Photoelectron microscopy at Elettra: Recent advances and perspectives

Journal of Electron Spectroscopy and Related Phenomena 224 (2018) 59–67

Contents lists available at ScienceDirect

Journal of Electron Spectroscopy and Related Phenomena journal homepage: www.elsevier.com/locate/elspec

Photoelectron microscopy at Elettra: Recent advances and perspectives M. Amati a , A. Barinov a , V. Feyer a , L. Gregoratti a , M. Al-Hada a,b , A. Locatelli a , T.O. Mentes a , H. Sezen a , C.M. Schneider c , M. Kiskinova a,∗ Elettra − Sincrotrone Trieste, Area Science Park, 34149 Basovizza, Trieste, Italy Department of Physics, College of Education and Linguistics, University of Amran, Yemen c Peter Grünberg Institut (PGI-6), Forschungszentrum Jülich, D-52425 Jülich, Germany a

b

a r t i c l e

i n f o

Article history: Received 20 October 2016 Received in revised form 23 June 2017 Accepted 27 June 2017 Available online 8 July 2017 Keywords: ␮-XPS ␮-ARPES ␮-XAS Imaging SPEM PEEM Graphene

a b s t r a c t The complementary capabilities of the Scanning PhotoElectron Microscopes (SPEM) and X-ray PhotoEmission Electron Microscopes (XPEEM), operated at Elettra, in terms of imaging and micro-spectroscopy have opened unique opportunities to explore properties of functional materials as a function of their morphology and dimensions and to follow modifications in their properties during their operation. This paper describes the present performance of SPEMs and XPEEMs at Elettra, illustrated by selected recent studies relevant to graphene science. Ongoing efforts for implementing SPEM set-ups allowing for insitu investigations under realistic operating conditions and PEEM set-up for spin-filtered momentum microscopy are outlined and discussed as well. © 2017 Elsevier B.V. All rights reserved.

1. Introduction The fast progress in the development and implementation of photoelectron spectromicroscopy techniques started in the last decade of 20th century with the construction of low emittance Xray synchrotron machines, providing tunable energy photon beams with very high brightness and variable polarization [1–4]. Thanks to the high photon intensity provided by insertion devices it has been possible to add sub-micrometer lateral resolution to X-ray photoelectron spectroscopy (XPS), which has enabled exploiting chemical, electronic and magnetic inhomogeneities at surfaces and interfaces as well as the properties of individual nanostructured objects. The tunability of the synchrotron light has allowed, by monitoring total or Auger electron yield, to complement XPS with another powerful spectroscopy method, X-ray absorption spectroscopy (XAS). Using synchrotron light and combining XPS and XAS, surface and interface phenomena with various sample probing depths (up to ∼ 10 nm with soft X-rays) can be now investigated with spatial resolution falling in the range between that of optical and electron microscopy.

∗ Corresponding author. E-mail address: [email protected] (M. Kiskinova). https://doi.org/10.1016/j.elspec.2017.06.006 0368-2048/© 2017 Elsevier B.V. All rights reserved.

In X-ray photoelectron microscopes high spatial resolution is achieved using full-field or scanning methods. The X-ray photoemission electron microscope (XPEEM) is a prominent example of the first type of instruments. It produces a magnified image of the sample area illuminated by the X-ray beam using electronoptics imaging systems employing electrostatic or electromagnetic lenses. Due to parallel imaging, XPEEM is an excellent tool for the investigation of dynamic processes in real time. In scanning instruments, called scanning photoelectron microscopes (SPEM) the photon beam is demagnified to sub-micrometer dimensions using suitable photon optics. Images are constructed pixel by pixel during X, Y scanning the sample with respect to the X-ray microprobe. The two spectromicroscopy approaches, XPEEM and SPEM, and their applications have repeatedly been reviewed in the past [5–10]. The unique opportunities for exploring surface and interfacial properties of technologically relevant materials through photoelectron spectromicroscopy were endorsed at the late eighties by the founders of Elettra synchrotron laboratory, where currently four photoelectron microscopes are in operation. Although none of these microscopes was invented at Elettra, most of them have undergone technical innovations and incremental improvements in order to reach or define the-state-of-the-art performance. In this work we will illustrate the current capabilities of chemical imaging

60

M. Amati et al. / Journal of Electron Spectroscopy and Related Phenomena 224 (2018) 59–67

thermore, the multi-channel imaging allows the reconstruction of the spectrum from micro-areas selected from the maps, so-called spectro-imaging mode [5,17]. The different type of optics determines the photon energy ranges for each of the two SPEMs. SPEM-1 operates at photon energies lower than ∼100 eV and is very well suited for valence band mapping [13,14], whereas SPEM-2 uses photon energies above 300 eV, being optimized for core level spectroscopy [16,17]. The lateral changes in the valence spectra and the core level intensities and energy shifts are fingerprints of the variations in local electronic structure and composition of the sample under investigation. 2.1. SPEM-1 for -ARPES at spetromicroscopy beamline

Fig. 1. Schematic sketch of a scanning photoelectron microscope (SPEM) set-up. The focused beam and the hemispherical electron analyzer are fixed while the sample is raster scanned. The principle of the multichannel detector covering a selected electron kinetic energy range is outlined.

