Superpoissonian shot noise in organic magnetic tunnel junctions

(1)

Superpoissonian shot noise in organic magnetic tunnel junctions

Juan Pedro Cascales,¹Jhen-Yong Hong,²Isidoro Martinez,¹Minn-Tsong Lin,^2,a) Tomasz Szczepanski,³Vitalii K. Dugaev,³Jozef Barnas,⁴and Farkhad G. Aliev^1,b)

1Dpto. Fisica Materia Condensada C3, Instituto Nicolas Cabrera (INC), Condensed Matter Physics Institute (IFIMAC), Universidad Autonoma de Madrid, Madrid 28049, Spain

2Department of Physics, National Taiwan University, Taipei 10617, Taiwan and Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei 10617, Taiwan

3Department of Physics, Rzeszow University of Technology, al. Powstancow Warszawy 6, 35-959 Rzeszow, Poland

4Faculty of Physics, Adam Mickiewicz University, ul. Umultowska 85, 61-614 Poznan, Poland and Institute of Molecular Physics, Polish Academy of Sciences, ul. Smoluchowskiego 17, 60-179 Poznan, Poland (Received 24 October 2014; accepted 25 November 2014; published online 8 December 2014) Organic molecules have recently revolutionized ways to create new spintronic devices. Despite intense studies, the statistics of tunneling electrons through organic barriers remains unclear. Here, we investigate conductance and shot noise in magnetic tunnel junctions with 3,4,9,10-perylene- teracarboxylic dianhydride (PTCDA) barriers a few nm thick. For junctions in the electron tunneling regime, with magnetoresistance ratios between 10% and 40%, we observe superpoissonian shot noise. The Fano factor exceeds in 1.5–2 times the maximum values reported for magnetic tunnel junctions with inorganic barriers, indicating spin dependent bunching in tunneling. We explain our main findings in terms of a model which includes tunneling through a two level (or multilevel) system, originated from interfacial bonds of the PTCDA molecules. Our results suggest that interfaces play an important role in the control of shot noise when electrons tunnel through organic barriers.V^C 2014 AIP Publishing LLC. [http://dx.doi.org/10.1063/1.4903739]

Organic elements in electronic devices have some advantages over inorganic ones, like the ability to chemi- cally adjust their electronic properties, their mechanical flexibility, and the capability to form self-assembled layers.

Exploring the unique properties of the organic world to improve and create new functionalities in spin related optics, electronics, and memory elements has been attract- ing considerable attention in the past decade.^1–7 Organic spintronics may lead to unique devices, for instance, organic light emitting diodes (OLEDs) based on magnetically controlled luminescence.⁸ A key limiting factor for the operation of these and related devices is their signal to noise ratio. Thus, the investigation of noise sources in organic tunnel junctions and spin valves is of fundamental and technological interest, as the noise ultimately deter- mines their practical applications.

Low frequency noise and shot noise (SN) measurements have been systematically used to characterize the electronic transport mechanisms in inorganic spintronics.^9–16 On the other hand, noise in organic-based devices, which could have 1/f and shot noise contributions, remains poorly under- stood. For example, 1/f noise measurements have been used to determine device quality,¹⁷ or transport features in graphene-based devices (including one or several layers).¹⁸ In another study, the 1/f noise and DC leakage measurements were used as a diagnostic tool for OLED reliability in a pro- duction line.¹⁹ Current 1/f noise measurements have been also used to identify individually contacted organic molecules.^20,21

Earlier noise measurements in organic spintronic devices were carried out at large applied voltages, where the 1/f noise is dominant so the role of shot noise could not be deter- mined. Apart from a technological view, precise knowledge of SN can provide a valuable information on electron correlations near the interfaces with organic barriers, especially in the regime of direct tunneling. In fact, the role of interfaces remains one of the central issues in organic spintronics.²²

We analyze the tunneling statistics of organic magnetic tunnel junctions (O-MTJs) by measuring shot noise, known to be an excellent tool for investigating the correlations and other details of electron tunneling, well beyond the capabilities of transport measurements.^23–33 Shot noise originates from the discrete nature of charge carriers, therefore, unlike thermal noise, its contribution to low frequency noise survives down to low temperatures. The normalized shot noise (or Fano factor F) indicates²³ whether the tunneling is uncorrelated (poissonian, F¼ 1), anti-bunched (sub-poissonian, typically due to negative correlations,F < 1) or bunched (super-poissonian, typically due to positive correlations,F > 1).

