Applications in Surface Science

  • Ernst Bauer
Chapter

Abstract

In this chapter we illustrate the applications of surface electron microscopy with slow electrons by examples selected from the extensive list of references, which should be accessible to most readers. The list has no claim to completeness, in particular because some important results have been published only in conference proceedings, which are hard to find, and some in the open literature may have been overlooked. Many application examples have already been mentioned in Chap. 3 in connection with the description of the methods and will not be repeated here. The examples are ordered primarily according to the material studied, an imperfect ordering system because certain phenomena are studied using several materials. This will become evident already in the first section, which is concerned with the microstructure of surfaces, studied on metals, semiconductors, and insulators. All methods, which contribute to the understanding of a given phenomenon will be discussed in the corresponding section to illustrate their complementarity whenever existing. Every method has its strengths and limitations and accordingly will dominate the various chapters. For example, LEEM is ideally suited for the study of the microstructure of surfaces and of processes that influence it, while of only limited use for the study of the surface composition. On the other hand, XPEEM gives little information on the microstructure but is powerful in chemical characterization. Some of the other methods such as MIEEM provide very limited information but will nevertheless be mentioned whenever they have been used in a given context.

5.1 Surface Microstructure

The microstructure of surfaces was one of the first applications of LEEM, using the two main contrast mechanisms, phase contrast and diffraction contrast. These show monoatomic steps, step bunches, facets, etch pits, the intersection of glide planes and screw dislocations with the surface and other surface features. The processes which determine the microstructure are imaged dynamically while heating, during sublimation, growth, electromigration, or other surface treatments such as annealing of the radiation damage caused by ion, electron, or photon bombardment. From these images the phenomenological quantities which characterize these processes and the microstructure can be extracted. The atomic aspects such as the nature of the diffusing species or the interactions between the adsorbed atoms, of course, are not accessible because of the limited resolution of LEEM. For example, scanning tunneling microscopy (STM) studies have clearly shown that the high temperature (1 × 1) phase of the Si(111) surface contains 10 % more atoms than the ideal (1 × 1) plane [1] and that these atoms form a two-dimensional gas of magic clusters [2], which play an important role in the step dynamics of the (1 × 1) phase and in the transition to the low-temperature (7 × 7) phase. In principle LEED could give some information on the adsorbed species but even high resolution diffractometer studies showed only a diffuse (2 × 1) structure and a high background around the (00) spot, superimposed on sharp (1 × 1) reflections, which indicate a well-ordered (1 × 1) structure [3]. The diffuse features are compatible with the magic clusters but do not allow extracting size and structure from the data.

With these limitations in mind we discuss in the following the information which can be and has been obtained from LEEM studies of the surface microstructure starting with metal surfaces, followed by Si surfaces, and ending with compound surfaces. Initially dynamic processes were recorded with video cameras on video tape or intermittently with still cameras. This made analysis tedious and led mostly to qualitative results. CCD cameras, data acquisition with computers, and the development of data analysis software eliminated this hurdle, allowing extraction of quantitative results from the images.

5.1.1 Metals

The first metal surface studied with LEEM was the Mo(110) surface [4]. Figure 5.1 shows round monoatomic sublimation steps and straight monoatomic glide steps introduced by glide during rapid cooling. This was the first demonstration of phase contrast, which was analyzed using Kastler’s calculations of Fresnel diffraction from two half-planes shifted by half a wavelength relative to each other [5]. Following Telieps’ original work Mundschau et al. studied surface step phenomena qualitatively not only with LEEM but also with PEEM (Fig. 5.2), making use of the preferential condensation of Cu at steps to decorate them and of the work function difference between Cu and Mo [6, 7, 8, 9, 10]. Other early work was concerned with the influence of steps on the orientation of reconstruction domains of Au(100), using dark field LEEM [11, 12] and with the influence of stress on reconstruction of Pb(110) [13, 14, 15]. Step structure studies continued in the late 1990s in Flynn’s group [16, 17, 18, 19, 20, 21, 22] with thick epitaxial Mo(110) and Nb(110) layers grown and annealed on Al2O3 \( \left(11\overline{2}0\right) \)surfaces. These studies were concerned in particular with faceting of slightly misoriented surfaces, resulting from the slight miscut of the substrate, and with the defects introduced by the misfit with the substrate. Microstructure changes caused by sputtering at elevated temperatures were later studied on Pd(111) [23] and Pt(111) [24] surfaces. Some of the references ([19, 22]) contain a wealth of information on the influence of carbon and oxygen on the surface microstructure, for example oxygen-induced domain formation.
Fig. 5.1

LEEM image of a clean Mo(110) surface taken with 14 eV electrons. The spots are defects in the channel plate image intensifier. Reproduced with permission from Ref. [4]. Copyright 2010 Elsevier

Fig. 5.2

PEEM image of a Mo(110) surface. Monoatomic steps are decorated with Cu deposited at 700 K. Mercury high-pressure arc lamp illumination with  ≈ 5 eV. From Ref. [6]

The studies mentioned up to now gave a good qualitative understanding of the microstructure of single crystals surfaces and the factors upon it depends, but a quantitative understanding requires a detailed study of the step structures as a function of time and temperature and analysis of the experimental results in terms of theoretical models. These have a long history, starting from the famous Burton-Cabrera-Frank (BFC) model [25], and have been reviewed repeatedly, e.g. in Refs. [26, 27, 28, 29], where the derivations of the formulas used below in the data analysis can be found. Both equilibrium and non-equilibrium situations have been studied, predominantly with two methods: step edge fluctuations of more or less straight steps and shrinking/growth of more or less circular two-dimensional islands. The first method, called step fluctuation spectroscopy, was pioneered by Bartelt et al. [30] for the analysis of step fluctuations on the Si(111) surface at 900 °C observed with high energy reflection electron microscopy (REM). Later it was applied by Ondrejcek et al. in Flynn’s group in LEEM studies of the microstructure of close-packed metal surfaces (Mo(110), Nb(110), Pt(111), Pd(111), Au(111), and Ni(111)) [31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43].

In step fluctuation spectroscopy the fluctuations y(x, t) around a step (Fig. 5.3a) are extracted from many LEEM images (typically several hundreds) and Fourier-analyzed:
Fig. 5.3

Fluctuations of a straight step on a Mo(110) surface. (a) Fluctuations y(x, t i ) at three moments t i separated by 2 s. (b) Fluctuations of the Fourier amplitudes y q (t) with different spatial frequency q. 1 s = 30 images. Reproduced with permission from Ref. [32]. Copyright 2003 Elsevier

$$ y\left( x, t\right)={\displaystyle \sum_q}{y}_q(t) \exp (iqx)\kern0.5em \left(0\le x\le L\right) $$
(5.1)
with y q  = y − q * , q = 2πq/L (q integer) and L the step length, which is limited by the image diameter and the magnification. Figure 5.3b shows y q (t) for three q values, indicating a rapid decrease with decreasing wavelength (increasing q). The temperature determines the amplitudes of the fluctuations q, whose energies U q add up to the excess free energy over that of the non-fluctuating step:
$$ U={\displaystyle \sum_q}{U}_q=\frac{L\tilde{\beta}}{2}{\displaystyle \sum_q}{q}^2{\left|{y}_q\right|}^2, $$
(5.2)
where
$$ \tilde{\beta}=\beta +{\mathrm{d}}^2\beta /{\mathrm{d}\Theta}^2 $$
(5.3)
is the step stiffness, which is a measure for the resistance to step fluctuations, and β(θ, T) is the step free energy or line tension per unit length. In thermal equilibrium on average U q = kT/2 so that from Eq. (5.2)
$$ {\left|{y}_q\right|}^2={k}_{\mathrm{B}} T/ L\tilde{\beta}{q}^2. $$
(5.4)
Thus by determining |y q |2 as a function of q 2, \( \tilde{\beta} \) can be determined as a function of step orientation and temperature as illustrated in Fig. 5.4a for two step orientations on Mo(110). Up to about q = 10 the data fit very well the q −2 dependence but at larger q the y q values decrease into the noise level. Nevertheless the fitting range is large enough to give step stiffness values with a small error bar. Measurements of several step orientations show a strong angular dependence of the step stiffness which can be described by \( \tilde{\beta}=0.26\pm 0.03 \cos 2\theta\;\mathrm{eV}/\mathrm{mm} \), with θ = 0° in the [001] direction, a consequence of the twofold symmetry of this surface. On the Nb(110) surface \( \tilde{\beta} \) shows a stronger angular dependence at about 1600 K: \( \kern0.5em \tilde{\beta}=0.32\pm 0.19 \cos 2\theta\;\mathrm{eV}/\mathrm{mm} \). On the threefold symmetric close-packed fcc(111) surfaces \( \tilde{\beta} \) was found in the case of Au to vary from 0.3 eV/nm to about 0.1 eV/nm almost up to the melting point T m, but to be isotropic in the case of Ni within the limits of error. Except for Ni, \( \tilde{\beta} \) decreased only little with temperature in the temperature range studied, 0.4 ≤ T/T m ≤ 0.96, varying with metal (Mo, Nb, Pd, Ni, Pt, Au). Thus, although some common trends were found, there are significant differences between the close-packed surfaces. On the Nb(110) surface another unrelated phenomenon was observed, the movement of steps by a moving edge dislocation intersecting the surface. The dislocation pulls the step along its path, deforms it until the step tension becomes too large, leaves it behind, and the process continues with the next step. From the step shape at the step-dislocation separation point the “plucking” force acting on the dislocation of 75 pN could be derived [39].
Fig. 5.4

(a) Fourier amplitudes of the fluctuations of straight steps with different orientations on a Mo(110) surface at 1545 K as a function of q. At large q values the amplitudes are so small that corrections are needed. The differing symbols indicate uncorrected and corrected values. The inset shows the step stiffness deduced from these data via Eq. (5.4). (b) Time dependence of the time correlation of the Fourier amplitudes F(t′ − t) for several q values at 1680 K. The relaxation time at this temperature is too short for q ≥ 8 for the determination of τ q . Reproduced with permission from Ref. [32]. Copyright 2003 Elsevier

Beyond the determination of the step stiffness the fluctuations provide information on the kinetics of the fluctuations, which are caused by the locally varying attachment and detachment of atoms. The kinetics can be extracted from the q-dependent relaxation of the various modes
$$ \left\langle {y}_q\left( t,^{\prime}\right)\right\rangle ={y}_q(t) \exp \left\{-\left( t^{\prime }- t\right)/{\tau}_q\right\} $$
(5.5)
The relaxation time τ q is extracted from the time correlation functions of the different modes y q (t)
$$ {F}_q\left( t^{\prime }- t\right)=\frac{y_q\left( t,^{\prime}\right){y}_q^{*}(t)+\mathrm{c}.\mathrm{c}.}{2{\left|{y}_q\right|}^2} \exp \left\{-\left( t^{\prime }- t\right)/{\tau}_q\right\}. $$
(5.6)

Figure 5.4b shows an example of such a fit for several q values of a step on Mo(110), from which τ q is obtained. As expected, the relaxation time decreases rapidly to the image acquisition time with increasing q, limiting the q range. Nevertheless the fit gives a clear q −2 dependence of τ q .

Inasmuch as the step edge fluctuations are caused by attachment/detachment of atoms they give information on the corresponding processes, such as surface diffusion or diffusion from the bulk. Assuming that the two processes are additive [34]
$$ {\tau}_q^{-1}=\left({D}_{\mathrm{b}}+\frac{2 qa}{\pi}{D}_{\mathrm{s}}\right)\frac{\pi {A}^2{q}^2}{\Omega {k}_{\mathrm{B}} T}\tilde{\beta}, $$
(5.7)
where D b and D s are the temperature-dependent diffusion coefficients for bulk and surface diffusion, a is the spacing of the surface planes, A is the surface area per atom and Ω is the atomic volume. A plot of \( {\left({\tau}_q{q}^2\tilde{\beta}\right)}^{-1}\Omega {k}_{\mathrm{B}} T/\pi {A}^2 \) as a function of q gives an effective diffusion coefficient D eff = D b + (2qa/π)D s, whose slope gives D s and whose value at q = 0 gives D b. The plot of D eff in Fig. 5.5 shows pure surface diffusion below about 1450 K (D eff (0) = 0) but an increasing contribution of bulk diffusion D eff (0) at temperatures above about half of the melting temperature, increasing with increasing temperature. Using this method the diffusion coefficients for the close-packed surfaces mentioned above have been determined. The results for D s are compiled in Fig. 5.6 and in Table 5.1 together with the step stiffness. These and general theoretical consideration of the mass transport processes involved [44, 45, 46] have led Flynn to the proposal of the “standard metal” model of diffusion with parameters relative to T m, which is indicated in the inset of Fig. 5.6, common to many fcc(111) surfaces with minor deviations.
Fig. 5.5

Effective diffusion coefficients D eff = D B + (2qa/π)D S on the Mo(110) surface as a function of q for various temperatures. At high temperatures the fast relaxation causes large uncertainties even for small q. Reproduced with permission from Ref. [41]. Copyright 2006 Elsevier

Fig. 5.6

Surface diffusion coefficients for the surfaces studied with step fluctuation spectroscopy as a function of T m/T. The dotted line is the “standard metal” surface diffusion coefficient, which is shown again in the inset together with the bulk “standard metal” diffusion coefficient. The dashed line marked IB-Pt(111) is from data obtained from an ion bombardment-modified Pt(111) surface. New figure based on data in Refs. [36, 37, 38, 39, 40, 41, 42, 43]

Table 5.1

Step stiffness and diffusion parameters

Surface

\( \tilde{\beta}\;\left(\mathrm{eV}/\mathrm{nm}\right) \)

α S (cm2/s)

E S (eV)

α B (cm2/s)

E B (eV)

Ni(111)

0.41

1 × 10−4

0.65

1.4

2.45

Pd(111)

0.27

3 × 10−3

1.2

0.2

2.75

Pt(111)

0.21

5 × 10−4

1.2

0.8

2.75

Au(111)

0.17

1.7 × 10−3

0.87

6 × 10−2

1.95

Mo(110)

0.27

4 × 10−4

1.13

0.4

3.1

Nb(110)

0.32

    

V(110)

0.66

0.8

1.43

  

The step stiffness \( \tilde{\beta} \) values are average values at T/T m = 0.6. Only Au, Mo, and Nb show a significant θ dependence and Ni a significant temperature dependence. Some pre-exponentials α and activation energies E have large error bars and are not listed therefore. Compiled from data in Refs. [36, 37, 38, 39, 40, 41, 42]

The second method using surface steps to extract quantitative information on surface energetics and kinetics is based on the growth or shrinking of two-dimensional islands under non-equilibrium conditions. Such conditions exist on flat surfaces at hills or shallow valleys consisting of a series of concentric monoatomic terraces. Driving force is the minimization of the step free energy by reducing the step length. In the study of metal surfaces with LEEM the method was first applied by McCarty [47] to a NiAl(110) surface. He noticed that all islands shrunk at the same rate, which is difficult to understand if the shrinking would be accomplished by surface diffusion from the edges of the smaller islands to the edges of the larger islands. In order to understand the material transport mechanism he oscillated the temperature and found that the normalized island area fluctuations dA/dtA were independent of island size, which means that the material transport occurred locally at the island edge. In the temperature range of his experiment (0.53 ≤ T/T m ≤ 0.63) there is already a significant temperature-dependent vacancy concentration in the bulk, which changes during the temperature fluctuations by vacancy creation or annihilation at the surface causing the surface steps to move. From the area fluctuations as a function of temperature the vacancy formation energy could be determined. Superimposed on the oscillating area changes is a monotonic area decrease due to the isothermal smoothing, which depends upon the temperature around which the oscillations occur. From the temperature dependence of this area decrease an activation energy of 2.54 eV was obtained, which gives an activation energy of 1.9 eV for the rate of vacancy exchange between the bulk and surface steps, taking into account the vacancy formation energy of 0.64 eV.

Further studies of two-dimensional island growth/shrinking have shown that the changes of the island size with time after sudden temperature changes can occur either by volume or surface diffusion. On the Pt(111) surface it was found that volume diffusion was responsible for the island size change in the temperature range studied (0.52 ≤ T/T m ≤ 0.62) and a bulk vacancy formation energy of 1.54 eV was derived from the temperature dependence of the changes [48]. On the other hand surface diffusion was deduced from the isothermal and post-temperature jump island shrinking on the Rh(100) surface in a similar temperature range (0.54 ≤ T/T m ≤ 0.65) with an unusually high energy for creation and diffusion of the active species [49]. Apparently the difficulties of interpreting such data have stopped further work along this line.

In the studies discussed up to now the step changes involved no changes in the total number of atoms in the crystal. Useful information on surface microstructure and energetics can also be obtained from step studies in which this is not the case, i.e., during deposition and sublimation. An example is the determination of the Ehrlich-Schwoebel barrier, i.e., the energy E ES needed to transfer an atom from the upper side of a step to the lower side on the terrace (where it is incorporated in the step). This energy can play an important role in growth at low temperatures and can be determined by step flow studies. These have been carried out by exposing concentric monatomic terraces on flat hills on a Cu(100) surface at temperatures between 360 and 400 K to a constant Cu atom flux and measuring the time dependence of the terrace radii. At these temperatures the thermal adatom concentration is negligible compared to the concentration produced by the incident flux. Assuming steady state conditions and precisely known Cu flux a value of 125 meV was obtained [50]. Surface step studies of the phenomenon opposite to condensation give useful information on the dynamics of sublimation, which proceeds via island decay, nucleation and growth of vacancy islands and in addition via formation and rotation of double spirals when dislocation loops intersect the surface. The analysis of these processes for a Cr(100) surface showed that no mass transfer with the bulk was involved and gave activation energies in agreement with the heat of sublimation [51].

A powerful method, which allows both addition and removal of atoms to the crystal in the same experimental setup was developed in Flynn’s group [24] but unfortunately used only for a short period due to termination of funding. It uses bombardment of the surface by “self-ions,” that is ions of the same material as the crystal, at elevated temperatures at which point defects anneal out rapidly forming two-dimensional structures, either vacancy islands or adatom islands, depending upon if the sputter rate (incoming flux/outgoing flux) is smaller or larger than 1 (Fig. 5.7). The transition between these two modes was found to occur for Pt at 245 eV ion energy. By proper choice of ion current density and temperature large flat regions up to 10 μm in diameter can be prepared, on which single or multiple adatom or vacancy islands can be nucleated and grown in a controlled manner by changing temperature, ion current density, or energy. This is illustrated in Fig. 5.7 [52] for the influence of energy. These processes have been studied in detail theoretically and experimentally [52, 53, 54, 55, 56, 57]. Alternately, on stepped surfaces a high density of ion bombardment-produced two-dimensional islands can coalesce with the existing steps by annealing at higher temperatures and produce strongly oscillating steps, whose decay can be analyzed by step fluctuation spectroscopy over a wider temperature range than in the experiments described above. In this manner the curve marked IB-Pt(111) in Fig. 5.6 was obtained. Concluding the subject step motions a study of the high temperature behavior of steps on i-Al-Pd-Mn quasicrystal surfaces should be mentioned. Two sets of steps were found, which cross each other forming terraces with identical surface composition in contrast to the alternating composition in the bulk. The steps move with different velocities in opposite directions when the temperature is changed due to addition or removal of atoms from the bulk and have different step height. For details see Ref. [58].
Fig. 5.7

Surface modification of a Pt(111) surface by bombardment with Pt ions at 1005 K. The single terrace with 5 μm diameter in (a) was formed initially from a preexisting hill by bombardment with 115 eV ions. Bombardment of this terrace with 115 eV ions created the adatom island shown in (b). Subsequent bombardment with 515 eV islands produced many vacancy islands and caused shrinking of the adatom island as shown in (c). (d) Illustrates how the size of an adatom or vacancy island can be varied at 1070 K by changing the ion energy. Adapted from Ref. [52] by permission of Taylor and Francis Group, LLC

Summarizing the quantitative studies of the surface microstructure, it is evident that only a few metals have been studied over a limited T/T m range. The temperature range is limited towards the top by the vapor pressure, towards the bottom by the magnitude of the step changes that can be detected by LEEM. The results show to some extent common features such as in the surface diffusion behavior of fcc(111) surfaces (“standard metal”) or the transition to bulk diffusion at high temperatures as dominating process, which determines the surface microstructure. Significant deviations from this trends have, however also been observed such as in the case of V(110) and Rh(100). Other unexpected results are the strong angle dependence of the step stiffness on Au(111) up to close to the melting point in contrast to the weak dependence on other fcc(111) surfaces even at lower temperatures as well as the weak temperature dependence of the step stiffness on most surfaces except Ni(111). Much work is still necessary for a comprehensive picture to emerge. In any case, LEEM has been demonstrated to be a powerful method for the study of surface energetics and kinetics of metals.

5.1.2 Semiconductors

Semiconductors have been one of the most studied materials with LEEM, mostly the Si(111) surface because of its interesting phase transition and the Si(100) surface because of its technological importance, followed by other Si surface orientations, Ge, GaAs and other compound semiconductors of practical importance such as SiC, GaN and other binary compounds. In the following only results of interest for the microstructure will be discussed while aspects of more practical interest will be described in Sect.  6.3.

5.1.2.1 Si(111)

The Si(111) was the first semiconductor surface studied with LEEM, mainly in order to settle a problem of the (7 × 7)↔(1 × 1) transition on this surface, which was much discussed in the early 1980s. LEED studies by Bennett and Web [59] had originally favored a second order phase transition but Witt in a very careful precision LEED diffractometer study [3], which was not accepted for publication, excluded a second order transition because of the absence of critical scattering. On the other hand reflection electron microscopy (REM) studies led originally [60] to the conclusion that the transition is first order but a later REM study [61] indicated that the transition is second order. REM studies are hampered by the strong foreshortening of the image and dynamical diffraction problems, which make contrast interpretation difficult. In LEEM neither problem exist because of the normal incidence of the beam and the easy distinction between (1 × 1) and (7 × 7) structure due the difference in the I(V) curves, even for the (00) beam, so that dark field imaging is not needed for contrast formation. Therefore, in a detailed study of the kinetics of the phase transition with LEEM, Telieps, and Bauer showed unambiguously that the transition is first order and determined the shape, orientation, nucleation, growth, and decay of the triangular (7 × 7) domains [62, 63]. Depending upon heating and cooling history many microstructural features were obtained, some of which are reported in later topical reviews [64, 65, 66, 67]. Examples are shown in Fig. 5.8 [66, 67]. The influence of surface defects was also studied, though mainly on Si(100) [68]. Si(111) served later for a demonstration of the possibilities of a commercial instrument (JEOL) [69].
Fig. 5.8

LEEM images of a Si(111) surface after different treatments. (a) Slowly cooled a few degrees below the transition temperature. (b) Rapidly quenched to low temperatures. (c) Annealed for several hours at about 1200 K. (d) Cooled to room temperature after many annealing and cooling cycles, which have led to the formation of SiC crystals which pin the steps. Bright region have (7 × 7) structure, dark regions have (1 × 1) structure (except the black dots in (d)). Electron energy 10–11 eV. The dark diffuse line seen in these and some later images is a defect in the channel plate image intensifier. (a) Reproduced from Ref. [67] with permission from the American Institute of Physics, © 1991; (bd) from Ref. [66] with permission from Springer Science + Business Media

The early work in Bauer’s group [62, 63, 64, 65, 68, 70, 71] was more or less exploratory, looking qualitatively at the phase transition, determining the shape and orientation of the (7 × 7) domains, their primary nucleation at steps as a function of undercooling and step orientation and the secondary nucleation and growth, finally ending with the surface completely covered with (7 × 7) domains separated by domain boundaries. The growth of these domains was later studied by Phaneuf et al. [72], who found that there was no anisotropy in the boundary energy and that grain growth was not purely driven by boundary curvature as predicted by theory. They also studied the phase transition on vicinal Si(111) surfaces with the result that the surface broke up very rapidly into step bunches and (7 × 7) structure-covered (1 × 1) terraces, stopping spontaneously at a well-defined width. This theoretically unexpected result of finite facet width was attributed to elastic interactions caused by the terrace boundaries [73, 74]. A closer examination of the influence of the quenching rate revealed at high rates a spinodal pattern, which they attributed to the crossing of the surface free energies of the (7 × 7) and (1 × 1) structures as a function of miscut, resulting in a spinodal region [75].

With exception of the work reported in Ref. [69] all studies discussed up to now were made with the first LEEM instrument, which was used for the study of many other surfaces. Once newer LEEM instruments became available in the early to middle 1990s, in particular in semiconductor industry laboratories such as IBM Yorktown Heights and later in the NTT Basic Research Laboratories, in-depth studies of Si surfaces started, both on Si(111) and Si(100). They may be roughly divided into two groups, one with the main attention focused on the (7 × 7)-(1 × 1) transition, the other focusing on surface diffusion in general and on its involvement in the phase transition.

We discuss first the former group of studies, which were done predominantly by Hannon et al. [76, 77, 78, 79, 80, 81, 82, 83, 84] but others have also contributed to this subject [85, 86]. They were directed on the one hand to test the validity of standard phase transition theory using the (7 × 7) ↔ (1 × 1) transition as test object, on the other to understand this transition within the framework of existing theory. Both aspects and the results obtained up to 2002 are thoroughly discussed in two excellent review papers [87, 88] so that they will be sketched here only briefly. The most striking aspect of the transition is the relatively wide coexistence range around the critical temperature T c, whose values varies somewhat from author to author due to the difficulty of accurate absolute temperature measurements in LEEM. The phase coexistence range can be seen on large step-free regions, on which (homogeneous) nucleation is not pre-emptied by (heterogeneous) nucleation at the upper side of the steps, where the nucleation barrier is lower. On the step-free regions considerable overheating, undercooling, and hysteresis occurs as shown in Fig. 5.9b, c, with a wide coexistence range, in particular upon cooling, in contrast to stepped regions as shown in Fig. 5.9a. This coexistence region allows now the study of the energetics and kinetics of domain nucleation, growth, and decay and to determine the physical parameters, which determine these processes.
Fig. 5.9

Fractional area of the (1 × 1) regions on a Si(111) surface during very slow heating and cooling to achieve equilibrium. (a) In a region with 0.3–0.6 μm step spacing; (b) in a vacancy island with 4 μm average diameter; (c) on a mesa with 15 μm diameter. Adapted with permission from Ref. [85]. Copyright 2001 Elsevier

Phase coexistence in two dimensions is usually determined by the competition between short range attractive and long range repulsive elastic interactions but electrostatic interactions can also produce coexistence when the work functions of the two phases differ sufficiently. Measurements of the relative size of (1 × 1) and (7 × 7) domains as function of temperature and terrace widths showed that in this system elastic interactions are more than an order of magnitude larger than electrostatic interactions, because of the low work function difference between the two phases as determined by the transition between mirror and LEEM mode of imaging [77]. The elastic interactions are caused by the surface stress difference between the two phases. Fluctuation spectroscopy (similar to the step fluctuation spectroscopy described in Sect. 5.1.1) of phase boundaries, was used to determine the stiffness and the surface stress difference between the two phases [78].

Measurements of the size of stable isolated triangular (7 × 7) domains on a large terrace as a function of temperature above the critical temperature T c showed at 5.3 K above T c the smallest stable domain size with edge lengths L 0 = 200 nm, corresponding to a free energy difference Δγ = 6.15 μeV/(1 × 1) unit cell. This gives a very small entropy difference ΔS = ΔγT = 0.013k B per (1 × 1) unit cell which illustrates how delicate this transition is. A detailed analysis of the elastic interactions showed that the stable domain size is determined by elastic interactions at the phase boundaries [81]. A later study of stable isolated (7 × 7) domains on a very large terrace [82] showed that their equilibrium shape changed from convex to concave with increasing domain size as illustrated in Fig. 5.10a. Concave shapes had already been observed in the very early work [63] but had been attributed to growth shapes. The equilibrium, characterized by the ratio b/L in Fig. 5.10b, is plotted in Fig. 5.10c as a function of domain area A. The experimental data can be well fitted by an azimuthal angle θ dependence of the boundary energy β(θ) = β 0 + β 1(1 − cos(3θ))/2 with β 0 = 15.5 meV/nm, β 1 = 3.2 meV/nm, assuming the surface stress difference of 30 meV/Å2 determined in the earlier work. Medhekar et al. [83] have generalized these calculations and found that the concave shape is actually metastable. Figure 5.11a shows the experimental equilibrium distribution of domains, Fig. 5.11b the calculated domain shape of an isolated domain and Fig. 5.11c the domain distribution as a function of system parameters. As seen in (5.11c) the stable configurations of large domains are corner-connected subdomains. Their formation requires overcoming a notch nucleation barrier, which however is so small that these domains do form as seen in the equilibrium configurations on a large terrace (Fig. 5.11a). Apparently the tendency to form interconnected domains above a given domain size is pre-emptying another phenomenon often seen in pattern formation in the presence of long range elastic interactions: arrangement of the (7 × 7) regions in a periodic pattern of stripes or droplets (if the domain boundary energy were isotropic), depending upon coverage. A theoretical study of this phenomenon showed that also the triangular (7 × 7) domains should order in this manner. The fact that this was not observed in experiment was attributed earlier to a too weak elastic driving force or too slow kinetics [80] but obviously it is the process described in [83].
Fig. 5.10

Size dependence of the equilibrium shape of an isolated (7 × 7) domain on a large Si(111) terrace. (a) Observed domain shapes; (b) definition of the curvature ratio b/L; (c) measured b/L ratio and b/L ratio calculated with the parameters derived from the fit. Adapted from Ref. [82]. Permission from Nature Publishing Group (UK)

Fig. 5.11

(7 × 7) domain equilibrium structures of domain groups on a large Si(111) terrace. (a) LEEM image showing different equilibrium domain configurations. (b and c) Morphological phase diagrams of the equilibrium configurations. The shapes in (b) develop continuously with size but become metastable with increasing size, those in (c) have to overcome nucleation barriers indicated by lines but are lower in energy. A 0 = 0.18 μm2, α is the anisotropy parameter related to the parameters of the preceding analysis by α = β 1/2β 0. For the (7 × 7) domains α = 0.28. Adapted with permission from Ref. [83]. Copyright 2007 by the American Physical Society

While a considerable amount of work has been done on the (quasi)equilibrium aspects of the (7 × 7) structure, which gave considerable insight in the energetics, the kinetics of the (7 × 7) ↔ (1 × 1) transition was studied much less. Telieps and Bauer had done only a few measurements of the dependence of nucleation and growth rate upon undercooling without detailed analysis [63]. Hannon et al. measured the decay rate of superheated (7 × 7) domains and found a linear decay of the domain size with time, independent of the size of the domains [76]. This implies a steady supply of atoms needed to convert the (7 × 7) structure into the (1 × 1) structure, which has a higher atomic density. They explained this with continuous vacancy-adatom creation in the (1 × 1) region surrounding the (7 × 7) domains, with the adatoms diffusing to the (7 × 7) domains and the vacancy to the steps, thus decreasing the size of the (7 × 7) domains and causing the steps to retreat. The activation energy for vacancy-adatom generation was estimated to be in the range 3.2–3.8 eV, much higher than the sum of the activation energy for adatom formation by detachment from a step E ad and of the energy barrier for surface diffusion E diff, which varies between 1.1 and 1.9 eV on the pure (1 × 1) surface [96]. It is well established that the (1 × 1) surface is covered in equilibrium by about 0.25 monolayers of fast moving atoms (or clusters). This adatom coverage is continuously replenished by supply from the steps causing their retreat. Apparently this process is suppressed at steps whose upper level has (7 × 7) structure as will be seen later. A detailed theoretical study of the decay of the (7 × 7) structure with the phase field method has shown that it proceeds in two steps: first the corner holes and dimer chains are broken, then the stacking fault is eliminated, which is the rate-limiting process [89]. The growth of (7 × 7) domains has also been studied with the phase field method and found to be in good agreement with experiment [84]. In particular the serrated domains formed at low growth rate and the pine tree-like shapes seen already in the first LEEM studies [62, 63, 64, 65] and in a compilation of experiments in the coexistence range [86] are reproduced by this theory.

Keeping the atomic density difference between (7 × 7) and (1 × 1) structure and the presence of the self-adsorption layer in the (1 × 1) surface in mind it is immediate obvious that their relative surface coverage can be manipulated in the coexistence range by changing the equilibrium between the two phases, either by adding or removing Si atoms. This has been accomplished by exposing the surface to silane—increasing the number of Si atoms—or by exposing to oxygen—removing Si atoms from the adsorption layer. These chemical potential changes have also been described by effective free energy differences between the two phases [79]. Concluding the discussion of the phase transition studies, LEEM has revealed a fascinating scenario of two-dimensional crystal phenomena, which has also contributed to the understanding of phase transitions in general.

The second group of studies, surface diffusion processes, has been done mainly by Altman’s group [90, 91, 92, 93, 94, 95, 96] and by Hibino et al. [97, 98, 99]. The goal of their studies was to understand the relative importance of surface diffusion, step attachment–detachment of atoms, and step barriers for the surface microstructure by studying the behavior of monatomic steps as a function of time and temperature. Most of the work was done by studying the decay of circular islands and holes (vacancy islands), which was already discussed in Sect 5.1.1, but step boundary fluctuation spectroscopy of circular and linear steps was used too. Hibino et al. used the decoration of the upper sides of steps by the (7 × 7) as step marker and were therefore limited to a temperature region near T C while Altman’s group made use of the step phase contrast, which allows a wider temperature range and excludes a possible influence of the (7 × 7) structure on step behavior.

The analysis of the experimental data is based as in the case of metals on the methods reviewed in Refs. [26, 27, 28, 29]. Briefly, the size of an island decays with time at a fixed temperature as dA/dt = A 0(t 0 − t) α , where A 0 is the initial area, t 0 the total decay time at which the island disappears and α the decay exponent, whose value is determined by the processes involved in the island decay. The limiting processes are adatom surface diffusion and attachment–detachment. The diffusion rate is characterized by the diffusion constant D, the step attachment–detachment rate by the step attachment–detachment kinetic coefficient K (assuming symmetry with respect to the attachment direction) and the relative importance of the two processes by the kinetic length d = D/K. For two concentric circular islands theory predicts for the decay of the island/hole area A
$$ \frac{\mathrm{d} A}{\mathrm{d} T}=-2\pi \sqrt{3}\Omega D{n}_{\mathrm{eq}}\frac{ \exp \left(\frac{\Omega \tilde{\beta}}{r{ k}_{\mathrm{B}} T}\right)- \exp \left(\frac{\Omega \tilde{\beta}}{R{ k}_{\mathrm{B}} T}\right)}{d/ r+ d/ R+ \ln \left( r/ R\right)}\kern0.5em. $$
(5.8)

Here r and R are the radii of the inner and outer island/hole, \( \tilde{\beta} \) is the step stiffness, Ω = √3a 2/2 the area per atom on the (111) lattice with lattice constant a. The prefactor n eq = 2Ω− 1 exp(−E ad/k B T) is the adatom concentration in equilibrium with a straight step, the prefactor \( D=\sqrt{3}\Omega \nu \exp \left(-{E}_{\mathrm{diff}}/{k}_{\mathrm{B}} T\right) \) is the diffusion constant and E ad and E diff are the energy needed for formation of an adatom by step detachment and the activation energy for surface diffusion, respectively. When βa 2/kT ≪ r ≪ R Eq. (5.8) can be simplified for the limiting cases of d = D/K. For K ≪ D the island decay is attachment-detachment-limited (ADL) and Eq. (5.8) can be reduced to \( \mathrm{d} A/\mathrm{d} t=-2\pi \tilde{\beta}{n}_{\mathrm{eq}}{\Omega}^2 K/{k}_{\mathrm{B}} T \), so that A(t) ~ t(α = 1). When K ≫ D then diffusing atoms can be attached/detached easily at the step and Eq. (5.8) can be reduced to \( \mathrm{d} A/\mathrm{d} t=-2\pi {c}_{\mathrm{eq}}\tilde{\beta}{\Omega}^2 D/{k}_{\mathrm{B}} Tr \ln R \). With A = r 2 π integration leads to \( A(t)=\pi {\left(3{c}_{\mathrm{eq}}\tilde{\beta}{\Omega}^2 D/{k}_{\mathrm{B}} T \ln R\right)}^{2/3}{\left({t}_0- t\right)}^{2/3} \), i.e., α = 2/3. With the concentric ring configuration the experiments are straightforward—the art is in the sample preparation, careful data acquisition, and analysis—and give α and the pre-exponential \( {n}_{\mathrm{eq}}\tilde{\beta} D \).

The results of the island decay measurements from many islands with R/r ratios ranging from 3 to 200, initial island and hole distributions and areas up to a few micrometer in the temperature region 1090–1150 K gave a decay exponent α = 0.76 ± 0.06 [97]. In this temperature region the rim of the steps was decorated by a band of (7 × 7) structure. An example of such an experiment is shown in Fig. 5.12a [97]. Measurements at temperatures in the absence of (7 × 7) step decoration gave values of 0.726 at 1143 K and 0.771 at 1183 K [94] and of 0.75 at 1163 K [95]. While the last three values suggest a temperature dependence, it appears that all values taken together congregate around a mean value of ¾ instead of 2/3. This problem is discussed in the cited publications based on Eq. (5.8) for the decay rate in terms of the variation of the kinetic length d, i.e., in terms of the competition between surface diffusion and step attachment–detachment.
Fig. 5.12

Island decay on Si(111) surface. (a) Top: Snapshots of an isolated island on a 20 μm wide terrace at T C (1133 K); bottom: area as a function of decay time with data fitted to A 0(t 0 − t) α . (b) Dn eq from island decay times measured at many temperatures, assuming a step stiffness of 137 meV/Å. Adapted with permission from Ref. [97]. Copyright 2001 by the American Physical Society

When \( \tilde{\beta} \) is known, then the sum of the energy for adatom formation by detachment from the step, E ad, and of the surface diffusion energy, E diff, can be obtained from the temperature dependence of the pre-exponential in Eq. (5.8). These energies are important parameters for the understanding of the mass transport on the surface. The island and hole shapes on Si(111) at high temperatures are more or less circular so that there is no anisotropy and \( \tilde{\beta}\approx \beta \). Hibino et al. [97] assumed in the temperature range of their measurements (1090−1150 K) a constant value of 137 meV/Å, while Altman’s group [96] measured \( \tilde{\beta} \) by step fluctuation spectroscopy and found values decreasing from 68 meV/Å at 1145 K to 56 meV/Å at 1233 K. Where the temperatures of the two experiments overlap, the \( \tilde{\beta} \) value of the (7 × 7) structure-decorated step [97] is by a factor of two larger than that of the pure (1 × 1) step [96]. If the larger \( \tilde{\beta} \) value for the decorated step is correct then the (7 × 7) structure would cause considerable step stiffening. However it has to be kept in mind that \( \tilde{\beta} \) and β values determined with different methods differ considerably.

Figure 5.12b shows the temperature dependence of Dn eq derived from measurements of the island area decay (Fig. 5.12a) at many temperatures, using Eq. (5.8). Dn eq = 2 √ 3ν  exp(−(E ad + E diff)/k B T) decreases in a narrow temperature range by a factor of 50 from above T C to below T C, indicating a significant change of E ad + E diff as the jump frequency ν is unlikely to vary much between the two surface structures. From the temperature dependence of Dn eq above T C the values ν = 1.1 × 1013 s−1 and E ad + E diff = 1.3 eV were obtained [97]. From island decay measurements at higher temperatures, at which the steps are not decorated, values of ν = 2.59 × 1013 s−1 and E ad + E diff = 1.53 eV [95, 96] were extracted. E ad can be estimated from n eq ≈ 0.2 monolayers to be 0.2 or 0.23 eV, which gives 1.3 eV for E diff for the (1 × 1) structure. Below T C the scatter of the data is too large to allow such an analysis. The much smaller Dn eq value on the (7 × 7) surface has, however, an important consequence for the decay of vacancy islands vs. adlayer islands: the rim of the vacancy islands is decorated by the (7 × 7) structure, which due to its lower Dn eq reduces the rate at which vacancy islands are filled. This reduction acts as an effective downhill Ehrlich-Schwoebel barrier, for whose height an upper limit of 0.6 eV was estimated. No such barrier exists for the decay of adlayer islands. This phenomenon and the influence of n eq differences in the decay of superheated and equilibrium vacancy and adlayer islands have been discussed in detail by Hibino et al. [98, 99].

At the higher temperatures studied by Altman’s group, at which this (7 × 7) barrier at the step does not exist, the question of the step permeability P and of the distinction between diffusion-limited and attachment–detachment-limited step movement as characterized by the kinetic length d = D/K can be studied more clearly [94, 95]. A numerical analysis of Eq. (5.8) for various d values as a function or r/R [95] shows that α depends upon the magnitude of d. For d ≈ 240a, where a is the lattice constant, α is 0.77 from the largest r/R values considered (0.4) down to r/R ≈ 0.1. When d < 240a, then α decreases with decreasing r/R but is still far away from 2/3 at r/R = 0.5, even at d = 10a, a value for a pure DL mechanism. When d > 240a, then α increases with decreasing r/R but still is very different from 1, the value for a pure attachment–detachment-limited process. The analysis of island decay at 1163 K gave d = 75a, so that diffusion is dominating the island decay. In the studies, described up to now, R was fixed. A study of the time dependence of the area of two concentric islands with radii r and R > r on a larger disc allows determination of the step permeability P. Assuming zero detachment from the outer island R will increase until r = 0 because all atoms detached from the inner island will be attached to the outer when P = 0. When P = 1 R will remain constant because all atoms detached from the inner island will diffuse beyond the step of the outer island. A detailed analysis clearly indicated negligible permeability [95].

Island decay experiments at the highest temperatures studied (up to 1380 K) showed above 1200 K increasing deviations from the behavior expected on the basis of Eq. (5.8), which were attributed to desorption. Therefore this equation was generalized by incorporating desorption via the lifetime τ = ν 0 − 1  exp(E des/k B T) of the adsorbed atoms. The diffusion length before desorption is \( {x}_{\mathrm{s}}=\sqrt{D\tau}=\left(\sqrt{3}/2\right) a \exp \left[\left({E}_{\mathrm{des}}-{E}_{\mathrm{diff}}\right)/2{k}_{\mathrm{B}} T\right] \), assuming the same value of ν 0 for diffusion and desorption. The inclusion of desorption gives excellent agreement between theory and experiment for island decay times from 1145 to 1380 K with the fit parameters ν 0 = 2.59 × 1013 s−1, E ad + E diff = 1.53 eV and E des − E diff = 2.56 eV. The last two values give a sublimation energy E subl = E ad + E des = 4.09 eV in good agreement with literature, supporting the consistency of the values [96].

Summarizing the island decay studies, LEEM has provided considerable insight into the processes that control the microstructure of the Si(111) surface both in the phase transition region and above it. These studies were done at constant atom density except in those, in which sublimation occurred [95]. Additional information can be obtained from growth studies. Altman’s group studied the competition between step flow growth and nucleation on the terraces on the (7 × 7) surface between 700 and 850 K [90, 91, 92, 93]. As a first step the so-called critical terrace width for step flow growth, λ c, defined by the last appearance of nuclei on the terrace (Fig. 5.13a) [92] was studied for various step orientations, with the result shown by the unmarked central curve in Fig. 5.13b [93]. With a number of assumptions theory relates λ c to the activation energy for surface diffusion E diff and the binding energy E i* of the critical nucleus consisting of i * atoms by
Fig. 5.13

Critical terrace width λ c for step flow growth on the Si(111)-(7 × 7) surface. (a) LEEM image of the transition between two-dimensional nucleation on the terrace and step flow growth at 800 K and a flux of 0.015 Si monolayers/min. The islands determining λ c are indicated by arrows. Energy 42.5 eV. (b) Critical terrace width as a function of 1/T for a flux of 0.1 monolayer/min. (a) Reprinted with permission from Ref. [92]. Copyright 2002 by the American Physical Society. (b) Reprinted with permission from Ref. [93]. Copyright 2012 World Scientific

$$ {\lambda}_{\mathrm{c}}^2= A\kern0.37em \exp \left[-\left({E}_{\mathrm{diff}}+{E}_{i^{*}}/{i}^{*}\right)/{k}_{\mathrm{B}} T\right], $$
(5.9)
with the pre-exponential containing the flux, the jump frequency, the density of the critical nuclei and their size. Fitting the data in Fig. 5.13b to this expression gave for sum of the two energies E = 2.05 and 0.72 eV, for the pre-exponential 7.0 × 1019 and 5.0 × 109 nm2 in the high and low temperature ranges, respectively. At high temperatures, at which nucleation occurs only at the (7 × 7) domain walls, nucleation was considered to be the dominating factor determining λ c, while at low temperatures, at which nuclei form also within the domains, diffusion was concluded to be dominating, with a transition between these two processes at about 750 K [90, 91]. Adsorption layers can either enhance or impede step flow growth (surfactants or anti-surfactants) by changing surface diffusion and/or step attachment, which should have a strong influence on the critical terrace width. This is evident in the curves marked Sb and In in Fig. 5.13b: the pre-factor in Eq. (5.9) are strongly modified but the activation energy in the presence of Sb is similar to that on the pure (7 × 7) structure at high temperature, while that in the presence of In is similar to that in the low temperature range on the (7 × 7) surface. This suggests that In (Sb) suppresses (enhances) island nucleation on the terraces and enhances (suppresses) Si adatom attachment [93].