and micro-spectroscopy using the SPEM and XPEEM instruments at Elettra and briefly review recent results placing the emphasis on graphene. Due to exotic electronic and robust mechanical properties graphene has become attractive for numerous applications, spanning over many domains such as electronics, photonics, energy storage, medicine, etc. [11]. Despite the huge efforts, considerable information has still to be attained in order to understand, predict and control specific properties of several graphene-based materials. In this respect both SPEM and XPEEM experiments have made significant contributions. Recent advances for overcoming the pressure gap limits and prospects for studies in magnetism will be also briefly conferred. 2. Scanning photoelectron microscopes at Elettra laboratory Currently Elettra hosts two SPEMs. Both microscopes utilize suitable photon optics to demagnify the monochromatized photon beam and focus it into a spot of sub-micrometer dimensions on the sample, as schematically illustrated in Fig. 1. One of the microscopes (referred here as SPEM-1) uses refractive near-normal-incidence mirrors, called Schwarzschild objective [12–14] and the other (SPEM-2) uses diffractive Fresnel zone plate (ZP) lenses [15–17]. For both instruments image formation occurs by collecting photoelectrons within a selected energy window, emitted from the specimen raster-scanned with respect to the microprobe. The lateral and spectral resolution of SPEM instruments are independent from each other, the former being determined only by the focusing optics. Both SPEM instruments utilize multichannel detectors in their electron analyzers [18] which reduces substantially the measurement time for chemical imaging. As schematically illustrated in Fig. 1 the number of images collected with a single raster scan is equal to the number of channels, each corresponding to a specific electron kinetic energy within the selected energy window. For instance, summing selected channels, corresponding to the energy window of core level features or background secondary photoelectron signal, chemically specific and topographic information can be obtained, respectively. Along with ‘surface roughness’ information, the topographic maps also are necessary for correction of the topographic artefacts in the chemical images [19]. Fur-

The low photon energy of SPEM-1 is imposed by the multilayer mirror Schwarzschild objective providing the microprobe [20], which makes this instrument attractive for sub-micron angleresolved photoemission spectroscopy (␮-ARPES). This instrument allows extracting high resolution spectroscopic information on the band structure of sub micrometer-sized areas or samples, including also tiny operative micro-devices. Similar SPEMs optimized for ␮-ARPES but using ZP optics are already under operation or in commissioning phase in other synchrotron laboratories [21–23] as well. Successful applications of SPEM-1 at Elettra in studies of 2D, layered and low dimensional materials, such as graphene, transition metal chalcogenides and selenides, were recently demonstrated [24–27]. The main challenges for realization of ␮-ARPES SPEM instruments are: (i) sufficiently high photon flux in a submicron spot on the sample, (ii) stability of the beam focusing during X, Y sample scanning and especially during spectrum acquisition from the selected micro-spot and (iii) necessity to acquire three dimensional (kx , ky , E) band structure maps with spectral quality comparable to that of classical ARPES experiments. In the SPEM-1 instrument, operated at Elettra Spectromicroscopy beamline, the Schwarzschild objectives focus ∼1011 photons s−1 into a submicron sized probe. Varying the photon energy is an advantage in order to extract kz (perpendicular to the sample surface) and optimize the photoemission yield for particular orbital features of interest but unfortunately each type of multilayered mirrors can work only with a selected photon energy, which excludes scanning the photon energy. That is why the beamline is equipped with two Schwarzschild mirror objectives, designed for working at 27 and 74 eV photon energy. The exchange between the two objectives can be performed within one hour without vacuum loss. Band structure mapping is performed with the sample kept fixed and rotating the hemispherical electron energy analyzer around two axes [14]. Thanks to the in-house designed ARPES lens column the electron analyzer collects at each angular position angle-resolved data within an energy window from 0.5 to 4 eV (depending on the electron analyzer pass energy) and an angle of 15◦ . This experimental set-up allows two measurement modes. The first is three-dimensional (kx , ky , E) ␮-ARPES maps from a selected sub-micron spot, comparable to standard ARPES measurements. The second is four-dimensional images (X, Y, k, E) acquired by X, Y scanning of the sample, where in the two-dimensional image each sample spot contains two-dimensional angle-resolved dispersion (k–E). Fig. 2 presents an example for ‘non-standard’ 4D ARPES imaging, obtained for a twisted few layer graphene. The sample, grown on SiC in Ar atmosphere at 1690–1730 K, consists of a few layered graphene flakes, which are not only rotationally misaligned but the layers are also twisted within certain flakes. The averaged over the entire sample 3D ␮-ARPES image reveals multiple branches of Dirac fermions and only thanks to using a microprobe smaller than a single graphene domain, the band structure of each individual domain can be addressed [25]. The 4D image, integrated over E and k of the

M. Amati et al. / Journal of Electron Spectroscopy and Related Phenomena 224 (2018) 59–67

61

Fig. 2. (a-c) 50 × 50 ␮m2 images of a multilayer twisted graphene sample, extracted from a four-dimensional ␮-ARPES dataset corresponding to the green, red and blue detector areas, respectively. (d-f) Detector snapshot spectra extracted from the green, red and blue domains indicated in the (a-c) images. The inset in a) shows the sum in k, E of the 4D ␮-ARPES image. For clarity, the arrows indicate domain to spectra (↓) and spectral region to image (↑) correspondence. The whole 4D dataset required acquisition time of one hour. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

detector with the analyzer tuned close to one of the K points of the graphene band structure (inset in Fig. 2a), is almost uniform and its intensity reflects exclusively the thickness or, more precisely, the number of graphene layers with an azimuthal alignment matching nearly that of the electron analyzer orientation. When we construct the images corresponding to various regions of interests (ROI) of the E-k windows, indicated by colored lines inside the multichannel detector, detailed information about the laterally-resolved domain structure can be obtained. The images corresponding to these ROIs, shown in Fig. 2(a–c), reveal in detail the contrast variations relevant to the domain structure, while the panels (d–f) illustrate the corresponding spectra, integrated from the areas within these selected in (a-c) domains. From the results shown in Fig. 2 we can conclude that the green domain is a single layer having a large twist angle with respect to the underlying layers, so the latter do not contribute to the detector signal at the current E-k position. The blue and red ROI domains are composed of at least twisted bilayers. These domains reveal the interactions between twisted layers, reflected by a distinct gap inside the Dirac fermions branch at E-Ef = −0.8 eV for the blue domain and a small gap at E-Ef = −0.3 eV for the red domain (indicated by yellow arrows). These findings demonstrate the potential of advanced ␮-ARPES SPEM instruments to get complete information studying spatially complex samples, which possess several types of morphological and electronic inhomogeneity.