We have investigated the conductance and shot noise of O-MTJs with PTCDA molecular barriers in the direct tunneling regime.³⁴In contrast to MTJs with inorganic barriers,^10,11 tunneling through molecular barriers showssuper-poissonian shot noise which additionally depends on the relative alignment of the electrodes’ magnetization. Our observations are qualitatively accounted for within a model based on spin dependent electron tunneling through an interacting two-level (or multi-level) system.

The layer sequence of the studied O-MTJs is:

NiFe(25 nm)/CoFe(15 nm)/AlO_x(0.6 nm)/PTCDA(1.2–5 nm)/

AlO_x(0.6 nm)/CoFe(30 nm). The structure was deposited onto a glass substrate, and prepared in a high-vacuum

a)Electronic mail: mtlin@phys.ntu.edu.tw

b)Author to whom correspondence should be addressed. Electronic mail:

farkhad.aliev@uam.es

0003-6951/2014/105(23)/233302/4/$30.00 105, 233302-1 VC2014 AIP Publishing LLC

APPLIED PHYSICS LETTERS 105, 233302 (2014)

This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:

140.112.101.91 On: Mon, 02 Feb 2015 02:40:27

(2)

environment with a base pressure lower than 10⁸mbar. The metallic layers were deposited by sputtering at an Ar working pressure of 5 10³mbar. The PTCDA layers were grown by thermal evaporation at 10⁸mbar, with a deposition rate of 0.1 nm/s. Thin AlO_xbuffer layers were grown between the PTCDA layer and both ferromagnetic layers by partially oxi- dizing Al in oxygen plasma for 5 s. XPS measurements of the Co/PTCDA interface have revealed that the addition of a buffer layer prevents the hybridization between both layers,³⁵ preserving an effective spin injection into the organic spacer.

Figure 1(a) shows a diagram of the investigated O-MTJs.

The PTCDA molecules have been found to lie essentially flat over both AlOx/Co or Co, with a tilt angle of 126 3.³⁵

The voltage noise was measured using a cross- correlation technique, described elsewhere.^10,11 The correct calibration of our setup has been confirmed by independent studies.^14,15The current noise power in the absence of correlations is Poissonian (full shot noise) and is given by SI¼ 2eI, where I is the average current and e the electron charge. The voltage full SN is then Sf ull¼ 2eIR²_d, with Rd

being the dynamic resistance obtained from the correspond- ingI – V curves. The experimental SN, Sexp, is obtained by fitting a Gaussian peak to the histogram of the part of the spectra independent of frequency (see the supplementary material³⁶). The Fano factor F is then calculated as F¼ Sexp=S_full.

Figure 1(b) shows that the resistance of the junctions increases exponentially with the thickness of the PTCDA barrier. This indicates that the PTCDA layer acts as a barrier in the single-step tunneling regime.³⁸ Further proof can be found in Ref.36, where the conductancevs. temperature is compared to hopping transport models. Figure 1(c) shows the tunneling magnetoresistance (TMR) for three different temperatures in a 2 nm PTCDA O-MTJ, where the parallel (P) and antiparallel (AP) magnetic alignment of the electrodes are indicated by arrows. The TMR decreases when the bias reaches 100 mV (see Fig.1(d)). Figure1(d)also presents the differential conductance in the P and AP states as a function of the bias voltage at T¼ 0.3 K for a 4 nm PTCDA

O-MTJ. We found that the magnetic tunnel junctions with PTCDA barriers were more robust than conventional inorganic MTJs, and typically did not experience dielectrical breakdown as readily. Out of 14 samples studied, only 3 have degraded during multiple bias sweeps up to 500 mV.

The experimental SN and Sfull and Fano factor at T¼ 0.3 K for the 2 nm PTCDA junction from Fig. 1(c) are shown in Figs. 2(a) and2(b) for the P and AP states. Figs.

2(c) and2(d)show similar graphs for a sample with a 5 nm thick PTCDA barrier. As can be seen, the F factor ranges from F¼ 1 at low voltages to F ’ 2 at higher voltages. All the O-MTJ samples measured displayed a qualitatively similar variation of the Fano factor with the bias voltage. The shot noise could be obtained for voltages up to a few tens of mV only. The maximum voltage for which the shot noise is measured corresponds to the energy at which the 1/f noise becomes dominant and obscures the frequency independent part of the noise spectrum. The spectra could be obtained up to 100 kHz, but a filtering arising from the sample capaci- tance (dependent on the PTCDA thickness) allowed shot noise measurements only between 1–10 kHz. The appear- ance of 1/f noise restricted SN measurements in all the studied samples, especially in the AP state.

