Final Report Summary - PARNASS (Parallel nano assembling directed by short-range field forces)
The experimental investigation was intended to acquire a quantitative knowledge of the forces acting on nanoparticles and their response to these forces. The work consisted of three steps. First, techniques to prepare suitable substrate on which nanoparticles can be deposited were developed. Second, qualitative knowledge of the forces on the nanoparticles was acquired. Finally, this qualitative knowledge was built upon to derive a quantitative model capable of describing and predicting the behaviour of nanoparticles when acted on by a design's force fields.
One important class of nanoparticles studied was semiconductor nanowires, i.e. long, thin rods of materials such as InAs that are typically a few tens of nanometres in diameter and several microns in length. Lund and Halmstad Universities have the capability to grow extremely high quality nanowires with a good degree of uniformity. A mix of growth techniques makes it possible to investigate a variety of semiconductor materials as well as more complex heterostructures and threedimensional nanoparticles such as branched nanotrees and core-shell nanowires. A range of methods, including high-resolution electron microscopy, scanning electron microscopy and various probe microscopy techniques are employed to characterise the basic structures and materials.
A second class of nanoparticles studied was single-wall carbon nanotubes, which are considerably thinner than nanowires and can be as small as one nanometre in diameter. Produced at the Universitat Rovira I Vigili in Tarragona, the nanotubes are purified in a complex series of steps and finally suspended in a carrier solvent. They can then be deposited on a suitable substrate and the solvent removed or allowed to evaporate.
Direct manipulation with the tip from an atomic force microscope was the technique applied most to acquire qualitative knowledge of the nanoparticles' reaction to forces. Direct manipulation can push nanowires and nanotubes laterally on a substrate surface with a high degree of control and can bend them into non-equilibrium shapes or push them so that they interact with one other or objects placed on the surface.
A measure of the deficiency of knowledge about the mechanical properties of semiconductor nanowires is the continued uncertainty whether macroscopic models of elasticity are appropriate at this length scale. Obviously this question needs to be answered before it will be possible to infer the forces acting on wires by analysing their movement or of the shapes they assume. Therefore, one fundamental task of these investigations was to characterise the wires' elastic behaviour as a function of their size, shape and processing history.
Studying the interplay of friction with the substrate surface and elastic forces in the wire has been particularly fruitful. The measured shape of the wire can be used to quantify information on the forces acting on it. Classical elasticity theory directly relates the stress in a distorted wire to its curvature. A key insight is that the intensity of the friction between a wire and its substrate determines the maximum curvature the wire can assume. If friction is weak, a tightly curved wire will simply straighten to relieve the elastic stress. The higher the friction, the more tightly a wire can be bent. The curvature of any wire can be measured directly from atomic-force microscopy (AFM) images of it, which can then be used to calculate the strain. Assuming the Young's modulus of the wire is the same as in bulk InAs makes it possible to directly derive numerical values for the shear stress and the coefficient of friction.
Just as the elastic properties, the applicability of existing models of friction to nanowires sliding on a planar surface was not at all clear at the outset of the project. The contact between the wire and the substrate perpendicular to the wire axis only covers a few tens of nanometres. However, the contact along the wire axis extends a few microns. Therefore, much effort to quantify the variation of nanowire friction with the nanowire diameter and on surface with various properties. This made it possible to investigate the influence (if any) of an adsorbed water layer on the substrate and to map the range of friction forces experienced by the nanowires.
The silanisation of silicon dioxide entails coating the oxide with a single layer of organic molecules. The molecules in a silanised layer stand upright like the bristles on a brush and form a layer with a thickness corresponding to the length of the molecule. In the case presented here, the chains were thirteen carbon atoms long, each carbon atom in the chain being connected to two fluorine atoms in addition to its neighbours in the chain.
As would be expected from relatives of Teflon, fluorinated silane layers exhibited very low coefficients of friction in macroscopic measurements. In the project's experiments, the silanised surfaces indeed had the lowest values for sliding friction. However, the values for static friction did not differ much from those measured on other substrates. From the perspective of application, this is a decided advantage. Once a nanoparticle has stopped moving, it will stay put, held in place by the significantly higher value of sliding friction.
