QTu3E.1.pdf CLEO Technical Digest © OSA 2012 Transition edge sensors with low timing jitter at 1550 nm Antia Lamas-Linares 1 , Thomas Gerrits 1 , Nathan A. Tomlin 1 , Adriana E. Lita 1 , Brice Calkins 1 ,J¨ orn Beyer 2 , Richard P. Mirin 1 , Sae Woo Nam 1 1 National Institute of Standards and Technology, 325 Broadway, Boulder CO 80305, USA 2 Phys. - Tech.Bundesanstalt (PT B), Berlin, Germany antia.lamas-linares@nist.gov Abstract: Transition edge sensors have demonstrated near unity detection efficiency of single photons and photon number resolving ability over a wide wavelength range (≈ 700 - 1600nm). They are also known to be slow, with timing jitter values in the ≈ 100 ns range, and the best current values in the ≈ 20ns range. We report measured jitter times below 5 ns for a tungsten TES. This is particularly relevant for experiments clocked from an 80 MHz Ti:Sa laser. © 2011 Optical Society of America 1. Introduction Superconducting transmission edge sensors (TES) for visible and near infra-red regions have shown near unity de- tection efficiency and very low error photon number resolving power [1,2], making them near ideal tools in quantum optics. One obstacle to a wider adoption is that they are considerably slower than superconducting nanowire single photon detectors or avalanche photodiodes [3]. In particular, the jitter time is not compatible with the ≈ 80 MHz repetition rate of many multiphoton experiments. Here we demonstrate a tungsten device and readout system with a FWHM (Full Width Half Maximum) detection time distribution of < 5 ns, thus opening the range of applicability to an important class of experimental settings. 2. Device details and readout The device under test is a tungsten TES optimized for detecting photons at a wavelength near 800nm. It is a 25 × 25 μ m square with a thickness of 20 nm and a weak cavity produced by an optical structure consisting of a metallic mirror and 2 dielectric layers on both sides of a trilayer of amorphous-Si/W/amorphous-Si [1]. We have not measured the detection efficiency (DE) of this particular device, but a companion from the same wafer was found to have a detection efficiency of 97%. Previous experience shows little variation in DE within a wafer. The device is installed in a commercial dilution refrigerator and tested at a temperature of 30 mK. The TES is wire-bonded directly to a single stage SQUID (Superconducting Quantum Interference Device) series array current sensor [4]; and read out by commercial high bandwidth amplifiers. These particular SQUIDS were chosen for their low input inductance of 3 nH. This configuration minimizes total inductance in the TES bias circuit, which has been a limiting factor in previous measurements of jitter with our devices. 3. Jitter measurements The TES was illuminated via a commercial telecom single mode fiber with laser pulses at a center wavelength of 1552 nm to allow energy resolved jitter analysis at 0.8 eV steps. Pulse duration was set to 1 ns, and a low repetition rate of 100kHz was used to avoid pulse pile up artifacts during measurement. The input level was adjusted so that most of the pulses contained no photons, but still maintaining a significant fraction of 2-photon and 3-photon pulses for photon number resolved jitter analysis. The output pulses from the amplifier electronics were recorded with a digitizing oscilloscope at a sampling rate of 1.25 GHz. The nominal bandwidth of the amplifying electronics was set at 20 MHz. Figure 1 shows the results of the jitter measurements with two different data processing methods. The first is a pure thresholding measurement, where we histogram the time at which the raw signal crosses a given threshold. Since we