Search

Measurements of the ESR linewidth of uncompensated Si:As close to ${n}_{c}$ for the metal-insulator transition have been made at 9.4 GHz in the temperature range 4.2-1.4 K. The linewidth exhibits a minimum value $\ensuremath{\Delta}{H}_{min}$ which occurs at $n={n}_{c}$ and an excess component [$\ensuremath{\Delta}{H}_{\mathrm{ex}}=\ensuremath{\Delta}H\ensuremath{-}\ensuremath{\Delta}{H}_{min}$] which exhibits scaling behavior of the form $B{(\frac{n}{{n}_{c}}\ensuremath{-}1)}^{p}$ for $ng{n}_{c}$. The scaling behavior, established for both Si:As and Si:P (from earlier data), can be explained as a simple extension of the Elliot-Yafet mechanism for conduction-electron spin-resonance linewidths. The strong donor dependence of $B$ is satisfactorily explained with impurity spin-orbit interaction parameters obtained from Orbach spin-lattice relaxation studies.

The mobility of electrons in $n$-type silicon has been calculated and compared with the available experimental data for temperatures between 5 and 80 K. Ionized impurity scattering is treated using the partial-wave phase-shift formalism. It is found that since the Born approximation is invalid at low temperatures, use of Brooks-Herring theory introduces serious error in that regime. For relatively pure samples excellent agreement between theory and experiment is observed, while the agreement is not as good for more highly doped samples. Possible reasons for this disagreement are discussed.

A theoretical investigation has been made of the effect of a drift current on surface polaritons in n-type silicon. The current direction is taken to be the same as the propagation direction of the surface polaritons. Retardation is included, but damping is neglected. In obtaining the dispersion relation, the specular-reflection/mirror-image technique of Kliewer and Fuchs is used to handle the boundary conditions. The results indicate that an interaction takes place between the current and polaritons which gives rise to polariton instabilities for certain frequency ranges. These instabilities are a consequence of the presence of the surface.

ESR linewidth, electronic g-value and ESR spin susceptibility of n-type silicon doped with phosphorus and arsenic are investigated at temperatures between 1.2 and 300 K near a metal-nonmetal transition. Temperature, impurity concentration, and donor dependences of the linewidths are discussed with regard to the spin relaxation mechanism in line-narrowing at low temperatures. ESR-Linienbreite, elektronischer g-Faktor und ESR-Spin-Suszeptibilitat von Phosphorund Arsen-dotiertem n-Silizium werden im Temperaturbereich von 1,2 bis 300 K in der Nahe des Metall-Isolator-Ubergangs untersucht. Die Abhangigkeit der Linienbreite von der Temperatur, der Konzentration der Verunreinigungen und der Art des Donators wird anhand der bei niedrigen Temperaturen vorliegenden Spin-Relaxations-Mechanismen diskutiert.

The two-centre electronic states used in the theories of the hopping conduction in n -type silicon and germanium were obtained until now using inconsistently the “valley-orbit splitting” potential. In the present paper the total effect of this potential was taken into account and the corrections to the previous results were obtained. In some cases the obtained corrections can rise the hopping rate by a factor of 2.3.

Spin-lattice relaxation rate, 1/ T s , of donor electrons in phosphorus and boron doped silicon was studied at liquid helium temperature as a function of the degree of compensation K = N A / N D ( N A , N D : acceptor and donor concentrations). Donor concentrations ranged from approximately 5×10 15 /cc to 3×10 16 /cc and the degree of compensation was varied up to 0.8. It was found that 1/ T s increased first with increasing K , reached a maximum at around K =0.5∼0.6 depending on N D , and then decreased with a further increase in K . If a power law is assumed for the dependence on temperature T , the relaxation rate goes as T n , n increasing with K but being independent of N D . The above experimental results are discussed assuming the hopping of donor electrons. A comparison is made with Sugihara's theory, which gave a rather satisfactory explanation of the experimental results.

The energy loss of fast electrons in heavily doped n-type silicon is calculated using optical constants which are deduced from the conductivity and carrier scattering time using classical dispersion theory. These results are compared with the theory of the collective longitudinal motion of electrons in solids, the so-called plasmon effect.

