Library

Notes: In this study the structural and optical properties of nanocrystalline Si/SiO2 superlattices have been investigated and discussed. Ordered planar arrays of silicon nanocrystals (Si-nc) have been formed by thermal annealing of ten period amorphous Si/SiO2 superlattices prepared by plasma enhanced chemical vapor deposition. Thermal processing of the superlattices results in well separated (by about 5 nm of SiO2) nanocrystalline Si layers, when the annealing temperature does not exceed 1200 °C. The photoluminescence (PL) properties of these layers have been studied in details. The PL peaks wavelength has been found to depend on the laser pump power; this intriguing dependence, consisting in a marked blueshift for increasing power, has been explained in terms of the longer lifetime characterizing larger Si-nc. It is also observed that these decay lifetimes exhibit a single exponential behavior over more than two orders of magnitude, in clear contrast with the typical, nonsingle exponential trends observed for Si-nc uniformly dispersed inside an insulating matrix. We attributed this peculiar behavior to the lack of interaction among nanocrystals, due to their large reciprocal distance. In agreement with the carrier quantum confinement theory, we have found that the wavelength of the PL peak can be properly tuned by changing the annealing temperature and/or the thickness of the Si layers of the superlattices, and, in turn, the Si-nc mean size. Moreover, the observed lifetimes remain very long (about 0.3 ms) even at room temperature, revealing the absence of relevant nonradiative decay processes in these samples. Furthermore, we have used the experimental PL intensities and decay times to evaluate the radiative rate as a function of the temperature; the obtained data are in good agreement with a model proposed by Calcott in the case of porous silicon. All of these data are presented, discussed, and explained within a consistent picture. © 2000 American Institute of Physics.

Notes: In this article the luminescence properties of Si nanocrystals (nc) formed by plasma enhanced chemical vapor deposition and their interaction with Er ions introduced by ion implantation are investigated in detail. Si nc with different size distributions and densities were produced and all show quite intense room temperature luminescence (PL) in the range 700–1100 nm. It is shown that the time-decay of the luminescence follows a stretched exponential function whose shape tends towards a single exponential for almost isolated nc. This suggests that stretched exponential decays are related to the energy transfer from smaller towards larger nc. Indeed, by comparing samples with similar nc size distributions, but with very different nc densities, it is demonstrated that the PL has a quite strong redshift in the high density case, demonstrating a clear energy redistribution within the sample. Excitation cross sections have been measured in all samples yielding a value of ∼1.8×10−16 cm2 for isolated nc excited with 2.54 eV photons. This effective excitation cross section is shown to increase by a factor of 4 in interacting nc as a result of the energy transfer within the sample. When Er ions are introduced in these samples a strong nc–Er interaction sets in and the energy is preferentially transferred from the nc to the Er ions. The nc-related luminescence is quenched and the Er-related luminescence at 1.54 μm appears. The effective excitation cross section of Er ions through Si nc has been determined to be ∼1.1×10−16 cm2. This number resembles the excitation cross section of nc themselves demonstrating that the coupling is extremely strong. Moreover, by increasing the Er content the effective excitation cross section is seen to increase. In the same concentration range the Er lifetime decreases demonstrating that "concentration quenching" effects, with the energy transferred among Er ions, are setting in. These Er–Er interactions are responsible for the effective increase of the cross section. However, since the increase in the cross section is related to a simultaneous decrease in lifetime the net effect for the luminescence efficiency is negative. The best Er content to take advantage of the sensitizer action of Si nc avoiding the detrimental Er–Er interactions has been determined to be ∼2×1020/cm3. These data are presented and their implications discussed. © 2001 American Institute of Physics.

Notes: The annealing behavior of the damage produced in Si single crystals by 150 keV Au implants at ∼2×1013 ions/cm2 and at substrate temperatures of 77 K (LN2T) and 300 K, room temperature (RT), has been investigated. The annealing kinetics has been studied in the temperature range between 493 and 623 K and for times up to 2 h. Data have demonstrated that, at a fixed temperature some damage anneals out fast and then a saturation occurs for both the LN2T and RT-damaged samples. This behavior has been tentatively interpreted assuming that the damage is mainly composed by amorphous-like material and that reordering is initiated at preferential growth sites present at nonplanar crystal-amorphous interfaces.

Notes: The interaction between high-energy ion irradiation and pre-existing damage clusters dispersed in single-crystal Si is discussed. Silicon substrates were predamaged by low-dose 150 keV Au ions. Post-irradiation by 600 keV Kr2+ ions resulted in either damage annealing or damage accumulation, depending on the substrate temperature. The transition temperature between these two different regimes is 420 K. These data are discussed and compared with the ion beam induced epitaxy and amorphization of continuous surface amorphous layers.

