Generation-recombination Noise is seen in photo-conductors in which the absorbed photons produce both positive and negative charge carriers. Some of the free carriers may recombine before they are collected. Thermal excitation may generate additional carriers. Both the generation and recombination occur randomly, resulting in noise fluctuations in the output current.
Johnson Noise or thermal noise is caused by the random motion of carriers in a conductor. The result is fluctuations in the detector’s internal resistance, or in any resistance in series with the detector's terminals.
Flicker or 1/f Noise is not well understood. It occurs in detectors such as photoconductors which require a biasing current. Its magnitude is proportional to 1/fB where B is usually between 0.8 and 1.2.
Readout Noise is a characteristic of array detectors and is associated with the uncertainties introduced during the transfer of charges between storage registers.
In addition to the sources described above, detector signals can be subject to microphonic noise, caused by vibration or shock and by post detector electronic noise. Often the circuitry after the detector determine the lowest measurable signal, particularly for detectors which do not provide some internal amplification of the photocurrent. By contrast, the almost noise-free internal amplification of photomultipliers accounts for their superb performance.
Finally, detectors are subject to temperature noise caused by fluctuations in their temperature. This can be a problem for small thermal detectors with low thermal mass.
There are a lot of measurement situations when the knowledge of the exact magnitude of an effect is not needed. A relative reading is all that is necessary. In that case we only have to make sure that the inputs are higher than the NEP level and lower than the damage level. Measurements become much simpler if the inputs are also within the linear range of the detection system response.
However, when absolute measurements are required to quantify the light flux, calibrated instruments are needed. Our Calibration Laboratory has acquired a number of National Institute of Standards and Technology (NIST) source and detector calibration standards, as well as additional standards traceable to international standards’ setting and disseminating bodies. We are also in the enviable position of being able to utilize our own broad range of products: UV and IR light sources, detectors, electronics, optics, positioning equipment, monochromators, spectrographs, and FT-IRs, to provide single point or spectrally resolved calibration for most of the detector products we ship. For some detectors we offer optional PROMS that store a wavelength-responsivity table. We use our calibration transfer standards, lamps and detectors, typically traceable to the national Institute of Standards and Technology (NIST) to ensure meaningful and reproducible calibrations.
Thermal detectors work by converting the incident radiation into a temperature rise. The temperature change can be measured in several ways. Our detectors use either the voltage generated at the junction of dissimilar metals, or the pyroelectric effect. In either case, we have a “sensitive element” where we measure temperature change.
We blacken the sensitive element to enhance the absorption of the radiation. We choose the blackening material for its high and nearly uniform absorption, and hence detector responsivity, over a wide spectral range. This is a major advantage of thermal detectors.
The thickness of the black absorber is controlled to avoid adding excessive thermal mass to some detectors. High thermal mass slows the response time and typically increases NEP.
We offer two types of Thermal Detectors:
1. Thermopile Detectors for DC radiation.
2. Pyroelectric Detectors for pulsed, chopped or modulated radiation.
One way to increase the output voltage is to connect a number of the thermocouple junctions (typically 20 to 120) in series. All the “hot” junctions are placed close together to collect the radiation. This constitutes a thermopile. Thermopiles can be produced economically by thin film processes, yielding rugged units suitable for field use. These have time constants ranging down to below 50 ms, in small sizes with metal oxide blackening. Larger thermopiles, made with individual wire junctions backing up a highly absorbing black disk, have long time constants, more than a second in many cases. Please see Thermopile Sensor Physics for additional information.
Thermopiles exhibit no flicker, 1/f, noise, since no current bias is needed to operate them. Highly sensitive measurements can be made from DC to the few Hz frequency response limit of a particular device.
These devices are quite sensitive in the infrared, thanks to their broadband absorption. Therefore, care must be taken to stabilize their field of view, since all near room temperature objects, including people, emit significant IR. Measurements are best made by shuttering the radiation falling on the detector and observing the change in output voltage. It is not uncommon to see "negative" radiation if the source is cooler or less emissive than the shutter used to get “zero” reading.