X-RAY SPECTROMETRY
X-Ray Spectrom. 28, 396–398 (1999)
High-Resolution X-Ray Spectrometry Using
Iridium–Gold Phase Transition Thermometers†
J. Höhne,* M. Altmann, G. Angloher, P. Hettl, J. Jochum, T. Nüssle, S. Pfnür, J. Schnagl,
M. L. Sarsa, S. Wänninger and F. v. Feilitzsch
Technische Universität München, Physik Department E15, 85748 Garching, Germany
Cryogenic detectors based on iridium–gold phase transition thermometers are being developed for various
applications in astroparticle physics, aiming at high energy resolution, low energy threshold and moderate
count rates. A microcalorimeter with an energy resolution of 15.5 eV (FWHM) for the 55 Mn Ka1 x-ray line
for tentative use in an analytical x-ray system (e.g. SEM, TRFA) is presented. Copyright  1999 John Wiley
& Sons, Ltd.
INTRODUCTION
Cryogenic detectors with excellent energy resolution and
low energy threshold far beyond the level of semicon-
ducting detectors open up a variety of new applications
in physics, including the search for dark matter in the
universe,2 neutrino physics3 and IR, UV and x-ray astro-
physics.4 Interdisciplinary fields where cryogenic detectors
have already shown promising results are the detection of
biomolecules5 and x-ray spectrometry at synchrotron beam
lines6 and in scanning electron microscopes (SEMs).7 For
both astrophysical and analytical use we started the devel-
opment of high-resolution microcalorimeters based on irid-
ium–gold phase transition thermometers for use in a mobile
detector system.
MICROCALORIMETER
Our microcalorimeters consist of an absorber, a ther-
mometer and a coupling to a heat sink. An incoming par-
ticle or x-ray photon deposits an energy E in the absorber.
The corresponding temperature rise T is then measured
using an iridium–gold superconducting phase transition
thermometer (SPT). After the particle interaction, the sys-
tem relaxes to its operating temperature via the coupling
to the heat sink. In a simple model the temperature rise
T of the thermometer can be described as T D E/C,
where C is the heat capacity of the thermometer–absorber
system. Since heat capacities become smaller on going
* Correspondence to: J. Höhne, CSP Crgogenic Spectrometers Gmb4,
Bahnhofstr. 18A, D-857371 smaning, Germany.
E-mail address: jens.hoehne@csp-munich.com
† Presented at the European Conference on EDXRS, Bologna, Italy,
7–12 June 1998.
Contract/grant sponsor: Deutsche Forschungsgemeimschaft.
Contract/grant sponsor: Sonderforschungsbereich Astro/Teilchen-
physik; Contract/grant number: SFB 375.
Contract/grant sponsor: European Community; Contract/grant num-
ber: ERB; Contract/grant number: CHRX CT93-0341 ERB FMRX
CT98-0167.
Contract/grant sponsor: Beschleuniger-Laboratorium Garching.
CCC 0049–8246/99/050396–03 $17.50
Copyright  1999 John Wiley & Sons, Ltd.
to lower temperatures, this detector concept is most sen-
sitive at operating temperatures below 100 mK. These
temperatures are reached with standard 3 He/4 He dilution
or adiabatic demagnetization refrigerators.
SUPERCONDUCTING PHASE TRANSITION
THERMOMETERS (SPTs)
Our phase transition thermometers are thin superconduct-
ing films (ca 150 nm) with areas ranging from µm2 to
cm2 depending on the application. The film is stabilized
in the region between the normal conducting and the
superconducting state, where the resistance shows a strong
dependence on temperature. Fig. 1 shows an example of a
transition curve of an iridium–gold proximity effect phase
transition thermometer at Tc D 39.5 mK. The width of the
transition region is about 2 mK. A typical temperature rise
due to absorption of a 6 keV x-ray is in the µK range.