2.2. SPEM-2 for core level spectroscopy at ESCA microscopy beamline The SPEM-2 at the ESCA Microscopy beamline at Elettra has been operated for already 22 years exploring various types of nanostructured surfaces and interfaces and the related phenomena occurring at mesoscopic length scales [16]. It uses a diffractive

zone-plate as focusing optics in normal photon incidence geometry, which imposes a grazing acceptance angle (∼ 45–50◦ ) of the analyzer. An order sorting aperture (OSA) to cut the undesired higher order diffraction signal is placed between the sample and the zone plate. The distance between the sample and OSA is determined by the photon energy and ZP parameters so using X-rays in the range 350–1200 eV it is rather short (∼ 5–12 mm), which imposes certain limits in sample manipulation and signal detection. The monochromaticity of the ZP optics requires moving the sample when scanning the photon energy in order to preserve the spot size, since the focal distance varies with the photon energy. This requirement impedes easy implementation of micro-XAS in total electron yield − to solve the problem interferometer controlled sample movements have to be implemented [21] but this is not done in our SPEM yet. However, it is possible to vary the probing depth of XPS by selecting the photon energy or comparing the signals measured for different core levels of the same element. The micro-spectroscopy and imaging capabilities of SPEM-2 instrument have been used for tracing the chemical and electronic properties of various materials, following lateral variations as a function of object size and shape, temperature, ambient, applied electric fields, etc. [5,17,28–30]. The best spatial resolution achieved in micro-spectroscopy operation is ∼ 100 nm, whereas in imaging mode features smaller than 100 nm can be readily distinguished. As reported below the spectral resolution is sufficient to probe the quality of a single free standing graphene layer placed on an amorphous carbon support. Free standing monolayer graphene flakes with a high structural quality have attracted intensive research interest. The flakes of micrometer dimensions, obtained by exfoliation methods, usually exhibit multi-thickness morphology and different crystal quality. This makes it extremely challenging to investigate single layer regions alone and probing thickness dependence of the film properties using laterally averaging techniques. Thanks to the sub-

62

M. Amati et al. / Journal of Electron Spectroscopy and Related Phenomena 224 (2018) 59–67

Fig. 3. a) Raw 12.8 × 12.8 ␮m2 SPEM image of a Lacey Carbon support partially covered by graphene flakes, obtained summing all detector channels within a 4.4 eV energy window covering the C 1s region; b) C 1s spectra measured in flake-covered and bare amorphous C support areas. The grey and black bars indicate the energy regions of the channels selected for the two images that were used for obtaining the ‘ratio’ image c); c) ‘Ratio’ image, obtained by dividing the ‘grey bar’ image to the ‘black bar’ one, where the bright areas highlight the presence of high quality graphene flakes; d) Deconvoluted C 1 s spectrum of a free standing flake. Photon energy 650 eV.

micrometer spatial resolution and sufficient spectral resolution provided by SPEM-2, we were able to investigate the free standing areas of single layer flakes supported on Lacey Carbon perforated films. This has opened new opportunities for shedding light on the effect of the support (onto which the graphene is grown or simply placed by transfer methods) on the film properties. As can be expected, using carbon supports is rather challenging because their spectroscopic signal overlaps with that of graphene. In this regard, high spectral resolution has crucial importance to identify unambiguously the graphene covered areas. The samples under investigation were prepared by transferring CVD grown graphene flakes onto a standard electron microscopy grid covered by a Lacey Carbon network. Prior to SPEM analysis, the samples were examined by high resolution scanning electron microscopy (SEM), which verified the thickness of the flakes. The monolayer thick areas were marked for easier location after the transfer into the SPEM, where the samples were heated in UHV at 973 K for 12 h to remove, as much as possible, undesired contaminant species from the graphene surface. Fig. 3a) shows a raw SPEM image obtained by collecting photoelectrons within an energy window covering the entire C 1s spectral region. The contrast here is dominated by the topography of the graphene flakes and the Lacey Carbon support. The great advantage of the multichannel acquisition is that the details of the C 1s spectra allow different C-based species to be identified. Indeed, the two C 1s spectra in Fig. 3b), recorded on the graphene flake (A) and bare support (B), are distinctly different: the B one is indicative for the dominance of amorphous carbon. This difference of the graphene and carbon support C 1s spectral features allowed for obtaining the ‘ratio’ image shown in Fig. 3c, by dividing the image corresponding to C 1s energy window of monolayer graphene (grey bar in Fig. 3b) to the image corresponding to the C 1s energy window of amorphous carbon (black bar in Fig. 3b). In the ‘ratio’ image in

Fig. 3c) the bright areas are the high quality monolayer graphene flakes, whereas the regions of amorphous carbon support with holes appear as a noisy background. Fig. 3d) shows the C 1s core level spectrum measured in the graphene flake region A, where a Doniach-Sunijc´ line profile was used in the fit. The three components correspond to a dominant graphene peak at binding energy (BE) of 284.5 eV and two weaker components at BE of 285.1 and 286.5 eV, attributed to some C-containing contaminant species on the surface. The full width half maximum of the graphene peak of 0.5 eV, limited by the overall energy resolution of the instrument, imply graphene flakes of high quality. The adaptability of the UHV SPEM-2, where focusing optics, sample handling and electron detection are decoupled, has made possible important upgrades and developments for overcoming the pressure gap. The noted above rather short focal distance means that the OSA is rather close to the sample, which impedes to position the analyzer very close to the sample, as is done in classic ambient pressure photoelectron spectroscopes [31–33]. Typically, the set-ups for near-ambient pressure photoelectron spectroscopy are based on a differentially pumped analyzer, that is commercially available nowadays [34]. The challenges and the solutions for near ambient pressure SPEM experiments, developed or tested at Elettra, are reviewed recently in ref. [35]. The first experimental setup available in the SPEM-2 uses a series of gas jet shots at a fixed repetition rate and duration using a computer-controlled pulsed valve [36]. The pulsed jet of gas is emitted by a thin needle, avoiding any interference with the X-ray optic system and electron analyzer. Each pulse generates gas pressure up to a few mbar at the sample without exceeding the pressure limits in the chamber required for the SPEM operation. Using this set-up we performed the first in-situ photoelectron microscopy study of an solid oxide fuel cell operated at 650 ◦ C with CH4 :O2 or H2 :O2 fuel, generating measurable electric current [37,38]. The continuous monitoring by SPEM the Ni 2p core level spectra of the Ni anode during electrochemical operation has provided the necessary information about the lateral evolution of the Ni chemical state and the overpotential, resulting from the current generated by the electrochemical reactions. The obtained results have provided insightful information on the electrocatalytic performance of the cell, which is correlated to the chemical state and morphology of key cell components. Another setup for overcoming pressure gap is placing the sample inside a near ambient pressure cell (NAPC) [39,40]. In this case, the sample is probed with the focused X-ray beam penetrating through a pin-hole, open or sealed by a graphene membrane. Using gaseous ambient one can control the leakage of gas molecules into the main chamber by a proper choice of the pin-hole dimensions. This allows running the whole system in regular operation mode without sealing with graphene membranes and without the necessity of a differentially pumped electron energy analyzer [40]. The recent new design of the NAPC [41], which is an upgrade of the previous version, employs two 500 ␮m diameter pin-holes, where the second allows more photoelectrons emitted from the probed spot to reach the analyzer (see the inset of Fig. 4c). The environmental cells also include heating and biasing the sample for in-situ and in-operando type studies. The performance of the current NAPC version was probed by oxidation and reduction of an unpolished Cu foil with micro-meter sized scratches. The sample mounted inside the cell was primarily annealed at 823 K in 1 mbar H2 ambient to remove contaminant adsorbates and to reduce the native Cu oxide surface layer. Subsequently, the Cu sample was exposed to 0.2 mbar O2 at 623 K for one hour. The measured O 1s, Cu 2p and Cu LMM Auger spectra confirmed that the Cu+1 oxidation state is attained. In a following step the Cu2 O was partially reduced by exposing the sample to 0.2 mbar H2 at 723 K, while acquiring photoemission maps and spectra in

