For many tasks in illumination and collection the acceptance angle is required to vary along the reflector. If the acceptance angle function is known, then the reflector profile can be calculated as a function of it. The total flux seen by an observer from a source of uniform brightness (radiance) is proportional to the sum of the view factor of the source and its reflection. This allows one to calculate the acceptance angle function necessary to produce a certain flux distribution and thereby construct the reflector profile. We demonstrate the method for several examples, including finite size sources with reflectors directly joining the source.

A novel design utilizing two Xenon flashlamps and two luminaires each of which collects from two sides of the lamps is described. The combination gives uniform 360 degree(s) coverage in azimuth and a very narrow elevation pattern. Good glare control is also maintained in the elevation plane. Theoretical considerations are briefly presented. Ray traces are also presented to show the uniformity of illumination and glare control as a function of luminaire depth.

Interest in improving backlight performance is increasing rapidly, driven by the large world-wide investment in high-perfonnance LCD display screens. At the simplest level, afl backlights consist of one or more light sources (usually tubular fluorescent lamps) and an optical system. The optical system collects the fluorescent light and causes it to pass through the LCD display screen towards the viewer. NiOptics' proprietary optical designs have been used to increase flat-panel backlight luminance by a factor of 2 to 2.5 over comparable conventional systems. This dramatic improvement results from matching illumination to preferred viewing angles. The maximum theoretical improvement is quantified by deriving the fundamental thermodynamic limits which restrict two key backlight sub-systems, the light source and the distribution panel. These sub-system analyses are then combined to show thesurprisingly complex set oftrade-offs intrinsic to all backlight designs, completely independent ofthe specific device architecture. Specifically, we show that backlight luminance has an upper bound given by Po ILot< x— D1D2Sin90t cxlcclL hsinOi 1—cx In this expression L0t is the output luminance, øout is the vertical output half-angle, P is the lamp luminous output, Di and D2 are the linear dimensions of the backlight emission surface, h is the backlight thickness, dL is the lamp I.D., and a is the lamp's effective self-absorptivity coefficient. Two limiting special cases of this expression are developed to illustrate the design trade-offs. The comparisondemonstrates that superior performance requires that angular reduction be achieved within the distribution panel. Finally, we compare these two theoretical limits to NiOptics' experimental results. The results show that NiOptics proprietary backlights perform at 66% of the theoretical limit, with shortfall due primarily to material absorption and reflection losses.

Luminaire designs based on multiple asymmetric nonimaging compound parabolic reflectors are proposed for 2-D illumination applications that require highly uniform far-field illuminance, while insuring maximal lighting efficiency and sharp angular cutoffs. The new designs derive from recent advances in nonimaging secondary concentrators for line-focus solar collectors. The light source is not treated as a single entity, but rather is divided into two or more separate adjoining sources. An asymmetric Compound Parabolic Luminaire is then designed around each half-source. Attaining sharp cutoffs requires relatively large reflectors. However, severe truncation of the reflectors renders these devices as compact as many conventional luminaires, at the penalty of a small fraction of the radiation being emitted outside the nominal cutoff. The configurations that maximize the uniformity of far-field illumination offer significant improvements in flux homogeneity relative to alternative designs to date.

Generalized nonimaging Compound Elliptical Luminaires (CELs) and Compound Hyperbolic Luminaires (CHLs) are developed for pair-overlap illumination applications. A comprehensive analysis of CELs and CHLs is presented. This includes the possibility of reflector truncation, as well as the extreme direction spanning the full range from positive to negative. Negative extreme direction devices have been overlooked in earlier studies, and are shown to be well suited to illumination problems were large cutoff angles are required. Flux maps can be calculated analytically without the need for computer ray tracing. It is demonstrated that, for a broad range of cutoff angles, adjacent pairs of CELs and CHLs can generate highly uniform far-field illuminance, while maintaining maximal lighting efficiency and excellent glare control. The tradeoff between luminaire compactness and flux homogeneity is also illustrated. For V-troughs, being a special case of CHLs and being well suited to simple, inexpensive fabrication, we identify geometries that closely approach the performance characteristics of the optimized CELs and CHLs.

The applicability of non imaging concentrators for solar thermal devices and laser pumping in a central receiver facility is discussed. A kilowatt scale solar pumped laser is described.

Corrections to classical radiometry are considered for the case of a uniform medium. The corrections represent wave effects of physical optics and the effects of partial coherence. New results include integral transforms and infinite series connecting the real and imaginary parts of Walther's distribution function and the Wigner function; infinite series representing all the corrections to the evolution of different distributions along rays; and estimates and comparisons of the degree to which different distribution functions are conserved along rays. In addition, a new distribution function is presented, which is exactly conserved along rays.

