Showing papers by "Carl Michael Stahle published in 2002"

Multiplexing of Radio-Frequency Single Electron Transistors

Abstract: We present results on wavelength division multiplexing of radio-frequency single-electron transistors. We use a network of resonant impedance matching circuits to direct applied rf carrier waves to different transistors depending on carrier frequency. A two-channel demonstration of this concept using discrete components successfully reconstructed input signals with small levels of crosscoupling. A lithographic version of the rf circuits had measured parameters in agreement with electromagnetic modeling, with reduced crosscapacitance and inductance, and should allow 20–50 channels to be multiplexed.

1/f Noise and hot electron effects in variable range hopping conduction

Abstract: In the course of developing microcalorimeters as detectors for astronomical X-ray spectroscopy, we have undertaken an empirical characterization of non-ideal effects in the doped semiconductor thermometers used with these detectors, which operate at temperatures near 50 mK. We have found three apparently independent categories of such behavior that are apparently intrinsic properties of the variable-range hopping conduction mechanism in these devices: 1/f fluctuations in the resistance, which seems to be a 2D effect; a departure from the ideal coulomb-gap temperature dependence of the resistance at temperatures below T 0 /24; and an electrical nonlinearity that has the time dependence and extra noise that are quantitatively predicted by a simple hot electron model. This work has been done largely with ion-implanted Si:P:B, but similar behaviors have been observed in transmutation doped germanium.

Performance of Mo/Au TES microcalorimeters

Abstract: We are developing X-ray calorimeters to meet the specifications of the Constellation-X mission. Each calorimeter consists of a transition-edge-sensor (TES) thermometer, which is suspended on a silicon-nitride membrane. Our TES thermometers are Mo/Au bilayer films that are biased in the sharp phase transition between the superconducting and normal-metal states. These calorimeters have demonstrated very good energy resolutions: 2.4 eV at 1.5 keV and 3.7 eV at 3.3 keV. The energy resolutions are limited by thermal noise and Johnson noise (which are intrinsic to any resistive calorimeter) plus excess noise. The excess noise, which is several times larger than the Johnson noise, is consistent with frequency-independent voltage noise in the TES. Detailed measurements of one Mo/Au TES demonstrate that the excess noise is independent of the voltage applied to the TES over a range of biases at the same TES resistance. The magnitude of the excess noise is smallest at the high-resistance end of the phase transition. We also compared noise in square Mo/Au TES’s ranging in size from 300 microns to 600 microns to learn how the excess noise is affected by the geometry of the TES.

Hot-electron model in doped silicon thermistors

Abstract: Non-ohmic behavior of doped silicon and germanium can be empirically explained using a hot-electron model, which is motivated by the hot-electron effect in metals at low temperatures. This model assumes that the thermal coupling between electrons and lattice at low temperatures is weaker than the coupling between electrons, so that the electric power applied to the electrons raises them to a higher temperature than the lattice. Although this model seems not suitable for semiconductors in the variable range-hopping regime, where the electrons are localized, it fits quite well the experimental data. To determine whether the hot-electron model in doped semiconductor is just an alternative way to parameterize the data or has some physical validity, we investigated the noise and the frequency-dependence of the impedance of doped silicon thermistors that are used for low temperature thermal X-ray detectors. The measured excess white noise at low frequencies is consistent with the predicted thermodynamic fluctuations of energy between electron and phonon systems. The non-ohmic behavior shows a characteristic time that can be interpreted as a C/G time constant in the hot-electron model. By measuring this time constant, we get a hot-electron heat capacity C that agrees with the measured excess heat capacity of the implants. These support the assumption of a hot-electron system thermally separated from the lattice system.

Detailed characterization of Mo/Au TES microcalorimeters

Abstract: We are optimizing Mo/Au transition-edge-sensor (TES) calorimeters to meet the specifications of NASA’s Constellation-X mission. Our calorimeters have already demonstrated very good energy resolution of X rays (2.4 eV at 1.5 keV). We wish to further improve the energy resolution by reducing excess noise in the calorimeters. Development of a detailed model and understanding of the noise is instrumental to reaching this goal. Towards that end, we employ a linear model that describes the response of a calorimeter to signal and various sources of noise. The model is based on detailed measurements of the parameters that affect the calorimeter’s performance, such as current-voltage characteristics of the TES, thermal conductance of our silicon-nitride membranes, and inductance in the electronic circuit used to bias the TES. We determine the sharpness of the superconducting phase transition by fitting the model to the measured responsivity of the calorimeter. The model relates sources of noise, such as phonon noi...

