X-ray Detected Magnetic Resonance: First evidence of forced precession of polarized orbital components

17-02-2006

X-ray Magnetic Circular Dichroism (XMCD) is inherently an element specific and orbital selective experimental technique that has greatly contributed to our understanding of magnetism. X-ray Detected Magnetic Resonance (XDMR) is a novel spectroscopy in which XMCD is used to probe the resonant precession of the magnetization caused by a microwave pump field hMW orthogonal to the static bias field H0.

  • Share

XDMR was measured on exciting the Fe K-edge in a thin film of Yttrium Iron Garnet (YIG) grown by liquid phase epitaxy along the (111) direction of a Gadolinium Gallium Garnet (GGG) substrate [1]. This experiment carried out on the ESRF beamline ID12 was performed in the longitudinal geometry, i.e. the wavevector kX of the incident, circularly-polarized X-rays being set parallel to the bias field H0. The XDMR signal was recorded in the X-ray fluorescence excitation mode. The microwave power could be increased up to 30 dBm at 9.5 GHz. Given the narrow FMR linewidth of the YIG film (Hfwhm = 3.64 Oe ), resonant pumping occurred in a non-linear foldover regime.


Fig. 1: Experimental geometry of the XDMR experiment.

 

The XDMR signal was detected as modulation side-bands of the X-ray macrobunch repetition frequency F = 710.084 kHz. The magnitude of the XDMR signal was peaking ca. 20 dBV above the noise floor. The real and imaginary parts of the spectrum confirmed the expected inversion of the XDMR signal when the helicity of the incident X-ray beam was changed from Left to Right. After proper renormalization, the small differential cross-section: XDMR ≈ 1.34·10-5 yielded a critical precession angle of ≈ 3.5° for the precession of the magnetisation at the Fe sites.

Since the effective operators accounting for XMCD at a K-edge are exclusively of orbital origin, the measured Fe K-edge XDMR signal produces clear evidence of the forced precession of magnetically polarized orbital components. This experiment led us to the conclusion that, in YIG, there should be no dynamical quenching of the magnetic orbital polarization components.

The impact of combining these two spectroscopies could be large and offers a new capability to study dynamics of orbital magnetization.

Principal Publication and Authors
[1] J. Goulon, A. Rogalev, F. Wilhelm, N. Jaouen, C. Goulon-Ginet, G. Goujon, J. Ben Youssef and M.V. Indenbom, JETP Letters, 82, 791-796 (2005).