Fig. 1. Our 18 mW, 402-nm laser.
The Laser
At the starting point of these experiments is the production of a pair of photons. They are produced by the process of spontaneous parametric down conversion. This is a well known nonlinear effect where two photons are produced from one “pump” photon. Since these photons are produced simultaneously (to within the length of the pump photon's wave packet), they are correlated in time. Coincidence detection of the two photons can be used to convince us of the quantum nature of the source, although more stringent conditions can be set up. The efficiency of this detection is the product of the efficiencies of the two detectors. Avalanche photodiodes (APD) have the highest efficiencies for single-photon detection. Current efficiencies of single-photon APD’s peak around 80% for 700 nm, dropping off rather quickly: above 1000 nm they are below 10%.1 This imposes severe restrictions on the pump source. The ideal one, widely used for research, is a UV argon-ion laser (350 nm). There are other possibilities, but for that wavelength currently there are no cheaper alternatives. We used two compromises: a 18-mW 402.36-nm GaN diode laser, shown below in Fig. 1, and a 100-mW 457.9-nm argon-ion laser, shown in Fig. 2.
APD’s have an efficiency of about 60% at twice the wavelength of the blue laser, at 804.72 nm. We bought this laser (Power Technologies) to do the experiments with "biphotons." The laser is very compact. Its price ($6k) includes current and temperature stability, plus optics to make the beam profile circular. This laser worked out very nicely. The APS's are very efficient at 804 nm. We note that the manufacturer now offers even better alternatives.2
Fig. 1. Our 18 mW, 402-nm laser.
Another compromise is possible for those who have a multiline argon-ion laser. We had a Coherent INNOVA-90 that was donated to us by Bell Labs.3 We used it to pump a cw dye laser. We got a wavelength selector and made the laser lase at its shortest wavelength 459 nm. We could get as high as 200 mW with at this wavelength. The APD efficiency for twice this wavelength, 915.8 nm, is 30%. With it we have been able to do very nice experiments with single photons.
Fig. 2 Our argon-ion laser. It is big and expensive, but it worked out very well
for us.
2Power Technology Inc. now offers 375-nm, 10mW and 405-nm, 60 mW lasers.
3Thanks to Scott Hinton and Steve Hinterlong.
E.J. Galvez/Colgate U.