WO1997039159A1

WO1997039159A1 - Coated substrate

Info

Publication number: WO1997039159A1
Application number: PCT/GB1997/001031
Authority: WO
Inventors: Brian L. Evans
Original assignee: The University Of Reading
Priority date: 1996-04-12
Filing date: 1997-04-14
Publication date: 1997-10-23
Also published as: AU2518197A; GB9607635D0

Abstract

Regular arrays of microscopic particles are provided on substrate surfaces by (a) providing a monolayer of close-packed spheres (S) on a substrate surface (e.g. by spin-coating with a suspension of polystyrene spheres); and (b) using the spheres as a mask for depositing a material on the substrate surface through the gaps (G) between the spheres (e.g. by physical vapour deposition). Possible further steps include: removal of the spheres (S); heating to melt the particles to covert them to hemispheres; and application of more material which may be the same or different.

Description

COATED SUBSTRATE

Technical Field

The present invention relates to a substrate coating process, and its application to the production of substrates bearing arrays of particles. In a further aspect it relates to such substrates themselves.

The invention is principally concerned to provide a substrate bearing a substantially regular array Of substantially uniform particles of very small size, typically of the order of a few nm or a few tens of nm. In the past it has only been possible to produce very small arrays of this nature, e.g. arrays of 100-600 particles formed by a laborious process involving STM decomposition of Fe(C0)₅. The present invention can provide arrays which are orders of magnitude greater than this, e.g. extending over more than lcm². Disclosure of Invention

Broadly the present invention provides a process in which: (a) a substrate is coated with a monolayer of close- packed spheres;

(b) a coating of the particle-forming material is applied over the sphere-coated substrate, the spheres acting as a mask such that discrete deposits of the particle-forming material reach the substrate surface; and

(c) the spheres are removed.

In general the spheres for forming the monolayer should be substantially uniform, since otherwise they will not form close-packed coatings. Generally, where the nominal diameter is D, a population of spheres of diameters 90-120%D is likely to be acceptable, with 90- 110%D preferred. (However a controlled level of non- uniformity can be exploited as described later.) . Suspensions of (e.g.) Polystyrene and silica spheres are commercially available for purposes such as electron microscope calibration and colloidal systems research. For example, Duke Scientific Corporation (California, USA) offer aqueous suspensions of polystyrene spheres, e.g. a 2% solids content suspension of polystyrene spheres of certified mean diameters D of 20 ± 1.5nm, 30 ± 1.3nm, and other sizes. DuPont supply silica particles as aqueous suspensions, with average particle diameters of 7,12 or 22nm. Furthermore methods of preparing suspensions of polystyrene latex particles of specific diameters are well known [R H Ottewill and R A Richardson, Colloid & Polymer Sci 260 708-719 (1982)] . Spheres of different materials may be used. The choice may affect the techniques applicable in steps (b) and (c) .

Adhesion of the polystyrene and silica spheres to the substrate can be due to electrostatic attraction (this is principally but not always the case for polystyrene spheres) and/or chemical bonding.

In the case of silica sols the particles are negatively charged due to the addition of small amounts of alkali which react with the silica surface to product a negative charge (hydroxyl ions react with surface silanol groups to create surface negative charges) . As the pH is reduced, particle charge and stability decrease. At pH < 5 particles become uncharged; however interparticle bonding decreases and the sol becomes more stable. There is a grade of colloidal silica in which trivalent Al atoms have been substituted for the tetravalent Si atoms on the particle surfaces creating a fixed negative charge independent of pH.

The invention will now be described in more detail with reference to the accompanying drawings. Brief Description of Drawings

Figs 1-3 are scanning electron microscope ("SEM") photographs of coatings of polystyrene spheres produced by spin coating;

Fig 4 is a diagrammatic representation of a close- packed array of spheres;

Figs 5 and 6 are SEM photographs of arrays of spheres showing the effects on close packed arrays of the presence of spheres of different sizes;

Fig 7 is a SEM photograph of an array of spheres,- and

Fig 8 is a SEM photograph of a substrate bearing a pattern of aluminium particles produced using the array of Fig 7. Modes for Carrying Out the Invention

Techniques suitable for applying a monolayer of spheres on a substrate include the following.

(i) A monolayer is formed on the surface of a Langmuir trough, compressed with barriers to form an ordered structure, then lifted off on a substrate. The monolayer will adhere if the substrate has a surface positive charge due to the aqueous medium having pH less than the isoelectronic point of the substrate surface.

