke disorder to explain, within a phenomenologicalmodel, the presence at EF of a localized molecular state that pro-motes spin-polarized resonant transport in the tunnelling regime8.Other reports have considered the impact of intrinsic molecularstates near EF (ref. 22) on spin polarization

21,36,45. Relative to thesereports, our results explicitly reveal the fundamental and appliedinterest in the direct coupling of two such ferromagnetmoleculeinterfaces. In this case, the spin selective broadening of theLUMO leads to a highly conductive situation in the parallel con-figuration with electronic states delocalized across the wholejunction, much like the classical and metallic GMR junctions3,4.The resulting single-molecule spintronic junction, operatingin the contact regime, combines a low resistance (52 kV) forhigh-frequency applications, a low resistancearea product(70 mVmm2) for increased density, and a large spintronic response(60%) for enhanced data-processing capabilities. This effect shouldalso be present in nanoscale solid-state devices consisting of two fer-romagnetic electrodes separated by a monolayer of organic mol-ecules useful for applications.

MethodsThe Cu(111) crystal was cleaned by several cycles of Ar sputtering and annealing.The molecules were evaporated in situ from a Knutsen cell heated to 500 K.

0.1

1

Antiparallel

2

Majority

Majority Majority

Majority

MinorityMinority

Minority Minority

11.5 1.51.5 1.50.5 0.50.5 0.50 1 2

E EF (eV)

2 1 0 1 2

E EF (eV)

0.1

1

T(E)perspin,e2h1

T(E)perspin,e2h1

Parallela b

Figure 4 | Molecular orbitals and transmission curves reveal the spin-selective LUMO broadening and the impact on GMR across the H2Pc molecule.a, Transmission probability T(E) of an electron with energy E through a parallel oriented molecular junction for the majority (red) and minority (orange) spin

channels and the current-carrying (LUMO) orbitals of the molecular junction at the Fermi energy. Hybridization of the LUMO orbitals with the cobalt states

is weak in the majority channel and strong in the minority channel. b, Corresponding data for an antiparallel oriented junction. Majority spin electrons are

injected from the cobalt surface to the minority spin band of the cobalt tip (violet), and minority spin electrons are injected from the cobalt surface to the

majority spin band of the cobalt tip (blue). In this configuration, the LUMO orbitals are hybridized in an asymmetric way between the cobalt surface and the

tip, thus accounting for the conductance decrease relative to that found for parallel oriented junctions.

ARTICLES NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2011.11

NATURE NANOTECHNOLOGY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturenanotechnology4


5/5

During the deposition process the sample was maintained at 270 K to reducethermal diffusion of the deposited molecules. dI/dV curves were measured on thebare islands with a lock-in technique.

DFT-based transport calculations were carried out with a homemade codebuilding upon the NEGF formalism and the TURBOMOLE package41. Ourimplementation enabled us to perform transport simulations with free boundaryconditions, which for the present case were extended to account for the spin-polarized electronic structure of the magnetic electrodes (for further details, seeSupplementary Information). The gradient-corrected approximation (GGA) DFTenergy was amended by empirical corrections47 to account for dispersive van der

Waals interactions between the molecule and the surface.

Received 22 November 2010; accepted 14 January 2011;

published online 20 February 2011

References1. Wolf, S. A. et al. Spintronics: a spin-based electronics vision for the future.

Science 294, 14881495 (2001).2. Zutic, I., Fabian, J. & Sarma, S. D. Spintronics: fundamentals and applications.

Rev. Mod. Phys. 76, 323410 (2004).3. Baibich, M. N. et al. Giant magnetoresistance of (001)Fe/(001)Cr magnetic

superlattices. Phys. Rev. Lett. 61, 24722475 (1988).4. Binasch, G., Grunberg, P., Saurenbach, F. & Zinn, W. Enhanced

magnetoresistance in layered magnetic structures with antiferromagneticinterlayer exchange. Phys. Rev. B 39, 48284830 (1989).

5. Julliere, M. Tunneling between ferromagnetic films. Phys. Lett. 54A,225226 (1975).

6. Moodera, J., Kinder, I., Wong, T. & Meservey, R. Large magnetoresistance atroom temperature in ferromagnetic thin film tunnel junctions. Phys. Rev. Lett.74, 32733276 (1995).

7. Xiong, Z., Wu, D., Vardeny, Z. & Shi, J. Giant magnetoresistance in organicspin-valves. Nature 427, 821824 (2004).

8. Barraud, C. et al. Unravelling the role of the interface for spin injection intoorganic semiconductors. Nature Phys. 6, 615620 (2010).

