In the transformation, a triangular cross-section in a virtual space (a) filled with isotropic materials is mapped to a quadrilateral (brown region in (b) with uniform and anisotropic optical properties. The cloaked region is defined by the small grey triangle wherein objects can be rendered invisible. (c) A photograph of the triangular cloak, which consists of two calcite prisms glued together, with the geometrical parameters indicated in the figure. The dimension of the cloak along z direction is 2 cm. The optical axis, represented by red arrows, forms an angle of 30° with the glued interface.

The refractive index of the surrounding media is assumed to be 1.532. (a, b) Light incident on a cloaked bump at incident angles of (a) 45° and (b) 15°, showing almost no scattering. (c, d) Without cloaking, the bump strongly scatters the light into different directions. Although the cloak is designed at 590 nm, in accordance with the experiment, the ray-tracing calculation is performed at 532 nm.

(a) (Left) Schematic of the experimental setup. A patterned laser beam is reflected by a calcite cloak (or a flat reflective surface as control sample) and projected onto a screen. (Right) The original pattern of the laser beam, which consists of a bright arrow in the centre and a number of flipped dim arrows on both sides. (b) The pattern of the laser beam as reflected by a flat surface. The size of the mirror only allows the central arrow to be reflected. The projected arrow image is about 1.2 cm long in the horizontal direction. (c, d) The projected image of the laser beam reflected by the calcite cloak for TE and TM polarizations, respectively. The TM measurement shows that the laser beam is not distorted by reflection by the triangular protruding surface. (e, f, g) the projected images for mixed TE and TM polarizations at incidence angles of 39.5°, 64.5° and 88°, respectively. For all incident angles, the central TM images are not distorted, the cloaked reflective bump appears to be a flat mirror to outside observers. Because of the limited size of the reflective surface, only the central arrow was reflected and subsequently changed its propagation direction, generally causing a large separation between its image projected on the screen and the others. However, in Figure 3g, for an incident angle close to the grazing angle, the change of direction is very small; therefore, the images of the reflected central arrow and the other dimmer arrows all appear in the field of view of the camera. (h, i) The photographs of a red laser beam with mixed TE and TM polarizations projected on the screen after being reflected by (h) calcite cloak and (i) a flat surface at an incident angle of 64.5°.

(a) Schematic of the optical setup. (b, c) The reflected image captured by the camera for TE (without cloaking) and TM polarizations, respectively. Note that the rainbow appears at the edge of each letter for TM polarization due to the optical dispersion of the calcite crystal.

(a) The deviation angles of the beams with TM polarization reflected by the cloak with the design parameters indicated in Figure 1c. (b) Same as a, but for a slightly modified cloak geometry so that perfect cloaking occurs at wavelength of 532 nm, consistent with the experimental observations for the green laser beam. (c) The effect of dispersion for TE polarization in terms of the variation of reflection angles relative to those at λ=590 nm. In all the figures, the incident angle was set to 64.5°.

(a) Schematic of the ray tracing for an incident beam with TE polarization, which is split into two because of reflection by the cloak. (b) The calculated angles of deviation, φ₁, φ₂, of the beams reflected by the protruding bottom surface, are plotted against the incident angle θ (solid lines), and found to be in good agreement with the measurement (solid squares).