“…The splitting is determined as 267 6 17 meV. Apart from the anomalous sign, the magnitude is comparable to splittings observed for image states at the surfaces of other ferromagnets [13 6 13 meV for Ni(001) [12], 18 6 3 meV for Ni(111) [13], and 57 6 5 meV for Fe(110) [15], all atḠ, and 125 6 24 and 96 6 30 meV atȲ for Co͑1010͒ [14] ]. However, it is not possible to immediately conclude that the Fe(001) image resonance actually has a negative exchange splitting.…”
Section: S Crampinsupporting
confidence: 71%
“…Hence the possibility of a spin splitting (a different energy for majority-and minority-spin levels) which is expected to be a useful probe of surface magnetism. Himpsel [10] has used a simple phase-shift model coupled with the two-band approximation to predict splittings for various ferromagnetic surfaces, and subsequently a number have been recorded experimentally [11][12][13][14][15]. There has also been a detailed theoretical calculation [16] for the Fe(110) surface which has predicted a splitting of 55 meV atḠ (the center of the surface Brillouin zone) for the n 1 image state, twice that given by the two-band model [10] but in excellent agreement with the recently measured value [15] of 57 6 5 meV.…”
We have observed a surprising negative exchange splitting of the n 1 image-potential surface state at Fe(001) using spin resolved inverse photoemission, indicating that the minority-spin level has a lower energy than the majority-spin level. Calculations show the negative sign results from two superimposing effects. A true reverse polarization of the image state, which hybridizes with bulk bands, which is then enhanced by matrix element effects in inverse photoemission. [S0031-9007(96)00790-9] 75.30.Pd, 79.60.Bm The long-ranged Coulombic attraction between an electron in front of a metal surface and the induced image charge gives rise to a hydrogeniclike infinite series of states loosely bound to the crystal at energies just below the vacuum level. These so-called image states have been extensively studied in recent years [1], and have been exploited in applications as diverse as monitoring the growth and morphology of ultrathin metal films [2], as a source of elemental contrast in the scanning tunneling microscope [3], and in the study of electron localization in insulators and at the metal͞dielectric interface [4]. Image states have been experimentally identified primarily through inverse photoemission (IPE) [5-7] and two-photon photoemission [8]. When the crystal lacks a suitable projected band gap to prevent penetration of the image state into the metal, hybridization with bulk states gives rise to image resonances. These are less pronounced but still visible in IPE spectra [9,10].In the case of a ferromagnetic surface the exchange interaction between the image state and the electrons in the crystal depends upon the spin of the electron. Hence the possibility of a spin splitting (a different energy for majority-and minority-spin levels) which is expected to be a useful probe of surface magnetism. Himpsel [10] has used a simple phase-shift model coupled with the two-band approximation to predict splittings for various ferromagnetic surfaces, and subsequently a number have been recorded experimentally [11][12][13][14][15]. There has also been a detailed theoretical calculation [16] for the Fe(110) surface which has predicted a splitting of 55 meV atḠ (the center of the surface Brillouin zone) for the n 1 image state, twice that given by the two-band model [10] but in excellent agreement with the recently measured value [15] of 57 6 5 meV. We report here the first result (both theoretical and experimental) concerning the spin splitting of an image resonance, that which is found at the Fe(001) surface. Remarkably, we find a negative splitting, corresponding to the minority level having lower energy than the majority level. Of all the electronic states in bcc Fe this is the first (either occupied or unoccupied) to be found to have a negative exchange splitting, and points to a complicated relationship between substrate and imagestate magnetism.Our experiments are spin resolved inverse photoemission [17] in which we collect photons emitted during the radiative decay of spin-polarized incident electrons into emp...