The location of the nuclei used in the determination of λ c with respect to the steps bounding the terrace allows determination of the asymmetry of the step attachment–detachment kinetic coefficient K. If attachment is symmetric from bottom and top side of the step then nucleation will occur halfway in between the two steps, otherwise it is shifted towards the step with smaller K, provided that the steps are completely impermeable, i.e., incorporate all arriving atoms. Such shift towards the step-up side was actually observed at λ c = 950 nm, 800 K, and a flux of 0.015 monolayers/min, indicating K (step-down) > K + (step-up) [92]. More data would be needed to determine how general these results are, whose explanation requires extensions of the Burton-Cabrera-Frank (BCF) theory. One example, in which the BCF theory completely describes step processes, is the growth and sublimation of Si(111) via spiral motion around screw dislocations. However for the same process on the Si(100) surface the theory has to be generalized [100].

5.1.2.2 Si(100)

The study of the Si(100) surface, the most important Si surface in semiconductor technology, began shortly after the nature of the (1 × 1) ↔ (7 × 7) phase transition on the (111) surface had been demonstrated. The first published images [65] showed very clear (2 × 1) structure contrast with slightly tilted illumination, which makes the (2 × 1) and (1 × 2) regions inequivalent. Steps in which the bonds are oriented along the step (type A) appeared smooth, steps with bonds perpendicular to the step (type B) appeared rough (Fig. 5.14) [101]. Depending upon cooling conditions from the high temperature annealing or flashing the roughness difference between the two steps can be much larger, indicating a strong temperature dependence of the stability of the two steps. Above 1300 K the steps were moving rapidly due to sublimation and at about 1350 K the steps disappeared. After Telieps’s initial experiments Mundschau et al. studied the dependence of the microstructure of this surface upon preparation conditions in considerable detail, in particular the influence of slip lines, partial and screw dislocations on step structure and dynamics [68] and the influence of step pinning centers such as SiC precipitates on the step dynamics during sublimation between 1100 and 1350 K [102]. Figure 5.15 [102] shows a later stage of the sublimation when already many steps have piled up in front of two pinning centers. In between these centers the steps have passed due to sublimation, forming sublimation hillocks. Sublimation via repeated Lochkeim formation at still higher temperature is illustrated in Fig. 5.16 [103]. Here the step free energy is isotropic, resulting in circular hole patterns.
Fig. 5.14

Si(100)-(2 × 1) surface. (a) LEEM image showing the contrast between (2 × 1) and (1 × 2) reconstructed regions, smooth SA and rough SB steps, obtained by tilting the beam into one of the 〈011〉 directions. Electron energy 6 eV. (b) Schematic of Si(100) surface showing the orientation of the steps with respect to the dimer directions. From Ref. [101] with permission from Springer Science + Business Media, Copyright 1988

Fig. 5.15

Hillock formation on Si(100) upon extended sublimation caused by step pile-up at pinning centers. Electron energy 7 eV. The sketches indicate the formation process of the resulting step/terrace structure. Adapted with permission from Ref. [102]. Copyright 1989 Elsevier

Fig. 5.16

LEEM video frames of the sublimation process via hole formation. Electron energy 5 eV. Adapted with permission from Ref. [103]. Copyright 1992 Elsevier

The anisotropy of the step energy at low temperatures suggested by the different roughness of A and B steps after rapid cooling (Fig. 5.14) can be estimated from the shape of two-dimensional islands formed during very slow growth. Figure 5.17 [104] shows frames from a movie taken during deposition of Si under conditions at which no two-dimensional nucleation on terraces occurred, but new terraces started only at defects. Growth is clearly strongly anisotropic. Assuming that growth is close enough to equilibrium, a step energy ratio β A/β B = 2.6 would explain the observed terrace shapes, a value much smaller than the theoretical value of 15 available at that time, but in qualitative agreement with STM results. Of course, the conclusion implies that surface diffusion, step attachment anisotropies, and energy barriers have no influence on the shape [104]. The work in Bauer’s group on clean Si surfaces, about which additional information can be found in some summaries [71, 103, 105, 106] turned after this study to metal layers on surfaces.
Fig. 5.17

Near equilibrium growth of Si on Si(100) with terrace nucleation at a defect at about 900 K and a deposition rate of 0.2 monolayers/min. LEEM video frames. Time elapsed between images: 180, 300, 480, 660, and 780 s. Reproduced with permission from Ref. [104]. Copyright 1991 Elsevier

The study of clean Si(100) surfaces was continued by Tromp et al. [107, 108, 109, 110, 111, 112, 113, 114, 115] and extended to the determination of quantitative values of the parameters, which determine the surface microstructure similar to the work discussed in Sects. 5.1.1 and 5.1.2.1. In one of their first studies they investigated the step morphology of surfaces with a miscut of less than 0.1° as a function of miscut and found that the resulting step structure could be classified into four phases with decreasing miscut: double height steps, single height straight steps, single height wavy steps, and a hilly structure [107, 108], in agreement with theory according to which step undulations reduce the size of stress domains.

As the preceding images show, the surface morphology of the Si(100) surface as seen in the terrace-step shapes, is strongly temperature-dependent, much more than that of the Si(111) and the metal surfaces discussed beforehand. This makes step fluctuation spectroscopy an ideal tool for the determination of the step dynamics and energetics on Si(100). In a first experiment Bartelt et al. [109] measured step fluctuations (Eq. (5.1)) from 910 to 1480 K and analyzed the data via the correlation function (Eq. (5.6)) with the pre-exponential \( A(q)=2{k}_{\mathrm{B}} T/ L\tilde{\beta}{q}^2 \). The stiffness \( {\tilde{\beta}}_{\mathrm{A}} \) of the SA steps was found to be much higher than that of the SB steps, \( {\tilde{\beta}}_{\mathrm{B}} \), for example by a factor of 7, and the ratio \( {\tilde{\beta}}_{\mathrm{A}}/{\tilde{\beta}}_{\mathrm{B}} \) strongly decreased with increasing temperature as expected. In this study also the ratio of the step free energies was determined from the shape of equilibrium shapes of islands produced by Si deposition and annealing. Between 1020 and 1330 K only elliptical shapes with temperature-dependent axis ratio were observed, but not the shape with sharp tips shown in Fig. 5.17. However, extrapolation of the linear increase of the β A/β B ratios derived from the equilibrium shape measurements and of those derived from the step stiffness to the lower temperature used in Fig. 5.17 [104] agrees well with the ratio reported in that study. Towards high temperatures there is still a significant anisotropy at 1330 K (β A/β B ≈ 0.7). At all temperatures the step stiffness is considerably smaller than on the (111) surface, e.g. \( {\tilde{\beta}}_{\mathrm{A}} \) by a factor of 8 compared to the values measured by Altman’s group and \( {\tilde{\beta}}_{\mathrm{B}} \) is also correspondingly smaller.

In an extension of this work Bartelt et al. [113] determined also another measure of the step stability, the step-edge mobility Γ, which is the rate at which adatoms attach/detach at step edges and which is related to the kink formation energy E k via Γ = Γ0 exp(−E 0/k B T). In the case of attachment–detachment limited kinetics, which is indicated by the observed 1/q 2 dependence of τ(q), Γ is related to the quantities obtained from step fluctuation spectroscopy A(q) and τ(q) via Γ = LA(q)/2τ(q). In the temperature and q range studied (990–1370 K, 0.01–0.035 nm−1) Γ was found to increase from 103 to about 105 nm3/s and to be independent of q. From the slope of the Arrhenius plot of Γ a value of E 0 = 1.45 ± 0.15 eV was derived, interestingly both for SA and SB steps. Thus the difference between the two steps is not caused by different attachment–detachment rates but by step stiffness and step free energy.

With steps well characterized, the interaction between steps and surrounding two-dimensional gas, i.e., the relative importance of step attachment/detachment kinetics and surface diffusion, was studied. This was done by analyzing the time dependence of the island areas in a high-density, non-equilibrium two-dimensional island distribution, the so-called Ostwald ripening, in a large step-free region at 940 K [110, 114]. The average area of all islands was found to increase linearly with time while that of the smallest islands decreased linearly with time. This confirms the earlier conclusion that the change of the island size is limited by attachment/detachment kinetics at the steps. The rate of change of the individual island areas ∂A/∂t is determined by the chemical potential difference between island and two-dimensional gas via
$$ \frac{\partial A}{\partial t}=\frac{L\Gamma}{\Omega {k}_{\mathrm{B}} T}\left({\mu}_{\mathrm{is}}-{\mu}_{\mathrm{ad}}\right), $$
(5.10)
where L is the island circumference and Ω the area per atom. The chemical potential of the islands is determined by the Gibbs-Thomson equation, which relates it to the size and shape of the island. Therefore measurements of ∂A/∂t of individual islands give the chemical potential surrounding it. The chemical potential maps generated in this manner showed strong variations depending upon the island surrounding. Thus the chemical potential of the adatom gas is not uniform with corresponding consequences for the adatom diffusion coefficients.
Other interesting experiments addressed the critical nucleus size in homogeneous nucleation [111] and the thermal adatom concentration [115]. The first problem was studied using silane and disilane gas as supply of Si, a substrate temperature of 920 K, which is above the H desorption temperature, and growth rates from 0.1 to 2 monolayers/min. Under these conditions the adatom concentration from the incoming flux is only about 2 % of that of thermal adatom concentration, i.e., the supersaturation is very small and consequently the critical nucleus size N c large. Careful analysis of the data, using classical homogeneous nucleation theory, gave N c ≈ 650 adatoms, generally accepted to be dimers. The thermal adatom concentration was studied between 1310 and 1560 K using large lithographically prepared [112] step-free regions with up to 10 μm edge length. Quenching from these temperatures to low temperature caused condensation of the two-dimensional adatom gas into many small two-dimensional crystals, whose total area coverage is a measure for the concentration c T of the gas at the starting temperature T, excluding edge effects (Fig. 5.18) [115]. c T was found to increase from 0.02 at 1310 K to 0.04 at 1560 K. This gave an activation energy for thermal adatom creation on terraces of 0.35 ± 0.05 eV, which fits best to the formation energy of a dimer instead of an atom.
Fig. 5.18

Thermal adatom concentration on a step-free Si(100) terrace. (a) (1/2,0) dark-field LEEM images before and after quenching taken with 3.5 eV electron energy. The islands seen after quenching have formed from the two-dimensional adatom gas, which does not produce contrast in LEEM. (b) Adatom concentration as function of 1/T from measurements of the total island area coverage shown in (a). The dashed lines show the concentrations calculated for different formation energies. Adapted with permission from Ref. [115]. Copyright 1998 by the American Physical Society

Another method for producing large flat areas for measurements minimizing edge effects should be mentioned briefly too: Si deposition onto the sublimating surface, which allows higher temperatures without increase of the net sublimation rate. This increases the diffusion rate (D) without decreasing the adatom lifetime (τ), resulting in a larger diffusion length λ = ()1/2. λ values from 10 to 30 μm at 1270 K have been observed under these conditions [116]. This method was used to study the change of the island shape with island size during Si deposition under quasi-equilibrium conditions at 960 and 1130 K at a growth rate of 0.1 monolayer/min from disilane gas in the absence of defects. To exclude edge effects, 10 × 15 μm2 terraces were prepared lithographically. With increasing size the island shape changes from elliptical to “American football” shape, a shape similar to that in Fig. 5.17, and at the lower temperature the aspect ratio approaches the value derived from this figure. The size dependence was interpreted in terms of a theoretical model, which takes into account not only the step free energy in the calculation of the equilibrium shape but also the surface stress anisotropy. By fitting the aspect ratios measured at the two temperatures to the theory the step free energies and the surface stress anisotropy could be determined. This experiment clearly shows that whenever stress is present, it has to be taken into account in the interpretation of the equilibrium shape [117]. To complete the discussion of studies with patterned surfaces, the work of Blakely’s group on photolithographically prepared biperiodic surfaces still needs to be mentioned. These surfaces consist initially of a square arrangement of flat top square pillars with a height of 0.1–0.25 μm and a periodicity of 4 μm, separated by flat trenches. When annealed at high temperatures surface diffusion and sublimation produce beautiful characteristic hill and valley structures whose shape can be explained in terms of the processes studied on flat surfaces and described above [118, 119, 120].

Silicon is frequently heavily boron-doped. When B diffuses to the surface it causes considerable tensile strain due to the short Si–B bond. The resulting changes in the surface stress anisotropy can cause the break-up of terraces into periodic step structures due to stress relaxation at the step edges. This is actually the case and has been the subject of several studies. In the first study in 1996 [121] of miscut surfaces the SA steps were smooth and the SB steps were rough at high temperatures (≥1240 K) as on undoped surfaces. With decreasing temperature the SB steps developed a periodic sawtooth shape, leading to triangular islands. The width of these islands at the smooth SA steps decreased continuously with decreasing temperature, resulting in a striped phase, which could not be resolved any longer at 1160 K. The cause of this process was attributed to increasing stress anisotropy with increasing B segregation to the surface but a lowering of the SA step energy by preferential B adsorption could not be excluded. Studies on larger terraces revealed a more complex behavior as illustrated in Fig. 5.19 [122]: striped phases on the same terrace orthogonal to each other, formation of elongated islands with temperature-dependent aspect ratio, which split to form new stripes and nucleation of stripes between existing stripes. These features were clarified later by more thorough studies of Hannon et al. [123, 124]. They derived the step stiffness from step fluctuation spectroscopy and came to the conclusion that the driving force for stripe formation was not surface stress but a strong decrease of the free energy of SA steps with decreasing temperature. The coexistence of orthogonal stripe domains was attributed to vacancy island formation. This suggests that a high vacancy concentration at high temperatures on the B-doped surface is necessary for the mass transport in the stripe formation process. In the studies just discussed pre-existing steps play an important role in the stripe formation.
Fig. 5.19

Striped phase and step shape transition on B-doped Si(100). From (a) to (f) the crystal is cycled from high to low temperature and back again. Due to the different response of the thermocouple and the crystal to temperature changes the temperatures shown are only rough numbers. Adapted from Ref. [122] with permission from Springer Science + Business Media

The lithographically pre-patterned surfaces with large atomically flat surfaces, mentioned earlier, allow avoiding the step effects by adding the material needed for stripe formation from the vapor phase. Initial experiments using diborane as Si supply were only partially successful but deposition of 0.5 monolayers of Si atoms and annealing at 1010 K brought success [125]. This method was later perfected by using higher temperatures at which sublimation creates vacancies and Si deposition creates adatoms. By varying temperature and Si deposition rate a wide variety of microstructures could be formed, including perfect stripe patterns in regions with more than 5 μm diameter or edge length. Stripe pattern formation was limited to temperatures above 1150 K and attributed to stress interactions as in the original interpretation. Upon prolonged annealing at 1120 K the stripe pattern broke up into oval islands with sharp tips seen in Fig. 5.19. Their number and aspect ratio decreased with annealing time via Ostwald ripening, driven by minimization of the step free energy [126].

In connection with B doping it is interesting to note that segregation to the Si(111) surface has quite different consequences for the surface microstructure. No nanofacetting occurs but a (√3 × √3) structure is formed. On this surface Si grows between about 800 and 900 K in double layers in twin orientation with respect to the substrate, above about 950 K in parallel orientation. The twinning allows producing an artificial twinning superlattice but occurs only when the twinned crystallites are small. Small isolated crystals were observed even at 1100 K but as soon as they coalesce they convert into parallel orientation, which explains the loss of twinning in continuous films above about 950 K. This interesting phenomenon was attributed to the reduction of the step energy by twinning, which overcompensates the energy increase from the twin boundary energy [127].

Unless passivated with hydrogen Si is always covered with an at least 1 nm thick SiO2 layer, which is usually removed by flashing in UHV to ≥1050 K. Slow removal leads to pit formation, whose mechanism and kinetics was thoroughly studied by Hibino et al. [128]. They found that pit growth proceeds via reaction of Si monomers with the edge of the pit with an activation energy of 4 eV. There have been several other studied of SiO2 on Si(100) but they were aimed mainly at the understanding of contrast information in LEEM, PEEM, and MIEEM [129, 130, 131], of the interaction of light with extreme UV light [132] and at the determination of the thickness and composition of the layer [133]. Two other studies involving oxygen aimed at the understanding of the etching process by O2 in the 10−8 mbar range. Below 850 K etching was found to occur randomly on the surface, producing vacancy clusters, above 850 K the vacancies diffused to the steps and caused effects similar to those seen in sublimation [134]. At higher temperatures (1000–1250 K) etching leads on large terraces to random vacancy island formation similar to what is shown in Fig. 5.16 during sublimation. The growth rate of the islands was found to depend upon their neighborhood, which shows that it is determined by vacancy diffusion from the environment. A study of the temperature dependence of the growth rate gave an effective activation energy of 2.0 eV for etching, an unexpectedly low value [135].

Finally some other studies involving the clean Si(100) surface and its interface with SiO2 need to be discussed. They concern the influence of interfacial strain on the LEEM contrast and on the stability of SOI (silicon-on-insulator) films, which have been studied in detail by Lagally’s group [136, 137, 138, 139, 140] and Müller’s group [141, 142, 143, 144]. The strain field originating at the film-substrate interface produces lattice distortions concentrated around dislocation lines which cause diffraction contrast in form of diffuse lines. These can be used to characterize the stress [136, 137, 138] and are particularly pronounced when the stress is enhanced by thermal shock [136]. When 10–20 nm thick Si(100) single crystalline films on a 150 nm thick amorphous SiO2 layer on a Si(100) surface are annealed at 1100–1200 K they break up into self-organized nanocrystals (“dewetting”), leaving voids in the Si film behind [139, 140]. This process has been studied with LEEM, in particular with dark field LEEM. Figure 5.20 [142] illustrates the void formation process. The speckle in the image is due to the contrast between small crystals with (2 × 1) and (1 × 2) orientation. The break-up forms a pattern with fourfold symmetry with agglomerated Si at the rim surrounding the exposed SiO2. With increasing dewetting the Si nanocrystals form fingers aligned along the 〈001〉 directions. The details of the kinetics and energetics have been studied in considerable detail [141, 142, 143, 144].
Fig. 5.20

Dewetting of a 22 nm thick SiO layer at 1100 K. (a) Dark-field LEEM images taken with 7.8 eV electrons during the dewetting process. The white arrow marks the nucleation of the void at a defect, the black arrow a nanocrystal distorted by charging. (b) Kinetic Monte Carlo simulation of the process shown in (a). Adapted with permission from Ref. [142]. © IOP Publishing & Deutsche Physikalische Gesellschaft (2011). All rights reserved

5.1.2.3 Other Si Surfaces

Very little work has been done with other surfaces because they are not practically important as the (100) surface or interesting as the (111) surface. The (110) surface has been studied by Swiech et al. [145, 146], motivated by the phase transition from its (16 × 2) reconstruction to the (1 × 1) structure at 1005 K and to understand step energetics and dynamics on a Si surface with lower symmetry. The phase transition was found to have a similar coexistence region as the Si(111) (7 × 7) ↔ (1 × 1) transition, with large aspect ratio (16 × 2) domains coexisting with (1 × 1) regions. The results were analyzed in the same manner as in Ref. [77]. The step stiffness was obtained from island decay measurements of islands on large terraces, which were produced by laser texturing. The islands are elliptical and the decay of their areas between 780 and 982 °C occurred with constant aspect ratio and followed the law dA/dt = A 0(t 0 − t) α discussed earlier with α = 0.65 ± 0.03. This means that diffusion is isotropic in this temperature range and the decay is diffusion-limited. From the shape of the islands the following values for the step energy β and step stiffness \( \tilde{\beta} \) were derived (in units of meV/Å): for the long axis 13.1 and 5.4, respectively, for the short axis 8.4 and 20.5, respectively, i.e., values between those cited earlier for the (111) and (100) surfaces.

The Si(311) surface was of interest because LEED and X-ray diffraction measurements had shown a second order phase transition from a (3 × 1) reconstruction to the (1 × 1) structure at 693 °C, in contrast to the first order transition on the (111) and (110) surfaces. Second order phase transitions show critical fluctuations and are characterized by critical exponents. While it is difficult to determine the critical exponents β, γ, and ν with LEED in LEEM, the critical exponent α of the specific heat, the correlation length ξ of the fluctuations and the critical exponent z of their relaxation time are accessible to LEEM. α can be obtained from the time average of the mean square fluctuations of the image intensity, ξ from the spatial autocorrelation function of the fluctuations g(r) via g(r) ~ exp(−r/ξ) and z from τ ~ ξ z . The relaxation time τ is extracted from the time autocorrelation function of the fluctuations c(t) via c(t) ~ exp(−t/τ). Tromp et al. [147] embarked successfully on this ambitious endeavor and obtained z = 1.9 ± 0.3. The ξ values were limited in one direction \( \left(\left[2\overline{3}\overline{3}\right]\right) \) to 80 nm by steps, while in the perpendicular direction (\( \left[01\overline{1}\right] \)) the instrument resolution gave an upper limit of 40 nm. α could be determined only with large uncertainty. With improved instrumentation, for example with aberration-corrected systems and better data acquisition, the accuracy of such experiments could be improved significantly.

5.1.3 Other Inorganic Semiconductors

Surprisingly little work has been done on the microstructure of other clean semiconductor surfaces. Only for Ge(100), GaAs(100), GaN(0001), and SiC(0001) are limited results available. On Ge(100) Poelsema’s group studied the (2 × 1) ↔ (1 × 1) phase transition, for which laterally averaging studies had given conflicting interpretations such as dimer breakup, step proliferation, or domain wall proliferation without dimer breakup. They interpreted their LEED (1/2,0) beam intensity measurements in the LEEM system in terms of dimer breakup followed by step proliferation, with the dimer concentration decreasing from 100 % at 900 K to 0 at 1100 K [148]. Extension of the LEED measurements to the (00) intensity and to the (00) and (1/2, 0) halfwidths supported this initial interpretation and lead to a new model of the phase transition, which assumes vibrating non-interacting dimers randomly distributed over lattice sites [149]. Although the detailed atomistic picture is not clear, the phenomenological aspects are in general supported by more recent phase field model calculations [150] with two exceptions: (1) while the dimer concentration as expressed by the order parameter decreases initially continuously with increasing temperature, it drops abruptly to zero at about 50 % concentration and (2) it shows considerable hysteresis upon cooling. This phenomenon possibly is not visible in experiment because of surface heterogeneity.

In the case of GaAs(100) only the decomposition [151] and congruent evaporation [152] has been studied, except for a brief MEM/LEEM inspection of a surface with a highly doped overlayer by the author. GaAs decomposes above 900 K in As, which sublimates and Ga, which forms droplets. The first paper [151] served to control droplet formation and find preconditions for congruent evaporation, the second [152] studied congruent evaporation by compensating the loss of As with an incoming As flux.

The GaN(0001) surface was studied only briefly in connection with GaN homoepitaxy. The surface was not from a bulk single crystal but from a 1.4 μm thick GaN film produced by OMVPE (Organo-Metallic Vapor Phase Epitaxy) on a AlN buffer layer on 6H-SiC(0001), the only GaN material available at that time. After cleaning with atomic nitrogen from a radiofrequency discharge at about 675 °C a step and terrace structure or spiral structure was obtained, depending upon substrate used [153]. The surface can be either Ga or N terminated. Distinction between different terminations is possible with dark field imaging with diffracted beams such as the (10) or (11) beams [154].

While GaN is of practical interest mainly for solar cells, SiC is not only important as substrate for nitride solar cells but also by itself as a material for high-power, high-frequency, high-temperature electronics. More recently it has become also important as template for graphene. SiC comes in a variety of polytypes, differing in the SiC bilayer stacking. They are characterized by their repetition frequency n and symmetry (H, C for hexagonal and cubic), with 4H-SiC and 6H-SiC the most used types. The first LEEM studies were made by Pavlovska et al. [155] in connection with GaN heteroepitaxy on 6H-SiC(0001) crystals, which had been etched at about 1600 °C in H2. In this study the various LEED patterns previously reported were correlated with LEEM, which showed a large variety of structures, depending upon heat treatment. Some of the images are shown in Fig. 5.21 [155]. The contrast differences between different regions are attributed to different stacking sequences. Imaging of larger areas with FEL-PEEM revealed beautiful patterns attributed to screw dislocations [156].
Fig. 5.21

LEEM images of a clean 6H-SiC(0001) surface showing domains with different surface terminations. (a) 13.0 eV; (b) 14.5 eV and (c) 15.5 eV. Reproduced with permission from Ref. [155]. Copyright 2002 Elsevier

More detailed LEEM studies of the SiC(0001) surface were made later in connection with its application for the growth of graphene [157, 158, 159]. In these studies the crystal was cleaned in disilane at a much lower temperature (700–800 °C), which produced larger terraces and apparently fewer stacking faults. Both 4H- and 6H-SiC(0001) surfaces were studied. On the 6H-SiC surface the same sequence of LEED and LEEM patterns as reported previously [155] was found. High resolution images of the step roughening observed previously and the subsequent growth of the (6√3 × 6√3)-R30° layer, the “buffer layer” in graphene growth, gave a much deeper insight in the graphitization of the surface than the earlier work [155]. Another study explored the process of polytype formation during homoepitaxial growth on the 4H-SiC. Upon annealing cubic phase monolayer islands formed, which grew upon further annealing, displacing the hexagonal phase and causing step bunching and thus increasing surface roughness, a technologically detrimental effect. This phenomenon was interpreted in terms of the various surface terminations possible on the 6H-SiC(0001) surface, based on first principle calculations of their surface energy [158]. The “softer” cleaning with silane compared to H2 produces apparently wider terraces which made local LEED from single 200 nm wide terraces possible. The wider terraces allowed μLEED of the 4H-SiC(0001) surface and its atomic structure with a dynamical LEED analysis of the I(V) curves [159].

This surface was also subject of a MEM study aiming at the understanding of tilted stacking faults in a homoepitaxial layer, which are detrimental in electronic devices. While pure MEM could not image them due to surface charge fluctuations, simultaneous illumination with light with energy slightly above the band gap clearly revealed the stacking faults. This was attributed to charge generation and transport in the layer, eliminating the surface charge. For contrast between stacking fault and surrounding, generation and/or transport must obviously be different [160].

5.1.4 Other Inorganic Compounds

Amongst the many inorganic materials that have been studied by surface microscopy with slow electrons, only two of them have been subject of thorough surface microstructure studies, titanium nitride and titanium oxide. Many compounds are not suitable for these methods, either because of charging or radiation damage by electrons and photons, such as halides and many oxides, or are not available in sufficiently perfect, large, and clean crystals, or are not of sufficient practical interest. Therefore the discussion of inorganic compounds in this section is limited to TiN and TiO2, with a few comments on CeO2.

TiN is a corrosion resistant, hard refractory compound (melting point T m = 3200 K) with NaCl structure. Changes of its surface microstructure occur at similar high temperatures as on the refractory metals discussed in Sect. 5.1.1 and the experimental procedures and goals are similar. In contrast to the alkali halides the (111) surface is the most stable surface and has been studied on thick epitaxial layers, whose surface structure is believed to be representative for that of bulk crystals. Like in the case of metals and semiconductors TiN surfaces had been studied before at lower temperatures (1050–1250 K) with STM. LEEM served to extend the measurements on the (111) surface to higher temperatures (1500–1700 K) at which different processes can determine the surface microstructure but at which sublimation is still negligible. Step–step interactions, step permeability, step mobility, and the possible contributions of bulk mass transport processes were deduced from the decay of conical stacks of bilayer-height islands [161]. The details of the decay process showed high step permeability, strong temperature-dependent step–step repulsion and attachment–detachment limited decay. From the temperature dependence of the island area decay an activation energy of 2.8 ± 0.3 eV was obtained, which is much smaller than that for bulk diffusion so that surface diffusion is the rate-controlling process. In an extension of this study the anisotropy of the step mobility Γ, which is a measure for the rate of attachment–detachment of atoms at the step, was deduced from island decay rate data. Γ is intimately related to the step energy β because the attachment–detachment probability increases with β, so that Γ is orientation-dependent when β is direction-dependent. Theory shows that Γ(φ) ~ β(φ)dA/dt. With the proportionality factor and β(φ) known from other experiments, Γ has been obtained from the angular dependence of the decay rate of the inequivalent \( \left\langle 1\overline{1}0\right\rangle \) steps of the strongly rounded triangular islands at 1560 K [162].

While the last two studies showed surface diffusion as the dominating mass transport mechanism, a study of dislocations terminating at surfaces [163] revealed that in the same temperature range bulk diffusion was causing microstructure changes as illustrated in Fig. 5.22 by the evolution of a screw dislocation (a) and a dislocation loop (b) intersection with the surface. In (a) the spiral step has made one full turn in the image sequence, the dislocation core has climbed one unit step height normal to the surface and then the process repeats itself with constant unwinding frequency ω. In (b) the spiral steps originating at the two end of the dislocation loop rotate in opposite direction and form a complete loop, which subsequently expands while the previous step repeats itself with constant period τ. A comparison with other step movements showed that the spiral in (a) is moving inward and that loop in (b) bounds a vacancy island. Because there was no sublimation this was explained by vacancy diffusion from the bulk. This was supported by measurements of ω at several temperatures T and plotting ω(T) = ω 0 exp(−E d/kT), which gave E d = 4.5 ± 0.2 eV and ω 0 = 1012 s−1. The high value of E d is typical for bulk processes such as point defect formation and diffusion, so that the change of the surface microstructure in this case is due to vacancy diffusion along the dislocations [163]. In order to ensure that sublimation made a negligible contribution to spiral step formation the experiments were repeated as a function of N2 pressure up to 5 × 10−7 mbar. While ω was found to decrease slightly with time from cycle to cycle, E d had the value mentioned before, independent of N2 pressure [164]. Of course this phenomenon is limited to materials which can sustain a high concentration of vacancies.
Fig. 5.22

Nucleation and growth of steps from a screw dislocation at 1688 K (a) and from a dislocation loop at 1653 K (b) on TiN(111) during annealing in N2.without deposition and evaporation. The LEEM images show the time evolution of the bilayer high steps. The step edges are pinned at the core(s) of the dislocation(s), which in (b) have opposite sign. The loop generation process observed in (b) is shown schematically in (c). Adapted from Ref. [163]. Permission from Nature Publishing Group (UK)

TiO2 also has a significant homogeneity range though not as wide as TiN, so that surface microstructure changes due to mass transport between volume and surface can be expected at elevated temperatures. Actually, changes of the microstructure of the (110) surface of the rutile phase of TiO2 were studied before those of TiN. In a series of experiments McCarty and Bartelt developed a very comprehensive picture of the processes occurring on this surface at temperatures between about 690 and 1230 K as a function of the deviation from exact stoichiometry [165, 166, 167, 168, 169]. The specific problem addressed was the transition between the (1 × 1) and the (1 × 2) surface terminations, which served as probe for the mass transport and can be easily distinguished in LEEM at selected energies. The crystals were heated in UHV to several high temperatures to reduce the oxygen content from 2 to values between 1.9977 and 1.9958. The phase transition kinetics was studied under several experimental conditions: isothermal annealing after quenching, heating, cooling, oxidation after reduction for several surface geometries such as step density, pits, and multilayer islands. Of the wealth of results only a few can be mentioned. The phase transition is always first order. Under isothermal conditions it occurs mainly via mass redistribution across the surface, upon heating or cooling mass flow between surface and bulk is dominating except on wide terraces, on which also the first process is involved. During oxidation of oxygen-deficient crystals surface restructuring occurred via periodic transitions between (1 × 1) and (1 × 2) termination, either in the step flow growth mode without nucleation (at high temperatures) or with two-dimensional nucleation and layer-by-layer growth (at low temperatures), with the details depending on the terrace size. Figure 5.23 [168] illustrates the oxidation of an oxygen-deficient crystal at 547 °C in 1 × 10−6 mbar O2. The surface oscillates with time between (1 × 1) termination (bright) and (1 × 2) termination (dark). For the many interesting details the reader is referred to the original publications.
Fig. 5.23

Bright field LEEM images taken during the oxidation of slightly reduced a TiO2 (110) surface in 1 × 10−6 mbar at 547 °C, starting from a (1 × 1) structure. At the electron energy used for imaging the (1 × 1) structure appears bright, the (1 × 2) dark and the surface oscillates between these two structures due to transport of Ti from the bulk and reaction with oxygen without step movement at these specific experimental conditions (see double arrows). Adapted with permission from Ref. [168]. Copyright 2003 Elsevier

In the work just discussed the (1 × 2) structure was intimately connected with oxygen deficiency in the bulk and produced by cooling an oxygen-deficient crystal. In another approach Menteş et al. [170] produced the (1 × 2) on stoichiometric TiO2(110) surfaces by electron-stimulated desorption (ESD) at about 720 K, at which the surface vacancies order in the (1 × 2) structure. In contrast to the added row model of this structure, which is generally accepted when produced in the manner discussed before, the (1 × 2) structure produced by ESD was interpreted in terms of the missing row model, based on a number of indirect indications. A later comparative analysis of the LEED I(V) curves of the (1 × 2) structure produced by the two methods, however, suggested that the ESD-generated (1 × 2) structure is also of the added row type. The (1 × 2) structure has also been produced by photon-stimulated desorption with synchrotron radiation [171] and studied with LEED and LEEM. Light from a frequency doubled Ti:sapphire laser was also used to produce the (1 × 2) structure in a UVPEEM study [172].

5.2 Adsorption

Adsorption layers can form either from the gas phase or by diffusion from the bulk. They are frequently the precursors of the growth of three-dimensional phases but differ so much from them that they will be treated here as a separate phase. Their interaction with the surface imposes usually a structure on them different from the bulk, in particular also a different electronic structure, and vice versa modifies also the properties and structure of the substrate, for example by reconstruction or by more extensive changes such as alloying and faceting. Many of the results presented in this section have been reported in connection with film growth but are included here for the reasons just mentioned. Organic adsorbates, however, will be discussed in connection with organic layers.

5.2.1 Adsorption on Metals

5.2.1.1 Nonmetallic Adsorbates

The adsorption of O2 on metal surfaces is a good example of the application of the various methods and of the variety of phenomena, which can be studied with them. The simplest phenomenon is the work function change ΔΦ caused by adsorption, which can be studied well with UVPEEM via the change of the emission current caused the work function change. PEEM has been used to image surface diffusion of oxygen on Ni [173], to study the influence of adsorption on various surface orientations of polycrystalline Cu and Ti surfaces [174] and to extract some information on the various binding states of oxygen on Pd(100) [175]. Faceting upon oxygen adsorption at high temperatures is a well-known process on refractory metal surfaces but only one attempt was made to study it with LEEM, with limited success because of the facet size on the crystal studied, Ir(210), were at the resolution limit of the instrument used [176]. LEEM played a limited but important role in some LEED studies. Examples are the structure analysis of the Ag(111)-(4 × 4)O structure, which required imaging for selection of a region with a minimum number of defects for microspot I(V) analysis [177] and studies of the strain in the W(110)-p(2 × 1)O structure as a function of p(2 × 1)O island size [178] and temperature [179], which combined with first principle calculations allowed to determine the associated stress. In this work it was necessary to select step-free regions for quantitative spot position and shape analysis. Figure 5.24 [179] illustrates the usefulness of LEEM not only for this selection but shows also that adsorbate domains are not necessarily bounded by surface steps.
Fig. 5.24

LEEM images taken with 35 eV electrons of a W(110) surface covered with 1/2 monolayer oxygen with (1 × 2) structure. (a) Bright field image, (b) and (c) dark-field images taken with the (0,1/2) and the (1/2,0) beams of the two domains of the (1 × 2) structure, respectively. The domains extend frequently across the monatomic steps. Adapted with permission from Ref. [179]. Copyright 2010 by the American Physical Society

LEEM plays a central role whenever obtaining the desired information requires surface-sensitive imaging. This is the case, for example, for the understanding of the later stages of oxygen adsorption on W(100) at high temperatures. This system had been studied earlier by LEED and other techniques. However, many of the structural changes occurring with increasing coverage, in particular between the (5 × 1) pattern at 1.0 ML and the saturation of the two-dimensional layer with p(2 × 1) and p(2 × 2) structure at about 1.25 ML could not be explained using laterally averaging methods. The first LEEM studies [180, 181] provided an explanation (Fig. 5.25). The clean surface with monatomic steps is shown in Fig. 5.25a. At the beginning of the transition small islands, identified by LEED as p(2 × 1) islands, grew rapidly into random chains of small islands, which were unrelated to the step structure of the substrate (Fig. 5.25b). They became more intense and grew laterally over the surface while simultaneously developing a p(2 × 2) structure. When this structure completely covered the surface at 1.25 ML the substrate steps became visible again (compare Fig. 5.25a, c). Dark field LEEM with (1/2,0) spots indicated that the chains grew at the domain boundaries between (5 × 1) and (1 × 5) domains, initiating the transition from a chemisorbed layer to a two-dimensional oxide. A detailed analysis of the LEEM images showed that the two (5 × 1) domains formed a two-dimensional percolating system [180, 181]. The formation of the chemisorbed oxygen layer and of the two-dimensional oxide was studied also briefly with LEEM on the Ru(0001) surface as a preliminary to CO oxidation on this surface [182].
Fig. 5.25

Bright field LEEM images taken from a study of the oxidation of the W(100) surface at 1060 K in 2.5 × 10−8 mbar O2. (a) Clean surface before oxidation (7.0 eV), (b) after exposure to 5.6 × 10−6 mbar s O2 (16 eV), (c) after completion of the oxidation (16 eV). The dark regions in (b) are from two domains with a (5 × 1) structure, the white chains decorate the domain walls and have p(2 × 2) structure. In (c) monatomic steps are visible again. Adapted with permission from Ref. [181]. Copyright 1996 Elsevier

On the close-packed refractory bcc (110) surfaces oxygen induces at high coverage and high temperatures complex superstructures with two domains corresponding to the twofold symmetry of the surface, which can be imaged with dark field LEEM similar to the p(2 × 1) domains shown in Fig. 5.25. On well-oriented surfaces the two domains form with equal probability but on Nb(110) films grown on Al2O3 \( \left(11\overline{2}0\right) \) this symmetry is broken in a temperature-dependent manner due to the strain in the films caused by the anisotropic thermal expansion of the substrate [16]. Near the disordering temperature a stripe pattern was observed which was attributed to the interaction of the strain introduced by the substrate with the anisotropic surface stress of the film [183].

Most of the work on CO adsorption was connected with CO oxidation which will be discussed in Sect. 5.2.1.2 but CO alone or with H2 has been studied too. Rotermund et al. [184, 185, 186] developed in the early 1990s a method for studying CO surface diffusion by desorbing CO locally from a saturated CO layer on Pt(110) and Pd(111) using a laser pulse and recording the subsequent filling of the region with reduced CO coverage with PEEM via work function contrast. This method, called laser-induced thermal desorption (LITD) was later applied in significantly improved version to the study of CO diffusion on Pt(111) using the LEEM intensity at the electron energy of maximum sensitivity to CO for local coverage measurement [187]. The work function change deduced from the MEM–LEEM transition energy served only to check the accuracy of the LEEM calibration. A detailed analysis of the time dependence of the coverage profile after LITD, which desorbed 0.06 ML in the center from an initial coverage of 0.31 ML, allowed to determine the coverage dependence of the CO diffusion coefficient between 0.25 and 0.31 ML. Beyond the demonstration of the use of LEEM for this type of measurement a comparison of the LEEM images before and after LITD showed that the laser pulse needed for desorption of a measurable amount of CO caused considerable surface damage, an effect which would be difficult to be seen with PEEM [187]. Another study of CO adsorption with LEEM was concerned with the interaction with hydrogen on Pd(111). Here LEEM served to illustrate the displacement of adsorbed hydrogen from the surface to below the first Pd layer, making use of low energy reflectivity differences. The structural details were obtained from I(V) curves and first principle calculations [188].

Sulfur adsorption by segregation from the bulk is an unwelcome process encountered with many metals, in particular also with Cu. In a LEEM study of intentional S adsorption on a well cleaned Cu(111) surface an interesting phenomenon was discovered: Cu surface self-diffusion was enhanced several orders of magnitude by less than 0.01 ML of S. This was deduced from monolayer island decay and coarsening, similar to the diffusion studies on clean surfaces discussed in Sect. 5.1 [189].

As a final example of the use of imaging in adsorption studies the PEEM study of Xe on graphite should be mentioned. Xe monolayer adsorption required cooling to 60 K and for photoemission a windowless low-pressure hydrogen discharge source was needed. The photo ionization cross section of Xe for the 10.2 eV photons from this source is very high, producing high contrast between graphite and Xe. From variations of the contrast, shape, and size of wave-like patterns as a function of Xe exposure, temperature and during desorption, conclusions regarding the phase of the Xe layer were drawn [190]. LEEM combined with diffraction, of course, would be much better suited for such a study.

5.2.1.2 Coadsorption and Reaction: Catalysis

Imaging of the coadsorption and reaction of gases on metal surfaces has long been and still is an exciting playground for surface microscopists and a gold mine for surface chemists. It reveals a large variety of time-dependent patterns depending upon fine details of the partial pressure of the adsorbed gases and the temperature, which are summarized under the term spatiotemporal patterns. They include periodic plane waves, soliton waves, spherical waves, spiral waves, turbulent and chaotic states, and other configurations, which under the same conditions vary from crystal plane to crystal plane. The field was started in Ertl’s group in the late 1980s, first with scanning PEEM [191] and has reached a high level of sophistication since. Initially PEEM was the main method because the first LEEM used in these studies had some technical problems which made the reflection methods LEEM and MEM difficult to use. Later LEEM and XPEEM made important contributions too in contrast to MIEEM, which was also tried [192]. The basics of surface reactions and the early phase of imaging have been reviewed by Imbihl and Ertl [193] and later developments are covered in several reviews [194, 195, 196, 197]. The literature in this field is too extensive so that we will restrict ourselves to a few general remarks, a few examples, and the most recent literature.

