摘要:

Two‐dimensional doping profiles can be determined from multiple one‐dimensional secondary ion mass spectroscopy (SIMS) profiles using computed tomography techniques. The chemical nature of SIMS enables the measurement of both as‐implanted and annealed profiles with this technique. Mechanical lapping was done of multiple samples to expose different faces of the substrates followed by one‐dimensional SIMS measurements. Measurements have been done with a Perkin Elmer Model 6300 SIMS and a Cameca IMS‐3f SIMS. Refinements have been made in the numerical alignment of the separate one‐dimensional SIMS data sets, which lead to more accurate reconstructions of the two‐dimensional profiles.

ISSN:0734-211X    

DOI:10.1116/1.587148

出版商:American Vacuum Society    

年代:1994

数据来源: AIP

摘要:

Collision cascade mixing in the sputter depth profiling process broadens the measured width of abrupt dopant structures by displacing the target atoms, predominantly through collisions with the disturbed matrix atoms prior to sputter removal. Sputtering theory predicts a profile broadening dependence on the nuclear stopping power of the projectile in the target matrix, but this is not supported by experimental data. In this experiment, abrupt boron profiles were measured in a matrix of silicon or germanium, at a range of energies from 2.5 to 10.5 keV. The higher mass matrix gives access to high stopping powers, particularly for Xe+projectiles. As with the projectile mass experiments, this work does not support the predicted change in decay length with the range of the projectile. The data indicate the decay length dependence on projectile energy alone is due to dopant displacements by primary recoils as well as the collision cascade mixing.  

ISSN:0734-211X    

DOI:10.1116/1.587149

出版商:American Vacuum Society    

年代:1994

数据来源: AIP

摘要:

Codeposited delta doping distributions of B and Sb in Si, with areal densities on the order of 1×1013atoms/cm2, were used to determine the resolution functions for depth profiling of these dopants under 2 and 5 keV O2bombardment at impact angles θ between 0° (normal incidence) and 60°. The resolution functions are described by the first and second moments and by the characteristic inverse slopes, which define the exponential rise and decay of the signal at the profile’s leading and trailing edges, respectively. The decay lengths λd(0°) agree very well with numbers reported recently. The up‐slope parameters λu(0°) are the same for B and Sb (0.65 and 1.0 nm at 2 and 5 keV, respectively), but significantly smaller than the previously published data, which showed differences for B, Sb, and Ge. This discrepancy could be due to differences in roughness of the as‐grown samples. Moreover, the high doping densities used in the previous work might have caused profile distortions during growth. Variations of θ between 0° and 60° have only a very small effect on λu(B, Sb) and on λd(B); e.g., at 5 keV λd(B)=3.3 nm at 0° and 60° with a broad maximum in between, where λd=3.7 nm. By contrast, λd(Sb) is quite large in the range 0°≤θ≤25°, 5.6–6.0 nm, then decreases rapidly with increasing impact angle to fall below λd(B) as θ exceeds 40°. Significant differences between B and Sb are also seen in the first moments 〈z〉 of the resolution function. At 5 keV, e.g., 〈z〉Sb−〈z〉Bvaries from 3.5 nm (0°) to −1.5 nm (60°). The square root of the second moments can be approximated as σ=λu+λd. The difference in the angular dependence of the resolution parameters for B and Sb is thought to be due to differences in the redistribution of the two dopants within the silicon dioxide layer (θ≤25°) and the silicon suboxide (θ≳25°) generated by oxygen bombardment.

ISSN:0734-211X    

DOI:10.1116/1.587150

出版商:American Vacuum Society    

年代:1994

数据来源: AIP

摘要:

Sputter‐initiated resonance ionization spectroscopy (SIRIS) is an emerging analytical technique for quantitative and sensitive measurements of impurities and ultra‐shallow doping profiles in semiconductors. SIRIS has almost all of the advantages of secondary ion mass spectroscopy (SIMS), while making significant improvements in the following SIMS shortcomings: efficiency, matrix dependence, isobaric and molecular interferences, sensitivity, dynamic range, and quantitation accuracy. In this article, the SIRIS technique is described and shows typical SIRIS depth profiles of boron implanted into silicon with dynamic ranges greater than ∼2×106as well as SIRIS aluminum depth profiles of layered GaAs/AlGaAs/GaAs samples with depth resolution of up to ∼2 nm at 0.5 keV Ar+primary ion energy. SIRIS’s dynamic range, sensitivity, depth resolution and its limitations, and quantitation accuracy will be discussed in detail.

