Correlated Chemical and Electrically Active Dopant Analysis in

www.acsnano.org

Correlated Chemical and Electrically Active Dopant Analysis in Catalyst-Free Si-Doped InAs Nanowires Jonathan Becker,†,∥ Megan O. Hill,‡,∥ Max Sonner,† Julian Treu,† Markus Döblinger,§ Alexander Hirler,† Hubert Riedl,† Jonathan J. Finley,† Lincoln Lauhon,*,‡ and Gregor Koblmüller*,† †

Walter Schottky Institut, Physik Department, and Center of Nanotechnology and Nanomaterials, Technische Universität München, Garching 85748, Germany ‡ Department of Materials Science and Engineering, Northwestern University, Evanston, Illinois 60208, United States § Department of Chemistry, Ludwig-Maximilians-Universität München, Munich 81377, Germany S Supporting Information *

ABSTRACT: Direct correlations between dopant incorporation, distribution, and their electrical activity in semiconductor nanowires (NW) are difficult to access and require a combination of advanced nanometrology methods. Here, we present a comprehensive investigation of the chemical and electrically active dopant concentrations in n-type Si-doped InAs NW grown by catalyst-free molecular beam epitaxy using various complementary techniques. N-type carrier concentrations are determined by Seebeck effect measurements and four-terminal NW field-effect transistor characterization and compared with the Si dopant distribution analyzed by local electrode atom probe tomography. With increased dopant supply, a distinct saturation of the free carrier concentration is observed in the mid-1018 cm−3 range. This behavior coincides with the incorporated Si dopant concentrations in the bulk part of the NW, suggesting the absence of compensation effects. Importantly, excess Si dopants with very high concentrations (>1020 cm−3) segregate at the NW sidewall surfaces, which confirms recent first-principles calculations and results in modifications of the surface electronic properties that are sensitively probed by field-effect measurements. These findings are expected to be relevant also for doping studies of other noncatalytic III−V NW systems. KEYWORDS: III−V nanowires, catalyst-free growth, doping, atom probe tomography, Seebeck effect, electrical transport demonstrated, for example, for VLS-grown Si NWs1,7 and GaAs NWs.8 In VLS-grown NWs it has been also shown that dopant incorporation and carrier activation are directly affected by the peculiar crystal phase polytypism commonly observed in III−V semiconductor NWs. Both experimental and theoretical studies revealed that Si-doping of wurtzite (WZ) phase GaAs NWs results in p-type conductivity as opposed to n-type conductivity in Si-doped bulk zincblende (ZB) GaAs.9,10 Doping of VLStype III−V NWs has also profound effects on the crystal phase stability and defect structure itself. For example, doping of III− V NWs can alter the crystal phase from WZ to ZB11 and vice

T

he realization of high-performance electronic and optoelectronic devices from semiconductor nanowires (NW) makes accurate control of charge carrier conductivity by intentional doping indispensable. Understanding doping behavior and establishing links between dopant distribution and carrier activation in free-standing NWs is, however, still a challenging task. The three-dimensional, high aspect ratio structure of NWs possesses axial and radial growth facets with different crystal orientations, surface reconstructions, and hence highly anisotropic growth dynamics, which are expected to influence dopant incorporation.1−3 In many cases, these characteristics are complicated by fundamentally different growth modes of the participating facets. For example, in the predominantly studied catalyst-assisted vapor− liquid−solid (VLS) growth of NWs, dopant incorporation can proceed either via the liquid catalyst interface,4,5 the solid sidewall facets,6 or by a combination of both facets, as recently © 2018 American Chemical Society

Received: November 18, 2017 Accepted: January 31, 2018 Published: January 31, 2018 1603