(ii) Spin coating: the substrate is attached to the vacuum chuck of a spin coater and a measured amount of particulate suspension (which must wet the substrate; the addition of a small amount of detergent to the suspension can assist in this) is dropped on to the centre of the spinning substrate. Spinning continues until the substrate surface appears dry. The spin speed required to form a monolayer coating is critically dependent on the sphere diameter and the physical and electrical properties of the suspension and substrate. This is illustrated by Figures 1 and 2 : Figure 1 shows the state of a monolayer prepared at too high a spin speed whereas in Figure 2 the speed was almost correct. It is a routine matter to adjust the spin speed or other variables to achieve a uniform coating.

(iii) The substrate is immersed in a suspension of the spheres and then slowly withdrawn, at an angle to the liquid surface.

Adhesion of spheres to the substrate is principally due to electrostatic attraction. In polystyrene suspensions (commonly referred to as polystyrene latices) the spheres have a surface negative charge which gives rise to interparticle repulsion which prevents coagulation.

It is generally prudent to filter out any * floes' that may be present in the suspension before attempting to form a monolayer, e.g. by filtration through glass wool. Fig. 3 shows the disruptive effect of floes.

In the step (b) the regular array of close packed spheres on the substrate is overcoated with one or more films of material The deposition technique employed should generally be such that the impinging atoms arrive normal to the substrate, otherwise incident atoms may not penetrate the sphere interstices. Appropriate methods of physical vapour deposition (PVD) are thermal evaporation and sputtering (although some forms of sputtering can damage or even remove the spheres.) In this respect silica spheres are more robust than those of polystyrene The deposition method employed is desirably capable of depositing a precise amount of material since this determines the vertical neight of the deposited particles, and thus the amount of material in a particle and its dimensions after annealing (if any) (see below)

In step (c) the spheres are removed Different techniques are appropriate in different circumstances (notably, for spheres of different materials) For example, spheres may be dissolved away (e.g. using dichloromethane to dissolve polystyrene spheres) , etched away (e.g using oblique incidence ion beam etching) , or physically lifted off, e.g. by application and removal of a flexible adhesive tape.

Fig. 4 is a diagrammatic plan view of a close packed array of spheres S of diameter D = 2r. A triangle such as CAB drawn between the centre C of a sphere and the centres A,B of two adjacent inter-sphere gaps G is equilateral . It is of height r and thus its sides are of length 1 = 2r/J 3 = D/J 3. This is the spacing of the centres of the gaps, and thus the spacing of particles deposited on the substrate. To calculate the area of a gap as seen in Fig. 4

(which is the area of material deposited through a gap by deposition normal to the substrate) , consider the hexagon

H formed by joining the centres of six spheres S surrounding a central sphere S. It is composed of six equilateral triangles T of side 2r and height J3r. Thus its area iε 6>J3r². This is made up of the area of six gaps (6G) , the area of the central sphere (τrr²) and the areas of one third of each of the six surrounding spheres (Vb X 6 X πr²) . (The "area of a sphere" refers to the area as seen in the plan view of

Fig. 4) .

Thus,

6G = 6 V^r² - 3τrr²

G = (V3 - π/2)r² = D² (V3 - τr/2)

4

If material is deposited to a thickness t, then the volume v of a particle is tG. The substrate bearing the particles may be "annealed" . If this leads to melting of the material of the particles, and the material is non-wetting of the substrate and remains unaltered on the surface, the particles will become substantially hemispherical, of radius R, where

% π R³ = tG The surface area of an annealed particle is given by a=2πR² = 0.446 [D²t] ^2/3 and the number of such particles/unit area of substrate is N=4/V3D². Thus the total surface area of particles/unit area of substrate is A=NA = 1.03 (t/D)^2/3. If the practical upper limit on t is D/2 then the maximum value of A is A_,_ax = 0.65. It may be desirable to use the smallest practical value of D (hence t) to reduce the mass of material required to achieve A_max . The particles cover a fraction f=NπR²=0.518 (t/D)^2/3 of the substrate ie f_max =0.326 for A=A_max. If the particle assembly is subsequently -overcoated then the total surface area = l-f_max+A_niax=l.32 ie the surface area has increased by ~ 32%.

Annealing may be carried out before or after removal of the spheres .

