Open Access
Add to BibSonomyAbstract
We present a method for high-resolution magnetic imaging at linearly polarized partially coherent x-ray sources. The magnetic imaging was realized via Fourier transform holography. In order to achieve elliptical x-ray polarization, three different filters were designed based on the x-ray magnetic circular dichroism effect. We present proof-of-principle images of magnetic nanostructures and discuss the application of the method for future experiments at free-electron laser sources.
©2010 Optical Society of America
One of the major objectives using x-ray light sources of the fourth generation, namely free-electron lasers (FELs), is the possibility for imaging physical and chemical processes on a femtosecond (fs) timescale with high spatial resolution. Especially magnetization and demagnetization phenomena are currently moving into the focus as the underlying spin dynamics is intrinsically on a sub-picosecond timescale and can be very selectively studied by x-rays [1, 2]. Intense fs x-ray pulses are thus an ideal probe to study fundamentals of spin dynamics. Unfortunately, all currently operating x-ray FEL sources, i.e. FLASH (Hamburg, Germany), and LCLS (Stanford, USA), solely provide linearly polarized x-rays. In contrast, the predominant x-ray imaging methods providing magnetic contrast depend on circularly polarized beams, as they are based on x-ray magnetic circular dichroism (XMCD) as contrast mechanism [3]. Mainly due to this discrepancy magnetic imaging at FEL sources has so far not been achieved. If an FEL-compatible polarizer device generating circularly polarized x-rays would be available in the soft x-ray regime of currently operating FELs, a new avenue for ultra-fast magnetic imaging experiments would be opened. Here we report on the design of such a system and present proof-of-principle imaging experiments on magnetic structures.
Different concepts for producing circularly polarized synchrotron radiation were already developed ten and more years ago, before APPLE-type undulators became common [4, 5]. Circular polarization of soft x-rays (100eV–2keV) is achieved either by using the birefringence for the s- and p-components by transmitting through or reflecting from multilayer structures near the Brewster angle. For the energy region of 640eV–850eV, where the important 2p absorption edges of Fe, Co and Ni are located, a suitable multilayer system was found only recently [6]. Alternatively, the XMCD effect has been employed to enhance one helicity component of the x-ray beam relative to the orthogonal component [7]. Today, both concepts are mainly applied in polarization state analysis, since the figure of merit TP 2, with the filter transmission T and degree of polarization P, is very limited.
In the present study, we demonstrate the feasibility of magnetic imaging with linearly polarized soft x-rays at the Co L3-edge. We combine the lensless imaging method of soft x-ray Fourier transform holography (FTH) [8] with Co thin-film polarizers. Especially in respect to FEL applications, we believe this tandem being a well suited candidate for magnetic imaging. As both sections, the polarizing and the imaging part, are exploiting the XMCD effect at the same resonant transition, the photon energy is simultaneously optimized for high contrast.
We have developed and tested polarizers operating in transmission geometry. The polarizers consist of thin magnetic films with a magnetization component in the photon propagation direction. Due to the XMCD effect, the transmission rates T ± for circularly polarized x-rays with positive or negative helicity are different. For this reason, linearly polarized light, which can be perceived as a superposition of right and left circularly polarized light, acquires an elliptical polarization. The XMCD contrast P = (T + − T −)/(T + + T −) and thereby the degree of polarization behind the filter varies with the scalar product of the film magnetization M and the wave vector of the incident light k and is thus maximized for a parallel arrangement and decreases with the cosine of the angle φ between both vectors. In particular, our transmission polarizers consist of magnetized Co thin films or Co/Pd multilayers deposited by magnetron sputtering on 100nm thick Si3N4 membranes. We investigated three different types of filters (Fig. 1). The Co/Pd multilayer system of polarizers A and B were designed to exhibit an out-of-plane magnetization and were used in normal incidence. Polarizer C consists of an in-plane magnetized Co layer and was operated at an incidence angle of 45°.