DOI: 10.1021/acsnano.7b08197 ACS Nano 2018, 12, 1603−1610

Article

Cite This: ACS Nano 2018, 12, 1603−1610

Article

ACS Nano versa,12 allowing the formation of complex twin-defect NW superlattice structures.11 In addition, the presence of specific planar defects in the NW may also trap dopant impurities via the catalyst interface,13 modifying their electrical activity. Compared to VLS-grown NWs, much less attention has been paid to dopant incorporation studies of catalyst-free, that is, non-VLS type NWs. This is because the noncatalytic spontaneous growth mode is limited mostly to NWs from the group-III nitride3,14,15 and III−V arsenide families grown under selective area epitaxial processes.16−19 While most doping studies of these classes of NWs focused on the effects of dopants on morphology, microstructure, and basic transport properties,17,20−23 correlated studies between dopant concentration and distribution and carrier activation has remained quite limited3 due to challenges in probing dopants with atomic resolution in NWs. In this work, we demonstrate a direct correlation between chemical and active dopant concentrations in n-type Si-doped InAs NWs grown by molecular beam epitaxy (MBE) in a completely catalyst-free, that is, vapor−solid (VS)-type growth mode. Essentially, we employ local electrode atom probe tomography (APT) to map the Si dopant density and distribution in single NWs and compare data with active carrier concentrations probed by Seebeck effect and fourterminal NW-field effect transistor (4T-NW-FET) measurements.

Figure 1. SEM micrographs of (a) undoped InAs NWs and (b) Sidoped InAs NWs (Si: 13 A) as grown on Si (111), exhibiting similar morphologies. The TEM micrograph in (c) and corresponding SAD pattern (d) recorded from a Si-doped InAs NW (Si: 13 A) illustrate a WZ-dominated crystal phase with many stacking along the [0001] NW growth direction.

RESULTS AND DISCUSSION The growth of the InAs NW samples was carried out in a solid source Gen-II MBE system equipped with standard effusion cells for group-III elements (Al, Ga, In) and a valve cracker cell supplying molecular arsenic (As4). 2 inch [111]-oriented Si wafers were used as growth substrates, which were covered with a ∼2−5 nm-thick and wet chemically etched SiO2 layer. The etched SiO2 layer provides a mask for nucleating NWs in a selective-area-like manner from random nm-sized pinholes.24 Typical examples of the noncatalytic vertical growth of InAs NWs on this growth substrate are shown by the scanning electron micrographs (SEM) in Figure 1 for cases of (a) undoped NWs and (b) highly Si-doped NWs for comparison. No metallic In droplets are present on the NW tips, confirming the catalyst-free VS growth mode of these NWs.24 The depicted NW samples are part of a growth series performed under identical growth conditions but variable Si-dopant fluxes (see Supporting Information for respective SEM images). Specifically, we used an In flux of 0.36 Å/s and an As4 flux of 14 Å /s (beam equivalent pressure of 2.6 × 10−5 mbar), corresponding to a V/III ratio of ∼40.25 Growths lasted for 90 min at a growth temperature of 500 °C. Atomic Si dopants were supplied continuously during growth, except for the very first 5 min to prevent any Si-induced effects on the nucleation. The Si dopants were provided by a solid dopant cell via thermal sublimation induced by resistive heating of the cell. Besides an undoped InAs NW reference, three Si-doped NW samples were grown with different cell heating currents (11 A, 12 A, 13 A), where the highest current corresponds to a Si dopant flux of ∼1.6 × 1012 cm−2 s−1.8 Note that a linear increase in heating current means an exponentially increasing Si dopant flux, as confirmed by secondary ion mass spectrometry (SIMS) on planar reference samples. Specifically, the three respective heating currents correspond to Si dopant concentrations of 2.1 × 1018 cm−3 (11 A), 6 × 10 18 cm−3 (12 A), and 1.4 × 1019 cm−3 (13 A) for thin (100)-oriented films grown at a growth