CO oxidation is by far the most intensely studied reaction followed by the water formation reaction (H2 + O2) and the reaction between NO and H2. The low index planes of Pt ((110), (100), (111)) are usually used for the CO oxidation, the Rh (110), and (111) surfaces for the other two reactions. Most of the work has been done and is being done with PEEM—a beautiful example of this work is shown in Fig. 5.26 [198]—but LEEM is increasingly used too and in cases, in which chemical information is essential, also XPEEM. It has to be kept in mind that the substrate is not static but rather is modified more or less by the reaction occurring on it. These changes are in general below the resolution limit of PEEM but in some cases occur over large enough regions to be detectable. An example is the formation of subsurface oxygen in the CO oxidation on Pt(110) in excess oxygen, which occurs over large enough regions to become visible via the more positive dipole moment [199]. In general, however, LEEM is the method of choice as illustrated by the same reaction on the same surface [200].
Fig. 5.26

UVPEEM images of the evolution of O2 + CO reaction spirals on a Pt(110) surface at 448 K, 4 × 104 mbar O2 and 4.3 × 10−5 mbar CO. The spirals start at defects and have different rotation periods and wave lengths. The image sequence shows how the faster rotating spirals with shorter wave length annihilate the slower one in the center. Adapted from Ref. [198] with permission from the American Institute of Physics, © 1993

LEEM images not only the microstructure but via μLEED also the local crystal structure. This is particularly useful when—as is often the case—the reacting species produce characteristic LEED patterns. In this case dark-field imaging with LEED spots of the various species make LEEM a chemically sensitive imaging method. The fast image acquisition of LEEM allows following the dynamics of each reactant, which is not possible with XPEEM because of its longer image acquisition time. As an example Fig. 5.27a–d shows dark field LEEM snapshots of the propagation fronts in the NO + H2 reaction on Rh(110) [201]. The images are taken with LEED spots corresponding to different compositions. Following the bright regions with time gives a vivid picture of how the reaction proceeds in space. The time sequence can be recorded quantitatively by selecting a small region of the crystal with the field limiting aperture and recording the intensities of the various LEED spots. This is illustrated in Fig. 5.27e. The time scale is obtained from the speed of the reaction fronts in the LEEM video [201].
Fig. 5.27

LEEM and LEED study of chemical reaction waves in the NO + H2 reaction on Rh(110). (ad) Dark-field images taken with diffraction beam characteristic for the various phases occurring in the reaction system about 1 min apart at 550 K, 3.5 × 10−7 mbar NO and 2 × 10−6 mbar H2. (e) LEED intensities of the characteristic diffraction beams acquired simultaneously during propagation of the reaction wave at 530 K, 3.2 × 10−7 mbar NO and 4.5 × 10−6 mbar H2. Adapted with permission from Ref. [201]. Copyright 2000 Elsevier

In cases in which the species involved in the surface reaction do not have characteristic LEED patterns XPEEM is called for. This can be done in the XANES-PEEM or threshold ionization PEEM mode, which requires resetting the photon energy for each species but has high intensity from the high secondary electron yield and does not need an energy filter. In instruments equipped with an energy filter it is more convenient to image with characteristic photoelectrons because only one photon energy is needed, which is chosen so that all species are ionized with sufficient probability. An example, in which XPEEM was needed for chemical analysis, is the multicomponent system of the water formation reaction on Rh(110) partially covered with Au and Pd, in which Au is passive and Pd active in the reaction [202, 203, 207]. The question arises how these two species are distributed. During reaction the lamellar structures shown in Fig. 5.28 [202, 203] form, which spread again upon reduction and become asymmetric in width upon oxidation. The composition of the various regions in the various states can be determined with XPEEM. In the state depicted in Fig. 5.28 it is evident that the oxygen is adsorbed in the Rh regions ((c) and (f)), that Au (a) and Pd (b) form alloy lamellas, which have a high reflectivity in LEEM, and that the higher work function in the oxygen regions (f) cause a higher reflectivity in MEM (d) [203].
Fig. 5.28

XPEEM, LEEM, and MEM images of the stationary patterns resulting from the water formation reaction on the Rh(110) surface originally uniformly covered with 1/2 monolayer of AuPd. Reaction parameters: 880 K, pH2/pO2 = 0.45, 5.3 × 10−7 mbar total pressure. The XPEEM images taken with the photoelectrons shown below the images show that Au and Pd form alloy stripes separated by oxygen-covered Rh regions, which are bright in MEM because the higher work function. The LEEM image taken with 8 eV electrons shows diffraction contrast. The \( \left[1\overline{1}0\right] \) direction is normal to the stripe direction. Adapted with permission from Ref. [202]. Copyright 2006 American Chemical Society and with permission from Ref. [203]. © IOP Publishing 2008. All rights reserved

Most work in the past was concerned with the three reactions on clean single crystal surfaces discussed up to now. More recently the attention has turned to the influence of catalytically active and/or passive adsorbates on these reactions as illustrated in the previous example and to catalysis on clean polycrystalline surfaces. The first development, advanced mainly by Locatelli, Kiskinova, Imbihl et al. [202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, 214], is motivated by the desire to tailor the activity and selectivity of the basic catalyst with co-adsorbates, the second route, followed mainly in Schlögl’s group [215, 216, 217, 218, 219], by simulating pure real catalysts, which always expose several surface orientations. Depending upon the reaction system and the instrument used in the study, the experiments have been performed between 500 and 1000 K at pressures ranging from 10−7 to 10−5 mbar, in the simple PEEM system used in [215, 216, 217, 218, 219] up into the low 10−4 mbar range.

The initial studies of the first group looked at the influence of an inhibitor species, Au, on the O2 + H2 reaction on Rh(110) [204, 205] and were then extended to include also an activator species, Pd, simulating a bimetallic catalyst [202, 206, 207]. The lamellar structure shown in Fig. 5.28 is only one of many found at different stages of the reaction. Another effort aimed at the understanding of the influence of coadsorbed K, which is known to be a strong promoter in practically important reactions, on pattern formation, in this case in the NO + H2 reaction on Rh(110) [208, 209, 220]. As in the case of Au and Pd, phase separation was observed, in this case not into pure K regions but into K + O regions due to the strong K–O interaction. Similar redistributions of catalytically active coadsorbates, called reactive phase separation, on single crystal surfaces have been observed in the O2 + H2 reaction on Rh(111) with vanadium oxide, also an industrially important catalyst in form of VO x (x ≈ 3) [210, 211, 212]. A more complicated situation was encountered in the O2 + H2 reaction on a bimetallic Rh(111)-Ni catalyst prepared by alloying the pure Rh surface with an unknown amount of Ni. Here the reaction produced many small three-dimensional NiO crystals, which acted as obstacles to the propagation of the spiral wave excited by the addition of Ni. By reducing the pressure and increasing the H2:O2 ratio the NiO crystals could be reduced again leaving an O-free alloy surface. Thus the NiO crystallites, believed to be catalytically inactive in this reaction, acted as reservoir of Ni, which could be brought to the surface by coadsorption of a proper amount of oxygen [213, 214]. An even more complex situation was found when VO x and Ni were present simultaneously in this reaction on this surface [214], which illustrates the challenges of simulating catalysis on real surfaces.

The second route in recent catalysis experiments, the study of reactions on polycrystalline surfaces, allows comparing different surfaces under identical conditions. Pt is particularly well suited for this purpose because polycrystalline foils with regions, which are sufficiently large for PEEM studies, can be prepared easily. The regions with different crystal orientation are usually tilted somewhat relative to each other so that PEEM is the preferred imaging method. In LEEM this tilt requires realignment or moving the contrast aperture when moving from grain to grain, which makes a quantitative comparison of the various surfaces difficult. The CO oxidation is an ideal reaction for this study because of the many results from single crystal studies, which can be used for comparison, in particular for the determination of the local coverage via the image intensity. Figure 5.29 illustrates how the contrast in the PEEM images of such a foil changes from orientation to orientation and with adsorbate [216]. Though the spatial resolution is poor, simultaneous data acquisition from several orientations is fast. This allows efficient determination of different reaction states such as high and low reactivity and bistability regions [215, 216, 217, 218]. The method was also applied to a Si-contaminated polycrystalline Pd foil. The adsorbed oxygen enhanced Si segregation to the surface, resulting in silicone oxide formation, which drastically inhibited CO oxidation and its dependence on crystal orientation [219].
Fig. 5.29

UVPEEM images of a clean, CO-covered, and O-covered polycrystalline Pt foil. The contrast varies not only with surface orientation but also slightly between surfaces of the same orientation due to step density differences, which influence the work function. These differences influence also adsorption as seen in the strong intensity differences between different (110) surfaces upon CO adsorption. Reproduced from Ref. [216] with permission from Springer Science + Business Media

Recently other reactions have been added to the three classical ones discussed up to now, the NH3 + O2 reaction on Rh(110) [221] and the reaction of ethylene with O2 on clean and Ag submonolayer-covered Ni(111) [222]. In the first case an interesting reaction front modulation was found. In the second experiment Ag increased the reactivity and shifted the stability of the reaction product to higher temperatures.

Of course, from the practical point of view all these studies suffer from the “pressure gap” which separates the low-pressure experiments from the high-pressure conditions of technical catalysis and which is frequently used to question the usefulness of these studies. Nevertheless these experiments make important contributions to the fundamental understanding of catalysis, in particular when complementary imaging and diffraction methods such as UVPEEM, LEEM, LEED, and XPEEM are combined.

5.2.1.3 Metallic Adsorbates

Metal adsorption on a clean metal surface under UHV conditions was imaged already in 1985 by Bethge et al. [223]. They studied the diffusion of Ag at submonolayer coverage on polycrystalline Ni with UVPEEM and correlated image intensity with Ag coverage using Auger electron spectroscopy, which also was used to characterize the cleanness of the surface. Although the first LEEM image of a metal adsorbate on a metal, Ag on Mo(110), was published in the same year it served mainly to demonstrate the usefulness of LEEM for such studies [4]. In the late 1980s the interest turned to submonolayer Cu and Au films on Mo(110) and W(110). Mundschau et al. imaged step decoration and step flow of Cu on Mo(110) with LEEM and PEEM [7], studied with LEEM the misfit flip transition [71, 103, 224] and the step roughening [8, 225] in the Cu double layer on Mo(110). On the W(110) surface the first two Cu monolayers have a different structure despite nearly identical lattice constants of W and Mo. Their formation, structure, and stability were studied by Altman et al. [226, 227]. A more detailed study of the evolution of these layers by Koshikawa et al. [228, 229] revealed additional features before the formation of three-dimensional crystals. Contrary to Cu, Au forms on Mo(110) a striped phase in the sub-monolayer range [71, 225, 230]. Later this growth mode was studied also on W(110) briefly by Duden [231] and thoroughly by de la Figuera et al. [232]. These studies stimulated the work of Menteş et al. and Stojić et al. on stress-induced stripe formation in Pd/W(110) [233, 234] and its stabilization by coadsorption of oxygen [235]. Figure 5.30 shows a few examples of submonolayer stripe formation of metals on bcc(110) surfaces. On Mo(110) Au stripes grow in the \( \left[1\overline{1}2\right] \) directions (a), on W(110) in the \( \left[1\overline{1}0\right] \) direction (b, c) [231] and Pd stripes on W(110) grow in the \( \left[1\overline{1}0\right] \), \( \left[1\overline{1}2\right] \) or [001] directions, depending upon oxygen coverage (d–f) [235]. The comparison of (b) and (c) shows that stripe formation is very temperature-dependent. The transition from stripes to elongated islands is similar to that on B-doped Si(100) shown in Fig. 5.19 [122], although in opposite direction. The mechanism is however the same: anisotropic surface stress and temperature-dependent boundary free energy. In contrast to Si(100), Au on W(110) shows no long range order in the striped phase as determined by LEED diffractometry [236]. Thus, long range order is not a necessary condition for stripe formation. This is in accord with the conclusion drawn from the detailed theoretical analysis [232] that the stripes are in the diffuse-interface limit of surface stress domains. A similar situation was found in the Pd on W(110) study [233]. The discovery that coadsorption of oxygen can be used to select orientation, width, and periodicity of metal stripe patterns [235] has opened the door to surface stress engineering of linear nanostructures and has actually already been used to produce ferromagnetic nanowires (see Chap.  7). Surface stress-induced pattern formation has been studied theoretically in several newer papers [237, 238, 239] in which also references to the pioneering work in this field can be found. A theoretical analysis of the influence of the elastic anisotropy of the substrate can be found in [240, 241]. The conditions for the formation of the herringbone structure seen in Fig. 5.30e are discussed in [237, 238].
Fig. 5.30

LEEM images of stripes phases at submonolayer metal coverages (a) Au/Mo (110), stripes parallel to \( \left\langle 11\overline{2}\right\rangle \); (b and c) Au/W (110), 0.3 monolayers at 1045 and 1015 K, respectively; (df) Pd/W(110) with 0, 0.2, and 0.3 monolayer oxygen coverage and 0.4–0.5 monolayers Pd, (d, e) at 1240 K, (f) quenched to room temperature. The dark metal stripes are in (a) and (e) parallel to \( \left[1\overline{1}2\right] \), in (bd) parallel to \( \left[1\overline{1}0\right] \) and in (f) parallel to [001]. The curved lines are monatomic steps. (a) E. Bauer, M. Mundschau et al., unpublished; (b, c) Adapted from Ref. [231] with permission from T. Duden; (df) adapted from Ref. [235] with permission from Europhysics Letters

Up to a temperature-dependent coverage metal atoms form initially a two-dimensional gas before nucleation of two-dimensional crystals. The two phases can be distinguished easily via the difference of the work function, the LEED background intensity, or also by the LEEM intensity caused by the attenuation of the (00) beam intensity due to the diffuse scattering of the adatoms. de la Figuera et al. [242] have used the latter method to determine the gas—condensate boundary for Ag on W(110) and in addition the Ag desorption energy, both in good agreement with earlier work function and thermal desorption measurements. Yasue et al. [243, 244] used the diffuse scattering contrast to understand abnormalities in the step contrast of W(110) covered by Pb gas. Little use has been made otherwise of this potential of LEEM and via the work function contrast in PEEM.

While the close-packed bcc(110) surfaces are relative stable in the presence of adsorbates, more open low index bcc surfaces are not. An extreme response to metal adsorption is encountered on the (111) surfaces. From field emission microscopy studies it has been known for some time that the W(111) surface facets upon Au and Pd adsorption or annealing at high temperatures into {112} planes, which have a lower surface energy. Later STM studies showed that also other metals have the same effect. Contrary to STM, LEEM, and LEED have the advantage that the faceting can be studied in real time as a function of metal coverage or temperature but have been used for only one metal, Pt [245, 246]. The bright field imaging and limited resolution used, however, did not deliver much information beyond that obtained from STM.

The bcc(100) metal surfaces are more stable than the (111) surfaces but can also rearrange easily as illustrated by the reconstruction caused by hydrogen adsorption. Many metal atoms produce at half a monolayer coverage c(2 × 2) structure, which had been attributed originally to atoms on top of the surface with (1 × 1) periodicity but later to two-dimensional ordered alloys. Altman’s group has extended the original LEED work to LEEM studies of Cu, Ag, Au, and Pd on Mo(100) and combined them with first principle calculations, which led to a deep insight in the kinetics and energetics of two-dimensional alloy formation [247, 248, 249]. While most of the information was extracted from LEED, no I(V) structure analysis was needed to demonstrate alloy formation because of the evolution of the LEEM images with coverage. This is illustrated in Fig. 5.31 for the case of Ag [249], which is representative for the other atoms. In order to form a two-dimensional alloy there must be an exchange between adsorbed and substrate atoms. This leads to alloy monolayer islands on top of the Mo surface and alloy regions in the Mo surface with the same composition, which are visible by step contrast. If pure c(2 × 2) islands would form on top of the (1 × 1) Mo surface, they would show diffraction contrast at least at some energies. The islands reach the maximum density at 0.5 monolayers. Al lower coverage step faceting and an island-denuded zone around the steps is seen (Fig. 5.31b) suggesting the two-dimensional alloying occurs also by place exchange at the steps. The many aspects of this alloying–dealloying process are discussed in great detail in the papers cited.
Fig. 5.31

LEEM images of the two-dimensional alloy formation of Ag on Mo(100) at 860 K. (a) 0, (b) 0.16, and (c) 0.38 monolayers Ag. For explanation see text. The dark lines are monoatomic steps. Electron energy 6.0 eV. Adapted with permission from Ref. [249]. Copyright 2007 Elsevier

A second metal on which surface alloying has been studied thoroughly is Cu. For comparison with the bcc(100) surface just discussed we start with the (100) surface, which, being fcc, is more densely packed. The apparently simplest adsorbed metal is Pb because it does not form bulk alloys with Cu. However, on the surface a variety of two-dimensional alloy phases exist. The LEEM studies of Kellogg’s group [250, 251, 252] have revealed the evolution of the various phases as function of Pb coverage and temperature and are an excellent demonstration of the power of LEEM in surface microstructure analysis. They have shown that the three superstructure phases known from previous LEED studies involve initially alloying followed by two stages of dealloying, each with specific LEED pattern and microstructure, which has led to a new interpretation of the alloying process. Too many images are needed to illustrate this process so that the reader is referred to the original publications, in particular to Ref. [250].

While Pb does not form a bulk alloy with Cu because of its much larger atomic diameter, which is also at least partially responsible for the complex surface alloy, Pd is completely miscible with Cu. Here a quite different question was addressed with LEEM and dynamical theory LEED I(V) analysis: how does the alloy composition evolve during Pd deposition at temperatures, at which no diffusion into the bulk occurs? [253, 254]. Because of the different backscattering cross sections of Cu and Pd the reflectivity of the surface is very sensitive to the Pd distribution normal to the surface so that it can be extracted from a careful I(V) analysis. While Pd is incorporated all over the surface, including at the steps, Cu atoms, which have been displaced on the terraces by Pd, not only form islands but also diffuse to the steps, where the largest changes occur in a narrow region close to the steps. Understanding these changes requires LEEM images with the highest lateral resolution as a function of energy for local I(V) analysis of the (00) beam. With the usual trial and error approach in LEED—dynamical LEED I(V) calculations for model distributions and comparison with experimental I(V) curves—it was found that at a Pd coverage of about 0.5 monolayer most Pd was in the second layer, while the top layer consisted almost completely of Cu (“buried Pd layer”). The Pd distribution around the step indicated effective Pd–Pd repulsion, resulting in step overgrowth, which was identified as the major cause of compositional heterogeneity in thin film growth.

The studies just discussed were made at a temperature (200 °C) at which diffusion into the bulk can be neglected. Pd diffusion into the bulk was addressed in a subsequent LEEM study [255] by annealing the buried surface alloy layer with c(2 × 2) structure at temperatures up to 370 °C. From the image intensity decay at an energy characteristic for this structure as function of temperature an activation energy for diffusion from the alloy into the bulk of only 1.6 eV was obtained. A dynamic LEED I(V) analysis of the image intensity changes near the steps gave a Pd distribution similar to that in an ordered Cu3Pd phase known from the bulk. Another LEEM study of the system Pd-Cu(100), combined with extensive first principle and kinetic Monte Carlo calculations looked at the influence of the buried Pd c(2 × 2) layer on the surface self-diffusion on the Cu(100) surface [256]. In this study the decay of Cu monolayer islands deposited onto the Pd-Cu surface alloy was studied as a function of Pd concentration up to the saturation of the c(2 × 2) structure and analyzed in a manner similar to the methods briefly discussed in Sect. 5.1.2.1. The results are illustrated by the following data. At 240 °C the product of the diffusion coefficient D and the concentration c of the diffusing species at 0.06 and 0.4 Pd monolayers decreases to 50 and 10 % of that on the Pd-free surface, respectively, while the activation energy for surface diffusion increases from 0.83 eV for the clean surface by about 20 % for a Pd coverage of 0.4 monolayers. The results for the lowest Pd coverage, 0.06 monolayers, could be modeled well by surface vacancy diffusion, impeded by the interaction with the Pd atoms buried in the second layer.

On the Cu(111) surface even more interesting phenomena occur. Starting with Sn, the formation of the Sn-Cu alloy, bronze, has been found in LEEM studies [257] to proceed in a rather unusual manner. Already at low coverages Sn forms clusters consisting of several hundred thousand atoms and identified by STM as two-dimensional crystalline Sn islands, which move more or less randomly across the surface, leaving an alloy trace behind. Their speed at 270 K is about 0.8 nm/s independent of particle size and increases by more than two orders of magnitude within 60 K. From the temperature dependence of the speed an activation energy of 0.9 eV was obtained. The driving force of the particle motion was attributed to repulsion between the Sn in the clusters and the Sn atoms in the alloy track left behind the clusters. This experiment is a nice demonstration of the application of the fast image acquisition (30 frames/s) which LEEM allows even at high magnification.

The most exciting phenomenon discovered on Cu surfaces upon alloying is the self-organization of submonolayer Pb films on the two-dimensional Pb-Cu alloy on the Cu(111) surface [258, 259]. The alloy contains 0.22 Pb monolayers (in units of the Cu(111) packing density) and has no long range order. On top of it pure Pb grows up to a coverage of 0.56 monolayers, which corresponds to a slightly compressed Pb(111) monolayer, before three-dimensional Pb crystals form. At partial coverages the theoretically predicted scenario for two-phase systems with competing short range and long range interactions in an elastically isotropic system was found as shown in Fig. 5.32 [258]: the transition from a Pb droplet phase (a) to a striped phase (b) to an anti-droplet (hole) phase (c) with increasing coverage. Here the Pb islands/stripes represent one phase, the uncovered alloy regions the other. The width of the stripes and their periodicity decreased strongly with increasing temperature as expected based on their decreasing boundary free energy and the islands elongated. The phases form surprisingly fast, keeping the size of the islands (50–200 nm diameter) in mind, which requires a significant driving force. By analyzing the magnitude of the fluctuations of the island distances the surface stress difference between the two phases was identified as source of this force and found to be 1.2 N/m independent of temperature [260]. The temperature dependence of the boundary free energy was determined by the fluctuation spectroscopy described in Sect. 5.1 of the boundaries of the striped phase, resulting in values decreasing from 22 meV/nm at 600 K to 10 meV/nm at 650 K [261]. Other LEEM studies, which have been made on this system, concerned the shape transition of Pb droplets and anti-droplets caused by the surface stress difference between them and the alloy phase [262], the effect of the Pb in the alloy layer on Cu surface diffusion as measured via the island decay rate [263] and the mechanism responsible for the fast movement of the large Pb droplets during self-assembly [264]. This last study came to the surprising conclusion that the mechanism not only involved fast transport of Pb atoms but in addition also diffusion of Cu atoms through the Pb islands, mediated by a high vacancy concentration in the islands.
Fig. 5.32

LEEM images of domain pattern evolution in two-dimensional Pb-Cu alloys on Cu(111) at 673 K. The Pb coverages are in (a) 0.33, in (b) 0.39, and in (c) 0.48 monolayers. Electron energy 18–20 eV. For explanation see text. Adapted from Ref. [258] with permission from Nature Publishing Group (UK)

Another usual phenomenon seen in submonolayer metal adsorption on a metal surface is the snake-like growth of two-dimensional Pd islands on Ru(0001). This growth mode has also been explained with the help of first principle calculations by slight alloying, with slow growth in alloy regions surrounding the sides of the snakes and fast growth along unalloyed surface along the head of the snake over which Pd diffuses easily. The details are complicated but the messages are clear: there are other mechanisms for pattern formation than surface stress phenomena and partial surface alloying can suppress step flow growth [265]. A final example in which LEEM was used for studying metal adsorption on a metal concerned the herringbone and triangular dislocation patterns of Ag, Au and AgAu monolayers, and double layers on Ru(0001). Ag monolayers developed two different herringbone structures with different coverage, Au and AgAu monolayers only one. In the double layer the dislocations of Ag and Au formed an interconnected network, that of AgAu was more complex [266].

5.2.2 Adsorption on Semiconductors

5.2.2.1 Adsorption of Metals

Au on Si(111) was first studied by Telieps and Bauer. They used the three (5 × 2) superstructure beams closest to the (00) beam in dark-field imaging to distinguish the three domains. With PEEM they used the work function difference between different surface phases ((√3 × √3)-R30°, (5 × 2)) to image their propagation during diffusion [4]. The first systematic video-LEEM study of the evolution of the three Au surface phases at various temperatures by Swiech et al. [267] showed a wide range of the morphologies, a strong influence of steps on nucleation, domain growth of the various phases and the phase transitions between them. Figure 5.33 [267] illustrates the growth of the (5 × 2) phase at various temperatures for the same coverage. At the lowest temperature many small (5 × 2) regions form except for isolated larger regions at steps. Diffusion of Au was studied later with the same coverage edge shift method used by Telieps in more detail in a SPELEEM, with PEEM complemented by XPEEM [268]. Starting from the edge of a region with monolayer coverage the diffusion coefficients derived from the edge displacement with time at 985 K were found to be significantly higher on the (5 × 2) surface than on the (7 × 7) surface. In another surface diffusion study [269] Au microspheres were deposited on the hydrogen-terminated surface. Upon hydrogen desorption the spheres converted into Au silicide particles surrounded by the (7 × 7) structure. The spreading of the (√3 × √3)-R30° structure from these particles over the surface was measured from 800 to 930 K where the higher coverage (6 × 6) structure is disordered. The spreading rate of the edge of the (√3 × √3)-R30° structure was found to be linear over the complete temperature range studied, independent of distance travelled, with spreading rates increasing from less than 0.1 to 50 nm/s in this temperature range. This indicates that it is not the diffusion rate of Au atoms on top of the (√3 × √3)-R30° structure but the rate of the structural rearrangement occurring during the propagation of the Au front at the border between the (√3 × √3)-R30° and the (7 × 7) regions, which determines the spreading rate [269].
Fig. 5.33

Frames from a LEEM video of the growth of Au on Si(111) at 620 K (a), 750 K (b), 880 K (c) and 1000 K (d). The Au coverage is approximately 0.15 monolayers in all images. LEED shows a (5 × 2) pattern. Nucleation and growth occurs preferentially at the steps but on the terraces (5 × 2) domains grow too elongated along the \( \left\langle 1\overline{1}0\right\rangle \) direction of the Si(111) surface. Adapted with permission from Ref. [267]. Copyright 1991 Elsevier

The adsorption of Ag on Si(111) was studied to a lesser extent. The first paper [270] clarified some apparent discrepancies of earlier LEED studies concerning the evolution of the (√3 × √3)-R30° and (3 × 1) structures. Later a spectroscopic PEEM study with a SPELEEM showed that the Ag 3d core level of the Ag in the (√3 × √3)-R30° structure was shifted 0.5 eV [271]. PEEM measurements of the spreading of Ag from a Ag island at temperatures at which both (√3 × √3)-R30° and (3 × 1) structures exist at their respective coverages, could be explained with diffusion coefficients from other studies and suggested that diffusion in the (√3 × √3)-R30° region is fast, slows down in the (3 × 1) region and speeds up again in the (7 × 7) region [272].

Cu adsorption on Si(111) was studied only in the context of three-dimensional silicide formation with LEEM [230, 273]. The two-dimensional (“5 × 5”) phase ((5.42 × 5.42) according to LEED) was found to nucleate at steps but to spread across both bordering terraces, to disorder above 850 K by dissolution in the bulk and to reappear upon cooling in the same manner. Later work [274, 275] looked at the formation of the adsorption layer in more detail, also at lower temperature at which nucleation occurred at the (7 × 7) domain boundaries and on the domains themselves. Hydrogen adsorption was found to suppress the formation of the (“5 × 5”) phase.

Adsorption of the transition metals Ni and Co does not produce superstructures visible in LEED but at high temperatures an “impurity-stabilized” (1 × 1) structure, which consists of a two-dimensional gas of ring clusters according to STM studies. This “(1 × 1)-RC” structure is in a temperature and cooling rate-dependent competition with the (7 × 7) structure, with the (7 × 7) regions growing and disappearing at the steps. Formation of the “(1 × 1)-RC” structure requires thermal activation so that at lower temperatures rather three-dimensional silicide crystals grow in the sea of the (7 × 7) structure [276]. A (√7 × √7)-R19.1° structure with a Co coverage of 1/7 monolayers was seen in LEEM via diffraction contrast and the phase diagram of the three surface phases (7 × 7), “(1 × 1)-RC” and (√7 × √7)-R19.1° could be determined [277, 278, 279, 280].

Aluminum forms initially a (√3 × √3)-R30° on Si(111) like Ag but at higher coverage a (9.4 × 9.4) structure. LEEM studies of a surface covered with many Si islands showed that the (9.4 × 9.4) structure started to grow from the island edges, consuming the islands and then to complete the layer with the Si between the islands. From the evolution of the surface morphology an Al coverage of 0.68 monolayers at the completion of this layer could be determined. The (9.4 × 9.4) periodicity agreed well with this coverage assuming misfit accommodation via a dislocation network with this periodicity [281]. A more detailed LEEM–LEED study [282] of this adsorption system revealed that it is more complex: there are two (√3 × √3)-R30° structures, a relatively unstable (√7 × √7) structure and the (9.4 × 9.4) structure with coverages of 0.25, 0.33, 0.43, and 0.68 monolayers in addition to a high temperature disordered phase. From studies at many temperatures and coverages the phase diagram was established and the interesting evolution and stability of the various phases deduced.

Pb adsorption on Si(111) is of particular interest because of the subsequent growth of three-dimensional Pb crystals with quantized thicknesses, which requires significant mass transport in the adsorption layer. This process was studied with LEEM using the LITD method for two phases, the low-temperature amorphous wetting layer and the α-(√3 × √3)-R30° phase with 4/3 monolayer coverage [283]. The measurements at 186 and 300 K showed identical unusual desorption profiles differing only in the time scales. As in the case of Au on Si(111) the profile moved at a constant rate. Detailed analysis suggested a convection-like mass transport, which was attributed to diffusion of thermally activated atoms on top of the wetting layer.

Adsorption of In on Si(111) had been studied extensively with STM and laterally averaging methods before the LEEM studies, which showed many but not all of the previously reported phases. Their evolution with increasing coverage, temperature, and other parameters is too complicated to be sketched here. Therefore the reader is referred to the original work [284, 285]. A spectroscopic PEEM study [286] showed a chemical shift of the In 3d peaks of 0.5 eV to higher binding in the (√3 × √3)-R30° and the (√31 × √31) structures, no shift in the In double layer and two peaks in the (4 × 1) structure, one unshifted, the other shifted as in the previous structures, indicating two binding sites as expected from the LEEM–LEED studies [284, 285]. Spectroscopic XPEEM combined with LEEM and LEED has also been used to look at the interaction of Sb and Ag with the (√3 × √3)-R30° and the (√31 × √31) structures. The data were interpreted in terms of displacement and intermixing processes [287, 288, 289]. Finally, Ga adsorption on Si(111) was subject of a LEEM study, which showed the successive replacement of the (7 × 7) structure by the (√3 × √3)-R30°-Ga structure with nice domain formation [290].

Metal adsorption on Si(100) surfaces has been studied much less extensively. The adsorption of Au has attracted most of the attention because the faceting it induces on slightly misoriented (100) surfaces at high temperatures. Several methods were combined to understand this process: bright-field and dark-field LEEM, LEED, STM for structural and spectroscopic PEEM for chemical characterization. Medium energy ion scattering served for absolute Au coverage determination [291, 292, 293, 294, 295]. 4° miscut surfaces consisting of 4 nm wide terraces separated by double steps, Au deposition rates of about 1.2 × 10−3 monolayers/s and temperatures between 1000 and 1150 K were used in all experiments. Driven by surface energy minimization, the Au lattice gas-covered surface breaks first up into (100) terraces covered with a (5 × 3.2) adsorption layer and step bunches which transform into (911) facets with increasing coverage. A few steps of the interesting dynamics, in which Au-decorated terraces spread with speeds up to 100 μm/s over the surface and reach aspect ratios of 10000:1, are shown in Fig. 5.34 [294, 295]. These experiments, which were performed in the first SPELEEM in Elettra, Trieste, are a good demonstration of the power of combining LEEM with spectroscopic PEEM.
Fig. 5.34

Faceting of a vicinal Si surface 4° off (100) with increasing Au deposition at 850 °C. (ac) LEEM images and sketches of the surface topography at three stages of the faceting process. The (100) surface is perpendicular to the electron beam so that (100) regions appear bright and tilted regions are dark. (a) (100) terrace nucleation during initial condensation stage, (b) step bunching connected with lateral terrace growth, (c) (119) facet formation from step bunches. (d) Evolution of the Au 4f μXPS signal from terraces and step bunches/facets and average 4f signal with deposition time illustrating the Au coverage difference between the different surface regions. (ac) Reproduced with permission from Ref. [294]. Copyright 2001 Elsevier. (d) Reproduced with permission from Ref. [295]. Copyright 2001 by the American Physical Society

Ag adsorption on Si(100) at high temperatures leads to a (2 × 3) superstructure whose composition and structure has been controversial. The solution of this controversy is another feat achieved by combining dark-field LEEM with LEED. With the same method mentioned already in connection with Al on Si(111), i.e., Si island deposition before Ag deposition, the amount of Si needed for completion of the (2 × 3) structure was determined. The amount of Ag was obtained from MEIS measurement, resulting in a 1:1 composition with three atoms each per unit mesh. From the analysis of the LEED patterns a detailed structural model was obtained [281, 296]. Other studies of Ag adsorption focused on surface diffusion on the (100) surface and some of its vicinals with PEEM [297, 298]. The spreading of the front of the (2 × 3) region originating from an Ag island was measured, similar to diffusion measurements mentioned earlier and activation energies were determined. On the 4° miscut surface an anisotropy of these energies parallel and perpendicular to the step direction of 0.7 eV was found. Ag diffusion on (100), 4° miscut (100), (911), and (111) surfaces is compared in Ref. [299].

Adsorption of In on Si(100) was studied with dark-field LEEM from room temperature up to 700 °C [300]. Below 600 °C and coverages up to 0.6 monolayers no dramatic changes in surface morphology were seen, only the transition from the (2 × 1) structure to the (4 × 3) structure with increasing coverage with varying degree of order. At higher temperatures, which required higher In fluxes because of desorption from the surface, major changes in the morphology occurred, similar to those seen with As adsorption, which will be discussed below: massive displacements of Si atoms, leading to island and hole formation and finally to progressive etching. Obviously In is much more aggressive than Au and Ag and so may be other non-noble metal adsorbates such as Sn.

A final interesting result concerning metal adsorption is the observation that Cs adsorption allows to image the (1 × 2) and (2 × 1) domains on Si(100) using linear polarized light. Between 0.3 and somewhat above 0.5 monolayer coverage the work function is low enough for photoemission excitation with 2.33 eV laser light, but not from states along the surface normal as the density functional theory calculations in this study show. However electrons from higher lying off-normal states are transmitted and allow imaging. The calculations identify the electrons as originating from π* anti-bonding Si dimer states, to which Cs has donated electrons without destroying the (2 × 1) periodicity, and explain the polarization dependence of the contrast [301].

Other surfaces on which metal atom adsorption has been studied in connection with its possible surfactant effect are the (311) and (211) surfaces. On the (311) surface, which is one of the equilibrium planes of Si, Ag adsorption causes at saturation of the monolayer faceting with a periodicity of about 40 nm into (111) facets covered with the Ag(√3 × √3)R30° superstructure and (511) facets covered with a (n × 2)-like superstructure [302]. Saturation of the (311) surface with Ga also causes faceting with about the same periodicity into (211) facets with (1 × 6) reconstruction and (511) facets with (1 × 4) reconstruction [303, 304]. At lower coverages intermediate phases form, preferentially at the steps. In all cases temperatures around 500 °C were used. On the non-equilibrium (211) surface, which in the clean state consists of (111) facets with (7 × 7) structure and (5,5,12) facets with (1 × 2) structure, the adsorption of Ga and In was studied [305, 306]. In both cases adsorption initially reduces the faceting and then induces new facets with new superstructures. Both the facets on the clean and on the Ga and In covered surfaces are, however, too narrow to be resolvable with the LEEM instruments used so that all information had to be obtained from LEED.

An interesting adsorption system on a compound semiconductor, Ga on GaN(0001), was studied by Pavlovska and Bauer [307]. Ga forms a double layer, which is stable up to about 680 °C and shows three phase transitions, a first order transition at 225 °C and two continuous transitions at higher temperatures. The LEED patterns connected with the first order transition shown in the LEEM images in Fig. 5.35 [307] indicate a material exchange between first and second layer, in which the first layer expels atoms into the second layer and becomes pseudomorphic upon heating; upon cooling it converts to the same misfitting packing density as the second layer. This transition is similar to the misfit phase transition in the Cu double layer mentioned earlier [224]. The importance of the Ga double layer for the growth of smooth GaN layers will be discussed below.
Fig. 5.35

Frames from a LEEM movie of the “(1 + 1/12)” to “(1 + 1/6)” phase transition of the Ga double layer on GaN(0001). (ad) Heating and (eh) cooling through the transition. Dark: “(1 + 1/12)” structure, bright “(1 + 1/6)” structure. The step structure is not influenced by the phase transition. Electron energy 11 eV. Adapted with permission from Ref. [307]. Copyright 2001 Elsevier

5.2.2.2 Nonmetallic Adsorbates

In view of the interesting surface structure of Si(111) and the practical importance of the Si(100) surface remarkably little work on the interaction of nonmetallic elements with these surface has been done, oxygen excepted. Arsenic has been studied on both surfaces. On the (100) surface As was found to displace Si atoms on the terraces at high temperatures (900 K) resulting in Si monolayer island formation. The driving force was identified to be due to a change of the surface stress anisotropy from the clean to the As-covered surface [107]. On the Si(111) surface MEM studies of As adsorption at low temperatures (<750 °C) also show displacive adsorption on the terraces but with increasing temperature and thus decreasing As coverage, when steps become the preferential adsorption sites, Si is increasingly displaced from the steps. This was attributed to surface stress relaxation on terrace regions near the steps [308].

Two more complicated adsorbates have been studied too: CCl4 [309] and benzoic acid (BA) (C6H5COOH) [310]. In the first study UVPEEM was used to compare the reactive adsorption on the Si(111)-(7 × 7), the Ag(111) and the Si(√3 × √3)-R30°-Ag surface. The coverage was determined separately with XPS under identical conditions. The reactivity decreased in the same sequence, with dissociate adsorption on the first two surfaces but molecular adsorption on the last, which was interpreted in terms of the different electronic surface structure. The second study addressed the question of the influence of OH adsorption on the bonding of benzoic acid to the Si(100) surface with XPEEM and XPS. A patterned surface with OH patches and carbon-contaminated SiO x regions was formed. In order to distinguish between adsorption on OH and C-contaminated regions 4-nitrobenzoic acid (4-NBA) was used and the N K absorption edge used for imaging, which clearly showed adsorption only in the OH regions.

Concluding the section adsorption on semiconductors some studies on Ge should be mentioned briefly. Apparently, no nonmetallic adsorbates have been studied and only a few metals on Ge(111). Adsorption of Ag has been the subject of a low-resolution LEEM–LEED study, which showed mainly the evolution of the various adsorption phases with coverage and temperature [311]. In a Pb adsorption layer a novel phase transformation was found at 189 °C at a coverage of 1.30 at which the high coverage (√3 × √3)R30° phase and the (1 × 1) phase coexist due to thermal fluctuations of rather large domains between the two phases without nucleation (“spontaneous domain switching”) [312].

5.3 Film Growth and Structure

In this section we will discuss not only the growth and properties of continuous films but also of “nanostructures” such as nanowires, quantum dots, and droplets on the clean and wetting layer-covered surfaces described in the previous sections.

Included are not only studies of purely scientific interest but also work motived by technological problems.

5.3.1 Films on Semiconductors

5.3.1.1 Metal Films

In the growth of Au on Si(111) three-dimensional particles form before the two-dimensional (6 × 6) structure. Below the eutectic temperature (632 K) they have hexagonal cross section, above this temperature they are round (liquid silicide droplets), which are mobile and suppress (6 × 6) formation in the tracks they leave behind during their migration across the surface. These droplets show an interesting reversible wetting transition to larger triangular-shaped particles, which have {112} interfaces with the substrate (Fig. 5.36a, b) [267]. With increasing temperature the speed of the droplets increases strongly so that at 1150 K, where Au sublimates, their motion can be followed with the help of the tracks they leave behind (Fig. 5.36c) [267]. The droplets move perpendicular to the surface steps uphill leaving a step-free track behind. A later study on a better equilibrated surface with lower step density showed that the droplets formed right after completion of the (6 × 6) structure at the steps [313]. Droplet formation is also briefly discussed in [314].
Fig. 5.36

Frames from a LEEM video of the temperature dependence of the shape of Au-Si eutectic particles on Si(111). (a) At the eutectic temperature (650 K), (b) at 900 K, (c) at 1150 K after most Au had evaporated during the particle movement across the surface. The transition between (a) and (b) is completely reversible. Adapted with permission from Ref. [267]. Copyright 1991 Elsevier

The growth of Ag on Si(111) and Si(100) has been studied in much more detail, starting with studies of crystal shape oscillations [315] and shape transition induced by stress relaxation, leading to nanowire formation [316]. Tromp et al. imaged the interface between Ag crystals and the Si(111) substrate making use of the strain field and the long inelastic mean free path of slow electrons in Ag [317]. With the growth of Ag on Si(111) covered with a (√3 × √3)R30°-Sb layer his group also gave the first demonstration of the surfactant effect in thin film growth. This effect reduces the tendency to three-dimensional growth, which is very pronounced in the growth of Ag films [318]. Altman’s group determined the three-dimensional shape of Ag crystals via the facet spot movements on Si(111) [319] and studied the influence of In co-deposition on their growth shape, which they attributed to the modification of the Ehrlich-Schwoebel barrier [320]. A PEEM study of the shape and orientation of Ag crystals from 200 to 640 °C showed that up to 240 °C the crystals had (100) and (111) orientations, from 240 to about 600 °C all crystals had (111) orientation with polygonic shape, above 600 °C with triangular shape [321]. In a different context, Ag films with different thickness and agglomeration were studied in an effort to understand the emission process as a function of photon energy and power [322].

On the Si(100) surface Ag crystals were observed to grow at about 250 and 300 °C in a bamboo-like manner along the 〈011〉 directions, with non-uniform width and height. This excludes a stress-induced mechanism as cause for the elongated growth. Instead a high density of nuclei along certain 〈011〉 directions followed by connection between the crystals during growth was proposed [323]. In PEEM studies at higher temperatures, around 600 °C, wires with uniform cross section have been reported, together with compact crystals. The ratio of wires to compact crystals increased with increasing deposition time and strongly with increasing temperature and miscut from 0° to 4° [324]. Specific studies have been made for a 4° miscut surface [325] and an Au-pre-facetted 4° miscut surface [326], in the last one also with LEEM. The wire growth has been attributed mainly to the anisotropic diffusion on the stepped surfaces.