Fig. 1. (a) Schematic of the beamline setup. The linearly polarized light was elliptically polarized by either polarizing unit A, unit B, or unit C and then incident on the samplemask structure. The hologram was recorded with a CCD camera. (b) For polarizer B a sliced electromagnet was used in order to saturate the magnetic film in the out-of-plane direction. (c) Polarizer C was magnetized in the in-plane direction by two permanent magnets. (d) The holography mask was processed into the gold film (yellow) on a Si3N4 membrane (gray) held by a silicon frame (blue). The magnetic film sample (magenta) was deposited on the opposite membrane side.
The expected transmission and polarization rates of the filters were calculated from the circular dichroism Δβ in the imaginary part of the refractive index [7, 9]: n ± = 1 − δ ± − iβ ± with β ± = β ± Δβ and δ ± = δ ± Δδ, where β and δ denote the absorption and the dispersion for unpolarized light, respectively. The magnetically dichroic contribution in the dispersion Δδ vanishes at the photon energy for a desired maximum Δβ. For a layer with thickness d and the incidence angle θ, the transmission rates are:
For in-plane magnetized systems φ equals θ, for out-of-plane magnetization φ = π/2 − θ. For calculation at the photon wavelength λ for the Co L3-edge, we used Δβ Co = −1.8 × 10−3 and β Co = 6.7 × 10−3 for Co [10] and β Pd = 1.7 × 10−3 for Pd. The total transmission of the multilayer systems was determined by the product of the single layer transmissions. The absorption of the substrate, seed and cap layers was not taken into account.
In principle, it is desirable to optimize the filters in a way that the figure of merit TP 2 reaches a maximum. In the case of the multilayers, this is achieved by (i) keeping the fraction of Co per layer as high as possible, since Pd does not show the XMCD effect and absorbs x-rays without polarizing the light and (ii) optimizing the number of repeats by still maintaining the desired properties of the multilayer. A pure Co filter is only optimized by tuning the layer thickness and the angle of incidence (Fig. 2).
Fig. 2. The theoretically calculated transmission rates T ± (thin green lines) using Eq. (1), the mean transmission T = (T + + T −)/2 (thick green line), degree of polarization P (red), and figure of merit 10 × TP 2 (blue) in dependence on the number of multilayer repeats for the multilayer system used for polarizers A (a), and B (b), respectively. Analogous plots for a pure Co layer (polarizer C) (c) in dependence on the layer thickness (fixed incidence angle 45°) and (d) in dependence on the incidence angle (fixed layer thickness 40nm). Calculations are made for λ at the Co L3-edge. The arrows point to the design properties of the particular polarizer.
For the multilayer structure of polarizer A the Ta(15 Å)/Pd(30 Å) seed layer and the Pd(20 Å) cap layer were sputtered at 3mTorr base pressure and the [Co(5 Å)/Pd(5 Å)]40 bilayer was sputtered at 20mTorr base pressure. The multilayer was designed for a maximum remanence in the out-of-plane direction by a combination of the multilayer composition and control of interfacial pinning centers via the sputtering parameters [Fig. 3(a)] [11, 12]. This allows the operation of the polarizer without any externally applied magnetic field, which makes it very compact and easy to introduce upstream in the vacuum system of a soft x-ray beamline. The filter has to be turned by 180° for changing the sign of the helicity of the light.
Fig. 3. Magnetic hysteresis loops of polarizers A (a), and B (b). The out-of-plane film magnetization M normalized to the magnetization in saturation M s is shown in dependence on the external magnetic field H applied in the out-of-plane direction. Polarizer A features full remanence and can be used without any external magnetic field, whereas for Polarizer B an electromagnet is needed. The arrows point to the external fields used in the imaging experiment.