“…The splitting is determined as 267 6 17 meV. Apart from the anomalous sign, the magnitude is comparable to splittings observed for image states at the surfaces of other ferromagnets [13 6 13 meV for Ni(001) [12], 18 6 3 meV for Ni(111) [13], and 57 6 5 meV for Fe(110) [15], all atḠ, and 125 6 24 and 96 6 30 meV atȲ for Co͑1010͒ [14] ]. However, it is not possible to immediately conclude that the Fe(001) image resonance actually has a negative exchange splitting.…”
Section: S Crampinsupporting
confidence: 71%
“…Hence the possibility of a spin splitting (a different energy for majority-and minority-spin levels) which is expected to be a useful probe of surface magnetism. Himpsel [10] has used a simple phase-shift model coupled with the two-band approximation to predict splittings for various ferromagnetic surfaces, and subsequently a number have been recorded experimentally [11][12][13][14][15]. There has also been a detailed theoretical calculation [16] for the Fe(110) surface which has predicted a splitting of 55 meV atḠ (the center of the surface Brillouin zone) for the n 1 image state, twice that given by the two-band model [10] but in excellent agreement with the recently measured value [15] of 57 6 5 meV.…”
We have observed a surprising negative exchange splitting of the n 1 image-potential surface state at Fe(001) using spin resolved inverse photoemission, indicating that the minority-spin level has a lower energy than the majority-spin level. Calculations show the negative sign results from two superimposing effects. A true reverse polarization of the image state, which hybridizes with bulk bands, which is then enhanced by matrix element effects in inverse photoemission. [S0031-9007(96)00790-9] 75.30.Pd, 79.60.Bm The long-ranged Coulombic attraction between an electron in front of a metal surface and the induced image charge gives rise to a hydrogeniclike infinite series of states loosely bound to the crystal at energies just below the vacuum level. These so-called image states have been extensively studied in recent years [1], and have been exploited in applications as diverse as monitoring the growth and morphology of ultrathin metal films [2], as a source of elemental contrast in the scanning tunneling microscope [3], and in the study of electron localization in insulators and at the metal͞dielectric interface [4]. Image states have been experimentally identified primarily through inverse photoemission (IPE) [5-7] and two-photon photoemission [8]. When the crystal lacks a suitable projected band gap to prevent penetration of the image state into the metal, hybridization with bulk states gives rise to image resonances. These are less pronounced but still visible in IPE spectra [9,10].In the case of a ferromagnetic surface the exchange interaction between the image state and the electrons in the crystal depends upon the spin of the electron. Hence the possibility of a spin splitting (a different energy for majority-and minority-spin levels) which is expected to be a useful probe of surface magnetism. Himpsel [10] has used a simple phase-shift model coupled with the two-band approximation to predict splittings for various ferromagnetic surfaces, and subsequently a number have been recorded experimentally [11][12][13][14][15]. There has also been a detailed theoretical calculation [16] for the Fe(110) surface which has predicted a splitting of 55 meV atḠ (the center of the surface Brillouin zone) for the n 1 image state, twice that given by the two-band model [10] but in excellent agreement with the recently measured value [15] of 57 6 5 meV. We report here the first result (both theoretical and experimental) concerning the spin splitting of an image resonance, that which is found at the Fe(001) surface. Remarkably, we find a negative splitting, corresponding to the minority level having lower energy than the majority level. Of all the electronic states in bcc Fe this is the first (either occupied or unoccupied) to be found to have a negative exchange splitting, and points to a complicated relationship between substrate and imagestate magnetism.Our experiments are spin resolved inverse photoemission [17] in which we collect photons emitted during the radiative decay of spin-polarized incident electrons into emp...