M. Amati et al. / Journal of Electron Spectroscopy and Related Phenomena 224 (2018) 59–67

63

Fig. 4. (a) Raw Cu LMM image obtained after partial reduction of Cu2 O formed on the Cu foil by introducing 0.2 mbar H2 in the cell and heating the foil to 723 K. (b) Processed ‘Cuo /Cu+1 ratio image’, emphasizing the chemical contrast between the two oxidation states (the reduced Cu areas appear brighter). The image is obtained by dividing the ‘grey-dashed bar’ image to the ‘black-dashed bar’ one, corresponding to the Cuo and Cu+1 energy windows, respectively. (c) Normalized Cu LMM spectra representing the reduced and still partially oxidized areas, measured in the locations indicated in image (a). The inset illustrates the recent NAPC design and working principle. Photon energy 1067 eV.

order to control the process. The SPEM image in Fig. 4a) of a partially reduced Cu2 O surface is obtained by collecting the Cu LMM Auger electrons. Regions of different brightness and roughness are clearly visible in the raw image, where the reduced domains appear brighter. The two domains also have different morphologic evolution, the reduced regions appearing with a higher roughness. This indicates that along with changing the chemical state the surface morphology is also drastically modified during the reduction cycle. Fig. 4b) shows the ratio image obtained by dividing the maps of oxidized and reduced Cu measured by properly selecting the energy windows of the metallic and oxide Cu components. It is apparent that the topographic contributions are removed and the chemical contrast is emphasized, i.e. the regions of the reduced state are highlighted. Micro-spectroscopy from selected regions defines in detail the Cu chemical state. In the present case the Cu LMM Auger spectrum has better fingerprints compared to the core level ones, as can be seen comparing the two representative spectra in Fig. 4c), where the solid-line spectrum corresponds to a fully reduced Cu◦ , whereas the dotted one is still dominated by the Cu+1 oxidation state. The metallic state (Cu◦ ) component in the latter spectra may be attributed to coexistence of reduced or oxide islands with dimensions smaller than the SPEM microprobe or a very thin oxide layer allowing electron emission from the metallic state below. 3. X-ray photoelectron emission microscopes at Elettra Synchrotron-based PEEM is a well-established cathode lens microscopy technique with lateral resolution currently surpass-

ing few tens of nanometers. PEEM has a rather long history, which started in the 30’s of the last century and these laboratory-based instruments frequently use UV light sources as a probe. The PEEM also ideally couples with tunable-energy X-ray synchrotron sources and it has been shown in the late 80’s of the last century that collecting secondary emitted electrons, PEEM instruments allow x-ray absorption spectroscopy to be performed in a laterally resolved manner [1,4,6]. Since then the instruments using X-ray are most often called XPEEM. With the subsequent development of XPEEM with bandpass energy filter, electrons emitted from core levels or the valence band could be used for imaging, enabling the technique to achieve sensitivity to the chemical state and electronic structure [6,10]. A few different types of energy-filtered instruments are nowadays commercially available allowing operation in imaging, as well as in diffraction mode [10]. In all modern XPEEMs reciprocal space maps can be obtained probing selected areas of less than 10 ␮m diameter, chosen by the insertion of a suitable aperture in the image plane. Elettra currently operates two different XPEEM instruments, located at the Nanospectroscopy [42] and NanoESCA [43] beamlines. Both instruments are served by a high flux, intermediate energy resolution beamline, which provides photons with variable polarization (linear horizontal, linear vertical and left or right circular) and tunable energy range up to ∼1000 eV, focused into a few micron sized spot onto the sample. These XPEEMs are most often employed for applications in the fields of surface and material sciences, spanning from nanoscale magnetism and dynamical pro-