Notes: We investigate in detail the effect of plasma processing on the transient enhanced diffusion of implanted boron in silicon. Thermally oxidized silicon wafers were first processed with CHF3/CF4 plasma and subsequently implanted with boron, with energies ranging from 3 to 20 keV and a dose of 1×1013/cm2. Chemical profiles were measured by secondary ion mass spectrometry while lattice extended defects induced in silicon by plasma processing were characterized by transmission electron microscopy. Secondary ion mass spectrometry measurements reveal that the transient enhanced diffusion of boron after rapid thermal annealing is strongly reduced in plasma processed samples with respect to unprocessed samples. Defects induced by plasma processing are responsible for the reduction by acting as very efficient traps for the interstitial atoms generated during the implant. We note that the trapping efficiency is critically dependent on the projected range of the boron implant, being extremely evident at low energies and less marked as the energy is increased (i.e., when the interstitials generated by the B implant are far away from the trapping sites). By varying the plasma conditions (an argon plasma is used instead of a CHF3/CF4 plasma), we are able to establish a general correlation between trapping defect centers and transient enhanced diffusion reduction. Finally, spreading resistance measurements reveal that the amount of electrically active boron in plasma processed pure epitaxial Si is almost equal to that obtained in samples not exposed to plasma bombardment, thus demonstrating that the plasma processing has no detrimental effect on the boron electrical activation. © 1998 American Institute of Physics.

Notes: Ion beam induced epitaxy of amorphous Si layers onto 〈100〉 substrates has been investigated by varying the As concentration. At As concentrations below 4×1018/cm3 no rate effect is observed. In the intermediate regime, between 4×1018/cm3 and 2×1021/cm3, the growth rate increases linearly with the logarithm of As concentration and reaches a value about a factor of 2 higher than that of intrinsic Si. At concentrations above 2×1021/cm3, the epitaxy experiences a sudden, severe retardation. Finally, at a concentration of ∼6×1021/cm3, twins are observed to form.

Notes: We have investigated the room-temperature migration properties of ion generated defects in crystalline Si. The defects were injected into the bulk of a pure epitaxial Si layer by a low energy (40 keV) Si implant and monitored using a preexisting defect distribution, produced by a high energy He implant, as a marker. The depth of this defective marker layer was changed by varying the He implant energy in the range 1–3 MeV. Spreading resistance measurements show that the injected defects produce a partial annihilation of the pre-existing damage. The magnitude of the annihilation process is strongly dependent on the depth of the defective marker, being very large when this marker is confined within 5 μm from the surface and negligible when it lies beyond a depth of ∼10 μm. From these results, detailed information on the nature of ion-generated defects which are injected and on their migration properties is obtained. It is found that the observed phenomena are due to the annihilation of divacancies and phosphorus-vacancy defect complexes, generated by the He implant, by Si self-interstitials injected by the shallow Si implant. © 1996 American Institute of Physics.

Notes: The electrical activation of B in Si after ion implantation in the energy range between 5 and 160 keV and rapid thermal annealing processes is investigated. It is found that it critically depends on the purity of the substrate as well as the distance from the surface. In particular, while in very pure epitaxial Si layers (where O and C contents are below ∼1×1015/cm3) typically the total B content is electrically active, in Czochralski Si (containing ∼1×1018 O/cm3 and ∼1×1017 C/cm3) the active fraction is very small at doses of 1×1012/cm2 and increases with increasing dose. For very shallow B implants (∼5 keV), the electrical activation in Czochralski Si further decreases to a few percent of the total amount. These results are interpreted in terms of the formation of boron-impurity complexes deactivating the dopant, the dose effect being a result of trap saturation. Vacancies can eventually dissolve some of the inactive complexes. However, close to the surface an enhanced vacancy annihilation process reduces the dissolution probability producing the observed dramatic effects on the electrically active profiles. Finally, at very low energies (∼5 keV), also in epitaxial Si layers, part of the dopant can be electrically inactive due to B–B interactions, the process being negligible at low doses and increasing with dose. The implications of these results on the formation of ultrashallow junctions are discussed. © 1998 American Institute of Physics.

Notes: Two-dimensional profiles of ultralow-energy B implants in Si after diffusion have been studied in detail by scanning capacitance microscopy in connection with a double beveling technique to enhance depth and lateral resolution. Implants have been made into patterned wafers with different feature sizes ranging from 0.8 to 5 μm. It is demonstrated that the B transient enhanced diffusion is strongly reduced with decreasing feature size below about 2 μm. This effect is related to the increasing effect of interstitial lateral out-diffusion under the SiO2 mask. The implication for the formation of ultrashallow junctions in device structures is discussed. © 2001 American Institute of Physics.

Notes: High-resolution scanning capacitance measurements were carried out magnifying the sample dimensions by a double beveling method. A magnification of ten times has been reached, but in principle even higher magnifications can be obtained. For depth magnifications the reverse junction carrier spilling has to be considered. The measurements indicate that the amount of the spilling effect is in agreement with the models developed to date. The method was successfully applied directly to silicon devices and it demonstrates that accuracy well below tip dimensions can be easily reached. Junction depths as well as channel lengths can be determined with a high resolution. © 2000 American Institute of Physics.