The atom probe microscope provides three-dimensional ...

8 downloads 371 Views 132KB Size Report
Atom probe development. The precursor to the atom probe microscope was the field ion microscope (FIM), which was origi- nally developed by Erwin Müller in ...
atom.qxd

1/17/04

4:53 PM

Page 1

ATOM PROBE ANALYSIS The atom probe microscope provides three-dimensional compositional and structural analysis at the atomicand near-atomic scales. Amy A. Gribb Thomas F. Kelly Imago Scientific Instruments Corporation Madison, Wisconsin

W

hen introducing the atom probe microscope, it is important to first address a common misconception that the atom probe microscope is related to the more widely known atomic force microscope (AFM), which operates by scanning a sharp tip across surfaces and provides atomicscale imaging. In contrast, the atom probe microscope operates by removing and analyzing individual atoms. The atom probe’s unique atom-by-atom analysis provides a map of the elemental and isotopic identity and position of individual atoms in volumes of material of up to 100 nm in diameter (parallel to specimen surface) and 100 nm in depth (normal to specimen surface). Thus, the basic differences between the atom probe and AFM are: • The atom probe analyzes three-dimensional volumes of material, whereas the AFM analyzes surface features only. • The atom probe provides both imaging and chemical analysis, whereas AFM provides imaging only. In this article, the operation and analytical output of the atom probe is compared with the more widely known analytical techniques AFM, TEM, and SIMS. An atom probe study of buried interfaces in multilayer thin metal films is presented as an example application of the technique.

that the atom probe microscope can provide. An analytical tool that can provide the equivalent of a ball-and-stick model of atomic structure with elemental and isotopic identification may seem too good to be true, and certainly too good to be relatively obscure. It might be anticipated that the above-mentioned “notable exceptions” will draw the atom probe back into familiar analytical regimes. However, although the exceptions introduce deviations from the ideal analytical results embodied by a ball-and-stick model, the atom probe comes closer than any other technique to combining chemical composition with structure at the atomic scale.

Analytical output Having established the atom probe’s distinctness from AFM, it is now worth restating the type of analytical information available by atom probe microscopy. Consider a ball-and-stick model of the atomic structure, in which different elements and isotopes are represented by balls of different colors. Now extend the model to include up to tens of millions of atoms. With two notable exceptions, this is the information

Atom probe development The precursor to the atom probe microscope was the field ion microscope (FIM), which was originally developed by Erwin Müller in the 1940s. During FIM analysis, gas atoms adsorbed to the surface of a sharply pointed specimen are ionized by means of a strong electric field. They are then accelerated toward a phosphor screen or other position-sensitive detector placed at some distance from the specimen. The image created by the gas

Fig. 1 — FIM image of tungsten specimen. Bright spots indicate individual tungsten atoms on the specimen surface.

ADVANCED MATERIALS & PROCESSES/FEBRUARY 2004

31

atom.qxd

1/17/04

4:54 PM

Page 2

High voltage

Position-sensitive detector

Needle-shaped specimen

Fig. 2 — Schematic of atom probe operation. Individual atoms undergo sequential field evaporation. Position-sensitive detector records time and position of impact.

atoms is a projection of the atomic-scale structure of the specimen surface. In 1955, Müller and a colleague used FIM to collect the first-ever images of individual atoms. Figure 1 shows a FIM image of a tungsten specimen. The bright spots in the image represent individual tungsten atoms. After years of working with FIM, Müller and his group took the technique several steps further by evaporating the specimen atoms themselves from the surfaces of materials, and analyzing these atoms by time-of-flight mass spectrometry. Drs. Müller, Panitz, and McLane published this new process of field evaporating and analyzing actual specimen atoms in 1968, and it was in this paper that the term “atom probe” was first used to describe the new technique. How the atom probe works The basic components of the atom probe are shown schematically in Fig. 2. The atom probe operates by cycling a high voltage pulse between a needle-shaped specimen and an opposing electrode, which in this case is also a position-sensitive detector. The significance of the needle geometry of specimens is twofold: • The sharp needle makes it possible to create the high electric fields at the specimen surface needed to induce field evaporation of atoms. • The highly divergent electric field emanating from the needle is the basis of the projection magnification of the image. For a specimen with a 100 nm tip radius, approximately 10 kilovolts creates sufficient electric field at the specimen tip to pull atoms from the surface in the form of positively charged ions. During atom probe analysis, the magnitude of the electric field is carefully controlled so that one atom at a time leaves the specimen. Note that atoms are ionized prior to evaporation from the specimen surface, such that atoms analyzed during atom probe microscopy are more precisely designated as ions. The process of field evaporating specimen atoms (as positively charged ions) continues until the specimen fails or until the specimen tip becomes too blunt and the applied voltage is insufficient to induce field evaporation. 32