rate of 1 μm/h. To this end, the range between ∼1018 and 1019 cm−3 is sufficiently large to cover the most interesting region in Si dopant concentrations of III-arsenide-based materials. Within this range many effects were observed in the past for thin films, for example, the onset of carrier compensation and dopant-induced changes in the lattice strain influencing the occupancy of group-III versus group-V atomic sites26 or, for instance, changes in the incorporation efficiency of group-V growth species.27 Interestingly, the aspect ratio and morphology of the Sidoped InAs NWs remains largely unchanged with respect to undoped InAs NWs (see Figure 1), independent of the supplied Si dopant flux. On average, the 90 min long growths resulted in ∼2.3−3.4 μm long NWs with diameters in between ∼60 and 110 nm for each of the four investigated NW samples, as evaluated from a statistical analysis of at least 20 NWs/ sample using SEM imaging. Any changes in growth dynamics due to the presence of Si doping are therefore smaller than the statistical dispersion of NW lengths and diameters within a given sample. Since this dispersion is relatively large in NWs nucleated from random pinholes, we also examined the growth kinetics in high-uniformity periodical NW arrays on lithographically prepatterned SiO2/Si(111) substrates as will be reported elsewhere. Although the length and diameter dispersion is substantially reduced in this control experiment, no marked change in the aspect ratios and morphologies was observed upon Si doping. This observation differs markedly from previous findings of both metal organic chemical vapor deposition (MOCVD)-17,22 and MBE-grown23 catalyst-free InAs NWs. Wirths et al.22 and Dimakis et al.23 observed a continuous decrease (increase) in NW length (diameter) with Si doping, irrespective of the different surface chemistries of the employed dopants (Si2H6 in MOCVD versus atomic Si in MBE). In particular, in MBE-based studies which use atomic Si dopants with similar dopant fluxes as in our work, the NW 1604


Article

ACS Nano length is dramatically decreased upon Si doping.23 We attribute the absence of such observation in our Si-doped NWs to the different growth conditions (V/III ratio, growth temperature) employed. Furthermore, Ghoneim et al.17 found that for MOCVD growth a change in aspect ratio of Si-doped InAs NWs (along with a deterioration of morphology) occurs only for very high Si dopant fluxes, corresponding to active dopant concentrations >5 × 1019 cm−3. Transmission electron microscopy (TEM) was performed on individual NWs transferred onto carbon-coated grids to identify the crystal phase and defect structure of the undoped and Sidoped InAs NWs. Based on an analysis of multiple NWs per sample, we found that the NWs crystallize preferentially in the WZ phase and exhibit a high density of stacking defects, similar to our previous findings of undoped InAs NWs.28,29 A representative TEM micrograph along with the respective selected area diffraction (SAD) pattern are shown in Figure 1c,d for NWs grown under the highest Si dopant flux (13 A). From the SAD pattern, distinct WZ-sensitive reflections (i.e., 011̅2 and 01̅12) are clearly visible, while the streaks between individual reflections indicate the high stacking disorder along the [0001] growth direction.29 Since no qualitative differences are found between undoped and Si-doped InAs NWs, the microstructure of noncatalytic InAs NWs is insensitive to Si doping under the given growth conditions. This is in good agreement with other Si-doping studies of InAs NWs reported in the literature.22,23 To identify the effect of the different Si dopant fluxes on the active carrier concentration, Seebeck effect measurements were conducted at room temperature. Thermoelectric device structures were fabricated from single NWs dispersed onto 200 nm SiO2/n++-Si substrates using electron beam lithography, metal evaporation, and lift-off. As shown in Figure 2a, the devices consist of an InAs NW with two Ni (nickel) contacts on each end that serve as probes for temperature (resistive thermometers) and Seebeck voltage. To induce a Seebeck voltage across a single NW, a temperature gradient was established by resistively heating one end of the NW using a meandering metal strip line consisting of five parallel 22 μmlong, 1 μm-wide, and 100 nm-thick strips of Ni.30 This generates a temperature gradient of a few Kelvin that was measured in a four-terminal sensing geometry. Hereby, the first resistive thermometer was biased via contacts 1 and 2, and the voltage drop was measured across contacts 3 and 4. For the second thermometer, contacts 5 and 6 were biased, and 7 and 8 were used for sensing. The resulting Seebeck voltage was measured between one of the blue-colored (cold) and one of the red-colored (hot) contacts, for example, between contacts 8 and 3. Measurements were performed on chip carriers connected to a 24 pin breakout, which was purged with N2 gas to ensure low humidity at a constant temperature close to room temperature. Two dc constant current supplies (Yokogawa 7651) were used to connect to the resistive heater and thermometer, respectively, while either a differential voltage amplifier (Femto DLPVA-100FD) in conjunction with a Keithley 2000 multimeter to probe the Seebeck voltage or an Agilent Model 34420A two channel nanovoltmeter to probe the voltage drops over the thermometers was connected to the device. Figure 2b shows the Seebeck voltage as a function of temperature difference (0−2 K) for representative devices fabricated from undoped and Si-doped InAs NWs. Note that the plotted Seebeck voltage data are already corrected by