In the case of polarizer B, we chose a multilayer (Pd(200 Å)/[Co(12 Å)/Pd(7 Å)]50/Pd(12 Å), sputtered at 3mTorr) featuring a high total magnetization and thus yielding a high degree of polarization. An external magnetic field of 0.7T was applied by an electromagnet in order to keep the magnetic film saturated [Fig. 3(b)]. The complete saturation, i.e. the absence of any magnetic domain structure, was checked by small-angle scattering measurements at the Co edge of the filter alone. The film magnetization and thereby the helicity of the exiting light were changed by flipping the direction of the external field.
The multilayer polarizers have the advantage of exhibiting an out-of-plane magnetization which is thus parallel to the x-ray beam in a normal incidence geometry. A certain drawback is the resulting relatively high Pd content. In contrast, polarizer C consists of a pure 40nm thick Co film with only 3nm Pd as seed layer and a 2nm Pd cap to prevent oxidation. The in-plane magnetization of the Co film was aligned using a pair of NdFeB permanent magnets generating a magnetic field of 0.1T at the position of the filter. The complete polarizer arrangement was rotated to invert the sign of the polarization.
All experiments were performed at the UE52-SGM undulator beamline at BESSY II, Germany, using the ALICE scattering chamber (Fig. 1). The photon energy was tuned to the Co L3-edge at E = 779.7eV (λ = 1.59nm) with a resolution of E/ΔE > 4000. The x-ray beam diverged with (6 × 1)mrad2 (hor. × vert.) after the beamline focus which was located 35cm upstream of the sample. In between, the polarization filters were placed. The partially polarized x-ray beam was then incident on a magnetic sample. Fourier transform holograms were recorded in the coherently small-angle scattered x-rays with a charge-coupled device (CCD) camera (Princeton Instruments, 2048 × 2048 pixels) 37cm behind the sample. The FTH geometry was realized by generating a suitable object-reference mask lithographically on the samples as described earlier [8]. For each sample, we compare holograms recorded with horizontally polarized light passing through the polarization unit with a data set measured with circularly polarized x-rays generated by the APPLE II undulator. The undulator parameters were tuned to reach a degree of polarization of 90%.
We imaged three different magnetic samples with a similar design [Fig. 1(d)]: Co/Pd or Co/Pt multilayers with perpendicular anisotropy are the magnetic units exhibiting labyrinth domain patterns. These multilayers on Si3N4 membranes were masked by a focused ion beam patterned gold film of 1µm thickness with an aperture defining the field of view (FOV) and a reference pinhole (diameter 60nm–100nm) which is responsible for encoding the phase information into the hologram. If the photon energy is tuned for maximum XMCD contrast and the holograms are properly centered, the real space image of the domain configuration is obtained in the real part of the Fourier-transformed difference of the holograms taken with opposite polarization [13, 14].
The magnetic layer of our first sample was similar to polarizer B. In remanence, the multilayer system [Co(12 Å)/Pt(7 Å)]50 is known to decay into labyrinth domains with either up or down out-of-plane magnetization [15]. Figure 4(a) shows the domain state of our sample imaged via FTH with circularly polarized x-rays generated in the APPLE II undulator. The stripe domains with a mean width of 70nm are clearly resolved. The same sample region imaged using linearly polarized x-rays incident on the polarization unit B results in the image shown in Fig. 4(b). Although the image contrast is reduced compared to Fig. 4(a), the domain configuration is exactly reproduced and clearly resolved. Due to spatial restrictions given by the electromagnet for polarizer B in the existing vacuum chamber, the sample had to be positioned 68cm downstream of the beamline focus. The resulting low flux density of approx. 3 × 105 photons/(s µm2) resulted in long exposure times of 80min for the measurement without polarizer and of 360min with polarizer B.
Fig. 4. Images of the magnetic domain pattern of the first (a), (b) and second sample (c), (d) achieved via FTH imaging and taken with either circularly polarized light from an undulator or linearly polarized x-rays passing through a polarizer. The FOV diameter is 1.5µm in both cases. The yellow boxes in panel (c) mark the pillar position of the second sample’s prepatterned substrate.
