The intellectual core of the work is an analogy that is as simple as it is powerful. In conventional deep-level transient spectroscopy, temperature is swept to probe the energy distribution of trap states: as thermal energy changes, carriers are emitted from traps at rates that reveal their depth in the band gap. Earlier work by Gulyamov and Sharibaev had shown that as temperature falls, the continuous density-of-states function gradually sharpens, with the derivative of the resistivity response tending toward a Dirac delta-function and the spectrum resolving into discrete peaks. The new study asks a provocative question: if thermal energy can act as a scanning parameter, why not deformation energy? When a crystal is elastically compressed, its band structure shifts, defect potentials are distorted, and localized levels move. Pressure, in other words, does not merely perturb the spectrum; it reshapes it.

To formalize this idea, the authors introduce an effective deformation energy, defined as Ed = κX, where X is the applied uniaxial pressure and κ is an effective deformation-energy coefficient expressed in electron-volts per pascal. This coefficient is deliberately not treated as a universal constant; it depends on the elastic response of the crystal, its crystallographic orientation, and the surrounding defect environment. Each localized level in the band gap is then assumed to shift linearly with deformation energy, Ei(X) = Ei(0) + αiEd, where αi is a sensitivity coefficient whose sign indicates whether the level moves toward the conduction band or toward the valence band under compression. The authors are careful to stress that these relations are first-order effective approximations, valid over a modeled pressure interval from zero to 6 × 10^8 pascals, and that a full tensorial deformation-potential treatment lies beyond the scope of the model.

The mathematical machinery of the reconstruction rests on a pressure-dependent release probability. The probability of liberating an electron from a state of energy Ei is written as ρ(Ei, Ed) = 1 − exp[−(t0/τ(Ei))F(Ed)], where t0 is the characteristic measurement time, τ(Ei) is the relaxation time of the level, and F(Ed) is a monotonic pressure-activation function, exemplified in its simplest phenomenological form by exp(Ed/Ea). Differentiating this probability with respect to deformation energy yields a localized sensitivity kernel. In the idealized limit where the kernel becomes infinitely narrow, it approaches a Dirac delta-function, and the measured derivative of the released charge with respect to pressure becomes directly proportional to the local density of states at the corresponding energy. In realistic, finite-temperature conditions, the spectrum must instead be reconstructed numerically, using a broadening kernel such as a Gaussian, with amplitudes representing the concentrations of individual defect centers.

What makes this framework potentially transformative for device engineering is the way it connects spectral shifts to measurable electrical quantities. The model predicts that for thermal donor centers, associated with oxygen, pressure pushes levels upward toward the conduction band, enhancing electron generation; the resistivity of p-type silicon containing boron and thermal donors falls from its zero-pressure value to roughly 0.74 of that value across the modeled pressure range at 77 kelvin, accompanied by rising carrier concentration and mobility. For samples containing manganese-related centers, the opposite occurs: levels shift downward toward the valence band, localization intensifies, and the relative resistivity climbs to approximately 1.55 times its initial value. These opposing trends, encoded in the signs of the respective αi coefficients, mean that a simple resistivity-versus-pressure curve carries fingerprints of which defect families are active in the crystal.

Perhaps the most intriguing result concerns nickel-contaminated n-type silicon. The modeled relative resistivity for these samples exhibits a striking two-step increase, rising from 1 to 9 and then to 15 as pressure grows, while the corresponding normalized carrier concentration drops in two stages from 1 to 0.11 and then to 0.07. The authors capture this behavior with a double-sigmoidal function featuring two characteristic activation pressures and two amplitudes. The two-step response is consistent with the presence of at least two distinct groups of nickel-related centers, each with its own effective pressure sensitivity. Crucially, however, the model cannot uniquely determine the microscopic structure of these centers; independent spectroscopy or atomistic calculations would be required for a definitive assignment. The authors explicitly note that their manuscript contains no cobalt-specific calculations and cannot establish any cobalt-versus-nickel mechanism, a caution that reflects the study’s deliberately restrained scope.