The typical number of ions per analysis varies with the configuration of the atom probe, specimen material type, specimen shape, and other factors. The largest atom probe data sets have been collected on metal, which is the most easily analyzed material type due to its high inherent electrical conductivity. For metal specimens, state-of-the-art atom probe microscopes routinely collect data sets containing more than 50 million ions. After field evaporation, specimen ions follow the electric field lines out to a position-sensitive detector that records both the time and the position of impact. The identity of the atoms (ions) is determined by measuring their flight time, which depends on their mass; and mass determines chemical identity. For this technique to be effective, atoms must be evaporated in a pulsed mode so that the departure time is known. By pulsing the voltage, the time of departure of an ion is known, and so the total time of flight of the ion from the specimen to the detector can be measured. This atom-by-atom mass spectrometry enables the atom probe to analyze composition at the atomic-scale. As previously mentioned, the unique power of atom probe microscopy lies in its ability to tie compositional information to structure. The atom probe achieves this by recording positional information in addition to time-of-flight for each ion analyzed. Positional information collected during atom probe analysis includes the two-dimensional hit position of the ion on the detector, and a one-dimensional sequence number. In a manner analogous to the FIM, ions removed from the specimen surface during atom probe analysis create a highly magnified projection of the atomic-scale structure of the specimen surface on the detector. Thus, the ion’s position in x,y (the plane parallel to the specimen surface) in the original specimen can be calculated from the two-dimensional hit position of the ion on the detector. Further, because atomic layers erode predictably from the specimen surface, a sequence number can be used to calculate an ion’s position in z (direction normal to the specimen surface) with high precision. One of the exceptions to the ball-and-stick model as an analogue for atom probe results should now be apparent. The source of positional information is the two-dimensional projection from the curved surface of the ion’s position in x,y. However, this two-dimensional projection is subject to an error that can reduce lateral (x,y) resolution to several atomic diameters. If the specimen surface were perfectly smooth, then the projection could be arbitrarily accurate. In reality, the surface is atomically rough and may have facets or grooves at certain crystallographic orientations or microstructural defects. It is these imperfections that displace the ions from a perfect projection. The other notable exception to ideal analytical output results from the fact that currently available position-sensitive detectors are based on microchannel plate (MCP) amplifiers as the first line of detection. MCPs detect about 60% of all atoms that

ADVANCED MATERIALS & PROCESSES/FEBRUARY 2004

10:11 AM

Page 3

Depth

1/26/04

0.2

a. Ideal

0.4 0.6 Composition

b. TEM

c. SIMS

0.8

d. Atom probe

Fig. 3 — Schematic representations of material structure and analytical results. 3a shows actual structure and hypothetical ideal analytical result for a two-component material. 3b, c, and d show stylized representation of analytical information available by TEM, SIMS, and atom probe, respectively.

strike them with equal probability. Thus, the atom probe records approximately 60% of the ions that evaporate from the specimen. (More detailed explanations of the atom probe technique may be found in texts by Miller et al.) Comparison with TEM and SIMS Although the atom probe microscope does not yield a perfect atomic-scale model of materials, it comes closer to achieving this ideal than other techniques. To illustrate this point, consider a hypothetical material comprised of two elements, as shown in Fig. 3a. Simplified representations of analytical results available from TEM, SIMS, and atom probe analysis of the material are presented to point out the basic differences in analytical output. As shown in Fig. 3b, TEM can provide accurate atomic-scale structure in projection, but only average composition. The TEM’s ability to resolve crystal structure derives from its mode of imaging, namely by electron diffraction from an intact specimen. Compositional analysis may be combined with TEM by ancillary techniques such as electron energy loss spectroscopy (EELS). EELS and other indirect compositional analysis techniques provide average composition. SIMS utilizes mass spectrometry for chemical identification of ions, and so provides compositional analysis superior to the indirect techniques used in conjunction with TEM (Fig. 3c). However, the improvement in compositional analysis comes at the cost of structural information. SIMS dislodges specimen atoms for analysis by sputtering. The progress of the sputtering can be controlled to determine the position in z (normal to the specimen surface) of analyzed ions to about 10 atomic layers, but the sputtering process is such that information about an atom’s position in x,y is limited by the spot size of the primary ion beam to about 100 nm. Like SIMS, the atom probe uses mass spectrometry for chemical analysis of specimen ions. Unlike SIMS, the atom probe removes specimen ions in a controlled manner that preserves positional information for individual ions in three dimensions. Spatial resolution of the atom probe can fairly be said to be as good as 0.2 nm in z (normal to the specimen surface) and less than 0.5 nm in x,y (parallel to the specimen surface) in general analyses. Significantly better spatial resolution has been demonstrated in special cases. As illustrated in Fig. 3d, the atom

