— physics pp. 1091–1095
Structure and magnetism of Ni/Ti multilayers on annealing
SURENDRA SINGH1,∗, SAIBAL BASU1 and P BHATT2
1Solid State Physics Division, Bhabha Atomic Research Centre, Mumbai 400 085, India
2Department of Science and Technology, Link¨oping University, Norrk¨oping, Sweden
∗Corresponding author. E-mail: surendra@barc.gov.in
Abstract. Neutron reflectometry study has been carried out in unpolarized (NR) and polarized (PNR) mode to understand the structure and magnetic properties of alloy for- mation at the interfaces of Ni/Ti multilayers on annealing. The PNR data from annealed sample shows a noticeable change with respect to the as-deposited sample. These changes are: a prominent shift of the multilayer Bragg peak to a higher angle and a decrease in the intensity of the Bragg peak. The PNR data from annealed sample revealed the formation of magnetically dead alloy layers at the interfaces. Changes in roughness parameters of the interfaces on annealing were also observed in the PNR data.
Keywords. Neutron reflectometry; multilayers and interfaces.
PACS Nos 61.05.fj; 68.35.Ct; 68.37.-d
1. Introduction
Ni/Ti multilayers have been studied extensively as ideal candidates for neutron op- tics [1,2] as supermirrors, polarizers, monochromators, etc., because they have an excellent contrast factor for thermal and cold neutrons. Ni/Ti bulk alloys also have good corrosion resistance and are of technological interest [3,4]. A recent study on Ti/Ni multilayer shows a drastic change in structure and magnetic properties of the sample at an annealing temperature of 300◦C as compared to the as-deposited sample [5]. In this paper we report neutron reflectometry (NR) data, both in un- polarized and polarized mode, from Ni/Ti multilayers of different thickness as well as same samples annealed at 300◦C. Layer thickness and their densities were ex- tracted from NR measurements. Polarized neutron reflectivity (PNR) experiments gave detailed magnetic moment depth profiles of the as-deposited as well as an- nealed samples. This attempt is to understand the nature of alloying at the Ni/Ti interfaces as a function of annealing.
2. Experimental details
Ni/Ti multilayers with bilayer thicknesses (Ni + Ti) 100 ˚A and 80 ˚A were deposited on a float glass substrate using e-beam evaporation technique under ultra-high vac- uum conditions at room temperature. First a Ni layer of thickness ∼50 ˚A was grown on a float glass substrate followed by Ti layer of thickness ∼50 ˚A (sample S1) and 30 ˚A (sample S2). Ten such bilayers were deposited for both the samples, in the present case. A thin layer of 20 ˚A of carbon was deposited as a protec- tive capping layer at the air–film interface for both the samples. The designed structures of the multilayer samples can be represented as: for S1, float glass sub- strate/ [Ni50˚A/Ti50˚A]×10/C20˚Aand for S2, float glass substrate/[Ni50˚A/Ti50˚A]× 10/C20˚A. During deposition, thickness of each layer was monitored using a water- cooled quartz crystal thickness monitor. Samples were annealed at a temperature of 300◦C for 1.5 h under a high vacuum of the order of 1×10−6Torr after performing experiments with as-deposited samples.
Neutron reflectometry experiments in both polarized and unpolarized mode were carried out at the polarized neutron reflectometer at Dhruva, India [6]. For PNR measurements the sample was placed in a vertical magnetic field of 0.20 T. The neutron reflectometry data, for both polarized and unpolarized modes have been analysed using a genetic algorithm (GA) basedχ2minimization program [7], which uses a matrix method [8] for generating the reflectivity pattern for a given set of physical parameters of the system. The fitting program also takes into account the interface roughness by introducing a multiplicative exponential factor, as proposed by N´evot and Croce [9].
3. Results and discussion
Figures 1A and 1B show the unpolarized NR pattern from as-deposited multilayer samples S1 and S2 respectively. Open circles in figures 1A and 1B are the experi- mental data from samples S1 and S2. Continuous lines are fit to the measured data.
The average thicknesses of Ni and Ti layers in a bilayer of S1, in the as-deposited sample, obtained from the fit to unpolarized NR measurement, were 58(±2) ˚A and 56(±2) ˚A respectively. For as-deposited S2, thickness of Ni and Ti layers were 55(±3) ˚A and 35(±2) ˚A respectively. In both the samples, the layer densities of Ni as well as Ti layers are close to their bulk values. The multilayers showed Bragg peaks corresponding to their bilayer thickness as well as Kiesig oscillations in un- polarized NR pattern as shown in figure 1. The insets of figures 1A and 1B show the corresponding scattering length density (SLD) profiles of samples S1 and S2, respectively as extracted from the fit to the measured unpolarized NR data from the corresponding samples.