The growth of Cu on Si(111) was already studied in the early years of LEEM but little has been published about it [225, 230, 273]. Irrespective whether the surface has initially the (7 × 7) structure or the hydrogen-terminated δ-(7 × 7) structure, the crystals grown on this surface were Cu silicides in various shapes. This is illustrated in the PEEM image of Fig. 5.37a [273] from a film grown at about 850 K. Figure 5.37b shows one of the triangular islands seen in PEEM, illustrating how much more information is contained in a LEEM image. Heating slightly above this temperature causes melting of the small polyhedral crystals, cooling below 850 K crystallization again. At high temperature the smaller particles migrate over the surface in the step-up direction leaving pronounced reaction tracks behind (Fig. 5.37c), similar to the Au silicide particles shown earlier. In the later studies of Cu growth on the hydrogen-terminated surface qualitatively the same distribution of particle shapes and sizes was found. The surface planes of the most frequent shape were found from the facet spot movements to be high index planes as expected for growth shapes [274, 275, 327].
Fig. 5.37

Growth of Cu silicide on Si(111). (a) PEEM image showing rods and triangles which are flat (no shadows) and smaller three-dimensional particles (with shadows). (b) LEEM image of a triangular particle and (c) three-dimensional particle after reactive migration at high temperature. Adapted from Ref. [273] with permission from the American Institute of Physics, © 1989

Even less has been published on the growth of Co on Si(111) than for Cu. Some results can be found in [71, 225, 328]. Like Cu, Co forms a silicide, CoSi2 that is very stable and has a higher surface energy than Si. Unlike Cu, Co does not form a wetting layer as soon as enough Si has been deposited for nucleation of three-dimensional crystals. These crystals are surrounded by regions with (7 × 7) and (1 × 1) structure (Fig. 5.38a) [328]. Depending upon deposition and annealing temperature a wide variety of growth shapes were observed. When heated to temperatures at which Si sublimates, large CoSi2-topped hillocks form on the Si surface with (1 × 1) and (7 × 7) because of the lower vapor pressure of CoSi2 (Fig. 5.38b) [67]. A later study found that the two-dimensional structures at low Co coverages [277, 278, 280] indeed disappear once silicide islands have nucleated, even at very low coverage, apparently having been absorbed into the islands. The influence of SiO2 layer on Si(100) on the formation of CoSi2 [329] and FeSi2 [330] has also been studied with LEEM and XPS in a SPELEEM, respectively.
Fig. 5.38

Co on /Si(111) (a) CoSi2 crystals (small dark triangles) on a Si(111) surface surrounded by regions with (7 × 7) (bright triangles) and (1 × 1) (dark areas in between) structure. Energy 10 eV. (b) Hillock formation on Si(111) due to Si sublimation suppression by CoSi2 islands. Energy 10 eV. (a) Reproduced from Ref. [328] with permission from Springer Science + Business Media. (b) Reproduced from Ref. [67] with permission from the American Institute of Physics, © 1991

The growth of Pb on Si(111) was subject of a study [331] aimed at the understanding and control of the three basic growth modes of thin films (Volmer-Weber, Stranski-Krastanov and Frank-van der Merwe described in Refs. [332, 333, 334]). On the Pb wetting layer on the Si(7 × 7) surface three-dimensional crystals form on the terraces and preferentially at the steps. Despite significant lateral growth and coalescence the film is still not continuous at 6 monolayers because of the simultaneous strong growth normal to the surface (Fig. 5.39a) [331]. Normal growth was reduced strongly by growing not directly on the (7 × 7) structure but on an Au wetting layer, which acts as an interfactant, encouraging lateral growth. This produced a continuous Pb film already at one monolayer, which continued to grow quasi-monolayer-by-monolayer via step flow and two-dimensional island formation up to the largest thickness of 5 ML studied (Fig. 5.39b [331]). Thus the interfactant converted the growth mode from Stranski-Krastanov to Frank-van der Merwe. A similar but less effective action could be achieved by growing on a Ag(√3 × √3)R30° wetting layer. The effect of the interfactant was attributed to misfit stress relief, increase of the nucleation rate and decrease of surface diffusion [331, 335].
Fig. 5.39

LEEM images of the growth of Pb on several Si surfaces at room temperature. (a and b) 5 monolayers on the S(111)-(7 × 7) and on the Si(111)-(6 × 6)Au surface, respectively, grown at room temperature. Electron energy 8 eV. (c) Pb islands after deposition of 12 Pb monolayers on the Si(100)-c(4 × 4)Pb surface, grown at room temperature. Electron energy 5 eV. (d) 6.6 monolayers on the Si(755) surface with 0.2 pre-adsorbed Au monolayers. 276 K. Electron energy 7.1 eV. (a, b) Adapted with permission from Ref. [331]. Copyright 2000 by the American Physical Society. (c) Reproduced with permission from Ref. [336]. Copyright 1994 by the American Physical Society. (d) Reproduced with permission from Ref. [338]. Copyright 2001 Elsevier

On the Si(100) surface Pb grows also in the Stranski-Krastanov mode, with the first crystals forming at room temperature already before completion of the wetting layer. At a deposition rate of 1 monolayer/min the density of the particles is about 0.2/μm2 so that they have a large supply region from their surroundings. This and the much faster diffusion on the wetting layer than on top of the Pb crystals leads to the formation of rims on the crystal as shown in Fig. 5.39c [336]. The resulting “ashtrays” fill up in the later stages of the growth when the crystals become so large that the direct supply of atoms from the vapor phase dominates that from the wetting layer [336]. On 4° miscut (100), (110), and high index surface such as (533) or (755) Pb also grows in the Stranski-Krastanov mode but with a completely different growth shape, mesoscopic wires [337, 338]. Length to width ratios of up to 130 with temperature-dependent widths were obtained but in general there is a wide distribution of lengths as illustrated in Fig. 5.39d [338] for a film on a (755) surface after deposition of nearly 7 monolayers at 276 K. The wires are not flat but have (111) and (100) side faces and tips. Anisotropic stress and one-dimensional diffusion were invoked as driving force for the formation of the wires.

The growth of In on Si(111) is the most complex one of all metals, which have been studied on this surface. The multitude of thickness- and temperature-dependent two-dimensional phases has already been discussed. Three-dimensional crystals grow on the double layer predominantly with (100) orientation, which can be explained by the very small misfit with the centered square unit mesh of the (√7 × √3) structure of the double layer. The (100) surface of the crystals is reconstructed. Deconstruction and melting occur at the same temperature as in the bulk [284, 285].

Several metals including Co and Ni are known to form silicide nanowires on Si surfaces in a certain temperature range, at least on those which a 〈110〉 direction in the surface [339]. A LEEM study of the growth and stability of TiSi2 nanowires [340] showed that they form on Si(111) at about 850 °C and become unstable above 900 °C. In FEL-PEEM studies [341] their growth on the same surface was studied as a function of deposition time at 1150 °C. The wires were found to develop from isometric particles at a critical size, after which the width remained constant while the length increased linearly with deposition time as expected for stress-induced wire formation [316]. Short depositions produced only isometric crystals [342]. On Si(100) ErSi2 nanowires developed when a 10–20 Å thick Er film deposited at 750 °C was annealed at 800 °C, with linearly increasing length and constant width as a function of annealing time, suggesting the same stress effect as in TiSi2. Annealing at 1050 °C caused only Ostwald ripening of the smaller particles [343]. Thus the formation of silicide nanowires is limited to a narrow temperature range.

When the temperature during deposition of silicide-forming metals is above or close to the bulk eutectic temperature then the silicide particles migrate across the surface. This has been seen in LEEM first with Au on Si(111) [267] and with Cu on Si(111) [273] but not studied in detail. However, it was noticed that the particles moved uphill perpendicular to the steps and the driving force was assumed to be dissolution of Si on the uphill side of the particle and expulsion on its downhill side. Detailed studies of the formation and migration of liquid silicide droplets were later performed with FEL-PEEM [344] and LEEM [345, 346], using Pt on Si(100). In these experiments a temperature gradient was assumed as driving force, with Si dissolution on the high temperature side and expulsion on the low temperature side of the particle. In the FEL-PEEM studies the migration rate increased independent of particle size (2–7 μm diameter) from 0.8 μm/s at 1085 ° C to 3.6 μm/s at 1210 °C, yielding an activation energy of 0.59 eV for migration. The later LEEM studies found that only larger particles moved normal to the step direction. Smaller particles are guided by the dragging force of the steps along the step direction. In all cases migration occurs from lower towards higher temperature regions on the substrate.

5.3.1.2 Ge on Si

The importance of Ge-Si nanostructures in electronics and optoelectronics [347] has made the growth and structure of Ge on Si and of Si on GeSi [348] an interesting subject for LEEM and XPEEM. The technologically important (100) surface has been studied by Tromp’s group [349, 350, 351, 352, 353] and Lagally’s group [354, 355, 356, 357, 358, 359], in part jointly [360]. The first study [349] compared Ge growth on the clean and As-terminated surface at 630 °C and found that on the clean surface three-dimensional islands developed after three monolayers while As induced at every level a high nucleation rate leading to quasi-monolayer-by-monolayer growth (surfactant-mediated growth). The high nucleation rate was attributed to exchange between As and Ge dimers during growth. Intermixing of Ge and Si occurs already during the formation of the first monolayer via Ge partially displacing Si on the terraces and both Ge and displaced Si diffusing to the steps [361]. A later study [353] provided detailed insight into the further intermixing. At higher temperatures (>750 °C) the GeSi film roughens already at 2 monolayers or less without involving the substrate with a thickness-dependent roughening temperature in between that of Si(100) and Ge(100) [360].

The study of the growth of technologically important Ge x Si1−x layers brought some interesting results. Growth by exposing the Si(100) surface at 650–700 °C to Si2H6 and Ge2H6 at growth rates of several monolayers/min leads after a few monolayers to roughness, which develops into mounds with long range order with a Ge concentration-dependent period. While the period is independent of thickness, the slopes of the mounds increase reaching a final angle of 11° and azimuthal orientation, resulting in pyramids bounded by {105} faces [351, 352, 357, 358]. With further increasing thickness the pyramids coarsen, decrease in density, and transform into circular domes bounded predominantly by {311} and {15,3,23} surfaces [350, 352]. Figure 5.40 [350] shows this final stage in the LEEM image (a) and from a similar film in the SEM image (b) for comparison. All these processes, the roughening, mound formation, and transition to the domes are strain-induced and occur without nucleation. When these layers are imbedded in Si by Si2H6 exposure at 650 °C the islands underwent considerably shape changes ending up with large (100) top faces [355, 356].
Fig. 5.40

(a) LEEM image, taken with 4 eV electrons, and (b) SEM image of Ge x Si1−x islands grown on Si(100) by exposure to a silane–germane mixture at 650–700 °C, showing the transition from pyramidal shapes (P) to domes shapes (D). From Ref. [350] with permission by the American Association for the Advancement of Science

The initial growth of Ge on Si(111) was studied already in 1993 on the Sb(√3 × √3)R30° surface during simultaneous Ge and Sb exposure and for comparison on the clean (7 × 7) surface, in both cases at 775 K [318]. On the (7 × 7) surface three-dimensional crystals formed after the completion of four two-dimensional layers, the first two with (7 × 7), the second two with (5 × 5) structure. On the Sb-covered surface growth proceeded by periodic two-dimensional island formation up to the highest coverage studied, demonstrating the surfactant action of Sb. No information on alloying of Ge with Si could be given at this early state of the art. Such information was obtained nearly 20 years later by combining LEEM with XPS in a SPELEEM instrument. Here only the first monolayer was studied at lower temperatures than in the original work. A detailed analysis of the Ge 3d/Si 2p intensity ratio showed that at 200 °C very little if any Ge-Si exchange occurred, while at 270 °C the data indicated considerable intermixing which decreased with coverage. Combining this with the evolution of the island sizes led to the conclusion that steps play the crucial role in intermixing [362], confirming earlier interpretations based on LEEM observations only. Another XPS-XPEEM [363] study looked at a phenomenon at the other extreme of Ge coverage: the dependence of the shape of islands upon their size up to lateral dimensions of tens of microns and thickness up to 500 nm. With increasing size a transition from compact to dendritic-looking islands occurred, which was attributed to strain relief at the island perimeters. Ge 2p and Si 1s XPEEM images clearly indicated a composition gradient from island center to edge but no quantitative information could be extracted [363]. Some information on the composition was obtained in a series of SPELEEM studies of crystals of various shapes grown at lower temperatures (460–560 °C) [364, 365, 366]. A LEEM study that looked at the distance distribution between three-dimensional Ge crystals found a substantial tendency to self-ordering without a significant influence of the size of the crystals on their distance [366].

Finally some studies of the influence of adsorbed metal layers on the growth of Ge on Si(111) should be mentioned. Growth on a surface incompletely covered with the Ga(√3 × √3)R30° structure produced preferred nucleation of Ge on the Ga-covered regions [367]. The situation is different when the surface is covered by a Ag(√3 × √3)R30° structure. On this surface large 3 monolayer thick islands grow after some intermediate stages, spread across the surface and coalesce, forming large flat regions, in contrast to the compact pyramidal crystals on the clean surface. Thus, Ag acts like a surfactant, though in a manner completely different from those mentioned earlier. In the case of As and Sb the surfactant reduces surface diffusion, which produced a large island density, while Ag enhances the surface diffusion, which drastically reduces the island density [368]. For a review of the processes involved in Ge growth on Si(111) see Ref. [369].

The growth of Ge on clean and metal adsorbate-covered (11h) surfaces (h = 2, 3) has also been studied with LEEM and LEED [370, 371, 372]. The (113) surface is the third most stable clean Si surface and is (3 × 2) or (3 × 1) reconstructed, depending upon temperature while the (112) surface breaks up into (111) and (337) facets. On the clean (113) surface Ge was found to form initially an about 4.4 monolayer thick wetting layer followed by the formation of three-dimensional clusters. These develop with increasing temperature and thickness into larger crystals, which are strongly elongated along the \( \left[33\overline{2}\right] \) direction and bounded by many facets (Fig. 5.41) [370]. On the clean (112) surface growth also proceeds via a wetting layer and three-dimensional islands. Their number density decreases and size increases with increasing thickness, from which—like in the growth on the (113) surface—activation energies for surface processes were derived [305].
Fig. 5.41

(a) LEEM image after deposition of 36 Ge monolayers on Si(113) at about 560 °C, showing 3D islands. Electron energy 1.7 eV. (b) Reciprocal lattice section in an azimuth in which one of the facet spots move with energy. The dotted circles are the Ewald spheres for the indicated energies. φ is the tilt angle of the facet. Adapted from Ref. [370] with permission from the American Institute of Physics, © 2002

Pre-adsorption of metals has a strong influence on the growth, at least when the adsorption layer is saturated. On the Ga-saturated (113) surface, which consists of (112) and (115) facets, Ge grows with a much higher nucleation density than on the clean, non-faceted surface with a particle distribution, which is closely correlated with the facet structure and periodicity [367, 372]. On the non-faceted Ga-saturated (112) surface with (6 × 1) reconstruction Ge grows in rod-shaped crystals with the long axis parallel to the \( \left[1\overline{1}0\right] \) direction instead of the isometric crystals seen on the clean faceted surface [305]. Pre-adsorption of In has a similar influence on the growth of Ge on the Si(112) surface, which after In adsorption is also flat and is (3.5 × 1) reconstructed. After completion of the wetting layer a high density of crystals grow, initially extended along the \( \left[1\overline{1}0\right] \) direction but transforming with increasing thickness and temperature into triangles with their apex in the \( \left[11\overline{1}\right] \) direction. They are bounded by (112) top faces and (111) and {130} side faces [306]. Finally, on the Ag-pre-adsorbed, faceted (113) surface growth proceeds in a similar manner with Ge crystals elongated along the facet direction \( \left[1\overline{1}0\right] \) [302].

Summarizing this section, Ge on Si, having been studied originally intensely because of its importance in semiconductor electronics, has developed in an interesting playground for the study of the influence of surface orientation, reconstruction, faceting, and adsorbates on epitaxy. Many detailed results of the structure and kinetics have been obtained which can be found in the original literature. For comparison with the growth on Si a brief LEEM study of the growth on GaAs(100) should be mentioned [373]. On this surface Ge grows at 420 °C layer-by-layer from small two-dimensional islands, which become anisotropic with increasing size and increasing temperature (450–480 °C). At higher temperatures growth proceeds via step flow until at 540 °C growth becomes unstable due to As desorption and surface roughening.

5.3.1.3 Other Films on Semiconductors

The growth of CaF2 on Si(111) was studied by Tromp et al. from the sub-monolayer range [374] to thick films [375] in two temperature ranges. Below about 700 °C they found an interesting de-wetting transition: initially monolayer CaF2 islands formed but once second layer islands nucleated on top of the first layer, the first layer contracted to form bilayer islands. This was interpreted as follows: below this temperature CaF2 does not dissociate, diffuses over large distances on the (7 × 7) structure and CaF2 bonding to the substrate is weaker than the bonding between the first and second layer. Above 700 °C CaF2 dissociates and forms a double layer with Ca bonded to Si and F on top of Ca, with one F atom being desorbed, possibly as SiF x molecule during the restructuring from the (7 × 7) to the (1 × 1) structure in the adsorption process. On top of this wetting layer initially isolated two-dimensional CaF2 islands grow without crossing the steps but the third layer crosses the steps and leads to step flow growth up to a critical thickness of about 30 Å, above which the misfit strain is released by formation of a dislocation network. Using different energies and focus the authors were able to correlate the interfacial dislocations with the surface structure. In combination with transmission microscopy they could determine the evolution of the dislocation structure in considerable detail. It should be noted that electron energies below 10 eV were used in these studies in order to avoid electron stimulated desorption of F, which starts at about 25 eV due to ionization of the Ca 2p level followed by interatomic Auger transitions.

Another interesting film growth process is that of silicon nitride on Si(111). It had been studied extensively with lateral averaging methods with contradicting results before laterally resolving studies gave some insight in the growth process. In the first such study the nitride layer was grown by reaction with ammonia in the low 10−7 mbar range [376], in the second one [377] nitrogen from a radio frequency (RF) gas discharge was used for nitridation in the low 10−5 mbar range. In both cases LEEM and LEED were used, in the first case in addition also STM. Saturation of the surface with NH3 followed by annealing to 1210 K led to triangular islands with (8 × 8) structure which decorated the steps and domain walls of the original (7 × 7) structure, with a total nitride coverage of 0.25 monolayers. During reaction at 1180 K, i.e., above the (7 × 7) ↔ (1 × 1) transition, a high density of concave triangular (8 × 8) islands forms, on which second layer islands form before the first layer covers the surface completely. At 1260 K the nucleation and growth of the (8 × 8) nitride islands is indistinguishable from that of the (7 × 7) Si islands during the (1 × 1) ↔ (7 × 7) transition. On slightly miscut surfaces the (8 × 8) islands are elongated along the steps and with increasing coverage pronounced step bunching occurs as seen in Fig. 5.42 [376]. This is much more pronounced on surfaces with larger miscut as seen in STM.
Fig. 5.42

Frames from a LEEM video taken during the growth of the Si-N (8 × 8) structure on a 0.5° miscut Si(111) surface exposed to 1 × 10−7 mbar ammonia at 1270 K, showing widening of terraces and step bunching as growth proceeds. Exposure times (ah) from 10 to 480 s. Electron energy 3.6 eV. Adapted from Ref. [376] with permission. Copyright 1995 by the American Physical Society

Once the (8 × 8) wetting layer is completed no three-dimensional Si nitride crystals form above 1200 K at the NH3 pressures used. They do nucleate and grow occasionally, however, before completion of the wetting layer at surface defects. On a heavily contaminated surface, which evolves after many growth and desorption cycles, they grow with a high density, in particular in the electron-irradiated region. This is attributed to elution of adsorbed gases from the walls by the NH3 and dissociation of these gases on the Si surface, enhanced by electron irradiation. At the higher pressures during reaction with nitrogen from the gas discharge this phenomenon is even more pronounced. The LEED patterns of the three-dimensional crystals are compatible with the (111)-oriented high-pressure phase of Si3N4 (γ-Si3N4). If contracted by 1 % it would fit perfectly in a 2:1 ratio in parallel orientation to the substrate but it is expanded 4.7 % and rotated several degrees relative to the substrate. In thicker films these crystals are also observed in exact parallel orientation with a (√3 × √3) superstructure. There are still many details of the high temperature growth of Si nitride, which are not understood yet despite of its technological importance.

Another practically important film on semiconductor surfaces is GaN on SiC(0001) or on thick GaN layers grown on various substrates. Similar to the Si nitride growth studies GaN growth is another example of in situ studies at high pressures of reactive gases and their limitations due to gas-induced contamination. The homoepitaxial growth of GaN has been studied with LEEM and LEED on GaN(0001) [153, 155, 378] and on SiC(0001) [155, 379, 380, 381] using more complex sources. Ga was evaporated from a BN crucible and nitrogen was supplied either from a RF plasma source as in the growth of Si nitride or from a supersonic He jet source seeded with up to 10 % NH3. The motivation for the second choice was the belief that hyperthermal NH3 would be more reactive and allow growth at lower temperature. Operation pressures were around 1 × 10−5 mbar with the RF discharge source and in the middle 10−8 mbar range with the jet source. During homoepitaxial growth on a GaN(0001) layer with (√3 × √3) superstructure at around 670 °C with a Ga/NH3 ratio above 2, a two-dimensional layer formed from a high density of two-dimensional islands, completely restoring the original surface configuration. At lower Ga excess the surface was etched and no growth occurred [153, 378]. A more detailed study showed that the Ga double layer mentioned earlier, which is liquid around 670 °C [307] is necessary for continuous growth and has to be maintained during growth by proper adjusting the Ga/N ratio. This type of growth may be considered as quasi-liquid phase epitaxy. If the Ga/N ratio is too high Ga droplets form, if it is too low the surface becomes rough [155, 307].

Similar considerations apply also to the growth of GaN on SiC(0001) [155, 379, 380, 381]. The most extensive information is contained in the last study [155], which combined LEEM and LEED with cross-sectional high resolution TEM, and in the preceding one [381], which also corrects misinterpretations in [380]. The essential results are as follows. GaN grows on SiC(0001) both with activated nitrogen from the RF plasma discharge and the NH3-seeded He jet source three-dimensionally, independent of Ga/N ratio. Ga excess produces flat epitaxial crystals with (0001) top faces and \( \left\{10\overline{1}1\right\} \) and \( \left\{10\overline{1}0\right\} \) side faces while with N excess the \( \left\{10\overline{1}1\right\} \) faces are dominating and the (0001) face is absent, leading to a rough surface. The polar (0001) surface is not an equilibrium plane and appears only because it is stabilized by Ga adsorption. Substrate defects and N excess can cause growth of the cubic phase of GaN and stacking faults. Activated N is better suited for GaN film growth than NH3, at least below 700 °C.

PEEM with the spontaneous radiation from a FEL has also been used in the study of GaN films, in this case, however, not for understanding of the growth but of inversion domain boundaries between Ga and N terminated surfaces [382]. The boundary was obtained by growing part of the film directly on sapphire, which gave N-face termination, the other part on an AlN nucleation layer, which resulted in Ga-face termination. Photoemission from the N-face started at 4.9 eV with emission from the conduction band, from the Ga-face at 6.3 eV from the valence band. The about 2 μm wide inversion domain boundary region showed strong emission above 4.9 eV from the space charge region between the two terminations, predominantly due to emission of electrons in an accumulation layer in the conduction band. Interestingly, the AFM and PEEM images suggest that the Ga-terminated surface is rougher than the N-terminated surface, in contrast to the LEEM, LEED, and TEM studies mentioned above.

5.3.1.4 Nanostructures and Droplets on Semiconductors

Nanostructures, when truly deserving of “nano” are at the limits of microscopes without aberration correction. Therefore mostly structures ranging from 50 nm to several 100 nm have been studied. A few structures, in general prepared ex situ, will be discussed here: dots, rings, and wires. An example for dots are the so-called InAs quantum dots on GaAs(100). They were grown either randomly epitaxial in the Stranski-Krastanov growth mode on a Se-terminated surface at 200 °C [383], on a Si doped surface at 540 °C [384, 385] or on a prepatterned GaAs surface [386]. After growth they were capped with As for transfer to the SPELEEM, decapped in it and studied with LEEM, XPEEM, μXPS, and LEED with the goal to determine the degree of intermixing between InAs and GaAs. This is a difficult experiment for two reasons: (1) the dots are surrounded by the wetting layer which contains both Ga and In and (2) the spherical aberration of the objective lens deflects electrons from the surrounding of the dots into the dots and vice versa. This makes quantitative chemical analysis on the nanoscale difficult. Nevertheless useful information can be extracted with this limitation in mind. The situation becomes more complicated in the analysis of quantum rings, which evolve when InAs quantum dots are overgrown with a thin GaAs layer and subsequently annealed properly [387, 388]. The transition from dot to ring was studied by XPEEM, LEEM, AFM, and low-temperature photoluminescence spectroscopy as a function of annealing time. The combination of AFM and photoluminescence spectroscopy showed a clear correlation with the shape of the dots and the rings which could be explained by the distribution of In and Ga as deduced from XPEEM [389]. The field of quantum dots and rings is reviewed in Ref. [390].

A quite different system, which demonstrates very nicely the power of XPEEM with core level photoelectrons is the formation of Ge nanoparticles on an array of Au discs on a hydrogen-terminated Si(100) surface after it had been exposed to air (Fig. 5.43) [391]. The air exposure caused Au-catalyzed local oxidation. Annealing this surface in UHV at 600 °C agglomerates the Au in the oxidized regions and desorbs the hydrogen between them, producing clean regions. When Ge is deposited on this surface at 450 °C it forms three-dimensional crystals only in the clean regions and Au diffuses from the oxidized regions onto the Ge, where it produces a much stronger Au signal than the large Au particles in the oxidized regions. Thus by oxidizing the surface Au not only produces the regions in which no Ge condenses but also has a major influence on the formation of the Ge crystals to which it migrates during their growth.
Fig. 5.43

Core level XPEEM images of Au-SiO2-induced Ge nanopyramid arrays. The regular pattern of Au islands, one of which is indicated by the circle in (a), was deposited on hydrogen-terminated Si(100) and locally oxidized during transfer through air (b). Ge grew subsequently on this surface in UHV at 400–600 °C only in the clean areas between the SiO2 regions while most of the Au diffused into the Ge regions, catalyzing the growth of Ge into pyramids (c). Photon energy 195 eV. Adapted with permission from Ref. [391]. Copyright 2007 American Chemical Society

The third group of semiconductor nanostructures, nanowires, cannot be studied directly with LEEM or XPEEM when grown perpendicular to the substrate. These nanostructures have to be removed from their substrate and remounted flat on another substrate for microscopic study. An example is the secondary electron PEEM, core photoelectron PEEM, and MEM study of InP nanowires, whose ends had been heavily n-doped. The combination of these methods gave detailed information on doping, surface chemistry, and morphology of wires [392]. Another example is the study of InAs nanowires, which had been grown through a SiO x mask on InAs(111) and transferred onto a Si(111) substrate for imaging with MEM. A careful analysis of through focus series images allowed a precise determination of the diameter and shape of the wires [393]. LEEM and XPEEM have also contributed directly to the understanding of the formation of these nanowires by studying the formation of the holes in the SiO x mask, from which they grow, as a function of temperature. By comparing these results with the number and distribution of long nanowires grown in a separate system, the growth was established to occur via a vapor–liquid–solid mechanism starting from liquid In droplets formed in the holes. The oxide layer immobilizes the droplets on the surface and restricts their size [394].

Studies of other nanostructures, oxide patches on GaAs [395, 396] and GaAs/AlAs/GaAs [397, 398, 399] heterostructures on GaAs(100), ranging from several 100 nm to a few micrometer, showed one of the limitations of XPEEM and μXPS, photon stimulated desorption. The nanostructures were produced by anodic oxidation in humid atmosphere by AFM with a voltage of 10 to several 10 V between tip and sample. A detailed μXPS analysis with Ga 3d, As 3d, and Al 2p photoelectrons combined with AFM height measurements showed a complex decomposition process of the nanostructures as a function of irradiation time and thickness, resulting in Al enrichment of the near surface region.

The formation, growth, shape, migration, coalescence, and evaporation of Ga droplets on GaAs and GaP surfaces has been studied in much more detail than the corresponding processes of liquid metal-Si eutectic droplets discussed in Sect. 5.3.1.1, using MEM, LEEM and UV-PEEM [151, 152, 308, 400, 401, 402, 403, 404, 405]. The droplets form during Langmuir (free) evaporation above the congruent evaporation temperature 625 °C of GaAs (638 °C of GaP) because of the higher evaporation rate of As (and P). They grow (decrease) in size at higher (lower) temperatures and move across the surface in a stick–slip motion driven by the chemical potential difference between their front and backside. The combination of experiment and theory resulted in a comprehensive picture of the thermodynamics and kinetics of droplets on surfaces, which is summarized in [405].

Metal droplets are used in the droplet epitaxy of nanowires. Au droplets have been studied to understand the growth of GaAs nanowires [406] and Ge nanowires [407]. In the first brief study LEEM and LEED confirmed that a Au wetting layer was necessary to stabilize the Au droplets for nanowire growth. The second very carefully executed study combined several complementary methods (STM, SEM, RHEED, TEM, and energy-dispersive X-ray spectroscopy) with XAS-PEEM, core level PEEM, and XPD in order to obtain deeper insight in the nanowire formation process. Of the many results obtained in this manner only a few salient features will be mentioned here. The Au nanodroplets were produced by annealing a 0.8–1.2 nm thick film deposited at room temperature on a well-cleaned Ge(111) surface at or somewhat above the eutectic temperature (350–400 °C), which resulted in alloy particles with sizes ranging from 5 to 200 nm, surrounded by a Au wetting layer. The Au 4f XPS spectrum from the particles differs little from that of pure Au while that from the wetting layer is shifted nearly 0.5 eV due to the strong bonding to the substrate. AFM and cross-sectional TEM images show that the particles are sitting on pedestals. The corresponding Ge and Au EDX data show that with increasing annealing temperature and increasing time an increasing amount of Ge is incorporated into the particles via material transport across the surface. Thus, vertical Ge nanowire growth is achieved even without an external Ge supply.

Nanowire/nanocolumn formation was studied for InN, starting from In droplets produced by annealing a 2.3 monolayer thick In film on Si(111) at 480 °C [408]. AlN nanodots, formed by ex situ nitridation of 30 nm thick Al islands with 90 nm × 130 nm lateral dimensions were subject of another study. The islands were produced by electron beam lithography and reactive ion etching on SiO2/Si(111) [409]. In both approaches the metals were nitrided in about 7 × 10−6 mbar NH3, in the first case at about 480 °C, in the second case by heating in steps from 680 to 800 °C. In the first study the combination of MEM, Si 2p XPEEM, In, Si and N XPS and AFM lead to the conclusion that not only the In droplets were nitrided, leading to InN nanowire formation, but also the In-covered Si surface, which caused In aggregation into droplets and partial In desorption. The second study was done with MEM, LEEM, XPEEM in the absorption mode and Al and Si in the core level mode as well as with Al, Si, and N XPS and found that nitridation did not change the structural order of the pattern but decreased the size of the individual nanopattern and made it more corrugated. Also the surrounding SiO2 layer was partially nitrided.

Summarizing this subsection, many of the studies reported in it are at the resolution limit of LEEM and XPEEM so that these methods served more or less to provide useful complementary information to that obtained with other methods such as STM, AFM, SEM, and TEM, in part with considerable, in part with limited success. Aberration-corrected instruments should make this kind of studies more successful.

5.3.2 Films on Metals

5.3.2.1 Metal Films

This subsection discusses the growth of metals on metals beyond the initial adsorption state, which was the subject of Sect. 5.2.1.3. In all film/substrate pairs studied the surface energy of the film material was higher than that of the substrate so that films grew in the Stranski-Krastanov mode (Refs. [332, 333, 334]) by forming first a two-dimensional adsorption or wetting layer before the growth of three-dimensional crystals. Under suitable growth conditions (temperature and deposition rate) a metastable wetting layer can be grown beyond the stable adsorption layer in the quasi-monolayer-by-monolayer (Frank-van der Merwe) growth mode before the layers in excess of the wetting layer coalesce into three-dimensional crystals. Many of the films had been studied previously with lateral averaging methods but the imaging methods discussed here have given a much deeper insight into their growth and microstructure. Most of the work has been done on close-packed surfaces of metals with high surface energy and high melting point, the bcc(110) surfaces of W and Mo, the hcp(0001) surface of Ru and the fcc(111) surface of Ni, but growth on bcc(100) and (111) and fcc(100) surfaces has also been studied. The film materials studied include Cu, Ag, Pd, Cr, Mg, Pb, and Sb as well as Fe, Ni, and Co. The latter will be discussed in Chap.  7 in connection with their ferromagnetic properties.

The growth of Cu on Mo(110) was the first system studied [230, 410] except for a cursory study of Ag on Mo(110) in the first LEEM paper [4]. In this study it had already been noticed that at high temperatures flat three-dimensional crystals, once nucleated on a terrace, grew rapidly along it due to supply of atoms diffusing rapidly on the wetting layer. Mundschau et al. [230, 410] found that crystals on different terraces developed different thickness due to step-limited growth as manifested by different reflectivities, caused by quantum size effects (QSE). The QSE effects in the growth of Cu on W(110) were studied in detail by Altman et al. [411] and analyzed in terms of a Koster-Kronig band structure model. The combined LEEM-UVPEEM studies of the same system by Koshikawa’s group [228, 412] found that also the PEEM intensity was layer-dependent and not simply determined by the work function but strongly influenced by the electronic structure of the film and in the wetting layer also of the substrate.

Ag grows on W(110) above two monolayers quasi-monolayer-by-monolayer with a distorted (111) plane parallel to the substrate. Chung et al. [413] have analyzed the resulting QSE effects by a dynamical theory analysis and—combined with first principle calculations—determined interlayer and film-substrate spacings. From the same experimental data Altman [414] determined the electronic band structure above the vacuum level using the phase accumulation model [415]. A quite different study of the growth of Ag on W(110) made use of the high diffusion rate of atoms on the wetting layer, the nucleation on steps and the tendency of growth over downhill steps. McCarty [416] deposited Ag on a W(110) surface on which 200 nm diameter holes had been produced by focused ion beam milling. At 290 °C crystals nucleate only at the step bunches of the holes and grow until they fill the holes (Fig. 5.44). At lower temperatures, e.g. at 110 °C, Ag crystals form everywhere but annealing at high temperature, e.g. 540 °C, transports the Ag into the holes. This procedure provides a method for producing three-dimensional metal nanostructures of controlled shape and position, not only of Ag on W(110) but also in other Stranski-Krastanov growth systems.
Fig. 5.44

LEEM images of the growth of Ag on W(110) with pits at 290 °C. Nucleation of three-dimensional crystals at the step bunch bordering the pits occurs shortly after completion of the wetting layer. Thereafter the crystals grow rapidly across the pits due to rapid diffusion of Ag atoms on the wetting layer surrounding the pits and fill the pits. Adapted with permission from Ref. [416]. Copyright 2006 American Chemical Society

Pd on W(110) has been subject of a brief study of the influence of the layer thickness on the valence band spectrum with μXPS [417] and of a μLEED I(V) study of the structure of different layer stacking sequences resulting in twin formation [418]. The growth, structure, and electronic structure of Mg on W(110) up to about 450 °C has been studied in great detail in a SPELEEM instrument with LEEM, LEED, and XPS from the submonolayer range to 12 monolayers [419]. Between 390 and 400 °C quasi-monolayer-by-monolayer growth occurs on top of the stable wetting layer, whose structure evolves similar to that of Pb submonolayers [420]. XPS showed clear differences between the bonding in the two-dimensional gas phase, the wetting layer, and thick films. The most detailed LEEM/LEED studies of metals on bcc(110) surfaces was made by McCarty’s group. Santos et al. [421] studied the first three monolayers and thick films of Cr on W(110) with LEEM and LEED I(V) measurements and derived their interlayer spacings with dynamical LEED analysis. Comparison with first principle calculations showed that these spacings were compatible only with antiferromagnetic ordering. McCarty et al. [422] made a very thorough study of the dewetting of thick Cr layers and of the dewetting during the transition from 3 to 4 monolayers, which gave deep insight into the energetics and kinetics of growth and dewetting.

Dewetting was also the subject of a study of Ag and Cu films on Ru(0001), which showed the importance of steps for the formation of three-dimensional crystals via downhill migration without the need of nucleation of two-dimensional islands on top of the crystal [423]. Other studies of the growth of Ag on Ru(0001) addressed the question how twin boundaries between crystals with different layer stacking sequence are eliminated during growth [424] and the decrease of the circular dichroism in the valence band photoemission with increasing thickness from 1 to 3 monolayers [425]. LEEM and LEED studies showed that Pd can be grown on Ru(0001) pseudomorphic quasi monolayer-by-monolayer up to 6 monolayers at 750 K, with twin formation-induced monatomic substrate steps. Interlayer spacings were obtained from I(V) data for these layers and for a thick film [418]. In a similar but more detailed study of the initial layer-by-layer growth of Co on Ru(0001) at 500 K only the first monolayer was found to be pseudomorphic, but quasi-monolayer-by-monolayer growth occurred also in the next few laterally relaxed monolayers. The island shapes reflected the stacking sequence of the layers whose spacings were determined via LEED I(V) analysis [426]. Contrary to Co, quasi-monolayer-by-monolayer growth was observed in a combined LEEM, LEED, and STM study of Mg on Ru(0001) up to 10 monolayers between room temperature and 390 K. Similar to Pb and Mg on W(110) the two-dimensional gas below one monolayer is compressed with increasing coverage into an ordered lattice. The spacing was found to decrease continuously to the lattice constant of a slightly compressed Mg(0001) plane at the completion of the monolayer. With increasing thickness stack faults develop [427].

On Ni(111) the growth of Ag, Au, and Pb has been studied with LEEM and LEED. In the case of Ag [428] and Au [429] growth on a 2 monolayer thick wetting layer occurred in the Stranski-Krastanov mode with multilayer islands at the temperatures studied. Ag, grew above 700 K along step edges before very thick crystals formed. The wetting layer had already bulk lateral periodicity, with two temperature dependent azimuthal orientations. LEED I(V) revealed a large film-substrate spacing. The study of Pb on Ni(111) focused on the influence of QSE effects on the thickness of the Pb islands, their movement across steps [430] and the rapid break-up of these metastable islands into hemispherical droplets above the growth temperature [431]. Interesting structural changes in the wetting layer were reported [430].

Summarizing the results of the studies of metals on densely packed surfaces of high surface energy metals, they all grow as expected in the Stranski-Krastanov mode, with wetting layers differing in thickness and structure depending upon the binding energy of the deposited metals and their misfit with the substrate. With proper growth conditions (temperature, deposition rate, terrace size) metastable growth in the quasi-Frank-van der Merwe mode has been achieved in most cases, sometimes up to 10 monolayers, which makes these films ideal model systems for the study of QSE effects. The LEEM studies have given detailed insight in the growth and structure, in particular in combination with STM, which were not accessible to the laterally averaging studies used in the past, such as island size, shape, and extended defects.

Similar phenomena were found in some less densely packed surfaces such as the (100) surface of bcc crystals. Although in some cases two-dimensional alloying occurs in the submonolayer range such as in Cu, Ag, and Au layers (see Sect. 5.2.1.3), pseudomorphic wetting layers form before dewetting occurs leading to three-dimensional crystals on top of a thinner wetting layer. For example, in the case of Cr on W(100) the wetting layer consists of three pseudomorphic layers at growth temperatures between 575 and 840 K but once islands of the forth layer form the somewhat compressed third layer dewetts resulting in three-dimensional crystals and a double layer [432]. Ag on Fe(100) turned out to be a particular interesting system with quasi-monolayer-by-monolayer growth at 300 K up to 18 monolayers thanks to the small lattice misfit, which allows nearly strain-free pseudomorphic growth. The first pure LEEM study of this system [433] showed well-pronounced QSE effects (Fig. 5.45a). Upon annealing to 500 K the 3 and the 4 monolayer thick films broke up into 2 monolayer thick regions (the wetting layer) and 5 monolayer thick regions (Fig. 5.45b), which are QSE-stabilized. During growth at temperatures above 400 K these preferred thicknesses also occur (“electronic growth” of the 5 monolayer thick regions). Kinetic effects in the growth accompanied the electronic growth. A later study [434] combined with detailed LEED spot profile analysis brought deeper insight in the mechanism of the thermal decomposition process into the electronically stabilized thicknesses. The originally flat films were found to have pinholes which were assumed to facilitate vacancy and adatom island nucleation. This study is a good illustration of the power of combining LEEM with quantitative LEED spot-profile analysis, but combining simple LEED with LEEM is frequently also helpful for the understanding of growth processes studied with LEEM. An example is the growth of Pb on Cu(100). In LEEM three-dimensional crystals with two azimuthal orientations are seen on top of the compressed Pb two-domain wetting layer. LEED allowed to correlate the orientations of the crystals with those of the domains [250]. All systems discussed up to now involve heteroepitaxial growth. However, once the influence of the substrate and possible QSE effects have faded away, growth occurs very similar to homoepitaxial growth, which has been studied theoretically by Tersoff et al. [435] and illustrated with the growth of Ag on Ag(111). They have developed a growth “phase diagram” for homoepitaxial growth, which predicts at high temperatures and high growth rates step flow growth, at lower growth rates with decreasing temperature layer-by-layer, multilayer, and reentrant layer-by-layer growth. Finally an example, Sm on Ag [436], should be mentioned yet in which the film does not break up but massive alloying occurs upon annealing above 400 °C as determined by XPS, accompanied by a strong decrease of the intensity in UVPEEM. Ultraviolet photoelectron spectroscopy (UPS) showed that the valence of Sm changes with coverage and temperature in a complex manner. Thus while XPS, UPS, and PEEM give some useful information little can be said about growth and structure without LEEM and LEED, which are the essential methods in this field.
Fig. 5.45

Quantum size effects in Ag films on Fe(100). (a) LEEM intensity as a function of energy and thickness measured during deposition at 300 K. (b) Top: LEEM images of initially uniform 3 (left) and 4 monolayer (right) thick films after annealing to 500 K. Bottom: local intensity spectra in the bright regions (left) and dark regions (right) of the two films (thin lines) compared to spectra of the 2 and 5 monolayer thick uniform films (thick lines). Adapted with permission from Ref. [433]. Copyright 2004 by the American Physical Society

Compared to the studies of metal film on metal surfaces discussed up to now and on semiconductor surfaces (Sect. 5.3.1.1) little work has been done on insulator surfaces. The structural changes of thick (50–135 nm) Mo films [19, 22] and thick (50–500 nm) Nb films [16, 17, 21] with (110) orientation on \( \left(11\overline{2}0\right) \) oriented Al2O3 surfaces at 1070–1200 K have been studied extensively with LEEM (Sect. 5.1.1) but the films were grown ex situ so that no information on growth was obtained. During growth studies it has to be kept in mind that the surface energy of insulator surfaces usually is lower than that of the film material. As a consequence, growth occurs in the Volmer-Weber mode described in Ref. [332], page 146 of Ref. [333] and page 171 of Ref. [334]. In a cathode lens electron microscope the grazing angle of incidence of the vapor beam is usually between 15° and 30°. Unless the surface mobility of the arriving atoms is low, three-dimensional crystals form, which shadow the surface behind the incident vapor so that the film is initially discontinuous and rough. Only once it has become continuous and thick enough so that it does not break up during annealing, the annealing can lead to surface smoothing, as it was done in the Nb and Mo films after transfer into the LEEM system. Another problem is that insulating substrates charge under the electron or photon beam, making imaging impossible. This problem can frequently alleviated by heating, producing some ionic or electronic conductivity but the increased surface mobility leads to larger crystals, which makes it more difficult to grow continuous films.

These problems can be reduced or even avoided by depositing a suitable nucleation layer, which ensures a high nucleation rate of the film and strong bonding to the substrate, or by surface modification so that the film material wets the substrate. An example of the second method is electron- or photon-stimulated desorption of oxygen on oxide surfaces via interatomic Auger transitions. The oxygen vacancies act as high binding energy sites for the arriving metal atoms and increase the nucleation density or can even lead to initial two-dimensional growth. An illustration of this approach is the study of the initial growth of Au on TiO2(110) with MEM, μLEED, and μXPS [170, 171]. In electron-irradiated regions, in which the surface has oxygen vacancy rows, Au grows in these rows, on unirradiated regions in three-dimensional crystals. In order to avoid charging, the crystal had to be heated to about 150 °C.

5.3.2.2 Inorganic Compound Films

Oxide films are the most-studied nonmetallic films because of their importance in corrosion and catalysis. Films grown simply by oxidizing bulk material (“homo-oxide films”), by oxidizing deposited films or by “reactive deposition” of metals in oxygen at low pressures (“hetero-oxide films”) have been studied. The first group includes oxide films on Ni, Cu, Ag, Ru, and NiAl. The oxidation of a Ni(111) surface has been studied as a function of oxygen dose and temperature with PEEM, MIEEM (metastable impact electron emission microscopy) and MIES (metastable impact electron spectroscopy) [437, 438]. A more thorough study with LEEM and μLEED determined the structure by comparing the (00) I(V) curves with band structure calculations [439]. NiO patches were observed already during room temperature oxidation after saturation of the chemisorption layer, with O2 exposure- and temperature-dependent patch size and density. The more detailed structural studies gave a more complex picture: at room temperature a continuous NiO layer 2–3 monolayers thick formed, while at 750 K a small number of (111)-oriented NiO crystals appear which spread across the surface leaving essentially clean Ni between them. Upon oxidation of the Cu(100) surface at 870 K (111)-oriented Cu2O crystals were observed already at an oxygen coverage of less than 2.6 monolayers on an ordered chemisorption layer [440].

Ag does not oxidize in molecular oxygen and forms only a chemisorbed oxygen layer when exposed to the more reactive NO2. However, during irradiation with electrons—and to a certain extent also with photons—a thick Ag2O film can be grown, which is initially crystalline but becomes amorphous with increasing thickness. This process has been studied thoroughly with LEEM, LEED, MEM, XPEEM, and valence band and core level μXPS and has been used for micropatterning of Ag(111) and polycrystalline Ag [441, 442]. The oxidation of Ru, whose oxide plays an important role in catalysis, has also been studied in O2 with UVPEEM and in NO2 with LEEM, LEEM-I(V), μLEED, and valence band XPS. The study in O2 [443, 444], which covered a wide temperature and pressure range (10−5–10−1 mbar, 300–1100 K) found three phases and determined the phase diagram. One of them, characterized by rod-shaped particles in three azimuthal orientations, existed only in a narrow pressure-dependent temperature range and was attributed to RuO2 crystals. The later studies in NO2 [182, 445] identified these particles via LEED clearly as (110)-oriented RuO2 crystals. LEEM-I(V) and valence band μXPS allowed the conclusion that the three coexisting phases at high temperature were chemisorbed O on top of Ru, O–Ru–O trilayer regions, and RuO2, a nice demonstration of the power of SPELEEM.