The particles may be of any chemical composition and each particle may itself contain a substructure. This can be produced by depositing a number of different material films on the initial layer of spheres. For example, the first deposition may be of material A. With the spheres in situ the array is annealed to give hemispherical particles of A. Depositing material B, then annealing if required, produces a particle consisting of A overcoated with B. The process may be repeated as required. Particles may be modified by various techniques including annealing as described above; by etching the particles, either before or after annealing; and/or by further deposition of material. If the particle separations are less than the atom diffusion length on the substrate then (under appropriate conditions) existing particles will grow in size. Further deposition may be carried out by a method of the type suitable for step (b) , using a low deposition rate such that the diffusion length exceeds the mterparticle separation. Some applications of the technique will now be outlined.

(a) A regular array of metal or semiconducting particles on an insulating or semiconducting substrate behaves as a semiconductor in that it exhibits thermally activated electrical conduction. The magnitude of the activation energy (ie the effective band gap) is determined by the particle separation 1 which may be tailored to suit a particular application The conductivity can also be enhanced by irradiating one or more particles with electromagnetic radiation

(influencing the particle work function φ) or by an e beam. This effect would be particularly strong in a linear particle chain. A linear particle chain, which may be a single chain or a series of parallel particle chains, can be obtained by selectively evaporating (using the e^" beam of an electron microscope) unwanted particles in the 2-D array. An alternative approach involves using, for formation of the monolayer of spheres, a population of spheres containing a minor percentage of mis-sized spheres. The hexagonal close packing of the spheres in the monolayer can be disrupted by the presence of larger spheres, see Figure 5, or smaller spheres, see Figure 6. The latter situation can be exploited by deliberately introducing a number of smaller spheres into the host suspension which gives rise to a random distribution of 'line-voids' .

It will be noted that although the larger sphere is immediately obvious in Fig 5, it does not significantly disturb the overall close packing of the spheres. The large sphere occupies a void in the regular array. In contrast the presence of the small sphere in Fig. 6 is revealed by the change in the regular close packing (as in the form of a dislocation) . The situation shown in Fig 6 is that of the maximum relative displacement in which the centres of the spheres defining the fault line (running SW from the smaller sphere, diameter d) define the corners of a square instead of the vertices of the equilateral triangle of the close packed arrangement. It can readily be shown that this situation occurs when d = -< 0.558D. As d increases from this value the relative displacement decreases and there will be a value of d < D for which it will be energetically favourable to recover the overall close-packing, leaving the small sphere in a void. We have not yet observed this situation but estimate that it will occur at d = 0.9D. Similarly, as d becomes greater than D the array will develop fault lines associated with the increasing strain which will be relieved when the 'area' occupied by the sphere, diameter d, is the same as that of the displaced spheres. This is almost the situation seen in Fig. 5 where the rogue sphere diameter d is slightly in excess of this value. The resulting strain is shown by the sphere immediately to the S of d which has been displaced by about 0.2D; this generates a fault line running SW which disappears after 8 sphere diameters ie the array has accommodated the strain over this length. Moving W from the seventh sphere on this fault line it is seen that the third sphere along has a diameter greater than average but exerts negligible effect on the packing. The diameter d of this sphere is 1.2D. The conclusion to be drawn is that the array can accommodate spheres of diameter 20% greater than average and about 10% less than average. A conservative estimate is that a regular array can be created from spheres of diameter D +_ 10%.

(b) An array of magnetic particles can be used as an information storage medium. Magnetic thin films currently used for data storage are generally polycrystalline Co based alloys in which interactions between domains gives rise to media noise which imposes the present limit on recording density. In this respect a regular ferromagnetic array will be particularly valuable as an information storage medium since the technique is capable of preparing large area arrays comparatively cheaply. As example an array with 1 = 11 πrn corresponds to an areal recording density of 6 terabits/inch² (1 terabit/cm² i.e. 10¹² bit/cm²) as compared with the present figure of - 1 gigabit/inch² (i.e. of the order of 10⁸ bit/cm²) . (c) Arrays of nanoscale (quantum dot) devices can be produced by thermally diffusing the particle array into the semiconducting substrate surface. These active elements may be used in integrated circuits, in position sensitive devices or image display/sensing systems. For a given particle array the effect of heat treatment will depend on the nature of the substrate. Thus, for platinum on glass, heating results in the Pt particles assuming a hemispherical form on the glass surface. For platinum on silicon, however, heat treatment results in the Pt diffusing into the Si surface to form platinum suicides. Similarly titanium on a GaAs substrate forms a Schottky barrier.

