The recently emerged fields of metamaterials and transformation optics promise a family of exciting applications such as invisibility, optical imaging with deeply subwavelength resolution and nanophotonics with the potential for much faster information processing. The possibility of creating optical negative-index metamaterials (NIMs) using nanostructured metal–dielectric composites has triggered intense basic and applied research over the past several years1, 2, 3, 4, 5, 6, 7, 8, 9, 10. However, the performance of all NIM applications is significantly limited by the inherent and strong energy dissipation in metals, especially in the near-infrared and visible wavelength ranges11, 12. Generally the losses are orders of magnitude too large for the proposed applications, and the reduction of losses with optimized designs seems to be out of reach. One way of addressing this issue is to incorporate gain media into NIM designs13, 14, 15, 16. However, whether NIMs with low loss can be achieved has been the subject of theoretical debate17, 18. Here we experimentally demonstrate that the incorporation of gain material in the high-local-field areas of a metamaterial makes it possible to fabricate an extremely low-loss and active optical NIM. The original loss-limited negative refractive index and the figure of merit (FOM) of the device have been drastically improved with loss compensation in the visible wavelength range between 722 and 738 nm. In this range, the NIM becomes active such that the sum of the light intensities in transmission and reflection exceeds the intensity of the incident beam. At a wavelength of 737 nm, the negative refractive index improves from −0.66 to −1.017 and the FOM increases from 1 to 26. At 738 nm, the FOM is expected to become macroscopically large, of the order of 106. This study demonstrates the possibility of fabricating an optical negative-index metamaterial that is not limited by the inherent loss in its metal constituent.
Figures at a glance
Figure 1: Schematic of the fabrication process. 
a, Unit cell of the fishnet structure with alumina as the spacer material between two silver layers. b, One-quarter of the fishnet structure with an alumina spacer. c, After etching the alumina, the fishnet structure has air or solvent as the spacer with alumina pillars as support. d, After coating with Rh800–epoxy, the fishnet structure has the dye–epoxy material in the spacer region and above the fishnet structure.
Figure 2: SEM images of the fishnet structure at different fabrication stages. 
a, Fishnet structure with alumina spacer. b, After etching the alumina and coating with Rh800–epoxy, the fishnet structure has the dye–polymer as the spacer and on the top. c, Tilt-view SEM image of the structure after coating with Rh800–epoxy and after a part of the top layer of silver has been removed by focused ion-beam milling. The scale is the same in all of the SEM images. d, The pump–probe experimental set-up. CCD, charge-coupled device; OPA, optical parametric amplifier.
Figure 3: Experimental results and simulation. 
a, Experimental far-field transmission (Te), reflection (Re) and absorptance (Ae) spectra of the sample, along with simulated results (Ts, Rs, and As) at the primary linear polarization shown in Fig. 2a. b, The transmission spectra without pumping (line 1), with the optimized delay between pump and probe (probe pulse is 54 ps later than the pump) and 1-mW pumping power (line 5), with the optimized delay and 0.12-mW pumping power (line 3), with the optimized delay and 0.16-mW pumping power (line 4), and with the pump preceding the probe by 6 ps and 1-mW pumping power (line 2). The wavelength-dependent relative transmission change from the pump–probe experiment is shown by the red solid line.
Figure 4: Simulation and determined parameters. 
a, The simulated refractive index, n′ (real part), and absorptance, A (in the forward direction), as functions of wavelength with (solid) and without (dashed) gain. b, The effective refractive index, n = n′ + in′′, determined with (solid) and without (dashed) gain. c, The effective FOM determined with (solid) and without (dashed) gain (the FOM is set to zero when the real part of the refractive index is positive). d, The effective permittivity, ε′ (real part), and permeability, μ′ (real part), determined with (solid) and without (dashed) gain.
right