The probably most interesting oxidation layer is that on NiAl(110), which has been studied extensively by MCarty’s group with LEEM. Usually the surface is exposed at about 550 K to a high O2 dose (>103 mbar s). This produces an amorphous layer, which crystallizes into an about 5 Å thick κ-Al2O3 film in two azimuthal orientations upon annealing to 1000–1200 K. Dark-field LEEM imaging with LEED spots of the two domains showed that the domain size increases strongly with annealing temperature, while imaging with a NiAl LEED beam reveals oxide-free islands in the film whose density (size) decreases (increases) simultaneously. Step bunches acted as sinks for oxide-free regions, which spread onto the terraces at the highest temperatures studied (≈1300 K). At low oxygen doses annealing at 1200 K lead to step decoration by the oxide [47]. Exposure to low O2 doses (4–20 × 15−6 mbar s) at 800−1200 K produced quite different features: thin rods aligned parallel to the [001] direction of the substrate, accompanied with κ-Al2O3 islands above 950 K. From the temperature dependence of the growth rate of the rods an activation energy for growth of 1.2 eV was derived. Steps were found to slow the growth, annealing above the growth temperature to shrink the rows accompanied by the growth of the islands [446]. The islands themselves developed internal translation domain boundaries, attributed to stress relief, which could be clearly imaged by defocusing [447]. In the absence of O2 the κ-Al2O3 islands were found to rapidly decrease with increasing temperature. A detailed analysis showed that the O2 pressure necessary to stabilize the oxide film is much higher than that of bulk Al2O3, which shows that the film is kinetically stabilized far from equilibrium [448]. These studies are a good illustration of the advantage of LEEM over STM when information can be obtained only from the study of large areas and from kinetics.

Oxidation of deposited films has been studied only seldom. One example is the oxidation of one and two monolayer thick pseudomorphic Cr films on W(100) between 375 and 740 K at very low pressures [432]. One monolayer was found to be stable below 630 K but to dewet above 790 K, which resulted in three-dimensional Cr2O3 clusters surrounded by a penetrated (2 × 2) oxide structure on W. Oxidation of the two monolayer film caused phase separation already below 630 K but the crystals were not oxidized. This illustrates the complexity of the oxidation process when the substrate oxidizes more easily than the film. The second example is the oxidation of Mg and Al films on W(110), which show well-pronounced QSE effects up to more than ten monolayers thanks to their quasi-monolayer-by-monolayer growth near room temperature. Aballe et al. [449, 450] combined LEEM with XPEEM, 2p core electron, and valence band XPS and found that the oxide layer thickness correlated with the QSE effects seen in LEEM. Figure 5.46 [449] shows that the thickness of the oxide layer on 7 and 14 monolayer thick regions of the Mg film is largest, indicating a higher oxidation rate than at other thickness. The maximum at 7 monolayers correlated well with the maximum of the density of states at the Fermi level. The experimental results for Al films were compared with first principle calculations of the electronic structure of the films and showed that the oxidation rate is correlated with the electronic decay length into the vacuum, which enhances the reaction probability before the O2 molecule is reflected from the surface [450]. This study shows the importance of combining LEEM with XPEEM and spectroscopies with first principle calculations for the fundamental understanding of surface processes.
Fig. 5.46

Quantum size effect in the oxidation of thin Mg films on W(110). (a) LEEM image, taken with 1.3 eV electrons, (b) XPEEM image, taken with the chemically shifted Mg 2p photoelectrons of MgO. The thickness of the various regions of the film, indicated by the numbers, has been determined during the growth via the QSE contrast in the LEEM image. The thickness of the oxide layer formed by oxidation of the completed film with about 10−6 mbar s O2 at 50 °C is proportional to the brightness in (b). Adapted with permission from Ref. [449]. Copyright 2004 by the American Physical Society

Oxide films produced by reactive growth have been studied for several materials on several surfaces: Fe oxides on Ru(0001) and Pt(111); Ce oxides on Ru(0001), W(110), and Re(0001); and Ti oxides on Pt(111). Four Fe oxides have been grown: FeO, Fe3O4, γ-Fe2O3, and α-Fe2O3, which are interesting because of their magnetic properties and as supports in catalysis. In the first study [451] films were grown at about 900 K in 10−6 mbar O2 and annealed at about 1200 K, which resulted in large flat triangular crystals surrounded by a FeO wetting layer as identified with μLEED. In a subsequent study [452] films grown in this manner were studied also with μXPS and XMCD, which identified the crystals as Fe3O4 and their thickness as about 1 nm. Oxidizing these films with the more reactive NO2 gas converted the Fe3O4 islands into Fe2O3 based on μXPS without changing the wetting layer [453]. μLEED gave some information in these studies too but only XPS and XMCD allowed a reliable identification of the crystals. Growth on Pt(111) was studied with the same combination of methods and found to be very similar (Bauer et al, unpublished).

Ceria (CeO2) is an important oxide in catalysis because of its easy oxygen exchange in the bulk. CeO2 growth on Ru(0001) has been studied with LEEM and in particular with LEED at temperatures between 575 °C and 1000 °C at oxygen pressures from 2 × 10−7 to 6 × 10−7 mbar in the highly oxidized state and at 1 × 10−8 mbar at 360 °C in a low oxidized state [454]. In all cases the film grew in the Volmer-Weber mode with triangular (111)-oriented crystals in the expected manner (decreasing particle density and increasing crystal size with increasing temperature) in a few azimuthal orientations surrounded by a two-domain Ru-p(2 × 2)-O layer. Annealing of films grown at low temperature (460 °C) in vacuum did not produce the p((2 × 2)-pattern and ended at 1000 °C with CeO2 in parallel orientation to the substrate. No reaction between layer and substrate was observed. The situation is different for growth on W(110). In this case the film was grown only at one temperature and pressure (870 K, 5 × 10−7 mbar) and was featureless in LEEM so that analysis had to be based only on LEED and XPS [455]. After saturation of the high temperature oxygen superstructure of W(110) a characteristic LEED pattern developed which could be explained equally well with epitaxial CeO2 and Ce6WO12, both in (100) orientation. Distinction between the two explanations was achieved with detailed XPS studies, in particular of Ce 3d spectra and fitting them with the known spectra for Ce3+, characteristic for Ce6WO12, and Ce4+, characteristic for CeO2. From the thickness dependence of the relative contributions of the two components it was concluded that the layer consisted up to about 0.9 nm of Ce6WO12, on top of which CeO2 grew epitaxial. Thus, in this system CeO2 layer grows via an interfacial reaction layer.

Ce oxide layers can also be grown by first depositing Ce at room temperature in UHV and subsequent annealing in O2. This method was used for growth on Re(0001) by depositing about 0.8 monolayers Ce and annealing in 1 × 10−7 mbar O2 to 850 K. LEEM and LEED showed then flat (111)-oriented crystals with slightly expanded in-plane lattice constants in four azimuthal orientations. Re 4f and O 1s XPS indicated that the Re surface was oxidized and that the Ce oxide crystals were slightly reduced. The degree of reduction was determined by valence band μXPS of the islands, resonantly excited at the Ce3+ and Ce4+ binding energies, resulting in nearly uniform composition CeO1.64 of all islands [456].

Finally a LEEM/μLEED study of Ti oxide films grown on Pt(111) in 5 × 10−5 Pa O2 at 450 °C and at room temperature followed by annealing should be mentioned yet [457]. Movies taken during growth and annealing showed the evolution of a number of complex phases with thickness and temperature. Ti diffusion into the substrate and electron stimulated oxygen desorption were considered to be involved in these transitions. Such processes have to be considered in all oxide studies. This is also true for films grown ex situ. Unpublished studies (Bauer et al) of thick epitaxial (111)-oriented CeO2 films on YSZ have shown in addition other problems: diffusion of substrate impurities (Al, Si) into the film upon annealing and strong defect generation in XPEEM by photon-stimulated dissociation.

Hydride films are of interest, in particular MgH2 because of its large storage capability. The growth and thermal decomposition of MgH2 films on Ru(0001) was studied [458] by exposing up to 10 monolayer thick Mg films grown between 300 and 430 K to atomic hydrogen and annealing the resulting MgH2 films up to 500 K. In LEEM, taken near the mirror mode because of the destructive effect of electrons with energies above a few eV, a low density of nuclei was observed right after the start of the H2 exposure from which large crystals with different azimuthal orientations developed with increasing exposure. Their structure could not be determined because of the destructive beam effect. Thermal decomposition was thickness-dependent and started around 450 K as deduced from MEM and thermal desorption of H2. In the thickest film it was completed at 480 K [458]. This system is an example for an extreme case of radiation damage, limiting the number of electron and photon beam methods useful for characterization.

Another film growth method is thermal decomposition of molecules, which contain the constituents of the film. BN films have been grown in this manner by decomposition of borazine (BHNH)3 on the Ru(0001) and on the Fe(110) surface at about 800 °C in the low 10−7 mbar range. The study on the Ru(0001) surface [459] looked at the reaction of the BN film, called h-BN nanomesh because of the pattern of the hexagonal film seen in STM, with O2. The combination of MEM, LEED, and Ru 3d, B 1s and N 1s XPS showed that the removal of BN started at defects and occurred in 5 × 10−8 mbar O2 at about 750 °C within several minutes with a well-defined two-step reaction front, leaving a Ru surface with a (2 × 2) LEED pattern behind. In the second study [460], the h-BN monolayer was grown on an epitaxial Fe film on W(110). Because of the twofold symmetry of this surface the hexagonal BN lattice grew in two orientations as seen in LEED and dark-field LEEM. Accompanying STM studies showed a one-dimensional “washboard” corrugation in agreement with the periodicity of the LEED superstructure spots. The varying film-substrate contact in this corrugation was also supported by comparing N 1s NEXAFS and N 1s XPS spectra of BN on substrates with different film-substrate contact [460].

Summarizing the section on inorganic compound films, this is a field in which the combination of structural and spectroscopic information is in general needed to obtain a full understanding of the film. The films are generally heterogeneous so that good resolution in LEEM and μLEED are required for structural analysis and μXPS for chemical identification. In poorly ordered systems as encountered here in some cases MEM and PEEM can give some information but XPS is the main tool for analysis. As a result, SPELEEM systems at synchrotron radiation facilities dominate this field.

5.3.3 Organic Films

The development of organic thin film semiconductor devices around the turn of the century such as organic field-effect transistors (OFETs) and organic light-emitting diodes has made the study of organic thin films an important application field of cathode lens electron microscopy and associated techniques. Film morphology, crystallinity, and the orientation of the molecules in the crystals determine to a large extent the properties of the devices, together with the interface of the organic film with the electrodes (metal or semiconductor) of the device. For example charge transport is very sensitive to crystallinity and best perpendicular to the molecular axis while light emission is polarized parallel to it. The contact resistance depends strongly on the molecule/electrode material interaction. Many of the crystals can grow in different polymorphic forms. All these features are controlled by film deposition and interface conditions, which make surface characterization and in situ deposition essential tools for device optimization.

AFM and STM are frequently used in this endeavor but LEEM, UVPEEM, XPEEM combined with LEED are major tools too, in particular when combined with linear polarized NEXAFS, which allows determination of the orientation of the molecules. Of course these experiments are limited by the requirement that the molecules do not break up during sublimation and condensation. Radiation damage is another limiting factor, not only at the high flux densities in XPEEM but also in UVPEEM. However, there are many aromatic molecules, which satisfy these requirements. The first studies of organic films compared the possibilities of different imaging modes (UVPEEM, MIEEM, LEEM) using the square chloroaluminum phthalocyanine molecule (C32H18AlClN8) on MoS2. No contrast was seen in LEEM [461] but weak contrast between first and second layer in PEEM and to a lesser extent in MIEEM [461, 462]. The second layer appeared brighter, which led to the conclusion that emission is not determined by the work function but by the electronic structure. Structure and growth could not be studied because of limited resolution.

No more PEEM work was done on phthalocyanine films until much later when more sophisticated equipment became available. In one study [463] a better PEEM instrument with tunable laser excitation was combined with high resolution UPS in a separate system to correlate image contrast with work function and electronic structure of up to two monolayers thick Pb phthalocyanine films on graphite. A second study [464], interesting from the methodical point of view, combined laterally averaging angle-dependent linear polarized NEXAFS in an XPS system with laterally resolved NEXAFS at fixed angle in a PEEM. The goal was to determine the molecular orientation of silicon phthalocyanine chloride via NEXAFS resonance absorption spectra. This molecule has a Si atom in the center to which a Cl atom is attached on each side of the molecule normal to the molecular plane. At the Si K absorption edge at 1840 eV there are two 1s–σ* transitions separated by about 2 eV, one to the in-plane Si–N bond, the other one to the Si–Cl bond normal to the plane. After determining the intensity ratio of these two transitions as function of angle in the angle-dependent NEXAFS measurements the authors could determine the tilt angle of the molecules on the microscopic level in PEEM by analyzing the local intensity in a series of spectroscopic images around the absorption edge. This method was applied to a study of five monolayers thick patterned films on a Au surface [464].

The first studies of the growth of organic films were made in Tromp’s group, stimulated by the development of pentacene transistors at IBM. The linear molecule pentacene (C22H14) can be easily sublimated from purified crystals at 200–300 °C. Meyer zu Heringdorf et al. [465, 466] deposited it on Si(100) surfaces and studied film growth with UVPEEM. In order to minimize radiation damage they reduced exposure time per image to about 1 s in 60 s. Figure 5.47 [466] shows some of their results. The top row (a–c) shows the dendritic quasi-monolayer-by-monolayer growth on the clean (2 × 1) surface on top of a wetting layer of flat lying molecules. μLEED showed that the individual islands are single crystals with random azimuthal orientation. From the rate dependence of the nucleation rate a critical nucleus size of six molecules was deduced. The bottom row (d–f) shows the influence of the surface condition on the growth. A comparison of the images shows that the number density of crystals is an order of magnitude smaller on the clean and cyclohexane-passivated surface than on the oxidized surface, resulting in a much smaller number of grain boundaries, which are detrimental to charge transport. A later combined PEEM, LEEM, and LEED study confirmed the beneficial effects of passivating layers of two other organic adsorbates and found that the pentacene crystals grew epitaxial on top of them to very large grain sizes [467].
Fig. 5.47

UVPEEM images of the growth of pentacene on Si at room temperature. (ac) growth on the clean Si(100)(2 × 1) surface: (a) one layer, (b) two layers, (c) three layers. (df) Growth on different surface terminations: (d) on plasma-cleaned oxide, (e) on the clean (2 × 1) surface, (f) on a well-ordered self-assembled cycloxane (C6H10) layer, which saturates the dangling bonds and suppresses the flat-lying pentacene wetting layer present in (e). Adapted from Ref. [466] with permission from Springer Science + Business Media

While the early work concentrated on the technologically important (100) surface, later work in Sakurai’s group studied pentacene growth on Si(111) surfaces and other surfaces with hexagonal surface nets, starting with the hydrogen-terminated Si(111) surface [468, 469]. These studies combined tilted-beam LEEM with μLEED, which showed complex anisotropic dendritic polycrystalline growth with the (001) plane parallel (molecules perpendicular to the surface) to the substrate and the b axis of the two-dimensional pentacene unit cell somewhat rotated relative to the \( \left\langle 1\overline{1}0\right\rangle \) and \( \left\langle 11\overline{2}\right\rangle \) directions of the substrate. The b axis was the main growth direction of both the main branch and of the side branches of the dendrites. Thus the direction of b determined together with the density gradient in the two-dimensional molecular gas phase the shape of the crystals. The crystals grew directly on the H-terminated surface from a small number of nuclei radially outward and not on a wetting layer as on the clean (111) or (100)-(2 × 1) surface.

The strong influence of the substrate on the growth mode is even more dramatically demonstrated by the comparison with the growth on the Bi(0001) surface [470, 471] on which pentacene grows from local defects in the step flow mode instead dendritic, with twinned domains. This is attributed to the fact that every second molecular row parallel to the a direction in the ab plane of pentacene coincides with the Bi atomic rows, resulting in only one-dimensional misfit. The preferred growth direction was perpendicular to these rows as in the dendritic growth on the H-terminated Si(111) surface although the bonding is strongest along the row direction (a axis). Similar growth behavior was found on the Bi-terminated surface (“α√3-Bi-Si(111)” [471, 472]. This shows that the preferential growth direction is determined by the molecular structure and bonding, which determines the crystal anisotropy. The detailed analysis of the data showed also a significant reorientation barrier for the transition from the diffusing molecules which are parallel to the substrate to the crystal in which they are upright.

Pentacene growth was also studied on several other surfaces with hexagonal surface net: on a fullerene (C60) film on the Bi(0001) surface [473] on the (0001) graphite surface [474], on the BN monolayer on Ru(0001) [475] and on the (√3 × √3)R30°-Ag/Si(111) surface [476]. On the fullerene surface, crystals with flat lying molecules formed initially dendritic islands, on top of which crystals with nearly perpendicular molecules in several azimuthal orientations grew as on the surfaces mentioned before. On graphite growth on the wetting layer occurred also mainly with nearly perpendicular molecules but continued partially also with flat lying molecules. The crystal size increased strongly with deposition temperature and all crystals had preferred azimuthal alignments. Growth on BN starts already in the submonolayer range with nearly perpendicular molecules when the density of the flat-lying molecules reaches a critical density. Large two-dimensional crystals with several azimuthal orientations form, similar to the growth on the Bi(0001) surface. In the second monolayer the nucleation density is much higher and the islands have fractal shapes. This study was combined with angle-dependent NEXASF spectroscopy, which clearly showed the orientation of the molecules. A e (√3 × √3)R30°-Ag/Si(111) surface: the molecules were lying flat on the substrate not only in the wetting layer but also in the several 10 nm thick crystals growing on it. Summarizing the growth on surfaces with hexagonal surface nets, the results show the strong influence of the molecule-substrate interaction, which determines the orientation of the molecules, the diffusion process, and the resulting island shape, but also influences the growth on top of the first layer. In all cases the film grows with preferred azimuthal orientations and orientation-dependent growth rates.

The results obtained from the growth studies on well-defined hexagonal surfaces are of fundamental interest and allow some general suggestions for optimizing organic thin film devices but practical devices usually are grown on SiO2 and polycrystalline metal surfaces. The problems encountered on such surfaces have also been studied with LEEM and PEEM. The LEEM study [477] addressed the question how self-assembled monolayers, which are frequently used to improve the performance of OFETs by reducing the trapping sites at the film-substrate interface, influence the growth. It found that hexamethyldisilazane ((CH3)3SiNHSi(CH3)3) and octadecyltrichlorosilane (CH3(CH2)17SiCl3) adsorption on SiO2 caused spontaneous aggregation deleterious to OFET performance due to the decrease and loss of the conduction path. This effect was attributed to the lower surface energy of the layers compared to that of pentacene, causing Volmer-Weber growth.

A PEEM study [478] was concerned with the problems, which can occur at the pentacene/electrode contact. Au contact patterns were grown on a SiO2 layer on a Si(111) wafer and pentacene was grown on this patterned surface. PEEM showed that a denuded zone (“groove”) at the SiO2–Au boundary developed during deposition, which persisted up to thick films, also when the SiO2 surface was precovered with a ((CH3)3SiNHSi(CH3)3 self-assembled (SAM) monolayer as in the preceding LEEM study. However when the Au surface was covered with a octanethiol (CH3(CH2)7SH) SAM no groove formed during growth. Conductivity measurements showed that this caused an increase of the charge-carrier mobility by orders of magnitude due to the elimination of the high resistance region at the contact. The groove formation was attributed to the high nucleation and growth rate of pentacene crystals with flat-lying molecules on the clean Au surface compared to that on SiO2.This reduced the molecule concentration in the surroundings of Au, while on the octanethiol-covered Au surface pentacene grew in the same manner as on SiO2 due to the lowering of the surface energy by the SAM, thus eliminating the concentration gradient. This experiment is a nice example of the usefulness of UVPEEM for the understanding of technological problems. Such problems stimulated also a LEEM study of the growth of 6,13-pentacenequinone (C22H12O2) on Si(111), which forms upon oxidation of pentacene and causes deterioration of OFETs exposed to air [479]. The study revealed a very complex growth process. Initially an amorphous wetting layer forms followed by the nucleation and growth of compact, amorphous islands, from which long curved feather-like crystals grow with the ab plane parallel to the substrate. They have left-handed and right-handed curvature, related to the tilt direction of the nearly upright molecules, and the in-plane orientation changes continuously with the curvature. During very slow growth large rectangular crystals formed, which indicates that the feather-like growth is determined by kinetics.

While pentacene was at the center of the interest in organic films, some others have been studied, at least briefly. A UVPEEM study of the growth of PTCDA (C24H8O6) on the Ag(111) surface looked at the influence of the surface morphology on the film structure [480]. In the first monolayer the nucleation rate in stepped regions was much larger than in flat regions, in the second monolayer the difference was much smaller and in the third monolayer three-dimensional crystals formed, indicating Stranski-Krastanov growth. Anthracene (C14H10) was found in standard UVPEEM studies to grow on Si(111) in the same manner as pentacene with a wetting layer of flat-lying molecules on which dendritic crystal layers grow with nearly perpendicular molecules [481]. 2PPE imaging with 400 nm fs laser light ( = 3.1 eV) gave much deeper insight, in particular by making use of the polarization of the light [482]. 3.1 eV is the energy of the lowest lying exciton level, which can be excited only if the E vector has a component along the b axis. Thus the intensity of the two-electron photoemission depends upon the orientation of the b axis relative to the E vector and allows determination of the azimuthal orientation of the crystals. This method of orientation determination, which is based on the selective excitation effect in a crystal, is different from the orientation determination with linear-polarized NEXAFS, which gives the orientation of a molecule via the dependence of the excitation probability of π and σ orbitals on the relative orientation of E and the orbital orientation under resonance condition.

Diindenoperylene (C32H16) films on polycrystalline, (111), and (100) Au surfaces were subject of a series of UVPEEM, XPEEM, and LEEM studies combined with linear polarized NEXAFS in a separate system by Casu et al. [483, 484, 485, 486]. Their results can be summarized as follows: the films grow in the Stranski-Krastanov mode with flat-lying molecules at sub-monolayer coverage and thereafter with fractal crystals in which the molecules are tilted (relative to the surface) by angles increasing from 23° to 46° with thickness from 0.6 to 3.6 nm. LEEM contrast in the three-dimensional crystals was attributed to QSE effects.

The initial growth, structure, and thermal stability of 4,4-biphenyldicarboxylic-acid (BDA, C14H12O4) on Cu(100) was studied in detail with LEEM and μLEED in Poelsema’s group [487, 488, 489, 490, 491]. These studies have given deep insight into the complex processes, which can occur in the submonolayer range. Nucleation of two-dimensional crystals with neighboring flat-lying molecules in orthogonal orientation occurred rapidly after a considerable incubation time (t i), followed by lateral growth without much further nucleation until renewed nucleation occurred at a critical crystal size (2 × 104 nm2) during growth at room temperature. The renewed nucleation at 1.5ti was attributed to stress build-up in the existing crystals, which made further growth energetically unfavorable. During growth at 448 K further growth was not suppressed indicating the absence of stress but at this temperature the crystals decayed with a decay exponent of 0.6. LEED showed that the crystals were epitaxial, with the long axis of the molecules parallel to the 〈110〉 directions of Cu [487]. Further LEEM intensity studies allowed determining the phase boundary between the two-dimensional gas phase and the crystal + gas coexistence region as a function of temperature from which a cohesive energy of 0.35 eV was derived based on a lattice gas model. The analysis gave also a large entropy contribution, which was attributed to the large size and shape of the molecule and considered to be responsible for the high density in the gas phase relative to that in the crystals. From island decay studies diffusion constants were derived, steps did not act as limiting diffusion barriers [488]. A detailed study of the formation and decay of crystal nuclei between 297 and 332 K gave critical island sizes between 400 and 600 nm2 and a Gibbs free energy for nucleation of 4.0 eV [489]. Another interesting observation, found by combining LEEM with μLEED, was the formation of a compressed phase inside large crystals during growth above 370 K, in four rotational domains and 14 % higher density than the uncompressed region. It was stabilized by the incoming molecular flux but disappeared without it [490]. At high temperature (410 K) growth was strongly influenced by steps: once a crystal had formed on a terrace it continued to grow only along the terrace. This is similar to the growth of metals on close-packed metal surfaces, as illustrated by Cu on W(110). At high temperature steps are no diffusion barriers so that atoms/molecules can easily surmount them get incorporated in the crystal while growth on neighboring terraces requires a nucleation event. There is, however, a difference between metals and large elongated molecules: they have to arrive at the growth front in the correct orientation or have to rotate, which results in a roughening of the growth front as observed [491].

Another interesting phenomenon, which has been found in BDA films on Au(111) in the submonolayer range, is the formation of three different crystalline phases, depending upon growth or annealing temperature [492]. The crystals grown at room temperature are needle-shaped and transform into compact ramified crystals upon annealing above about 330 K with a somewhat different crystal structure. Growth above about 330 K produces compact crystals with well-defined straight edges and a completely different structure. In all structures the molecules are lying flat. The transition between the first two structures has been attributed to loss of the H at the carboxyl group at the ends of the molecules (“deprotonation”), which changes the lateral interactions between them, causing the structural changes. This suggests that on the Cu(100) surface, with which the molecule interact much stronger, significant deprotonation should occur too. Another substrate with which BDA interacts more weakly than with Cu is graphene. On graphene supported on Ir(111) the crystals nucleated at room temperature at the wrinkles and protrusions of the graphene layer and grew across it in azimuthal orientations determined by the wrinkle direction. During growth at 370 K no crystals formed on graphene but decorated the substrate at the edges of the graphene flakes. During annealing of room temperature-grown films the crystals disappeared above 350 K but nucleated again upon cooling. This indicates that no sublimation occurred and molecules diffused back from the surrounding Ir onto the graphene. No second layer growth was observed [493].

Para-sexiphenyl (6P, C22H12O2) is another model molecule, whose growth was studied with several imaging methods in Poelsema’s group with LEEM and μLEED on graphene and the supporting Ir(111) surface in the temperature range from 240 K [494, 495] to 405 K [496]. 6P forms initially an epitaxial layer of flat-lying molecules. This converts into an alternating row of flat-lying and upright (edge-on) epitaxial layer, which nucleates at the wrinkles of the graphene and completes the first layer. This molecular arrangement is maintained with further growth. At 240 K this occurs in a quasi-monolayer-by-monolayer mode as evidenced by characteristic changes of the image intensity, suggesting quantum size contrast, up to the largest thickness studied (4.35 monolayers) but at 320 K the crystals grow as needles across the surface, starting from wrinkles again. Above 380 K the molecules diffuse onto the Ir surface and form near the graphene edge crystals with upright standing molecules. At lower temperatures these crystals grow all over the Ir surface with ramified shape.

Fleming et al. [497] studied the growth of the same molecule (6P) on the Cu(110)(2 × 1)-O surface from 25 to 150 °C with UVPEEM. It shows some similarities to that on graphene but also distinct differences. The mode of film growth, which probably occurs also in many other organic films, is illustrated in Fig. 5.48 [497]. The initially flat-lying molecules tilt upward forming a monolayer of tilted molecules, which is stable up to 220 °C. On this monolayer a second layer grows, also with initially flat-lying molecules which turn upright with increasing coverage to form a second monolayer of molecules tilted into the opposite direction. This layer is metastable and breaks up into needles along the Cu \( \left[1\overline{1}0\right] \) direction, which grow in length and thickness with increasing deposition time, initially by the supply of molecules from the second monolayer until dewetting is completed. From a detailed study and analysis of the evolution of the image intensity and the needle dimensions as a function of deposition rate, temperature and annealing temperature, combined with AFM measurements and results from other studies the authors have developed a very detailed model of the growth of these films.
Fig. 5.48

(a) Two typical plots of the photoelectron intensity as a function of the thickness of para-sexiphenyl (6P) films deposited at 140 °C at a rate of about 1/3 monolayer/min. Unless otherwise stated, all discussions in the text refer to the darker curve. The change in work function from clean substrate (4.8 eV) to 1 monolayer at point B is about −0.4 eV. The evaporator shutter was continuously open. The deposition rate had no significant influence on the trends of the curves. Inset: schematic illustrating the structure of the various layers that assemble. (b) Frames from the PEEM video taken at the points marked in (a) at the darker curve. (c) Schematic diagrams illustrating the orientation and packing of the molecules that explain the photoelectron intensity at the points marked in (a). Only the molecular end-on view is shown; the short axis of the 6P molecule is illustrated by a double-headed arrow. The packing angle θ is the angle between the molecular short axis and the surface. The O-induced corrugated surface reconstruction is indicated by the O rows which are normal to the plane of the drawing and have a distance of 5.1 Å. Adapted with permission from Ref. [497]. © IOP Publishing 2009. All rights reserved

Another film/substrate system studied with PEEM in connection with STM is α-sexithiophene (α-6T, C24H14S6) on Ag(110) [498, 499, 500]. The dependence of the PEEM intensity upon deposition time showed the formation of two monolayers. With STM they were identified to consist of flat-lying molecules, which were partially aligned in the first layer but perfectly aligned in the double layer. On top of this wetting layer a third layer formed in which needle-shaped crystals grew across the surface, partially depopulating the third layer. As judged by the coverage of needles their height was estimated to be 100 nm after deposition of 33 monolayers. Thin platelets were observed too occasionally. The use of D2 and He I light ( = 6.4 and 21.2 eV, respectively) did not produce significant information in addition to imaging with the Hg lamp but the polarization had a strong influence on the emission, in particular from the needles. Local UPS spectra obtained by imaging with a retarding grid energy filter and varying the retarding voltage showed little difference between wetting layer and needles suggesting the same molecular orientation in both of them. The spectrum of the platelets, however, was very different and assigned to nearly upright standing molecules.

Summarizing the in situ work on molecular layers, cathode lens electron microscopy methods have made important contributions although most of the understanding of the molecular structure and orientation has been obtained by other methods such as STM, AFM, Fourier transform infrared spectroscopy, UPS, to name the most important ones. Cathode lens electron microscopy allows the study of the kinetics of film growth, from which energetic parameters can be extracted, at least in the early growth stages. It is indispensable in the later growth stages, when three-dimensional crystals form. Used alone, it gives an incomplete picture. Therefore, in most of the studies mentioned the interpretation of the results relies more or less heavily on information from the other methods mentioned above. Although the general picture of growth and structure of molecular layers, that has emerged, is similar to what is known from the growth of atomic layers, it is strongly modified by the shape and bonding of molecules. Many questions are still open, whose answer will require systematic studies of more molecule-substrate combinations over a wider temperature range, with improved methods such as excitation with polarized tunable light in PEEM or addition of I(V) measurements to μLEED.

The organic film studies discussed up to now were made in situ with PEEM, LEEM, and μLEED but XPEEM and XPS were used only occasionally. There are two other groups of organic films, prepared ex situ, in which XPEEM and XPS are the main laterally resolving and only chemically resolving surface characterization methods: Langmuir-Blodgett films and polymer films. The interest in these fields is in phase-separating two- or more components systems. Two-component Langmuir-Blodgett film samples are prepared by spreading a mixture of mutually immiscible surfactant molecules on water and transferring the Langmuir-Blodgett film onto a Si wafer or other substrate. This is done by submerging the wafer below the film and pulling it from the water, followed by drying before transfer into the PEEM system. Chemical contrast in imaging is achieved via NEXAFS spectromicroscopy at the C K edge making use of the energy difference between the absorption resonances of different bonds, for example of the C 1s → σ*C–F resonance at  = 282.8 eV in a perfluorinated region and of the C 1s → σ*C–H resonance at  = 287.9 eV in a hydrocarbon-rich region. Excitation at these two photon energies allows a clear distinction between the corresponding regions. This method has been used very successfully, together with AFM, in studies of phase-separated Langmuir-Blodgett films of fluor-free and fluorinated organic acids [501, 502] and even of more complex systems, in which the spectra of the two components were not clearly separated but differed only in width [503].

Polymers were studied for a variety of reasons, ranging from technical to medical applications. A polymer was actually the first non-biological organic material studied with cathode lens microscopy [504, 505]. The subject of this study was a polyimide film which is used to orient the molecules in the liquid crystal layer of flat panel displays. In order to achieve this, the molecules in the polyimide film have to be oriented by rubbing parallel to the direction of rubbing. As the intensity of the NEXAFS resonances depends upon the angle between E and the orientation of the final state orbital (σ* or π*) the intensity of the resonance peaks allows determination of the orientation of the molecules as discussed before. The image series (“stacks”) taken in a 50 eV wide photon energy range around the C K absorption edge with the rubbing trace parallel and perpendicular to the in-plane component of E showed clear intensity differences at the π* resonance. From these intensity differences the minimum pressure needed for orientation of the molecules could be determined. In a later study [506] molecular orientation was produced by irradiation with linear polarized 4.8 eV laser irradiation, which preferentially decomposes polymer chains that are parallel to E. As a result the surface layer, in which decomposition occurs is enriched with molecules aligned perpendicular to E. In this study not only the resonance at the C K edge was used as before but also resonances at the N K edge and the O K edge, with the latter showing the largest dichroism.

Semiconducting polymers are also important in electronics. High mobility is usually attributed to high intrachain and interchain order in the film. This is achieved by annealing the film, which causes considerable changes in the film morphology such as grain boundary formation and break-up into separate ordered regions, resulting in a deterioration of the conductivity. A UVPEEM study of these morphology changes in 18 nm thick poly(3-hexylthiophene) films on Si upon annealing at 170 and 220 °C [507] indeed showed dramatic changes in the image contrast connected with the grain growth. They were interpreted to differences of the work function and ionization potential caused by variations of the molecule orientation. A study of a doped polypyrrole film on an insulating fluorinated ethylene propylene substrate, in part covered with a Cu grid pattern, combined XPEEM with NEXAFS and some AFM studies [508]. The film was found to be chemically homogeneous in XPEEM images at the absorption edges of C, N, and S. NEXAFS did not show dichroism, indicating that the molecules were not ordered. The study was hampered by considerable radiation damage by the high photon flux density of the undulator beamline used in these experiments. This example illustrate one of the limitations of cathode lens microscopy.

The polymer films studied most intensely with XPEEM consist of two immiscible polymers, which phase separate upon annealing and are used mainly for competitive adsorption studies on the phase-separated mixtures. XPEEM is frequently combined with scanning transmission X-ray microscopy (STXM), which gives the structure of the bulk material for comparison with the surface structure seen in XPEEM. The first combination studied was polystyrene (PS) and poly(methylmethacrylate) (PMMA) [509], a system also used later for protein adsorption [510]. Figure 5.49 [510] illustrates this kind of studies for an annealed PS:PMMA film on Si. The NEXAFS spectrum of the two polymers is very different as seen in the absorption spectra (b) so that the two components can be distinguished easily by imaging at the corresponding resonance transitions, C 1s → π*C–C of PS at 285.1 eV and of PMMA at 288.4 eV, as shown in the NEXAFS images. Quantitative information has been extracted by detailed image analysis, taking into account several possible artifacts. This is of particular importance in studies of protein adsorption on these polymer blends because of the small signal from the proteins and the small difference of their resonance energies from those of the substrate. For example, the resonance of human serum albumin protein is shifted only 0.25 eV relative to that of PMMA. Nevertheless preferred adsorption of the protein at the interfacial PS/PMMA region could be established with a resolution of better than 100 nm [510, 511, 512]. Other polymer combinations and protein adsorption were studied too, for example polystyrene (PS)-polylactide (PLA) blends [513] and PS blended with a copolymer of polymethylmethacrylate and polyacrylic acid (PMMA-b-PAA) [514]. In both studies the XPEEM and XSTM were combined with AFM in order to obtain also information on the film morphology. The first one was concerned with the dependence of the morphology and chemical composition upon the polymer ratio and annealing conditions (time and temperature) and found that the AFM images did not correspond to the XPEEM images. The second one studied the adsorption of positively charged human serum albumin and a negatively charged peptide. Quantitative analysis showed that the thickness of the adsorbate film was strongly correlated with its charge, which suggests that adsorption is driven by electrostatic interactions with the negatively charged surface during adsorption. A quite different method of two-component polymer pattern formation was used to study the adsorption of the protein ubiquitin. It used two superimposed polymer films. The water-soluble top film was patterned by electron beam lithography, which formed cross-linked patterns, which remained after washing. Protein was found to adsorb only on the cross-linked polymer [515]. For a review of the work on polymer films with XPEEM and XSTM see Hitchcock et al. [516].
Fig. 5.49

(a) X-PEEM images at 283, 285.1, 288.4, and 290 eV, 4 of the 40 images in the C 1s region of an annealed 28:72 (w/w) PS:PMMA blend thin film spun cast on native oxide Si (the average surface composition is 57:43). (b) Spectra from the indicated spots. (c and d) Component maps of PS and PMMA derived by singular value decomposition of the C 1s image sequence. (e) Color coded composite map (red: PS; green: PMMA). Adapted with permission from Ref. [510]. Copyright 2004 Elsevier