Figure 2. (a) Color-coded SEM image of a typical single NW device for Seebeck effect measurements containing lithographically defined resistive heater and thermometers at both NW ends. (b) Seebeck voltage versus applied temperature gradient at room temperature as obtained from three Si-doped InAs NWs with different dopant fluxes (11 A, 12 A, 13 A) in comparison with an undoped InAs NW. The Seebeck coefficients are extracted from linear fits to the data. (c) Calculated Seebeck coefficient of n-type InAs as a function of carrier concentration; the blue curve is numerically calculated, while the red and yellow curves are analytical approximations for the nondegenerate and degenerate case, respectively. Experimental data points as obtained from (b) are also plotted (black symbols), with corresponding carrier concentrations ranging from ∼5 × 1016 cm−3 (undoped InAs NWs) to 2−5 × 1018 cm−3 (doped InAs NWs).

subtraction of the respective Seebeck voltage induced by the Ni contacts (Seebeck coefficient SNi = −19 μV/K).30 The Seebeck voltage V varies linearly with the measured temperature difference ΔT, with the slope giving directly the Seebeck coefficient of the InAs NWs. As expected, all investigated NWs exhibit a negative slope indicative of n-type conduction. While the Seebeck coefficient of the undoped InAs NW is as high as 1605


Article

ACS Nano

Table 1. Summarized Data of the Undoped and Three Si-Doped InAs NW Samples Grown under Three Different Si Dopant Currents with Corresponding Nominal Si Dopant Concentrations Calibrated on Planar Films (at 1 μm/h growth rate)a Si flux (A)

nom. [Si] (cm−3)

bulk [Si] (APT) (cm−3)

0 11 12 13

− 2.1 × 1018 6.0 × 1018 1.4 × 1019

− 1.5 × 1018, 5 × 1018 1.8 × 1018, 2.5 × 1018 4.3 × 1018

|S| (μV/K) 299−352 41−62 44−104 38−49

n (Seebeck) (cm−3) 4.0−8.7 2.5−4.6 1.0−4.1 3.4−5.3

× × × ×

16

10 1018 1018 1018

n (NW-FET) (cm−3) 1.0−1.2 1.0−4.0 6.8−10 0.7−3.6

× × × ×

17

10 1018 1018 1019

μFE (cm2/(V s)) 1205 (±70) 542 (±114) 406 (±75) 243 (±133)

The APT data give the measured Si dopant concentration in the bulk region of the NW (accuracy of ±5 × 1017 cm−3) for 1−2 NWs measured per samples; the typical range of Seebeck coefficients S (absolute values) and n-type carrier concentrations from Seebeck effect and NW-FET measurements at room temperature are listed for comparison (measured on 2−6 NWs/sample); μ is the corresponding field effect mobility extracted from the NW-FET transconductance at room temperature. a

Figure 3. (a) Color-coded SEM image and (b) schematic illustration of a typical 4T InAs NW-FET, consisting of four ohmic contacts (outer contacts for source and drain and inner probe contacts). The n++-Si substrate serves as a global backgate. (c) Transfer characteristics (conductivity versus gate voltage) as obtained from an intrinsically undoped InAs NW-FET at different measurement temperatures using a source−drain voltage of VSD = 0.5 mV. The threshold voltage Vth at pinch-off is used to extract the carrier concentration. (d) Inverse temperature dependence of the carrier concentration of four representative InAs NWs under different doping conditions.