First results from Position-Sensitive quantum calorimeters using Mo/Au Transition-Edge Sensors

Abstract: We report the first results from a high-energy-resolution imaging spectrometer called a Position-Sensitive Transition-Edge Sensor (PoST). A PoST is a quantum calorimeter consisting of two Transition Edge Sensors (TESs) on the ends of a long absorber to do one dimensional imaging spectroscopy. Comparing rise time and energy information, the position of the event in the PoST is determined. Energy is inferred from the sum of the two pulses. We have fabricated 7- and 15-pixel PoSTs using Mo-Au TESs and Au absorbers. We have achieved 32 eV FWHM energy resolution at 1.5 keV with a 7-pixel PoST calorimeter.

Non-ideal effects in doped semiconductor thermistors

Abstract: Semiconductor thermistors have been used for several years and their ideal behavior is well known both experimentally and theoretically. Their current performance is limited by non-ideal behaviors. These include 1/f noise and non-ohmic effects. We find that the 1/f noise appears to be a 2-D effect, and can be greatly reduced by fabricating thicker thermistors. Eliminating this noise could improve the intrinsic detector resolution as much as 40%. It also allows us to study other sources of excess noise in the thermometer. The non-ohmic behavior can be empirically explained using a hot-electron model. Although this model does not seem suitable for semiconductors in the variable range-hopping regime, where the electrons are localized, it fits the experimental data quite well. We measured an excess white noise at low frequencies consistent with the predicted thermodynamic fluctuations between electrons and phonons. We also measured a characteristic time of the non-ohmic behavior that is consistent with a C/G ...

Antenna-Coupled Superconducting Tunnel Junctions with Single-Electron Transistor Readout for Detection of Sub-Mm Radiation

Abstract: Antenna-coupled superconducting tunnel junction detectors have the potential for photon-counting sensitivity at sub-mm wavelengths. The device consists of an antenna structure to couple radiation into a small superconducting volume and cause quasiparticle excitations, and a single-electron transistor to measure currents through tunnel junction contacts to the absorber volume. We will describe optimization of device parameters, and recent results on fabrication techniques for producing devices with high yield for detector arrays. We will also present modeling of expected saturation power levels, antenna coupling, and rf multiplexing schemes.

1/f noise in doped semiconductor thermistors

Abstract: We have characterized the 1/f noise in standard ion-implanted silicon thermistors, which are about 250 nm thick. We find that it is associated with the bulk of the implant, and is interpretable as a ∆R/R fluctuation that is independent of the bias and depends only on the doping density and resistivity, or electron temperature. This excess noise is large enough that it has a significant effect on the energy resolution or NEP of a detector using these thermistors. The very steep temperature dependence of the 1/f noise suggested that it might be related to the conduction becoming two- dimensional, and we have fabricated thicker detectors to test this hypothesis. Similar doped silicon thermistors that are 1500 nm thick show negligible 1/f noise, but otherwise behave identically to the thinner thermistors of the same volume. This simple change could provide a 40% improvement in resolution for some existing X-ray detectors.

Background Measurements from Balloon-Borne CZT Detectors

Abstract: We report detector characteristics and background measurements from two prototype imaging CZT detectors flown on a scientific balloon payload in May 2001. The detectors are both platinum-contact 10mm x 10mm x 5mm CZT crystals, each with a 4 $\times$ 4 array of pixels tiling the anode. One is made from IMARAD horizontal Bridgman CZT, the other from eV Products high-pressure Bridgman material. Both detectors were mounted side-by-side in a flip-chip configuration and read out by a 32-channel IDE VA/TA ASIC preamp/shaper. We enclosed the detectors in the same 40deg field-of-view collimator (comprisinga graded passive shield and plastic scintillator) used in our previously-reported September 2000 flight. I-V curves for the detectors are diode-like, and we find that the platinum contacts adhere significantly better to the CZT surfaces than gold to previous detectors. The detectors and instrumentation performed well in a 20-hour balloon flight on 23/24 May 2001. Although we discovered a significant instrumental background component in flight, it was possible to measure and subtract this component from the spectra. The resulting IMARAD detector background spectrum (from 30 keV to ~450 keV) reaches ~5 x 10^{-3}$ counts/cm^2 -sec-keV at 100 keV and has a power-law index of ~2 at high energies. The eV Products detector has a similar spectrum, although there is more uncertainty in the energy scale because of calibration complications.

Wavelength division multiplexing scheme for radio-frequency single electron transistors

Abstract: We describe work on a wavelength division multiplexing scheme for radio-frequency single electron transistors. We use a network of resonant impedance matching circuits to direct applied rf carrier waves to different transistors depending on carrier frequency. Using discrete components, we made a two-channel demonstration of this concept and successfully reconstructed input signals with small levels of cross coupling. A lithographic version of the rf circuits had measured parameters in agreement with electromagnetic modeling, with reduced cross capacitance and inductance, and should allow 20 to 50 channels to be multiplexed.