The magnetoresistance effect has been investigated in $n$-type silicon at 4.2, 77, and 300\ifmmode^\circ\else\textdegree\fi{}K. At 77 and 300\ifmmode^\circ\else\textdegree\fi{}K, the low-field magnetoresistance coefficients have been evaluated in the impurity concentration range from approximately 1\ifmmode\times\else\texttimes\fi{}${10}^{15}$ to 1\ifmmode\times\else\texttimes\fi{}${10}^{20}$ ${\mathrm{cm}}^{\ensuremath{-}3}$. At 300\ifmmode^\circ\else\textdegree\fi{}K, the symmetry relation among the coefficients found in pure silicon is obeyed at all impurity concentrations, while at 77\ifmmode^\circ\else\textdegree\fi{}K, the symmetry relation is fulfilled only up to impurity concentrations of 1\ifmmode\times\else\texttimes\fi{}${10}^{18}$ ${\mathrm{cm}}^{\ensuremath{-}3}$. The mobility anisotropy has been evaluated from the low-field coefficients in the impurity concentration ranges where the symmetry relation is obeyed. At 4.2\ifmmode^\circ\else\textdegree\fi{}K, the magnetoresistance effect is negative and tends toward saturation with increasing magnetic field strength for impurity concentrations greater than 5\ifmmode\times\else\texttimes\fi{}${10}^{18}$ ${\mathrm{cm}}^{\ensuremath{-}3}$. Below this concentration, a large positive magnetoresistance is observed and the change in sign is correlated with the appearance of an impurity ionization energy.

Measurements of the Faraday rotation in n-type silicon have been performed at room and at low (93 °K) temperatures. A crossed-waveguide technique (18.0 to 26.5 GHz), similar in principle (but in a simplified form) to the technique of Bouwknegt and Volger [4], has been used for the measurements. The rotation has been studied as a function of de magnetic field up to 10 kOe. The effect of multiple reflections has also been studied at low temperatures. The theory of the Faraday effect in anisotropic n-silicon, previously developed by the authors [3], is modified taking into account the contribution of mixed scattering of lattice and impurities. The modified theory is compared with the experiment, and an estimation is made of the relative importance of the impurity over the lattice scattering. It is shown that impurity scattering gives an appreciable contribution to the rotation. Bei Zimmertemperatur und niedrigeren Temperaturen (93 °K) wurden Messungen der Faraday-Rotation in n-leitendem Silizium durchgefuhrt. Eine Technik gekreuzter Hohl-leiter (18,0 bis 26,5 GHz), die im Prinzip der Technik von Bouwknegt und Volger [4], jedoch in vereinfachter Form, ahnlich ist, wurde fur die Messungen benutzt. Die Rotation wurde als Funktion des konstanten Magnetfeldes bis zu 10 kOe untersucht. Desgleichen wurde der Effekt der Mehrfachreflexionen bei niedrigen Temperaturen gemessen. Die Theorie des Faraday-Effekts im anisotropen, n-leitenden Silizium, die von den Autoren [3] entwickelt wurde, wird modifiziert und berucksichtigt den Beitrag der gemischten Streuung von Gitter und Storstellen. Die modifizierte Theorie wird mit dem Experiment verglichen und die relative Bedeutung der Storstellen-gegenuber der Gitterstreuung berechnet. Es wird gezeigt, das Storstellenstreuung einen bedeutenden Beitrag zur Drehung liefert.

The photomagnetoelectric (PME) effect was observed between 21 and 84\ifmmode^\circ\else\textdegree\fi{}K in silicon doped with gold and phosphorus. A generalized diffusion equation is formulated and solved with the impurity centers included in the charge-balance equation and arbitrary intensity of steady illumination. When the carrier densities are much smaller than the gold density, the charged impurities maintain charge neutrality, and this results in a range of injection where $\ensuremath{\Delta}p={\ensuremath{\Gamma}}_{1}\ifmmode\times\else\texttimes\fi{}\ensuremath{\Delta}n+{\ensuremath{\Gamma}}_{2}\ensuremath{\Delta}{n}^{2}$. If the quadratic term dominates, the PME current is proportional to the $\frac{4}{3}$ power of the photoconductance. This power law is observed in our measurements, from which we deduce values for the product $\ensuremath{\gamma}_{\frac{1}{2}}^{\frac{1}{2}}{\ensuremath{\tau}}_{n}$ between 36 nsec and 63 \ensuremath{\mu}sec.