The readout arrangement is shown in the inset of
Fig. 1. A constant current I0 is passed through both the
thermometer RT and a shunt resistor RS placed in parallel.
A resistance change of the thermometer due to the heating
from energy deposition in the absorber causes a change
in current in the SQUID input coil. Assuming linearity
of the transition curve and a calorimetric operating mode
of the detector, the signal height is directly proportional
to the deposited energy.
Our SPTs, consisting of a bilayer of iridium and gold,
make use of the proximity effect. The critical temperature
of the iridium film (bulk Tc D 112 mK) is reduced
by the gold film depending on the film thicknesses of
both films.8 This bilayer technique permits us to fabricate
thermometers with operating temperatures between 20 and
112 mK. The reproducibility of the transition temperature
is about š5 mK above 40 mK. Fabrication is done under
UHV conditions in a dedicated system.
EXPERIMENTAL
Figure 2 shows the experimental setup for a microcalori-
meter consisting of a gold absorber (250 ð 250 ð 1 µm3 )
Received 13 October 1998
Accepted 4 June 1999
 HIGH-RESOLUTION XRS USING Ir–Au PHASE TRANSITION THERMOMETERS
Figure 1. Transition curve for an iridium–gold phase transition
thermometer. Inset: readout circuit for our SPTs.
Figure 2. Experimental setup for our microcalorimeter.
and an iridium–gold thermometer (1 mm ð 1 mm ð
150 nm). The sapphire substrate underneath served as a
heat sink, being coupled to the thermometer via its sur-
face. Its heat capacity can be neglected for the evaluation
of T since its coupling to the thermometer–absorber
system is slow compared with the heat distribution in
this system. Electrical contact to the thermometer was
made via evaporated aluminum leads from where contact
was made to the SQUID circuit using 25 µm diame-
ter wedge bonded aluminum wires. Detector fabrication
was done in three major steps: evaporation of the irid-
ium–gold thermometer, sputtering of the gold absorber
and evaporation of the aluminum leads. Device patterning
processes were done via photolithography and ion beam
etching.
The readout circuit consisted of a 6 m shunt resis-
tor and a single-stage HYPRES SQUID system. The
microcalorimeter and the SQUID were mounted in a
3 Figure 3. Spectrum of 55 Mn measured with our microcalorime-
He/4 He dilution refrigerator. The microcalorimeter was
operated at 40 mK with a temperature stability of better
than 100 µK.
Copyright  1999 John Wiley & Sons, Ltd.
397
The gold absorber was irradiated through a 100 µm
diameter copper collimator with a rate of about 50 cps.
The SQUID signal was recorded using a 12-bit transient
recorder and a 50 kHz anti-aliasing filter.
RESULTS
Since the absorption efficiency of our gold absorber is
about 50% for 6 keV x-rays, two different classes of
events were expected: absorptions in the gold absorber
or in the thermometer and absorptions in the underly-
ing sapphire substrate. For further processing the signals
were fitted using one exponential rise time and one expo-
nential decay time. The two classes of events could be
distinguished by analyzing the decay time of the signals:
absorptions in the gold absorber are described best with
a rise time of 30 µs and a decay time of 630 µs, whereas
absorptions directly underneath the thermometer have a
much longer decay time and can be seen at a much lower
pulse height in the spectrum. Figure 3 shows the spectrum
of a 55 Fe source measured with our microcalorimeter. The
55
Mn K˛ line complex was fitted using two Gaussians of
equal width, fixed intensity ratio and fixed relative energy
difference. The 55 Mn K˛1 line could be resolved with an
energy resolution of 15.5 eV (FWHM) (Fig. 3, inset). Two
features in the structures of the K˛ and Kˇ peaks should
be noticed: there is a small peak with about one tenth of
the intensity immediately below the K˛ and Kˇ peaks,
which might be due to absorption of x-rays in the Ir–Au
thermometer since the ratio of thicknesses of the gold
absorber and the Ir–Au thermometer is in this range. Just
below these lines there is another small peak of unknown
origin.
