Antonino Miceli FNAL Research Techniques Seminar April 24 2012 Outline Why Superconducting Detectors Detector Requirements Applications Overview of MKIDs MKID activities at Argonne MKID resonator readout ID: 733949
Download Presentation The PPT/PDF document "Microwave Kinetic Inductance Detectors f..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Microwave Kinetic Inductance Detectors for X-ray Science
Antonino
Miceli
FNAL
Research Techniques Seminar
April 24, 2012Slide2
Outline
Why Superconducting Detectors?
Detector Requirements
Applications
Overview of MKIDs
MKID activities at Argonne
MKID resonator readout Slide3
Acknowledgements
Tom Cecil (XSD Staff)
Orlando
Quaranta
(Post-doc
)
Lisa
Gades
(XSD Staff)
Outside Collaborators:
Professor Ben
Mazin
(UCSB
)
MSD/
UChicago
TES Team (
Novosad
et al)
CNM Cleanroom StaffSlide4
Superconductors Detectors for X-ray Detector R&D
Energy dispersive
semiconductor detectors
have almost reached their theoretical
limits
e.g., Silicon Drift Diodes have energy resolution ~
150
eV
at 6
keV
Limited R&D on spectroscopic detectors
Only effort is
Silicon array detector
of Peter Siddons (BNL) and Chris Ryan (Australia
)
Using silicon arrays to achieve large collection solid angles for fluorescence experiments.
Leverages local facilities and existing projects.
ANL’s Center for
Nanoscale
Materials for device fabrication
Many groups with thin film deposition experience
Superconducting Transition Edge Sensors for
UChicago’s
SPTpol
ANL’s Material Science DivisionSlide5
Detector Requirements
Solid angle coverage
Pixel area and sample-detector distance
Count-rate
Need lots of pixels (i.e. multiplexing) if you going to deal with
superconducting detectors, which tend to be slow
Peak-to-background ratio
Maximize
charge
collection Minimum of 1000:1 Limits overall sensitivity Energy resolutionDepends on applications (1-50eV at 6 keV)Slide6
Applications for superconducting x-ray detectors?
X-ray Inelastic Scattering
Access wide range of excitations.
Superconducting detectors allows broadband and efficient measurement compared to crystal analyzers (i.e., wavelength dispersive spectrographs).
X-ray
Fluorescence
F
luorescence line overlap in complex biological samples
Fluorescence Tomography
Needs pixelated energy-dispersive detectorsWhite-Beam Diffraction (ED-XRD)Versus angle-dispersive diffraction Using monochromatic incoming beam and area detector
Complex sample environments
(e.g., high-pressure cells)Slide7
Competing with x-ray spectrographs (like astronomers…)Slide8
Quasiparticle detectors
Use small Superconducting Energy Gap
Break Cooper Pairs
Use Cooper Pairs as detection mechanism (like electron hole pairs in a semiconductor)
Cooper Pair
Semiconductor Energy Gap: ≈ 1
eV
Superconductor
Energy Gap:
≈ 1
meV
How to measure
change in
quasiparticle
number? Slide9
Microwave Kinetic Inductance Detectors
Quasiparticle
generated by x-ray causes an inductance increase (i.e., “kinetic inductance”)
Measure inductance change in a LC resonating circuit
Multiplexing:
Lithographically vary geometric inductance/resonant frequency…
D
L
s
D
R
s
Observables….
1024 pixels demonstrated in 2011
People are contemplating 10k pixels now
Limited by room temperature electronicsSlide10
What has been shown already for x-rays?
62eV
Mazin et al 2006
Material
Atomic Number Z
Transition Temperature
Theoretical Energy Resolution at 6
keV
Attenuation Length at 6
keV
Lead
82
7.2 Kelvin
3.53 eV
1.86
m
m
Tantalum
73
4.5 Kelvin
2.79 eV
1.79
m
m
Tin
50
3.7 Kelvin
2.53
eV
2.60
m
m
Indium
49
3.4 Kelvin
2.43
eV
2.77
m
m
Rhenium
75
1.7 Kelvin
1.72
eV
1.31
m
m
Molybdenum
42
.92 Kelvin
1.26
eV
2.94
m
m
Osmium
76
.66 Kelvin
1.07
eV
1.18
m
mSlide11
MKIDs @ Argonne for synchrotrons
The goal is moderate energy resolution (< 20eV) with good count rate capabilities (> 200kcps) (i.e., ~200-500 pixels)
Leverages Argonne’s micro/
nano
-fab facilities (CNM) and superconductivity expertise (MSD); Astronomical TES bolometer program(MSD &
UChicago
).
Three Main Aspects:
Device Fabrication
Fabrication is completely in-house
Relatively “simple”… patterning of metal (deposition, photolithography, etching)
Film quality is very important!
Initially aim a simple device, then progress to more complex designs (e.g., membrane-suspended)
Dedicated deposition system on order.Cryogenics and Device CharacterizationWe are mostly limited by how fast we can test devices.