ISSN:0734-211X    

DOI:10.1116/1.587151

出版商:American Vacuum Society    

年代:1994

数据来源: AIP

摘要:

Understanding the limitations of depth profiling with ion sputtering is essential for accurate measurements of atomically abrupt interfaces and ultra‐shallow doping profiles. The effects of cascade mixing, sputtering statistics, ion‐induced roughness, the inhomogeneity of ion beams, and sample rotation on the depth resolution of Si δ‐doped, AlAs, and InAs monolayers in GaAs and an AlGaAs(5 nm)/GaAs(5 nm) superlattice were investigated. Atomic force microscopy (AFM) investigation of the ion‐induced surface ripple formation on a GaAs substrate sputtered with 3 keV O+2at angle of incidence θ=40° showed that ripples form rapidly below 200 nm depth. AFM measured root mean square roughness of Si δ‐doped GaAs sputtered with 2 keV O+2was 0.8 and 2.6 nm with and without sample rotation showing that ripples play a dominant role in depth resolution degradation at shallow depth under these conditions of bombardment. Sample rotation yielded the lowest full width at half‐maximum, 4.1 nm for a Si δ layer at 120 nm depth corresponding to a depth resolution ΔZ=3.5 nm. Use of AFM enabled determination of the atomic mixing ΔZmand sputtering statistics ΔZsscomponents of depth resolution to be identified directly for the first time. These components were 3.1 and 1.5 nm, respectively.

ISSN:0734-211X    

DOI:10.1116/1.587152

出版商:American Vacuum Society    

年代:1994

数据来源: AIP

摘要:

Conceptually the use of spreading resistance measurements for two‐dimensional profiling has been introduced two years ago. In the mean time our understanding has improved towards a better interpretation of the experimental results. A simulation program has been realized which allows to simulate the spreading resistance scans along the specially beveled surface. An ‘‘insitu’’ calibration procedure for measuring with high accuracy, the probe size and separation has been developed and a new formula, with less experimental involved parameters, for calculating the lateral spread has been derived. The technique has been applied to a study of the lateral diffusion of B as a function of annealing time. The effect of different masking material (polycrystalline silicon, thick oxide) has been studied. Other recent applications of this technique include the determination of the lateral straggling of channeling implants and the lateral diffusion of transition metals.

ISSN:0734-211X    

DOI:10.1116/1.587154

出版商:American Vacuum Society    

年代:1994

数据来源: AIP

摘要:

In this article, a system is described for the analysis of spreading resistance data which includes carrier spilling through the solution of Poisson’s equation. The Poisson equation solution requires boundary conditions for the charge density on the beveled surface, which was modeled by considering the presence of surface states with levels within the band gap. Considering surface states allows the deviation from ideality of bothn‐ andp‐type calibration curves to be accounted for consistently without empirical contact resistance or variable radius corrections. In addition, the effects of finite probe penetration and change in energy gap due to probe pressure were considered to account for the reduction in peak resistance near junctions observed experimentally as compared to theoretical calculations. The software (SERAPES) can be used both to calculate simulated resistance profiles from doping profiles for process model characterization, as well as to calculate doping profiles directly from spreading resistance measurements.

ISSN:0734-211X    

DOI:10.1116/1.587155

出版商:American Vacuum Society    

年代:1994

数据来源: AIP

摘要:

The calculation of the electrically active dopant profile from spreading resistance measurements requires not only the calculation of the sampling volume correction factor but also a deconvolution of the carrier spilling effects. A completely automatic Poisson‐based correction package has been developed which performs this calculation–deconvolution within 1 min on a 486‐based microcomputer and supports all the currently known contact models. The capabilities of the package will be demonstrated on an ultra‐shallow opposite type implant, a narrow‐base transistor and a submicron source‐drain implant. Furthermore, a series of requirements is being proposed for the accurate correction of carrier spilling effects. The currently available Poisson contact models will be reviewed with respect to these requirements starting from secondary ion mass spectrometry measurements and an improved contact model involving the parallel circuit of an elastically deformed stress layer and a plastically deformed metallic phase layer will be proposed.

ISSN:0734-211X    

DOI:10.1116/1.587156

出版商:American Vacuum Society    

年代:1994

数据来源: AIP

摘要:

Despite the rapidly increasing capabilities of Poisson deconvolution schemes for spreading resistance (SR) measurements, carrier spilling effects will remain a serious problem for the interpretation of SR profiles for structures involving significant forward spilling such as epilayers. In order to remedy this situation a spreading impedance probe is being developed allowing to influence the internal spilling by an external dc field, while performing an impedance measurement at a small ac bias. Three‐dimensional calibration surfaces representing resistance versus resistivity and dc bias determine the probe characteristics. Results for apwell and ap⋅p+epilayer will be discussed.

ISSN:0734-211X    

DOI:10.1116/1.587157

出版商:American Vacuum Society    

年代:1994

数据来源: AIP

摘要:

Spreading resistance has become a standard in semiconductor resistivity determination. The method is based on metal–semiconductor point contact measurements. The resulting spreading resistance is generally described byRs=ρ/4a, wherein ρ refers to the bulk resistivity andato the radius of the contact area. This relation is only valid in the ideal case where the contact area can be considered as a homogenous circular contact. New contact models have introduced the concept of multiple microcontacts in spreading resistance probe technique (SRP) measurements. The physical nature of these contacts has now been studied by an atomic force microscope (AFM). The AFM images show indentations of multiple point contacts penetrating 10–100 nm into the silicon depending on the applied force and the probe characteristics. Time‐dependent measurements at constant force revealed a gradual increase in penetration depth and in contact area. An AFM with conducting tip has been developed to study the behavior of a single contact point.I–Vcurves for such a metallic tip are compared withI–Vcurves for standard SRP probes.

ISSN:0734-211X    

DOI:10.1116/1.587158

出版商:American Vacuum Society    

年代:1994

数据来源: AIP