Indeed, intellectual honesty is a defining feature of the paper. The authors present the work as an analytical proof of concept rather than a validated new spectroscopy. No independent DLTS, Hall-effect, or capacitance measurements are included; no first-principles calculations back the effective parameters; and no replicate raw measurements or uncertainty distributions are available to support confidence intervals. The representative value of the deformation-energy coefficient, κ = 1.0 × 10^−10 eV/Pa, is used as a plausible model parameter, not asserted as universal. The delta-function limit is invoked only as an idealized zero-width approximation of the sensitivity kernel, not as an exact mathematical identity. Figures showing the narrowing of the kernel and the temperature evolution of the spectrum are explicitly schematic rather than fits to data. This transparency distinguishes the work from overclaimed spectroscopy papers and lays out a clear roadmap for what experimental validation must follow.

The broader context makes the proposal timely. Decades of literature have established that mechanical stress is far more than a nuisance in silicon technology. Early stress-DLTS experiments showed that uniaxial stress splits and linearly shifts trap peaks, allowing researchers to determine defect symmetry and pressure sensitivity. Subsequent studies demonstrated that hydrostatic pressure accelerates the formation of oxygen thermal donors and alters oxygen clustering, explaining observed decreases in resistivity and increases in mobility in doped samples. Other work has shown that stress depassivates interface centers, distorts silicon-oxygen bonds, and activates otherwise silent traps at the Si-SiO2 interface, effectively creating a new pressure-dependent density of surface states. Piezoresistance research, particularly in silicon-on-insulator structures, has further confirmed that the combination of defects and stress radically restructures charge transport. The new model unifies these threads by treating pressure as a spectrum-forming factor rather than a mere external parameter.

If the tenso-DLTS concept survives experimental calibration, its applications could be significant. Non-destructive, pressure-sensitive defect profiling would be attractive for quality control in MEMS fabrication, where devices are permanently subjected to mechanical stress during operation, and for assessing strain-engineered transistors in which deliberately applied stress is used to boost carrier mobility. Because the reconstruction requires only electrical measurements as a function of applied force, it could in principle be implemented with tooling far simpler than conventional deep-level transient spectroscopy setups. The framework may also prove useful for Si-SiO2 interface characterization, where stress-activated surface states degrade MOS device reliability, provided the reconstruction kernel and effective parameters are calibrated for the specific material and loading configuration.

For now, the study stands as a carefully bounded theoretical advance: within its stated assumptions and modeled pressure range, uniaxial deformation acts as an additional, mechanically tunable parameter that modifies the effective spectrum of localized electronic states in silicon. It extends the classical temperature-based picture by adding a mechanical energy scale, κX, alongside the familiar thermal scale kT, and it demonstrates that opposite-sign level shifts for different defect families translate directly into opposite resistivity trends that experiments can target. Whether pressure can ultimately claim a place beside temperature as a fundamental spectroscopic coordinate for defect physics will depend on the calibration and validation studies the authors themselves call for. But the core message is already compelling: squeeze a silicon crystal, and it will tell you what is hiding inside it.

Subject of Research: Analytical reconstruction of the pressure-dependent density of localized electronic states in silicon under uniaxial mechanical stress

Article Title: Analytical reconstruction of the density-of-states spectrum in silicon under uniaxial pressure

Article References: Rakhmanov, M., Tursunov, I., Mamatkarimov, O., Sharibaev, N., & Sharipbaev, S. (2026). Analytical reconstruction of the density-of-states spectrum in silicon under uniaxial pressure. Results in Engineering, 32, Article 113273. https://doi.org/10.1016/j.rineng.2026.113273

Image Credits: AI Generated

DOI: 10.1016/j.rineng.2026.113273

Keywords: silicon, density of states, uniaxial pressure, deep-level transient spectroscopy, defect states, deformation energy, thermal donors, piezoresistance, semiconductor physics, tenso-DLTS, nickel contamination, Si-SiO2 interface