Most probe-based approaches to data storage rely on heating a polymer substrate with a tip, but a new approach relies on pressure instead.
Does ice melt when we skate? If the skater is heavy enough, the skates are sharp enough and the air temperature is not too cold, localized melting will occur. The thin layer of water formed beneath the skates reduces the friction of the ice and allows the skater to glide easily over the surface (depending on the skater's skill level, of course). A similar situation is found with block copolymers (polymers in which two dissimilar polymer chains or 'blocks' are joined together at one end). Usually, non-favourable interactions between the blocks drives them as far away from each other as possible, which is limited by the length of the block copolymer, and results in the formation of microdomains of one block, tens of nanometres in size, embedded in a matrix of the second block.
The volume fraction of each block dictates whether the polymer self-assembles into arrays of spherical, cylindrical, gyroidal or lamellar microdomains with a characteristic period, L0, where the size of the microdomains and L0 are controlled by the molecular weight of the polymer (see Fig. 1 and ref. 1). On heating, the non-favourable interactions decrease to the point where the blocks mix and the self-assembled spherical, cylindrical or lamellar structures are lost. Like melting, this order-to-disorder transition is accompanied by a dramatic change in the rheological properties of the polymer (Fig. 1). In the ordered state the microdomains retard flow, but when they are mixed, the polymer can readily flow, provided it is above its glass transition temperature, Tg.
Some block copolymers, specifically polystyrene-block-poly(n-alkylmethacrylates), become ordered on heating2 through a lower disorder-to-order transition (Fig. 1). Unlike the classical order-to-disorder transition, where enthalpy governs the phase behaviour, the lower disorder-to-order transition is driven by a volume change and is therefore dominated by entropy3. This family of block copolymers is quite remarkable because the nature of the phase transition depends on the length of the alkyl group: when the alkyl group is methyl, a classic order-to-disorder transition is seen on heating; for ethyl, propyl or butyl, a lower disorder-to-order transition is observed4; for pentyl, closed-loop behaviour is seen; and for longer alkyl chains, an order-to-disorder transition is again observed4 (Fig. 1). Unlike the order-to-disorder transition, the lower disorder-to-order transition and closed-loop behaviour are very sensitive to pressure5,6 so an ordered block copolymer can be disordered simply by applying pressure. At high enough pressures, only the classic order-to-disorder transition remains (Fig. 1).
On page 727 of this issue Jin Kon Kim and co-workers7 at Pohang University of Science and Technology and the LG Electronics Advanced Research Institute, both in Korea, demonstrate that it might be possible to exploit the pressure-dependent properties of block copolymers for use in data storage. Kim and co-workers use a scanning probe tip to apply a very localized pressure to a surface coated with polystyrene-block-poly(n-pentyl methacrylate) copolymer (PS-b-PnPMA). As with ice skating, the pressure causes a local order-to-disorder transition or melting that allows the polymer to flow8 making it possible for the tip to make an impression in the polymer. The glass transition temperatures for both the ordered and the disordered block copolymers are well above room temperature, so the mark in the polymer made by the tip is frozen in place when the tip is removed and the pressure decreased. This mark can be used for information storage. Kim and co-workers produced features that were 2 nm deep with a pitch of 37 nm, which translates into a storage density of ∼1 Tbit in−2. The elegance of this process is that entropy (pressure) rather than enthalpy (heat) enables the writing to be done at room temperature.
The maximum storage density achievable will depend on the sharpness of the tip and also on the amount of surface energy that is needed to generate such small features. Sharper tips lead to smaller features and higher storage densities, but nature abhors such small features and a Laplace pressure operates to smooth the surface. If the temperature is not sufficiently below the Tg of the polymer, the polymer will relax and the features will be lost. This is where a second important characteristic of block copolymers comes into effect. The different interactions of the blocks with the substrate and the difference in the surface energies of the blocks cause the blocks to segregate. However, as the blocks are linked together, in thin films this segregation leads to a multilayered structure consisting of alternating layers of the two blocks9.
The films used by Kim and co-workers initially consisted of alternating layers of polystyrene, with Tg ∼100 °C, and PnPMA (Tg ∼7 °C). The Tg of the mixed block copolymer is ∼70 °C, so the marks made by the tip were frozen into place during operation at room temperature and atmospheric pressure. Outside the indentation region the polymer is ordered and the flow is retarded, thus removing the need for crosslinkers (which are needed when homopolymer films are used10). The multilayered structure of the film also means that the film thickness at any point is defined in terms of L0 (see Fig. 1 and ref. 9). This means that the planarity of the film surface is guaranteed by this layering as long as the thickness of the original film placed on the substrate is commensurate with L0. If not, a terraced surface topography will form with steps L0 in height9.
The results of Kim and co-workers7 show clear advantages in using block copolymers that show lower disorder-to-order transition or closed-loop behaviour. With advantages also come disadvantages. Lower disorder-to-order transitions and/or closed-loop behaviour are only seen in some of the polystyrene-block-poly(n-alkylmethacrylates). Although the syntheses of this family of block copolymers is well known, this limitation is disadvantageous, particularly as the poly(n-alkylmethacrylates) have low values of Tg that can give rise to a long-term instability of the written features. Furthermore, methacrylate-based polymers can degrade. However, these potential drawbacks should not detract from the importance of the work of Kim and co-workers7 in developing a process that has the potential to advance the present state-of-the-art in probe-based non-volatile memory technology.