Engineering topics
Chervenak, James
Publications and source records attributed to Chervenak, James.
The Impact of Transition Edge Sensor Design on Achievable Performance Uniformity of Kilo-Pixel Arrays
Future astronomy missions using x-ray transition-edge sensor (TES) microcalorimeters, such as X-IFU on Athena, will require large arrays of 1000s of pixels fabricated on a single wafer. To wire out so many pixels the current array designs have pixels with different rotational orientations. Fabrication is done in multiple layers and so, dependent on method, there is potential for spatial misalignment between layers. Because of the variation of orientation of pixels, misalignment may not impact each pixel equally. This has the potential to degrade the achievable uniformity of performance across an array. How well aligned do different layers need to be? How does sensitivity to misalignment depend on choice of pixel design?
Flexible Microstrip Circuits for Superconducting Electronics
Flexible circuits with superconducting wiring atop polyimide thin films are being studied to connect large numbers of wires between stages in cryogenic apparatus with low heat load. The feasibility of a full microstrip process, consisting of two layers of superconducting material separated by a thin dielectric layer on 5 mil (approximately 0.13 mm) Kapton sheets, where manageable residual stress remains in the polyimide film after processing, has been demonstrated. The goal is a 2-mil (approximately 0.051-mm) process using spin-on polyimide to take advantage of the smoother polyimide surface for achieving highquality metal films. Integration of microstrip wiring with this polyimide film may require high-temperature bakes to relax the stress in the polyimide film between metallization steps.
Fabrication Method for LOBSTER-Eye Optics in <110> Silicon
Soft x-ray optics can use narrow slots to direct x-rays into a desirable pattern on a focal plane. While square-pack, square-pore, slumped optics exist for this purpose, they are costly. Silicon (Si) is being examined as a possible low-cost replacement. A fabrication method was developed for narrow slots in <110> Si demonstrating the feasibility of stacked slot optics to replace micropores. Current micropore optics exist that have 20-micron-square pores on 26-micron pitch in glass with a depth of 1 mm and an extent of several square centimeters. Among several proposals to emulate the square pore optics are stacked slot chips with etched vertical slots. When the slots in the stack are positioned orthogonally to each other, the component will approach the soft x-ray focusing observed in the micropore optics. A specific improvement Si provides is that it can have narrower sidewalls between slots to permit greater throughput of x-rays through the optics. In general, Si can have more variation in slot geometry (width, length). Further, the sidewalls can be coated with high-Z materials to enhance reflection and potentially reduce the surface roughness of the reflecting surface. Narrow, close-packed deep slots in <110> Si have been produced using potassium hydroxide (KOH) etching and a patterned silicon nitride (SiN) mask. The achieved slot geometries have sufficient wall smoothness, as observed through scanning electron microscope (SEM) imaging, to enable evaluation of these slot plates as an optical element for soft x-rays. Etches of different angles to the crystal plane of Si were evaluated to identify a specific range of etch angles that will enable low undercut slots in the Si <110> material. These slots with the narrow sidewalls are demonstrated to several hundred microns in depth, and a technical path to 500-micron deep slots in a precision geometry of narrow, closepacked slots is feasible. Although intrinsic stress in ultrathin wall Si is observed, slots with walls approaching 1.5 microns can be achieved (a significant improvement over the 6-micron walls in micro - pore optics). The major advantages of this technique are the potential for higher x-ray throughout (due to narrow slot walls) and lower cost over the existing slumped micropore glass plates. KOH etching of smooth sidewalls has been demonstrated for many applications, suggesting its feasibility for implementation in x-ray optics. Si cannot be slumped like the micropore optics, so the focusing will be achieved with millimeter-scale slot plates that populate a spherical dome. The possibility for large-scale production exists for Si parts that is more difficult to achieve in micropore parts.
Bismuth Passivation Technique for High-Resolution X-Ray Detectors
The Athena-plus team requires X-ray sensors with energy resolution of better than one part in 3,000 at 6 keV X-rays. While bismuth is an excellent material for high X-ray stopping power and low heat capacity (for large signal when an X-ray is stopped by the absorber), oxidation of the bismuth surface can lead to electron traps and other effects that degrade the energy resolution. Bismuth oxide reduction and nitride passivation techniques analogous to those used in indium passivation are being applied in a new technique. The technique will enable improved energy resolution and resistance to aging in bismuth-absorber-coupled X-ray sensors. Elemental bismuth is lithographically integrated into X-ray detector circuits. It encounters several steps where the Bi oxidizes. The technology discussed here will remove oxide from the surface of the Bi and replace it with nitridized surface. Removal of the native oxide and passivating to prevent the growth of the oxide will improve detector performance and insulate the detector against future degradation from oxide growth. Placing the Bi coated sensor in a vacuum system, a reduction chemistry in a plasma (nitrogen/hydrogen (N2/H2) + argon) is used to remove the oxide and promote nitridization of the cleaned Bi surface. Once passivated, the Bi will perform as a better X-ray thermalizer since energy will not be trapped in the bismuth oxides on the surface. A simple additional step, which can be added at various stages of the current fabrication process, can then be applied to encapsulate the Bi film. After plasma passivation, the Bi can be capped with a non-diffusive layer of metal or dielectric. A non-superconducting layer is required such as tungsten or tungsten nitride (WNx).
