Leap to .
The CNL x-ray scattering research team performs x-ray diffraction experiments on metal, insulator (C60) and semiconductor (III-V's and group IV) surfaces and interfaces. The interest in surfaces and interfaces lies not in merely determining t he atomic coordinates, but in relating that structural information to the physical properties of that surface or interface. There are a number of excellent and complementary methods for obtaining this structural information from a surface; where we use the term surface to refer to a specific type of interface, specifically a vacuum-crystal interface. However, the same strengths that make these techniques well suited for surface studies impede them in interface investigations. For example, electron s used in low-energy electron diffraction (LEED) strongly interact with atoms limiting their penetration depth. This results in only a couple of techniques that can probe and determine the structure of a buried interface.
X-rays are particularly well suited for these investigations. Their characteristic wavelengths, on the order of an angstrom, make them a nearly ideal diffraction probe of atomic dimensions. In addition, their weak interaction with matter not only contributes to their unmat ched penetration power, but also allows straightforward analysis using single scattering kinematic theory. Unfortunately, this strength is also their greatest drawback for use in surface studies. However, augmentation of the low surface signal rates attributable to the x-rays' weak interaction with the "small" number of surface atoms can be achieved by using a glancing incidence geometry and/or with the higher intensity beams available at x-ray synchrotrons.
We primarily use sy nchrotron radiation to perform our scattering experiments. There are many advantages to using synchrotron x-rays over conventional laboratory based sources. The availability of high intensity, tunable monochromatic x-rays or broad spectrum "white" radiation permit numerous experiments once technically formidable or infeasible to become routine. These include extended x-ray absorption fine structure (EXAFS), anomalous or resonant scattering and protein crystallography. The advent of x-ray sync hrotron sources has revolutionized materials research using x-rays.
A qualitative overview of x-ray diffraction is available.
The majority of our experiments to date fall into three main categories. These are:
The structure of the metal/semiconductor interface is a subject of continued interest in gaining a greater fundamental understanding of these systems. The confirmed correlation between macroscopic effects and the atomic structure at an interface only further enhances this interest. The noble metals (Au, Ag and Cu) exhibit different characteristic interactions with Si(111) substrates. Briefly, Cu alloys with Si(111), Au indiffuses into Si(111), while Ag/Si(111) behaves like a proto-typical non-reactive ideal binary system.
We measured the Si(1 0 L)hex truncation rod of these interfaces. Crystal trun cation rod modeling revealed the preservation of the Si(111)-(7×7) stacking fault in the Ag-modified 7×7) interface structure, while a Ag-Si mixed (alloy) layer was found at the annealed Ag/Si(111)-(1×1) interface.4,5 These results may provide some insight into the observed Schottky barrier height difference for these two interfaces.
The abstract for this experiment as well as a comparison graph of the data to the Ag/Si(111) case is still available on the National Synchrotron Ligh t Source (NSLS) on-line version of the 1994 NSLS activity report as Au/Si(111) Anomalous truncation rods.
In the fall of 1991, Prof. J. H. Weaver, presented a physics department colloquium. In it, he extolled the virtues of a wonderful new molecule he and his collaborators had found called Buc kminster fullerene or "Bucky-balls", named after the architect Buckminster Fuller who designed geodesic domes. The reason stems from the molecular structure of C60 being similar to the appearance of a soccer ball with C atoms at the intersections of the pentagons and hexagons; the is the same pattern which comprises a geodesic dome. Of all of the characteristics he mentioned, the one that intrigued us was C60 's inert nature.
This fact in conjunction with it being a van der Waals molecular solid, made it a good candidate to use for the encapsulation or preservation of clean surface atomic structures which can only be generated and otherwise kept clean in an ultra-high vacuum (UHV) enviroment. Even in UHV, clean surfaces usually deteriorate by residual gas contamination in a matter of a few hours because of the presence of chemically active dangling bonds on the surface. This contamination problem and the need for bulky UHV chambers to carry out many surface science experiments (including surface x-ray diffraction) lead to severe constraints. It would be highly desirable to find an inert capping material suitable for preserving the clean surface structure. This inert cap (or insulating layer) might also prove relevant in electronic device applications, especially as devices continue to shrink. The structural and chemical integrity of the substrate under such a cap would be highly desirable.