Figure3(a)presents the average saturation value of the Fano factor in the P state for the samples which presented the frequency-independent spectra. Figure 3(a) also shows the variation of TMR with the PTCDA thickness. Control junctions, with only a 1.2 nm AlOxlayer, show TMRs below 1%, and a metallic-like electron transport (see Ref.36). This points to diffusive electron transport, for which the theory²³ predicts the Fano factor equal to 1/3. Thus, control measurements prove that the super-poissonian SN is due to the PTCDA barriers. Our O-MTJs with PTCDA thicknesses between 1.2 and 5 nm show relatively high TMR and superpoissonian tunneling statistics with the Fano factor approach- ing 2, indicative of co-tunneling or tunneling with bunching.

Eight O-MTJs of different barrier thicknesses, from four sample sets, have shown qualitatively similar SN values (Fig.3(a)).

FIG. 1. (a) Sketch of the sample’s lay- ered structure. Diagram of the PTCDA molecule.³⁷The molecules lay almost in plane, with a 128 tilt angle.³⁵ (b) Experimental dependence of the resistance on the PTCDA thickness. (c) TMR curves at different temperatures for a sample with 2 nm of PTCDA. (d) Dependence of the TMR and differential conductance on the bias voltage in the P and AP states for a 4 nm PTCDA O-MTJ.

233302-2 Cascales et al. Appl. Phys. Lett. 105, 233302 (2014)

140.112.101.91 On: Mon, 02 Feb 2015 02:40:27

(3)

A number of electron tunneling mechanisms (Kondo effect,³⁹ co-tunneling,^40,41 and others^42,43) are capable of producing super-poissonian SN, relevant mostly for small quantum dots. Inelastic electron tunneling spectroscopy (IETS) spectra from our samples reveal multiple features (peaks) which could have an electronic and/or vibrational origin (see Ref.36). The observed SN has been accounted for in terms of the approach developed by Belzig,²⁹extended to spin dependent transport. The corresponding model is based on tunneling through a spin dependent, two-level system (or multi-level system in a more general case), with remarkably different tunneling rates. The statistics of the transport process are described as a sum of independent Poissonian events during which a group or “bunch” ofn electrons transfer inde- pendently.²⁹ This arises from the difference in tunnel rates between the two levels and the transfer of one or more ele- mentary charges in each process leads to an enhanced noise.

Details of the model and description will be presented elsewhere. In Fig.3(b), we show the calculated (lines) and experimental (points) TMR and Fano factor in the P and AP states as a function of the parameter b, which describes the spin asymmetry in tunneling rates.

Physically, the two or more levels with different cou- plings can have their origin in localized states arising from interfacial bonds between the PTCDA molecules and the AlO_x buffer layers. The following arguments suggest that the localized states in the model have an interfacial nature:

(i) the exponential dependence of the tunneling resistance on PTCDA thickness (Fig. 1(b)) including the metallic character of the conductance when only the AlO_x buffer layer is present (see Ref.36); (ii) a lateral size of the junctions larger than a micron, for which the influence of Coulomb blockade is minimized. We believe that the main role of the AlO_xlayers in the superpoissonian shot noise is providing the two (or multi) levels localized at the AlO_x/ PTCDA interface. Therefore, the tunneling process takes place through two parallel channels with statistics controlled by interfacial two-level systems. The origin of the interfacial states could be a charge neutrality level,⁴⁴ or gap states,⁴⁵ which appear due to the alignment of the energy levels at metal/organic interfaces.⁴⁶ The bias dependence of the interfacial density of states could explain the suppression of the Fano factor at large voltages (Fig.2(b)).

FIG. 3. (a) Maximum Fano factor in the P state and the zero bias TMRvs the PTCDA thickness, shown to be well above the range for inorganic barriers. (b) Fit of the theory to experimental values ofF and TMR for the samples with 2 and 5 nm of PTCDA.

The points are obtained from the average saturation value in the Fano factor vs bias plots, and the error bars indi- cate the dispersion of the average.

FIG. 2. Voltage dependence of the noise at T¼ 0.3 K in the P and AP states of the (a) experimental (dots) and expected full shot noise (lines) and (b) Fano factor for a 2 nm PTCDA O-MTJ. (c) and (d) present similar results for a 5 nm PTCDA O-MTJ.