Magnetic Field Dependence of the Critical Current in S-N Bilayer Thin Films
Here we investigate the effects a non-uniform applied magnetic field has on superconducting transition-edge sensors (TESs) critical current. This has implications on TES optimization. It has been shown that TESs resistive transition can be altered by magnetic fields. We have observed critical current rectification effects and explained these effects in terms of a magnetic self-field arising from asymmetric current injection into the sensor. Our TES physical model shows that this magnetic self-field can result in significantly degraded or improved TES performance. In order for this magnetically tuned TES strategy to reach its full potential we are investigating the effect a non-uniform applied magnetic field has on the critical current.
The Magnetically Tuned Transition-Edge Sensor
Here we present our latest results on a magnetically tuned transition-edge sensor (MTES).
Fabrication of Compact Superconducting Lowpass Filters for Ultrasensitive Detectors
It is extremely important for current and future far-infrared and sub-millimeter ultrasensitive detectors, which include transition edge sensors (TES) and microwave kinetic inductance detectors, to be adequately filtered from stray electromagnetic radiation in order to achieve their optimal performance. One means of filtering stray radiation is to block leakage associated with electrical connections in the detector environment. Here we discuss a fabrication methodology for realizing non-dissipative planar filters imbedded in the wall of the detector enclosure to limit wave propagation modes up to far-infrared frequencies. Our methodology consists of fabricating a boxed stripline transmission line, in which a superconducting (Nb, Mo, or Al) transmission line is encased in a silicon dioxide dielectric insulator coated with a metallic shell. We report on achieved attenuation and return loss and find that it replicates the simulated data to a high degree.
Fabrication of a Cryogenic Terahertz Emitter for Bolometer Focal Plane Calibrations
A fabrication process is reported for prototype emitters of THz radiation, which operate cryogenically, and should provide a fast, stable blackbody source suitable for characterization of THz devices. The fabrication has been demonstrated and, at the time of this reporting, testing was underway. The emitter is similar to a monolithic silicon bolometer in design, using both a low-noise thermometer and a heater element on a thermally isolated stage. An impedance-matched, high-emissivity coat ing is also integrated to tune the blackbody properties. This emitter is designed to emit a precise amount of power as a blackbody spectrum centered on terahertz frequencies. The emission is a function of the blackbody temperature. An integrated resistive heater and thermometer system can control the temperature of the blackbody with greater precision than previous incarnations of calibration sources that relied on blackbody emission. The emitter is fabricated using a silicon- on-insulator substrate wafer. The buried oxide is chosen to be less than 1 micron thick, and the silicon device thickness is 1-2 microns. Layers of phosphorus compensated with boron are implanted into and diffused throughout the full thickness of the silicon device layer to create the thermometer and heater components. Degenerately doped wiring is implanted to connect the devices to wire-bondable contact pads at the edge of the emitter chip. Then the device is micromachined to remove the thick-handle silicon behind the thermometer and heater components, and to thermally isolate it on a silicon membrane. An impedance- matched emissive coating (ion assisted evaporated Bi) is applied to the back of the membrane to enable high-efficiency emission of the blackbody spectrum.
Fabrication of a Cryogenic Bias Filter for Ultrasensitive Focal Plane
A fabrication process has been developed for cryogenic in-line filtering for the bias and readout of ultrasensitive cryogenic bolometers for millimeter and submillimeter wavelengths. The design is a microstripline filter that cuts out, or strongly attenuates, frequencies (10 50 GHz) that can be carried by wiring staged at cryogenic temperatures. The filter must have 100-percent transmission at DC and low frequencies where the bias and readout lines will carry signal. The fabrication requires the encapsulation of superconducting wiring in a dielectric-metal envelope with precise electrical characteristics. Sufficiently thick insulation layers with high-conductivity metal layers fully surrounding a patterned superconducting wire in arrayable formats have been demonstrated. A degenerately doped silicon wafer has been chosen to provide a metallic ground plane. A metallic seed layer is patterned to enable attachment to the ground plane. Thick silicon dioxide films are deposited at low temperatures to provide tunable dielectric isolation without degrading the metallic seed layer. Superconducting wiring is deposited and patterned using microstripline filtering techniques to cut out the relevant frequencies. A low Tc superconductor is used so that it will attenuate power strongly above the gap frequency. Thick dielectric is deposited on top of the circuit, and then vias are patterned through both dielectric layers. A thick conductive film is deposited conformally over the entire circuit, except for the contact pads for the signal and bias attachments to complete the encapsulating ground plane. Filters are high-aspect- ratio rectangles, allowing close packing in one direction, while enabling the chip to feed through the wall of a copper enclosure. The chip is secured in the copper wall using a soft metal seal to make good thermal and electrical contact to the outer shield.