64

M. Amati et al. / Journal of Electron Spectroscopy and Related Phenomena 224 (2018) 59–67

cesses, such as growth and self-organization, to low-dimensional systems and nanomaterials [6,42,43]. Notably, the polarized x-rays provided by the undulator source make it possible to probe the magnetic state of solid surfaces and interfaces by x-ray dichroism methods. The control over photon polarization and energy is crucial in studying antiferromagnetic and ferromagnetic surfaces using x-ray magnetic linear and circular dichroism (XMLD and XMCD), in which the photon energy is tuned to a certain absorption threshold and the photon polarization is varied in order to enhance the magnetic signal [44]. X-ray dichroism has been successfully applied in XPEEM instruments operated at Elettra, e.g. for studies on exotic skyrmion states in ferromagnetic films [45] or hidden states in artificial spin-ice arrays [46] among others. Recent instrumental developments at the Nanospectroscopy beamline are taking advantage of the time structure of the synchrotron photon bunches to open up the possibility for applying dynamic pump-probe measurements to dichroism studies on time-resolved magnetic phenomena. The instrument operated at the Nanospectroscopy beamline, the so-called Spectroscopic Photoemission and Low Energy Electron Microscope (SPELEEM), combines photon and electron probes. After the introduction of the first prototype in the mid 90’s [47], commercial SPELEEMs have benefited from various advances in instrument design and operation. The optimization of experimental protocols has been crucial to the development of the technique, which has now reached maturity. The second instrument, operated at the NanoESCA beamline, combines the electrostatic PEEM column with a doublehemispherical analyzer [48]. The angle of incidence of the synchrotron beam is 65◦ with respect to the surface normal. This particular geometry provides homogeneous photon illumination to an area of about 15 ␮m in diameter. In this instrument the electrostatic objective lens is kept at a high voltage, leaving the sample close to ground potential. Recently the entrance slit of the analyzer was upgraded, allowing an improvement of energy and k resolution in ␮-ARPES mode up to ∼50 meV and 0.05 Å−1 , respectively. 3.1. SPELEEM studies of complex interfaces and 2D systems The main strength of the SPELEEM resides in its intrinsic multitechnique ability: along with spectroscopic imaging at high lateral resolution, microprobe versions of Low Energy Electron Diffraction (LEED) and Angle-Resolved Photoemission Spectroscopy (ARPES) are readily available. These methods can deliver direct information on the crystal quality and electronic structure. In this manner, the SPELEEM allows characterizing individual, micron-sized crystals both in real and momentum space, addressing issues which could not be tackled using standard laterally averaging techniques. As a result, SPELEEM has found frequent applications in the study of graphene [49,50] and other 2D materials, such as transition metal dichalcogeneides [51]. The availability of LEEM is an essential advantage since the favorable backscattering conditions allows imaging at video rate, an ideal feature for monitoring growth and phase transition processes. Indeed, LEEM studies have played a major role in disclosing the physics of graphene epitaxy on a multitude of crystalline substrates [52,53]. The epitaxial relation between graphene and support is typically determined using LEED [54,55] and, as will be shown in the following, graphene doping and interfacial interactions are instead probed by measuring valence or core level photoelectrons in ARPES or XPEEM experiments. The multi-technique capabilities of the SPELEEM operating at Elettra are well illustrated by our study on graphene on a square lattice support, Ir(001) [56]. A striking feature of graphene on Ir(001) is that distinct physisorbed and chemisorbed phases coexist at room temperature, showing flat and buckled morphology.

The buckled phase (BG) exhibits one-dimensional ripples with periodicity of 2.1 nm, giving origin to a coincidence structure in LEED. The BG phase nucleates during cooling from growth temperature to ambient conditions. At room temperature about 20% of graphene exhibits the BG phase. As revealed by LEEM, the flat and buckled phases organize into alternating, stripe-shaped domains. Since they have slightly different C densities, the arrangement into stripes provides an effective strain relief mechanism compensating the different thermal contraction of film and support. The flat and buckled graphene phases are characterized by subtle differences in the film-substrate interaction. Laterally-resolved measurements of the C 1s core level emission have demonstrated that a small fraction of C atoms in the BG unit cell (∼11%) is chemisorbed to Ir. To determine whether and how such bonds affect the electronic structure of the film, angle-resolved photoemission experiments were carried out. The full ARPES pattern and momentum distribution curves for graphene on Ir(001) are shown in Fig. 5(a)–(c). Note that the Dirac point is shifted above the Fermi level, indicating positive doping. The problem arises whether the BG phase preserves the structure of the ␲ band intact or not. In fact, BG couldn’t be tested using microprobes, due to the small size of its domains, reaching at most a few hundred nanometers in width. An imaging approach was therefore necessary. For this purpose, a dark-field (df) imaging method was developed [57]. In df-PEEM, the off-axis electrons emitted from the graphene ␲ band are used for imaging. Fig. 5(d) shows a df-PEEM image of a graphene flake. The flat (FG) and buckled (BG) phases are indicated by labels. As can be easily seen, the flat graphene regions show up with high intensity. This reflects the high density of states of the Dirac cones close to the Fermi level. Conversely, the BG regions appear completely dark, suggesting that the Dirac cones are here disrupted. The inversion of contrast with the XPEEM image measured at the  point is evident, see Fig. 5(e). This indicates that the BG phase exhibits metallicity, a finding which was also confirmed by ab-initio calculations. Therefore, chemisorption, even in small amounts, profoundly alters the local electronic structure of BG, changing the density of states from semi-metal to metallic type. Indeed, depositing Au at temperatures nearing 870 K can modify permanently the graphene-substrate interactions. The Au deposition monitored using LEEM revealed that, initially, Au adsorbs on the bare Ir regions solely. The Au growth front expands into the graphene flakes only after the graphene-free regions are completely filled. To find out whether Au was intercalated, microprobe–XPS measurements (not shown here) were carried out. They showed that the C 1s emission from graphene is not attenuated upon Au deposition, whereas the Ir 4f appeared strongly attenuated. This clearly indicates that Au is intercalated at the graphene-Ir interface. Consistently, microprobe-ARPES revealed that complete Au buffer layer recovers the free-standing behavior of graphene [58]. Graphene/Ni(111) provides yet another remarkable example of a system where the electronic properties can be tuned by varying the sample temperature. In this case, LEEM and XPEEM experiments demonstrated the reversible formation of a carbidic buffer layer, located under the flakes of rotated graphene [59]. This buffer is grown “from below”, via the segregation of C atoms that were previously dissolved in the Ni bulk. By tuning annealing conditions, the buffer layer may be made continuous, extending on length scales of several microns. Microprobe-ARPES measurements showed that it causes graphene to electronically decouple from the substrate, resulting in the recovery of free-standing-like characteristics of the ␲ band. The same study has provided a second successful application of the df-PEEM technique, used to map contiguous regions of epitaxial and rotated graphene exhibiting different local electronic properties.

