Specific resistance of Pd/Ir interfaces

Abstract

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...

CPP magnetoresistance of magnetic multilayers: A critical review

“…Co Ni (Ni ) +0.5; +0.75 (2) (Fe) +0. 85 (Fe) +0.76; +0. 90 (3) (Co) 0; +0.57 (2) (Mn) -0.11-(Co) +0.86;+0.94 (6) (Cr) -0.46; -0.71 (2) (Cr) -0.54 (Cr) -0.38;-0.67 (6) (V) -0.77, -0.78 (2) In addition to giving a list of spin-flip crosssections, σ sf , (which determine λ sf ) for various Cuand Ag-based alloys, Monod and Schultz [46] also compared several of them with the momentum (i.e., resistivity) cross-sections that determine λ.…”

“…The good agreement for Pd/Pt stimulated a second double-blind test, with Pd/Ir [85]. For Pd/Ir, the difference in lattice parameters is just over 1% and solubility is opposite to Pt/Pd; the two metals separate except for < 2% Ir solubility in Pd and < 5% Pd solubility in Ir [209].…”

A way to measure electron spin-flipping at ferromagnetic/nonmagnetic interfaces and application to Co/Cu

“…(2) From a combination of experiment and theory, Ref. [20] shows that the other basic interface parameter, the interface specific resistance, is often only weakly sensitive to intermixing when the two metals are close to lattice matched, as are Co and Cu. From these results, we conclude that the sensitivity of δ F/N to interfacial intermixing and roughness is not yet known.…”

CPP-GMR: Materials and Properties

“…[133]. An additional quantitative CPP-MR result is that the thermal conductances of sputtered Pd/Ir interfaces at temperatures from 78 to 295 K were found [134] to agree with the Wiedemann-Franz law to within 10 %, assuming a temperature-independent interface specific resistance for the same samples at 4.2 K [62].…”

“…Early calculations used a basis set with spd and linear muffin-tin orbitals (LMTO) [59][60][61]. Later calculations used spdf and MTO orbitals without linearization [62]. The spdf and MTO results will be given in section "Interface Parameters: g F/N , 2AR …”