The low temperature detector project



alt

Abstract

The development of integrated silicon microcalorimeters with high resolving power for low-energetic single quanta of radiation is being carried out. The optimum thermistor implantation dose has been found, and measurements have shown, that our detectors are very sensitive. The experimental set-up, however, suffers from mechanical vibrations, which are picked up by the cryostat and amplified in there. Measurements of the frequency spectrum of the noise gave evidence for this interpretation. All necessary steps to improve the damping are being done. Furthermore, the development of a cryogenic preamplifier has been started, meant to reduce electronic noise. The influence of wafer purity which was observed earlier, could be determined and verified by measuring the implantation profiles by the secondary ion mass spectrometry method. Beside that, studies of hopping conduction phenomena, the processes responsible for the observed temperature dependence of resistivity, have been performed. It could be concluded that the model of Coulomb gap variable range hopping is supported by our experimental data over a wide temperature range. Recently, the study of magnetoresistance effects was started.

1. Introduction

The measurement of nuclear radiation by using calorimetric methods is an idea which already became reality in the early decades of this century. It is kept alive by the challenging quest for the optimum energy resolution, in principle possible to achieve just with this type of detectors. This results from the fact that the average creation energy of phonons is hundred thousand times smaller than of electron-hole pairs in usual semiconductor detectors, leading to less statistical fluctuations and therefore to better energy resolution. Another big advantage of low-temperature microcalorimeters is given by their feature that there is no dead layer, so even very low-energetic particles can be measured without energy loss.
The Subatomic Physics group at Chalmers University of Technology in Göteborg, Sweden, started a program for the development of microcalorimeters in 1985 in collaboration with the Laboratoire de Physique Stellaire et Planétaire (LPSP), Verrières le Buisson, France, and the Institute for Physics and Astronomy at Aarhus University, Denmark. Since the working principle of these microcalorimeters is explained elsewhere, we may restrict us here to give only a brief and schematic description. When a quantum of radiation hits the absorber, its kinetic energy is deposited there partly as ionization energy, e.g. creating electron-hole pairs, or as heat. It is of major importance to keep the detector at very low temperatures (less than 1 K) in order to be able to detect the increase of temperature caused by the incident particle's energy (typically in the order of some keV). However, these low working temperatures require a special type of - high sensitive - thermometers. One possibility is to use superconducting transition edge thermometers or Josephson junctions, where the resistance change due to the transition from the superconducting to the normalconducting state is used to determine the amount of incident energy. Another quite common type of thermometers - the thermistor - is based on the strong temperature dependence of the resistivity in heavily doped semiconductors. Here the incident energy leads to a raise in temperature, which is reflected by a decreasing resistivity (see fig. 1). Biasing the thermistor with a constant current, the change in resistance leads to a voltage pulse. Since the sensitivity of these thermistors is dependent on the derivative dR/dT, the optimum thermistor doping has to be found, that gives a reasonably high resistance in the temperature regime of interest.
Fig. 1. Schematic view of the working principle of thermal detectors.

2. Experimental status

In contrast to conventional composite bolometers, where radiation absorber and thermistor are separate components which are usually glued together, our philosophy is to integrate both parts.This could be achieved by implanting the thermistor region on the surface of the absorber body. By using silicon wafers we profit from well known micromachining methods for the fabrication. The entire fabrication process consists of more than 70 different steps, a simplified scheme is shown in fig.2.


 




Fig. 2. Fabrication scheme for microcalorimeters by silicon micromachining.

In order to determine the optimum thermistor doping for our thermal detectors (refer to fig.1), the temperature dependence of the resistance was measured in a temperature range between 40 mK and 4.2 K for various phosphorus doses between 7.5 and 11.25 x 1013 cm-2. In this region of low temperatures we use a commercial 3He/4He Oxford TLE 200 dilution refrigerator. It is situated in a shielded room and is equipped with a top loading facility, which makes it possible to change the sample while the cryostat remains at low temperature.
Since the signal from a thermal detector is of completely different nature than that from charge sensing detector systems, a special type of data acquisition was necessary in order to sample the complete pulse shape with a fast ADC. Based on these facts it was decided, at an early stage, to develop a data acquisition system especially for thermal detectors. During the last year the development of implanted thermistors was complemented by investigations of Josephson junction based thermometers. Furthermore, we succeeded to find the optimum implantation dose for our thermistors. According to calculated predictions, experimental evidence was found for a value of 1.025 x 1014 phosphorus ions / cm2. Very recently new experiments were performed to determine the actual energy resolution by measuring spectra of conversion electrons and X-rays from the decay of 109Cd and X-rays from the decay of 55Fe. Careful studies of the influence of bias voltage, signal filtering procedures and working temperature (around some tens of mK) were done. Preliminary results are shown in fig. 3.
 