9. Pople, J. A. & Walmsley, S. H. Bond alternation defects in long polyenemolecules. Mol. Phys. 5, 1520 (1962).

10. Boukari, S. et al. Electrical transport across a structurally ordered phthalocyaninefilm: role of defect states. Phys. Rev. B 76, 033302 (2007).

11. Bowen, M. et al. Using half-metallic manganite interfaces to reveal insights intospintronics. J. Phys. Condens. Matter19, 315208 (2007).

12. Greullet, F. et al. Evidence of a symmetry-dependent metallic barrier in fullyepitaxial MgO based magnetic tunnel junctions. Phys. Rev. Lett. 99,187202 (2007).

13. Santos, T. et al. Room temperature tunnel magnetoresistance and spin polarizedtunneling studies with organic semiconductor barrier. Phys. Rev. Lett. 98,

016601 (2007).14. Szulczewski, G., Tokuc, H., Oguz, K. & Coey, J. M. D. Magnetoresistance in

magnetic tunnel junctions with an organic barrier and an MgO spin filter.Appl. Phys. Lett. 95, 202506 (2009).

15. Nagamine, Y. et al. Ultralow resistancearea product of 0.4 V (mm)2 and highmagnetoresistance above 50% in CoFeB/MgO/CoFeB magnetic tunnel junction.

Appl. Phys. Lett. 89, 162507 (2006).16. Fukuzawa, H., Yuasa, H., Hashimoto, S., Iwasaki, H. & Tanaka, Y. Large

magnetoresistance ratio of 10% by Fe50Co50 layers for current-confined pathcurrent-perpendicular-to-plane giant magnetoresistance spin-valve films.

Appl. Phys. Lett. 87, 082507 (2005).17. Elbing, M. et al. A single molecule diode. Proc. Natl Acad. Sci. USA 102,

88158820 (2005).18. Park, J. et al. Coulomb blockade and the Kondo effect in single-atom transistors.

Nature 417, 722725 (2002).19. Osorio, E. A., Bjrnholm, T., Lehn, J.-M., Ruben, M. & van der Zant, H. S. J.

Single-molecule transport in three-terminal devices. J. Phys. Condens. Matter20,

374121 (2008).20. Yu, L. H. et al. Kondo resonances and anomalous gate dependence in theelectrical conductivity of single-molecule transistors. Phys. Rev. Lett. 95,256803 (2005).

21. Javaid, S. et al. Impact on interface spin polarization of molecular bonding tometallic surfaces. Phys. Rev. Lett. 105, 077201 (2010).

22. Takacs, A. F. et al. Electron transport through single phthalocyanine moleculesstudied using scanning tunneling microscopy. Phys. Rev. B 78, 233404 (2008).

23. Wende, H. et al. Substrate-induced magnetic ordering and switching of ironporphyrin molecules. Nat. Mater. 6, 516520 (2007).

24. Iacovita, C. et al. Visualizing the spin of indivitual cobaltphthalocyaninemolecules. Phys. Rev. Lett. 101, 116602 (2009).

25. Joachim, C., Gimzewski, J. K., Schlittler, R. R. & Chavy, C. Electronictransparence of a single C60 molecule. Phys. Rev. Lett. 74, 21022105 (1995).

26. Haiss, W. et al. Precision control of single-molecule electrical junctions.Nat. Mater. 5, 9951002 (2006).

27. Neel, N. et al. Controlled contact to a C60 molecule. Phys. Rev. Lett. 98,065502 (2007).

28. Venkataraman, L. et al. Single-molecule circuits with well-defined molecularconductance. Nano Lett. 6, 458462 (2006).

29. Li, Z. et al. Conductance of redox-active single molecular junctions: anelectrochemical approach. Nanotechnology18, 044018 (2007).

30. Balashov, T., Takacs, A. F., Wulfhekel, W. & Kirschner, J. Magnon excitationwith spin-polarized scanning tunneling microscopy. Phys. Rev. Lett. 97,187201 (2006).

31. Kuch, W. et al. Magnetic dichroism study of the valence-band structure ofperpendicularly magnetized Co/Cu(111). Phys. Rev. B 57, 53405346 (1998).

32. Pietzsch, O., Kubetzka, A., Bode, M. & Wiesendanger, R. Spin-polarizedscanning tunneling spectroscopy of nanoscale cobalt islands on Cu(111).Phys. Rev. Lett. 92, 057202 (2004).

