Doping and radiation damage profiles of P+ ions implanted in silicon along the [110] axis

Several doses of 200 KeV phosphorus ions have been implanted under channeling conditions along the [110] direction in silicon.

Range distribution has been determined for the three implant doses 10¹³, 10¹⁴, 10¹⁵ P⁺/cm² both with the electrical measurements and the neutron activation techniques.

The radiation damage distribution has been determined both with 290 KeV proton back-scattering analysis and with transmission electron microscopy (TEM) observations.

Good agreement has been found between electrical and neutron activation profiles in the samples where 100% of the implanted dose had been electrically activated by means of annealing.

Carrier concentration profiles, from samples implanted with 10¹⁵ P⁺/cm², determined after two different annealing temperatures (500°C and 700°C) have bcen compared with the radiation damage distribution and a correlation between damage and phosphorus electrical activation process seems to be possible.

Maximum damage peak, as determined by back-scattering analysis, shifts from ～0.4 μ depth in the lower dose(5 × 10¹⁴ P⁺/cm²), to ～0.22 pm depth in the higher implanted dose (4 × 10¹⁵ P⁺/cm²). Damage distribution of phosphorus ions random implanted in the same experimental conditions shows 3 peak at ～0.2 μn depth.

In accordance with the back-scattering analysis, T.E.M. observations on 10¹⁵P⁺/cm² implanted samples show the presence of amorphous regions at depth between 0.25 and 0.5 μm from the surface. In the most damaged layer ～0.3μm in depth, a surface density of ～10¹²/cm² amorphous regions 25-50 A diameter was observed.     

Low energy boron and phosphorus implants in silicon (a) electrical sheet measurements

Silicon wafers were implanted in 〈111〉-direction with boron and phosphorus ions of 7 keV at room temperature. Doses between 10¹² and 10¹⁸ ions/cm² were applied. After successive annealing steps the electrical properties of the implanted layers have been determined by Hall effect and sheet resistivity measurements. The annealing characteristics of the implants depend on ion dose and species. Three annealing stages can be distinguished: (I) the temperature range below 500°C, (II) 500—700°C, (III) 700—900°C.

After annealing at 90°C the apparent electrical yield is proportional to dose for all implants and amounts to approx. 80 per cent for boron and 40 per cent for phosphorus.

Sheet resistivity vs. dose curves were derived for the annealing temperature of 400°C and used for the fabrication of position sensitive detectors. The position characteristics were found to be linear within ～1 per cent for resistive layers as long as 20 cm.     

Low energy boron and phosphorus implants in silicon (b) doping profiles

Silicon wafers were implanted in 〈111〉-and 〈110〉-direction with boron ions of 6 keV and phosphorus ions of 20 keV at room temperature. Doses of 10¹⁴ ions/cm² were applied. At four different temperatures, 300, 420, 600, and 900°C. a few samples of each type of implant were annealed. Standard electrical techniques combined with successive layer removals were used to determine the depth distribution of electrically active centers. Since the method of using non type inverting implanted layers was applied, the local annealing behavior over the whole penetration region could be measured.

For both Si(B)-and Si(P)-implants the part of the profile beyond approx. 0.15 μm, i.e. the deeper part of the channeling and the whole supertail region, is unaffected by going from 300 to 900°C. All additional annealing, with respect to the electrical yield, takes place in the amorphous range and the adjacent part of the channeling range. After raising the temperature from 600 to 900°C both B-and P-profiles undergo diffusion controlled changes in this latter region.     

Impurity redistribution in Pb implanted and annealed silicon

Reordering of 〈111〉 silicon, implanted with Pb ions at energies >100 keV and fluences ～5 × 10¹⁵ cm^?2 is accompanied by substantial impurity indiffusion in addition to pronounced outdiffusion and accumulation at the near surface region.     

Structure, electrical and magnetic properties, and the origin of the room temperature ferromagnetism in Mn-implanted Si

The structure and the electrical and magnetic properties of Mn-implanted Si, which exhibits ferromagnetic ordering at room temperature, are studied. Single-crystal n- and p-type Si wafers with high and low electrical resistivities are implanted by manganese ions to a dose of 5 × 10¹⁶ cm^?2. After implantation and subsequent vacuum annealing at 850°C, the implanted samples are examined by various methods. The Mn impurity that exhibits an electric activity and is incorporated into the Si lattice in interstitial sites is found to account for only a few percent of the total Mn content. The main part of Mn is fixed in Mn₁₅Si₂₆ nanoprecipitates in the Si matrix. The magnetization of implanted Si is found to be independent of the electrical resistivity and the conductivity type of silicon and the type of implanted impurity. The magnetization of implanted Si increases slightly upon short-term postimplantation annealing and disappears completely upon vacuum annealing at 1000°C for 5 h. The Mn impurity in Si is shown to have no significant magnetic moment at room temperature. These results indicate that the room temperature ferromagnetism in Mn-implanted Si is likely to be caused by implantation-induced defects in the silicon lattice rather than by a Mn impurity.     