The PNR data, before and after annealing, from samples S1 and S2 are shown in figures 2A and 2B, respectively. We obtained an average magnetic moment of 0.55 µB for Ni atoms in Ni layers for both the as-deposited multilayer samples S1 and S2 from fits to these data (continuous lines in figures 2A and 2B), comparable to its bulk value of 0.60µB. Closed and open circles represent the measured reflectivity data for spin-up and spin-down neutrons. Figures 3A and 3B show the scattering
Figure 1. Unpolarized neutron reflectivity pattern from as-deposited sam- ples S1 (A) and S2 (B). Inset shows the SLD model used to analyse the reflectivity data for the as-deposited S1 (A) and S2 (B).
Figure 2. Polarized neutron reflectivity (PNR) pattern from as-deposited and annealed samples S1 (A) and S2 (B). Closed and open circles are the spin-up and spin-down experimental neutron reflectivity data. Continuous lines are a fit to the measured data.
length densities that have been obtained from the fits to the PNR data shown in figures 2A and 2B, for the as-deposited and annealed samples S1 and S2. Both the nuclear (dash-dot-dot line) and magnetic (continuous line) scattering length density (SLD) profiles obtained from the measured PNR data have been plotted in figures 3A and 3B for S1 and S2 respectively.
The SLD profile and the magnetic moment density profile, obtained from the PNR data for annealed samples, clearly indicate alloy layers at the Ni and Ti interfaces in both the samples S1 (figure 3A) and S2 (figure 3B). An interesting difference between these two samples is the differences in the physical density and
Figure 3. The nuclear (dash-dot-dot line) and magnetic (continuous line) SLD depth profiles obtained from the measured PNR data for the as-deposited and annealed samples S1 (A) and S2 (B).
the thickness obtained for the interface alloy layers. In both the cases, these alloy layers are found to be non-magnetic from the magnetic moment density profile (continuous lines in figures 3A and 3B). In case of sample S1, where the intial thickness of Ni and Ti layers are nearly equal (∼55 ˚A), there is a large asymmetry in the physical density and thickness of the interface layer for Ni on Ti and Ti on Ni. Thickness of Ni on Ti layer is 54 ˚A as compared to Ti on Ni (30 ˚A) in this sample. The SLD of these layers are also quite different. The thicker interface alloy layer has a SLD of 4.09×10−6˚A−2and the other has a SLD of 5.26×10−6˚A−2. In the case of S2 the as-deposited Ti layer was 35 ˚A thick. In this sample there was a difference in thickness of the alloy layers at the two interfaces, i.e Ni on Ti and Ti on Ni. These thicknesses were 20 ˚A and 49 ˚A respectively with almost the same value of SLD (∼6.3 × 10−6 ˚A−2). Such differences due to the difference in the initial layer thicknesses is an interesting feature of asymmetric diffusion of Ni in Ti and Ti in Ni. Further systematic studies are being attempted in this regard.
We obtained an average magnetic moment of 0.25±0.03µB and 0.15±0.03µB per Ni atom for thin central Ni layers sandwiched between the alloy layers for samples S1 and S2 respectively. The SLD for Ni3Ti and NiTi aloys, for their bulk density layers should be 5.76×10−6˚A−2and 2.6×10−6˚A−2respectively. The SLDs that we have obtained from PNR are in the range of 4.09×10−6 ˚A−2 to 6.29 ×10−6
˚A−2. It is difficult to identify a specific alloy composition from the present data.
4. Summary and conclusion
Depth profiles of the sample composition, averaged across the sample plane, were extracted from specular polarized neutron reflectometry measurements. Scatter- ing length density profiles obtained from annealed samples using polarized neutron
reflectivity measurements shows asymmetric nature, which is different for two sam- ples. PNR measurement, on annealed samples, also shows reduction in magnetic moment of Ni atoms. The alloy layers at the interfaces are magnetically dead.
References
[1] F Mezei,Proc. Soc. Photo-Opt. Instrum. Eng.83, 10 (1988) [2] M Senthil, P Boni and D Clemens,J. Appl. Phys.84, 6940 (1998) [3] M Bouhki, A Bruson and P Guilmin,Solid State Commun.83, 5 (1992) [4] B M Clemens,Phys. Rev.B33, 7615 (1986)
[5] P Bhatt, A Sharma and S M Chaudhari,J. Appl. Phys.97, 043509 (2005) [6] Saibal Basu and Surendra Singh,J. Neutron Research44(2), 109 (2006) [7] Surendra Singh and Saibal Basu,Solid State Physics (India)44, 257 (2001) [8] S J Blundell and J A C Bland,Phys. Rev.B46, 3391 (1992)
[9] L N´evot and P Croce,Rev. Phys. Appl.15, 761 (1980)