140.112.101.91 On: Mon, 02 Feb 2015 02:40:27

(4)

In conclusion, super-poissonian statistics in tunneling events through the PTCDA molecular barriers have been unveiled by shot noise measurements. A superpoissonian shot noise has been found, which is likely due to localized states originated from interfacial bonds of the PTCDA molecules. For a technological application, the shot noise could be reduced or controlled, for instance, by the growth of dou- ble-barrier⁴⁷ O-MTJs. Challenges for further work include extending the bias range where the shot noise could be investigated and comparing the role of the organic layers in the superpoissonian SN by the study of O-MTJs with different organic layers.

The authors acknowledge support by the Spanish MINECO (MAT2012-32743), UAM-SANTANDER, NANOFRONTMAG-CM (S2013/MIT-2850) and NSC 102- 2120-M-002-005 (Taiwan) Grants. This work was also partly supported by the National Center of Research and Development in Poland in the frame of the EU project Era.Net.Rus “SpinBarrier.”

1V. Dediu, M. Murgia, F. Matacotta, C. Taliani, and S. Barbanera,Solid State Commun.122, 181 (2002).

2Z. H. Xiong, D. Wu, Z. Valy Vardeny, and J. Shi,Nature427, 821 (2004).

3A. R. Rocha, V. Garcia-Suarez, S. W. Bailey, C. J. Lambert, J. Ferrer, and S. Sanvito,Nat. Mater.4, 335 (2005).

4T. S. Santos, J. S. Lee, P. Migdal, I. C. Lekshmi, B. Satpati, and J. S.

Moodera,Phys. Rev. Lett.98, 016601 (2007).

5S. Sanvito,Chem. Soc. Rev.40, 3336 (2011).

6J. S. Jiang, J. E. Pearson, and S. D. Bader,Phys. Rev. Lett.106, 156807 (2011).

7R. Vincent, S. Klyatskaya, M. Ruben, W. Wernsdorfer, and F. Balestro, Nature488, 357 (2012).

8T. D. Nguyen, E. Ehrenfreund, and Z. V. Vardeny, Science 337, 204 (2012).

9E. R. Nowak, R. D. Merithew, M. B. Weissman, I. Bloom, and S. S. P.

Parkin,J. Appl. Phys.84, 6195 (1998).

10R. Guerrero, F. G. Aliev, Y. Tserkovnyak, T. S. Santos, and J. S. Moodera, Phys. Rev. Lett.97, 266602 (2006).

11R. Guerrero, D. Herranz, F. G. Aliev, F. Greullet, C. Tiusan, M. Hehn, and F. Montaigne,Appl. Phys. Lett.91, 132504 (2007).

12J. Scola, H. Polovy, C. Fermon, M. Pannetier-Lecoeur, G. Feng, K. Fahy, and J. M. D. Coey,Appl. Phys. Lett.90, 252501 (2007).

13J. M. Almeida, P. Wisniowski, and P. Freitas,IEEE Trans. Magn.44, 2569 (2008).

14K. Sekiguchi, T. Arakawa, Y. Yamauchi, K. Chida, M. Yamada, H.

Takahashi, D. Chiba, K. Kobayashi, and T. Ono,Appl. Phys. Lett. 96, 252504 (2010).

15T. Arakawa, K. Sekiguchi, S. Nakamura, K. Chida, Y. Nishihara, D.

Chiba, K. Kobayashi, A. Fukushima, S. Yuasa, and T. Ono,Appl. Phys.

Lett.98, 202103 (2011).

16T. Tanaka, T. Arakawa, M. Maeda, K. Kobayashi, Y. Nishihara, T. Ono, T.

Nozaki, A. Fukushima, and S. Yuasa,Appl. Phys. Lett.105, 042405 (2014).

17N. Clement, S. Pleutin, O. Seitz, S. Lenfant, and D. Vuillaume,Phys.

Rev. B76, 205407 (2007).

18A. A. Balandin,Nat. Nanotechnol.8, 549 (2013).

19P. Rocha, H. Gomes, L. Vandamme, D. De Leeuw, S. Meskers, and P. van de Weijer, in22nd International Conference on Noise and Fluctuations (ICNF) (IEEE, 2013), pp. 1–4.

20M. Tsutsui, M. Taniguchi, and T. Kawai,Nat. Commun.1, 138 (2010).

21J. Schaffert, M. Cottin, A. Sonntag, H. Karacuban, C. Bobisch, N. Lorente, J.-P. Gauyacq, and R. Muller,Nat. Mater.12, 223 (2012).

