Deep-level transient spectroscopy (DLTS) is a high-resolution technique for characterizing deep-level defects in semiconductors, particularly silicon. It provides detailed information on defect energy levels, capture cross-sections, and concentrations, making it indispensable for identifying and quantifying electrically active traps that degrade carrier lifetime and material quality. In silicon, key defects include oxygen vacancies, transition metal impurities such as iron (Fe) and copper (Cu), and dislocation-related traps, all of which originate from crystal growth conditions, processing, or contamination.

Oxygen vacancies are among the most common intrinsic defects in silicon, particularly in Czochralski (CZ)-grown material where oxygen is incorporated from the quartz crucible. These vacancies form complexes such as the A-center (vacancy-oxygen, VO), which introduces an energy level at approximately Ec -0.17 eV. DLTS reveals this defect through its characteristic emission signature, with a capture cross-section around 10^-15 cm². The concentration of VO defects correlates with the cooling rate during crystal growth; slower cooling allows oxygen to diffuse and form larger aggregates, reducing electrically active VO centers. High concentrations of VO defects reduce minority carrier lifetime by acting as recombination centers, impacting device yield in applications requiring long diffusion lengths.

Transition metal impurities, particularly Fe and Cu, are detrimental to silicon performance due to their deep energy levels and high diffusivity. Iron introduces a well-studied defect level at Ec -0.39 eV (Fe-B pair in boron-doped silicon) with a capture cross-section of about 10^-14 cm². DLTS identifies Fe contamination through its distinct peak, and the concentration can be quantified down to parts-per-billion levels. Iron contamination often arises from furnace components, crucibles, or handling during wafer processing. Copper, another common contaminant, creates multiple energy levels, including a prominent one at Ev +0.1 eV. Unlike Fe, Cu tends to precipitate at room temperature, complicating DLTS analysis unless measurements are performed at low temperatures or with rapid quenching. Both metals degrade carrier lifetime by introducing Shockley-Read-Hall recombination centers, with Fe being particularly harmful due to its high recombination activity.

Dislocation-related traps are another class of defects detectable by DLTS, especially in multicrystalline or deformed silicon. Dislocations introduce bandgap states through dangling bonds and strain fields, with energy levels distributed across the forbidden gap. DLTS spectra of dislocated silicon show broad peaks corresponding to a range of activation energies, typically between Ec -0.1 eV and Ec -0.5 eV. The density of these traps depends on the dislocation density, which is influenced by growth parameters such as thermal gradient and cooling rate during solidification. High dislocation densities lead to severe lifetime degradation, limiting the efficiency of solar cells and other optoelectronic devices.

The DLTS measurement process involves filling traps by applying a forward bias pulse to a Schottky or p-n junction, then monitoring the capacitance transient as carriers emit from the traps at a fixed rate. By varying the temperature and rate window, an Arrhenius plot is constructed to extract the defect’s activation energy and capture cross-section. The amplitude of the transient provides the defect concentration, typically expressed in cm^-3. For example, a DLTS peak at 200 K with an emission rate of 100 s^-1 might correspond to the Fe-B pair, with its concentration calculated from the capacitance step height.

Crystal growth conditions strongly influence the type and density of deep-level defects. In CZ silicon, oxygen content is controlled by the crucible rotation rate and argon flow, with higher oxygen leading to more VO defects unless properly managed. Float-zone (FZ) silicon, being oxygen-free, avoids VO-related traps but is more susceptible to metal contamination due to the absence of oxygen gettering. Dislocation densities are minimized in single-crystal growth by optimizing the seed crystal pull rate and thermal profile, whereas multicrystalline silicon exhibits higher dislocation densities due to uncontrolled grain boundaries.

Quantitative DLTS analysis has shown that Fe concentrations above 10^12 cm^-3 can reduce carrier lifetime by an order of magnitude, while VO defects at 10^14 cm^-3 have a measurable but less severe impact. Dislocation-related traps are more detrimental at densities exceeding 10^5 cm^-2, where they dominate recombination. These findings underscore the importance of defect control during crystal growth and wafer processing to ensure high-quality material for semiconductor applications.

In summary, DLTS is a powerful tool for identifying and quantifying deep-level defects in silicon, linking their origins to growth conditions and contamination. Oxygen vacancies, metal impurities, and dislocation-related traps each have distinct DLTS signatures, enabling precise characterization of their impact on carrier lifetime. By correlating DLTS data with crystal growth parameters, manufacturers can optimize processes to minimize these defects and improve material performance.