The first candidate for this test was the Si(1 11)-(7×7) clean surface reconstruction. The structure of this surface is described by the dimer-adatom-stacking fault model of Takayanagi et al. 6 This surface is highly reactive. Most elemental materials deposited on it react or alloy with the substrate, resulting in drastic structural and chemical modifications. All previous attempts to cap this surface have failed to preserve the adatom feature, and in only a few cases has the stacking fault been retained.
We discovered the Si(111)-(7×7) clean surface structure was preserved at the interface.7 In fact, the correlation between the diffracted intensities from the fractional-order peaks was surprisingly similar to that of those measured in UHV by Robinson et al.8 A difference Fourier map highlighting the difference between the measured and proposed simplified stacking-fault structure revealed additional peaks at the adatom sites indicating these features needed to be added to the model structure. Subsequent experiments have shown the buried interface retains the periodicities of the dimerized Si(001)-(2×1)9 and Ge(001)-(2×1)10 clean surface reconstructions. These successful preservations buoyed our hopes that this was a universal property of C60 interfa ces.
The abstract as well as a surface diffraction rod graph of the C60/Ge(001) data is still available on the NSLS on-line version of the 1994 NSLS activity report as C60 /Ge(100) boundary structure
Our own studies as well as further investigations into the characteristics of C60 films has proven the interaction between C60 and various substrates, including the above mentioned substrates, is beyond mere van der Waals bonding. This was supported by the recent discovery that the Ge(111)-c(2×8) adatom reconstruction transformed to a Ge(111)-(1×1) structure after C60 adsorption.11 This reversion to a bulk-like (1×1) structure at the C60 interface is especially true for metal substrates and adsorbate-induced semiconductor reconstructions where C60 lifts the surface reconstruction. C60 encapsulation of other reconstructed surfaces is currently inconclusive.
Our experience, along with that of other groups, of the Herculean efforts
that are necessary to remove the last monolayer of C60 from
various surfaces coupled with the near invulnerability of C60
prompted us to consider the atmospheric durability of the C60
(4×4) monolayer structure on
Cu(111). The stability of this layer to
atmospheric exposure was investigated with reflection high-energy electron
diffraction (RHEED) and x-ray diffraction. The results indicate that the
C60 (4×4) overlayer structure is partially preserved under
the atmospheric adsorbed layer. This (4×4) structural remnant was
metastable and had a limited lifetime. This is consistent with previous
findings, and yield a clue as to the durability of C60 monolayer
structures.
The abstract as well as a linescan of the data is still available on the NSLS on-line version of the 1994 NSLS activity report as Atmospheric durability of a (4×4) C60 monolayer on Cu(111)
The x-ray transmission geometry has a numbe r of strengths. A notable benefit lies in the collection of a significant section of the diffraction pattern all at once. This pattern would yield the same qualitative information as the popular LEED patterns that are nearly indispensable to the surface scientist in his/her investigations. Quantitative analysis of this x-ray pattern would not only be simplified by the use of kinematic scattering theory rather than multiple scattering dynamic theory necessary in LEED, but would also significantly improve the precision of LEED structural determinations. Furthermore, depending upon the complexity of the system, Bragg and truncation rod profiles can be collected in the traditional manner as before.13 Another major benefit lies in the enormous reduction in data collection time. A x-ray crystallography experiment could be carried out in only a fraction of the time currently necessary. This would allow studies of more reactive surfaces and enhance i nvestigations of in situ real-time crystal growth. Unfortunately, there are also notable weaknesses in this technique.
