Search

We describe our experience on design and fabrication, on high-resistivity silicon substrates, of microstrip detectors and integrated electronics, devoted to high-energy physics experiments and medical/industrial imaging applications. We report on the full program of our collaboration, with particular regards to the tuning of a new fabrication process, allowing for the production of good quality transistors, while keeping under control the basic detector parameters, such as leakage current. Experimental results on JFET and bipolar transistors are presented, and a microstrip detector with an integrated JFET in source–follower configuration is introduced.

Tetra-methyl ammonium hydroxide, or TMAH, is an anisotropic silicon etchant that is gaining more and more attention in the fabrication process of mechanical microstructures and device isolation, as an alternative to the more usual KOH and EDP etchants [1]: because of its high compatibility with conventional IC processes, due to the absence of metal ions in it. The possibility to passivate the aluminum metalisation in properly saturated TMAH solution has also been demonstrated by doping the solution with appropriate amounts of silicon or silicic acid [2, 3]. This increases the range of application of these etchants, simplifying both the post processing and the etch set-up configuration. In this paper we present some different technological solutions adopted for the fabrication of 3D structures using as anisotropic etchant doped TMAH solutions. The effects of the additives on etch uniformity and surface roughness are studied. Using the etching results under different doping conditions of the TMAH solutions, we obtained silicon bulk-micromachined structures with controlled surface roughness. Furthermore the aluminium passivation permits to etch devices with no protection of the metalisation layers as some applications require (i.e. bolometers).

We show that in high-resistivity silicon bulk generation lifetime and surface generation velocity can not be measured with acceptable accuracy using a single gated diode, unless the gate length is suitably tailored. The accuracy with which bulk generation lifetime can be evaluated is, in particular, limited by nonidealities, contributing extra components to the gated-diode current, which are not accounted for in the standard extraction procedure.

PIN diodes and other test structures have been fabricated on both n- and p-type, high-resistivity, Floating-Zone (FZ) silicon substrates. Different alternative extrinsic-gettering techniques have been adopted to the purpose of meeting the required specification of a detector leakage current density lower than 1 nA/cm2. Phosphorus-doped polysilicon gettering provided the best results on n-type Si with a leakage current density lower than 0.2 nA/cm2 at 100 μm depletion width. On the contrary, devices made on p-type substrates exhibited a leakage current density two orders of magnitude higher. A proper control of the oxide charge at the silicon-silicon dioxide interface was found to be crucial in obtaining a predictable behavior of PIN diode detectors. Some degradation of the reverse leakage current has been observed after device dicing and bonding.

In current particle physics experiments, silicon strip detectors are widely used as part of the inner tracking layers. A foreseeable large-scale application for such detectors consists of the luminosity upgrade of the Large Hadron Collider (LHC), the super-LHC or sLHC, where silicon detectors with extreme radiation hardness are required. The mission statement of the CERN RD50 Collaboration is the development of radiation-hard semiconductor devices for very high luminosity colliders. As a consequence, the aim of the RandD programme presented in this article is to develop silicon particle detectors able to operate at sLHC conditions. Research has progressed in different areas, such as defect characterisation, defect engineering and full detector systems. Recent results from these areas will be presented. This includes in particular an improved understanding of the macroscopic changes of the effective doping concentration based on identification of the individual microscopic defects, results from irradiation with a mix of different particle types as expected for the sLHC, and the observation of charge multiplication effects in heavily irradiated detectors at very high bias voltages. © 2011 Elsevier B.V.

Silicon bolometers obtained by P implantation on single crystal substrates and compensated with boron in the range 10–50% have been characterized from a technological point of view. Special emphasis was laid on the doping profile of the thermistor, the most critical point of the whole process for an optimal performance of the device as thermal detector. In particular, the effect of post implant anneal on the recovery of the radiation damage and on the modification of the as implanted doping profile is reported. An anomalous enhanced diffusion at 920°C was observed in uncompensated samples. The effect of compensation with B is to reduce the P diffusion, thereby preserving the original as implanted “box” profile almost in haltered. A reproducibility of ±5% of the resistance vs. temperature R(T) characteristic of different bolometers at temperatures below 4.2K has been achieved. This represents an important advancement in the realization of Si bolometers to be used as thermal detectors.

