Carrier trapping and activation at short-period wurtzite/zinc-blende stacking sequences in polytypic InAs nanowires

We explore the effects of random and short-period crystal-phase intermixing in InAs nanowires (NW) on the carrier trapping and thermal activation behavior using correlated optical and electrical transport spectroscopy. The polytypic InAs NWs are grown by catalyst-free molecular beam epitaxy under different temperatures, resulting in different fractions of wurtzite (WZ) and zincblende (ZB) and variable short-period (∼1–4 nm) WZ/ZB stacking sequences. Temperature-dependent microphotoluminescence $\left(\mu \text{PL}\right)$ studies reveal that variations in the WZ/ZB stacking govern the emission energy and carrier confinement properties. The optical transition energies are modeled for a wide range of WZ/ZB stacking sequences using a Kronig-Penney type effective mass approximation, while comparison with experimental results suggests that polarization sheet charges on the order of ∼0.0016–0.08 C/m along the WZ/ZB interfaces need to be considered to best describe the data. The thermal activation characteristics of carriers trapped inside the short-period WZ/ZB structure are directly reproduced in the temperature-dependent carrier density evolution (4–300 K) probed by four-terminal (4T) NW-field effect transistor measurements. In particular, we find that activation of carriers in-between $\sim {10}^{16}-{10}^{17}\mathrm{c}{\mathrm{m}}^{-3}$ follows a two-step process, with activation at low temperature attributed to WZ/ZB traps and activation at high temperature being linked to surface states and electron accumulation at the InAs NW surface.

Figure 1

(a) Selective area diffraction (SAD) patterns and representative HR-TEM images of InAs NWs grown at three different temperatures of [(a)–(c)] $T^{\mathrm{G}}=430{}^{\circ }\mathrm{C}$, [(d)–(f)] 490 °C, and [(g)–(i)] 530 °C. WZ sensitive reflections $01\overline{1}2$ and $0\overline{1}12$ are marked by orange and green arrows, respectively, and are clearly visible in all SAD patterns—except for the InAs NW grown at the lowest $T^{\mathrm{G}}$ that exhibits continuous streaks signifying the heavily disordered stacking. The HR-TEM images (center and right hand columns) show the characteristic layer stacking, where clear, distinguishable WZ segments are marked in blue and stacking defects by arrows. The layer stacking is also labelled in “ABC” notation, facilitating identification of ZB-type and WZ-type stacking. The scale bar for all images is 5 nm.

Figure 2

Temperature-dependent PL spectra as measured for NW arrays grown at (a) $T^{\mathrm{G}}=430{}^{\circ }\mathrm{C}$, (b) 490 °C, and (c) 530 °C, at a fixed excitation power of 3.35 mW; (d) PL peak energy as a function of temperature, illustrating a consecutive blue/redshift with increasing temperature; the critical temperature $\left(T_{\mathrm{c}}\right)$ for the transition between blue- and redshifted PL increases as the growth temperature of the NWs is increased.

Figure 3

(a) Scanning electron micrograph and (b) schematic illustration of a typical InAs 4T-NWFET. The four terminals consist of individual Ni/Au contacts for source and drain (outer contacts) and inner probe contacts (H,L), while the ${\mathrm{n}}^{++}-\mathrm{Si}$ substrate serves as a global backgate. (c) Temperature-dependent transfer curves (conductivity σ vs $V_{\mathrm{g}})$ from an InAs NW grown at $T^{\mathrm{G}}=430{}^{\circ }\mathrm{C}$. (d) Typical transfer curve of the same device (recorded at 122 K) depicting the linear regression of the conductivity and extraction of threshold (pinch-off) voltage $V_{\mathrm{th}}$ from the intersection with the $V_{\mathrm{g}}$ axis.

Figure 4

(a) Temperature dependence of $n$-type carrier density of a representative InAs NW from the sample grown at $T^{\mathrm{G}}=430{}^{\circ }\mathrm{C}$. The inset shows an Arrhenius plot illustrating two specific regions for carrier activation, i.e., at low temperature $\left(E_{\mathrm{A}}^{\mathrm{I}}\sim 4.6\mathrm{meV}\right)$ and at high temperature $\left(E_{\mathrm{A}}^{\mathrm{II}}\sim 39\mathrm{meV}\right)$ with activation energies obtained form best fits to the data. Error bars are due to the uncertainties in extracting the threshold voltage $V_{\mathrm{th}}$ during pinch-off. (b) Comparison of the temperature-dependent carrier density from NWs grown at the three growth temperatures $T^{\mathrm{G}}=430{}^{\circ }\mathrm{C}$, 490 °C, and 530 °C, which exhibit different WZ/ZB stacking order and hence different thermal carrier activation characteristics. Solid lines represent guides to the eye. For ease of comparison, we plot here not the absolute carrier density but the relative change in carrier density for each sample.

Figure 5

Inverse temperature dependence of carrier density (Arrhenius plot) as obtained from two NWs with different diameter (68 nm, 151 nm) from sample A $\left(T^{\mathrm{G}}=430{}^{\circ }\mathrm{C}\right)$, illustrating the different activation energies in the high-temperature region. Best fits to the data (red lines) clearly demonstrate higher activation energy for the thinner NW.

Figure 6

(a) Idealized conduction-band (CB) and valence-band (VB) profiles, i.e., without polarization sheet charges, for the three different WZ/ZB polytype SL structures (samples A, B, C) using the averaged WZ/ZB-stacking lengths as defined in Table 1. The squared wave functions for electron and hole ground states are depicted with color codes (red/blue: highest/lowest intensity). (b) Calculated optical transition energies between CB and VB ground states mapped as a function of WZ segment length and two different ZB segment lengths (solid versus dashed curves) at 75 K. Each data set depicts also different polarization sheet charge densities, i.e., 0 (black), $1\times {10}^{13}$ (0.016 C/m, dark grey), $2\times {10}^{13}$ (0.032 C/m, medium grey), $3\times {10}^{13}$ (0.048 C/m, grey), and $5\times {10}^{13}\mathrm{c}{\mathrm{m}}^{-2}$ (0.08 C/m) (light grey). The data points (in color) indicate the experimentally observed PL peak energies at 75 K at low excitation power. The error bars delineate the standard deviation from the average WZ segment length obtained from statistical analysis as well as the typical minimum and maximum segment lengths observed in these samples. (c) CB and VB profiles and respective electron and hole ground states for the three different SL structures, assuming the presence of polarization sheet charges, that give closest agreement to the experimental PL peak energies; used values for the polarization sheet charges are 0.04, 0.014, and 0.063 C/m for $T^{\mathrm{G}}=430{}^{\circ }\mathrm{C}$, 490 °C, and 530 °C, respectively.