(d) If a regular ferromagnetic array is overcoated with a continuous non-magnetic film then the composite system will exhibit Giant magneto resistance analogous to that exhibited by magnetic/non-magnetic multilayers. These thin film composite array systems can be used as the sensing component in close proximity magnetic read- out systems . e) Catalysis: as explained above, an annealed coating substrate has a surface area about one third greater than that of the original substrate. It may be coated with a catalytically active material e.g. a noble metal such as Pt or Pd.

Figs 7 and 8 show an actual embodiment of the invention. An aqueous suspension (or latex) of highly uniform polystyrene spheres (D = 480nm) was prepared by the method of Ottewill and Richardson referred to above. The surface of the spheres is negatively charged (due to the persulphate initiator) - surface charge density about 4 μC cm^"2) . The suspension was applied to a silicon substrate by spin coating, leading to a regular close packed array, pictured in Fig 7. Next, aluminium was deposited over the coating of spheres by vapour deposition. Finally the spheres, with their aluminium coatings, were removed by dissolving the polystyrene in dichloromethane. Fig 8 shows the substrate after this removal, with its regular array of aluminium deposits or 'particles' defining a mosaic of hexagons each like the hexagonal array of gaps G visible in Fig 4. The spacing of the particles is 480/V3 = 280nm.

Claims

1. A process for applying a coating of discrete particles to a substrate comprising the steps of: a) providing a monolayer of close-packed spheres on the substrate; b) applying a coating of a material over said monolayer of spheres under conditions such that the spheres act as a mask and discrete deposits of said material reach the substrate surface.

2. A process according to claim 1 wherein the spheres used for providing the monolayer are of nominal diameter D and they are substantially uniform such that all have diameters in the range 90-120%D.

3. A process according to claim 2 wherein all said spheres have diameters in the range 90-110%D.

4. A process according to any preceding claim wherein said spheres are of silica or polystyrene.

5. A process according to any preceding claim wherein said step (a) comprises forming a monolayer on the surface of a liquid in a Langmuir trough, compressing the monolayer to form an ordered structure, and using the substrate to lift the monolayer off the liquid surface.

6. A process according to any of claims 1-4 wherein said step (a) employs spin coating.

7. A process according to claim 6 wherein for said spin coating, the substrate is spun and a predetermined amount of a suspension of the spheres which is capable of wetting the substrate is dropped onto the spinning substrate; the speed of spinning being adjusted to achieve a uniform coating on the substrate.

8. A process according to any of claims 1-4 wherein said step (a) involves providing a suspension of the spheres in a liquid, immersing at least part of the substrate in the suspension, and slowly withdrawing the substrate with a substrate surface at an angle to the liquid surface.

9. A process according to any preceding claim wherein said step (b) employs physical vapour deposition.

10. A process according to claim 9 wherein said step (b) employs thermal evaporation.

11. A process according to any preceding claim including a subsequent step (c) of removing the spheres.

12. A process according to claim 11 wherein said step (c) involves dissolution of the spheres.

13. A process according to claim 11 wherein said step (c) involves etching.

14. A process according to claim 13 wherein said etching is oblique incidence beam etching.

15. A process according to claim 11 wherein said step (c) involves application and removal of a flexible adhesive tape.

16. A process according to any preceding claim including a step of annealing the coated substrate so that the discrete deposits melt.

17. A process according to claim 16 wherein the material is non-wetting of the substrate, and said annealing converts the deposits into substantially hemispherical deposits.

18. A process according to any preceding claim including a plurality of steps (b) applying coatings of different materials.

19. A process according to any preceding claim wherein the deposited material is subjected to etching.

20. A process according to any preceding claim wherein further material is deposited on the material deposited in step (b) under conditions such that the diffusion length exceeds the interparticle separation.

21. A process according to any preceding claim including a step of selectively removing some of the deposited particles so that the remaining particles constitute one or more linear particle chains.

22. A process according to any preceding claim wherein the substrate is an electrical insulator and the deposited material (s) comprise one or more metals and/or semiconductors .

23. A process according to any preceding claim wherein the particles are magnetic.

24. A process according to any preceding claim including a subsequent step of thermally diffusing the particles into the substrate.

25. A process according to claim 24 wherein the substrate is a semiconductor.

26. A substrate bearing an array of discrete particles as producible by the method of any of claims 1- 23 .

27. A substrate having an array of surface portions containing diffused material as producible by the method of claim 24 or claim 25.