No abstract

We describe a technique, using the current-perpendicular-to-plane (CPP) geometry, to measure the parameter δ F/N , characterizing flipping of electron spins at a ferromagnetic/non-magnetic (F/N) metallic interface. The technique involves measuring the CPP magnetoresistance of a sample containing a ferromagnetically coupled [F/N] n multilayer embedded within the 20 nm thick central Cu layer of a symmetric Py-based, double exchange-biased spin-valve. To focus on δ F/N , the F-and N-layers are made thin compared to their spin-diffusion lengths. We test the technique using F/N = Co/Cu. Analysing with no adjustable parameters, gives inconsistency with δ Co/Cu = 0, but consistency with our prior value of δ Co/Cu = 0.25 ± 0.1. Taking δ Co/Cu as adjustable gives δ Co/Cu = 33 . 0

From measurements of the current-perpendicular-to-plane (CPP) total specific resistance (AR = area times resistance) of sputtered Pd/Ir multilayers, we derive the interface specific resistance, 2AR Pd/Ir = 1.02 ± 0.06 fΩm 2 , for this metal pair with closely similar lattice parameters. Assuming a single fcc crystal structure with the average lattice parameter, no-free-parameter calculations, including only spd orbitals, give for perfect interfaces, 2AR Pd/Ir (Perf) = 1.21 ± 0.1 fΩm 2 , and for interfaces composed of two monolayers of a random 50%-50% alloy, 2AR Pd/Ir (50/50) = 1.22 ± 0.1 fΩm 2 . Within mutual uncertainties, these values fall just outside the range of the experimental value. Updating to add f-orbitals gives 2AR Pd/Ir (Perf) = 1.10 ± 0.1 fΩm 2 and 2AR Pd/Ir (50-50) = 1.13 ± 0.1 fΩm 2 , values now compatible with the experimental one. We also update, with f-orbitals, calculations for other pairs.In electronic transport with current-flow perpen-dicular to the layer planes (CPP geometry) of a metallic multilayer, the interface specific resistance AR (area A through which the CPP-current flows times the sample resistance R) is a fundamental quantity. In the past few years, measurements of AR have been published for a range of metal pairs [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15]. Special interest focuses upon pairs M1 and M2 that have the same crystal structure and closely the same lattice parameters a o -i.e., Δa/a o ≤ 1%, since AR for such pairs can be calculated with no free parameters. That is, taking a given crystal structure and a common a o as known, the electronic structures for M1 and M2 can be calculated without adjustment using the local density approximation, and then AR M1/M2 can be calculated without adjustment using a modified Landauer formula for either interfaces that are perfectly flat and not intermixed (perfect interfaces), or for interfaces composed of two or more monolayers (ML) of a 50%-50% random alloy (50-50 alloy) [16][17][18][19][20][21]. For all four such pairs (Ag/Au [18,20], Co/Cu [18,20,21], Fe/Cr [17,18], and Pd/Pt [12]) where experimental values of 2AR M1/M2 have been published, Table I shows that the previously calculated values for perfect and 2 ML thick alloyed interfaces of these pairs are not very different, and that the experimental values are generally consistent with both values to within mutual uncertainties.In contrast, when Δa/a o is ~ 10%, the agreement between experiment and theory is only semi-quantitativeexperiment and calculations differ by amounts as low as 50% to more than factors of two [11]. Moreover, a test [11] of decreasing the difference in lattice parameter from ~ 10% for Pd/Cu to ~ 5% for Pd/Ag and Pd/Au gave no improvement in agreement between experiment and theory. A subsequent comparison between calculations and experimental data on residual resistivities of a variety of impurities in different hosts showed that those calculations could be very sensitive to local strains [22]. Given these results, it seemed worthwhile to t...

Current-perpendicular-to-plane magnetoresistance measurements let us quantify conduction electron scattering and spin-flipping at a sputtered ferromagnetic/ferromagnetic (F1/F2 = Co/Ni) interface, with important consequences for CPP-MR and spin-torque experiments with perpendicular anisotropy. We use ferromagnetically coupled [Ni/Co] n Ni multilayers, and Py-based, symmetric double exchange-biased spin-valves (DEBSVs) containing inserts of ferromagnetically coupled [Co/Ni] n Co or [Ni/Co] n Ni multilayers, to derive Co/Ni interface specific resistances AR ↑ Co/Ni = 03 . 0 02 . 0 03 . 0 + − fΩm 2 and AR ↓ Co/Ni = 1.00 ± 0.07 fΩm 2 , and interface spinflipping parameter δ Co/Ni = 0.35±0.05. The specific resistances are consistent with our no-free-parameter calculations for an interface thickness between 2 and 4 monolayers (ML) that is compatible with expectations.

To study spin flipping within the antiferromagnet IrMn, we extended prior Current-Perpendicular-to-Plane (CPP) Giant Magnetoresistance (GMR) studies of Py-based exchange-biased-spin-valves containing IrMn inserts to thicker IrMn layers-5 nm ≤ t IrMn ≤ 30 nm. Unexpectedly, AΔR= A(R AP -R P )--the difference in specific resistance between the anti-parallel (AP) and parallel (P) magnetic states of the two Py layers-did not decrease with increasing t IrMn , for t IrMn ≥ 5 nm, but rather became constant to within our measuring uncertainty. This constant looks to be due mostly to a new, small MR in thin Py layers. The constant complicates isolating the spin-diffusion length, l IrMn sf , in bulk IrMn, but l IrMn sf is probably short, ≤ 1 nm. Similar results were found with FeMn.