ABSTRACT
Designing micro-optics for light-emitting diodes must take into account the near-field radiance and relative spectral power distributions of the emitting LED die surfaces. We present the design and application of a near-field goniospectroradiometer for this purpose.
Keywords: Near-field Photometry, Goniophotometer, LED Measurement, Spectroradiometer
One of the often-touted advantages of high-flux light-emitting diodes (LEDs) is that they are “point sources” of light, which in theory greatly simplifies the design of non-imaging optical systems such as architectural light fixtures (or “luminaries”) and LED backlight systems.
The situation in practice is more complex. It is often necessary to employ arrays of LEDs in order to compete with traditional incandescent and fluorescent light sources. The size and density of the array is usually limited by LED thermal considerations – LEDs generate considerable amounts of heat – and so the aggregate light source can no longer be considered a “point source” for optical design purposes.
One solution to this problem is to couple each LED die to a micro-optical element with refractive and possibly diffractive microstructures. These elements can be mounted close to the LED die or molded directly onto the surface of its optical epoxy encapsulate. They can also be applied directly to the encapsulate surface using ultraviolet replication techniques.
In order to design efficient refractive and diffractive optical elements, we need to know the LED die geometry, including its size, shape, and position within its encapsulate. We also need to know the 2D radiance distribution across the surface of the emitting die surface and its relative spectral power distribution. It is not sufficient to model the LED die as a uniform Lambertian emitter, as the spatial distribution of radiance can, depending on the current spreading layer design and bond wire placement on the die, be highly non-uniform.
A useful technique for measuring these parameters can be developed from instrumentation that was originally developed for near-field photometry of architectural luminaries.
Near-field photometry was first developed to measure and model the near-field luminous flux distribution of architectural luminaries. It uses a digital camera to measure the four-dimensional scalar field of light surrounding a volume light source such as a lamp or luminary. With these measurements, it is possible to accurately model the illuminance distribution over any surface, regardless of its distance, orientation or curvature with respect to the light source.
The key concept is the geometric ray of light. The IESNA Lighting Handbook defines illuminance E as “the luminous flux per unit area incident at a point on a surface.” That is, illuminance is due to light coming from all directions above a surface and intersecting a point, where the “surface” can be real or imaginary. We can think of this light as an infinite number of geometric rays coming from every direction above the surface plane, each with its own quantity of luminous flux.
If we can measure the luminance of each ray, we can calculate the illuminance at the point. A ray of light travels in a straight line through an optically homogeneous medium such as air. Because the quantity of luminous flux within the ray does not change (neglecting scattering and absorption), neither does its luminance. We can therefore measure the luminance of a ray anywhere along its length.
Now, consider a planar or volumetric light source surrounded by an imaginary sphere. Every ray of light emitted by the light source will have to intersect this sphere at some point. We can therefore think of the light source being surrounded by a four-dimensional scalar photic field, wherein each point has two position coordinates and two direction coordinates.
We can measure the luminance of a single ray of light with a lens-type luminance meter. More accurately, we can measure the average luminance of a bundle of rays contained within a cone defined by the photo sensor area, the lens aperture and focal length, and the focus point. In practical terms, this may still be considered representative of a single ray for a luminance meter with a sufficiently narrow field of view. (The “surface” that the luminance meter is focused on can be real or imaginary.)
For the purpose of a practical near-field goniophotometer, we can replace the lens-type luminance meter with a photometrically-or radiometrically-calibrated digital camera, wherein each image pixel measures the luminance or radiance of a unique geometric ray. With an image resolution of (say) 1024 × 1024 pixels, a digital camera can simultaneously measure the luminance of over one million rays that converge on the camera lens.
If we mount the camera on a moveable arm that rotates in the vertical plane about the light source and rotate the light source in the horizontal plane, the camera will circumscribe an imaginary sphere about the light source. By capturing images at closely spaced intervals in the vertical and horizontal planes, we can thus adequately sample the surrounding photic field. The luminance of an arbitrary geometric ray can then be interpolated from this set of measured rays.
Practical near-field goniophotometers have been constructed. For example, Ashdown used the resultant ray set to predict the luminance distribution of architectural surfaces near linear fluorescent luminaries, while Rykowski characterized incandescent and high-intensity discharge lamps for non-imaging optical design using ray-tracing techniques. Other examples include near-field goniophotometers designed and constructed for high-intensity discharge lamps, automotive headlights, and computer graphics applications.