The experiments conducted in the project applied the quantitative knowledge of friction acquired to the nanoscale to perform directed self-assembly in which nanowires move to their intended positions without individual manipulation. The experiments have delivered quite promising results and will be continued beyond the conclusion of the PARNASS project.
One important class of nanoparticles studied was semiconductor nanowires, i.e. long, thin rods of materials such as InAs that are typically a few tens of nanometres in diameter and several microns in length. Lund and Halmstad Universities have the capability to grow extremely high quality nanowires with a good degree of uniformity. A mix of growth techniques makes it possible to investigate a variety of semiconductor materials as well as more complex heterostructures and threedimensional nanoparticles such as branched nanotrees and core-shell nanowires. A range of methods, including high-resolution electron microscopy, scanning electron microscopy and various probe microscopy techniques are employed to characterise the basic structures and materials.
A second class of nanoparticles studied was single-wall carbon nanotubes, which are considerably thinner than nanowires and can be as small as one nanometre in diameter. Produced at the Universitat Rovira I Vigili in Tarragona, the nanotubes are purified in a complex series of steps and finally suspended in a carrier solvent. They can then be deposited on a suitable substrate and the solvent removed or allowed to evaporate.
Direct manipulation with the tip from an atomic force microscope was the technique applied most to acquire qualitative knowledge of the nanoparticles' reaction to forces. Direct manipulation can push nanowires and nanotubes laterally on a substrate surface with a high degree of control and can bend them into non-equilibrium shapes or push them so that they interact with one other or objects placed on the surface.
A measure of the deficiency of knowledge about the mechanical properties of semiconductor nanowires is the continued uncertainty whether macroscopic models of elasticity are appropriate at this length scale. Obviously this question needs to be answered before it will be possible to infer the forces acting on wires by analysing their movement or of the shapes they assume. Therefore, one fundamental task of these investigations was to characterise the wires' elastic behaviour as a function of their size, shape and processing history.
Studying the interplay of friction with the substrate surface and elastic forces in the wire has been particularly fruitful. The measured shape of the wire can be used to quantify information on the forces acting on it. Classical elasticity theory directly relates the stress in a distorted wire to its curvature. A key insight is that the intensity of the friction between a wire and its substrate determines the maximum curvature the wire can assume. If friction is weak, a tightly curved wire will simply straighten to relieve the elastic stress. The higher the friction, the more tightly a wire can be bent. The curvature of any wire can be measured directly from atomic-force microscopy (AFM) images of it, which can then be used to calculate the strain. Assuming the Young's modulus of the wire is the same as in bulk InAs makes it possible to directly derive numerical values for the shear stress and the coefficient of friction.
Just as the elastic properties, the applicability of existing models of friction to nanowires sliding on a planar surface was not at all clear at the outset of the project. The contact between the wire and the substrate perpendicular to the wire axis only covers a few tens of nanometres. However, the contact along the wire axis extends a few microns. Therefore, much effort to quantify the variation of nanowire friction with the nanowire diameter and on surface with various properties. This made it possible to investigate the influence (if any) of an adsorbed water layer on the substrate and to map the range of friction forces experienced by the nanowires.
The silanisation of silicon dioxide entails coating the oxide with a single layer of organic molecules. The molecules in a silanised layer stand upright like the bristles on a brush and form a layer with a thickness corresponding to the length of the molecule. In the case presented here, the chains were thirteen carbon atoms long, each carbon atom in the chain being connected to two fluorine atoms in addition to its neighbours in the chain.
As would be expected from relatives of Teflon, fluorinated silane layers exhibited very low coefficients of friction in macroscopic measurements. In the project's experiments, the silanised surfaces indeed had the lowest values for sliding friction. However, the values for static friction did not differ much from those measured on other substrates. From the perspective of application, this is a decided advantage. Once a nanoparticle has stopped moving, it will stay put, held in place by the significantly higher value of sliding friction.
The experiments conducted in the project applied the quantitative knowledge of friction acquired to the nanoscale to perform directed self-assembly in which nanowires move to their intended positions without individual manipulation. The experiments have delivered quite promising results and will be continued beyond the conclusion of the PARNASS project.