Resistivity measurements carried out in the temperature range 80 mK to 1 K on silicon bolometers-phosphorous doped by ion implantation - at concentrations near the metal-insulator transition, exhibit variable range hopping (VRH) conduction in the whole observed temperature range, with a T-1/2 dependence, in accordance with the Coulomb interaction model for low-temperature conductivity in disordered systems. Samples which apparently show a different behaviour, intermediate between VRH and metallic conduction, can be modelled by a metallic resistance in parallel with an active layer which follows the classic exp(T(O)/T)1/2 law. The observed behaviour can be explained in terms of residual radiation damage induced by the ion implantation process.

In this work we have investigated the feasibility of atmospheric pressure chemical vapor deposition of phosphorus and boron doped silicon oxides from a liquid source—namely TMP (tri-methyl-phosphite)—an alternative to phosphine as a P dopant source'. The most important results on process characterization, dopant incorporation, and film properties are presented. In addition to the easy operation of the liquid source, a remarkable improvement in step coverage has been found, in comparison with PSG deposition from conventional silane and phosphine chemistry.

We report on a new method to measure the electron antineutrino mass, consisting of the end-point study of 187 Re β-spectrum (Q-value=2.6 keV) collected calorimetrically (source=detector approach) with high energy resolution and statistics. The proposed detector is a micro-bolometer, single or in an array structure, containing a total rhenium mass of ∼ 1 mg, operated at a temperature of a few tens of mK and able to reach 10 eV energy resolution. As phonon sensors, we are developing silicon thermistors doped by implantation; some tests are performed also with neutron transmutation doped germanium thermistors. We have obtained up to now promising energy resolutions down to ∼ 11 eV with tin absorbers (as test devices) and ∼ 70 eV with rhenium absorbers.

The accuracy of the gated-diode method for extracting bulk generation lifetime and surface generation velocity in high resistivity silicon is shown to depend critically on the gate length of the adopted test device, as a result of nonidealities which are not accounted for by the measurement technique. Minimization of the surface generation velocity measurement error requires the gate length to be suitably reduced, while long gate devices are needed for accurate bulk generation lifetime extraction. Both parameters can be measured from a single test structure obtained by compenetrating a short gate device with a long gate one.

We report on Junction Field Effect Transistors and PIN diodes monolithically integrated on high-resistivity silicon by adopting a non-standard technology recently developed at IRST. In particular, a test structure, consisting of a small PIN diode DC-coupled to an integrated n-channel JFET in the double-gate configuration was fully characterised and spectroscopic measurements were carried out by adopting a novel double-feedback charge amplifier circuit. An ENC of about 60 electrons r.m.s. has been obtained at room temperature and at 10 μs shaping time; such a resolution is shown to be determined by the relatively high total capacitance present in this preliminary set-up, associated with the 1/ f series noise of the transistor.

We report on the design, fabrication, and characterization of a microheater module for chemoresistive, metal-oxide semiconductor gas sensors, consisting of a dielectric stacked membrane, micromachined from bulk silicon and with an embedded polysilicon resistor heater. Fabricated structures exhibit excellent heating efficiency, requiring only 30 mW to achieve a temperature of 500 C. Measured electrothermal characteristics are in good agreement with the outcomes of 3D numerical simulations.

We describe an activity aimed at developing npn silicon bipolar phototransistors to be used in optical position encoders. Extensive numerical proces and device simulations have been carried out, providing the guidelines for the definition of the phototransistor fabrication process. Results from the electrical and optical characterization of manifactured devices are shown. With respect to the designed fabrication process, only a little adjustment of the technological parameters has been to be necessary to achieve a final product suitable for the intended application.

We report on the development of a radiation-detector compatible JFET technology on high-resistivity silicon for monolithic integration of detectors and front-end electronics. A dedicated test-chip has been designed and fabricated for process and device characterization. Results from the electrical characterization of a first fabrication run show that good values of detector leakage current ( ≈1 nA/cm 2 ) can be obtained in spite of the relatively high thermal budget characterizing the process. As far as the JFET performance is concerned, a problem of insufficient device isolation at high substrate voltages has been evidenced. A second run is currently being carried on with the aim of optimizing the JFET structure.

