3 S ep 2 00 8 Nickel on Lead , Magnetically Dead or Alive ?

Two atomic layers of Ni condensed onto Pb films behave, according to anomalous Hall effect measurements, as magnetic dead layers. However, the Ni lowers the superconducting Tc of the Pb film. This has lead to the conclusion that the Ni layers are still very weakly magnetic. In the present paper the electron dephasing due to the Ni has been measured by weak localization. The dephasing is smaller by a factor 100 than the pair-breaking. This proves that the Tc-reduction in the PbNi films is not due magnetic Ni moments. PACS: 75.20.Hr, 73.20.Fz, 74.45.+c, 74.78.Fk When a thin superconductor is condensed onto a normal conducting film then the first layers loose their superconductivity. This phenomenon is called the ”superconducting proximity effect”. A similar question arises if one condenses a thin ferromagnetic metal onto a normal metal. If the first layers of the ferromagnet loose their magnetism they are denoted as ”magnetic dead layers” (MDL). The first dead layers were observed almost 30 years ago for two to three atomic layers of Ni condensed onto amorphous Bi films [1]. For Ni layers on top of noble metals there were originally contradicting results. Liebermann et al. [2] observed two dead layers of Ni on Cu and Au while Pierce and Siegmann [3] observed ferromagnetism already in mono-layers of Ni on Cu (by means of spin polarized photo electrons). Kramer and Bergmann [4] investigated the magnetic properties of Ni on the surface of Mg, In, Sn, and the noble metals Cu, Ag and Au by means of the anomalous Hall effect (AHE). They observed between two and three dead Ni layers on the polyvalent substrate, while Ni on top of the noble metals showed ferromagnetism already for the first Ni mono-layer. However, the electronic properties of the Ni appeared to be modified for the first two to three Ni layers because the AHE had the wrong (positive) sign in this range of Ni thickness. Meservey et al. [5], [6] used spin-polarized tunneling to investigate the proximity effect of the ferromagnetic metals. They observed about three dead layers of Ni on Al. A number of theoretical papers [7], [8], [9], [10], [11], [12], [13] investigated the question of a magnetic proximity effect. 1 The occurance of magnetic dead Ni layers is still a question under investigation [14], [15] while double layers of PbNi and other pairs of superconductor and ferromagnetic metal experienced a new interest in the superconducting proximity effect [16]. But in this paper we want to address a claim made by Moodera and Meservey (MM) [17], [18], [19], [20], [21], [22] about the properties of single and double layers of Ni on Pb. MM increased the sensitivity in their investigation of PbNi double layers by using a 9nm thick Pb film as part of 14MHz oscillator. The frequency of the oscillator changed by about 60kHz when the Pb film made a transition from superconducting to normal state. They observed that the deposit of 0.4nm of Ni onto the Pb substrate reduced the transition temperature of the Pb below 4.2K. A similar effect can be produced by the deposition of Fe. However, the pair breaking effect of Fe is about 80 times stronger than that of Ni. MM gave their results the following interpretation: The Ni atoms on top of the Pb films do not completely loose their moments, even for the smallest coverages, and their magnetic scattering dephases (depairs) the superconducting Cooper pairs, even at their smallest thickness of 0.2 atomic layers of Ni. MM did not try to give a value for this reduced moment. In this paper we revisit the PbNi system. We have measured the magnetic scattering of Ni using the method of weak localization (quantum interference). It is well known and discussed below that the pair-breaking mechanism in superconductivity and the dephasing in weak localization are in many aspects identical. There is, however, an experimental difficulty in measuring weak localization in superconducting Pb films since the magnetoresistance is overshadowed by the Azlamazov-Larkin fluctuations [23]. Therefore we use only very thin Pb films between 1 and 10 atomic layers which are condensed onto a Ag film. Then the proximity effect suppresses the superconductivity of the Pb layers. The experimental procedure is the following. A Ag film with a thickness of 35 atomic layers is quench condensed onto a quartz plate at He temperature in an UHV of better than 10 torr. The Ag film is covered in different experiments with Pb films whose thicknesses lie between one and five atomic layers. Then the Pb film is covered in several steps with Ni. The first Ag film is chosen for three reasons: (i) Even for quenched condensation it is not possible to condense a homogeneous film of a few atomic layers of Pb onto a quartz plate. This requires a homogeneous conducting metal film of sufficient thickness as a substrate. (ii) The Ag film suppresses the superconductivity of the extremely thin Pb film. (iii) Ni on top of Ag shows magnetism already for a mono-layer of Ni. Therefore the observation of MDLs of Ni on AgPb can only be due to the Pb film in between the Ag and the Ni. (It also proves that there are no holes in the thin Pb film). We use two experimental methods to investigate the magnetic properties of the AgPbNi multi-layers, (i) the anomalous Hall effect and (ii) weak localization. In Fig.1 the anomalous Hall resistance Ryx (B) is plotted as a function of B for dPb = 2 atomic layers and different Ni thicknesses in a perpendicular magnetic field. The AHE curves can be extrapolated to zero magnetic field, yielding R yx. This AHE resistance R 0 yx is plotted in Fig.2. It measures the magnetization perpendicular to the film plane in zero magnetic field. The surprising result is that the thickness of the MDL is (almost) independent of the Pb thickness. Even