Earlier measurements have been extended to a larger number of samples of high crystal-line perfection, high purity, and well-defined homogeneous doping. Furtheron the measuring equipment has been improved. The before madc statement could be confirmed, that the transverse electron mass as derived from the magnetic susceptibility in silicon–germaniuni alloys with a silicon content between 65 and 100 at% Si is constant with a value m⟂ = 0.27 m (m free electron mass). Fruliere Messungen wurden mit einer verbesserten expcrimentellen Anordnung auf eine grose Zahl von Proben ausgedehnt, die eine hohe Kristallqualitat. hohe Reinheit und hornogene wohldefinierte Dotierung besasen. DAS fruhere Ergebnis wurde bestlitigt : die transversale effektive Masse m, ⟂ = 027 m, wie sie sich aus der niagnetischen Suszeptibilitat ergibt, ist fur Silizium–Germanium-Legierungen mit einem Siliziamgehalt zwischeri 65 und 100 At% si konstant (m Masse des freien Elektrons).

Introduction of impurity atoms in a semiconductor produces strain in the lattice. This strain depends on the nature and concentration of impurity for concentrations up to about 1019 atoms cm-3 in silicon. The expression for the low field mobility of electrons which takes into account the departure from periodicity due to strain is studied for As, P and Sb impurities in silicon. The numerical values of mobility calculated from experimental values of strains due to impurities in silicon agree well with the experimentally observed values of the electron mobility.

The g-factor of conduction electrons in the surface inversion layer on a silicon (100) surface has been determined using the tilted magnetic field method developed by Fang and Stiles. The value of (m∗/m0)·g at the fixed magnetic field was independent of surface carrier density ns, whereas it had a sharp peak at about 97 koe. At strong magnetic field limit the value was constant and 0.4. If we take the effective mass of conduction electrons in the inversion layer on the (100) surface as 0.2m0, the g-factor is about two which is the same as that for conduction electrons in bulk silicon.

Results of a detailed investigation of recombination in phosphorus‐doped Co60 γ‐irradiated n‐type silicon are reported. An analytical procedure is employed which allows, in principle, a complete and unambiguous determination of parameters associated with recombination centers. Properties of two recombination centers are determined. The level most important at low excess densities is determined to be 0.40 ±0.015 eV from the conduction band.cn varies as exp(0.06±0.015eV/kT), and cp varies as exp(0.03±0.015eV/kT). The capture‐probability ratio cp/cn varies from 8 ± 1 at 303°K to 9.5 ± 1 at 370°K. The level dominant at high excess densities is placed at 0.21 ± 0.03 eV below the conduction band with cn ∼exp(0.01±0.01eV/kT) and cp ∼exp(−0.05±0.03eV/kT). The corresponding capture probability ratio, cp/cn , varies from ∼5.6 at 303 °K to ∼8 at 370 °K. Based upon dependence on phosphorus concentration and annealing behavior, the deeper level appears to belong to the E center. There is some evidence that the other level is also associated with phosphorus‐containing defects.

In this work the photodielectric effect in antimony doped silicon has been studied by measuring the real and imaginary parts of the dielectric constant of the sample in the frequency range from 0·5 to 30 MHz and for a wide range of incident radiation using Schering Bridge and calibrated attenuators. The results have been interpreted in terms of Maxwell-Wagner Effect in hetrogeneous dielectric having regions with different conductivity, dielectric constant or both.