Multicolor Detectors for Ultrasensitive Long-Wave Imaging Cameras
A document describes a zeptobolometer for ultrasensitive, long-wavelength sensors. GSFC is developing pixels based on the zeptobolometer design that sense three THz wavelengths simultaneously. Two innovations are described in the document: (1) a quasiparticle (QO) filter arrangement that enables a compact multicolor spectrum at the focal plane, and (2) a THz antenna readout by up to three bolometers. The innovations enable high efficiency by greatly reducing high, frequency-dependent microstrip losses, and pixel compactness by eliminating the need for bulky filters in the focal plane. The zeptobolometer is a small TES bolometer, on the scale of a few microns, which can be readily coupled through an impedance-matching resistor to a metal or dielectric antenna. The bolometer is voltage-biased in its superconducting transition, allowing the use of superconducting RF multiplexers to read out large arrays. The antenna is geometrically tapped at three locations so as to efficiently couple radiation of three distinct wavelengths to the individual TESs. The transition edge hot electrons in metals offer a simple, compact arrangement for antenna readout, which can be crucial in the THz where line losses at high frequencies can be substantial. A metallic grill filter acts as a high-pass filter and directs the low-frequency components to a location where they will be absorbed. The absorption spectrum shows that three well-separated THz bands are feasible. The filters can be made from high-purity dielectrics such as float zone silicon or sapphire.
Fabrication of a Kilopixel Array of Superconducting Microcalorimeters with Microstripline Wiring
A document describes the fabrication of a two-dimensional microcalorimeter array that uses microstrip wiring and integrated heat sinking to enable use of high-performance pixel designs at kilopixel scales (32 X 32). Each pixel is the high-resolution design employed in small-array test devices, which consist of a Mo/Au TES (transition edge sensor) on a silicon nitride membrane and an electroplated Bi/Au absorber. The pixel pitch within the array is 300 microns, where absorbers 290 microns on a side are cantilevered over a silicon support grid with 100-micron-wide beams. The high-density wiring and heat sinking are both carried by the silicon beams to the edge of the array. All pixels are wired out to the array edge. ECR (electron cyclotron resonance) oxide underlayer is deposited underneath the sensor layer. The sensor (TES) layer consists of a superconducting underlayer and a normal metal top layer. If the sensor is deposited at high temperature, the ECR oxide can be vacuum annealed to improve film smoothness and etch characteristics. This process is designed to recover high-resolution, single-pixel x-ray microcalorimeter performance within arrays of arbitrarily large format. The critical current limiting parts of the circuit are designed to have simple interfaces that can be independently verified. The lead-to-TES interface is entirely determined in a single layer that has multiple points of interface to maximize critical current. The lead rails that overlap the TES sensor element contact both the superconducting underlayer and the TES normal metal
Ultrasensitive Superconducting Transition Edge Sensors Based On Electron-Phonon Decoupling
We have successfully fabricated the superconducting transition edge sensor (TES), bolometer technology that centers on the use of electron-phonon decoupling (EPD) to thermally isolate the bolometer. Along with material characterization for large format antenna coupled bolometer arrays, we present the initial test results of bolometer based on EPD designed for THz detection. We have selected a design approach that separates the two functions of photon absorption and temperature measurement, allowing separate optimization of the performance of each element. We have integrated Molybdenum/Gold (Mo/Au) bilayer TES and ion assisted thermally evaporated (IAE) Bismuth (Bi) films as radiation absorber coupled to a low-loss microstripline from Niobium (Nb) ground plane to a twin-slot antenna structure. The thermal conductance and the time constant of these devices have been measured, and are consistent with our calculations. The device exhibits a single time constant at 0.1 K of approx.160 IlS, which is compatible with readout by a high-bandwidth single SQUID or a time domain SQUID multiplexer. The effects of thermal conductance and electrothermal feedback are major determinants of the time constant, but the electronic heat capacity also plays a major role. The NEP achieved in the device described above is 2.5x10(exp -17)W(gamma)Hz. Our plan is to demonstrate a reduction of the volume in the superconducting element to 5 microns x 5 microns in films of half the thickness at Tc = 60mK. By calculation, this new geometry corresponds to an NEP reduction of two orders of magnitude to 2.5x10(exp -19)W/(gamma)Hz, with a time constant of ~130/ls.