Our second test sample consists of a [Co(5.5 Å)/Pd(8 Å)]24 multilayer, sputtered on a prepatterned Si3N4 membrane substrate. The pillar array pattern with a pillar size of (80 × 80 × 35)nm3 (width × depth × height) and a period of 240nm was produced via electron-beam lithography. On top of the pillars, magnetic islands are formed while leaving a continuous magnetic film in the trenches. A detailed description of the system can be found elsewhere [16]. The domain configuration presented in Fig. 4(c) was recorded holographically using circularly polarized x-rays with 25min total exposure at a flux density of 2 × 106 photons/(sµm2). The image was taken in remanence after first saturating the sample at 0.6T and then applying a counter field of −0.1T. Although at this field the islands have still not switched, the film in the trenches shows a domain structure which is pinned at the magnetic islands. While maintaining this magnetic configuration, we then imaged the sample using linearly polarized x-rays incident on polarizer A [Fig. 4(d)]. Obviously, the image quality has deteriorated in comparison to our measurement of the first sample with polarizer B. The reasons are: (i) the amount of Co in the second sample is reduced by 78% compared to the first sample, (ii) the degree of polarization after polarizer A was only 19%, as determined in XMCD measurements, and (iii) the exposure time of 160min was shorter than in the previous case. Apparently, under these conditions we reach a borderline signal to noise level which is just about sufficient to image the domain structure using a linearly polarized source.
The magnetic layer of the third sample consisted of two [Co(4 Å)/Pt(7 Å)]25 bilayers coupled via a Ru(9 Å)/[Co(4 Å)/Ni(10 Å)]2/Co(4 Å)/Ru(9 Å) multilayer. For more information about sample details see [17]. In the present case, the magnetic system was measured in remanence, where it exhibited labyrinth domains with 100nm width similar to the first sample. The measurements with circular polarization generated by the undulator or the polarizer unit C both reveal the identical domain pattern with strong contrast in both cases (Fig. 5). The exposure times were 67min and 200min, respectively, at the same incident photon density quoted above. XMCD measurements of the filter alone yielded 38% polarization contrast.
Fig. 5. Labyrinth domains of the third sample imaged in a 0.8µm FOV via FTH with and without polarizer. Panels (a) and (d) show the image results by Fourier-transforming the difference of the holograms taken with opposite helicity. Additionally, successful image reconstruction was also achieved solely from the holograms taken with positive (b), (e) or negative (c), (f) helicity.
With focus on FEL applications, we also show the holograms stemming from each helicity alone. Destructive single-shot imaging at maximum photon densities is an important mode for FEL imaging experiments. In this mode, it will not be possible to record holograms with both helicities, unless a beam splitter is used to generate two x-ray pulses in a much more complicated experiment. In Fig. 5 we compare the real space images obtained by Fourier-transforming the holograms taken with either positive or negative helicity alone. We achieve successful single-helicity reconstructions as seen in panels (b), (c), (e), and (f). As expected, the image contrast is reversed for opposite x-ray helicity. The images recorded with linearly polarized x-rays plus polarizer feature a lower contrast, a higher noise level and, thus, a lower contrast-limited spatial resolution, but nevertheless the domain structure is clearly resolved.
Based on the photon flux densities used in our synchrotron studies, it is possible to estimate the photon flux needed in a similar single-shot FEL experiment. If an FEL beam is focused down to an easily achievable (10 × 10)µm2 spot size, 1012 photons per pulse will be sufficient to generate magnetic images of samples comparable to the ones presented here, but recorded with a single fs x-ray pulse. For example, the LCLS FEL source is expected to deliver ten times more photons per pulse with 100fs pulse length at the Co L3-edge [18]. Additionally, the FTH method is fully compatible with single-fs-shot destructive imaging, because (i) the photon-in photon-out principle allows to record all information before the sample starts to explode [19], (ii) FTH allows an unambiguous reconstruction of the object retaining the complete wave field, i.e. magnitude and phase, for information formation [20], and iii) the method takes full advantage of the higher degree of coherence at FEL sources.