Spin susceptibility, $g$-value, and ESR linewidth data are presented for (Si:P) samples having impurity concentrations ${10}^{18}l{N}_{D}l{10}^{19}$ ${\mathrm{d}\mathrm{o}\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{s}/\mathrm{c}\mathrm{m}}^{3}$. These results and those previously obtained for ${N}_{D}g{10}^{19}$ ${\mathrm{d}\mathrm{o}\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{s}/\mathrm{c}\mathrm{m}}^{3}$ samples are compared with the predictions of alternative simple models of the semiconductor-metal transition in heavily doped semiconductor materials. At low concentrations (${N}_{D}l3\ifmmode\times\else\texttimes\fi{}{10}^{18}$ ${\mathrm{d}\mathrm{o}\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{s}/\mathrm{c}\mathrm{m}}^{3}$) our samples behave as semiconductors in transport experiments and are characterized by Curie-Weiss-law spin susceptibilities. Near the upper limit of our sample impurity concentration range (${N}_{D}\ensuremath{\sim}{10}^{20}$ ${\mathrm{d}\mathrm{o}\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{s}/\mathrm{c}\mathrm{m}}^{3}$), transport measurements indicate metallic properties and spin susceptibilities which follow the Pauli-law expression with an effective mass which is slightly smaller than that usually associated with the silicon conduction band. A good representation of the experimental spin susceptibilities of samples having intermediate impurity concentrations can be achieved through an expression which is the sum of two terms: one of which is of the Curie-Weiss form and the other of the Pauli type. An analysis of previously published Hall carrier-density data is offered which would appear to allow the identification of these two components with nonconducting (or partially localized) and conducting extrinsic electrons, respectively. The inferred presence of partially localized electrons even in samples with ${N}_{D}g{10}^{19}$ ${\mathrm{d}\mathrm{o}\mathrm{n}\mathrm{o}\mathrm{r}\mathrm{s}/\mathrm{c}\mathrm{m}}^{3}$ is discussed with respect to previously published magnetoresistance and Knight-shift data.

The magnitude of the subband splitting in inversion layers was investigated with piezoresistance measurements of n-type inversion channels on (100) and (110) silicon surfaces. Below 70 K the piezoresistance effect is no longer linear with stress and surface field. At high tensile and compressive stress two saturation regions occur corresponding to the magnitude of the subband splitting. The experimental results were compared with several models describing surface quantization. A self-consistent calculation shows the best agreement, whereas the triangular potential well model deviates more than 100%. For a (100) transverse configuration an anomalous behaviour was observed which is explained in terms of different relaxation times in the various subbands. Es wurde die Grose der Subbandaufspaltungen in Inversionsschichten mit Piezowiderstandsmessungen von n-leitenden Inversionskanalen auf (100)- und (110)-Siliziumoberflachen untersucht. Unterhalb 70 K ist der Piezowiderstandseffekt nicht langer linear mit der Spannung und dem Oberflachenfeld. Bei hohen Dehnungs- und Kompressionsspannungen treten zwei Sattigungsbereiche auf, die der Grose der Subbandaufspaltung entsprechen. Die experimentellen Ergebnisse werden mit verschiedenen Modellen der Oberflachenquantisierung verglichen. Eine selbstkonsistente Berechnung zeigt die beste Ubereinstimmung, wahrend das Dreiecks-Potentialtopfmodell um mehr als 100% abweicht. Fur eine transversale (100)-Konfiguration wurde ein anomales Verhalten beobachtet, das mit unterschiedlichen Relaxationszeiten in den verschiedenen Subbandern erklart wird.

Films of Ni and Co have been grown on the (111) surface of phosphorus doped n -type silicon by evaporation in high vacuum. For both systems polycrystalline films resulted for all substrate temperatures except in the range 250° to approximately 400°C when f.c.c. Ni or Co grew with the orientation Ni/Co (110)‖Si (111); Ni/Co [ 1 10]‖Si [ 2 11].

The theory of free carrier Faraday effect in anisotropic semiconductors, previously developed by Donovan and Webster [1,2], is extended to the case of n-type silicon. The high-frequency magnetoconductivity tensors are derived taking into account the detailed structure of conduction band. The Faraday rotation measurements in some [111] oriented samples of n-silicon are performed at room temperature and microwave frequencies. These results and the results of previous workers with different orientation of silicon samples, show reasonable agreement with the theory. Die fruher von Donovan und Webster [1,2] entwickelte Theorie des Faraday-Effekts freier Ladungstrager in anisotropen Halbleitern wird auf n-leitendes Silizium ausgedehnt. Die Tensoran der Magneto-Hochfrequenzleitfahigkeit werden abgeleitet, wobei die detaillierte Struktur des Leitungsbandes berucksichtigt wurde. Die Messungen der Faraday-Drehung werden in einigen in [111]-Richtung orientierten Proben von n-leitendem Silizium bei Zimmertemperatur und Mikrowellenfrequenzen durchgefuhrt. Diese Ergebnisse und Ergebnisse fruherer Autoren bei verschiedenen Orientierungen der Si-Proben zeigen befriedigende Ubereinstimmung mit der Theorie.