Damage anneal of antimony/phosphorus double implants in silicon

A method is presented for avoiding the dislocation generation in (100) silicon implanted with phosphorus doses up to 5×10¹⁵ ions/cm² at 50 keV. The residual defects after the damage anneal are considerably reduced if the phosphorus implant is combined with a low dose, e.g. 1×10¹⁴ ions/cm², antimony implant which produces a deeper surface layer of amorphous silicon. It is essential that the phosphorus ions are implanted shallower than the antimony ions, and come to rest within the amorphous layer. Subsequent thermal annealing proceeds by a solid phase epitaxial regrowth mechanism.     

Compound synthesis in TiO2 implanted with Rubidium

Rubidium ions, with energy in the range 0.1 MeV, 2.0 MeV have been implanted in TiO₂ single crystals at RT and LNT.

Defects induced by implantation have been studied by optical spectroscopy, X-ray diffraction, RBS, TEM and electrical conductivity.

During implantation, the implanted samples are blue colored after irradiation. This coloration is due to an optical absorption band localized at 900 nm which corresponds to optical transition of intrinsic defects identified as Ti³⁺. These defects are induced by a chemical reaction between the implanted ions and the oxygen of the lattice as in the case of D⁺, H⁺, Li⁺, Na⁺ and K⁺ implanted in rutile.^1-3

The synthesis of a new phase in heavily implanted rutile is exhibited by using a thermal treatment and by combining techniques such as RBS, TEM and X-ray diffraction at glancing angle in the temperature range 300°C-700°C.

This compound does not correspond to metallic precipitates of rubidium which are observed in MgO implanted with Rb ions.

Planar defects have been observed in the implanted area. A correlation is exhibited between these defects and the precipitates of the new phase. From X-ray diffraction measurements and TEM observations, the composition of the synthetized compound is likely Rb₂TiO₃.     

Antimony implanted silicon: A comparison between the total implanted concentration profile and the donor concentration profile

Abstract

The total concentration profiles of various doses of antimony, implanted into silicon at 100 keV, have been determined by a new technique, using Kr⁺ ions to detect selectively the antimony (as Sb-M X-rays) at the expense of the silicon. Since most of these X-rays arise from only a few tens of Angstroms below the surface of the silicon, this allows the X-ray generation to be used in conjunction with an anodic stripping technique to obtain the antimony depth distribution. These profiles are compared with others, obtained by measuring the donor concentration as a function of depth, using standard Hall effect and conductivity measurements. A significant difference between these profiles was observed, which is thought to be due to the suppression of electrical activity which occurs as the result of lattice damage. Confirmatory evidence is presented in the form of electron microscope observations of the implanted region at various depths below the silicon surface.     

Formation and nature of swirl defects in silicon

Point defect agglomerates in dislocation-free silicon crystals, usually called “swirls”, have been investigated by means of high-voltage electron microscopy. It was found that a single swirl defect consists of a dislocation loop or a cluster of dislocation loops. By contrast experiments it could be shown that these loops are formed by agglomeration of self-interstitial atoms. Generally the loops have a/2〈110〉 Burgers vectors, but in specimens with high concentrations of carbon (～10¹⁷ cm^?3) and oxygen (～10¹⁶ cm^?3) also dislocation loops including a stacking fault were observed. In crystals grown at growth rates higher thanv=4 mm/min no swirls are observed; lower growth rates do not markedly affect the size and shape of the dislocation loops. With decreasing impurity content (particulary of oxygen and carbon) the swirl density decreases, whereas the dislocation loop clusters become larger and more complex. A model is presented which describes the formation of swirls in terms of agglomeration of silicon self-interstitials and impurity atoms.     

The electrical behaviour of abrupt ion implanted and diffused P+N junctions

Abstract

Results are reported of measurements of the properties of diodes formed by ion implantation, and for comparison boron diffused P⁺N diodes of similar area close by on the same chip. The four group III acceptor impurities were implanted separately to a dose of 5 × 10¹⁵ ions/cm² at room temperature into similar samples of suitably masked silicon. Boron ions were also implanted at liquid nitrogen temperature and 450°C. Annealing was limited to a maximum temperatare of 550 °C.

Measurements have been made of sheet resistance, forward and reverse I-V characteristics (from 10^?9 amps/cm²), reverse breakdown voltage, noise, minority carrier storage time and junction series resistance.