We conclude that by refining and applying the polarization and imaging concepts demonstrated in the present study, magnetic imaging with < 100fs time resolution and < 50nm spatial resolution will be enabled at FEL sources even before dedicated FEL undulators with full polarization control will become available in the future. As demonstrated, this applies as well to destructive single-shot imaging with one helicity only.
Acknowledgments
We thank Professor Dr. Zabel and his group for making the ALICE scattering chamber available. Focused ion beam lithography was carried out at the Stanford Synchrotron Radiation Lightsource, a national user facility operated by Stanford University on behalf of the U.S. Department of Energy, Office of Basic Energy Sciences.
References and links
1. E. Beaurepaire, J.-C. Merle, A. Daunois, and J.-Y. Bigot, “Ultrafast spin dynamics in ferromagnetic nickel,” Phys. Rev. Lett. 76, 4250–4253 (1996). [CrossRef] [PubMed]
2. C. Stamm, T. Kachel, N. Pontius, R. Mitzner, T. Quast, K. Holldack, S. Khan, C. Lupulescu, E. F. Aziz, M. Wietstruk, H. A. Dürr, and W. Eberhardt, “Femtosecond modification of electron localization and transfer of angular momentum in nickel,” Nat. Mater. 6, 740–743 (2007). [CrossRef] [PubMed]
3. H. A. Dürr, T. Eimuller, H.-J. Elmers, S. Eisebitt, M. Farle, W. Kuch, F. Matthes, M. Martins, H.-C. Mertins, P. Oppeneer, L. Plucinski, C. Schneider, H. Wende, W. Wurth, and H. Zabel, “A closer look into magnetism: Opportunities with synchrotron radiation,” IEEE Trans. Magn. 45, 15–57 (2009). [CrossRef]
4. J. B. Kortright, M. Rice, S. K. Kim, C. C. Walton, and T. Warwick, “Optics for element-resolved soft x-ray magneto-optical studies,” J. Magn. Magn. Mater. 191, 79–89 (1999). [CrossRef]
5. F. Schäfers, H.-C. Mertins, A. Gaupp, W. Gudat, M. Mertin, I. Packe, F. Schmolla, S. Di Fonzo, G. Soullié, W. Jark, R. Walker, X. Le Cann, R. Nyholm, and M. Eriksson, “Soft-x-ray polarimeter with multilayer optics: Complete analysis of the polarization state of light,” Appl. Opt. 38, 4074–4088 (1999). [CrossRef]
6. M. A. MacDonald, F. Schäfers, R. Pohl, I. B. Poole, A. Gaupp, and F. M. Quinn, “A W:B4C multilayer phase retarder for broadband polarization analysis of soft x-ray radiation,” Rev. Sci. Instrum. 79, 025108 (2008). [CrossRef] [PubMed]
7. J. B. Kortright, S.-K. Kim, T. Warwick, and N. V. Smith, “Soft x-ray circular polarizer using magnetic circular dichroism at the Fe L3 line,” Appl. Phys. Lett. 71, 1446–1448 (1997). [CrossRef]
8. S. Eisebitt, J. Lüning, W. F. Schlotter, M. Lörgen, O. Hellwig, W. Eberhardt, and J. Stöhr, “Lensless imaging of magnetic nanostructures by x-ray spectro-holography,” Nature 432, 885–888 (2004). [CrossRef] [PubMed]
9. H.-C. Mertins, F. Schäfers, X. Le Cann, A. Gaupp, and W. Gudat, “Faraday rotation at the 2p edges of Fe, Co, and Ni,” Phys. Rev. B 61, R874–R877 (2000). [CrossRef]
