13.07.2015 Views

ARPES in strongly correlated 4f and 5f systems: Comparison to the ...

ARPES in strongly correlated 4f and 5f systems: Comparison to the ...

ARPES in strongly correlated 4f and 5f systems: Comparison to the ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

A.J. Arko et al. / Journal of Alloys <strong>and</strong> Compounds 271 –273 (1998) 826 –830 829Fig. 4. <strong>ARPES</strong> spectra at hn540 eV <strong>and</strong> T520 K for UPt3at <strong>the</strong>Fig. 5. Normal emission spectrum of UBe13at hn548 eV. The feature at20.5 eV was observed <strong>to</strong> disappear with 1 Langmuir of oxygen dop<strong>in</strong>g[29], <strong>and</strong> is thus presumed <strong>to</strong> be due <strong>to</strong> <strong>the</strong> surface. Only <strong>the</strong> 6d statesdisplay actual dispersion.<strong>in</strong>dicated po<strong>in</strong>ts <strong>in</strong> <strong>the</strong> Brillou<strong>in</strong> zone. The data, taken at 23TK, showhuge variation with momentum. Note <strong>the</strong> sudden disappearance of peak Aat 3.58, precisely where LDA predicts dispersion above EF. Inset much speculation about <strong>the</strong> possibility of multichannelemphasizes <strong>the</strong> dramatic difference <strong>in</strong> both peak position <strong>and</strong> <strong>in</strong>tensity Kondo phenomena responsible for <strong>the</strong> non-Fermi-liquidbetween 18 <strong>and</strong> 258. UBe13properties, our data cannot shed much light on thiss<strong>in</strong>ce <strong>the</strong> relatively broad, dispersionless nature of <strong>the</strong> <strong>5f</strong>From <strong>the</strong> above one is led <strong>to</strong> <strong>the</strong> conclusion that <strong>the</strong> peak can also result from disorder.electronic structure of both Ce <strong>and</strong> U compounds is We have also obta<strong>in</strong>ed <strong>the</strong> first ever spectra at resonancedom<strong>in</strong>ated by <strong>the</strong> existence of extremely narrow f-b<strong>and</strong>s at for d-Pu us<strong>in</strong>g our newly-commissioned laser plasma light<strong>the</strong> Fermi energy, <strong>and</strong> this already above <strong>the</strong> characteristic source, more fully described <strong>in</strong> [30]. In Fig. 6 (bot<strong>to</strong>mTK. Temperature dependence studies <strong>to</strong> 300 K <strong>in</strong> Ce, Yb, spectrum) we show this data taken at hn5111 eV <strong>and</strong><strong>and</strong> U compounds [6–12] show that <strong>the</strong>se b<strong>and</strong>s are nearly compare it <strong>to</strong> spectra taken at lower pho<strong>to</strong>n energies us<strong>in</strong>gtemperature <strong>in</strong>dependent with no spectral weight loss, a helium resonance lamp. There is recent speculation thatalthough <strong>the</strong>re is broaden<strong>in</strong>g of <strong>the</strong> features [10,12] <strong>and</strong> d-Pu is a heavy fermion material [31]. However, while aloss due <strong>to</strong> truncation by <strong>the</strong> 300 K Fermi function. While sharp <strong>5f</strong>-related feature is <strong>in</strong>deed evident at <strong>the</strong> Fermifor very low-TKmaterials, or, equivalently, for very high energy similar <strong>to</strong> those <strong>in</strong> Ce <strong>and</strong> U heavy fermions, <strong>the</strong>temperatures, <strong>the</strong> PAM <strong>and</strong> SIM results aga<strong>in</strong> converge, <strong>in</strong> PES cross-section would <strong>in</strong>dicate a substantial 6d admixorder<strong>to</strong> underst<strong>and</strong> why <strong>the</strong> f-b<strong>and</strong>s <strong>in</strong> CeSb2 <strong>and</strong> CePt2.2ture for this feature [27]. This is evidenced by <strong>the</strong> fact thatrema<strong>in</strong> strong even above 80 K, one must assume that <strong>the</strong> broad peak at 21.7 eV grows more rapidly with pho<strong>to</strong>nhybridization parameters, as determ<strong>in</strong>ed from <strong>the</strong> SIM, are energy relative <strong>to</strong> <strong>the</strong> near 2EFfeature. This behaviorgrossly underestimated. appears consistent with recent b<strong>and</strong> calculations [32]Very weakly hybridized compounds such as UBe <strong>and</strong> which predict that <strong>the</strong> broad 21.7 eV peak is derived from13UAl2 rema<strong>in</strong> a puzzle. In both cases most of <strong>the</strong> f-<strong>in</strong>tensity nearly localized pure <strong>5f</strong> states, while <strong>the</strong> near-EFpeak isis concentrated <strong>in</strong> <strong>the</strong> moderately narrow <strong>in</strong>tense peak near <strong>in</strong>deed <strong>strongly</strong> f–d hybridized. In its present form, <strong>the</strong><strong>the</strong> Fermi energy (see <strong>the</strong> normal emission UBe13(100) PAM is unable <strong>to</strong> contribute <strong>to</strong> <strong>the</strong> underst<strong>and</strong><strong>in</strong>g of d-Pu.<strong>ARPES</strong> spectrum <strong>in</strong> Fig. 5) with no clear evidence of Prelim<strong>in</strong>ary results from <strong>the</strong> PAM calculations, <strong>the</strong>n,dispersion, while <strong>the</strong> temperature dependence (not shown) suggest that <strong>the</strong> previous discrepancies encountered beisnever<strong>the</strong>less m<strong>in</strong>imal <strong>and</strong> no different from high-TKtween experimental <strong>ARPES</strong> data <strong>and</strong> <strong>the</strong> SIM are substancompounds.The broad feature at 20.5 eV <strong>in</strong> Fig. 5 is also tially dim<strong>in</strong>ished with <strong>the</strong> <strong>in</strong>clusion of <strong>the</strong> lattice. Theof f-symmetry, but is clearly surface-related s<strong>in</strong>ce it rapidly weak temperature dependence, as well as dispersion fardim<strong>in</strong>ishes with surface contam<strong>in</strong>ation [29]. While <strong>the</strong>re is above TKis <strong>the</strong>oretically reproduced, so that PAM is

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!