Its rarity in x-ray experiments is not without cause. A major drawback arises from x-ray absorption in the sample. This effect requires careful consideration be given to the film and substrate to be studied, in particular, its composition and thickness. Sample fluorescence is another obstacle which must be considered. Of greater consequence is the bulk scatter ing which in most surface/interface experiments is minimized by working in the grazing-incidence geometry. The limited rod intensity profile that can be collected in a surface/interface experiment is another drawback. For instance, if the sample is fixed, then the energy range determines what portion of the rod can be collected. Furthermore, the complexity of the analysis increases if this method of collection is used over a large enough energy range that absorption and fluorescence, which depend upon the wavelength, become an issue. Nonetheless, we believe these pitfalls can be overcome and the substantial benefits of surface/interface transmission diffraction reaped.
It is in the "Bragg" geometry that perhaps greater rewards lie. This reflection geometry solves both the absorption and bulk scattering problems encountered in the transmission geometry. It is analogous to the orientational transition from LEED to reflection high-energy electron diffraction (RHEED). As in RHEED, this geometry is better suited to certain experiments, such as in situ growth. The primary drawback in this geometry arises from the increased complexity of the analysis.
Finally, we would like to close with a quote from Andrew Zangwill's book Physics at Surfaces in which he writes,
"...bulk structural issues normally are resolved by x-ray diffraction. Unfortunately, the extremely large penetration depth and mean free path of x-rays severely limits the ir routine use for surface crystallography. Consequently, much effort has been devoted to the invention and application of alternative experimental approaches to surface-specific structural analysis. Although a number of common techniques will be discussed below, it is a sobering fact that no single surface structural tool has emerged that can be used as easily and reliably as x-rays are used for the bulk."14We are optimistic th at this technique will bring this hope closer to reality.
The complete text of the transmission diffraction project is available.
X-ray Scattering Research Team
1. F. J. Walker, E. D. Specht, and R. A. McKee, Phys. Rev. Lett. 67, 2818 (1991).
2. Hawoong Hong, R. D. Aburano, D.-S. Lin, H. Chen, T.-C. Chiang, P. Zschack, and E. D. Specht, Phys. Rev. Lett. 68, 507 (1992).
3. See for example, T. Takahashi, S. Nakatani, N. Okamoto, T. Ishikawa, and S. Kikuta, Jpn. J. Appl. Phys . 27, L753 (1988); E. Vlieg, A. W. Denier van der Gon, J. F. van der Veen, J. E. MacDonald, and C. Norris, Surf. Sci. 209, 100 (1989).
4. R. D. Aburano, Hawoong Hong, J. M. Roesler, D.-S. Lin, T.-C. Chiang, and P. Zschack, Surf. Sci. Lett. 339, L891 (1995).
5. R. D. Aburano, Hawoong Hong, J. M. Roesler, K.-S. Chung, D.-S. Lin, P. Zschack, H. Chen, and T.-C. Chiang, Phys. Rev. B 52, 1839 (199 5).
6. K. Takayanagi, Y. Tanishiro, S. Takahashi, and M. Takahashi, Surf. Sci. 164, 367 (1985).
7. Hawoong Hong, W. E. McMahon, P. Zschack, D.-S. Lin, R. D. Aburano, H. Chen, and T.-C. Chiang, Appl. Phys. Lett. 61, 3127 (1992).
8. I. K. Robinson, W. K. Waskiewcz, P. H. Fuoss, and L. J. Norton, Phys. Rev. B 27, 4325 (1988).
9. Hawoong Hong, R. D. Aburano, E. S. Hirschorn, P. Zschack, H. Chen, and T.-C. Chiang, Phys. Rev. B 47, 6450 (1993).
10. R. D. Aburano, Hawoong Hong, K.-S. Chung, M. C. Nelson, H. Chen, T.-C. Chiang, and P. Zschack, (unpublished).
11. R. D. Aburano, Hawoong Hong, P. Zschack, T. Gog, and T.-C. Chiang, (unpublished).
12. R. D. Aburano, Hawoong Hong, M. C. Nelso n, T. Miller, T.-C. Chiang, and P. Zschack, (unpublished).
13. A simple system with no sample motions could accomplish this by changing the energy of the incident x-rays. This would restrict the length of the rod collected to the energy range limit of the monochromator as well as require careful accounting of other factors, such as absorption.
14. A. Zangwill, Physics at Surfaces (Cambridge, New York, 1988) , p. 28.