The total internal reflection lens is a multi-faceted non-imaging device that was introduced in a paper we presented at the SPIE Non-Imaging Optics conference two years ago, with emphasis on solar-energy concentration. This paper will discuss the concentration on a target of the light emitted omnidirectionally from a compact source, such as an incandescent filament, a light- emitting diode, or an arc lamp. The converging type of TIR lens can efficiently concentrate this light into the relatively restricted range of acceptance angles typical of optical fibers and image illumination subsystems. It can replace the conventional ellipsoidal reflector and is superior to it in two ways: (1) its interception efficiency is considerably higher, when used with a back mirror; (2) its aberrations are lower in magnitude, because of the additional degrees of design freedom possible in the three-faced facets of the TIR lens. However, it is more sensitive to lens optical quality, so that theoretically possible results are best achieved with diamond-turning of lenses or molds. Also, TIR lenses are sensitive to errors in source position.

A new Intense Point Source (IPS) is capable of focusing twice as much light into an optical fiber as the traditional elliptical reflector system. The special feature of the IPS system is the ability to focus a linear light source to an extremely small point, such as at the tip of a optical fiber bundle. The scaling law of the regular reflector requires the increase of the physical dimension of the light source and thus leads also to a larger focused image. Unfortunately, the light intensity does not scale with the increased power. The new IPS system consists of a linear light source and a compound Orthogonal Parabolic Reflector (OPR), properly matched to produce an intense image at the focal point of the system. Intensity at the focal point can be increased simply by increasing the linear length of the light source.

A review of the nonimaging concentrator method of design called Poisson brackets method is presented. The method is valuable for 2D and 3D concentrators. The 2D version of the method has provided an interesting concentrator: the CTC (Compound Triangular Concentrator). Unfortunately, so far only impractical 3D concentrators (with variable index of refraction) have been obtained with the 3D version of the method. Nevertheless, these 3D concentrators prove that there are optimal and ideal 3D concentrators besides the trumpet concentrator and the trivial cases.

A new method of design of optical systems are presented in this paper. This method is a generalization of one recently developed at Instituto de Energia Solar (I.E.S.) of Univ. Polit. de Madrid. The method referenced has been used on the design of non-imaging concentrators for finite and infinite source. Outstanding results at the 3D performance have been achieved. The method now presented has also been used in the concentrators design, getting the same results as former's. The aim of this method is the design of two surfaces optically active (lens, two mirror or lens-mirror combination) to transform one incoherent radiation of light arriving at the system's aperture into a different radiation at the system's outlet. A computer program has been built for the method application, and some examples have been included.

Two new nonimaging concentrators called RX and RXI are presented. Both of them have been designed with a method of design recently developed. The concentrators, formed from a single dielectric piece, achieve the optimum relation between concentration and acceptance angle in 2D. Their performance in 3D is very good when the acceptance angle of the concentrators is small (less than 5 degrees for a source at infinity): total transmissions above 95% are usually achieved (total transmission is the ratio of the etendue of those rays impinging on the entry aperture, within the acceptance angle, which reach the receiver, over the etendue of the bundle formed by all the rays impinging the receiver). These values are similar or even higher than those achieved by an equivalent CPC (same acceptance angle and same refractive index). The RX shown here have been designed for a finite source and the RXI for a source at infinity.

A non imaging integrated evacuated solar collector for solar thermal energy collection is discussed which has the lower portion of the tubular glass vacuum enveloped shaped and inside surface mirrored to optimally concentrate sunlight onto an absorber tube in the vacuum. This design uses vacuum to eliminate heat loss from the absorber surface by conduction and convection of air, soda lime glass for the vacuum envelope material to lower cost, optimal non imaging concentration integrated with the glass vacuum envelope to lower cost and improve solar energy collection, and a selective absorber for the absorbing surface which has high absorptance and low emittance to lower heat loss by radiation and improve energy collection efficiency. This leads to a very low heat loss collector with high optical collection efficiency, which can operate at temperatures up to the order of 250 degree(s)C with good efficiency while being lower in cost than current evacuated solar collectors. Cost estimates are presented which indicate a cost for this solar collector system which can be competitive with the cost of fossil fuel heat energy sources when the collector system is produced in sufficient volume. Non imaging concentration, which reduces cost while improving performance, and which allows efficient solar energy collection without tracking the sun, is a key element in this solar collector design.

The goal of the optical design of luminaires and other radiation distributors is to attain the desired illumination on the target with a given source, while minimizing losses. While the required design procedure is well known for situations where the source can be approximated as a point or as a line, the development of general design methods for extended sources has begun only very recently. A solution for extended sources can be obtained by establishing a one-to-one correspondence between target points and edge rays. In the present paper the possible solutions in two dimensions (cylindrical sources) are identified, and are based on only one reflection for the edge rays. The solutions depend on whether the 'image' on the reflector is bounded by rays from the near or from the far edge of the source. For each case there are two solutions that could be called converging and diverging by analogy with imaging optics. Hence four building blocks emerge from which luminaires can be designed. Interesting hybrid configurations can be constructed by combining these building blocks. Thus one can gain a great deal of flexibility for tailoring designs to specific requirements. The differential equation for the reflector is shown to have an analytical solution. Explicit results are presented for symmetric configurations with target at infinity.