−352 μV/K, its magnitude decreases for the Si-doped InAs NWs, yielding S = −62 μV/K for the lowest doped sample (11 A), S = −46 μV/K for the medium doped sample (12 A), and S = −38 μV/K for the highest doped sample (13 A) (see also Table 1). The relatively small change in Seebeck coefficient among the three Si-doped InAs NW samples suggests that the free carrier concentration does not change significantly with increasing Si flux beyond 11 A. To verify this, we calculated the dependence of the Seebeck coefficient on n-type carrier concentration by solving the electronic part of the Boltzmann transport equation according to ref 31 (see also Supporting Information). Using a lattice temperature of 300 K, an electron effective mass of m*/m0 = 0.023, a unity scattering parameter,31 and assuming bulk-like electronic properties (no 1D electron density of states (DOS)), we derived a numerical solution as plotted on Figure 2c by the blue curve. The assumption of bulk-like electronic DOS is fully valid here, since the diameters of our investigated NWs are well above the dimensions for 1D quantum confinement.28 We further note that our numerically calculated data agrees well

with experimental data obtained for bulk InAs at roomtemperature.32 For comparison, we also plotted analytical solutions for the degenerate and nondegenerate conduction cases, which are approximations based on the assumption of parabolic band dispersion.31 In order to get a direct estimate of the carrier concentration of the investigated InAs NWs, we plotted the as-measured Seebeck coefficients onto the numerically calculated Seebeck coefficient data. From this we find that the undoped InAs NW has a carrier concentration of ∼5 × 1016 cm−3, whereas the Si-doped InAs NWs exhibit carrier concentrations of ∼2−5 × 1018 cm−3. A statistical analysis of the range of carrier concentrations obtained from several NWs per sample is further presented in Table 1 and shown in Figure 5. Two important messages can be derived from the extracted carrier concentration dependency. First, our undoped NWs reveal some of the lowest free carrier concentrations ever measured in intrinsic InAs NWs,33 which is likely a result of the pure MBE growth environment. MOCVD-grown InAs NWs have carrier concentrations typically 1−2 orders of magnitude larger.28,34,35 Second, the free carrier concentration for the Si1606


