Introduction

The future of nanotechnology ultimately rests on the controllable fabrication, integration, and stability of nanoscale devices. We are investigating the induced self-organized growth of nanoscale features (with and without an imposed template) in wafer-compatible processes, which are important because they increase the prospects for large-scale manufacturability.

Nanofeatures synthesized within and on solid surfaces hold promise for unprecedented functionality. However, understanding of the fundamental phenomena leading to the formation, stability, and morphological evolution of nanoscale features is lacking. As dimensions shrink into the nanoscale, many classical macroscopic models for morphological evolution lose their validity (see, for example, publication #117, our recent experiment on Si(001): J.D. Erlebacher, M.J. Aziz, et al. "Non-Classical Smoothening of Nano-Scale Surface Corrugations", Phys. Rev. Lett. 84, 5800 (2000); http://www.deas.harvard.edu/matsci/people/aziz/azizp1.html#117). We are studying, through a combination of experiment and modeling, the configurational forces and laws governing mass transport involved in the synthesis and organization of nanoscale features in solid state materials.

Fig. 1.  The first three pictures are a simulation of self-organized nanodot formation evolving from the gridded template on the left. The specific model (Z. Suo and W. Lu, J. Nanoparticles Res., in press) is for phase separation in a surface alloy monolayer, but its general features may apply to many spontaneous pattern formation processes.  The pattern that evolved is reminiscent of a state-of-the-art SRAM memory cell (small figures at right, 1 "unit cell", from H. Shimizu et al. (Fujitsu), 1999 IEEE International Solid State Circuits Conference).  Will we ever be able to cover entire wafers with nano-SRAM cells without being slowed down by the requirements of lithography?

We are exploring the possibilities using templates to guide the self-organization of these nanofeatures during their growth, with minimal recourse to lithography, thereby creating the possibility for parallel nanofabrication of entire circuits. Figure 1 compares a state-of-the-art SRAM memory cell with a pattern produced by templated self-organized growth; the cell size in self-organized growth can be far smaller than in current technology. From a regular starting template such as the grid shown on the left, a highly organized uniform pattern can form during subsequent processing as a result of surface growth kinetics, guided by the boundary conditions imposed by the template. We expect that self-organized nanostructures can be grown on a variety of planar substrates using wafer-compatible parallel processing techniques by understanding and exploiting a variety of physical and chemical phenomena including elastic stress, surface energy, chemical bonding and alloying, selective reactivity, Ostwald ripening, sputter instabilities, and boundary conditions imposed by templating. The ultimate goal is to develop the science base for rational synthesis of designed structures at the nanoscale which will permit their unique properties to be exploited in future devices.

3D nanoporosity induced by selective electrochemical dissolution

Selective dissolution of the less noble element of a binary alloy leads to 3D nanoporous structures with huge specific surface areas, potentially important for catalysts and sensors. Ligament widths as small as 5 nm have been observed. We are just beginning to understand the physical and chemical origin of this structure, and lots more work needs to be done before we have a really good understanding of what's going on. For a description of some recent insights we've gained into this process, click here.

To see a really cool (but huge, about 93 megabytes) movie (.AVI format) of a Kinetic Monte Carlo simulation of the early stages of this process, click here.
 
 

Ion sputter-induced nano-ripples and nano-dots

Low energy ion sputtering produces arrays of "quantum dots", "quantum dashes", or "quantum wires" on surfaces. We are exploring the mechanisms of and limits to miniaturization and regularity using templating, the interaction of coherency stress and sputter effects in attempts to fabricate heterophase nanofeatures, and the use of the resulting structures in electronic devices and as templates for the assembly of nanoparticles from solution. For a description of what we've done already, click here.
 

Si(111) (above): highly regular ripples on our first try!  From publication #95.	Si(001) (above): irregular ripples: why?  From publication #111.
	Figure at left: Normal incidence on GaSb creates dots instead of ripples. From S. Facsko et al., Science 285, 1551 (1999)

Semiconductor quantum dots

Figure:  Self-organized SiGe quantum dots grown by MBE on Si (from J.A. Floro et al, Phys. Rev. Lett. 79, 3946 (1997)).

     Semiconductor quantum dots can be grown into surfaces by molecular beam epitaxy using coherency stress on lattice-mismatched substrates to make dots more thermodynamically stable than flat films (i.e. dot formation thermodynamics). We have been learning about the effect of stress-modified interface and atomic mobilities on dot formation kinetics. For more on dot formation kinetics vs. thermodynamics, including a description of what we've already done, click here. We also plan to try out some new ideas to make uniform periodic arrays of quantum dots thermodynamically stable against coarsening and coalescence, and therefore make huge coherence lengths (see figure below for typical coherence lengths) attainable upon annealing.
 

Figure: How well organized can these grown-in quantum dots be made?  The coherence length is a measure of that.  The best that's been done so far is just a few island diameters.  Several applications have been envisaged for structures with small coherence lengths, but if coherence lengths could be boosted up to wafer dimensions then these structures would be useful for multiple-layer deposition, where registry between different layers is essential.  Figure from J.A. Floro et al., Appl. Phys. Lett. 73, 951 (1998)..  Scale bar: 1 micron.
 

Self-limited nano-island formation in vapor deposition

Figure:  STM topograph of Co deposited on S-terminated Mo (110) showing two-dimensional rectangular islands. With increasing coverage they get more numerous, but hardly any bigger!  Why???  a)  0.1 monolayer (ML) Co, average island size 200 atoms. b) 0.5 ML Co, average island size 250 atoms.  From P.G. Clark and C.M. Friend, J. Chem. Phys. 111, 6991 (1999).

It appears that there's potential for manipulating surface stress in the formation and control of metal nanostructures of self-limiting size during deposition on semiconductors and metals with chemically tailored surfaces. The figure above shows Co islands on S-terminated Mo(110), grown in the laboratory of Professor Cynthia M. Friend (Dept. of Chemistry and Chemical Biology). If such islands could be built up a few layers high (while keeping their nanometer lateral dimensions) and integrated into semiconductor wafers they could be extremely useful as nanomagnets in the manipulation of spins. We are collaborating on understanding the physical and chemical origin of what appears to be a strain-induced self-limitation to the lateral island size.
 

Ion implanted nano-precipitates for smart surfaces
Oriented Fe nanocrystals produced by ion implantation into cubic ZrO₂ have potential applications in the development of new integrated-optical and information-storage devices (from S. Honda et al., Appl. Phys. Lett. 77 (5), (2000)).

Ion implantation produces arrays of nanoprecipitates displaying exciting properties, even in their current poorly-organized form. Examples include. optically-switchable metal-insulator transition in Vanadium oxide [L.A. Gea et al., J. Mater. Res. 14, 2602 (1999)]; and vertically-oriented and therefore more densely-packable magnetic domains for Fe or Co-Pt in insulators [S. Honda et al., Mat. Res. Soc. Symp. Proc. 581, 71 (2000)]. We need to understand the energetics and kinetics of nanoprecipitate formation, orientation, spacing, size distribution and regularity, and the relation to properties. We are collaborating with our colleagues at ORNL in an effort to provide a physical and chemical science base for the formation of ion implanted metal and semiconducting nanoparticles. See, for example, our early published work, publication #85, on GaAs nanoparticles in Si: C.W. White, J.D. Budai, J.G. Zhu, S.P. Withrow, and M.J. Aziz, "Ion Beam Synthesis and Stability of GaAs Nanoclusters in Silicon", Applied Physics Letters, 68, 2389 (1996); addendum 69, 2297 (1996).