Amplification of Acoustic Evanescent Waves Using Metamaterial Slabs

We amplified acoustic evanescent waves using metamaterial slabs with a negative effective density. For the amplifying effect of the slab to overcome the dissipation, it is necessary that the imaginary part of the effective density is much smaller than the real part, a condition not satisfied so far. We report the construction of membrane-based two-dimensional negative-density metamaterials which exhibited remarkably small dissipation. Using a slab of this metamaterial we realized a 17-fold net amplitude gain at a remote distance from the evanescent wave source. Potential applications include acoustic superlensing.

Figure 1

Progress of the idea for a negative-density metamaterial with small dissipation. (a) Lattice of double resonators embedded in a fluid. (b) Membranes are installed to make the fluid move together with the double resonator. (c) Structure with the membranes that generates a negative density.

Figure 2

(a) The unit cell consists of four identical windows of thin elastic membranes. Photo of two-dimensional $\rho$-NG acoustic metamaterial with the top cover removed. (b) The equivalent circuit of the unit cell.

Figure 3

Amplification of evanescent wave with wave vector $k=1.6k_0$ and $f=462\mathrm{Hz}$. Pressure amplitude was normalized with the amplitude at the source without a metamaterial. Color maps represent the distribution of acoustic pressure. (a) The amplitude of the evanescent pressure field without a metamaterial. (b) The amplitude of the evanescent pressure field with a metamaterial.

Figure 4

(a) Schematic diagram of the experimental setup to measure the dispersion relation of the acoustic surface waves. A point source was used to excite the surface waves. (b) Pressure distribution of the surface wave in the air at $f=462\mathrm{Hz}$. The interface between the air ($z>0$) and the metamaterial ($z<0$) is indicated by the $z=0$ plane. (c) Theoretical and experimental dispersion relation of the acoustic surface wave. Experimental results agree excellently with the theoretical curve calculated from Eq. (5). The wave vector diverges at the frequency where ${\rho }_{\mathrm{eff}}=-{\rho }_a$.