The bulk properties of boron implanted diodes were found to be reproducible. The introduction of recombination centres by implantation is the major factor influencing variation in these properties between one implantation condition and another. Changes in surface oxide conditions probably affect reverse leakage currents and breakdown voltages.

The properties of boron implanted diodes are considered suitable for applications such as the MOSFET, and are superior to those of the AI, Ga and In implanted diodes.     

The measurement of cone length and range in etched plastics

Abstract

Group V impurities implanted at 400 keV into silicon have been detected in substitutional lattice positions by EPR. Three samples of VFZ, p-type 1200–1500 ohm-cm silicon from the same ingot were implanted with As⁷⁵, Sb¹²¹, and Sb¹²³, respectively. The EPR spectrum of each implanted substitutional impurity was observed after annealing the lattice damage. Only the isotope implanted in each sample was seen. Since only those donors which are electrically active can be observed, this technique measures the electrically active fraction of the implanted species. Upon annealing to 970°C, most of the antimony was active whereas only about 1/5 of the arsenic was observed. Comparisons with backscattering results indicate that between 350 and 600°C, ～95 per cent of the implanted antimony is substitutional but ～0 per cent is electrically active. The increase in electrical activity at 600°C is due to the rise of the Fermi level to the donor level as the residual lattice damage anneals. The paramagnetic damage centers observed were those also seen in oxygen-implanted silicon, Si-P3 and Si-Pl, but the Si-P3 center was not as well resolved and grows upon annealing to 200°C.     

Magnetic and transport properties of transition-metal implanted ZnO single crystals

ZnO single crystals were implanted with Mn, Co and Ni with fluences between 1 × 10¹⁶ cm^-2 and 1 × 10¹⁷ cm^-2 and energy of 200 keV. Results indicate that aggregation of transition metal ions in the as implanted state occurs only in the case of Ni. After an annealing stage to recover the ZnO structure aggregation occurs for the higher fluences of all implanted species. For lower concentrations paramagnetic behaviour with magnetic moments close to those of individual ions is observed. No polarised impurity band is formed as a result of the presence of transition metal ions and all samples show electrical conduction by carriers in extended states of ZnO. Significant values of magnetoresistance are measured at low temperatures, where electrical transport is described by hopping mechanisms between localized states. The sign of the magnetoresistance is dependent of the doping ion and is correlated with the observed aggregation.     

Lattice location and dopant behavior of group II and VI elements implanted in silicon

Outdiffusion, lattice location and electrical behavior of Zn, Cd, Hg and Se, Te implanted into silicon at 50 keV were investigated by backscattering and channeling effect of 1 MeV He⁺ ions and by Hall effect and sheet resistivity measurements. All the species exhibited outdiffusion with thermal processing. A significant fraction of Zn, Cd and Hg, when implanted into a substrate of 350°C, occupied regular interstitial lattice sites, while 50–60 per cent of the Se and Te atoms were on substitutional lattice sites. Selenium implanted at room temperature and mercury implanted into a substrate of 350°C exhibited depth dependent lattice location. The implanted layers showed n-type behavior: the maximum value of number of carriers/cm² was less than the number of implanted ions/cm² in all cases. The highest electrical activity was observed for Se corresponding to 25 per cent of the substitutional component.     

Mössbauer study of119Sn defects in silicon from ion implantations of radioactive119In

Radioactive¹¹⁹In⁺ ions (T _1/2=2.1 min) obtained from the ISOLDE facility at CERN have been implanted into silicon single crystals at room temperature. Mössbauer emission spectra from the 24 keV -radiation of the daughter¹¹⁹Sn have been measured by fast resonance-counting technique. Five independent lines, characterized by their hyperfine parameters and Debye temperatures, have been found in the spectra. From the bonding configurations, deduced for the Sn impurity atoms, these are concluded to be located in four different defects in the silicon lattice. Simple models are proposed for the defects.     

Electron irradiation assisted annealing of boron and phosphorus implanted silicon layers

Radíatíon annealing due to a 1.0 MeV election beam of intensity 25 μA/cm² was studied in silicon samples implanted with phosphorus and boron ions and annealed at 350–500°C. A significant annealing enhancement as compared to thermal annealing has been observed in phosphorus-implanted samples. In boron-implanted samples, a fast initial rise of electrical activity is followed by a continuous decrease of carrier concentration. The results are interpreted in terms of two competing processes: electron irradiation induced removal of post-implantation defects and introduction of simple electrically active defects.     