24 nm

a 16

12

Depth, nm

atom.qxd

Cobaltiron Copper

8

4

b

0

0.2

0.4 0.6 0.8 Concentration, at. fraction

1

Fig. 4 — Results of atom probe analysis of multilayer thin film stack of copper (red) and cobalt-iron (blue) layers.

probe provides more detailed structural information than SIMS, and more detailed compositional information than TEM. Figure 4 shows actual results of atom probe analysis of a multilayer metal thin film stack. Many repeats of copper with cobalt-iron layers approximately 2 nm thick are shown in the 3-D reconstruction of atom probe data. Each dot in the image represents an individual atom. Copper atoms are

ADVANCED MATERIALS & PROCESSES/FEBRUARY 2004

33

atom.qxd

1/17/04

4:54 PM

The atom probe’s unique atom-byatom analysis provides a map of the elemental and isotopic identity and position of individual atoms in volumes of material of up to 100 nm in diameter.

34

Page 4

shown in red, and for clarity, both and cobalt and iron are shown in blue. The copper layers are clearly visible and distinct from the copper-iron layers. Thus, the atom probe emerges as demonstrating unique analytical capabilities that come closest to replicating the ideal represented by a ball-and-stick model of materials when both the compositional and structural components of the model are considered. Atom probe limitations Current limitations of atom probe microscopy are the requirements that specimens possess a certain minimum level of electrical conductivity and have the capacity to be formed into the needleshaped geometry. These limitations increase the difficulty of analyzing low-conductivity materials. Alternatives to voltage pulsing as the means of inducing pulsed field evaporation and improved methods for specimen preparation are under investigation to resolve these limitations. However, the unique information available by atom probe microscopy can provide new insights into structure-property relationships at the atomic- and near■ atomic scales. For more information: Thomas F. Kelly is Chairman, Founder, and Chief Technology Officer at Imago Scientific Instruments, 6300 Enterprise Lane, Madison, WI 53719; tel: 608 / 274-6880; fax: 608 / 442-0622; e-mail: tkelly@ imago.com; Web site: www.imago.com.

Acknowledgements The work shown in Figure 4 was done in collaboration with Peter F. Ladwig and Y. Austin Chang of the University of Wisconsin Madison and David J. Larson, Martin C. Bonsager, Bharat B. Pant, and Allan E. Schultz of Seagate Technology, Inc. Bibliography 1. Z. Naturforsch, by E. W. Müller, Vol. 11a, 1956, p, 88. 2. J. Appl. Phys., by E. W. Müller, Vol. 27, 1956, p. 474. 3. “The atom probe field ion microscope,” by E.W. Müller, J.A. Panitz, and S. B. McClean: Rev. Sci. Instrum., Vol. 39, 1968, p. 83. 4. Atom Probe Tomography, by M.K. Miller: Kluwer Academic/Plenum Publishers, New York, 2000. 5. Atom Probe Field Ion Microscopy, by M.K. Miller, A. Cerezo, M.G. Hetherington, and G.D.W. Smith: Oxford Science Publications, New York, 1996. 6. “First data from a commercial local electrode atom probe (LEAP),” by T.F. Kelly, et al., Microscopy and Microanalysis, accepted for publication, 2004.

How useful did you find the information presented in this article? Very useful: Circle 272 Of general interest: Circle 273 Not useful: Circle 274

Circle 17 or visit www.adinfo.cc