M. Amati et al. / Journal of Electron Spectroscopy and Related Phenomena 224 (2018) 59–67

65

Fig. 5. (a) ␮-ARPES pattern (near EF ) of a graphene island on Ir(001) acquired with photon energy of 40 eV (b) Cross section through a graphene Dirac cone along the profile indicated by the red dashed line in (a). (c) Intensity profile as a function of electron energy. (d) Dark-field XPEEM image of a graphene island obtained by imaging photoelectrons at the K point in the diffraction plane. (e) XPEEM image at EF acquired by collecting electrons at normal emission. Reprinted with permission from ref. [56]. Copyright (2013) American Chemical Society. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

It is important to note that XPEEM warrants both elemental and surface sensitivity. These capabilities have allowed us probing the stoichiometry of complex interfaces obtained after low energy ion irradiation of graphene. Exo-species are here implanted into or below the lattice mesh. In a proof-of-principle study, n-type doping patterns were demonstrated in neutrally-doped single-layer graphene grown on Ir(111), after irradiating the film with nitrogen ions through a mask and probing it with ␮-ARPES [60]. XPEEM imaging has shown that the boundary between irradiated and nonirradiated regions is stable upon annealing to 1020 K. In another study, the morphology and spatial distribution of Ar intercalated under graphene on Ir(100) were investigated, specifically addressing the formation of nanobubbles (NB) and their evolution upon high-temperature annealing [61]. Fig. 6(a–f) shows room-temperature LEEM and PEEM images of a graphene flake after 0.1 keV Ar+ ion irradiation and subsequent thermal treatment to 1320 K, illustrating well the high-lateral resolution capabilities of the SPELEEM. In the LEEM image in Fig. 6(a) the image contrast

is purely structural: the bright and grey regions correspond to flat and buckled graphene, the dark patch to the bare Ir surface, respectively. The small black dots, highlighted by the red circles, identify the NBs. Fig. 6(b) shows the corresponding Ar L3 edge XAS-PEEM image, the image intensity being proportional to the Ar concentration. Spectra obtained from inside and outside the red circles are shown in Fig. 6(c). The three resonances, indicated by the arrows, are a clear spectroscopic fingerprint of Ar. The inversion of contrast between the Ar L3 XAS-PEEM image, shown in (d), and the Ir 4f7/2 XPEEM in (e) indicates that the substrate core level emission is screened by Ar. It follows that the Ar is located at the graphene/Ir interface, at variance with softer substrates where it forms bubbles buried in the metal. Notably, the larger Ar aggregates display a lateral size up to few tens of nanometers. We underline the advantages of low energy electron microscopy and diffraction, which are essential probes when tackling structure related issues. These methods were extensively employed when assessing the crystal quality and rotational orientation of

66

M. Amati et al. / Journal of Electron Spectroscopy and Related Phenomena 224 (2018) 59–67

Fig. 7. Experimental setup of the imaging spin filter installed at the PEEM. The quantization axis P is normal to the scattering plane.

3.2. Spin-resolved photoelectron spectroscopy at the NanoESCA beamline

Fig. 6. (a) LEEM image of graphene on Ir(100) after Ar+ irradiation (0.1 kV, 150 s at 1.5 × 10−5 mbar Ar) and subsequent annealing to 1050 ◦ C. (b) X-ray absorption images at the Ar L3 threshold obtained as a difference of PEEM images acquired at photon energies corresponding to the absorption peak and baseline. (c) X-ray absorption spectra from inside and outside the red circles. (d) Ar L3 XAS-PEEM image of another region along with (e) Ir 4f7/2 and (f) C 1 s XPEEM images. Reprinted with permission from ref. [61]. Copyright (2015) American Chemical Society. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

graphene grown on a bimetallic Ni3 Al alloy. Notably, the samples were annealed in oxygen at 520 K, resulting in the intercalation of oxygen and the formation of a thin alumina nanosheet below graphene, which was shown to cover most of the support [62]. Part of the interest in the field of two-dimensional layers is directed towards heterostructures consisting of sheets of different materials. Hexagonal boron nitride (hBN) is a close relative of graphene with a very similar honeycomb lattice. In spite of the similarity in the geometrical structure, the electronic properties of graphene and hBN differ greatly. In a recent study, the two materials were combined in a lateral heterostructure through a growth mechanism involving a single molecular precursor, dimethylamine borane (DMAB) [63]. SPELEEM methods were crucial in demonstrating the presence of laterally separated domains of graphene and hBN with domain sizes reaching 100 nm. Dark-field LEEM and XPEEM were used to identify the two materials structurally and chemically, which were shown to cover the entire surface as a single heterogeneous atomic layer.

A very recent development at the NanoESCA PEEM instrument involves the insertion of 2-dimensional spin polarimeter [42] after the energy filter (see Fig. 7). This new spin filter is based on the spin asymmetry in elastic scattering of electrons from a W(100) target. The image, both in real space and reciprocal space imaging modes, is conserved upon specularly reflecting off the W(100) crystal [64,65]. The quantization axis (P) is oriented normally to the scattering plane (see Fig. 7). The spin sensitivity necessitates a low scattering energy. Therefore, the photoelectrons are decelerated as they arrive at the W(100) crystal, and upon scattering they are accelerated again by an electrostatic lens, and the spin filtered image is formed on the detector in the 90◦ mirror geometry. The spin asymmetry in the image is induced by the spin–orbit interaction in the high-Z tungsten target. When spin polarimeter target is retracted from the beam path, the spin- integrated image can be obtained on the detector in the straight-line geometry. The present experimental setup has been tested by measurements on the spin-resolved three-dimensional Fermi surface of ferromagnetic fcc cobalt [43,66] using synchrotron radiation. The spin filter is already available for the facility users. 4. Conclusions Photoelectron spectromicroscopy has already become a routine microscopic tool at Elettra laboratory for probing exotic properties of matter down to nanoscales. There are continuous efforts for improving the lateral resolution, marching into sub–10 nm range by developing new optical elements for SPEM and by adding aberration correction to XPEEM. Adapting the present SPEMs for experiments at ambient pressures and pushing further their spatial and spectral resolution will open the route to explore chemical states and evolution under working conditions at the scale of individual nano-structures in complex functional systems. Faster and more efficient detection systems should be implemented for pushing the temporal resolution to ps range and spectral resolution to meV in ␮-ARPES. XPEEM instruments can take advantage of recent advances in detector technology. CMOS imaging detectors have superior efficiency and versatility with respect to the channelplates in particular for time resolved experiments, owing to the relative ease of implementing gating and synchronization with external stimuli, light, electric or magnetic fields. Magnetism will continue to be one of the mainstream applications of XPEEMs and we expect that imaging and ␮-APRES experiments will greatly benefit of most recently developed spin filtering approaches.

