Figure 1

(Color online) Calorimeter operation. Energy is deposited in the absorber and the thermometer reads out the change in temperature. The absorber cools down slowly through the weak link to the thermal heat sink. The tungsten transition-edge sensors discussed in this paper are quantum calorimeters, with the tungsten electrons acting as both the absorber and thermometer. A voltage bias keeps the tungsten electrons on the edge of the superconducting-to-normal transition, and the steep dependence of resistance on temperature provides precise thermometry. The anomalously low electron-phonon coupling in tungsten is the weak thermal link.Reuse & Permissions

Figure 2

(Color online) Histogram of corrected pulse heights. The sensor was illuminated with a 1550-nm (0.8-eV) gain-switched laser diode pulsed at 50 kHz with a pulse duration of 4 ns. The pulses from the detector were shaped with a $2\text{−}\mu \mathrm{s}$ semi-Gaussian filter and the maximum value for each pulse was determined. (For events in the zero-photon peak, the amplitude of the signal was measured at the expected arrival time of the photon.) The relationship between filtered pulse height and energy was determined by measuring the mean of each histogram peak, and the data were then corrected to linearize the response of the detector. The full width at half maximum of the zero-, one-, two-, three-, and four-photon peaks are 0.13, 0.20, 0.25, 0.34, and 0.45 eV, respectively. The inset shows typical unfiltered pulses for zero-, one-, two-, three-, and four-photon events.Reuse & Permissions

Figure 3

Experimental apparatus for the detector efficiency measurement. Light from the laser was split into two paths by a fiber coupler. The top path contained three programmable attenuators. A calibrated power meter at point $A$ was used to measure the actual attenuation value at every attenuation set point used in the experiment. After the attenuators were calibrated, the fiber exiting the last attenuator was cut at point $A$ and fused to the fiber (point $B$) leading to the detector. The power in the bottom arm of the fiber coupler was continuously monitored to ensure that the laser power did not drift. All measurements were performed within the linear range of the power meter.Reuse & Permissions

Figure 4

Detection efficiency. The efficiency was measured at several different average power levels with the laser in continuous-wave operation. The error bars show the statistical counting errors only. The slope of a weighted linear fit to the data is zero within the error bars, indicating that there is no dependence of efficiency on power level. A weighted average of the data yields an efficiency of 88.6% with a combined uncertainty of 0.4%.Reuse & Permissions