A common origin of photoplastic and electroplastic effects in ZnS

TL;DR

The study investigates why ZnS dislocations respond to optical and electric stimuli by carrier concentration, linking photoplastic and electroplastic effects to charged dislocations. Through imaging of misfit dislocations in ZnS/GaP epilayers and controlled Al-doping, it demonstrates that Zn-core dislocations are strongly inhibited by light or electron doping while S-core dislocations remain largely unaffected, validating a common carrier-driven mechanism. The work shows that carrier concentration engineering can substantially alter dislocation glide and strain relaxation, even reversing anisotropy in MD/TD glide, and provides a pathway to tailor dislocation content in compound semiconductors. Overall, the findings establish a unified framework for defect-mediated plasticity in semiconductors with potential implications for optoelectronic device performance and materials design.

Abstract

Dislocation motion--the atomic-scale mechanism of crystal plasticity--governs the strength and ductility of materials. In functional materials, external stimuli beyond mechanical stress can also affect dislocation glide. In the wide band gap semiconductor ZnS, optical illumination suppresses plasticity, whereas electric fields can enhance dislocation motion. Here, we show that the common underlying mechanism for these phenomena is the charged dislocations that respond to the changes in carrier concentration. Our prior theoretical work showed that locally charged dislocations in ZnS trap excess carriers, triggering core reconstructions that modify their mobility, with the positively charged Zn-rich core dislocations showing the most drastic change. Here, we validate this prediction experimentally by showing that either optical excitation or electronic doping selectively inhibits the glide of Zn-rich dislocations in epitaxially grown ZnS. First, imaging individual interface misfit dislocations under different optical excitation conditions shows that Zn-core glide is strongly reduced as optical power is increased, while the S-core dislocations show negligible sensitivity to light, marking the first, single misfit dislocation imaging of the photoplastic effect. Next, we show that a similar behavior is observed with direct electron (n-type) doping of ZnS epitaxial layers grown beyond the critical thickness. As the n-type dopant density is increased, the resulting Zn-core dislocation density is reduced by more than one order of magnitude, while the S-core density remains essentially unchanged, causing a sign reversal of the strain-anisotropy with n-type doping. These results demonstrate a common origin for the opto-electronic sensitivity of dislocations in ZnS and provide a pathway for the engineering of dislocation content in compound semiconductors.

Figures (7)

• Figure 1: Charge-carrier controlled plasticity in ZnS epitaxial layers. (Top row) Perspective view and (bottom row) corresponding plan-view schematic of dislocation glide in ZnS epitaxially grown on GaP (001). Dashed lines in the bottom panels illustrate how misfit dislocation (MD) segments along the epitaxial interface lengthen by glide (green arrows) of surface terminating threading dislocation (TD) segments, whose (a) charge states and resulting glide mobilities are altered by (b) light or (c) electronic impurity doping. Each MD segment and its TD pair define a dislocation half-loop with a common {111} plane and burgers vector. (a) The core structure of an MD and one of its TDs ($\theta =60$°) are identical, having the edge-type characteristic associated with an extra half-plane terminating at an ionic core, the composition of which is locked to the MD line direction $\vec{u}$ due to the $\bar{4}$ symmetry of the crystal structure. In ZnS, the $\vec{u}$ = [110] MDs and one of its TDs are S$^-$-core ($MD^-_{S}$ and $TD^-_{S}$) and for $\vec{u}$ = [$\bar{1}$10], they are Zn$^+$-core ($MD^+_{Zn}$ and $TD^+_{Zn}$). The other TD segment is screw-type ($\theta =180$°) with a neutral ZnS composition ($TD^0_{ZnS}$). (b) 365 nm optical excitation generates photo-electrons (-) and holes (+) that are trapped by the Zn$^+$ and S$^-$ dislocations, respectively. (c) Al-doped ZnS leads to conduction band electrons (-) that get trapped at Zn$^+$ MDs and TDs leaving bound Al$^+$ donor-ions (+) in the bulk ZnS. Electrons trapped in Zn$^+$-core trigger new core reconstructions ($MD^0_{Zn}$ and $TD^0_{Zn}$) decreasing the glide mobility.
• Figure 2: Single dislocation photoplastic effect from misfit dislocation (MD) imaging. Electron channeling contrast imaging (ECCI) of 20 nm undoped ZnS grown on GaP (001) (a) before and (b) after a glide cycle under 365 nm wavelength-irradiation at a power density, P = 0.07 mW/cm2. Insets in (a) and (b) show the longest MD in the image from which the change in length after a glide cycle is measured, $\Delta L^{max}_{MD}$, plotted in (c) as solid symbols (with solid lines) in reverse experimental sequence for both the [$\bar{1}$10] Zn-core MDs and [110] S-core MDs. Top x-axis plots the P condition of each glide cycle. Open symbols are averages of the two max $\Delta L_{MD}$. Boxes (25%-75%) represent the full distribution of $\Delta L_{MD}$ measurements. (d) The average in-plane strain before and after each glide cycle $\epsilon$ (right axis) is calculated from the ECCI data and plotted for the two directions, as well as the strain-factor $\epsilon/f$ . The top x-axis shows P during each cycle. Normalizing $\Delta$$L_{MD}^{max}$ data in (c) by the $\epsilon$ data in (d) yields the relative TD mobility $\mu$, which is plotted in (e) as a function of P for the two directions showing a strong reduction in $\mu_{Zn}$ and not $\mu_{S}$ with photoexcitation, see Fig. \ref{['fig:photoplastic']}(b). Lines are guides to the eye.
• Figure 3: ZnS growth morphology as a function of n-type Al-doping concentration ($N_{Al}$). (a) AFM surface topology of 30-nm thick ZnS epilayers grown by Molecular beam epitaxy (MBE) on GaP (001). (b) RMS roughness ($R_q$) variation with $N_{Al}$. Line is a guide to the eye.
• Figure 4: ZnS epitaxial strain-state as a function of n-type (Al) doping concentration ($N_{Al}$) measured by x-ray diffraction (XRD) reciprocal space mapping (RSM). (a) and (b) show the RSM at the (224) and ($\bar{2}24$) Bragg peaks, respectively, of a ZnS epitaxial layer with $N_{Al} = 4\times10^{17}$ cm$^{-3}$. Lines extending through the GaP Bragg peak represent various possible relaxation states of the ZnS layer. (c) and (d) show results of the same RSM scans for a ZnS sample with $N_{Al} = 4\times10^{19}$ cm$^{-3}$, showing a strong reduction in the in-plane strain $\epsilon$ along [110], with increased $N_{Al}$, but not along [$\bar{1}$10] indicative of an n-type doping suppression of Zn-core MDs (u=[$\bar{1}$10]), see Fig. \ref{['fig:photoplastic']}(c). (e) The corresponding MD densities $\rho_{MD}$ calculated from the RSM data as a function of $N_{Al}$ for Zn-core versus S-core MDs. Error bars are calculated by propagating the uncertainty in Gaussian peak fits of the RSM peak positions. Lines guide the eye.
• Figure 5: Impact of n-type (Al) doping on absorption spectra measured by spectroscopic ellipsometry. (a) Absorption spectra in ZnS epilayers grown with different Al doping concentrations (NAl). The three absorption regions (i)-(iii) described in the text, are labeled. Two absorption features in region (iii) are labeled according to the matching partial dislocation band gaps from Ref. genlik_origin_2025. (b) Urbach energy fit (lines) of the below band gap band tail absorption. (c) Tauc fit (lines) of the absorption spectra above band gap. (d) Bandgap (Eg) and Urbach energy (EU) as a function of NAl (lines are guides to the eye).