 




Fig. 3. Spectrum from an uncollimated 109Cd source recorded using a thermal detector with a thermistor doping of 1.025 x 1014 phosphorus ions / cm2. Variation of bias voltage, filtering and trigger level gives optimum energy resolution with different parameters in different regions. For the low-energy region a) a bias voltage of 1 V was applied, while 4 V was used in part c). The energy resolution was about 10 keV for the conversion electrons at 63 and 85 keV in c), 2 keV for the Ag X-rays at 22 and 25 keV in b), and around 400 eV in a) for Cu X-rays at 8 and 8.9 keV resulting from excitations in the detector holder.

The not yet satisfying energy resolution can be explained by: i) pronounced microphonic effects give rise to noise, ii) the conventional preamplifier is working at room temperature and is connected with the detector by a 4 m long coaxial cable, and iii) the working temperature of the detector is too high due to the biasing current. Thorough measurements of the frequency spectrum with different bias voltages and filterings indicated that these effects emanate from external sources. Since we now - for the first time - succeeded to localize the origin of these problems, all necessary efforts to improve shielding and damping will be taken.

3. Studies of intrinsic properties in heavily doped silicon

The measurements of the temperature dependence of the thermistor resistance for different implantation doses enabled us to study the underlying physical processes. This allows us to explain the experimental observations. Figure 4 shows the measured resistance values as a function of the inverse square root of temperature for different samples.


 




Fig. 4. Measured resistance values as a function of the inverse square root of temperature according to the model of Coulomg gap variable range hopping model for a selection of nine silicon samples with different phosphorus implantation doses. The full drawn lines represent fits for T>120 mK.

This way of data presentation is suggested by the model of Coulomb gap variable range hopping (CGH). The good agreement for the entire temperature region above 120 mK confirms the predictions of the CGH model, the deviations at very low temperatures could recently be explained by a screening effect on the Coulomb potential.
As reported recently, we observed a strong influence of the purity of the used silicon wafers on the thermistor behavior. The measured resistance values as well as the temperature dependence for wafers with the same doping but different purities differed quite a lot from each other. Obviously, the doping dose is not sufficient to characterize the samples. Therefore the doping profiles were measured by the method of secondary ion mass spectrometry (SIMS) and some examples of the experimental results are shown in fig. 5.
 



 




Fig. 5. Measured phosphorus implantation profiles in silicon. Both lines correspond to an implantation dose of 9 x 1013 cm-2 for wafers with different resistivities, r1 = 5 kOhm cm and r2 = 2 Ohm cm, respectively.

The two curves represent phosphorus distributions after an implantation dose of 9 x 1013 cm-2, for wafers of different intrinsic purity. The full drawn line corresponds to wafer material of a resistivity r1 = 5 kOhm cm, the dashed line to r2 = 2 Ohm cm. The corresponding impurity concentrations are 1012 cm -3 and 2 x 1015 cm-3, respectively. The dark hatched area indicates a SiO2 - layer of 400 Å thickness, which was grown before the ion implantation was performed, in order to optimize the maximum of the profiles towards the silicon surface. The discrepancy between the two distributions concerning the position of the maximum as well as their shapes is remarkable. The present status of our investigations leads to the conclusion that the diffusion process, caused by annealing right after the ion implantation, seems to be blocked in the more impure material. Nevertheless, our experimental data allow to estimate the critical doping concentration nc forthe metal-insulator transition (MIT). The results of the fits for tc and nc for the two sets of wafers with respect to their purities are shown in fig. 6.
 



 




Fig. 6. Temperature parameter To as a function of the maximum phosphorus concentration in uncompensated silicon. The data are plotted separately for both wafertypes, corresponding to their intrinsic resistivities, r1 = 5 kOhm cm and r2 = 2 Ohm cm, respectively.

It can be seen that the obtained values for nc differ quite a lot, however they match very well with the values published so far for phosphorus doped silicon, lying between 3.4 and 6.8 x 1018 cm-3.

4. Summary and outlook

This project can be considered to be a connecting link between nuclear physics and solid state physics. It is dedicated to develop an unique type of detectors for nuclear physics experiments, allows however - and even requires it - thorough studies of the low temperature conduction processes being responsible for the working principle of the detectors. The development of the integrated silicon microcalorimeters has reached an important stage. For the first time the energy resolution is not any longer restricted by the detector itself, but by noise coming in from the surroundings. In order to improve energy resolution the tackling of the following steps has been started and will be dominating our future work:
  1. Optimization of thermistor dimensions in order to reduce the heating due to the bias voltage, studies of the influence of thickness, width and number of bridges on the shape of the thermal pulses.

  2.  
  3. Improvement of the mechanical damping of the experimental set-up in order to eliminate the influence of vibrations. Exchange SiO2 layers by other materials like e.g. Si3N4, to avoid piezoelectric effects possibly interfering with the electronic signal.