A synthesis is given of the experimental evidence concerning the nature of the transition from semiconducting to metallic behavior in n-type silicon as impurity concentration is varied. Two stages in the transition are described; first, the delocatization transition and a second distinct change associated with movement of the Fermi level into the conduction band of the host silicon.

High field current—voltage characteristics of N-type silicon have been calculated by the Monte Carlo method for the and directions at different temperatures. The curve exhibits a region of negative differential conductivity for temperatures below about 50°K. Coherent high frequency current oscillations have been found experimentally for samples at and below about 50°K.

The Boltzmann equation describing the warm-electron case is discussed and a review is given of the scattering mechanisms for n-Ge and n-Si with relatively low doping levels. Taking into account the known band structure, the Boltzmann equation is solved by a numerical iteration method under the assumption of weak intervalley scattering. It is shown that this condition can be relaxed for special symmetry directions. The warm-electron coefficient beta has been measured in the temperature range from 77 to 250°K by an audio-frequency method based on analysis of nonlinear distortion. Good agreement between measured and calculated results is obtained using the accepted values of the deformation-potential constants.

Comparison of low-field mobility and piezoresistance data with calculations gives support to the importance of quadrupole and octupole electron-phonon interactions in n-type silicon.

The absolute reflectivity of a series of p and n-type silicon crystals has been measured in the region from 3.4 eV to 5.3 eV. The energies of two maxima in reflectivity did not shift for impurity concentrations up to 1020/cm3. No essential difference has been found between n and p-type samples. The 3.4 eV peak is showing a growing asymmetry at higher concentrations. The maximum in reflectivity in the 4.5 eV peak occurs at 4.54 ± 0.02 eV.

The (111) surface of n-type Si containing 1 part in 2×103 of phosphorus has been examined by LEED and Auger electron spectroscopy (AES). A Si(111)-P LEED pattern was observed after heating for 10 sec at 1000 °C followed by rapid cooling to room temperature. With prolonged heat treatment this was replaced by the characteristic Si(111)-7 pattern, but phosphorus was still detectable on this surface. Several models are proposed to explain the observed Si(111)-P LEED pattern. The LEED and Auger data appear to be consistent with incomplete coverage of the silicon by phosphorus or a silicon-phosphorus compound in parallel orientation and with a lattice spacing 18% different from that of silicon.

Experimental evidence of temporary trapping of electrons in the volume of $p$-type silicon in two different traps is presented. The evidence is shown to lead to a multiple trapping model the kinetics of which explains the data. The trap parameters are determined by the fit of experimental data to the theory. The trap density, energy level, cross section for capture, $S$, and the time which an electron spends in a trap, ${\ensuremath{\tau}}_{g}$, are determined for both traps. For the deeper traps, $S{\ensuremath{\tau}}_{g}$, a property of the trap itself, is shown to have the probable value 0.81\ifmmode\pm\else\textpm\fi{}0.25\ifmmode\times\else\texttimes\fi{}${10}^{\ensuremath{-}13}$ ${\mathrm{cm}}^{2}$ sec with ${\ensuremath{\tau}}_{g}\ensuremath{\cong}0.3$ sec for all specimens examined. In low-resistivity crystals there is evidence for a loss mechanism for electrons (recombination) from the deeper traps at a rate found proportional to the square of the hole concentration. The deep-trap concentration is found to be roughly proportional to the sample conductivity.

A number of methods has been devised in the past few years for measuring the lifetime of minority carriers in semiconductors. With germanium all these methods give results consistent with each other. When another semiconductor, namely silicon, is studied quite large differences in the results from different methods are found. These discrepancies are described and their origin discussed. It is concluded that an uncritical application of well-established germanium technology to any other semiconductor may lead to erroneous conclusions.

The factor relating the Hall coefficient to carrier concentration has been determined in n-type silicon by combining optical measurements of the carrier concentration at infrared frequencies with conventional dc Hall effect measurements. Interpreting this factor in terms of electronic charge, mobility anisotropy and the usual scattering constant, the latter is found to reach values much greater than unity, especially at 77 °K, when with a doping level of 2.6 × 1017 cm−3 a value of 7.3 is obtained. Fur n-leitendes Silizium wurde der Faktor bestimmt, der den Hall-Koeffizienten mit der Tragerkonzentration verknupft, indem optische Messungen der Tragerkonzentration bei Infrarot-Wellenlangen mit herkommlichen Gleichstrom-Halleffektrnessungen kombiniert wurden. Bei der Interpretation dieses Faktors hinsichtlich Elektronenladung, Anisotropie der Beweglichkeit sowie der ublichen Streukonstante ergeben sich fur die letztere Werte groser als eins; insbesondere wurde bei 77 °K und einer Dotierungskonzentration von 2,6 × 1017 cm−3 ein Wert von 7,3 gefunden.

