Inelastic electron tunneling spectra of MgO-based magnetic tunnel junctions with different electrode designs Volker Drewello,* Markus Schäfers, Oliver Schebaum, Ayaz Arif Khan, Jana Münchenberger, Jan Schmalhorst, Günter Reiss, and Andy Thomas Thin Films and Physics of Nanostructures, Bielefeld University, 33615 Bielefeld, Germany Received 14 August 2008; published 12 May 2009 MgO-based magnetic tunnel junctions with up to 230% tunnel magnetoresistance ratio at room temperature and up to 345% at 13 K are prepared. The lower electrode is either exchange-biased or free, while the top electrode is free or an exchanged-biased artificial ferrimagnet, respectively. Additionally, a pseudo-spin-valve hard-soft switching design with two unpinned electrodes is used. Inelastic electron-tunneling spectra for each of these systems show a strong variation in the zero-bias anomaly with a reduced peak for some of the junctions. At voltages around 200 mV additional structures are found, which are not known from junctions with lower magneto resistance, such as alumina-based junctions. We discuss the spectra for the different electrode types and compare our findings with respect to barrier material and magnetoresistance ratio. DOI: 10.1103/PhysRevB.79.174417 PACS numbers: 73.40.Gk, 73.43.Qt, 75.47.-m, 75.70.Cn I. MOTIVATION Magnetic tunnel junctions MTJs with MgO as a crystal- line barrier have been predicted to show very large tunnel magnetoresistance TMR ratios. 1,2 Recently, TMR ratios larger than 1000% at low temperature have been shown by Lee et al. 3 Nevertheless, the TMR ratio still significantly decreases if higher temperatures or voltages are applied; the room-temperature TMR ratio of the above system is about 500%, limiting the applicability of those systems. One rea- son for the decreasing TMR values is intrinsic excitations within the junctions which can be studied by inelastic electron-tunneling spetroscopy IETS. IETS is a well-established method to characterize nonmagnetic-tunnel junctions 4–6 and was applied to MTJs as well. 7–9 This technique has not only a resolution that is lim- ited only by the intrinsic temperature driven energy broaden- ing of the spectra. The bias-voltage range of the spectra is also only limited by the breakdown voltage of the junctions typically in the range of a few volts. 10 It is much simpler than laterally resolved methods in terms of sample prepara- tion. Furthermore, it is also closer to applications as it pro- vides information about MTJs that could be used as the base for reconfigurable magnetic logic, 11 magnetic sensors, or magnetic random-access memory. 12 As indicated by the name, IETS can in principle reveal all inelastic processes in which electrons take part in the tunnel- ing process. An overview can be found in Ref. 13. It is es- pecially possible to excite and identify phonons of the barrier 14 and the electrodes 15 as well as magnons in ferro- magnetic materials. 16 Another prominent feature in IET spec- tra is the zero-bias anomaly. In the dI / dV characteristics a sharp dip at zero bias up to a few mV is usually found which results in large peaks in the IET spectrum. In nonmagnetic-tunnel junctions this effect was discovered by Wyatt 17 and has been attributed to single-magnetic impurities. 18,19 A qualitative study of scattering at such im- purities, however, has proven to be difficult. 20,21 In MTJs the zero-bias anomaly has always been found since IETS was first applied to MTJs by Moodera et al. 7 Recently, also struc- tures at bias voltages higher than 200 mV have been discussed. 22,23 They are of interest because they are presum- ably connected to the coherent tunneling process which is the base of the high TMR ratios of crystalline MgO barriers. Here, we measured IET spectra of several tunnel junc- tions, including MgO-based MTJs and alumina-based sys- tems. We will show differences and similarities of these sys- tems, especially with respect to different electrode types in MgO systems and the different barrier materials. Since the growth of the tunnel barrier is crucial in preparing high TMR MTJs we will describe the layer stacks of the different samples. We will compare our findings to results found in literature and discuss the similarities and specific differences. II. PREPARATION The magnetic tunnel junctions are prepared in a magne- tron sputter system with a base pressure better than 1  10 -7 mbar. We used different layer stacks—an overview is given in Table I. The stacks are sputtered on top of a thermally oxidized 50 nm SiO 2  silicon 100 wafer. Stack 1 is a typical system with MgO barrier and Co-Fe-B elec- trodes. Stack 2 incorporates a pinned artificial ferrimagnet AFi as the top electrode. Hard-soft-switching is used to get an antiparallel state in stack 3 usually called pseudo-spin- valve. Layer stacks 1—3 are annealed after sputtering for 60 min in a magnetic field of 6500 Oe. This activates the exchange bias and initiates the crystallization of the MgO barrier. Layer stack 1 is annealed at 648 K, stack 2 at 623 K, and stack 3 at 673 K. The different annealing temperatures are chosen to get highest TMR ratios at room temperature RT and good magnetic separation in the antiparallel state of the two electrodes at low temperatures. Spectra for the different samples are taken and evaluated. The strength of different inelastic contributions may depend on the annealing tem- perature. The described approach gives the opportunity to compare the limiting factors for each of the layer stacks. Alternatively, the annealing temperature could haven been identical for all samples. Then the different evolution of in- PHYSICAL REVIEW B 79, 174417 2009 1098-0121/2009/7917/1744176 ©2009 The American Physical Society 174417-1