References

  1. 1.
    Yang, Y.-N., Williams, E.D.: High atom density in the “1 × 1” phase and origin of the metastable reconstructions on Si(111). Phys. Rev. Lett. 72, 1862–1865 (1994)Google Scholar
  2. 2.
    Ong, W.J., Tok, E.S.: Real time dynamics of Si magic clusters mediating phase transformation: Si(111)-(1 × 1) to (7 × 7) reconstruction revisited. Surf. Sci. 606, 1037–1044 (2012)Google Scholar
  3. 3.
    Witt, W.: Untersuchungen 2-dimensionaler Phasenübergänge mit einem hochauflösenden LEED-Diffraktometer. Ph.D. thesis, Technische Universität Clausthal (1984)Google Scholar
  4. 4.
    Telieps, W., Bauer, E.: An analytical reflection and emission UHV surface electron microscope. Ultramicroscopy 17, 57–66 (1985)Google Scholar
  5. 5.
    Kastler, A.: Un systeme de franges de diffraction à grand contraste. Rev. Opt. 29, 307–314 (1950)Google Scholar
  6. 6.
    Bauer, E., Mundschau, M., Swiech, W., Telieps, W.: Low energy electron microscopy. In: Goodhew, P.J., Dickinson, H.G. (eds.) Proc. 9th European Congress on Electron Microscopy (EUREM 88), 1988. IOP Publ., Inst. Phys. Conf. Ser. No 93, vol. 1, Chapter 4, pp. 213–218. York, England (1988)Google Scholar
  7. 7.
    Mundschau, M., Bauer, E., Swiech, W.: Photoemission microscopy and atomic steps on Mo(011). Surf. Sci. 203, 412–422 (1988)Google Scholar
  8. 8.
    Mundschau, M., Bauer, E., Swiech, W.: Modification of atomic steps by adsorbates observed by low energy electron microscopy and photoemission microscopy. Catal. Lett. 1, 405–412 (1988)Google Scholar
  9. 9.
    Mundschau, M., Bauer, E., Swiech, W.: Defects on the surface of Mo{011} observed by low-energy electron microscopy. Phil. Mag. A 59, 217–226 (1989)Google Scholar
  10. 10.
    Mundschau, M., Bauer, E., Swiech, W.: The role of ledges in vapor/solid phase transformations observed by low-energy electron microscopy and photoemission electron microscopy. Metall. Trans. 22A, 1311–1315 (1991)Google Scholar
  11. 11.
    Telieps, W., Mundschau, M., Bauer, E.: Dark field imaging with LEEM. Optik 77, 93–97 (1987)Google Scholar
  12. 12.
    Telieps, W., Mundschau, M., Bauer, E.: Surface domain structure of reconstructed Au(100) observed by dark field low energy electron microscopy. Surf. Sci. 225, 87–96 (1990)Google Scholar
  13. 13.
    Altman, M.S., Chiang, S., Statiris, P., Gustafsson, T., Bauer, E.: Stress-induced microfacetted reconstructions of the Pb(110) surface. In: Xi, X.D., Tong, S.Y., van Hove, M.A. (eds.) The Structure of Surfaces IV, pp. 183–191. World Scientific Publ, Singapore (1994)Google Scholar
  14. 14.
    Altman, M.S., Bauer, E.: The reconstruction of the Pb(110) surface studied by low energy electron microscopy. Surf. Sci. 344, 51–64 (1995)Google Scholar
  15. 15.
    Altman, M.S.: Evidence of a Pb(110) (2x2) reconstruction. Surf. Sci. 344, 65–69 (1995)Google Scholar
  16. 16.
    Swiech, W., Appleton, R.S., Wiemeyer, B.D., Flynn, C.P.: LEEM investigation of refractory metals grown on insulating substrates: Nb(110) on sapphire (1120). Surf. Rev. Lett. 5, 1221–1231 (1998)Google Scholar
  17. 17.
    Flynn, C.P., Świech, W.: Periodic states in the constrained equilibrium of vicinal Nb (011) miscut along [100]. Phys. Rev. Lett. 83, 3482–3485 (1999)Google Scholar
  18. 18.
    Mundschau, M., Swiech, W., Durfee, C.S., Flynn, C.P.: Slip propagation in epitaxial Mo (011) studied by low-energy electron microscopy. Surf. Sci. 440, L831–L834 (1999)Google Scholar
  19. 19.
    Swiech, W., Mundschau, M., Flynn, C.P.: Characterization of single crystal films of molybdenum (011) grown by molecular beam epitaxy on sapphire (1120) and studied by low-energy electron microscopy. Surf. Sci. 437, 61–74 (1999)Google Scholar
  20. 20.
    Swiech, W., Mundschau, M., Flynn, C.P.: Interfacial defects in thin refractory metal films imaged by low-energy electron microscopy. Appl. Phys. Lett. 74, 2626–2628 (1999)Google Scholar
  21. 21.
    Flynn, C.P., Swieçh, W., Appleton, R.S., Ondrejcek, M.: Nanofaceting of vicinal Nb(011). Phys. Rev. B 62, 2096–2107 (2000)Google Scholar
  22. 22.
    Swiech, W., Ondrejcek, M., Appleton, R.S., Durfee, C.S., Mundschau, M., Flynn, C.P.: LEEM investigations of bcc metals grown heteroepitaxially on sapphire. In: Bentley, J., Allen, C., Dahmen, U., Petrov, I. (eds.) Proc. MRS Symposium advances in materials problem solving with the electron microscopy, 2001, pp. 27–38. MRS Symposium Proceedings. Boston, USA (2001)Google Scholar
  23. 23.
    Ondrejcek, M., Swiech, W., Rajappan, M., Flynn, C.P.: Ripples formed in the sputter erosion of Pd(111). J. Phys. Condens. Matter 15, L735–L742 (2003)Google Scholar
  24. 24.
    Swiech, W., Rajappan, M., Ondrejcek, M., Sammann, E., Burdin, S., Petrov, I., Flynn, C.P.: Real-time imaging of surface evolution driven by variable-energy ion irradiation. Ultramicroscopy 108, 646–655 (2008)Google Scholar
  25. 25.
    Burton, W.K., Cabrera, N., Frank, F.C.: The growth of crystals and the equilibrium structure of their surfaces. Phil. Trans. Royal Soc. London Ser A Math. Phys. Sci. 243, 299–358 (1951)Google Scholar
  26. 26.
    Jeong, H.-C., Williams, E.D.: Steps on surfaces: experiment and theory. Surf. Sci. Rep. 34, 171–294 (1999)Google Scholar
  27. 27.
    Giesen, M.: Step and island dynamics at solid/vacuum and solid/liquid interfaces. Prog. Surf. Sci. 68, 1–153 (2001)Google Scholar
  28. 28.
    Flynn, C.P.: Step-edge fluctuations on crystal surfaces. Phys. Rev. B 66, 155405, 13 pages (2002)Google Scholar
  29. 29.
    Misbah, C., Pierre-Louis, O., Saito, Y.: Crystal surfaces in and out of equilibrium: a modern view. Rev. Mod. Phys. 82, 981–1040 (2010)Google Scholar
  30. 30.
    Bartelt, N.C., Goldberg, J.L., Einstein, T.L., Williams, E.D., Heyraud, J.C., Metois, J.J.: Brownian-motion of steps on Si(111). Phys. Rev. B 48, 15453–15456 (1993)Google Scholar
  31. 31.
    Ondrejcek, M., Swiech, W., Yang, G., Flynn, C.P.: Low energy electron microscopy studies of steps on single crystal thin films of refractory metals. J. Vac. Sci. Technol. B 20, 2473–2477 (2002)Google Scholar
  32. 32.
    Ondrejcek, M., Swiech, W., Durfee, C.S., Flynn, C.P.: Step fluctuations and step interactions on Mo(011). Surf. Sci. 541, 31–45 (2003)Google Scholar
  33. 33.
    Flynn, C.P., Ondrejcek, M., Swiech, W.: Capillary waves and thermodynamics of multisteps on Pt(111). Chem. Phys. Lett. 378, 161–166 (2003)Google Scholar
  34. 34.
    Ondrejcek, M., Swiech, W., Yang, G., Flynn, C.P.: Crossover from bulk to surface diffusion in the fluctuations of step edges on Pt(111). Phil. Mag. Lett. 84, 69–77 (2004)Google Scholar
  35. 35.
    Ondrejcek, M., Swiech, W., Flynn, C.P.: Studies of step stiffnesses and relaxation on Pt(111), Pd(111) and Mo(011). Surf. Sci. 566, 160–164 (2004)Google Scholar
  36. 36.
    Ondrejcek, M., Rajappan, M., Swiech, W., Flynn, C.P.: Step fluctuation spectroscopy of Au(111) by LEEM. Surf. Sci. 574, 111–122 (2005)Google Scholar
  37. 37.
    Ondrejcek, M., Swiech, W., Rajappan, M., Flynn, C.P.: Fluctuation spectroscopy of step edges on Pt(111) and Pd(111). Phys. Rev. B 72, 085422, 12 pages (2005)Google Scholar
  38. 38.
    Ondrejcek, M., Rajappan, M., Swiech, W., Flynn, C.P.: Low energy electron microscopy investigations of kinetics and energetics on clean close-packed metal surfaces. J. Phys. Condens. Matter 17, S1397–S1406 (2005)Google Scholar
  39. 39.
    Ondrejcek, M., Swiech, W., Flynn, C.P.: From step fluctuations to nanostructure energetics on Nb(011). Phil. Mag. 86, 1387–1400 (2006)Google Scholar
  40. 40.
    Ondrejcek, M., Rajappan, M., Swiech, W., Flynn, C.P.: Step fluctuation studies of surface diffusion and step stiffness for the Ni(111) surface. Phys. Rev. B 73, 035418, 8 pages (2006)Google Scholar
  41. 41.
    Ondrejcek, M., Swiech, W., Flynn, C.P.: Surface mass diffusion and step stiffness on an anisotropic surface: Mo(011). Surf. Sci. 600, 4673–4678 (2006)Google Scholar
  42. 42.
    Ondrejcek, M., Rajappan, M., Swiech, W., Flynn, C.P.: Surface mass diffusion and step stiffness on V(011). J. Appl. Phys. 100, 083523, 6 pages (2006)Google Scholar
  43. 43.
    Rajappan, M., Swiech, W., Ondrejcek, M., Flynn, C.P.: Surface mass diffusion over an extended temperature range on Pt(111). J. Phys. Condens. Matter 19, 226006, 8 pages (2007)Google Scholar
  44. 44.
    Flynn, C.P.: Point defect reactions at surfaces and in bulk metals. Phys. Rev. B 71, 085422, 16 pages (2005)Google Scholar
  45. 45.
    Flynn, C.P.: Coupling among bulk, surface and edge diffusion. Phys. Rev. B 73, 155417, 12 pages (2006)Google Scholar
  46. 46.
    Flynn, C.P.: Exact linear response of reacting thermal defects driven by creation processes. Phys. Rev. B 75, 134106, 11 pages (2007)Google Scholar
  47. 47.
    McCarty, K.F., Nobel, J.A., Bartelt, N.C.: Vaccine in solids and the stability of surface morphology. Nature 412, 622–625 (2001)Google Scholar
  48. 48.
    Poelsema, B., Hannon, J.B., Bartelt, N.C., Kellogg, G.L.: Bulk-surface vacancy exchange on Pt(111). Appl. Phys. Lett. 84, 2551–2553 (2004)Google Scholar
  49. 49.
    Kellogg, G.L., Bartelt, N.C.: Surface-diffusion-limited island decay on Rh(001). Surf. Sci. 577, 151–157 (2005)Google Scholar
  50. 50.
    Gerlach, R., Maroutian, T., Douillard, L., Martinotti, D., Ernst, H.J.: A novel method to determine the Ehrlich-Schwoebel barrier. Surf. Sci. 480, 97–102 (2001)Google Scholar
  51. 51.
    Tang, S.J., Kodambaka, S., Swiech, W., Petrov, I., Flynn, C.P., Chiang, T.C.: Sublimation of Atomic Layers from a Chromium Surface. Phys. Rev. Lett. 96, 126106, 4 pages (2006)Google Scholar
  52. 52.
    Rajappan, M., Swiech, W., Ondrejcek, M., Flynn, C.P.: Surface sink action during irradiation for Pt on Pt(111) by LEEM. Phil. Mag. 87, 4501–4518 (2007)Google Scholar
  53. 53.
    Flynn, C.P., Swiech, W., Ondrejcek, M., Rajappan, M.: Surface islands nucleated by a beam of energetic self-ions. Phys. Rev. B 77, 045406, 7 pages (2008)Google Scholar
  54. 54.
    Flynn, C.P., Ondrejcek, M., Swiech, W.: Synthesis of pans and mesas using a beam of self-ions. J. Phys. Condens. Matter 20, 395001, 8 pages (2008)Google Scholar
  55. 55.
    Flynn, C.P., Swiech, W., Ondrejcek, M.: Universal growth of islands driven by ion beams: theory and measurements on Pt(111). Phys. Rev. B 78, 075420, 9 pages (2008)Google Scholar
  56. 56.
    Ondrejcek, M., Swiech, W., Petrov, I., Rajappan, M., Flynn, C.P.: LEEM investigations of surfaces using a beam of energetic self-ions. Microsc. Res. Tech. 72, 197–207 (2009)Google Scholar
  57. 57.
    Swiech, W., Ondrejcek, M., Flynn, C.P.: LEEM investigations of ion beam effects on clean metal surfaces: quantitative studies of the driven steady state. J. Phys. Condens. Matter 21, 314021, 10 pages (2009)Google Scholar
  58. 58.
    Sato, Y., Ünal, B., Lograsso, T.A., Thiel, P.A., Schmid, A.K., Duden, T., Bartelt, N.C., McCarty, K.F.: Periodic step arrays on the aperiodic i-Al-Pd-Mn quasicrystal surface at high temperature. Phys. Rev. B 81, 161406, 4 pages (2010)Google Scholar
  59. 59.
    Bennett, P.A., Webb, M.W.: The Si(111) 7x7 to “1x1” transition. Surf. Sci. 104, 74–104 (1981)Google Scholar
  60. 60.
    Osakabe, N., Tanishiro, Y., Yagi, K., Honjo, G.: Direct observation of the phase transition between the (7x7) and (1x1) structures of clean (111) silicon surfaces. Surf. Sci. 109, 353–366 (1981)Google Scholar
  61. 61.
    Tanishiro, Y., Takayanagi, K., Yagi, K.: On the phase transition between the (7x7) and (1x1) structures of the silicon (111) surface studied by reflection electron microscopy. Ultramicroscopy 11, 95–102 (1983)Google Scholar
  62. 62.
    Telieps, W., Bauer, E.: Kinetics of the (7x7) (1x1) transition on Si(111). Ber. Bunsenges. Phys. Chem. 90, 197–200 (1986)Google Scholar
  63. 63.
    Telieps, W., Bauer, E.: The (7x7)↔(1x1) phase transition on Si(111). Surf. Sci. 162, 163–168 (1985)Google Scholar
  64. 64.
    Bauer, E., Telieps, W.: Low energy electron microscopy. In: Scanning Microscopy, pp. 99–108. Scanning Microscopy Int., Chicago (1987)Google Scholar
  65. 65.
    Telieps, W.: Surface imaging with LEEM. Appl. Phys. A 44, 55–61 (1987)Google Scholar
  66. 66.
    Bauer, E., Mundschau, M., Swiech, W., Telieps, W.: Low energy electron microscopy (LEEM) and photoemission microscopy (PEEM) of semiconductor surfaces. In: Cherns, D. (ed.) NATO Advanced research workshop on the evaluation of advanced semiconductor materials by electron microscopy, 1988. NATO ASI series B, vol. 203, pp. 283–294. Plenum Publ. Corp. Bristol, England (1989)Google Scholar
  67. 67.
    Bauer, E., Mundschau, M., Swiech, W., Telieps, W.: Low energy electron microscopy of semiconductor surfaces. J. Vac. Sci. Technol. A 9, 1007–1013 (1991)Google Scholar
  68. 68.
    Mundschau, M., Bauer, E., Telieps, W., Swiech, W.: Atomic step and defect structure on surfaces of Si(100) and Si(111) observed by low energy electron microscopy. Phil. Mag. A 61, 257–280 (1990)Google Scholar
  69. 69.
    Sakai, Y., Kato, M., Masuda, S., Harada, Y., Ichinokawa, T.: Dynamical change of step and domain-structures on the Si(111) surface with heat-treatment observed by a low energy electron microskope. Surf. Sci. 336, 295–302 (1995)Google Scholar
  70. 70.
    Telieps, W., Bauer, E.: Low energy electron microscopy (LEEM). In: van der Veen, J.F., Hove, M.A. (eds.) Structure of surfaces. Springer series in surface science, vol. 11, pp. 53–57. Springer, Berlin (1988)Google Scholar
  71. 71.
    Bauer, E., Mundschau, M., Swiech, W.: Low energy electron microscopy of nanometer scale phenomena. J. Vac. Sci. Technol. B 9, 403–406 (1991)Google Scholar
  72. 72.
    Phaneuf, R.J., Bartelt, N.C., Williams, E.D., Swiech, W., Bauer, E.: LEEM Investigations of the domain growth of the (7x7) reconstruction on Si(111). Surf. Sci. 268, 227–237 (1992)Google Scholar
  73. 73.
    Phaneuf, R.J., Swiech, W., Bartelt, N.C., Williams, E.D., Bauer, E.: LEEM investigations of orientational phase separation on vicinal Si(111) surfaces. Phys. Rev. Lett. 67, 2986–2989 (1991)Google Scholar
  74. 74.
    Williams, E.D., Phaneuf, R.J., Bartelt, N.C., Swiech, W., Bauer, E.: The role of surface stress in the facetting of stepped Si(111) surfaces. MRS Symp. Proc. 1992, 219–227 (1992)Google Scholar
  75. 75.
    Phaneuf, R.J., Bartelt, N.C., Williams, E.D., Swiech, W., Bauer, E.: Crossover from metastable to unstable facet growth on Si(111). Phys. Rev. Lett. 71, 2284–2287 (1993)Google Scholar
  76. 76.
    Hannon, J.B., Hibino, H., Bartelt, N.C., Swartzentruber, B.S., Ogino, T., Kellogg, G.L.: Dynamics of the silicon (111) surface phase transition. Nature 405, 552–554 (2000)Google Scholar
  77. 77.
    Hannon, J.B., Meyer zu Heringdorf, F.-J., Tersoff, J., Tromp, R.M.: Phase coexistence during surface phase transitions. Phys. Rev. Lett. 86, 4871–4874 (2001)Google Scholar
  78. 78.
    Hannon, J.B., Tromp, R.M.: Phase boundary fluctuations on Si(111). J. Vac. Sci. Technol. A 19, 2596–2600 (2001)Google Scholar
  79. 79.
    Hannon, J.B., Tersoff, J., Reuter, M.C., Tromp, R.M.: Influence of supersaturation on surface structure. Phys. Rev. Lett. 89, 266103, 4 pages (2002)Google Scholar
  80. 80.
    Hannon, J.B., Tersoff, J., Tromp, R.M.: The stability of triangular ‘droplet’ phases on Si(111). J. Cryst. Growth 237–239, 181–187 (2002)Google Scholar
  81. 81.
    Hannon, J.B., Tersoff, J., Tromp, R.M.: Surface stress and thermodynamic nanoscale size selection. Science 295, 299–301 (2002)Google Scholar
  82. 82.
    Thayer, G.E., Hannon, J.B., Tromp, R.M.: Shape and stability of self-assembled surface domains. Nat Mater. 3, 95–98 (2004)Google Scholar
  83. 83.
    Medhekar, N.V., Shenoy, V.B., Hannon, J.B., Tromp, R.M.: Metastability in 2D self-assembling systems. Phys. Rev. Lett. 99, 156102, 4 pages (2007)Google Scholar
  84. 84.
    Medhekar, N.V., Shenoy, V.B., Hannon, J.B., Tromp, R.M.: Self-assembling surface stress domains far from equilibrium. Appl. Phys. Lett. 91, 253101, 4 pages (2007)Google Scholar
  85. 85.
    Hu, C.W., Hibino, H., Ogino, T., Tsong, I.S.T.: Hysteresis in the (1 x 1)-(7 x 7) first-order phase transition on the Si(111) surface. Surf. Sci. 487, 191–200 (2001)Google Scholar
  86. 86.
    Hibino, H., Homma, Y., Hu, C.W., Uwaha, M., Ogino, T., Tsong, I.S.T.: Structural and morphological changes on surfaces with multiple phases studied by low-energy electron microscopy. Appl. Surf. Sci. 237, 51–57 (2004)Google Scholar
  87. 87.
    Tromp, R.M., Hannon, J.B.: Thermodynamics of nucleation and growth. Surf. Rev. Lett. 9, 1565–1593 (2002)Google Scholar
  88. 88.
    Hannon, J.B., Tromp, R.M.: Low-energy electron microscopy of surface phase transitions. Annu. Rev. Mater. Res. 33, 263–288 (2003)Google Scholar
  89. 89.
    Xu, Y.-C., Liu, B.-G.: Two-speed phase dynamics in the Si(111) (7x7) - (1x1) phase transition. Phys. Rev. Lett. 100, 056103, 4 pages (2008)Google Scholar
  90. 90.
    Altman, M.S.: The role of diffusion on step flow growth. In: Tringides, M.C. (ed.) Proc. NATO ASI conference on surface diffusion: atomistic and collective processes rhodes, Greece 1996. Plenum Press, NATO ASI series, vol. B360, pp. 227–233, New York (1997)Google Scholar
  91. 91.
    Altman, M.S., Chung, W.F., Franz, T.: LEEM determination of critical terrace widths for Si/Si(111) step flow growth. Surf. Rev. Lett. 5, 27–30 (1998)Google Scholar
  92. 92.
    Chung, W.F., Altman, M.S.: Kinetic length, step permeability, and kinetic coefficient asymmetry on the Si(111) (7x7) surface. Phys. Rev. B 66, 075338, 5 pages (2002)Google Scholar
  93. 93.
    Chung, W.F., Bromann, K., Altman, M.S.: The transition to step flow growth on the clean and surfactant covered Si(111) surface studied by in-situ LEEM. Int. J. Mod. Phys. B 16, 4353–4362 (2002)Google Scholar
  94. 94.
    Man, K.L., Tang, W.X., Jin, X.F., Altman, M.S.: Kinetic regime of step motion on the Si(111) (1x1) surface. Surf. Interface Anal. 38, 1632–1635 (2006)Google Scholar
  95. 95.
    Man, K.L., Pang, A.B., Altman, M.S.: Kinetic length and step permeability on the Si(111) (1x1) surface. Surf. Sci. 601, 4669–4674 (2007)Google Scholar
  96. 96.
    Pang, A.B., Man, K.L., Altman, M.S., Stasevich, T.J., Szalma, F., Einstein, T.L.: Step line tension and step morphological evolution on the Si(111) (1x1) surface. Phys. Rev. B 77, 115424, 12 pages (2008)Google Scholar
  97. 97.
    Hibino, H., Hu, C.W., Ogino, T., Tsong, I.S.T.: Decay kinetics of two-dimensional islands and holes on Si(111) studied by low-energy electron microscopy. Phys. Rev. B 63, 245402, 8 pages (2001)Google Scholar
  98. 98.
    Hibino, H., Hu, C.-W., Ogino, T., Tsong, I.S.T.: Diffusion barrier caused by 1x1 and 7x7 on Si(111) during phase transition. Phys. Rev. B 64, 245401–245405 (2001)Google Scholar
  99. 99.
    Hibino, H., Watanabe, Y., Hu, C.-W., Tsong, I.S.T.: Thermal decay of superheated 7x7 islands and supercooled “1x1” vacancy islands on Si(111). Phys. Rev. B 72, 245424, 6 pages (2005)Google Scholar
  100. 100.
    Hannon, J.B., Shenoy, V.B., Schwarz, K.W.: Anomalous spiral motion of steps near dislocations on silicon surfaces. Science 313, 1266–1269 (2006)Google Scholar
  101. 101.
    Bauer, E., Telieps, W.: Emission and low energy reflection electron microscopy. In: Howie, A., Valdre, U. (eds.) Surface and Interface Characterization by Electron Optical Methods, vol. 191 NATO ASI Series B: Physics, pp. 195–233. Plenum Press, New York (1988)Google Scholar
  102. 102.
    Mundschau, M., Bauer, E., Telieps, W., Swiech, W.: Atomic steps on Si(100) and step dynamics during sublimation studied by low energy electron microscopy. Surf. Sci. 223, 413–423 (1989)Google Scholar
  103. 103.
    Bauer, E.: Low energy electron microscopy of surface processes. Appl. Surf. Sci. 60/61, 350–358 (1992)Google Scholar
  104. 104.
    Swiech, W., Bauer, E.: The growth of Si on Si(100): a video-LEEM study. Surf. Sci. 255, 219–228 (1991)Google Scholar
  105. 105.
    Bauer, E., Mundschau, M., Swiech, W., Telieps, W.: Low energy electron microscopy of surface processes. Vacuum 41, 5–10 (1990)Google Scholar
  106. 106.
    Bauer, E., Mundschau, M., Swiech, W., Telieps, W.: Low energy electron microscopy and surface processes on clean Si(111) and Si(100). In: Bringans, R.D., Feenstra, R.M., Gibson, J.M. (eds.) MRS symposium atomic scale structure of interfaces, vol. 159, pp. 225–233, Boston, Massachusetts 1989. Materials Research Society (1990)Google Scholar
  107. 107.
    Tromp, R.M., Dernier van der Gon, A.M., Reuter, M.C.: Surface stress as a driving force for interfacial mixing. Phys. Rev. Lett. 68, 2313–2316 (1992)Google Scholar
  108. 108.
    Tromp, R.M., Reuter, M.C.: Step morphologies on small-miscut Si(001) surfaces. Phys. Rev. B 47, 7598–7601 (1993)Google Scholar
  109. 109.
    Bartelt, N.C., Tromp, R.M., Williams, E.D.: Step capillary waves and equilibrium island shapes on Si(001). Phys. Rev. Lett. 73, 1656–1659 (1994)Google Scholar
  110. 110.
    Theis, W., Bartelt, N.C., Tromp, R.M.: Chemical potential maps and spatial correlations in 2D-island ripening on Si(001). Phys. Rev. Lett. 75, 3328–3331 (1995)Google Scholar
  111. 111.
    Theis, W., Tromp, R.M.: Nucleation in Si(001) homoepitaxial growth. Phys. Rev. Lett. 76, 2770–2773 (1996)Google Scholar
  112. 112.
    Tanaka, S., Umbach, C.C., Blakely, J.M., Tromp, R.M., Mankos, M.: Fabrication of arrays of large stepfree regions on Si(001). Appl. Phys. Lett. 69, 1235–1237 (1996)Google Scholar
  113. 113.
    Bartelt, N.C., Tromp, R.M.: Low-energy electron microscopy study of step mobilities on Si(001). Phys. Rev. B 54, 11731–11740 (1996)Google Scholar
  114. 114.
    Bartelt, N.C., Theis, W., Tromp, R.M.: Ostwald ripening of two-dimensional islands on Si(001). Phys. Rev. B 54, 11741–11751 (1996)Google Scholar
  115. 115.
    Tromp, R.M., Mankos, M.: Thermal adatoms on Si(001). Phys. Rev. Lett. 81, 1050–1053 (1998)Google Scholar
  116. 116.
    Nielsen, J.-F., Pelz, J.P., Hibino, H., Hu, C.-W., Tsong, I.S.T.: Enhanced terrace stability for preparation of step-free Si(001)-(2x1) surfaces. Phys. Rev. Lett. 87, 136103, 4 pages (2001)Google Scholar
  117. 117.
    Zielasek, V., Liu, F., Zhao, Y., Maxson, J.B., Lagally, M.G.: Surface stress-induced island shape transition in Si(001) homoepitaxy. Phys. Rev. B 64, 201320 (2001)Google Scholar
  118. 118.
    Tanaka, S., Umbach, C.C., Blakely, J.M., Tromp, R.M., Mankos, M.: Atomic step distributions on annealed periodic Si(001) gratings. J. Vac. Sci. Technol. A 15, 1345–1350 (1997)Google Scholar
  119. 119.
    Tanaka, S., Bartelt, N.C., Umbach, C.C., Tromp, R.M., Blakely, J.M.: Step permeability and the relaxation of biperiodic gratings on Si(001). Phys. Rev. Lett. 78, 3342–3345 (1997)Google Scholar
  120. 120.
    Blakely, J.M., Tanaka, S., Tromp, R.M.: Atomic step dynamics on periodic semiconductor surface structures. J. Electron. Microsc. 48, 747–752 (1999)Google Scholar
  121. 121.
    Jones, D.E., Pelz, J.P., Hong, Y., Bauer, E., Tsong, I.S.T.: Striped phase and temperature dependent step shape transition on highly B-doped Si(001)-(2x1) surfaces. Phys. Rev. Lett. 77, 330–333 (1996)Google Scholar
  122. 122.
    Bauer, E.: LEEM and SPLEEM. In: Hawkes, P., Spence, J. (eds.) Science of Microscopy, vol. I, chapter 8, pp. 606–656. Springer, New York (2007)Google Scholar
  123. 123.
    Hannon, J.B., Bartelt, N.C., Swartzentruber, B.S., Hamilton, J.C., Kellogg, G.L.: Step faceting at the (001) surface of boron doped silicon. Phys. Rev. Lett. 79, 4226–4229 (1997)Google Scholar
  124. 124.
    Hannon, J.B., Swartzentruber, B.S., Kellogg, G.L., Bartelt, N.C.: LEEM measurements of step energies at the (001) surface of heavily boron-doped silicon. Surf. Rev. Lett. 5, 1159–1165 (1998)Google Scholar
  125. 125.
    Nielsen, J.-F., Pelz, J.P., Hibino, H., Hu, C.-W., Tsong, I.S.T., Kouvetakis, J.: Controlled striped phase formation on ultraflat Si(001) surfaces during diborane exposure. Appl. Phys. Lett. 79, 3857–3859 (2001)Google Scholar
  126. 126.
    Ermanoski, I., Bartelt, N.C., Kellogg, G.L.: Self-assembly of defect-free nanostripe arrays on B-doped Si(001). Phys. Rev. B 83, 205432, 5 pages (2011)Google Scholar
  127. 127.
    Hibino, H., Watanabe, Y.: Growth of twinned epitaxial layers on Si(111)√3x√3-B studied by low-energy electron microscopy. Jpn. J. Appl. Phys. 44, 358–364 (2005)Google Scholar
  128. 128.
    Hibino, H., Uematsu, M., Watanabe, Y.: Void growth during thermal decomposition of silicon oxide layers studied by low-energy electron microscopy. J. Appl. Phys. 100, 113519, 8 pages (2006)Google Scholar
  129. 129.
    Yamamoto, S., Masuda, S., Yasufuku, H., Ueno, N., Harada, Y., Ichinokawa, T., Kato, M., Sakai, Y.: Study of solid surfaces by metastable electron emission microscopy: energy-filtered images and local electron spectra at the outermost surface layer of silicon oxide on Si(100). J. Appl. Phys. 82, 2954–2960 (1997)Google Scholar
  130. 130.
    Yasufuku, H., Okumura, M., Ibe, T., Okudaira, K., Harada, Y., Ueno, N.: Surface images of SiO2/Si(100) pattern using electron emission microscopy with metastable atoms, photons and low-energy electrons. Jpn. J. Appl. Phys. 40, 2447–2450 (2001)Google Scholar
  131. 131.
    Fukidome, H., Tanaka, K., Yoshimura, M., Ueda, K., Guo, F.-Z., Kinoshita, T., Kobayashi, K.: Microscopically controlled oxidation of H/Si(100) by lateral surface electric field studied by emission electron microscopies. Surf. Sci. 601, 4675–4679 (2007)Google Scholar
  132. 132.
    Heun, S., Kremmer, S., Ercolani, D., Wurmbauer, H., Teichert, C.: Behavior of SiO2 nanostructures under intense extreme ultraviolet illumination. J. Appl. Phys. 97, 104333, 4 pages (2005)Google Scholar
  133. 133.
    de la Peña, F., Barrett, N., Zagonel, L.F., Walls, M., Renault, O.: Full field chemical imaging of buried native sub-oxide layers on doped silicon patterns. Surf. Sci. 604, 1628–1636 (2010)Google Scholar
  134. 134.
    Wurm, K., Kliese, R., Hong, Y., Rottger, B., Wei, Y., Neddermeyer, H., Tsong, I.S.T.: Evolution of surface morphology of Si(100)-(2x1) during oxygen adsorption at elevated temperatures. Phys. Rev. B 50, 1567–1574 (1994)Google Scholar
  135. 135.
    Hannon, J.B., Bartelt, M.C., Bartelt, N.C., Kellogg, G.L.: Etching of the Si(001) surface with molecular oxygen. Phys. Rev. Lett. 81, 4676–4679 (1998)Google Scholar
  136. 136.
    Sutter, P., Lagally, M.G.: Quantitative determination of dislocation-induced strain at the surface of (001) silicon-on-insulator. Phys. Rev. Lett. 82, 1490–1493 (1999)Google Scholar
  137. 137.
    Euaruksakul, C., Li, Z., Savage, D.E., Lagally, M.G.: X-ray absorption spectroscopy of strained-Si nanomembranes. ECS Trans 6, 257–262 (2007)Google Scholar
  138. 138.
    Seo, S., Euaruksakul, C., Savage, D.E., Lagally, M.G., Evans, P.G.: Nanostructure formation in the initial roughening of a thin silicon sheet. Phys. Rev. B 81, 041302, 4 pages (2010)Google Scholar
  139. 139.
    Flack, F.S., Yang, B., Huang, M., Marcus, M., Simmons, J., Castellini, O.M., Eriksson, M.A., Liu, F., Lagally, M.G.: Pattern formation on silicon-on-insulator. In: Chason, E., Gilmer, G.H., Huang, H., Wang, E. (eds.) Proc. MRS symposium KK—kinetics-driven nanopatterning on surfaces, vol. 849, pp. KK1.3.1–KK1.3.9, Boston, Massachusetts, USA 2004. Materials Research Society (2005)Google Scholar
  140. 140.
    Yang, B., Zhang, P., Savage, D.E., Lagally, M.G., Lu, G.-H., Huang, M., Liu, F.: Self-organization of semiconductor nanocrystals by selective surface faceting. Phys. Rev. B 72, 235413, 4 pages (2005)Google Scholar
  141. 141.
    Bussmann, E., Cheynis, F., Leroy, F., Müller, P.: Thermal instability of silicon-on-insulator thin films measured by low-energy electron microscopy. In: Henrion, G., Belmahi, M., Andrieu, S. (eds.) Proc. innovations in thin film processing and characterisation (ITFPC 2009), Nancy, France 2009. IOP Conf. Series: Materials Science and Engineering, vol. 12, pp. 1–4. Institute of Physics, Bristol, England (2010)Google Scholar
  142. 142.
    Bussmann, E., Cheynis, F., Leroy, F., Müller, P., Pierre-Louis, O.: Dynamics of solid thin-film dewetting in the silicon-on-insulator system. New J. Phys. 13, 043017, 9 pages (2011)Google Scholar
  143. 143.
    Cheynis, F., Bussmann, E., Leroy, F., Passanante, T., Müller, P.: Dewetting dynamics of silicon-on-insulator thin films. Phys. Rev. B 84, 245439, 9 pages (2011)Google Scholar
  144. 144.
    Leroy, F., Cheynis, F., Passanante, T., Müller, P.: Dynamics, anisotropy, and stability of silicon-on-insulator dewetting fronts. Phys. Rev. B 85, 195414, 6 pages (2012)Google Scholar
  145. 145.
    Swiech, W., Schwarz-Selinger, T., Cahill, D.G.: Phase coexistence and morphology at the Si(110) surface phase transition. Surf. Sci. 519, L599–L603 (2002)Google Scholar
  146. 146.
    Watanabe, F., Kodambaka, S., Swiech, W., Greene, J.E., Cahill, D.G.: LEEM study of island decay on Si(110). Surf. Sci. 572, 425–432 (2004)Google Scholar
  147. 147.
    Tromp, R.M., Theis, W., Bartelt, N.C.: Real-time microscopy of two-dimensional critical fluctuations: disordering of the Si(113)-(3x1) reconstruction. Phys. Rev. Lett. 77, 2522–2525 (1996)Google Scholar
  148. 148.
    van Vroonhoven, E., Zandvliet, H.J.W., Poelsema, B.: (2x1)-(1x1) phase transition on Ge(001): dimer breakup and surface roughening. Phys. Rev. Lett. 91, 116102, 4 pages (2003)Google Scholar
  149. 149.
    van Vroonhoven, E., Zandvliet, H.J.W., Poelsema, B.: A quantitative evaluation of the dimer concentration during the (2x1)-(1x1) phase transition on Ge(001). Surf. Sci. 574, L23–L28 (2005)Google Scholar
  150. 150.
    Xu, Y.-C., Liu, B.-G.: Phase-field-crystal modelling of the (2x1)-(1x1) phase transitions of Si(001) and Ge(001) surfaces. J. Phys. D Appl. Phys. 42, 035402, 5 pages (2009)Google Scholar
  151. 151.
    Tersoff, J., Jesson, D.E., Tang, W.X.: Decomposition controlled by surface morphology during langmuir evaporation of GaAs. Phys. Rev. Lett. 105, 035702, 4 pages (2010)Google Scholar
  152. 152.
    Zhou, Z.Y., Zheng, C.X., Tang, W.X., Jesson, D.E., Tersoff, J.: Congruent evaporation temperature of GaAs(001) controlled by As flux. Appl. Phys. Lett. 97, 121912, 3 pages (2010)Google Scholar
  153. 153.
    Pavlovska, A., Torres, V.M., Edwards, J.L., Bauer, E., Smith, D.J., Doak, R.B., Tsong, I.S.T., Thomson, D.B., Davis, R.F.: Homoepitaxial GaN layers studied by low-energy electron microscopy, atomic force microscopy and transmission electron microscopy. Phys. Status Solidi A 176, 469–473 (1999)Google Scholar
  154. 154.
    Maxson, J.B., Perkins, N., Savage, D.E., Woll, A.R., Zhang, L., Kuech, T.F., Lagally, M.G.: Novel dark-field imaging of GaN {0001} surfaces with low-energy electron microscopy. Surf. Sci. 464, 217–222 (2000)Google Scholar
  155. 155.
    Pavlovska, A., Bauer, E., Smith, D.J.: In situ studies of the role of excess Ga on the growth morphology of thin GaN layers. Surf. Sci. 496, 160–178 (2002)Google Scholar
  156. 156.
    Hartman, J.D., Naniwae, K., Petrich, C., Nemanich, R.J., Davis, R.F.: Photo-electron emission and atomic force microscopies of the hydrogen etched 6H-SiC(0001) surface and the initial growth of GaN and AlN. Appl. Surf. Sci. 242, 428–436 (2005)Google Scholar
  157. 157.
    Hannon, J.B., Tromp, R.M.: Pit formation during graphene synthesis on SiC(0001): in situ electron microscopy. Phys. Rev. B 77, 241404, 4 pages (2008)Google Scholar
  158. 158.
    Hannon, J.B., Tromp, R.M., Medhekar, N.V., Shenoy, V.B.: Spontaneous formation and growth of a new polytype on SiC(0001). Phys. Rev. Lett. 103, 256101, 4 pages (2009)Google Scholar
  159. 159.
    Sun, J., Hannon, J.B., Tromp, R.M., Pohl, K.: Highly uniform step and terrace structures on SiC(0001) surfaces. IBM J. Res. Dev. 55, 7, 6 pages (2011)Google Scholar
  160. 160.
    Hasegawa, M., Ohno, T.: Non-destructive observation of in-grown stacking faults in 4H-SiC epitaxial layer using mirror electron microscope. J. Appl. Phys. 110, 073507, 6 pages (2011)Google Scholar
  161. 161.
    Kodambaka, S., Israeli, N., Bareno, J., Swiech, W., Ohmori, K., Petrov, I., Greene, J.E.: Low-energy electron microscopy studies of interlayer mass transport kinetics on TiN(111). Surf. Sci. 560, 53–62 (2004)Google Scholar
  162. 162.
    Bareno, J., Kodambaka, S., Khare, S.V., Swiech, W., Petrov, I., Greene, J.E.: Orientation-dependent mobilities from analyses of two-dimensional TiN(111) island decay kinetics. Thin Solid Films 510, 339–345 (2006)Google Scholar
  163. 163.
    Kodambaka, S., Khare, S.V., Swiech, W., Ohmori, K., Petrov, I., Greene, J.E.: Dislocation-driven surface dynamics on solids. Nature 429, 49–52 (2004)Google Scholar
  164. 164.
    Kodambaka, S., Bareño, J., Khare, S.V., Świech, W., Petrov, I., Greene, J.E.: Nucleation and growth kinetics of spiral steps on TiN(111): an in situ low-energy electron microscopy study. J. Appl. Phys. 98, 034901, 5 pages (2005)Google Scholar
  165. 165.
    McCarty, K.F., Bartelt, N.C.: Role of bulk thermal defects in the reconstruction dynamics of the TiO2 (110) surface. Phys. Rev. Lett. 90, 046104, 4 pages (2003)Google Scholar
  166. 166.
    McCarty, K.F., Bartelt, N.C.: The 1x1/1x2 phase transition of the TiO2(110) surface–variation of transition temperature with crystal composition. Surf. Sci. 527, L203–L212 (2003)Google Scholar
  167. 167.
    McCarty, K.F., Bartelt, N.C.: Spatially resolved dynamics of the TiO2(110) surface reconstruction. Surf. Sci. 540, 157–171 (2003)Google Scholar
  168. 168.
    McCarty, K.F.: Growth regimes of the oxygen-deficient TiO2(110) surface exposed to oxygen. Surf. Sci. 543, 185–206 (2003)Google Scholar
  169. 169.
    McCarty, K.F., Bartelt, N.C.: Crystal growth rate limited by step length - the case of oxygen-deficient TiO2 exposed to oxygen. J. Cryst. Growth 270, 691–698 (2004)Google Scholar
  170. 170.
    Menteş, T.O., Locatelli, A., Aballe, L., Pavlovska, A., Bauer, E., Pabisiak, T., Kiejna, A.: Surface modification of oxides by electron-stimulated desorption for growth-mode control of metal films: experiment and density-functional calculations. Phys. Rev. B 76, 155413, 14 pages (2007)Google Scholar
  171. 171.
    Locatelli, A., Pabisiak, T., Pavlovska, A., Menteş, T.O., Aballe, L., Kiejna, A., Bauer, E.: One-dimensional Au on TiO2. J. Phys.: Condens. Matter 19, 082202, 8 pages (2007)Google Scholar
  172. 172.
    Xiong, G., Joly, A.G., Beck, K.M., Hess, W.P.: Laser-induced oxygen vacancy formation and diffusion on TiO2(110) surfaces probed by photoemission electron microscopy. Phys. Status Solidi C 3, 3598–3602 (2006)Google Scholar
  173. 173.
    Bethge, H., Klaua, M.: Photo-emission electron microscopy of work function changes. Ultramicroscopy 11, 207–214 (1983)Google Scholar
  174. 174.
    Przychowski, M.D., Marx, G.K.L., Fecher, G.H., Schönhense, G.: A spatially resolved investigation of oxygen adsorption on polycrystalline copper and titanium by means of photoemission electron microscopy. Surf. Sci. 549, 37–51 (2004)Google Scholar
  175. 175.
    Huang, W., Zhai, R., Bao, X.: Investigation of oxygen adsorption on Pd (100) with defects. Appl. Surf. Sci. 158, 287–291 (2000)Google Scholar
  176. 176.
    Ermanoski, I., Swiech, W., Madey, T.E.: Nucleation and coexistence of nanometer-scale facet domains on O/Ir(210). Surf. Sci. Lett. 592, L299–L304 (2005)Google Scholar
  177. 177.
    Reichelt, R., Günther, S., Wintterlin, J., Moritz, W., Aballe, L., Mentes, T.O.: Low energy electron diffraction and low energy electron microscopy microspot I/V analysis of the (4x4)O structure on Ag (111): surface oxide or reconstruction? J. Chem. Phys. 127, 134706, 8 pages (2007)Google Scholar
  178. 178.
    Menteş, T.O., Stojić, N., Binggeli, N., Niño, M.A., Locatelli, A., Aballe, L., Kiskinova, M., Bauer, E.: Strain relaxation in small adsorbate islands: O on W(110). Phys. Rev. B 77, 155414, 9 pages (2008)Google Scholar
  179. 179.
    Stojić, N., Menteş, T.O., Binggeli, N., Niño, M.A., Locatelli, A., Bauer, E.: Temperature dependence of surface stress across an order–disorder transition: p(1x2)O/W(110). Phys. Rev. B 81, 115437, 7 pages (2010)Google Scholar
  180. 180.
    Altman, M.S., Bauer, E.: The reaction of oxygen with the hot W(001) surface studied by low-energy electron microscopy. J. Vac. Sci. Technol. A 9, 659–660 (1991)Google Scholar
  181. 181.
    Altman, M.S., Bauer, E.: LEEM/LEED investigation of reconstruction and initial oxidation of the W(001). Surf. Sci. 347, 265–279 (1996)Google Scholar
  182. 182.
    Flege, J.I., Hrbek, J., Sutter, P.: Structural imaging of surface oxidation and oxidation catalysis on Ru(0001). Phys. Rev. B 78, 165407, 5 pages (2008)Google Scholar
  183. 183.
    Ondrejcek, M., Appleton, R.S., Swiech, W., Petrova, V.L., Flynn, C.P.: Thermally Activated Stripe Reconstruction Induced by O on Nb (011). Phys. Rev. Lett. 87, 116102, 4 pages (2001)Google Scholar
  184. 184.
    Rotermund, H.H., Nettesheim, S., von Oertzen, A., Ertl, G.: Observation of surface diffusion of adsorbates on metal surfaces by photoemission electron microscopy (PEEM). Surf. Sci. Lett. 275, L645–L649 (1992)Google Scholar
  185. 185.
    Rotermund, H.H.: Investigation of dynamic processes in adsorbed layers by photoemission electron microscopy (PEEM). Surf. Sci. 283, 87–100 (1993)Google Scholar
  186. 186.