10. H.-C. Mertins, S. Valencia, D. Abramsohn, A. Gaupp, W. Gudat, and P. M. Oppeneer, “X-ray kerr rotation and ellipticity spectra at the 2p edges of Fe, Co, and Ni,” Phys. Rev. B 69, 064407-6 (2004). [CrossRef]
11. M. S. Pierce, C. R. Buechler, L. B. Sorensen, J. J. Turner, S. D. Kevan, E. A. Jagla, J. M. Deutsch, T. Mai, O. Narayan, J. E. Davies, K. Liu, J. H. Dunn, K. M. Chesnel, J. B. Kortright, O. Hellwig, and E. E. Fullerton, “Disorder-induced microscopic magnetic memory,” Phys. Rev. Lett. 94, 017202 (2005). [CrossRef] [PubMed]
12. C. M. Günther, F. Radu, A. Menzel, S. Eisebitt, W. F. Schlotter, R. Rick, J. Lüning, and O. Hellwig, “Steplike versus continuous domain propagation in Co/Pd multilayer films,” Appl. Phys. Lett. 93, 072505 (2008). [CrossRef]
13. A. Scherz, W. F. Schlotter, K. Chen, R. Rick, J. Stöhr, J. Lüning, I. McNulty, C. Günther, F. Radu, W. Eberhardt, O. Hellwig, and S. Eisebitt, “Phase imaging of magnetic nanostructures using resonant soft x-ray holography,” Phys. Rev. B 76, 214410 (2007). [CrossRef]
14. S. Streit-Nierobisch, D. Stickler, C. Gutt, L.-M. Stadler, H. Stillrich, C. Menk, R. Frömter, C. Tieg, O. Leupold, H. P. Oepen, and G. Grübel, “Magnetic soft x-ray holography study of focused ion beam-patterned Co/Pt multilayers,” J. Appl. Phys. 106, 083909 (2009). [CrossRef]
15. O. Hellwig, S. Eisebitt, W. Eberhardt, W. F. Schlotter, J. Lüning, and J. Stöhr, “Magnetic imaging with soft x-ray spectroholography,” J. Appl. Phys. 99, 08H307 (2006). [CrossRef]
16. O. Hellwig, T. Hauet, T. Thomson, E. Dobisz, J. D. Risner-Jamtgaard, D. Yaney, B. D. Terris, and E. E. Fullerton, “Coercivity tuning in Co/Pd multilayer based bit patterned media,” Appl. Phys. Lett. 95, 232505 (2009). [CrossRef]
17. O. Hellwig, C. M. Günther, F. Radu, A. Menzel, W. Schlotter, R. Rick, J. Lüning, and S. Eisebitt, in preparation.
18. “Soft x-ray materials instrument (SXR) summary of technical design,” http://lcls.slac.stanford. edu/sxr/docs/sxr_tdr_webpage_version20090303.pdf.
19. H. N. Chapman, A. Barty, M. J. Bogan, S. Boutet, M. Frank, S. P. Hau-Riege, S. Marchesini, B. W. Woods, S. Bajt, W. H. Benner, R. A. London, E. Plönjes, M. Kuhlmann, R. Treusch, S. Düsterer, T. Tschentscher, J. R. Schneider, E. Spiller, T. Möller, C. Bostedt, M. Hoener, D. A. Shapiro, K. O. Hodgson, D. van der Spoel, F. Burmeister, M. Bergh, C. Caleman, G. Huldt, M. M. Seibert, F. R. N. C. Maia, R. W. Lee, A. Szöke, N. Timneanu, and J. Hajdu, “Femtosecond diffractive imaging with a soft-x-ray free-electron laser,” Nature Phys. 2, 839–843 (2006). [CrossRef]
20. E. Guehrs, C. M. Günther, R. Könnecke, B. Pfau, and S. Eisebitt, “Holographic soft x-ray omni-microscopy of biological specimens,” Opt. Express 17, 6710–6720 (2009). [CrossRef] [PubMed]
