A wide bandwidth detection and display system for use with TEA CO2 lasers , A. F. Gibson, M. F. Kimmitt, P. N. D. Maggs, and B. Norris, Journal of Applied Physics, Vol. 46, No. 3, March 1975
(See below to download the original paper)
TEA CO2 lasers were invented in the late 1960s by Beaulieu in Canda. By discharging tens of kilovolts across a centimetre or so of a CO2, nitrogen and helium gas mixture at atmospheric pressure, probed by a suitable laser cavity, it was possible to generate pulses of 10.6 micron radiation at megawatt powers. The pulses were short, only a few hundred nanoseconds long, but by using mode-locking, pulses of less than a nanosecond duration could be obtained. This was exciting, because it meant that high speed phenomena taking place in various semiconductors could be probed optically. However, there was a problem. In order for any measurements to be of practical use, it was necessary to know the precise time-varying shape of the laser pulse. At that time, only pyroelectric detectors and photon-drag detectors could operate with response times faster than one nanosecond at 10 micron wavelengths. Pyroelectric detectors were intrinsically capacitative, and limited to around 500 ps or so at best. Photon-Drag detectors, using P-type germanium, had an intrinsic response limit in the picosecond regime, but could produce only 100 mv signals requiring amplification to be viewed on an oscilloscope. The bandwidth of the oscilloscope's amplifier had to be fast enough to amplify the signal without distortion in order to drive the deflection plates of the cathode ray tube.
The fastest real-time oscilloscope available at the time was the Tektronix 7904 with a bandwidth of 500 MHz. This limited the effective response time to around 800 picoseconds. There was a way of taking response times at least a factor of two faster, but this involved using the Tektronix 519 oscilloscope. It had a bandwidth of 1GHz. There was a rumour, possibly apocryphal, that it was developed to observe very fast phenomena for the US nuclear weapons programme. The Tektronix 519 dated from the early '60s, and the reason it was so fast was that it did not use a vertical amplifier. The input signal had to be large enough to drive the vertical deflector plates directly, and these were arranged in the form of a travelling wave transmission line. In order to get a decently visible 1 cm deflection on the screen, a 10 volt signal was needed. Since photon drag detectors could only provide 100 mv, there was a factor of 100 adrift in signal strength.
My work on photoconductivity at 10 micron wavelengths, showed that the effect in P-type germanium had some similarities to photon-drag and it had a similar intrinsic speed. Photoconductivity, the change in conductance resulting from the presence of photon-induced charge carriers, could provide an arbitrarily large signal by increasing the externally-applied bias voltage. Photon-drag was an intrinsic effect and did not require a bias voltage. However, since the photoconductive signal could be increased simply by increasing the applied voltage, here was a possibility for producing a fast signal, large enough to directly drive the deflector plates of the 519 oscilloscope.
several hundred volts of bias was needed to produce a decent signal. The voltage had to be pulsed, because the thermal dissipation in the germanium crystal would have been around 5kW. Electronics was not my forte, and the first attempt at producing 500 V pulses for the detector bias used a relay. The thermal shock during the several ms pulse was such that the first time it was switched on the germanium detector exploded in a shower of bits. A high voltage transistor circuit was eventually designed with the help of some colleagues, and the application of microsecond pulses of voltage was able to provide the required bias without the pyrotechnical side effects.
Brian Norris, an old friend, a research student from Ardrossan, built the short pulse laser, and we were able to measure 380 ps pulses at 10.6 microns; a world record at the time.
Cick below to read the note.
(See below to download the original paper)
TEA CO2 lasers were invented in the late 1960s by Beaulieu in Canda. By discharging tens of kilovolts across a centimetre or so of a CO2, nitrogen and helium gas mixture at atmospheric pressure, probed by a suitable laser cavity, it was possible to generate pulses of 10.6 micron radiation at megawatt powers. The pulses were short, only a few hundred nanoseconds long, but by using mode-locking, pulses of less than a nanosecond duration could be obtained. This was exciting, because it meant that high speed phenomena taking place in various semiconductors could be probed optically. However, there was a problem. In order for any measurements to be of practical use, it was necessary to know the precise time-varying shape of the laser pulse. At that time, only pyroelectric detectors and photon-drag detectors could operate with response times faster than one nanosecond at 10 micron wavelengths. Pyroelectric detectors were intrinsically capacitative, and limited to around 500 ps or so at best. Photon-Drag detectors, using P-type germanium, had an intrinsic response limit in the picosecond regime, but could produce only 100 mv signals requiring amplification to be viewed on an oscilloscope. The bandwidth of the oscilloscope's amplifier had to be fast enough to amplify the signal without distortion in order to drive the deflection plates of the cathode ray tube.
The fastest real-time oscilloscope available at the time was the Tektronix 7904 with a bandwidth of 500 MHz. This limited the effective response time to around 800 picoseconds. There was a way of taking response times at least a factor of two faster, but this involved using the Tektronix 519 oscilloscope. It had a bandwidth of 1GHz. There was a rumour, possibly apocryphal, that it was developed to observe very fast phenomena for the US nuclear weapons programme. The Tektronix 519 dated from the early '60s, and the reason it was so fast was that it did not use a vertical amplifier. The input signal had to be large enough to drive the vertical deflector plates directly, and these were arranged in the form of a travelling wave transmission line. In order to get a decently visible 1 cm deflection on the screen, a 10 volt signal was needed. Since photon drag detectors could only provide 100 mv, there was a factor of 100 adrift in signal strength.
My work on photoconductivity at 10 micron wavelengths, showed that the effect in P-type germanium had some similarities to photon-drag and it had a similar intrinsic speed. Photoconductivity, the change in conductance resulting from the presence of photon-induced charge carriers, could provide an arbitrarily large signal by increasing the externally-applied bias voltage. Photon-drag was an intrinsic effect and did not require a bias voltage. However, since the photoconductive signal could be increased simply by increasing the applied voltage, here was a possibility for producing a fast signal, large enough to directly drive the deflector plates of the 519 oscilloscope.
several hundred volts of bias was needed to produce a decent signal. The voltage had to be pulsed, because the thermal dissipation in the germanium crystal would have been around 5kW. Electronics was not my forte, and the first attempt at producing 500 V pulses for the detector bias used a relay. The thermal shock during the several ms pulse was such that the first time it was switched on the germanium detector exploded in a shower of bits. A high voltage transistor circuit was eventually designed with the help of some colleagues, and the application of microsecond pulses of voltage was able to provide the required bias without the pyrotechnical side effects.
Brian Norris, an old friend, a research student from Ardrossan, built the short pulse laser, and we were able to measure 380 ps pulses at 10.6 microns; a world record at the time.
Cick below to read the note.
| hot_hole_detector.pdf |