22T. Keevers, A. Danos, T. Schmidt, and D. McCamey,Nat. Nanotechnol.8, 886 (2013).

23Y. Blanter and M. Bttiker,Phys. Rep.336, 1 (2000).

24B. R. Bułka, J. Martinek, G. Michałek, and J. Barnas,Phys. Rev. B60, 12246 (1999).

25Y. Tserkovnyak and A. Brataas,Phys. Rev. B64, 214402 (2001).

26R. Lopez and D. Sanchez,Phys. Rev. Lett.90, 116602 (2003).

27A. Thielmann, M. H. Hettler, J. K€onig, and G. Sch€on,Phys. Rev. B68, 115105 (2003).

28A. Cottet, W. Belzig, and C. Bruder,Phys. Rev. Lett.92, 206801 (2004).

29W. Belzig,Phys. Rev. B71, 161301 (2005).

30F. M. Souza, A. P. Jauho, and J. C. Egues, Phys. Rev. B 78, 155303 (2008).

31A. L. Chudnovskiy, J. Swiebodzinski, and A. Kamenev,Phys. Rev. Lett.

101, 066601 (2008).

32F. G. Aliev, E. Kunnen, K. Temst, K. Mae, G. Verbanck, J. Barnas, V. V.

Moshchalkov, and Y. Bruynseraede,Phys. Rev. Lett.78, 134 (1997).

33N. L. Schneider, J. T. L€u, M. Brandbyge, and R. Berndt,Phys. Rev. Lett.

109, 186601 (2012).

34K.-S. Li, Y.-M. Chang, S. Agilan, J.-Y. Hong, J.-C. Tai, W.-C. Chiang, K.

Fukutani, P. A. Dowben, and M.-T. Lin,Phys. Rev. B83, 172404 (2011).

35J.-Y. Hong, K.-H. Ou Yang, B.-Y. Wang, K.-S. Li, H.-W. Shiu, C.-H.

Chen, Y.-L. Chan, D.-H. Wei, F.-H. Chang, H.-J. Linet al.,Appl. Phys.

Lett.104, 083301 (2014).

36See the supplementary material athttp://dx.doi.org/10.1063/1.4903739for details about: estimation of shot noise, dependence of the conductance with the PTCDA thickness, comparison of the conductance vs. temperature with hopping models, and IETS results.

37D. A. Tenne, S. Park, T. U. Kampen, A. Das, R. Scholz, and D. R. T.

Zahn,Phys. Rev. B61, 14564 (2000).

38J. J. H. M. Schoonus, P. G. E. Lumens, W. Wagemans, J. T. Kohlhepp, P.

A. Bobbert, H. J. M. Swagten, and B. Koopmans,Phys. Rev. Lett.103, 146601 (2009).

39Y. Yamauchi, K. Sekiguchi, K. Chida, T. Arakawa, S. Nakamura, K.

Kobayashi, T. Ono, T. Fujii, and R. Sakano,Phys. Rev. Lett.106, 176601 (2011).

40E. Onac, F. Balestro, B. Trauzettel, C. F. J. Lodewijk, and L. P.

Kouwenhoven,Phys. Rev. Lett.96, 026803 (2006).

41Y. Okazaki, S. Sasaki, and K. Muraki,Phys. Rev. B87, 041302 (2013).

42N. Lambert, R. Aguado, and T. Brandes,Phys. Rev. B75, 045340 (2007).

43G. Kießlich, E. Sch€oll, T. Brandes, F. Hohls, and R. J. Haug,Phys. Rev.

Lett.99, 206602 (2007).

44H. Vazquez, R. Oszwaldowski, P. Pou, J. Ortega, R. Perez, F. Flores, and A. Kahn,EPL65, 802 (2004).

45S. Yogev, R. Matsubara, M. Nakamura, U. Zschieschang, H. Klauk, and Y. Rosenwaks,Phys. Rev. Lett.110, 036803 (2013).

46S. Braun, W. R. Salaneck, and M. Fahlman,Adv. Mater.21, 1450 (2009).

47J. P. Cascales, D. Herranz, F. G. Aliev, T. Szczepanski, V. K. Dugaev, J.

Barnas, A. Duluard, M. Hehn, and C. Tiusan, Phys. Rev. Lett. 109, 066601 (2012).

140.112.101.91 On: Mon, 02 Feb 2015 02:40:27