Readout electronics
Initially the analog readout for characterization.
Digital FPGA-based array readout
in the near future.Slide12
Anatomy of an MKID – Our work (one design)
1
pixel
15 pixels
2 mm
First x-ray pulses at APS in January 2012!
Fe-55 and Cd-109
1 micron WSi
2
(XSD)
Inductor/Absorber
CapacitorSlide13
Detector Testing
Be window
(x-ray transparent)
Cryostat
Microwave Electronics
Cryostat
Cryogen Free ADR
T = 100
mK
for 2 days
3-4 hour recycle time
Microwave Electronics
Vector Network Analyzer
IQ mixing
Control & Data Analysis SoftwareSlide14
Inside the cryostat…
LNA
Stainless Steel
coax
NbTi
(superconducting)
coax
Attenuator
Sample box
@ 60
mK
m
-metal shield
0.5K stage
60 K stage
4 K stageSlide15
VNA
Measure the transmission through the device (S21)
Wide range of frequency: 0.3 - 20 GHz
Easily identify resonators:
R
esonance frequencies, Quality Factors (
Q
r
, Q
c
, Q
i
)
Cannot measure pulses nor noiseSlide16
Signal
G
enerators
IQ Mixer
Variable
attenuator
IQ mixing (Pulses and Noise)
Homodyne mixing
Demodulate to 0 Hz (DC) and monitor for changes in Re(S21) and
Im
(S21) (i.e.,
I & Q)
Amplitude =
sqrt
(I
2
+ Q
2
)
Phase=
arctan
(Q/I
)
IQ MixerSlide17
IQ Fitting and parameter extraction
Frequency (GHz)
Real (S21)
Imag
(S21)Slide18
Pulses
Pulse fitting
Decay Time ->
Quasiparticle
life time
Phase pulse height -> Energy of the incident photon
The signal in phase is larger than amplitude
Phase
I &Q
Amplitude
Time (
m
s
)Slide19
Tungsten Silicide MKIDs
We have been searching for dense materials for x-rays.
WSi
x
is a material with low
T
c
, high kinetic inductance fraction and good quality factors.
Similar to Titanium NitrideMaybe also be interesting for optical MKIDs.arXiv:1203.5064v1Slide20
Optical MKIDs (Mazin et al) --
Spectrophotometry
Mazin
et al.
, Optics Express 2012
R=E/
Δ
E=16 at 254
nm
Limited by LNA/power handling/QSlide21
A readout for large arrays of Microwave Kinetic Inductance Detectors (McHugh, Mazin
et al, arXiv:1203.5861v1)
How to readout 1024 MKIDs….Slide22
Another technology – Transition Edge Sensors
TES have the best energy resolution (not
Fano
limited)
However, TES require relatively complex cryogenic electronics which makes multiplexing difficult.
Limited to 200 pixels right now.
Also, larger pixel means slower response….Slide23
TES are evolving towards MKID/resonator readout
“Microwave SQUIDs”
Eventually would like a wideband, quantum-limited LNA for MKIDs (i.e., don’t need SQUID anymore)
Josephson parametric
amp (J.
Gao
et al)
arXiv:1008.0046v1
Kinetic Inductance
parametric amp (P. Day et al)arXiv:1201.2392v1
Superconducting microwave resonant circuits for the detection of photons from microwaves through gamma
rays
, Irwin, K.D
., et al.
Microwave Symposium Digest (MTT), 2011 IEEE MTT-S International , vol., no., pp.1-4, 5-10 June 2011
B. Mates et al
ApL
2008 (NIST)Slide24
Conclusions
Superconducting detector development has started at the APS.
Testing infrastructure (
cryo
, electronics, analysis software) is complete.
Now focusing on device fabrication and iterating on designs
MKIDs are a path towards high count rates and higher solid angle coverage.
Has the potential to provide a very unique capability (detector/instrument) for the APS.
MKIDs are a relatively young technology and there is room for R&D.
Thank you!Slide25
The EndSlide26
Extra SlidesSlide27
Comparison of Silicon Drift Diode, MAIA, TES and MKID
Area
Sensor-Window Distance
Energy Resolution @ 6keV
Max Count rate (per pixel)
P/B ratio
Energy Range
State of maturity
Silicon
Drift Diode
40mm
2
5mm
250eV
250kHz
5000:1 (?)
< 20keV
Commercially
Available
175eV
100kHz
MAIA
1mm
2
x 384
5mm
250eV20kHz per pixel5000:1 (?)< 20keV
Semi-commercial??NIST TES0.14 mm2 per pixel (256+ pixels)
1 cm
3eV @ 6keV
18eV @ 40keV
70eV @ 100keV
10-100 Hz/pixel
(256+ pixels)
??
300eV
– 100keV
Semi-commercial
MKIDs
0.14 mm
2
per pixel
1 cm
60eV (2006)
2eV is theoretical
limit
4000Hz/pixel
(256+ pixels)
??