Article

ACS Nano doped InAs NWs saturates quickly in the low- to mid-1018 cm−3 range, despite the increasing Si dopant flux. This behavior suggests either strong carrier compensation effects or a significant dopant incorporation limit under high Si doping conditions. To further confirm this trend, we have also conducted NWFET measurements on single NW devices to allow an independent estimation of the carrier concentration. Backgated NW-FETs were fabricated using a similar device geometry as for the Seebeck effect measurements. The NWFET devices consisted of individual InAs NWs drop-cast directly on 200 nm SiO2/n+-Si (i.e., global back gate) and contacted by four Ni/Au ohmic contacts to realize 4-terminal (4T) transport measurements. An SEM image along with a schematic illustration of a typical 4T-NW-FET device is shown in Figure 3a,b. The 4T-device geometry allows us to separate intrinsic NW conductivity from NW contact resistance. This was achieved by applying a dc voltage (VSD) to an outer contact and measuring the respective current between the outer source and drain contacts (1,4), whereas the voltage drop due to the intrinsic NW conductivity was measured via the inner probe contacts (2,3). Temperature-dependent 4T measurements were performed between 6 K and 290 K in a He-4 flow cryostat where the NW-FET samples were placed on custom-made chip carriers and connected by wire bonding. Representative T-dependent transfer characteristics of the intrinsic NW conductivity σ are shown in Figure 3c for the case of an undoped InAs NW using a source−drain voltage of VSD = 0.5 mV. As expected, intrinsic n-type carrier conduction is confirmed by the increase in conductivity with increasing positive VG. At low temperature (e.g., 8 K) we also observe step-like pinch-off characteristics, which may be due to the depopulation of individual subbands as previously noted, for example, InAs- and GaAs-based NW-FETs.36,37 Note that the pinch-off and the respective threshold voltage Vth shift to increasingly negative VG with rising temperature. From the threshold voltage Vth, which is derived from a linear regression of σ = gm × (VG − Vth) (gm, transconductance), we can directly extract the carrier concentration n using Cg × Vth/(e × l × A), where Cg is the gate capacitance, which is estimated based on the widely used metal plate capacitor model of ref 38, and l and A are the length and cross-section area of the NW channel as measured by scanning electron microscopy. Figure 3d compares the calculated carrier concentration as a function of inverse temperature for the undoped, intrinsic NW and typical NWs from the three Si-doped samples analyzed in the same way. Additional data obtained from several other NWs for each sample are further illustrated in Figure 5 and Table 1, listing also values for the field effect mobilities that were extracted from the transconductance, that is, μ = gm × l2/(Cg × VSD). For all investigated NWs, the carrier concentration is fairly constant over the measured temperature range indicating the overall degenerate nature of both the undoped and Sidoped InAs NWs. Only a marginal increase in carrier concentration may be noticed for temperatures above ∼60 K, best illustrated for the undoped InAs NW. This behavior is likely due to a weak activation process from traps caused by either stacking defects or surface-related defect states. Most importantly, as in the Seebeck effect measurements, we observe a similar trend in the increase and saturation of carrier concentration with Si doping. However, compared to the Seebeck effect data, the carrier concentrations extracted from the NW-FET measurements are on average higher (up to factor

of 2−3 depending on sample) (see also Figure 5). We suggest two reasons for this observation: First, the metal plate capacitor model used to derive the gate capacitance38 slightly overestimates the capacitance and, thus, the charge carrier concentration by ∼10−20%. Second, and more significantly, field-effect measurements are very sensitive to spatially fixed surface/interface charges that occupy the surface and interface states between the NW surface and metal gate. These states cause Fermi level pinning at the NW surface39and trap additional charges induced by the applied gate voltage.40 As a result, only a fraction of the induced charge contributes free carriers to the NW channel, which in turn leads to an overestimation of the free carrier concentration. Moreover, we observe that the deviation in carrier concentration between NW-FET and Seebeck effect data is more significant for increased Si dopant fluxes, with estimated field-effect concentrations exceeding 1019 cm−3 for the highest doped sample. This suggests that excess Si dopants may accumulate in the near surface region of the NW channel and modify the Fermi level pinning, which is sensitively probed by the NW-FET measurements. This hypothesis is further examined below by APT measurements of the Si dopant distribution across individual NWs. Moreover, we found that along with the evolution of carrier concentration upon doping, the field-effect mobility decreases significantly, that is, from ∼1200 cm2/(V s) (undoped NWs) to ∼540 cm2/(V s) and below (Si-doped NWs) at room temperature (see also Supporting Information). This behavior mimics closely the mobility versus n-type doping dependence of bulk InAs41 as observed also in previous investigations of MOCVD-grown Sidoped InAs NWs.17 To further explain the characteristic saturation behavior in carrier concentration with increasing Si doping, we performed APT on selected Si-doped InAs NWs to directly compare chemical with electrically active dopant concentrations. APT specimens were prepared by picking and Pt welding individual NWs onto tungsten probe tips using a nanomanipulator in vacuum. Details of the specimen preparation as well as specifics of the APT experiments can be found elsewhere42 (also see Supporting Information). A representative map of the Si dopant distribution across a single NW is illustrated in Figure 4a for the highest doped sample (Si: 13 A). Similar dopant distribution maps were also recorded from NWs with lower Si doping and are shown in the Supporting Information. Clearly, we observe an overall homogeneous distribution of Si dopants inside the NW, but a distinctly higher Si concentration near the NW surface. Note that the contrast facilitated by the enriched Si layer reproduces closely the hexagonal NW geometry, however, the enrichment is visible only along one side of the NW due to the limited field of view in APT for specimens that are not perfectly vertically oriented on the probe tip. Due to the limited field of view, it is therefore helpful to probe also the presence of oxygen (O) impurities arising from the inherent surface oxide layer on the {2-110} InAs NW sidewalls40 and thereby delineate the exact position of the NW surface (Figure 4b). We note that the O content is reported for comparison, but the concentration should not be taken as quantitive at the very edge of the reconstruction. We observe direct coincidence between the positions of the surface oxide layer and the enriched Si dopant layer, which is also illustrated by the proximity histogram plotted in Figure 4c. The proximity histogram shows the Si dopant and O impurity distribution together with the measured molar fraction of InAs when 1607


