Two Curie temperatures in a single iron thin film

Abstract

Due to size effecf the decrease of the Curie temperature of a ferromagnetic film with its thickness is now well established on both experimental and theoretical points of view [I]. However, the Curie temperature variation with the thickness was always considered as continuous when the thickness was varied from n lo n+l atomic planes. In other words, the Curie temperature was assumed to only depend on the quantity of deposited material. We have shown that in the particular case of thin Fe films grown on Ir(100) by MBE, two magnetic transitions are observed in a thin Fe film with a fractional number of deposited monolayers (ml) [I]. The Curie temperahlres (Tc) were determined by both Mossbauer spectroscopy and ac susceptometry. For an integer number of Fe atomic planes, we observed only one peak in the susceptibility measurement, corresponding to Tc=60K for 4 ml and Tc=14OK for 5 ml, as shown in fig. la.However, for a series of sample with a Fe thickness varying from 4 to 5 ml, we did not observe a unique peak located between 60 and 140K as usually observed, but two peaks, one located near 60K and the other near 140K. as shown in fig. Ib. ,. I.'-'--1 _ _~ . ., ... InIan8 Solans 6 rlans : susceptibility measuremen@ showing the Fimre Ib : same measurements variation of the Curie temperature for an integer with varying the Fe thickness number of Fe planes. between 4 and Splanes.In fact, as the Fe growth on Ir(lO0) is layer by layer, the morphology of a 4 t x ml (O<x<l) thick film is a mixing of areas of 4 ml thick and areas of 5 ml thick. If we assume that the magnetic coherence length is smaller than the 4 and 5ml thick areas, these areas seem to be magnetically "independent" and two Curie temperatures should be observed. We performed Monte carlo simulations to justify this assumption ( fig. 2), and we actually found for a sufficiently largc system (128x64 atoms) that the 4 and 5 ml thick areas had their proper Curie temperature. Fimre 2 : Monte Cod0 simulation showing the presence of 2 magnetic transitions in LI single layer, corresponding to the areas of 4 and 5 ml thick. ? 4 8 ? 2 + 5In order to confirm this behavior, we pcrformcd similar cxpcriments on the FeN(001) system. This system is a model one since the growth of Fe on V(oO1) layer (free of oxygen) is perfectly layer by layer up to 5 atomic planes. Indeed, RHEED oscillations were observed during the growth and no damping of the RHEED intensity was detected during the growth of the first 5 planes (fig. 3). The fact that the V buffer layers were free of oxygen was also cmcial since we observed that the Fe growth mode drastically changed on V(OO1) surface usually contaminated with 0. A fimt series ofexpcrimcnts were performed on superlattices (SLs). A Curie temperature near 150K was observed on samples with 4 ml thick Fe layer, and near 320K for 5 ml. For a 4.5 ml thick Fe layers sample, these former two contributions were observed in SQUID measurements -uncorrected from the para-and diamagnetism of MgO-( fig. 4). Finally, X-Ray magnetic circular dichroism (XM...