M. Amati et al. / Journal of Electron Spectroscopy and Related Phenomena 224 (2018) 59–67

A distinct progress is expected with the improved illumination efficiency using the new diffraction limited storage rings. Beneficial effects in terms of spectral, lateral and time resolution are expected. Acknowledgements This work was supported by Sincrotrone Trieste SCpA and Forschungszentrum Jülich. References [1] B.P. Tonner, G.R. Harp, Rev. Sci. Instrum. 59 (1988) 853. [2] G. Margaritondo, F. Cerrina, Nucl. Instrum. Methods A 191 (1990) 26–35. [3] H. Ade, J. Kirz, S. Hulbert, E. Johnson, E. Anderson, D. Kern, Vac. Sci. Technol. A 9 (1991) 1902. [4] E. Bauer, Rep. Prog. Phys. 57 (1994) 895. [5] S. Günther, B. Kaulich, L. Gregoratti, M. Kiskinova, Prog. Surf. Sci. 70 (2002) 187 (and references therein). [6] A. Locatelli, E. Bauer, J. Phys. Condens. Matter 20 (2008) 093002. [7] A. Barinov, P. Dudin, L. Gregoratti, A. Locatelli, T.O. Mentes, M.A. Nino, M. Kiskinova, Nucl. Instr. Methods Phys. Res. A 601 (2009) 195. [8] G. Margaritondo, J. Electron. Spectrosc. Relat. Phenom. 178–179 (2010) 273–291. [9] A.P. Hitchcock, J. Electron. Spectrosc. Relat. Phenom. 200 (2015) 49–63. [10] E. Bauer, Surface Microscopy with Low Energy Electrons, Springer NY, 2014 (and references therein). [11] D. Haag, H.H. Kung, Top. Catal. 57 (2014) 762, http://dx.doi.org/10.1007/ s11244-013-0233-9. [12] F. Cerrina, J. Electron. Spectrosc. Relat. Phenom. 76 (1995) 9. [13] P. Dudin, P. Lacovig, C. Fava, E. Nicolini, A. Bianco, G. Cautero, A. Barinov, J. Synchrotron Radiat. 17 (4) (2010) 445. [14] http://www.elettra.eu/lightsources/elettra/elettra-beamlines/spectromicroscopy/beamline/all.html. [15] H. Ade, J. Kirz, S. Hulbert, E. Johnson, E. Anderson, D. Kern, J. Vac. Sci. Technol. A 9 (1991) 1902. [16] M. Marsi, L. Casalis, L. Gregoratti, S. Günther, A. Kolmakov, J. Kovac, D. Lonza, M. Kiskinova, J. Electron. Spectrosc. Relat. Phenom. 84 (1997) 73. [17] http://www.elettra.eu/lightsources/elettra/elettra-beamlines/escamicroscopy/scanning-photoelectron-microscope-spem.html. [18] L. Gregoratti, A. Barinov, E. Benfatto, G. Cautero, C. Fava, P. Lacovig, D. Lonza, M. Kiskinova, R. Tommasini, S. Mähl, W. Heichler, Rev. Sci. Instrum. 75 (2004) 64. [19] S. Gunther, A. Kolmakov, J. Kovac, M. Kiskinova, Ultramicroscopy 75 (1998) 35. [20] F. Barbo, M. Bertolo, A. Bianco, G. Cautero, S. Fontana, T.K. Johal, S. La Rosa, G. Margaritondo, K. Kaznacheyev, Rev. Sci. Instrum. 71-1 (2000) 5. [21] J. Alvia, et al., J. Phys.: Conf. Ser. 425 (PART 19) (2013) 192023. [22] E. Rotenberg, A. Bostwick, J. Synchrotron Radiat. 21 (2014) 1048–1056. [23] M. Hoesch, et al., Rev. Sci. Instr. 88 (2017) 013106. [24] H. Yuan, et al., Nano Lett. 16 (8) (2016) 4738. [25] V. Kandyba, M. Yablonskikh, A. Barinov, Sci. Rep. 5 (2015) 16388. [26] J. Yin, H. Wang, H. Peng, Z. Tan, L. Liao, L. Lin, X. Sun, A. Koh, Y. Chen, H. Peng, Z. Liu, Nat. Commun. 7 (2016) 10699. [27] Neil R. Wilson, et al., Sci. Adv. 3 (2017) 2. [28] M. Kiskinova, Surf. Interface Anal. 30 (2000) 464. [29] M.K. Abyaneh, L. Gregoratti, M. Amati, M. Dalmiglio, M. Kiskinova, e-J. Surf. Sci. Nanatechnol. 9 (2011) 158. [30] M. Amati, B. Aleman, B. Bozzini, L. Gregoratti, H. Sezen, M. Kiskinova, Surf. Sci. 652 (2016) 20 (and references therein). [31] H. Siegbahn, K. Siegbahn, J. Electron Spectrosc. Relat. Phenom. 2 (1973) 319. [32] R.W. Joyner, M.W. Roberts, Surf. Sci. 87 (1979) 501. [33] H.-J. Ruppender, M. Grunze, C.W. Kong, M. Wilmers, Surf. Interface Anal. 15 (1990) 245.