Gold was diffused into arsenic‐doped silicon slices with initial resistivity of 0.3 or 5 Ω·cm at temperatures between 850° and 1250°C for times sufficiently long that the gold concentration was essentially constant, except for a very thin surface region which had a much larger concentration. The gold concentration was determined by means of neutron activation analysis. It was found that the gold concentrations were lower than would be expected from published solubility data when diffusion occurred at temperatures below 1100°C. Several surface treatments were found to result in similar gold concentrations when an evaporated gold film was used as the source. The measured room‐temperature resistivity was found to be in fair agreement with computations using accepted values for the various silicon parameters and the gold energy levels. Although refinements in the calculations and improved experimental techniques are both still necessary, it has been shown that the qualitative features of the model on which th...

Light-induced degradation of the carrier lifetime in silicon due to the formation of boron-oxygen defects has been investigated using photoluminescence measurements. A finite difference simulation method has also been created to test theoretical models of the factors controlling defect generation. Experimental photoluminescence data is compared with simulated results. It is found that the maximum defect density in both p- and n-type material can be described as a function of the hole density during degradation.

An effective passivation property on n-type silicon surface was obtained using liquid source SiO 2 combining with a hot steam post annealing. Perhydropolysilazane (PHPS), which reacts with water in air, changing into pure quartz was used as a source material. The passivation property was estimated by effective lifetime and interface state density (D it ). The D it of the sample after the hot steam annealing at 600 °C was around 1x10 12 cm 2 , though the effective lifetime reached a higher value than the chemical passivation using 1% quinhydron methanol solution. On the other hand samples annealed in air showed a much lower value. The film after the hot steam annealing has a large number of positive charges, whose field effect might improve the effective lifetime. Regarding the long term stability, samples with thin passivation layer of less than 50 nm showed high lifetime stability at room temperature.

Photocatalytic decomposition of aqueous formic acid was successfully achieved by using platinized n-type silicon powder as a photocatalyst under the conditions that the photocatalyst was spread on the bottom of a reaction cell.

An additional splitting of Shubnikov-de Haas oscillations in (lll) n-type silicon inversion layers in observed. We attribute it to valley splitting, analogous to that previously observed on (100) surface layers, and point out that theoretical explanations of valley splitting, restricted to the valley configuration in (100) inversion layers, are not applicable.

Etude des profils de concentrations des niveaux profonds introduits dans Si n par implantation par He a des doses de 10 14 -10 15 cm −2 , par mesures C-V avec contact de Schottky. Variation du pic de distribution en profondeur avec l'energie d'impact; effet d'un recuit thermique dans N 2 entre 500 et 700°C

The electrical characteristics of the sputtering induced defects have been studied using the DLTS measurements. A band of defect states located at Ec-0.3 eV to Ec-0.5 eV was found. The measured capture cross-section for majority carriers at this band is 1.3 × 10−19 cm2. The concentration profile of this band of defects decreases exponentially from the surface. This initial concentration of the defects near the surface of a sputtered platinum silicide diode is extrapolated to 7 × 1017 cm−3. The defects anneal out after 10 min of annealing time at 400 °C.

To overcome difficulties which arise when either the series resistance is too high or the Schottky barrier height is too low, a new and simple method was developed to measure barrier height. Crucial point is the application of a ‘‘one‐sided’’ configuration. This method was successfully applied to the GdSi2 system, where φB =0.38±0.005 eV was measured.

The low temperature specific heat of an n-type phosphorous doped silicon sample containing 5 × 10 18 charge carriers/cm 3 in zero magnetic field and in a magnetic field of 28.5 kGauss has been measured. Within the experimental accuracy of the experiments no magnetic contribution to the specific heat has been detected. This result is discussed in the light of recent E.S.R. experiments on the same impurity banded silicon sample.