Development of Superconducting Transition Edge Sensors Based on Electron-Phonon Decoupling
We have successfully fabricated a superconducting transition edge sensor (TES), bolometer that centers on the use of electron-phonon decoupling (EPD) for thermal isolation. We have selected a design approach that separates the two functions of far-infrared and THz radiative power absorption and temperature measurement, allowing separate optimization of the performance of each element. We have integrated molybdenum/gold (Mo/Au) bilayer TES and ion assisted thermally evaporated (IAE) bismuth (Bi) films as radiation absorber coupled to a low-loss microstripline from niobium (Nb) ground plane to a twin-slot antenna structure. The thermal conductance (G) and the time constant for the different geometry device have been measured. For one such device, the measured G is 1.16 x 10(exp -10) W/K (plus or minus 0.61 x 10(exp- 10) W/K) at 60 mK, which corresponds to noise equivalent power (NEP) = 1.65 X 10(exp -18)W/vHz and time constant of approximately 5 microseconds.
Superconducting Detector Arrays for Astrophysics
The next generation of astrophysics instruments will feature an order of magnitude more photon sensors or sensors that have an order of magnitude greater sensitivity. Since detector noise scales with temperature, a number of candidate technologies have been developed that use the intrinsic advantages of detector systems that operate below 1 Kelvin. Many of these systems employ of the superconducting phenomena that occur in metals at these temperatures to build ultrasensitive detectors and low-noise, low-power readout architectures. I will present one such system in use today to meet the needs of the astrophysics community at millimeter and x-ray wavelengths. Our group at NASA in collaboration with Princeton, NIST, Boulder and a number of other groups is building large format arrays of superconducting transition edge sensors (TES) read out with multiplexed superconducting quantum interference devices (SQUID). I will present the high sensitivity we have achieved in multiplexed x-ray sensors with the TES technology and describe the construction of a 1000-sensor TES/SQUID array for microwave measurements. With our collaboration's deployment of a kilopixel TES array for 2 mm radiation at the Atacarna Cosmology Telescope in November 2007, we have first images of the lensed Cosmic Microwave Background at fine angular scales.
Modelling of TES X-ray Microcalorimeters with a Novel Absorber Design
Our development of a novel x-ray absorber design that has enabled the incorporation of high-conductivity electroplated gold into our absorbers has yielded devices that not only have achieved breakthrough performance at 6 keV, but also are extraordinarily well modelled. We have determined device parameters that reproduce complex impedance curves and noise spectra throughout transition. Observed pulse heights, decay time and baseline energy resolution were in good agreement with simulated results using the same parameters. In the presentation, we will show these results in detail and we will also show highlights of the characterization of our gold/bismuth-absorber devices. We will discuss possible improvement of our current devices and expected performance of future devices using the modelling results.
Heat Sinking, Cross Talk, and Temperature Stability for Large, Close-Packed Arrays of Microcalorimeters
We are developing large, close-packed arrays of x-ray transition-edge sensor (TES) microcalorimeters. In such a device, sufficient heat sinking is important to to minimize thermal cross talk between pixels and to stabilize the bath temperature for all pixels. We have measured cross talk on out 8 x 8 arrays and studied the shape and amount of thermal crosstalk as a function of pixel location and efficiency of electrothermal feedback. In this presentation, we will compare measurements made on arrays with and without a backside, heat-sinking copper layer, as well as results of devices on silicon-nitride membranes and on solid substrates, and we will discuss the implications for energy resolution and maximum count rate. We will also discuss the dependence of pulse height upon bath temperature, and the measured and required stability of the bath temperature.
Tunable Antireflection Layers for Planar Bolometer Arrays
It remains a challenge to obtain high-efficiency coupling of far-infrared through millimeter radiation to large-format detector arrays. The conventional approach of increasing detector coupling is to use reflective backshorts. However, this approach often results in excessive systematic errors resulting from reflections off the backshort edge. An alternate approach to both increasing quantum efficiency and reducing systematics associated with stray light is to place an antireflective coating near the front surface of the array. When incorporated with a resistive layer and placed behind the detector focal plane, the AR coating can serve to prevent optical ghosting by capturing radiation transmitted through the detector. By etching a hexagonal pattern in silicon, in which the sizes of the hexes are smaller than the wavelength of incident radiation, it is possible to fabricate a material that has a controllable dielectric constant, thereby allowing for simple tunable optical device fabrication. To this end, we have fabricated and tested tunable silicon "honeycomb" AR layers and AR/resistive layer devices. These devices were fabricated entirely out of silicon in order to eliminate problems associated with differential contraction upon detector cooling.