  4.  
  5. Tests of new detectors where the silicon absorber is replaced by an evaporated tin layer in order to avoid trapping effects in silicon to increase the intrinsic resolution

  6.  
  7. Development of a cryogenic preamplifier based on a GaAs-FET, designed for temperatures less than 1 K (in the dilution refrigerator). This work is topic of a diploma work. The amplifier is about to be tested and will be mounted inside the cryostat in the near future.

  8.  
  9. Improvement of the data acquisition procedure for the thermal pulses by using a digitizing oscilloscope and evaluating with the help of a LabVIEW® environment.

  10.  
  11. New experiments are planned at IKS Leuven, Begium. Our detectors will be tested in a similar cryostat, whose damping and shielding is much better than ours, in order to avoid the influence of vibrations. This will make it possible to determine the real performance of our microcalorimeters with respect to energy resolution. These tests are going to be performed in March 1996
Beside the more application oriented work dealing with the development and improvement of our low temperature detectors, the investigation of the temperature dependence of the electrical resistance allowed the study of hopping conduction processes. Clear evidence for Coulomb gap variable range hopping (CGH) was observed between 120 mK and 4.2 K. However, at temperatures below, deviations from this model appear, which - after ruling out all possible experimental error sources - give rise to the interpretation that screening of the Coulomb potential occurs due to the increasing hopping distances. Furthermore, the critical concentration nc for the metal-insulator transition (MIT) was determined, while the influence of the wafer purity was an important parameter. Differences with respect to nc led to the assumption that surface states might deliver additional sites for the variable range hopping. Detailed studies of the surface effects are planned. Currently, the investigation of the influence of magnetic fields on the characteristic R-T curves (magnetoresistance) is going on, also dedicated to be applied to the use of our thermal detectors. With these studies it will be possible to manipulate the sensitivity of the detectors by affecting the R-T dependence of the thermistors. This work is topic of a diploma work, too.
 

Related publications at CTH

  1. A composite bolometer as a charged-particle detector.

  2. N. Coron, G. Dambier, G. J. Focker, P. G. Hansen, G. Jegoudez, B. Jonson, J. Leblanc, J. P. Moalic, H. L. Ravn, H. H. Stroke and O. Testard,
    Nature 314 (1985) 75.
     
  3. Bolometers as particle detectors.

  4. H. H. Stroke, G. Artzner, N. Coron, G. Dambier, P. G. Hansen, G. Jegoudez, B. Jonson, J. Leblanc, J. P. Lepeltier, G. Nyman, H. L. Ravn and O. Testard,
    IEEE Trans. Nucl. Sci. NS-33 (1986) 759
     
  5. Micromachining of silicon for thermal and position-sensitive nuclear detector application.

  6. Y. Bäcklund, N. J. Coron, P. Delsing, B. Jonson, M. Lindroos, G. Nyman, H. Ravn, K. Riisager and H.H. Stroke,
    Nucl. Instr. and Meth. A 279 (1989) 555.
     
  7. A new temperature sensor in low-temperature bolometers for high resolution spectroscopy of nuclear radiation.

  8. P. Delsing, C. D. Chen, T. Claeson, P. Davidsson, B. Jonson, M. Lindroos, S. Norrman, G. Nyman and S. Qutaishat,
    Physica B 194-196 (1994) 27.
     
  9. Design of microelectronic thermal detectors for high resolution radiation spectroscopy.

  10. S. Qutaishat, P. Davidsson, P. Delsing, B. Jonson, R. Kroc, M. Lindroos, S. Norrman and G. Nyman,
    Nucl. Instr. and Meth. A342 (1994) 504.
     
  11. Silicon thermal detectors for single quanta of radiation: fabrication, statistical fluctuations of phonons, physical properties and operation.

  12. P. Davidsson, P. Delsing, B. Jonson, R. Kroc, M. Lindroos, S. Norrman, G. Nyman, A. Oberstedt and S. Qutaishat,
    Nucl. Instr. and Meth. A350 (1994) 250.
     
  13. Silicon microcalorimeters for high resolution spectroscopy.

  14. A. Oberstedt, P. Davidsson, P. Delsing, B. Jonson, R. Kroc and G. Nyman,
    Nucl. Instr. and Meth. A370 (1996) 206.

Ph D thesis at CTH

  1. Design and development of crystalline thermal detectors for single quanta of radiation and high resolution spectroscopy.

  2. S. Qutaishat, 1994.

Diploma thesis at CTH

  1. Thermal detection of charged particles, a first approach.

  2. Hans Cronberg and Leif Raknes, 1986.
     
  3. Multi Channel Analyser for Thermal Radiation Detector.

  4. Richrd Kroc, 1991.
     
  5. A Cryogenic Preamplifier using a GaAs Field Effect Transistor Input Stage.

  6. Björn Starmark, 1996.
     
  7. Victoria Svenmyr, in preparation

  8.  

 
 

For questions or comments send a mail to Andreas Oberstedt.