Our energy resolution is still worse than the theoretical
limit owing to thermodynamic fluctuations for this kind of
detector. One major contribution was the SQUID noise,
being 20 pA Hz 1/2 . The energy threshold of this detector
was measured to be below 30 eV.
ter. To the left of the main K˛ peak a satellite peak can be seen
which might be due to absorptions in the thermometer. Further
left an even smaller satellite of unknown origin can be seen.
X-Ray Spectrom. 28, 396–398 (1999)
398 J. HÖHNE ET AL.
CONCLUSION
With a cryogenic microcalorimeter with an active area of
250 ð 2502 µm and an absorption efficiency of 50% for
6 keV x-rays, an excellent energy resolution of 15.5 eV
(FWHM) for the 55 Mn K˛1 x-ray line could be shown.
The use of a SQUID system with much lower noise could
further increase the sensitivity of our detector. Techniques
for shortening signal lengths are being developed having
as a goal count rates in the kHz region. A mobile detec-
tor system is being constructed and will be available for
use in various experiments in the field of x-ray spectrome-
try. Especially for x-ray spectrometry for material analysis
REFERENCES
1. Proceedings of 7th Workshop on Low Temperature Detectors
(LTD7), MPI for Physics, Munich (1997).
2. S. W. Nam et al., in Proceedings of the 7th Workshop on
Low Temperature Detectors (LTD7), p. 217. MPI for Physics,
Munich (1997); D. L’Hote et al., in Proceedings of the 7th
Workshop on Low Temperature Detectors (LTD7), p. 237. MPI
for Physics, Munich (1997); M. Sisti, in Proceedings of the 7th
Workshop on Low Temperature Detectors (LTD7), p. 232. MPI
for Physics, Munich (1997).
3. M. Altmann et al., in Proceedings of 4th Solar Neutrino
Conference, Heidelberg, (1997); J. Höhne et al., in Proceedings
of the 7th Workshop on Low Temperature Detectors (LTD7),
p. 184. MPI for Physics, Munich (1997).
4. P. Hettl et al., in Proceedings of the 7th Workshop on Low
Temperature Detectors (LTD7), p. 20. MPI for Physics, Munich
Copyright  1999 John Wiley & Sons, Ltd.
applications, where there is a need for high-energy reso-
lution in the energy region below 2 keV to identify low-Z
elements, cryogenic detectors will bring major improve-
ments in sensitivity.
Acknowledgements
We gratefully acknowledge the support of the Deutsche Forschungsge-
meinschaft (DFG), the Sonderforschungsbereich Astro/Teilchenphysik
(SFB 375), the European Community under Grants ERB CHRX CT93-
0341 and ERB FMRX CT98-0167 and the Beschleuniger-Laboratorium
Garching. We are indebted to F. Pröbst and W. Seidel of the Max-
Planck-Insitut für Physik in Munich for many helpful discussions and
practical support.
(1997); S. Deiker et al., in Proceedings of the 7th Workshop on
Low Temperature Detectors (LTD7), p. 108. MPI for Physics,
Munich (1997); P. Verhoeve et al., Appl. Phys. Lett. 72, 3359
(1998).
5. G. Hilton et al., Nature (London) 391, 672 (1998).
6. M. Frank et al., Rev. Sci. Instrum. 69, (1998); S. Labov et al.,
in Proceedings of the 7th Workshop on Low Temperature
Detectors (LTD7), p. 82. MPI for Physics, Munich (1997).
7. D. Wollman et al., J. Micros. 188, 196 (1997); E. Silver et al.,
X-Ray Spectrory 26, 265 (1997).
8. J. Höhne et al., Nucl. Instrum. Methods 370, 160 (1996);
U. Nagel et al., J. Appl. Phys. 78, 4262 (1994).
X-Ray Spectrom. 28, 396–398 (1999)