Formation and diffusion of self-interstitial atoms in silicon crystals under hydrostatic pressure: Quantum-chemical simulation

A theoretical modeling of the formation of Frenkel pairs and the diffusion of a self-interstitial atom in silicon crystals at normal and high (hydrostatic) pressures has been performed using molecular dynamics, semiempirical quantum-chemical (NDDO-PM5, PM6), and ab initio (SIESTA) methods. It is shown that, in a silicon crystal, the most stable configuration of a self-interstitial atom in the neutral charge state (I ⁰) is the split configuration 〈110〉. The shifted tetrahedral configuration (T ₁) is stable in the singlet and triplet excited states, as well as in the charge state Z = +2. The split 〈110〉 interstitial configuration remains stable under hydrostatic pressure (P ≤ 80 kbar). The activation barriers for diffusion of self-interstitial atoms in silicon crystals are determined to be as follows: ΔE _a (Si)(〈110〉 → T ₁) = 0.59 eV, ΔE _a (Si)(T ₁ → T′₁) = 0.1 eV, and ΔE _a (Si)(T ₁ → 〈110〉) = 0.23 eV. The hydrostatic pressure (P ≤ 80 kbar) increases the activation barrier for diffusion of self-interstitial atoms in silicon crystals. The energies of the formation of a separate Frenkel pair, a self-interstitial atom, and a vacancy are determined. It is demonstrated that the hydrostatic pressure decreases the energy of the formation of Frenkel pairs.     

Beryllium-hydrogen and zinc-hydrogen shallow acceptor complexes in germanium

New shallow acceptor complexes with hydrogenically-spaced excited states have been discovered in intentionally-doped crystals of otherwise ultra-pure germanium. The doping consists of >10¹⁴ cm⁻³ of group II impurities, and the crystals were grown under hydrogen atmosphere. The identification proposed in this paper is that of a hydrogen-group II impurity complex, with piezospectroscopic behavior of a 〈111〉-oriented defect.     

Investigation of gallium arsenide single crystals with various crystallographic orientations, after implantation of silicon ions and pulsed photon annealing

The results of experimental investigations of gallium arsenide single crystals with the orientations (100), (311)A, (211)A, (111)A, and (221)A are presented. The crystals were doped with silicon ions on the Iolla-3M setup (ion energy 75 keV, ion beam density 1 μA/cm², implantation dose 1.2×10³ cm⁻²) at room temperature and annealed on the Impul’s-5 setup at 950°C. Raman scattering and low-temperature photoluminescence methods established that the highest electrical activity of the implanted silicon under identical implantation and annealing conditions obtains for (100) and (311)A gallium arsenide. In the process n-type layers are produced. Zh. Tekh. Fiz. 69, 78–82 (May 1999) Deceased.     

Interference of arsenic diffusion by argon implantation

In the study of ion implantation, electrically active ions or noble gas ions are often used for damage study, range profiling, etc. Very seldom are both electrically active ions and noble gas ions implanted at about the same depth. In the work reported here, argon and arsenic ion implants and their interference in diffusion were studied by using backscattering, electrical measurements, and transmission electron microscopy (TEM). Several unexpected phenomena were observed.

First, when both Ar and As are implanted in high doses (about 10¹⁶/cm²), at depths around a few hundred nanometers, the Ar significantly hampers the As diffusion, and the As prevents the outdiffusion of the Ar. The interference occurs regardless of which ion is implanted first.

Second, when Si wafers uniformly doped with about 4 × 10¹⁹ As/cm³ are ion-implanted with log¹⁶ Ar/cm² at 130 keV, the As atoms stay uniformly distributed. When the sample is annealed at a temperature between 900 and 1100°C in a nitrogen ambient, however, double peaks for both Ar and As are observed by backscattering. The nonuniform distribution of As after the heat treatment of the uniformly distributed As is puzzling.

Finally, the As profile for an As capsule diffused on a silicon wafer is greatly altered when the wafer has been pre-implanted with Ar. Arsenic atoms tend to build up at the same depth as the Ar atoms.

Several other observations concerning Ar and As are equally puzzling. This paper discusses the observations and some plausible explanations.     

Crystalline to amorphous transformations in ion implanted GaP

Abstract

The amorphization process of GaP by ion implantation is studied. The samples of 〈111〉 oriented GaP were implanted at 130 K with various doses 5 × 10¹³-2 × 10¹⁶ cm^?2 of 150 keV N⁺ ions and with the doses of 6 × 10¹²-1.5 × 10¹⁵ cm^?2 of 150 keV Cd⁺ ions. Room temperature implantations were also performed to see the influence of temperature on defect production. Rutherford backscattering and channelling techniques were used to determine damage in crystals. The damage distributions calculated from the RBS spectra have been compared with the results of Monte-Carlo simulation of the defect creation.

The estimated threshold damage density appeared to be independent on ion mass and is equal 6.5 × 10²⁰ keV/cm³. It is suggested that amorphization of GaP is well explained on the basis of a homogenous model.