33. Tersoff, J. & Hamann, D. R. Theory of the scanning tunneling microscope.Phys. Rev. B 31, 805813 (1985).

34. Diekhoner, L. et al. Surface states of cobalt nanoislands on Cu(111). Phys. Rev.Lett. 90, 236801 (2003).

35. Lippel, P. H., Wilson, R. J., Miller, M. D., Woll, C. & Chiang, S. High-resolutionimaging of copperphthalocyanine by scanning-tunneling microscopy.Phys. Rev. Lett. 62, 171174 (1989).

36. Atodiresei, N. et al. Design of the local spin polarization at the organic-ferromagnetic interface. Phys. Rev. Lett. 105, 066601 (2010).

37. Moresco, F. Manipulation of large molecules by low-temperature STM: modelsystems for molecular electronics. Phys. Rep. 399, 175225 (2004).

38. Temirov, R., Lassise, A., Anders, F. B. & Tautz, F. S. Kondo effect by controlledcleavage of a single-molecule contact. Nanotechnology19, 065401 (2008).

39. Joachim, C., Gimzewski, J. K., Schlittler, R. R. & Chavy, C. Electronictransparence of a single C60 molecule. Phys. Rev. Lett. 74, 21022105 (1995).

40. Oka, H. et al. Spin-dependent quantum interference within a single magneticnanostructure. Science 327, 843846 (2010).

41. Arnold, A., Weigend, F. & Evers, F. Quantum chemistry calculations formolecules coupled to reservoirs: formalism, implementation and application tobenzenedithiol. J. Chem. Phys. 126, 174101 (2007).

42. Heinrich, B. W. et al. Direct observation of the tunnelling channels of achemisorbed molecule. Phys. Chem. Lett. 1, 15171523 (2010).

43. Rosa, A. & Baerends, E. J. Origin and relevance of the staggering in one-dimensional molecular metals. A density functional study ofmetallophthalocyanine model dimers. Inorg. Chem. 31, 47174726 (1992).

44. Rosa, A., Ricciardi, G., Baerends, E. J. & van Gisbergen S. J. A. The opticalspectra of NiP, NiPz, NiTBP, and NiPc: electronic effects of meso-tetraaza

substitution and tetrabenzo annulation. J. Phys. Chem. A 105, 33113327 (2001).45. Brede, J. et al. Spin- and energy-dependent tunneling through a single moleculewith intramolecular spatial resolution. Phys. Rev. Lett. 105, 047204 (2010).

46. Breit, G. & Wigner, E. Capture of slow neutrons. Phys. Rev. B 49,519531 (1936).

47. Grimme, S. Semiempirical GGA-type density functional constructed with along-range dispersion correction. J. Comp. Chem. 27, 17871799 (2006).

AcknowledgementsThe authors thank O. Hampe, J. Kortus, K. Fink, S. Boukari, Xi Chen, M. Alouani,R. Mattana, J. van Ruitenbeek and P. Seneor for useful communications. The authors alsoacknowledge support from the Deutsche Forschungsgemeinschaft (WU 349/3-1 andSPP1243), the Center for Functional Nanostructures, the FrenchGerman University, theAlexander vonHumboldtfoundation,and theAgenceNationale de la Recherche(ANR-06-NANO-033-01) as well as from the Yamada Science Foundation and the Asahi GlassFoundation.

Author contributionsS.S. and W.W. conceived and designed the experiments. S.S., Y.N., T.K.Y. and An.B.performed the experiments. S.S., Y.N. and An.B. analysed the data. Al.B. and F.E. designedand performed the calculations. M.B. and E.B. provided purified molecules. S.S., Al.B.,T.K.Y., F.E., M.B., E.B. and W.W. co-wrote the paper. All authors discussed the results andcommented on the manuscript.

Additional informationThe authors declare no competing financial interests. Supplementary informationaccompanies this paper at www.nature.com/naturenanotechnology. Reprints andpermission information is available online at http://npg.nature.com/reprintsandpermissions/.

Correspondence and requests for materials should be addressed to W.W.

NATURE NANOTECHNOLOGY DOI: 10.1038/NNANO.2011.11 ARTICLES

http://www.nature.com/naturenanotechnologyhttp://npg.nature.com/reprintsandpermissions/mailto:[email protected]://www.nature.com/doifinder/10.1038/nnano.2011.11http://www.nature.com/naturenanotechnologyhttp://www.nature.com/naturenanotechnologyhttp://www.nature.com/doifinder/10.1038/nnano.2011.11mailto:[email protected]://npg.nature.com/reprintsandpermissions/http://www.nature.com/naturenanotechnology

Documents

Giant Magneto Resistance Through a Single Molecule