    von Oertzen, A., Rotermund, H.H., Nettesheim, S.: Investigation of diffusion of CO adsorbed on Pd(111) by a combined PEEM/LITD technique. Chem. Phys. Lett. 199, 131–137 (1992)Google Scholar
  187. 187.
    Yim, C.M., Man, K.L., Xiao, X., Altman, M.S.: Low-energy electron microscopy of CO/Pt(111) surface diffusion by nonequilibrium coverage profile evolution. Phys. Rev. B 78, 155439, 9 pages (2008)Google Scholar
  188. 188.
    Cerdá, J.I., Santos, B., Herranz, T., Puerta, J.M., de la Figuera, J., McCarty, K.F.: CO-assisted subsurface hydrogen trapping in Pd(111) films. J. Phys. Chem. Lett. 3, 87–91 (2012)Google Scholar
  189. 189.
    Ling, W.L., Bartelt, N.C., Pohl, K., de la Figuera, J., Hwang, R.Q., McCarty, K.F.: Enhanced self-diffusion on Cu(111) by trace amounts of S: chemical-reaction-limited kinetics. Phys. Rev. Lett. 93, 166101, 4 pages (2004)Google Scholar
  190. 190.
    Przychowski, M.D., Wiechert, H., Marx, G.K.L., Schönhense, G.: Real-space observation of xenon adsorption and desorption kinetics on graphite (0001) by photoemission electron microscopy. Surf. Sci. 541, 46–58 (2003)Google Scholar
  191. 191.
    Rotermund, H.H., Ertl, G., Sesselmann, W.: Scanning photoemission microscopy of surfaces. Surf. Sci. 217, L383–L390 (1989)Google Scholar
  192. 192.
    Lilienkamp, G., Wei, H., Maus-Friedrichs, W., Kempter, V., Marbach, H., Günther, S., Suchorski, Y.: Metastable impact electron emission microscopy of the catalytic H2 oxidation on Rh(110). Surf. Sci. 532–535, 132–136 (2003)Google Scholar
  193. 193.
    Lmbihl, R., Ertl, G.: Oscillatory kinetics in heterogeneous catalysis. Chem. Rev. 95, 697–733 (1995)Google Scholar
  194. 194.
    Rotermund, H.H.: Imaging of dynamic processes on surfaces by light. Surf. Sci. Rep. 29, 265–364 (1997)Google Scholar
  195. 195.
    Imbihl, R.: Non-linear dynamics in catalytic reactions. In: Hasselbrink, E., Lundquist, B. (eds.) Handbook of Surface Science, vol. 3, pp. 341–428. Elsevier, Amsterdam (2008)Google Scholar
  196. 196.
    Imbihl, R.: Nonlinear dynamics on catalytic surfaces: the contribution of surface science. Surf. Sci. 603, 1671–1679 (2009)Google Scholar
  197. 197.
    Rotermund, H.H.: Real time imaging of catalytic reactions on surfaces: past, present and future. Surf. Sci. 603, 1662–1670 (2009)Google Scholar
  198. 198.
    Nettesheim, S., von Oertzen, A., Rotermund, H.H., Ertl, G.: Reaction diffusion patterns in the catalytic CO-oxidation on Pt(110): front propagation and spiral waves. J. Chem. Phys. 98, 9977–9985 (1993)Google Scholar
  199. 199.
    Sadeghi, P., Dunphy, K., Punckt, C., Rotermund, H.H.: Inversion of pattern anisotropy during CO oxidation on Pt(110) correlated with appearance of subsurface oxygen. J. Phys. Chem. C 116, 4686–4691 (2012)Google Scholar
  200. 200.
    Wei, H., Lilienkamp, G., Imbihl, R.: Surface topographical changes and chemical wave patterns in catalytic CO oxidation on Pt(110). Chem. Phys. Lett. 389, 284–288 (2004)Google Scholar
  201. 201.
    Schmidt, T., Schaak, A., Guenther, S.G., Ressel, B., Bauer, E., Imbihl, R.: In situ imaging of structural changes in a chemical wave with low energy electron microscopy: the system Rh(110)/NO + H2. Chem. Phys. Lett. 318, 549–554 (2000)Google Scholar
  202. 202.
    Locatelli, A., Mentes, T.O., Aballe, L., Mikhailov, A., Kiskinova, M.: Formation of regular surface-supported mesostructures with periodicity controlled by chemical reaction rate. J. Phys. Chem. B 110, 19108–19111 (2006)Google Scholar
  203. 203.
    Locatelli, A., Bauer, E.: Recent advances in chemical and magnetic imaging of surfaces and interfaces by XPEEM. J. Phys. Condens. Matter 20, 82202–82224 (2008) (and references therein)Google Scholar
  204. 204.
    Locatelli, A., Heun, S., Kiskinova, M.: Direct observation of reaction-induced lateral redistribution of sub-monolayers of Au deposited on a Rh(110) surface. Surf. Sci. 566–568, 1130–1136 (2004)Google Scholar
  205. 205.
    Locatelli, A., Sbraccia, C., Heun, S., Baroni, S., Kiskinova, M.: Energetically driven reorganization of a modified catalytic surface under reaction conditions. J. Am. Chem. Soc. 127, 2351–2357 (2005)Google Scholar
  206. 206.
    Locatelli, A., Aballe, L., Menteş, T.O., Kiskinova, M., Bauer, E.: Photoemission electron microscopy with chemical sensitivity: SPLEEM methods and applications. Surf. Interface Anal. 38, 1554–1557 (2006)Google Scholar
  207. 207.
    Locatelli, A., Aballe, L., Menteş, T.O., Guo, F.Z., Kiskinova, M.: A spectro-microscopic study of the reactive phase separation of Au + Pd and O on Rh(110). Surf. Sci. 601, 4663–4668 (2007)Google Scholar
  208. 208.
    Hong, L., Uecker, H., Hinz, M., Qiao, L., Kevrekidis, I.G., Günther, S., Mentes, T.O., Locatelli, A., Imbihl, R.: Mass transport of alkali metal with pulses in a surface reaction. Phys. Rev. E 78, 055203(R), 4 pages (2008)Google Scholar
  209. 209.
    Hong, L., Hesse, M., Imbihl, R.: Patterns in the NO + H2 reaction on Rh(110) modified by coadsorbed potassium. J. Phys. Chem. C 113, 4174–4183 (2009)Google Scholar
  210. 210.
    Lovis, F., Hesse, M., Imbihl, R.: Redistribution of supported vanadium oxide catalysts by pattern formation. Catal. Lett. 136, 171–176 (2010)Google Scholar
  211. 211.
    Lovis, F., Imbihl, R.: Self-organization of ultrathin vanadium oxide layers on a Rh(111) surface during a catalytic reaction. Part I: a PEEM study. J. Phys. Chem. C 115, 19141–19148 (2011)Google Scholar
  212. 212.
    Lovis, F., Hesse, M., Locatelli, A., Mentes, T.O., Nino, M.A., Lilienkamp, G., Borkenhagen, B., Imbihl, R.: Self-organization of ultrathin vanadium oxide layers on a Rh(111) surface during a catalytic reaction. Part II: a LEEM and spectromicroscopy study. J. Phys. Chem. C 115, 19149–19157 (2011)Google Scholar
  213. 213.
    Lovis, F., Smolinsky, T., Locatelli, A., Niño, M.Á., Imbihl, R.: Chemical waves and rate oscillations in the H2 + O2 reaction on a bimetallic Rh(111)/Ni catalyst. J. Phys. Chem. C 116, 4083–4090 (2012)Google Scholar
  214. 214.
    Imbihl, R.: Chemical selforganization of composite catalysts during catalytic reactions. J. Electron. Spectrosc. Rel. Phenom. 185, 347–355 (2012)Google Scholar
  215. 215.
    Suchorski, Y., Spiel, C., Vogel, D., Drachsel, W., Schlögl, R., Rupprechter, G.: Local reaction kinetics by imaging: CO oxidation on polycrystalline platinum. Chem. Phys. Chem. 11, 3231–3235 (2010)Google Scholar
  216. 216.
    Spiel, C., Vogel, D., Suchorski, Y., Drachsel, W., Schlögl, R., Rupprechter, G.: Catalytic CO oxidation on individual (110) domains of a polycrystalline Pt foil: local reaction kinetics by PEEM. Catal. Lett. 141, 625–632 (2011)Google Scholar
  217. 217.
    Vogel, D., Spiel, C., Suchorski, Y., Urich, A., Schlögl, R., Rupprechter, G.: Mapping the local reaction kinetics by PEEM: CO oxidation on individual (100)-type grains of Pt foil. Surf. Sci. 605, 1999–2005 (2011)Google Scholar
  218. 218.
    Vogel, D., Spiel, C., Suchorski, Y., Trinchero, A., Schlögl, R., Grönbeck, H., Rupprechter, G.: Local catalytic ignition during CO oxidation on low-index Pt and Pd surfaces: a combined PEEM, MS, and DFT study. Angew. Chem. Int. Ed. 51, 10041–10044 (2012)Google Scholar
  219. 219.
    Vogel, D., Budinska, Z., Spiel, C., Schlögl, R., Suchorski, Y., Rupprechter, G.: Silicon oxide surface segregation in CO oxidation on Pd: an in situ PEEM, MS and XPS study. Catal. Lett. 143, 235–240 (2013)Google Scholar
  220. 220.
    Günther, S., Liu, H., Menteş, T.O., Locatelli, A., Imbihl, R.: Spectromicroscopy of pulses transporting alkali metal in a surface reaction. Phys. Chem. Chem. Phys. 15, 8752–8764 (2013)Google Scholar
  221. 221.
    Rafti, M., Uecker, H., Lovis, F., Krupennikova, V., Imbihl, R.: Traveling interface modulations in the NH3 + O2 reaction on a Rh(110) surface. Phys. Chem. Chem. Phys. 14, 5260–5264 (2012)Google Scholar
  222. 222.
    Rettew, R.E., Meyer, A., Senanayake, S.D., Chen, T.-L., Petersburg, C., Flege, J.I., Falta, J., Alamgir, F.M.: Interactions of oxygen and ethylene with submonolayer Ag films supported on Ni(111). Phys. Chem. Chem. Phys. 13, 11034–11044 (2011)Google Scholar
  223. 223.
    Bethge, H., Krajewski, T., Lichtenberger, O.: Investigation of ultra-thin Ag films on Ni with the photoelectron emission microscope. Ultramicroscopy 17, 21–30 (1985)Google Scholar
  224. 224.
    Tikhov, M., Stolzenberg, M., Bauer, E.: New type of two-dimensional phase transition. Phys. Rev. B 36, 8719–8724 (1987)Google Scholar
  225. 225.
    Bauer, E., Mundschau, M., Swiech, W., Telieps, W.: Surface Studies by low energy electron microscopy (LEEM) and conventional UV photoemission electron microscopy. Ultramicroscopy 31, 49–57 (1989)Google Scholar
  226. 226.
    Altman, M.S., Luo, E.Z., Chung, W.F., Orr, B.G.: Growth kinetics and structure of ultrathin copper films on the W(110) surface studied by LEEM. In: Zangwill, A. (ed.) Proc. MRS symposium evolution of epitaxial structure and morphology, vol. 399, pp. 523–528, Boston, Massachusetts, USA 1995. MRS Symposium Proceedings Series, Pittsburgh (1996)Google Scholar
  227. 227.
    Luo, E.Z., Cai, Q., Chung, W.F., Orr, B.G., Altman, M.S.: Competing desorption pathways during epitaxial growth: LEEM investigation of Cu/W(110) heteroepitaxy. Phys. Rev. B 54, 14673–14678 (1996)Google Scholar
  228. 228.
    Koshikawa, T., Shimizu, H., Nakaguchi, A., Yasue, T., Bauer, E.: Dynamic measurements of ultra thin film growth by using LEEM and PEEM. In: Proc. atomic level characterizations for new materials and devices (ALC’05) Big Island, Hawaii, USA 2005, pp. 653–657. JSPS 141 Activity Report (2006)Google Scholar
  229. 229.
    Shimizu, H., Nakaguchi, A., Yasue, T., Koshikawa, T., Bauer, E.: Dynamic observation of structure change on double layer Cu/W(110) by means of LEEM and selected area LEED. In: Proc. atomic level characterizations for new materials and devices (ALC’05) Big Island, Hawaii, USA 2005, pp. 694–697. JSPS 141 Activity Report (2006)Google Scholar
  230. 230.
    Mundschau, M., Bauer, E., Telieps, W., Swiech, W.: In situ studies of epitaxial growth in the low energy electron microscope. Surf. Sci. 213, 381–392 (1989)Google Scholar
  231. 231.
    Duden, T.: Entwicklung und Anwendung der Polarisationsmanipulation in der Niederenergie-Elektronenmikroskopie. Ph.D. thesis, Technical University Clausthal, Germany (1996)Google Scholar
  232. 232.
    de la Figuera, J., Leonard, F., Bartelt, N.C., Stumpf, R., McCarty, K.F.: Nanoscale Periodicity in Stripe-Forming Systems at High Temperature: Au/W(110). Phys. Rev. Lett. 100, 186102, 4 pages (2008)Google Scholar
  233. 233.
    Menteş, T.O., Locatelli, A., Aballe, L., Bauer, E.: Stress Induced Stripe Formation in Pd/W(110). Phys. Rev. Lett. 101, 085701, 4 pages (2008)Google Scholar
  234. 234.
    Stojić, N., Menteş, T.O., Binggeli, N.: Self-organization in Pd/W(110): interplay between surface structure and stress. J. Phys. Condens. Matter 25, 355010, 12 pages (2013)Google Scholar
  235. 235.
    Menteş, T.O., Stojić, N., Locatelli, A., Aballe, L., Binggeli, N., Nino, M.A., Kiskinova, M., Bauer, E.: Stress engineering at the nanometer scale: two-component adlayer stripes. Europhys. Lett. 94, 38003, 6 pages (2011)Google Scholar
  236. 236.
    Georgiev, N.: LEED-diffractometry of phase transitions in chemisorbed layers and on single crystal surfaces. Ph.D. thesis, Technische Universität Clausthal, Germany (1996)Google Scholar
  237. 237.
    Gao, Y.F., Lu, W., Suo, Z.: A mesophase transition in a binary monolayer on a solid surface. Acta Mater. 50, 2297–2308 (2002)Google Scholar
  238. 238.
    Lu, W., Suo, Z.: Symmetry breaking in self-assembled monolayers on solid surfaces: anisotropic surface stress. Phys. Rev. B 65, 085401, 9 pages (2002)Google Scholar
  239. 239.
    Gao, Y.F., Suo, Z.: The orientation of the self-assembled monolayer stripes on a crystalline substrate. J. Mech. Phys. Solids 51, 147–167 (2003)Google Scholar
  240. 240.
    Lu, W., Suo, Z.: Symmetry breaking in self-assembled monolayers on solid surfaces. II. Anisotropic substrate elasticity. Phys. Rev. B 65(205418, 7 pages) (2002)Google Scholar
  241. 241.
    Léonard, F., Bartelt, N.C., Kellogg, G.L.: Effects of elastic anisotropy on the periodicity and orientation of striped stress domain patterns at solid surfaces. Phys. Rev. B 71, 045416, 7 pages (2005)Google Scholar
  242. 242.
    de la Figuera, J., Bartelt, N.C., McCarty, K.F.: Electron reflectivity measurements of Ag adatom concentrations on W(110). Surf. Sci. 600, 4062–4066 (2006)Google Scholar
  243. 243.
    Yasue, T., Amakawa, R., Shimizu, H., Nakaguchi, A., Koshikawa, T., Bauer, E.: Anomalous step contrast in LEEM during growth of Pb on W(110). In: Proc. atomic level characterizations for new materials and devices (ALC’05) Big Island, Hawaii, USA 2005, pp. 680–683. JSPS 141 Activity Report (2006)Google Scholar
  244. 244.
    Yasue, T., Amakawa, R., Shimizu, H., Nakaguchi, A., Koshikawa, T., Bauer, E.: Step contrast reversal in LEEM during Pb deposition on W(110). J. Phys. Condens. Matter 21, 314024, 5 pages (2009)Google Scholar
  245. 245.
    Pelhos, K., Hannon, J.B., Kellogg, G.L., Madey, T.E.: LEEM investigation of the faceting of the Pt covered W(111) surface. Surf. Sci. 432, 115–124 (1999)Google Scholar
  246. 246.
    Pelhos, K., Madey, T.E., Hannon, J.B., Kellogg, G.L.: Nucleation and growth of the platinum-covered W(111) surface. Surf. Rev. Lett. 6, 767–774 (1999)Google Scholar
  247. 247.
    Wu, D., Lau, W.K., He, Z.Q., Feng, Y.J., Altman, M.S., Chan, C.T.: Ordered alloying of Pd with the Mo(100) surface. Phys. Rev. B 62, 8366–8375 (2000)Google Scholar
  248. 248.
    Man, K.L., Feng, Y.J., Altman, M.S.: Formation kinetics of the Mo(100)-Ag c(2 x 2) surface alloy. Phys. Rev. B 74, 085420, 11 pages (2006)Google Scholar
  249. 249.
    Man, K.L., Feng, Y.J., Chan, C.T., Altman, M.S.: Vibrational entropy-driven dealloying of Mo(100) and W(100) surface alloys. Surf. Sci. 601, L95–L101 (2007)Google Scholar
  250. 250.
    Kellogg, G.L., Plass, R.: The relationship between the growth shape of three-dimensional Pb islands on Cu(100) and the domain orientation of the underlying c(5√ 2 x √ 2)R45 structure. Surf. Sci. 465, L777–L782 (2000)Google Scholar
  251. 251.
    Plass, R., Kellogg, G.L.: Surface morphology changes during Pb deposition on Cu(100): evidence for surface alloyed Cu(100)-c(2 x 2) Pb. Surf. Sci. 470, 106–120 (2000)Google Scholar
  252. 252.
    Kellogg, G.L., Plass, R.A.: Mesoscopic scale observations of surface alloying, surface phase transitions, domain coarsening, and 3D island growth: Pb on Cu(100). Surf. Rev. Lett. 7, 649–655 (2000)Google Scholar
  253. 253.
    Hannon, J.B., Sun, J., Pohl, K., Kellogg, G.L.: Origins of nanoscale heterogeneity in ultrathin films. Phys. Rev. Lett. 96, 246103, 4 pages (2006)Google Scholar
  254. 254.
    Sun, J., Hannon, J.B., Kellogg, G.L., Pohl, K.: Local structural and compositional determination via electron scattering: heterogeneous Cu(001)-Pd surface alloy. Phys. Rev. B 76, 205414, 10 pages (2007)Google Scholar
  255. 255.
    Bussmann, E., Sun, J., Pohl, K., Kellogg, G.L.: Palladium diffusion into bulk copper via the Cu(100) surface. J. Phys. Condens. Matter 21, 314016, 7 pages (2009)Google Scholar
  256. 256.
    Bussmann, E., Ermanoski, I., Feibelman, P.J., Bartelt, N.C., Kellogg, G.L.: Buried Pd slows self-diffusion on Cu(001). Phys. Rev. B 84, 245440, 10 pages (2011)Google Scholar
  257. 257.
    Schmid, A.K., Bartelt, N.C., Hwang, R.Q.: Alloying at surfaces by the migration of reactive two-dimensional islands. Science 290, 1561–1564 (2000)Google Scholar
  258. 258.
    Plass, R., Last, J.A., Bartelt, N.C., Kellogg, G.L.: Nanostructures: self-assembled domain patterns. Nature 412, 875 (2001)Google Scholar
  259. 259.
    Plass, R., Bartelt, N.C., Kellogg, G.L.: Dynamic observations of nanoscale self-assembly on solid surfaces. J. Phys. Condens. Matter 14, 4227–4240 (2002)Google Scholar
  260. 260.
    van Gastel, R., Plass, R., Bartelt, N.C., Kellogg, G.L.: Thermal motion and energetics of self-assembled domain structures: Pb on Cu(111). Phys. Rev. Lett. 91, 055503, 4 pages (2003)Google Scholar
  261. 261.
    van Gastel, R., Bartelt, N.C., Feibelman, P.J., Leonard, F., Kellogg, G.L.: Relationship between domain-boundary free energy and the temperature dependence of stress-domain patterns of Pb on Cu(111). Phys. Rev. B 70, 245413, 6 pages (2004)Google Scholar
  262. 262.
    van Gastel, R., Bartelt, N.C., Kellogg, G.L.: Reversible shape transition of Pb islands on Cu(111). Phys. Rev. Lett. 96, 36106, 4 pages (2006)Google Scholar
  263. 263.
    Anderson, M.L., Bartelt, N.C., Feibelman, P.J., Swartzentruber, B.S., Kellogg, G.L.: The effect of embedded Pb on Cu diffusion on Pb/Cu(111) surface alloys. Surf. Sci. 600, 1901–1908 (2006)Google Scholar
  264. 264.
    Anderson, M.L., Bartelt, N.C., Feibelman, P.J., Swartzentruber, B.S., Kellogg, G.L.: How Pb-overlayer islands move fast enough to self-assemble on Pb-Cu surface alloys. Phys. Rev. Lett. 98, 096106, 4 pages (2007)Google Scholar
  265. 265.
    Rougemaille, N., El Gabaly, F., Stumpf, R., Schmid, A.K., Thürmer, K., Bartelt, N.C., de la Figuera, J.: Labyrinthine island growth during Pd/Ru(0001) heteroepitaxy. Phys. Rev. Lett. 99, 106101, 4 pages (2007)Google Scholar
  266. 266.
    Ling, W.L., Hamilton, J.C., Thurmer, K., Thayer, G.E., de la Figuera, J., Hwang, R.Q., Carter, C.B., Bartelt, N.C., McCarty, K.F.: Herringbone and triangular patterns of dislocations in Ag, Au, and AgAu alloy films on Ru(0001). Surf. Sci. 600, 1735–1757 (2006)Google Scholar
  267. 267.
    Swiech, W., Bauer, E., Mundschau, M.: A low energy electron microscopy study of the system Si(111)-Au. Surf. Sci. 253, 283–296 (1991)Google Scholar
  268. 268.
    Slezak, J., Ondrejcek, M., Chvoj, Z., Chab, V., Conrad, H., Heun, S., Schmidt, T., Ressel, B., Prince, K.C.: Surface diffusion of Au on Si(111): a microscopic study. Phys. Rev. B 61, 16121–16128 (2000)Google Scholar
  269. 269.
    Ferralis, N., El Gabaly, F., Schmid, A.K., Maboudian, R., Carraro, C.: Real-time observation of reactive spreading of gold on silicon. Phys. Rev. Lett. 103, 256102, 4 pages (2009)Google Scholar
  270. 270.
    Dernier van der Gon, A.W., Tromp, R.M.: Phase formation sequence and domain structure: a low-energy-electron-microscopy study of the Si(111)-(√3 x√3) Ag surface. Phys. Rev. Lett. 69, 3519–3522 (1992)Google Scholar
  271. 271.
    Hashimoto, M., Guo, F.Z., Suzuki, M., Ueda, M., Matsuoka, Y., Kinoshita, T., Kobayashi, K., Shin, S., Oura, M., Takeuchi, T., Saito, Y., Matsushita, T., Yasue, T., Koshikawa, T.: Mapping of chemical bonding states of Ag/Si(111) with synchrotron radiation photo emission electron microscopy. Surf. Interface Anal. 40, 1772–1776 (2008)Google Scholar
  272. 272.
    Wall, D., Lohmar, I., Roos, K.R., Krug, J., Horn-von Hoegen, M., Meyer zu Heringdorf, F.-J.: Imaging diffusion fields on a surface with multiple reconstructions: Ag/Si(111). New J. Phys. 12, 103019, 8 pages (2010)Google Scholar
  273. 273.
    Mundschau, M., Bauer, E., Telieps, W., Swiech, W.: Initial epitaxial growth of copper silicide on Si(111) studied by low energy electron microscopy and photoemission electron microscopy. J. Appl. Phys. 65, 4747–4752 (1989)Google Scholar
  274. 274.
    Yasue, T., Koshikawa, T., Jalochowski, M., Bauer, E.: Dynamic LEEM observation of Cu nanostructure formation processes on Si(111) with hydrogen. Surf. Rev. Lett. 7, 595–599 (2000)Google Scholar
  275. 275.
    Yasue, T., Koshikawa, T., Jalochowski, M., Bauer, E.: Dynamic observations of the formation of thin Cu layers on clean and hydrogen-terminated Si(111) surfaces. Surf. Sci. 480, 118–127 (2001)Google Scholar
  276. 276.
    Bennett, P.A., Lee, M.Y., Parikh, S.A., Wurm, K., Phaneuf, R.J.: Surface phase transformations in the Ni/Si(111) system observed in real time using low-energy electron microscopy. J. Vac. Sci. Technol. A 13, 1728–1732 (1995)Google Scholar
  277. 277.
    Phaneuf, R.J., Hong, Y., Horch, S., Bennett, P.A.: Two dimensional phase separation for Co adsorbed on Si(111). Phys. Rev. Lett. 78, 4605–4608 (1997)Google Scholar
  278. 278.
    Phaneuf, R.J., Bennett, P.A.: LEEM studies of phase separations and surface depletion of Co and Ni on Si(111). Surf. Rev. Lett. 5, 1179–1188 (1998)Google Scholar
  279. 279.
    Phaneuf, R.J., Bennett, P.A., Marsi, M., Gunther, S., Gregoratti, L., Casalis, L., Kiskinova, M.: Equilibration of ring-cluster surface phases and silicide islands for Co adsorbed on Si(111). Surf. Sci. 431, 232–241 (1999)Google Scholar
  280. 280.
    Phaneuf, R.J., Hong, Y., Horch, S., Bennett, P.A.: Observations of reversible and irreversible structural transitions of cobalt on Si(111) with LEEM. Micron 30, 13–20 (1999)Google Scholar
  281. 281.
    Tromp, R.M., Michely, T.: Atomic-layer titration of surface reactions. Nature 373, 499–501 (1995)Google Scholar
  282. 282.
    Michely, T., Reuter, M.C., Tromp, R.M.: Al on Si(111): phase diagram and atomic mechanisms. Phys. Rev. B 53, 4105–4108 (1996)Google Scholar
  283. 283.
    Man, K.L., Tringides, M.C., Loy, M.M.T., Altman, M.S.: Anomalous Mass Transport in the Pb Wetting Layer on the Si(111) Surface. Phys. Rev. Lett. 101, 226102, 4 pages (2008)Google Scholar
  284. 284.
    Pavlovska, A., Bauer, E., Giesen, M.: Low energy electron microscopy study of In on Si(111). J. Vac. Sci. Technol. B 20, 2478–2491 (2002)Google Scholar
  285. 285.
    Pavlovska, A., Bauer, E.: Surface reconstructions in two and three dimensions: In on Si(111). Surf. Interface Anal. 37, 110–114 (2005)Google Scholar
  286. 286.
    Shimizu, H., Nakaguchi, A., Yasue, T., Koshikawa, T., Guo, F-Z., Tamenori, Y., Matsushita,T., Kinoshita,T., Kobayashi, K., Bauer, E.: Surface chemical analysis on the selected area of In/Si(111) by means of LEEM, LEED and SR-XPEEM. In: Proc. atomic level characterizations for new materials and devices (ALC’05) Big Island, Hawaii, USA 2005, pp. 710–713. JSPS 141 Activity Report (2006)Google Scholar
  287. 287.
    Nakaguchi, A., Guo, F.Z., Hashimoto, M., Ueda, M., Yasue, T., Matsushita, T., Tamenori, Y., Kinoshita, T., Kobayashi, K., Koshikawa, T.: Sb on In/Si(111) processes with dynamically observable LEEM, selected area LEED and chemically analyzed SR-XPEEM. Surf. Interface Anal. 38, 1773–1776 (2006)Google Scholar
  288. 288.
    Ueda, M., Guo, F-Z., Nakaguchi, A., Suzuki, M., Hashimoto, M., Matsuoka, T., Kinoshita, T., Kobayashi, K., Shin, S., Oura, M., Takeuchi, T., Saito, Y., Yasue, T., Koshikawa, T.: Observation of the growth process Sb on In/Si(111) by observation of the growth process Sb on In/Si(111) by the synchrotron radiation XPEEM and LEEM. In: Proc. atomic level characterizations for new materials and devices (ALC’07), Kanasawa, Japan 2007, pp. 216–219. JSPS 141 Activity Report (2008)Google Scholar
  289. 289.
    Matsuoka, Y., Guo, F-Z., Hashimoto, M., Suzuki, M., Ueda, M., Yasue, T., Kinoshita, T., Kobayashi, K., Shin, S., Oura, M., Takeuchi, T., Saito, Y., Koshikawa, T.: Ag on In/Si(111) processes by means of LEEM, selected area XPS and chemically analyzed SR -XPEEM. In: Proc. atomic level characterizations for new materials and devices (ALC’7), Kanasawa, Japan 2007, pp. 212–215. JSPS 141 Activity Report (2008)Google Scholar
  290. 290.
    Schmidt, T., Gangopadhyay, S., Flege, J.I., Clausen, T., Locatelli, A., Heun, S., Falta, J.: Self-organized 2D nanopatterns after low-coverage Ga adsorption on Si(111). New J. Phys. 7, 193, 11 pages (2005)Google Scholar
  291. 291.
    Horn-von Hoegen, M., Meyer zu Heringdorf, F.-J., Kaehler, D., Schmidt, T., Bauer, E.: Adsorption induced giant faceting of vicinal Si(100). Thin Solid Films 336, 16–21 (1998)Google Scholar
  292. 292.
    Meyer zu Heringdorf, F.-J., Kaehler, D., Schmidt, T., Bauer, E., Copel, M., Minoda, H., Horn-von Hoegen, M.: Giant faceting of vicinal Si(001) induced by Au adsorption. Surf. Rev. Lett. 5, 1167–1178 (1998)Google Scholar
  293. 293.
    Horn-von Hoegen, M., Meyer zu Hennigdorf, F.-J., Hild, R., Zahl, P., Schmidt, T., Bauer, E.: Au-induced giant faceting of vicinal Si(100). Surf. Sci. 433–435, 475–480 (1999)Google Scholar
  294. 294.
    Meyer zu Heringdorf, F.-J., Hild, R., Zahl, P., Schmidt, T., Ressel, B., Heun, S., Bauer, E., Horn-von Hoegen, M.: Local Au coverage as driving force for Au induced faceting of vicinal Si(001): a LEEM and XPEEM study. Surf. Sci. 480, 103–108 (2001)Google Scholar
  295. 295.
    Meyer zu Heringdorf, F.-J., Schmidt, T., Heun, S., Hild, R., Zahl, P., Ressel, B., Bauer, E., Horn-von Hoegen, M.: Spatial variation of Au coverage as driving force for nanoscopic pattern formation. Phys. Rev. Lett. 86, 5088–5091 (2001)Google Scholar
  296. 296.
    Michely, T., Reuter, M.C., Copel, M., Tromp, R.M.: Symmetry, structure, and step induced ordering of the Si(001)-(2 x 3)Ag surface. Phys. Rev. Lett. 73, 2095–2098 (1994)Google Scholar
  297. 297.
    Roos, K.R., Roos, K.L., Lohmar, I., Wall, D., Krug, J., Horn-von Hoegen, M., Meyer zu Heringdorf, F.-J.: Real-time view of mesoscopic surface diffusion. Phys. Rev. Lett. 100, 016103, 4 pages (2008)Google Scholar
  298. 298.
    Wall, D., Roos, K.R., Horn-von Hoegen, M., Meyer zu Heringdorf, F.-J.: LEEM/PEEM study of anisotropic diffusion fields in the Ag/Si(001) system. In: Adler, D.L., Bauer, E., Kellogg, G.L., Scholl, A. (eds.) Proc. MRS symposium advances and applications of surface electron microscopy, San Francisco 2008. MRS Proceedings, vol. 1088, pp. W05–04 (2008)Google Scholar
  299. 299.
    Sindermann, S., Wall, D., Roos, K.R., Horn-von Hoegen, M., Meyer zu Heringdorf, F.-J.: Anisotropy of Ag diffusion on vicinal Si surfaces. e-J. Surf. Sci. Nanotech. 8, 372–376 (2010)Google Scholar
  300. 300.
    McKay, H.A., Feenstra, R.M.: Low energy electron microscopy of indium on Si(001) surfaces. Surf. Sci. 547, 127–138 (2003)Google Scholar
  301. 301.
    Thien, D., Kury, P., Horn-von Hoegen, M., Meyer zu Heringdorf, F.-J., van Heys, J., Lindenblatt, M., Pehlke, E.: Domain sensitive contrast in photoelectron emission microscopy. Phys. Rev. Lett. 99, 196102, 4 pages (2007)Google Scholar
  302. 302.
    Speckmann, M., Schmidt, T., Locatelli, A., Menteş, T.O., Niño, M.Á., Falta, J.: Silver: a novel growth catalyst for Ge nanoislands on Si(113). Phys. Status Solidi RRL 3, 305–307 (2009)Google Scholar
  303. 303.
    Clausen, T., Schmidt, T., Flege, J.I., Falta, J., Locatelli, A., Mentes, T.O., Heun, S., Guo, F.Z.: Real-time low-energy electron microscopy study of Ga adsorption and facet array formation on Si(113). e-J. Surf. Sci. Nanotech. 3, 379–383 (2005)Google Scholar
  304. 304.
    Schmidt, T., Clausen, T., Flege, J.I., Gangopadhyay, S., Locatelli, A., Mentes, T.O., Guo, F.Z., Heun, S., Falta, J.: Adsorbate induced self-ordering or germanium nanoislands on Si(113). New J. Phys. 9 (392), 1–13 (2007)Google Scholar
  305. 305.
    Speckmann, M., Schmidt, T., Flege, J.I., Sadowski, J.T., Sutter, P., Falta, J.: Temperature dependent low energy electron microscopy study of Ge island growth on bare and Ga terminated Si(112). J. Phys. Condens. Matter 21, 314020, 7 pages (2009)Google Scholar
  306. 306.
    Speckmann, M., Schmidt, T., Flege, T.J., Falta, J.: In adsorption on Si(112) and its impact on Ge growth. IBM J. Res. Dev. 55, paper 11, 8 pages (2011)Google Scholar
  307. 307.
    Pavlovska, A., Bauer, E.: Wetting and reactive thin film growth. Surf. Sci. 480, 128–136 (2001)Google Scholar
  308. 308.
    Zheng, C.X., Tang, W.X., Jesson, D.E.: Relief of surface stress at steps during displacive adsorption of As on Si(111). Appl. Phys. Lett. 100, 201602, 4 pages (2012)Google Scholar
  309. 309.
    Yao, Y., Fu, Q., Tan, D., Bao, X.: A comparative study of CCl4 reactions on Ag and Si surfaces by in situ ultraviolet photoemission electron microscopy. J. Phys. Condens. Matter 21, 314014, 7 pages (2009)Google Scholar
  310. 310.
    Ihm, K., Han, J.H., Kim, B., Chung, S., Hwang, C.-C., Kang, T.-H., Kim, K.-J., Jung, Y.J., An, K.-S.: Selective adsorption of benzoic acid species on patterned OH/Si(100) surface. J. Appl. Phys. 100, 043510, 5 pages (2006)Google Scholar
  311. 311.
    Mullet, C.H., Chiang, S.: High temperature growth of Ag phases on Ge(111). J. Vac. Sci. Technol. A 31, 020602, 5 pages (2013)Google Scholar
  312. 312.
    Sato, Y., Chiang, S., Bartelt, N.C.: Spontaneous domain switching during phase separation of Pb on Ge(111). Phys. Rev. Lett. 99, 096103, 4 pages (2007)Google Scholar
  313. 313.
    Hibino, H., Watanabe, Y.: Arrangement of Au-Si alloy islands at atomic steps. Surf. Sci. 588, L233–L238 (2005)Google Scholar
  314. 314.
    Hannon, J.B., Kodambaka, S., Ross, F.M., Tromp, R.M.: The influence of the surface migration of gold on the growth of silicon nanowires. Nature 440, 69–71 (2006)Google Scholar
  315. 315.
    Tersoff, J., Dernier van der Gon, A.W., Tromp, R.M.: Shape oscillations in growth of small crystals. Phys. Rev. Lett. 70, 1143–1146 (1993)Google Scholar
  316. 316.
    Tersoff, J., Tromp, R.M.: Shape transition in growth of strained islands: spontaneous formation of quantum wires. Phys. Rev. Lett. 70, 2782–2785 (1993)Google Scholar
  317. 317.
    Tromp, R.M., Denier van der Gon, A.W., LeGoues, F.K., Reuter, M.C.: Observation of buried interfaces with low energy electron microscopy. Phys. Rev. Lett. 71, 3299–3302 (1993)Google Scholar
  318. 318.
    Denier van der Gon, A.W., Tromp, R.M., Reuter, M.C.: Low energy electron microscopy studies of Ge and Ag growth on Si(111). Thin Solid Films 236, 140–145 (1993)Google Scholar
  319. 319.
    Tang, W.X., Man, K.L., Huang, H., Woo, C.H., Altman, M.S.: Growth shapes of Ag crystallites on the Si(111) surface. J. Vac. Sci. Technol. B 20, 2492–2495 (2002)Google Scholar
  320. 320.
    Man, K.L., Tang, W.X., Huang, H., Altman, M.S.: Kinetic Limitations in Two-and Three-Dimensional Growth. In: Chason, E., Gilmer, G.H., Huang, H., Wang, E. (eds.) MRS Symposium KK—kinetics-driven nanopatterning on Surfaces, vol. 849, pp. KK5.3.1–KK5.3.9, Boston, Massachusetts, USA 2004. Materials Research Society (2005)Google Scholar
  321. 321.
    Wall, D., Tikhonov, S., Sindermann, S., Spoddig, D., Hassel, C., Horn-von Hoegen, M., Meyer zu Heringdorf, F.-J.: Shape, orientation, and crystalline composition of silver islands on Si(111). IBM J. Res. Dev. 55, Paper 9, 6 pages (2011)Google Scholar
  322. 322.
    Gloskovskii, A., Valdaitsev, D., Nepijko, S.A., Schönhense, G., Rethfeld, B.: Coexisting electron emission mechanisms in small metal particles observed in fs-laser excited PEEM. Surf. Sci. 601, 4706–4713 (2007)Google Scholar
  323. 323.
    Li, B.Q., Swiech, W., Venables, J.A., Zuo, J.M.: A LEEM study of bamboo-like growth of Ag crystals on Si(001) surfaces. Surf. Sci. 569, 142–148 (2004)Google Scholar
  324. 324.
    Wall, D., Sindermann, S., Roos, K.R., Horn-von Hoegen, M., Meyer zu Heringdorf, F.-J.: The influence of anisotropic diffusion on Ag nanowire formation. J. Phys. Condens. Matter 21, 314023, 4 pages (2009)Google Scholar
  325. 325.
    Roos, K.R., Roos, K.L., Horn-von Hoegen, M., Meyer zu Heringdorf, F.-J.: High temperature self-assembly of Ag nanowires on vicinal Si(001). J. Phys. Condens. Matter 17, S1407–S1414 (2005)Google Scholar
  326. 326.
    Meyer zu Heringdorf, F.-J., Roos, K.L., Wiethoff, C., Horn-von Hoegen, M., Roos, K.R.: Growth of Ag nanowires on Au-pre-facetted 4° vicinal Si(001). Surf. Sci. 602, 1852–1857 (2008)Google Scholar
  327. 327.
    Yasue, T., Koshikawa, T., Jalochowski, M., Bauer, E.: LEEM observation of formation of Cu nano-islands on Si(111) surface by hydrogen termination. Surf. Sci. 493, 381–388 (2001)Google Scholar
  328. 328.
    Bauer, E.: Low energy electron microscopy. In: Vanselow, R., Howe, R. (eds.) Chemistry and Physics of Solid Surfaces VIII, pp. 267–288. Springer, Berlin (1990)Google Scholar
  329. 329.
    Maeda, F., Hibino, H., Suzuki, S., Kobayashi, Y., Watanabe, Y., Guo, F.Z.: Surface reactions of Co on SiO2 thin layer/Si substrate studied by LEEM and PEEM. e-J. Surf. Sci. Nanotech. 4, 155–160 (2006)Google Scholar
  330. 330.
    Maeda, F., Hibino, H., Suzuki, S., Guo, F.Z.: Oxide-mediated formation of α-FeSi2 on Si(001) studied by X-ray adsorption near edge structure analysis using SPELEEM. Surf. Interface Anal. 40, 1747–1750 (2008)Google Scholar
  331. 331.
    Schmidt, Th., Bauer, E.: Interfactant-mediated quasi-Frank-van der Merwe growth of Pb on Si(111). Phys. Rev. B 62, 15815–15825 (2000)Google Scholar
  332. 332.
    Bauer, E.: Phänomenologische Theorie der Kristallabscheidung an Oberflächen I. Z. Kristallogr. 110, 372–394 (1958)Google Scholar
  333. 333.
    Venables, J.A.: Introduction to Surface and Thin Films Processes. Cambridge University Press, Cambridge, UK (2000). p. 146Google Scholar
  334. 334.
    Ibach, H.: Physics of Surfaces and Interfaces. Springer Verlag, Berlin (2006). p. 171Google Scholar
  335. 335.
    Schmidt, Th., Bauer, E.: Influence of interfactants on thin metal film growth. Surf. Sci. 480, 137–144 (2001)Google Scholar
  336. 336.
    Li, L., Koziol, C., Wurm, K., Hong, Y., Bauer, E., Tsong, I.S.T.: Surface morphology of Pb overlayers grown on Si(100)-(2x1). Phys. Rev. B 50, 10834–10842 (1994)Google Scholar
  337. 337.
    Jalochowski, M., Bauer, E.: Self-assembled parallel mesoscopic Pb-wires on Au-modified Si(533) substrates. Prog. Surf. Sci. 67, 79–97 (2001)Google Scholar
  338. 338.
    Jalochowski, M., Bauer, E.: Growth of metallic nanowires on anisotropic Si substrates: Pb on vicinal Si(001), Si(755), Si(533) and Si(110). Surf. Sci. 480, 109–117 (2001)Google Scholar
  339. 339.
    Bennett, P.A., He, Z., Smith, D.J., Ross, F.M.: Endotaxial silicide nanowires: a review. Thin Solid Films 519, 8434–8440 (2011)Google Scholar
  340. 340.
    Bennett, P.A., Ashcroft, B., He, Z., Tromp, R.M.: Growth dynamics of titanium silicide nanowires observed with low-energy electron microscopy. J. Vac. Sci. Technol. B 20, 2500–2504 (2002)Google Scholar
  341. 341.
    Yang, W.-C., Ade, H., Nemanich, R.J.: Shape stability of TiSi2 islands on Si (111). J. Appl. Phys. 95, 1572–1576 (2004)Google Scholar
  342. 342.
    Yang, W.-C., Zeman, M., Ade, H., Nemanich, R.J.: Attractive migration and coalescence: a significant process in the coarsening of TiSi2 islands on the Si(111) surface. Phys. Rev. Lett. 90, 136102, 4 pages (2003)Google Scholar
  343. 343.
    Fitting, L., Zeman, M.C., Yang, W.-C., Nemanich, R.J.: Influence of strain, surface diffusion and Ostwald ripening on the evolution of nanostructures for erbium on Si(001). J. Appl. Phys. 93, 4180–4184 (2003)Google Scholar
  344. 344.
    Yang, W.-C., Ade, H., Nemanich, R.J.: Stability and dynamics of Pt-Si liquid microdroplets on Si(001). Phys. Rev. B 69, 045421, 8 pages (2004)Google Scholar
  345. 345.
    Bennett, P.A., Chobanian, J., Flege, J.I., Sutter, E., Sutter, P.: Surface thermomigration of nanoscale Pt-Si droplets on stepped Si(100). Phys. Rev. B 76, 125410, 6 pages (2007)Google Scholar
  346. 346.
    Sutter, P., Bennett, P.A., Flege, J.I., Sutter, E.: Steering liquid Pt-Si nanodroplets on Si(100) by interactions with surface steps. Phys. Rev. Lett. 99, 125504, 4 pages (2007)Google Scholar
  347. 347.
    Shiraki, Y., Usami, N. (eds.): Silicon-Germanium (SiGe) Nanostructures: Production, Properties and Applications in Electronics. Woodhead Publishing, Cambridge, UK (2011)Google Scholar
  348. 348.
    Aqua, J.-N., Berbezier, I., Favre, L., Frisch, T., Ronda, A.: Growth and self-organization of SiGe nanostructures. Phys. Rep. 522, 59–189 (2013)Google Scholar
  349. 349.
    Tromp, R.M., Reuter, M.C.: Local dimer exchange in surfactant-mediated epitaxial growth. Phys. Rev. Lett. 68, 954–967 (1992)Google Scholar
  350. 350.
    Ross, F.M., Tromp, R.M., Reuter, M.C.: Transition states between pyramids and domes during Ge/Si island growth. Science 286, 1931–1934 (1999)Google Scholar
  351. 351.
    Tromp, R.M., Ross, F.M., Reuter, M.C.: Instability-driven SiGe island growth. Phys. Rev. Lett. 84, 4641–4644 (2000)Google Scholar
  352. 352.
    Tromp, R.M., Ross, F.M.: Advances in in situ ultra-high vacuum electron microscopy: growth of SiGe on Si. Annu. Rev. Mater. Sci. 30, 431–449 (2000)Google Scholar
  353. 353.
    Hannon, J.B., Copel, M., Stumpf, R., Reuter, M.C., Tromp, R.M.: Critical role of surface steps in the alloying of Ge on Si(001). Phys. Rev. Lett. 92, 216104, 4 pages (2004)Google Scholar
  354. 354.
    Sutter, P., Mateeva, E., Lagally, M.G.: Si growth on partially relaxed Ge islands. J. Vac. Sci. Technol. B 16, 1560–1563 (1998)Google Scholar
  355. 355.
    Sutter, P., Lagally, M.G.: Embedding of nanoscale 3D SiGe islands in a Si matrix. Phys. Rev. Lett. 81, 3471 (1998)Google Scholar
  356. 356.
    Sutter, P., Mateeva, E., Sullivan, J.S., Lagally, M.G.: Low-energy electron microscopy of nanoscale three-dimensional SiGe islands on Si(100). Thin Solid Films 336, 262–270 (1998)Google Scholar
  357. 357.
    Sutter, P., Lagally, M.G.: Nucleationless three-dimensional island formation in low-misfit heteroepitaxy. Phys. Rev. Lett. 84, 4637–4640 (2000)Google Scholar
  358. 358.
    Sutter, P., Lagally, M.G.: Nucleationless island formation in SiGe/Si(100) heteroepitaxy. Mater. Sci. Eng. B 89, 45–48 (2002)Google Scholar
  359. 359.