PIN radiation detectors and other test-structures have been fabricated on a FZ, high resistivity (2 k/spl Omega//spl middot/cm), n-type silicon substrate by a process that features three different, alternative extrinsic-gettering techniques. Extremely-low leakage-current values have been measured regardless of the gettering technique adopted, this confirming the effectiveness of gettering procedures in limiting the detector leakage current. In particular, devices exploiting a phosphorus-doped polysilicon backside layer as gettering site have shown the best results in terms of leakage-current and generation-lifetime values. Results from the electrical and optical characterization of such devices are reported and discussed. X-ray detection testing is under way.

Silicate glasses, both undoped and lightly doped with phosphorus, prepared by a low pressure plasma enhanced chemical vapor deposition (PECVD) system using tetraethylorthosilicate or silane as silicon sources and atmospheric pressure chemical vapor deposition technique using silane, were evaluated. The three analyzed phosphosilicate films were nominally 4.0 wt % doped. All the samples were deposited in the temperature range of 350-420-degrees-C. Their main properties such as refractive index, density, wet etch rate, stress, step coverage, moisture resistance, and infrared absorption are reported. Infrared absorption measurements were performed on both as-deposited and annealed films (430-degrees-C for 15 min in N2) to investigate the presence of water and Si-OH groups, as well as P=O and Si-O bond interactions. The P-doped films were tested as final passivation layers on 1 Mbit erasable programmable read only memory devices mounted in ceramic packages. Similar electrical results have been obtained with three different phosphosilicate glass films. However, the most reliable results after assembly and packaging were obtained for films deposited by PECVD technique using tetraethylorthosilicate as the source of silicon in the films

A new picture for acceptors in silicon is proposed. Based on the assumption that the hydrogen‐like model holds true irrespective of the value of ionization energy, it implies an effective mass depending on the impurity. This leads to the prediction of new equilibrium and transport properties for moderately doped p‐type silicon.

Variable temperature Hall‐effect measurements on low‐concentration Si:In samples show the existence of supershallow levels with an activation energy for ionization of about 18 meV. For high concentration Si:In samples, the levels at 18 meV tend to disappear and transform into levels with energy of about 160 meV.

The strain entropy of substitutional Group III, IV and V impurities in silicon has been calculated from the experimental solid solubility taking into account the electronic, vibrational and polarization contributions. A correlation between differences of tetrahedral radii and strain entropy is found. Silicon is perhaps the most studied substance after water. In spite of this, there is, at present, no unified description of the solid solubility of impurities in silicon. The usual description correlates the distribution coefficient of an impurity between solid and liquid silicon to its tetrahedral radius. Impurities of the same group lie on the same experimental line (Trumbore 1960). In this work we shall give a correlation between strain entropy and tetrahedral radius that brings together On the same line 10 elements of Groups IIIA, IVA and VA (about 10% of the Periodic Table), and allows the prediction of solid solubility for two other elements.

We show that the accuracy of the gated diode method for measuring bulk generation lifetime and surface generation velocity in high resistivity silicon depends critically on the gate length of the test device, as a result of nonidealities affecting the gated diode operation. Minimization of the surface generation velocity measurement error requires the gate length to be suitably decreased, while long gate length structures are needed for accurate bulk generation lifetime extraction.

A new test structure intended as a pixel for segmented X- and /spl gamma/-ray detectors has been designed fabricated and tested. The structure consists of a PIN diode of 0.8 mm/sup 2/ area and of an n-channel JFET integrated on the same high resistivity (6 k/spl Omega/ cm) silicon chip. The electrical parameters-leakage current, the transconductance and the capacitance have good expected values. Instead, noise in excess is present in the transistor. Operated at room temperature as an X-ray detector with the integrated frontend transistor in the charge sensitive configuration, the new test structure shows an equivalent noise charge of about 60 electrons rms at the optimum shaping time of 3-6 /spl mu/s.

Anomalies concerning the expected values of resistivity and mobility of indium‐doped silicon in the range where hole scattering is dominated by ionized impurities have been observed. These anomalies can be removed assuming a value of hole effective mass in agreement with the hydrogenlike model of shallow impurities in semiconductors. This suggests a reanalysis of the current picture of hole transport properties in moderately doped semiconductors.