300eV-100keV
R&D
+Slide28
Silicon Drift Diodes
Commercially available
E.g., Vortex 4-pixel SDD (4 x 40mm
2
)
Performance of single pixel SDDs (measured)
Mn
K
a FWHM ~ 250 eV (Max Count rate ~ 250 kHz per pixel)Mn Ka FWHM ~ 175
eV
(Max Count rate ~ 100 kHz per pixel)
SDD have reached their limits of energy resolution. Only improvement to be made is more pixels to increase total count rate throughput and solid angle collection.
Thus, the MAIA detector (BNL/Australia)!Slide29Slide30Slide31
Speed and resolution are inverse proportional.
Controllable
thermal conductance to bath (“G”)
SiN
membrane geometrySlide32
Phonon-Coupled MKIDs -- Single Photon Emission CT -- SPECT
SPECT Imaging (for animals)
Patrick La
Riviere
thinks phonon-coupled MKIDs might be good for this.
Also interesting for high-energy x-ray applications
Sunil
Golwala
et al Slide33
Current Status of MKID Research at APS II
Exploring various detector configurations
Quasiparticle Trapping
Tradition Configuration
60eV resolution
and 4
000cps per pixel proven in 2006
30eV
should be “easily”
possible More is understood about noise sources
2eV is theoretical limit at 6keV
Lump-element design
(WSix
) No charge diffusion High Peak-to-BackgroundPhonon-Coupled Large pixels and thick silicon
(or other dielectric material for
absorption
Silicon Absorber (1mm)
SiN
membrane
Resonator
Inductor/Absorber
CapacitorSlide34
Pump-probe XAS w/o mono
3eV at 7keV
80Hz with 40 pixels (256 possible today)
80Hz * 256 = 20kHz (TDM)
80Hz * 1000 = 80kHz (CDM) This starts to be very interesting for RIXS/XES/XANESSlide35
MKIDs – Readout Scheme
Multiplexing is main advantage of MKIDs over STJs and TES
Should be able to readout
~4000
pixelsSlide36
MKID Array readout
Room temperature electronics
Transfer complexity from cryogenic electronics (TES & SQUIDs) to room temperature.
Scalable system
Array size limitation is room temperature electronics (512 is practical today)
Riding on Moore’s Law for room temperature microwave integrated circuits developed for the wireless communications industry.
X
X
IQ Mixer
Low Pass Filters
16-bit D/A
16-bit D/A
14-bit A/D
14-bit A/D
FPGA-based IQ demodulation
(Xilinx SX95T RF Engines Channel Core)
GHz SynthsizerSlide37
Quasiparticle Generation – A Cascade Process
Incoming Photon
absorbed, ionizing atom and releasing inner-shell electron
First Stage:
Rapid energy down-conversion (electron-electron interactions secondary ionization and cascade
plasmon
emission)
e.g. 10keV photoelectron down-conversion to a thermal population of electrons and holes at a characteristics energy ~ 1eV takes 0.1
ps
(
Kozorezov
et al )
1
st stage ends when electron-phonon inelastic scattering rate dominates electron-electron interactions. Second Stage: ~ 1eV down-convert to large number of Debye energy phononsEnergy of phonon distribution exceeds energy of electron distributionDebye energy of superconductors is larger than SC gap energy
Phonons with energy > 2
D
will generate quasiparticles.
Finally, we have a mixed distribution of quasiparticle and phonons:
N
qp ~ Ephoton/DBut scaled down because large percentage of photon energy stays in the phonon systemFor Tantalaum, 60% energy resides in qp system. (Kurakado) (Efficiency h
)
Nqp
= hhn
/D
At 6keV and Ta absorber, D = 0.67 meV 5 Million qp
!!Slide38
Phonon Trapping
Dominant phonon loss mechanism for Ta and Al is through the substrate.
Copper Pair breaking phonons Mean Free path ~ 50nm
Effective quasiparticle lifetime is lengthened by phonon trapping.
Acoustic match between superconducting film and substrate.
Thus thicker films or membrane-suspended absorbers. Slide39
Two major classes of Superconducting Detectors
Thermal Detectors (i.e., Transition Edge Sensors)
Quasiparticle DetectorsSlide40
Transition-Edge Sensors (TES)
1/e response time = 100 μs - 1 ms
Kent Irwin, NISTSlide41
Transition-Edge Sensors (TES)
TES have the best energy resolution (not
Fano
limited)
However, TES require relatively complex cryogenic electronics which makes multiplexing difficult.
Limited to 200 pixels right now.
Also, larger pixel means slower response….
Irwin &
Ullom
, NISTSlide42
TES array at NSLS (NIST U7A)
Commissioned at NSLS this year…. Waiting for results…Slide43
Gamma-Ray TES Array
(NIST + Los Alamos)
Resolution
25
eV
at 100
keV
for 1
mm
2
20eV at 40keV for 1 mm
2
70
eV
at 100
keV
for 2 mm
2
Count Rate
10Hz/pixel (now)
25Hz/pixel (soon)
Pixels
256 (now)
1024 (soon)