    Huang, M., Ritz, C.S., Novakovic, B., Yu, D., Zhang, Y., Flack, F., Savage, D.E., Evans, P.G., Knezevic, I., Liu, F., Lagally, M.G.: Mechano-electronic superlattices in silicon nanoribbons. ACS NANO 3, 721–727 (2009)Google Scholar
  360. 360.
    Maxson, J.B., Savage, D.E., Liu, F., Tromp, R.M., Reuter, M.C., Lagally, M.G.: Thermal roughening of a thin film: a new type of roughening transition. Phys. Rev. Lett. 85, 2152–2155 (2000)Google Scholar
  361. 361.
    Tromp, R.M.: Surface stress and interface formation. Phys. Rev. B 47, 7125–7127 (1993)Google Scholar
  362. 362.
    Moutanabbir, O., Ratto, F., Heun, S., Scheerschmidt, K., Locatelli, A., Rosei, F.: Dynamic probe of atom exchange during monolayer growth. Phys. Rev. B 85, 201416(R), 5 pages (2012)Google Scholar
  363. 363.
    MacLeod, J.M., Lipton-Duffin, J.A., Lanke, U., Urquhart, S.G., Rosei, F.: Shape transition in very large germanium islands on Si(111). Appl. Phys. Lett. 94, 103109, 3 pages (2009)Google Scholar
  364. 364.
    Ratto, F., Rosei, F., Locatelli, A., Cherifi, S., Fontana, S., Heun, S., Szkutnik, P.-D., Sgarlata, A., De Crescenzi, M., Motta, N.: Composition of Ge(Si) islands in the growth of Ge on Si(111). Appl. Phys. Lett. 84, 4526–4528 (2004)Google Scholar
  365. 365.
    Ratto, F., Rosei, F., Locatelli, A., Cherifi, S., Fontana, S., Heun, S., Szkutnik, P.-D., Sgarlata, A., De Crescenzi, M., Motta, N.: Composition of Ge(Si) islands in the growth of Ge on Si(111) by x-ray spectromicroscopy. J. Appl. Phys. 97, 043516, 8 pages (2005)Google Scholar
  366. 366.
    Ratto, F., Locatelli, A., Fontana, S., Kharrazi, S., Ashtaputre, S., Kulkarni, S.K., Heun, S., Rosei, F.: Diffusion dynamics during the nucleation and growth of GeSi nanostructures on Si(111). Phys. Rev. Lett. 96, 096103, 4 pages (2006)Google Scholar
  367. 367.
    Schmidt, Th., Flege, J.I., Gangopadhyay, S., Clausen, T., Locatelli, A., Heun, S., Falta, J.: Alignment of Ge nanoislands on Si(111) by Ga-induced substrate self-patterning. Phys. Rev. Lett. 98, 066104, 4 pages (2007)Google Scholar
  368. 368.
    Schmidt, Th., Speckmann, M., Falta, J., Mentes, T.O., Nino, M.A., Locatelli, A.: Ge growth on partially and entirely Ag covered Si(111). e-J. Surf. Sci. Nanotech. 8, 221–226 (2010)Google Scholar
  369. 369.
    Wang, Y.-L., Guo, H.-M., Qin, Z.-H., Ma, H.-F., Gao, H.-J.: Toward a detailed understanding of Si(111)-7 × 7 surface and adsorbed Ge nanostructures: fabrications, structures, and calculations. J. Nano Mater. 2008, 874213, 19 pages (2008)Google Scholar
  370. 370.
    Yasue, T., Koshikawa, T., Bauer, E.: LEEM and LEED study of the growth of Ge on the Si(113) surface. J. Vac. Sci. Technol. B 20, 2496–2499 (2002)Google Scholar
  371. 371.
    Clausen, T., Schmidt, T., Flege, J.I., Locatelli, A., Mentes, T.O., Heun, S., Guo, F.Z., Falta, J.: Temperature dependent low energy electron microscopy.study of Ge growth on Si(113). Appl. Surf. Sci. 252, 5321–5325 (2006)Google Scholar
  372. 372.
    Schmidt, T., Flege, J.I., Speckmann, M., Clausen, T., Gangopadhyay, S., Locatelli, A., Mentes, T.O., Heun, S., Guo, F.Z., Sutter, P., Faltaa, J.: From nanoislands to nanowires: Growth of germanium on gallium-terminated silicon surfaces. Phys. Stat. Sol. A 206, 1718–1722 (2009)Google Scholar
  373. 373.
    Falta, J., Reuter, M.C., Tromp, R.M.: Growth modes of Ge on GaAs(001). Appl. Phys. Lett. 65, 1680–1682 (1994)Google Scholar
  374. 374.
    Tromp, R.M., Reuter, M.C.: Kinetic Instability in the Growth of CaF2 on Si(111). Phys. Rev. Lett. 73, 110–113 (1994)Google Scholar
  375. 375.
    Tromp, R.M., LeGoues, F.K., Reuter, M.C.: Strain relief during growth: CaF2 on Si(111). Phys. Rev. Lett. 74, 2706–2709 (1995)Google Scholar
  376. 376.
    Bauer, E., Wei, Y., Muller, T., Pavlovska, A., Tsong, I.S.T.: Reactive crystal growth in two dimensions: silicon nitride on Si(111). Phys. Rev. B 51, 17891–17901 (1995)Google Scholar
  377. 377.
    Bauer, E., Pavlovska, A., Tsong, I.S.T.: In situ growth studies by low energy electron microscopy (LEEM) and low electron diffraction (LEED). In: Bailey, G.W., Dimlich, R.V.W., Alexander, K.B., McCarty, J.J., Pretlow, T.P. (eds.) Proc. MSA meeting microscopy and microanalysis ’97, Cleveland, Ohio 1997. Microscopy and Microanalysis, vol. 3, suppl. 2, pp. 611–612 (1997)Google Scholar
  378. 378.
    Pavlovska, A., Torres, V.M., Bauer, E., Doak, R.B., Tsong, I.S.T., Thomson, D.B., Davis, R.F.: Low-energy electron microscopy observations of GaN homoepitaxy using a supersonic source. Appl. Phys. Lett. 75, 989–991 (1999)Google Scholar
  379. 379.
    Pavlovska, A., Bauer, E., Torres, V.M., Edwards, J.L., Doak, R.B., Tsong, I.S.T., Ramachadran, V., Feenstra, F.M.: In situ real-time studies of GaN growth on 6H-SiC(0001) by low energy electron microscopy (LEEM). J. Cryst. Growth 189/190, 310–316 (1998)Google Scholar
  380. 380.
    Hu, C., Smith, D.J., Doak, R.B., Tsong, I.S.T.: Morphological control of gan buffer layers grown by molecular beam epitaxy on 6H-SiC (0001). Surf. Rev. Lett. 7, 565–570 (2000)Google Scholar
  381. 381.
    Pavlovska, A., Bauer, E.: Low energy electron microscopy studies of the growth of GaN on 6H-SiC(0001). Surf. Rev. Lett. 8, 337–346 (2001)Google Scholar
  382. 382.
    Yang, W.C., Rodriguez, B.J., Park, M., Nemanich, R.J., Ambacher, O., Cimalla, V.: Photoelectron emission microscopy observation of inversion domain boundaries of GaN-based lateral polarity heterostructures. J. Appl. Phys. 94, 5720–5725 (2003)Google Scholar
  383. 383.
    Heun, S., Watanabe, Y., Ressel, B., Bottomley, D., Schmidt, T., Prince, K.C.: Core-level photoelectron spectroscopy from individual heteroepitaxial nanocrystals on GaAs(001). Phys. Rev. B 63, 125335, 8 pages (2001)Google Scholar
  384. 384.
    Biasiol, G., Heun, S., Golinelli, G.B., Locatelli, A., Mentes, T.O., Guo, F.Z., Hofer, C., Teichert, C., Sorba, L.: Surface compositional gradients of InAs/GaAs quantum dots. Appl. Phys. Lett. 87, 223106, 3 pages (2005)Google Scholar
  385. 385.
    Biasiol, G., Heun, S., Golinelli, G.B., Locatelli, A., Mentes, T.O., Guo, F.Z., Sorba, L.: Surface concentration mapping of InAs/GaAs quantum dots. In: Jantsch, W., Schaffler, F. (eds.) AIP Conf. Proc., pp. 49–50 (2007)Google Scholar
  386. 386.
    Biasiol, G., Baranwal, V., Heun, S., Prasciolu, M., Tormen, M., Locatelli, A., Mentes, T.O., Nino, M.A., Sorba, L.: Composition uniformity of site-controlled InAs/GaAs quantum dots. J. Cryst. Growth 323, 176–179 (2011)Google Scholar
  387. 387.
    Biasiol, G., Magri, R., Heun, S., Locatelli, A., Mentes, T.O., Sorba, L.: Surface compositional mapping of self-assembled InAs/GaAs quantum rings. J. Cryst. Growth 311, 1764–1766 (2009)Google Scholar
  388. 388.
    Magri, R., Heun, S., Biasiol, G., Locatelli, A., Mentes, T.O., Sorba, L.: Surface compositional profiles of self-assembled InAs/GaAs quantum rings. In: Caldas, M.J., Studart, N. (eds.) AIP Conf. Proc., pp. 3–4 (2010)Google Scholar
  389. 389.
    Baranwal, V., Biasiol, G., Heun, S., Locatelli, A., Mentes, T.O., Niño, M., Sorba, L.: Kinetics of the evolution of InAs/GaAs quantum dots to quantum rings: a combined x-ray, atomic force microscopy, and photoluminescence study. Phys. Rev. B 80, 155328, 7 pages (2009)Google Scholar
  390. 390.
    Biasiol, G., Heun, S.: Compositional mapping of semiconductor quantum dots and rings. Phys. Rep. 500, 117–173 (2011)Google Scholar
  391. 391.
    Robinson, J.T., Ratto, F., Moutanabbir, O., Heun, S., Locatelli, A., Mentes, T.O., Aballe, L., Dubon, O.D.: Gold-catalyzed oxide nanopatterns for the directed assembly of Ge island arrays on Si. Nano Lett. 7, 2655–2659 (2007)Google Scholar
  392. 392.
    Hjort, M., Wallentin, J., Timm, R., Zakharov, A.A., Andersen, J.N., Samuelson, L., Borgstrom, M.T., Mikkelsen, A.: Doping profile of InP nanowires directly imaged by photoemission electron microscopy. Appl. Phys. Lett. 99, 233113, 3 pages (2011)Google Scholar
  393. 393.
    Kennedy, S.M., Hjort, M., Mandl, B., Marsell, E., Zakharov, A.A., Mikkelsen, A., Paganin, D.M., Jesson, D.E.: Characterizing the geometry of InAs nanowires using mirror electron microscopy. Nanotechnology 23, 125703, 9 pages (2012)Google Scholar
  394. 394.
    Mandl, B., Stangl, J., Hilner, E., Zakharov, A.A., Hillerich, K., Dey, A.W., Samuelson, L., Bauer, G., Deppert, K., Mikkelsen, A.: Growth mechanism of self-catalyzed group III-V nanowires. Nano Lett. 10, 4443–4449 (2010)Google Scholar
  395. 395.
    Ercolani, D., Lazzarino, M., Mori, G., Ressel, B., Sorba, L., Locatelli, A., Cherifi, S., Ballestrazzi, A., Heun, S.: GaAs oxide desorption under extreme ultraviolet photon flux. Adv. Funct. Mater. 15, 587–592 (2005)Google Scholar
  396. 396.
    Mori, G., Lazzarino, M., Ercolani, D., Sorba, L., Heun, S., Locatelli, A.: Desorption dynamics of oxide nanostructures fabricated by local anodic oxidation nanolithography. J. Appl. Phys. 97, 114324, 8 pages (2005)Google Scholar
  397. 397.
    Mori, G., Lazzarino, M., Ercolani, D., Biasiol, G., Sorba, L., Heun, S., Locatelli, A.: Evidence of material mixing during local anodic oxidation nanolithography. J. Appl. Phys. 98, 114303, 6 pages (2005)Google Scholar
  398. 398.
    Mori, G., Lazzarino, M., Ercolani, D., Biasiol, G., Locatelli, A., Sorba, L., Heun, S.: X-ray induced variation of the chemistry of GaAs/AlAs oxide nanostructures. Nucl. Instrum. Method Phys. Res. B 246, 39–44 (2006)Google Scholar
  399. 399.
    Lazzarino, M., Mori, G., Sorba, L., Ercolani, D., Biasiol, G., Heun, S., Locatelli, A.: Chemistry and formation process of Ga(Al)As oxide during local anodic oxidation nanolithography. Surf. Sci. 600, 3739–3743 (2006)Google Scholar
  400. 400.
    Jesson, D.E., Tang, W.X.: Surface electron microscopy of Ga droplet dynamics on GaAs (001). Microsc Sci Technol Appl Educ 4, 1608–1619 (2010)Google Scholar
  401. 401.
    Tersoff, J., Jesson, D.E., Tang, W.X.: Running droplets of gallium from evaporation of gallium arsenide. Science 324, 236–238 (2009)Google Scholar
  402. 402.
    Hilner, E., Zakharov, A.A., Schulte, K., Kratzer, P., Andersen, J.N., Lundgren, E., Mikkelsen, A.: Ordering of the nanoscale step morphology as a mechanism for droplet self-propulsion. Nano Lett. 9, 2710–2714 (2009)Google Scholar
  403. 403.
    Tang, W.X., Jesson, D.E., Pavlov, K.M., Morgan, M.J., Usher, B.F.: Ga droplet morphology on GaAs(001) studied by Lloyd’s mirror photoemission electron microscopy. J. Phys. Condens. Matter 21, 314022, 5 pages (2009)Google Scholar
  404. 404.
    Zhou, Z.Y., Tang, W.X., Jesson, D.E., Tersoff, J.: Time evolution of the Ga droplet size distribution during Langmuir evaporation of GaAs(001). Appl. Phys. Lett. 97, 191914, 3 pages (2010)Google Scholar
  405. 405.
    Tang, W.X., Zheng, C.X., Zhou, Z.Y., Jesson, D.E., Tersoff, J.: Ga droplet surface dynamics during Langmuir evaporation of GaAs. IBM J. Res. Dev. 55(4), 10:1–10:7 (2011)Google Scholar
  406. 406.
    Hilner, E., Mikkelsen, A., Eriksson, J., Andersen, J.N., Lundgren, E.: Au wetting and nanoparticle stability on GaAs(111)B. Appl. Phys. Lett. 89, 251912, 3 pages (2006)Google Scholar
  407. 407.
    Hajjar, S., Garreau, G., Josien, L., Bubendorff, J.L., Berling, D., Mehdaoui, A., Pirri, C., Maroutian, T., Renard, C., Bouchier, D., Petit, M., Spiesser, A., Dau, M.T., Michez, L., Le Thanh, V., Menteş, T.O., Niño, M.Á., Locatelli, A.: Morphology and composition of Au catalysts on Ge(111) obtained by thermal dewetting. Phys. Rev. B 84, 125325, 10 pages (2011)Google Scholar
  408. 408.
    Qi, B., Ólafsson, S., Göthelid, M., Gislason, H.P., Agnarsson, B.: Photoemission and low energy electron microscopy study on the formation and nitridation of indium droplets on Si (111)7 × 7 surfaces. Thin Solid Films 531, 61–69 (2013)Google Scholar
  409. 409.
    Qi, B., Ólafsson, S., Zakharov, A.A., Göthelid, M., Agnarsson, B., Gislason, H.P.: Study of spatial homogeneity and nitridation of an Al nanopattern template with spectroscopic photoemission and low energy electron microscopy. Appl. Surf. Sci. 264, 349–357 (2013)Google Scholar
  410. 410.
    Mundschau, M., Bauer, E., Swiech, W.: Initial epitaxial growth of Cu on Mo(011) by low energy electron microscopy and photoemission electron microscopy. J. Appl. Phys. 65, 581–584 (1989)Google Scholar
  411. 411.
    Altman, M.S., Chung, W.F., Liu, C.H.: LEEM phase contrast. Surf. Rev. Lett. 5, 1129–1141 (1998)Google Scholar
  412. 412.
    Shimizu, H., Nakaguchi, A., Yasue, T., Koshikawa, T., Bauer, E.: Abnormal contrast of PEEM images on Cu/W(110). In: Proc. atomic level characterizations for new materials and devices (ALC’05), Big Island, Hawaii, USA 2005, pp. 702–709. JSPS 141 Activity Report (2006)Google Scholar
  413. 413.
    Chung, W.F., Feng, Y.J., Poon, H.C., Chan, C.T., Tong, S.Y., Altman, M.S.: Layer spacings in coherently strained epitaxial metal films. Phys. Rev. Lett. 90, 216105, 4 pages (2003)Google Scholar
  414. 414.
    Altman, M.S.: Low energy electron microscopy of quantum well resonances in Ag films on W(110). J. Phys. Condens. Matter 17, 1305–1310 (2005)Google Scholar
  415. 415.
    Smith, N.V., Brookes, N.B., Chang, Y., Johnson, P.D.: Quantum-well and tight-binding analyses of spin-polarized photoemission from Ag/Fe(001) overlayers. Phys. Rev. B 49, 332–338 (1994)Google Scholar
  416. 416.
    McCarty, K.F.: Deterministic positioning of three-dimensional structures on a substrate by film growth. Nano Lett. 6, 858–861 (2006)Google Scholar
  417. 417.
    Aballe, L., Barinov, A., Locatelli, A., Heun, S., Cherifi, S., Kiskinova, M.: Spectromicroscopy of ultrathin Pd films on W(110): interplay of morphology and electronic structure. Appl. Surf. Sci. 238, 138–142 (2004)Google Scholar
  418. 418.
    Santos, B., Puerta, J.M., Cerda, J.I., Herranz, T., McCarty, K.F., de la Figuera, J.: Structure of ultrathin Pd films determined by low-energy electron microscopy and diffraction. New J. Phys. 12, 023023, 21 pages (2010)Google Scholar
  419. 419.
    Aballe, L., Barinov, A., Locatelli, A., Mentes, T.O., Kiskinova, M.: Initial stages of heteroepitaxial Mg growth on W(110): early condensation, anisotropic strain, and self-organized patterns. Phys. Rev. B 75, 115411, 9 pages (2007)Google Scholar
  420. 420.
    Amakawa, R., Shimizu, H., Yasue, T., Koshikawa, T., Bauer, E.: Dynamic observation of nano-structure growth by LEEM and PEEM. In: Proc. 4th international symposium on atomic level characterization of new materials and devices (ALC’03), Kauai, Hawaii 2003, pp. 573–576. JSPS 141 Activity Report (2004)Google Scholar
  421. 421.
    Santos, B., Puerta, J.M., Cerda, J.I., Stumpf, R., von Bergmann, K., Wiesendanger, R., Bode, M., McCarty, K.F., de la Figuera, J.: Structure and magnetism of ultra-thin chromium layers on W(110). New J. Phys. 10, 013005, 16 pages (2008)Google Scholar
  422. 422.
    McCarty, K.F., Hamilton, J.C., Sato, Y., Saá, A., Stumpf, R., de la Figuera, J., Thürmer, K., Jones, F., Schmid, A.K., Talin, A.A., Bartelt, N.C.: How metal films de-wet substrates-identifying the kinetic pathways and energetic driving forces. New J. Phys. 11, 043001, 38 pages (2009)Google Scholar
  423. 423.
    Ling, W.L., Giessel, T., Thürmer, K., Hwang, R.Q., Bartelt, N.C., McCarty, K.F.: Crucial role of substrate steps in de-wetting of crystalline thin films. Surf. Sci. 570, L297–L303 (2004)Google Scholar
  424. 424.
    Ling, W.L., Bartelt, N.C., McCarty, K.F.: Twin boundaries can be moved by step edges during film growth. Phys. Rev. Lett. 95, 166105, 4 pages (2005)Google Scholar
  425. 425.
    Mascaraque, A., Mentes, T.O., Mentes, T.O., McCarty, K.F., Marco, J.F., Schmid, A.K., Locatelli, A., de la Figuera, J.: Valence band circular dichroism in non-magnetic Ag/Ru(0001) at normal emission. J. Phys. Condens. Matter 23, 305006, 5 pages (2011)Google Scholar
  426. 426.
    El Gabaly, F., Puerta, J.M., Klein, C., Saa, A., Schmid, A.K., McCarty, K.F., Cerda, J.I., de la Figuera, J.: Structure and morphology of ultrathin Co/Ru(0001) films. New J. Phys. 9, 80, 23 pages (2007)Google Scholar
  427. 427.
    Herranz, T., Santos, B., McCarty, K.F., de la Figuera, J.: Real-space study of the growth of magnesium on ruthenium. Surf. Sci. 605, 903–911 (2011)Google Scholar
  428. 428.
    Meyer, A., Flege, J.I., Rettew, R.E., Senanayake, S.D., Schmidt, T., Alamgir, F.M., Falta, J.: Ultrathin silver films on Ni(111). Phys. Rev. B 82, 085424, 9 pages (2010)Google Scholar
  429. 429.
    Hashimoto, M., Umezawa, K., Yasue, T., Koshikawa, T.: Growth of Au on Ni(111) Surface studied with low energy electron microscopy. In: Proc. 8th international symposium on atomic level characterization of new materials and devices (ALC’11) Seoul, Korea 2011, pp. 300–302. JSPS 141 Activity Report (2012)Google Scholar
  430. 430.
    Bollmann, T.R.J., van Gastel, R., Zandvliet, H.J.W., Poelsema, B.: Quantum size effects on surfaces without a projected bandgap: Pb/Ni(111). New J. Phys. 13, 103025, 16 pages (2011)Google Scholar
  431. 431.
    Bollmann, T.R.J., van Gastel, R., Zandvliet, H.J.W., Poelsema, B.: Anomalous decay of electronically stabilized lead mesas on Ni(111). Phys. Rev. Lett. 107, 136103, 4 pages (2011)Google Scholar
  432. 432.
    Man, K.L., Guo, Q., Altman, M.S.: Growth and oxidation of Cr films on the W(100) surface. Surf. Sci. 600, 1060–1070 (2006)Google Scholar
  433. 433.
    Man, K.L., Qiu, Z.Q., Altman, M.S.: Kinetic limitations in electronic growth of Ag films on Fe(100). Phys. Rev. Lett. 93, 236104, 4 pages (2004)Google Scholar
  434. 434.
    Man, K.L., Qiu, Q., Altman, M.S.: Quantum size effect-driven thermal decomposition of Ag films on Fe(100) in the presence of pinhole growth morphological defects. Phys. Rev. B 81, 045426, 10 pages (2010)Google Scholar
  435. 435.
    Tersoff, J., Denier van der Gon, A.W., Tromp, R.M.: Critical island size for layer-by-layer growth. Phys. Rev. Lett. 72, 266–269 (1994)Google Scholar
  436. 436.
    Juel, M., Samuelsen, B.T., Kildemo, M., Raaen, S.: Valence variations of Sm on polycrystalline Ag. Surf. Sci. 600, 1155–1159 (2006)Google Scholar
  437. 437.
    Sogo, M., Kamada, T., Aoki, M., Masuda, S.: Initial oxidation of Ni(111) observed by electron emission microscopy: PEEM and MEEM. Surf. Sci. 601, 3988–3991 (2007)Google Scholar
  438. 438.
    Kamada, T., Sogo, M., Aoki, M., Masuda, S.: Mesoscopic pattern formation during initial oxidation of Ni(111) observed by electron emission microscopy. Surf. Sci. 602, 724–732 (2008)Google Scholar
  439. 439.
    Flege, J.I., Meyer, A., Falta, J.: Self-limited oxide formation in Ni(111) oxidation. Phys. Rev. B 84, 115441, 10 pages (2011)Google Scholar
  440. 440.
    Devlin, C.L.H., Sato, Y., Chiang, S.: Morphology of the Cu2O surface oxide phase formed on Cu(100) at high temperature. J. Appl. Phys. 105, 123534, 5 pages (2009)Google Scholar
  441. 441.
    Günther, S., Reichelt, R., Wintterlin, J., Barinov, A., Menteş, T.O., Niño, M.Á., Locatelli, A.: Chemical patterning of Ag(111): spatially confined oxide formation induced by electron beam irradiation. Appl. Phys. Lett. 93, 233117, 3 pages (2008)Google Scholar
  442. 442.
    Günther, S., Bocklein, S., Reichelt, R., Wintterlin, J., Barinov, A., Mentes, T.O., Nino, M.A., Locatelli, A.: Surface patterning of silver using an electron- or photon-assisted oxidation reaction. Chem. Phys. Chem. 11, 1525–1532 (2010)Google Scholar
  443. 443.
    Böttcher, A., Krenzer, B., Conrad, H., Niehus, H.: Mesoscopic-scale pattern formation induced by oxidation of Ru(0001). Surf. Sci. Lett. 466, L811–L820 (2000)Google Scholar
  444. 444.
    Böttcher, A., Krenzer, B., Conrad, H., Niehus, H.: Mesoscopic-scale growth of oxygen-rich films on Ru(0001) investigated by photoemission electron microscopy. Surf. Sci. 504, 42–58 (2002)Google Scholar
  445. 445.
    Flege, J.I., Sutter, P.: Nanoscale analysis of Ru(0001) oxidation using low-energy and photoemission electron microscopy. J. Phys. Condens. Matter 21, 314018, 6 pages (2009)Google Scholar
  446. 446.
    Pierce, J.P., McCarty, K.F.: Self-assembly and dynamics of oxide nanorods on NiAl(110). Phys. Rev. B 71, 125428, 11 pages (2005)Google Scholar
  447. 447.
    McCarty, K.F., Pierce, J.P., Carter, C.B.: Translation-related domain boundaries form to relieve strain in a thin alumina film on NiAl (110). Appl. Phys. Lett. 88, 141902, 4 pages (2006)Google Scholar
  448. 448.
    Pierce, J.P., Bartelt, N.C., Stumpf, R., McCarty, K.F.: Stability of ultrathin alumina layers on NiAl(110). Phys. Rev. B 77, 195438, 7 pages (2008)Google Scholar
  449. 449.
    Aballe, L., Barinov, A., Locatelli, A., Heun, S., Kiskinova, M.: Tuning surface reactivity via electron quantum confinement. Phys. Rev. Lett. 93, 196103, 4 pages (2004)Google Scholar
  450. 450.
    Aballe, L., Barinov, A., Stojić, N., Binggeli, N., Menteş, T.O., Locatelli, A., Kiskinova, M.: The electron density decay length effect on surface reactivity. J. Phys. Condens. Matter 22, 015001, 6 pages (2010)Google Scholar
  451. 451.
    Santos, B., Loginova, E., Mascaraque, A., Schmid, A.K., McCarty, K.F., de la Figuera, J.: Structure and magnetism in ultrathin iron oxides characterized by low energy electron microscopy. J. Phys. Condens. Matter 21, 314011, 10 pages (2009)Google Scholar
  452. 452.
    Monti, M., Santos, B., Mascaraque, A., de la Fuente, O.R., Niño, M.Á., Menteş, T.O., Locatelli, A., McCarty, K.F., Marco, J.F., de la Figuera, J.: Magnetism in nanometer-thick magnetite. Phys. Rev. B 85, 020404, 5 pages (2012)Google Scholar
  453. 453.
    Monti, M., Santos, B., Mascaraque, A., de la Rodriguez Fuente, O., Niño, M.Á., Menteş, T.O., Locatelli, A., McCarty, K.F., Marco, J.F., de la Figuera, J.: Oxidation pathways in bicomponent ultrathin iron oxide films. J. Phys. Chem. C 116, 11539–11547 (2012)Google Scholar
  454. 454.
    Kaemena, B., Senanayake, S.D., Meyer, A., Sadowski, J.T., Falta, J., Flege, J.I.: Growth and morphology of ceria on ruthenium (0001). J. Phys. Chem. C 117, 221–232 (2013)Google Scholar
  455. 455.
    Skala, T., Tsud, N., Niño Orti, M.Á., Menteş, T.O., Locatelli, A., Prince, K.C., Matolın, V.: In situ growth of epitaxial cerium tungstate (100) thin films. Phys. Chem. Chem. Phys. 13, 7083–7089 (2011)Google Scholar
  456. 456.
    Grinter, D.C., Yim, C.-M., Pang, C.L., Santos, B., Menteş, T.O., Locatelli, A., Thornton, G.: Oxidation state imaging of ceria island growth on Re(0001). J. Phys. Chem. C 117, 16509–16514 (2013)Google Scholar
  457. 457.
    Agnoli, S., Menteş, T.O., Niño, M.Á., Locatelli, A., Granozzi, G.: A LEEM/micro-LEED investigation of phase transformations in TiOx/Pt(111) ultrathin films. Phys. Chem. Chem. Phys. 11, 3727–3732 (2009)Google Scholar
  458. 458.
    Herranz, T., McCarty, K.F., Santos, B., Monti, M., de la Figuera, J.: Real space observations of magnesium hydride formation and decomposition. Chem. Mater. 22, 1291–1293 (2010)Google Scholar
  459. 459.
    Goriachko, A., Zakharov, A.A., Over, H.: Oxygen-etching of h-BN/Ru(0001) nanomesh on the nano- and mesoscopic scale. J. Phys. Chem. C 112, 10423–10427 (2008)Google Scholar
  460. 460.
    Vinogradov, N.A., Zakharov, A.A., Ng, M.L., Mikkelsen, A., Lundgren, E., Mårtensson, N., Preobrajenski, A.B.: One-dimensional corrugation of the h-BN monolayer on Fe(110). Langmuir 28, 1775–1781 (2012)Google Scholar
  461. 461.
    Yasufuku, H., Okumura, M., Kera, S., Okudaira, K.K., Harada, Y., Ueno, N.: PEEM and MEEM of chloroaluminum phthalocyanine ultrathin film on MoS2. J. Electron. Spectrosc. Rel. Phenom. 114–116, 1025–1030 (2001)Google Scholar
  462. 462.
    Yasufuku, H., Ibe, T., Okumura, M., Kera, S., Okudaira, K.K., Harada, Y., Ueno, N.: Diffusion of chloroaluminum phthalocyanine on MoS2 surface detected by photoemission electron microscopy and metastable electron emission microscopy. J. Appl. Phys. 90, 213–216 (2001)Google Scholar
  463. 463.
    Yamamoto, I., Matsuura, N., Mikamori, M., Yamamoto, R., Yamada, T., Miyakubo, K., Ueno, N., Munakata, T.: Imaging of electronic structure of lead phthalocyanine films studied by combined use of PEEM and Micro-UPS. Surf. Sci. 602, 2232–2237 (2008)Google Scholar
  464. 464.
    Baba, Y., Sekiguchi, T., Shimoyama, I., Honda, M., Hirao, N., Narita, A., Deng, J.Z.: Real-time observation on surface diffusion and molecular orientations for phthalocyanine thin films at nanometer spacial resolution. Surf. Sci. 603, 2612–2618 (2009)Google Scholar
  465. 465.
    Meyer zu Heringdorf, F.-J., Reuter, M.C., Tromp, R.M.: Growth dynamics of pentacene thin films. Nature 412, 517–520 (2001)Google Scholar
  466. 466.
    Meyer zu Heringdorf, F.-J., Reuter, M.C., Tromp, R.M.: The nucleation of pentacene thin films. Appl. Phys. A 78, 787–791 (2004)Google Scholar
  467. 467.
    Weidkamp, K.P., Tromp, R.M., Hamers, R.J.: Epitaxial growth of large pentacene crystals on Si(001) surfaces functionalized with molecular monolayers. J. Phys. Chem. C 111, 16489–16497 (2007)Google Scholar
  468. 468.
    Sadowski, J.T., Sazaki, G., Nishikata, S., Al-Mahboob, A., Fujikawa, Y., Nakajima, K., Tromp, R.M., Sakurai, T.: Single-nucleus polycrystallization in thin film epitaxial growth. Phys. Rev. Lett. 98, 046104, 4 pages (2007)Google Scholar
  469. 469.
    Nishikata, S., Sazaki, G., Sadowski, J.T., Al-Mahboob, A., Nishihara, T., Fujikawa, Y., Suto, S., Sakurai, T., Nakajima, K.: Polycrystalline domain structure of pentacene thin films epitaxially grown on a hydrogen-terminated Si(111) surface. Phys. Rev. B 76, 165424, 10 pages (2007)Google Scholar
  470. 470.
    Al-Mahboob, A., Sadowski, J.T., Nishihara, T., Fujikawa, Y., Xue, Q.K., Nakajima, K., Sakurai, T.: Epitaxial structures of self-organized, standing-up pentacene thin films studied by LEEM and STM. Surf. Sci. 601, 1304–1310 (2007)Google Scholar
  471. 471.
    Al-Mahboob, A., Fujikawa, Y., Sadowski, J.T., Hashizume, T., Sakurai, T.: Formation of giant crystalline grain via delayed growth process driven by organic molecular anisotropy. Phys. Rev. B 82, 235421, 6 pages (2010)Google Scholar
  472. 472.
    Al-Mahboob, A., Sadowski, J.T., Fujikawa, Y., Nakajima, K., Sakurai, T.: Kinetics-driven anisotropic growth of pentacene thin films. Phys. Rev. B 77, 035426, 6 pages (2008)Google Scholar
  473. 473.
    Al-Mahboob, A., Sadowski, J.T., Fujikawa, Y., Sakurai, T.: The growth mechanism of pentacene-fullerene heteroepitaxial films. Surf. Sci. 603, L53–L56 (2009)Google Scholar
  474. 474.
    Liu, H.W., Al-Mahboob, A., Fujikawa, Y., Fukui, N., Hitosugi, T., Hashizume, T., Xue, Q.K., Sakurai, T.: Pentacene growth on graphite investigated by low-energy electron microscope. J. Cryst. Growth 312, 967–970 (2010)Google Scholar
  475. 475.
    Ng, M.L., Preobrajenski, A.B., Zakharov, A.A., Vinogradov, A.S., Krasnikov, S.A., Cafolla, A.A., Mårtensson, N.: Effect of substrate nanopatterning on the growth and structure of pentacene films. Phys. Rev. B 81, 115449, 7 pages (2010)Google Scholar
  476. 476.
    Wall, S., Thien, D., Meyer zu Heringdorf, F.-J.: The interplay of topography and energy dissipation in pentacene thin films. J. Electron. Spectrosc. Rel. Phenom. 185, 436–440 (2012)Google Scholar
  477. 477.
    Yoshikawa, G., Sadowski, J.T., Al-Mahboob, A., Fujikawa, Y., Sakurai, T.: Spontaneous aggregation of pentacene molecules and its influence on field effect mobility. Appl. Phys. Lett. 90, 251906, 3 pages (2007)Google Scholar
  478. 478.
    Tsuruma, Y., Al-Mahboob, A., Ikeda, S., Sadowski, J.T., Yoshikawa, G., Fujikawa, Y., Sakurai, T., Saiki, K.: Real-time observation and control of pentacene film growth on an artificially structured substrate. Adv. Mater. 21, 4996–5000 (2009)Google Scholar
  479. 479.
    Al-Mahboob, A., Sadowski, J.T., Nishihara, T., Fujikawa, Y., Xue, Q.K., Nakajima, K., Sakurai, T.: Mechanism of two-dimensional chiral growth of 6,13-pentacenequinone thin films on Si(111). Surf. Sci. 601, 1311–1318 (2007)Google Scholar
  480. 480.
    Marchetto, H., Groh, U., Schmidt, T., Fink, R., Freund, H.J., Umbach, E.: Influence of substrate morphology on organic layer growth: PTCDA on Ag(111). Chem. Phys. 325, 178–184 (2006)Google Scholar
  481. 481.
    Buckanie, N.M., Meyer zu Heringdorf, F.-J.: Photoemission electron microscopy study of anthracene growth on Si(111). Surf. Sci. 601, 1701–1704 (2007)Google Scholar
  482. 482.
    Buckanie, N.M., Meyer zu Heringdorf, F.-J.: Exciton sensitive microscopy of anthracene monolayers on Si(111). Org. Electron. 10, 446–452 (2009)Google Scholar
  483. 483.
    Casu, M.B., Biswas, I., Nagel, M., Nagel, P., Schuppler, S., Chassé, T.: Photoemission electron microscopy of diindenoperylene thin films. Phys. Rev. B 78, 075310, 7 pages (2008)Google Scholar
  484. 484.
    Casu, M.B., Schuster, B.-E., Biswas, I., Nagel, M., Nagel, P., Schuppler, S., Chassé, T.: From interfaces to surfaces: soft x-ray spectromicroscopy investigations of diindenoperylene thin films on gold. J. Phys. Condens. Matter 21, 314017, 5 pages (2009)Google Scholar
  485. 485.
    Casu, M.B., Schuster, B.-E., Biswas, I., Raisch, C., Marchetto, H., Schmidt, T., Chassé, T.: Locally resolved core-hole screening, molecular orientation, and morphology in thin films of diindenoperylene deposited on Au(111) single crystals. Adv. Mater. 22, 3740–3744 (2010)Google Scholar
  486. 486.
    Casu, M.B., Savu, S.-A., Hoffmann, P., Schuster, B.-E., Menteş, T.O., Niño, M.A., Locatelli, A., Chassé, T.: Direct observation of step-edge barrier effects and general aspects of growth processes: morphology and structure in diindenoperylene thin films deposited on Au(100) single crystals. CrystEngComm. 13, 4139–4144 (2011)Google Scholar
  487. 487.
    Khokhar, F.S., van Gastel, R., Schwarz, D., Zandvliet, H.J.W., Poelsema, B.: A low energy electron microscopy study of the initial growth, structure, and thermal stability of 4,4′-biphenyldicarboxylic acid domains on Cu(001). J. Chem. Phys. 135, 124706, 7 pages (2011)Google Scholar
  488. 488.
    Schwarz, D., van Gastel, R., Zandvliet, H.J.W., Poelsema, B.: Phase transformations of 4,4'-biphenyldicarboxylic acid on Cu(001). Phys. Rev. B 85, 235419, 10 pages (2012)Google Scholar
  489. 489.
    Schwarz, D., van Gastel, R., Zandvliet, H.J.W., Poelsema, B.: Size fluctuations of near critical nuclei and gibbs free energy for nucleation of BDA on Cu(001). Phys. Rev. Lett. 109, 016101, 4 pages (2012)Google Scholar
  490. 490.
    Schwarz, D., van Gastel, R., Zandvliet, H.J.W., Poelsema, B.: Formation and decay of a compressed phase of 4,4′-biphenyldicarboxylic acid on Cu(001). Phys. Chem. Chem. Phys. 15, 5007–5016 (2013)Google Scholar
  491. 491.
    Schwarz, D., van Gastel, R., Zandvliet, H.J.W., Poelsema, B.: Growth anomalies in supramolecular networks: 4,4′-biphenyldicarboxylic acid on Cu(001). Phys. Rev. Lett. 110, 076101, 5 pages (2013)Google Scholar
  492. 492.
    Schwarz, D., van Gastel, R., Zandvliet, H.J.W., Poelsema, B.: In situ observation of a deprotonation-driven phase transformation: 4,4′-biphenyldicarboxylic acid on Au(111). J. Phys. Chem. C 117, 1020–1029 (2013)Google Scholar
  493. 493.
    Khokhar, F.S., van Gastel, R., Poelsema, B.: Role of topographical defects in organic film growth of 4,4′-biphenyldicarboxylic acid on graphene: a low-energy electron microscopy study. Phys. Rev. B 82, 205409, 8 pages (2010)Google Scholar
  494. 494.
    Hlawacek, G., Khokhar, F.S., van Gastel, R., Poelsema, B., Teichert, C.: Smooth growth of organic semiconductor films on graphene for high-efficiency electronics. Nano Lett. 11, 333–337 (2011)Google Scholar
  495. 495.
    Hlawacek, G., Khokar, F.S., van Gastel, R., Teichert, C., Poelsema, B.: Diffusion and submonolayer growth of para-sexiphenyl on Ir(111) and Ir(111)-supported graphene. IBM J. Res. Dev. 55(4), 15:1–15:7 (2011)Google Scholar
  496. 496.
    Khokhar, F.S., Hlawacek, G., van Gastel, R., Zandvliet, H.J.W., Teichert, C., Poelsema, B.: The influence of substrate temperature on growth of para-sexiphenyl thin films on Ir{111} supported graphene studied by LEEM. Surf. Sci. 606, 475–480 (2012)Google Scholar
  497. 497.
    Fleming, A.J., Netzer, F.P., Ramsey, M.G.: Nucleation and 3D growth of para-sexiphenyl nanostructures from an oriented 2D liquid layer investigated by photoemission electron microscopy. J. Phys. Condens. Matter 21, 445003, 11 pages (2009)Google Scholar
  498. 498.
    Wagner, T., Fritz, D.R., Zeppenfeld, P.: Standing and flat lying alpha-6T molecules probed by imaging photoelectron spectroscopy. Org. Electron. 12, 442–446 (2011)Google Scholar
  499. 499.
    Wagner, T., Fritz, D.R., Zeppenfeld, P.: α-6T on Ag(110): the formation of the wetting layer. Synth. Met. 161, 2006–2010 (2011)Google Scholar
  500. 500.
    Wagner, T., Fritz, D.R., Zeppenfeld, P.: Probing organic nanostructures by photoelectron-emission microscopy. Appl. Surf. Sci. 267, 26–29 (2013)Google Scholar
  501. 501.
    Christensen, S., Lanke, U.D., Haines, B., Qaqish, S.E., Paige, M.F., Urquhart, S.G.: Structural and compositional mapping of a phase-separated Langmuir-Blodgett monolayer by X-ray photoelectron emission microscopy. J. Electron. Spectrosc. Rel. Phenom. 162, 107–114 (2008)Google Scholar
  502. 502.
    Qaqish, S.E., Urquhart, S.G., Lanke, U., Brunet, S.M.K., Paige, M.F.: Phase separation of palmitic acid and perfluorooctadecanoic acid in mixed Langmuir-Blodgett monolayer films. Langmuir 25, 7401–7409 (2009)Google Scholar
  503. 503.
    Leung, B.O., Hitchcock, A.P., Won, A., Ianoul, A., Scholl, A.: Imaging interactions of cationic antimicrobial peptides with model lipid monolayers using X-ray spectromicroscopy. Eur. Biophys. J. 40, 805–810 (2011)Google Scholar
  504. 504.
    Cossy-Favre, A., Diaz, J., Anders, S., Padmore, H.A., Liu, Y., Samant, M., Stöhr, J., Brown, H., Russell, T.P.: Photoelectron emission microscopy and its application to the study of polymer surfaces. In: Proc. third international school and symposium on synchrotron radiation in natural science, Jaszowiec, Poland 1996. Acta Physica Polonica A 91 (5) 10 pages (1997)Google Scholar
  505. 505.
    Cossy-Favre, A., Díaz, J., Liu, Y., Brown, H.R., Samant, M.G., Stöhr, J., Hanna, A.J., Anders, S., Russell, T.P.: X-PEEM study on surface orientation of stylized and rubbed polyimides. Macromolecules 31, 4957–4962 (1998)Google Scholar
  506. 506.
    Zharnikov, M., Ouchi, Y., Hasegawa, M., Scholl, A.: X-ray absorption spectromicroscopy study of UV-photoinduced surface modification and anisotropy in polyimide films. J. Phys. Chem. B 108, 859–863 (2004)Google Scholar
  507. 507.
    Kanai, K., Miyazaki, T., Wakita, T., Akaike, K., Yokoya, T., Ouchi, Y., Seki, K.: Lateral inhomogeneity in the electronic structure of a conjugated poly(3-hexylthiophene) thin film. Adv. Funct. Mater. 20, 2046–2052 (2010)Google Scholar
  508. 508.
    Kappen, P., Hale, P.S., Brack, N., Prissanaroon, W., Pigram, P.J.: X-PEEM/NEXAFS and AFM of polypyrrole and copper micro-patterns on insulating fluoropolymer substrates. Appl. Surf. Sci. 253, 1473–1479 (2006)Google Scholar
  509. 509.
    Morin, C., Ikeura-Sekiguchi, H., Tyliszczak, T., Cornelius, R., Brash, J.L., Hitchcock, A.P., Scholl, A., Nolting, F., Appel, G., Winesett, D.A., Kaznacheyev, K., Ade, H.: X-ray spectromicroscopy of immiscible polymer blends: polystyrene-poly(methyl methacrylate). J. Electron. Spectrosc. Rel. Phenom. 121, 203–224 (2001)Google Scholar
  510. 510.
    Morin, C., Hitchcock, A.P., Cornelius, R.M., Brash, J.L., Urquhart, S.G., Scholl, A., Doran, A.: Selective adsorption of protein on polymer surfaces studied by soft X-ray photoemission electron microscopy. J. Electron. Spectrosc. Rel. Phenom. 137–140, 785–794 (2004)Google Scholar
  511. 511.
    Li, L., Hitchcock, A.P., Robar, N., Cornelius, R., Brash, J.L., Scholl, A., Doran, A.: X-ray microscopy studies of protein adsorption on a phase-segregated polystyrene/polymethyl methacrylate surface. 1. Concentration and exposure-time dependence for albumin adsorption. J. Phys. Chem. B 110, 16763–16773 (2006)Google Scholar
  512. 512.
    Li, L., Hitchcock, A.P., Cornelius, R., Brash, J.L., Scholl, A., Doran, A.: X-ray microscopy studies of protein adsorption on a phase segregated polystyrene/polymethylmethacrylate surface. 2. Effect of pH on site preference. J. Phys. Chem. B 112, 2150–2158 (2008)Google Scholar
  513. 513.
    Leung, B.O., Hitchcock, A.P., Brash, J.L., Scholl, A., Doran, A.: Phase segregation in polystyrene-polylactide blends. Macromolecules 42, 1679–1684 (2009)Google Scholar
  514. 514.
    Leung, B.O., Hitchcock, A.P., Cornelius, R.M., Brash, J.L., Scholl, A., Doran, A.: Using X-PEEM to study biomaterials: protein and peptide adsorption to a polystyrene-poly(methyl methacrylate)-b-polyacrylic acid blend. J. Electron. Spectrosc. Rel. Phenom. 185, 406–416 (2012)Google Scholar
  515. 515.
    Ceccone, G., Leung, B.O., Perez-Roldan, M.J., Valsesia, A., Colpo, P., Rossi, F., Hitchcock, A.P., Scholl, A.: X-ray spectromicroscopy study of ubiquitin adsorption to plasma polymerized microstructures. Surf. Interface Anal. 42, 830–834 (2010)Google Scholar
  516. 516.
    Hitchcock, A.P., Stöver, H.D.H., Croll, L.M., Childs, R.F.: Chemical mapping of polymer microstructure using soft X-ray spectromicroscopy. Aust. J. Chem. 58, 423–432 (2005)Google Scholar

Copyright information

© Springer Science+Business Media New York 2014

Authors and Affiliations

  • Ernst Bauer
    • 1
  1. 1.Department of PhysicsArizona State UniversityTempeUSA

Personalised recommendations