High Spectral Resolution High Cadence Imaging Xray Microcalorimeter

High Spectral Resolution, High Cadence, Imaging X-ray Microcalorimeter Arrays for Solar Physics Simon Bandler, Catherine Bailey, Jay Chervenak, Megan Eckart, Fred Finkbeiner, Caroline Kilbourne, Daniel Kelly, Richard Kelley, F. Scott Porter, Jack Sadleir, Stephen Smith - X-ray Astrophysics Laboratory at GSFC Jay Bookbinder, Ed De. Luca, Randall Smith - SAO Supported by NASA ROSES: Solar & Heliospheric Physics • Original motivation: RAM - Microcalorimeters to study dynamics & energetics of solar corona • Simultaneous imaging, spectroscopy & high cadence • Pixels need to be small and fast

Resistance How do TES microcalorimeters work ? Transition at ~ 100 m. K and less than 1 m. K wide. Temperature V TES Thermal relaxation time: shunt Thermal conductance SQUID

Solar TES microcalorimeters design : on solid substrate Pixel designs : Current Absorber contact region Gold absorber Interdigitated gold stripes ( 35 mm TES) Thermal link (G): Kapitza

How small are the pixels ? 42 mm 9. 1 mm

What energy resolution is achievable ? • Mn Ka 1 & Ka 2 x-rays at 6 ke. V from an 55 Fe internal conversion source • Instrumental broadening consistent with a gaussian response with 2. 13 e. V resolution FWHM

Heat bath temp. ~ 50 m. K

Resistance What is the linearity of the response like ? Temperature 35 mm TES, 57 mm absorber

Possibility to detect higher energy photons: Energy information from time in saturation Saturated pulses 10 - 50 ke. V

• Thermal ground plane - thermal cross-talk < 0. 01% between nearest neighbors • No collimation !!! • Superconducting plane under TES removes all sensitivity to external magnetic fields

What count rates are achievable with high energy resolution ? Optimal filtering : 35 mm device: (The f=0 term in the optimal filter, which must be discarded, contains less info as record grows longer).

What count rates are achievable with high energy resolution ? tetf = 210 ms tc = 27 ms Dead time = T 1 + T 2 T 1: So not on tail of previous pulse T 2: Pulse record length for optimal filter T 1: = 9 tc, 10 tc, 12 tc T 2: = 300 ms, 500 ms and 1500 ms

Next : Close-pack arrays TES size Absorber Size Gap with 75 µm pitch Pitch of stripline for 32 x 32 array 12 µm 34 µm x 34 µm 63 µm ~ 8 µm 20 µm 42 µm x 42 µm 50 µm ~ 6 µm 35 µm 57 µm x 57 µm 40 µm ~ 5 µm 50 µm 75 µm x 75 µm 25 µm ~ 3 µm ~ N/4 wire pairs per muntin for NXN array 8 wire pairs for 32 x 32 array (More exactly : (N-2) 2 / (4*(N-3)), - goes to N/4 for large N) Prototype IXO 32 x 32 array Plans : Build 8 x 8, 20 x 20, 32 x 32, 64 x 64 arrays of close-packed pixels

Why X-ray microcalorimeters for solar physics ? • Detect & resolve hundreds of atomic transition lines from different ionization states of many elements in the solar atmosphere • Determine the abundances of elements from carbon to nickel relative to hydrogen • Determine thermal & non-thermal electron distributions from both lines & continuum • Determine plasma densities from ~ 109 to 1014 cm-3 • Determine temperature from < 0. 7 to >10 million degrees • Detect flows with velocities down to < 50 km/s • Precision timing of photons from flares

Conclusions: • Excellent energy resolution achieved in small, low heat capacity x-ray microcalorimeters ~ 2 e. V • Count rates > 300 cps achievable • No energy resolution broadening due to being on solid substrate • Design suitable for Solar Physics

Very stable :

Assume exponential decay time = 150 µs. Taucrit ~ 50 µs T 1 = 7*Taucrit = 350 µs and Trec = 650 µs max output rate = 1/[(T 1+Trec)*e(1)] = 368 cps for input rate = 1/[(T 1+Trec)] = 1000 cps This gives optimal throughout of 1/e = 37% acceptance

Maybe a box car filter is sufficient ? Time for 100 samples is 100 µs.

Suggested requirements : 50 µm => 4” for a 3 m focal length (FOV=1’ => 15 x 15 array) => 0. 8” for a 15 m focal length

CDM Plausible: Pulse: 60 u. A size, Tau before damping ~ 200 us, Tau_crit~ 27 us Time to peak ~ 3 Tau_crit ~ 80 us Assume 160 ns row times & 32 pixels muxed => Sample rate of 5 us, This Nyquist sample rate would give us 16 samples on the rise. For 60 u. A pulse, the rise between samples might be ~ 60/16 = 3. 75 u. A. M of the chip is ~ 35 p. H. 3. 75 u. A * 35 p. H = 0. 063 Phi 0. This appears to be a small enough error signal (normally try to keep error < Phi 0/4).

Polarization sensitivity using microcalorimeters behind dichroic crystals :