67

[34] D.E. Starr, Z. Liu, M. Hävecker, A. Knop-Gericke, H. Bluhm, Chem. Soc. Rev. 42 (2013) 5833. [35] A. Kolmakov, L. Gregoratti, M. Kiskinova, S. Günther, Top. Catal. 59 (2016) 448. [36] M. Amati, M. Kazemian Abyaneh, L. Gregoratti, J. Instrum. 8 (5) (2013) T05001. [37] B. Bozzini, M. Amati, L. Gregoratti, M. Abyaneh, M. Prasciolu, M. Kiskinova, Electrochem. Commun. 24 (2012) 104. [38] B. Bozzini, M. Amati, L. Gregoratti, M. Kiskinova, Sci. Rep. (2013), http://dx.doi. org/10.1038/srep02848. [39] A. Kolmakov, D.A. Dikin, L.J. Cote, J. Huang, M.K. Abyaneh, M. Amati, L. Gregoratti, S. Gunther, M. Kiskinova, Nat. Nanotechnol. 6 (2011) 651. [40] H. Sezen, B. Alemán, M. Amati, M. Dalmiglio, L. Gregoratti, ChemCatChem 7 (2015) 3665. [41] H. Sezen, M. Al-Hada, M. Amati, L. Gregoratti, Surf. Interface Anal.- – Spec. Issue Near Ambient Synchrotron Surf. Anal. (2017) ()in press. [42] http://www.elettra.eu/lightsources/elettra/elettra-beamlines/ nanospectroscopy/beamline-description/all.html. [43] http://www.elettra.eu/lightsources/elettra/elettra-beamlines/nanoesca/ specifications/all.html. [44] C.M. Schneider, G. Schonhense, Rep. Prog. Phys. 65 (2002) R1785. [45] O. Boulle, J. Vogel, H. Yang, S. Pizzini, D. Chaves, A. Locatelli, T. Mentes¸, A. Sala, L. Prejbeanu, O. Klein, M. Belmeguenai, Y. Roussigné, A. Stashkevich, S. Chérif, L. Aballe, M. Foerster, M. Chshiev, S. Auffret, I. Miron, G. Gaudin, Nat. Nanotech. 11 (2016) 449. [46] B. Canals, I. Chioar, V. Nguyen, M. Hehn, D. Lacour, F. Montaigne, A. Locatelli, T. Mentes¸, B. Burgos, N. Rougemaille, Nat. Comm. 7 (2016) 11446. [47] Th. Schmidt, S. Heun, J. Slezak, J. Diaz, K.C. Prince, Surf. Rev. Lett. 5 (1998) 1287. [48] C.M. Schneider, C. Wiemann, M. Patt, V. Feyer, L. Plucinski, I.P. Krug, M. Escher, N. Weber, M. Merkel, O. Renault, N. Barrett, J. Electron. Spectrosc. Relat. Phenom. 185 (2012) 330–339. ˜ P. [49] K.R. Knox, S. Wang, A. Morgante, D. Cvetko, A. Locatelli, T. Mentes, M. Nino, Kim, R. Osgood Jr., Phys. Rev. B 78 (2008) 201408. ˜ P. Kim, A. [50] K.R. Knox, A. Locatelli, M. Yilmaz, D. Cvetko, T. Mentes, M. Nino, Morgante, R. Osgood Jr., Phys. Rev. B 84 (2011) 115401. [51] W. Jin, P. Yeh, N. Zaki, D. Chenet, G. Arefe, Y. Hao, A. Sala, T. Mentes, J. Dadap, A. Locatelli, J. Hone, R. Osgood, Jr., Phys. Rev. B 92 (2015) 201409 (R). [52] M.S. Altman, J. Phys. Condens. Mater. 22 (2012) 084017. [53] M. Batzill, Surf. Sci. Rep. 67 (2012) 83. [54] E. Miniussi, M. Pozzo, A. Baraldi, E. Vesselli, R.R. Zhan, G. Comelli, T.O. Mentes¸, ˜ A. Locatelli, S. Lizzit, D. Alfè, Phys. Rev. Lett. 106 (2011) 216101. M.A. Nino, [55] J. Kraus, S. Böcklein, R. Reichelt, S. Günther, B. Santos, T.O. Mentes¸, A. Locatelli, Carbon 64 (2013) 377. [56] A. Locatelli, C. Wang, C. Africh, N. Stojic´ı, T.O. Mentes¸, G. Comelli, N. Binggeli, ACS Nano 7 (2013) 6955. [57] T.O. Mentes¸, A. Locatelli, J. Electron. Spectrosc. Relat. Phenom. 185 (2012) 323. [58] T.O. Mentes¸, G. Zamborlini, A. Sala, A. Locatelli, Beilstein J. Nanotechnol. 5 (2014) 1873. [59] C. Africh, C. Cepek, L.L. Patera, G. Zamborlini, P. Genoni, T.O. Mentes¸, A. Sala, A. Locatelli, G. Comelli, Sci. Rep. 6 (2016) 19734. [60] A. Sala, G. Zamborlini, T.O. Mentes¸, A. Locatelli, Small 11 (2015) 5927. ´ C. Africh, A. Sala, N. [61] G. Zamborlini, M. Imam, L. Patera, T. Mentes¸, N. Stojic, Binggeli, G. Comelli, A. Locatelli, Nano Lett. 15 (2015) 6162–6169. [62] L. Omiciuolo, E. Hernández, E. Miniussi, F. Orlando, P. Lacovig, S. Lizzit, T. Mentes¸, A. Locatelli, R. Larciprete, M. Bianchi, S. Ulstrup, P. Hofmann, D. Alfè, A. Baraldi, Nat.Comm 5 (2014) 5062. [63] S. Nappini, I. Píˇs, T. Mentes¸, A. Sala, M. Cattelan, S. Agnoli, F. Bondino, E. Magnano, Adv. Func. Mat. 26 (2016) 1120. [64] Ch. Tusche, M. Ellguth, A.A. Ünal, C.-T. Chiang, A. Winkelmann, A. Krasyuk, M. Hahn, G. Schönhense, J. Kirschner, App. Phys. Lett. 99 (2011) 032505. [65] Ch. Tusche, M. Ellguth, A. Krasyuk, A. Winkelmann, D. Kutnyakhov, P. Lushchyk, K. Medjanik, G. Schönhense, J. Kirschner, Ultramicroscopy 130 (2013) 70–76. [66] Ch Tusche et al., in preparation.