“…Further elucidating results about the spin dependence of IS are expected in the future, may be not only for ferromagnetic surfaces but also for spin-orbit-split IS on high-Z materials [45]. Table 1 Summary of experimentally determined exchange splittings of image states (n = 1 unless otherwise noted) on ferromagnetic surfaces from spin-resolved inverse photoemission (SR-IPE), two-photon photoemission (2PPE), polarization-dependent 2PPE (PD-2PPE), and spin-resolved 2PPE (SR-2PPE) Sample k k DE ex (meV) Technique Reference fcc Ni(0 0 1) C None SR-IPE [26] fcc Ni(0 0 1) C <35 2PPE [34] fcc Ni(0 0 1) C 13 ± 13 SR-IPE [27] fcc Ni(1 1 0)-c(2 · 2)-S C 32 ± 13 SR-IPE [28] fcc Ni(1 1 1) C <40 2PPE [35] fcc Ni(1 1 1) C 18 ± 3 SR-IPE [18] hcp Co(0 0 0 1) C <30 2PPE [36] hcp Co(0 0 0 1) on W(1 1 0) C 78 ± 7 SR-IPE [31], this work hcp Co(1 0 1 0) Y 125 ± 24 (n = 1 À ) SR-IPE [29,30] …”
Section: Discussionmentioning
confidence: 99%
“…For nickel surfaces, as expected, the smallest values have been determined: 13 ± 13 meV for Ni(0 0 1) [27], 32 ± 13 meV for Ni(1 1 0)-c(2 · 2)-S [28], and 18 ± 3 meV for Ni(1 1 1) [18]. Cobalt surfaces, however, exhibit the largest exchange splittings: 125 ± 24 and 96 ± 30 meV for n = 1 À and n = 1 + states on Co(1 0 1 0), respectively, [29,30], and 78 ± 7 meV for Co(0 0 0 1) [31]. On iron surfaces, 57 ± 5 meV was observed for Fe(1 1 0) [32] and À67 ± 17 meV for Fe(0 0 1) [33].…”
“…This behavior has been observed before, but only at a zone boundary. Spin-polarized inverse photoemission measurements of Co(1010) have recently found evidence for a large (0.6 eV) symmetry splitting of the n 1 image state at theȲ zone boundary [32]. In our experiments, we have not been able to resolve symmetry split image states at the superlattice zone boundary.…”
Indium can be grown on Si(111) in a 4 3 1 pattern that contains rows of In atoms spaced 13.3 Å apart that have quasi-one-dimensional electronic structure. This ordered array of metallic wires produces an image-induced surface state series. We have measured the dispersion of the most tightly bound ͑n 1͒ image state band and found it to be unconventional because it falls below the free electron parabola perpendicular to the In atom rows. The most straightforward explanation for this is that the electrons feel the surface corrugation potential produced by the rows of In atoms. We were able to infer the form of the potential from our measurements. [S0031-9007 (99)08625-1] PACS numbers: 73.20.At, 73.20.Dx Indium can be grown on Si(111) in a 4 3 1 pattern [1], that contains long rows of In atoms [2-6] that are either 2 or 3 atoms wide [3] with a center-to-center separation of 13.3 Å. The surface reconstruction is complex and still under active study. Candidate structures conveniently divide into two categories. There are those that place the In atoms on a largely unreconstructed Si(111) surface [3,7,8] and those that require the uppermost Si layer to reconstruct [6]. The electronic structure of the 4 3 1 phase has also been the subject of intense study. Recent photoemission [9], inverse photoemission [10-12], and STM [4] studies have provided compelling evidence that the 4 3 1 phase is both quasi-1D and metallic. It is the latter property that sets the 4 3 1 system apart from other quasi-1D systems like the insulating Si(111)-M͑3 3 1͒ (where M ͕Li, Na, K, and Ag͖) overlayer systems [13]. Both photoemission [9] and inverse photoemission [10][11][12] have detected flat bands perpendicular to the atom rows and a Fermi level crossing in single domain samples at ഠ0.6 GX parallel to the atom rows. Consequently, the experimental evidence suggests that the 4 3 1 system may contain the smallest known metallic wires in existence. The possibility of integrating these naturally occurring wires into practical electronic devices remains a dream. However, recent developments in lithography [14] have demonstrated that it should be possible to continue to reduce the size of electronic devices down to the atomic or molecular limit where the wires could be used as atomic scale interconnects.In the course of our study of the 4 3 1 system [10-12], we discovered an image-induced surface state located 0.67 6 0.15 eV below the vacuum level at theḠ point. Image-induced surface states are spatially confined by both the surface barrier and the long-range image potential [15][16][17]. What makes these states particularly interesting is that the image potential is associated with electrons in an array of ordered metallic wires that are only a few atoms wide. In fact, the surface structure resembles an atomic scale diffraction grating with a translational symmetry that is quite unlike that of the low index metal surfaces where image states have been extensively studied [18]. Consequently, the form of the surface potential may also be different, and...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.