Article

ACS Nano

Figure 5. Statistical distribution of the concentrations of bulk Si dopants (APT data) and n-type charge carriers (Seebeck and NWFET data) as a function of nominal Si concentration (Si dopant current) at room temperature. The blue and black dashed curves are guides to the eye describing the evolution of Seebeck and NWFET data, respectively. The data points shown here are also summarized in Table 1.

rich surface layer may increase the downward band bending at the NW surface, similar to observations in, for example, sulfurpassivated InAs NW,43 thus leading to larger trap state densities which obscure the free carrier concentration. A fully quantitative correlation between the excess Si accumulation in the surface oxide and active carriers within the surface layer is, however, difficult due to the presence of other donor-type surface state defects39 and the fact that the Si enrichment is not homogeneous across the surface region. To resolve this, systematic X-ray photoelectron spectroscopy (XPS) experiments need to be performed to directly correlate the modified core level spectra arising from the Si enrichment and the induced changes in respective surface band bending. To finally discuss the origin of the excess Si accumulation on the WZ-{2-110} InAs NW surfaces and the observed dopant incorporation limit in the NWs, a few kinetic and thermodynamic considerations can be undertaken. First of all, the underlying noncatalytic VS growth mode implies that the majority of Si dopant incorporation occurs via the sidewall facets, since on slow growing facets, dopant incorporation rates are generally enhanced compared to fast growing facets.44 This behavior assumes a unity sticking coefficient of Si and a very low Si surface diffusion, irrespective of the given facet, just as in planar III−V semiconductor films.45,46 Notably, in planar III− V-based films, precipitation of Si at the growth surface has also been found under high Si dopant fluxes, limiting the incorporation of Si to concentrations below the mid-1018 cm−3 range.26 It has been speculated that such Si-rich phases reduce the As coverage at the growth surface, inducing changes in the dynamic surface structure, that is, surface reconstructions and surface strain, which may inhibit effective dopant incorporation.47 Such saturated dopant-rich phases at the surface of growing layers have also been found in other III−V systems where the dopant incorporation becomes independent of the dopant arrival rate.48 To further examine the link between segregated phases and associated modifications of the surface structure in the present InAs NWs requires additional studies, such as scanning tunneling microscopy along with theoretical modeling. Finally, we wish to emphasize that dopant

Figure 4. APT reconstruction of a Si-doped InAs NW (13 A) shown as projected 2D compositional maps of (a) silicon (Si) and (b) oxygen (O). The full nanowire cross section is not visible in the reconstruction, but the location of the InAs-native oxide interface is identified by the increase in O. Enrichment of Si is found at the NW surface. Scale bars are 10 nm. (c) Corresponding 1D composition profile moving from the nanowire center to the surface along one of the major {2-110} directions (white box in (a)). The red dashed line indicates the approximate location of the interface with the native oxide.

moving from center to surface of the NW along one of the {2110} directions. The average Si dopant concentration in the bulk part of the NW was determined to be ∼4.3 × 1018 cm−3, while it increases by at least 2 orders of magnitude at the surface. A similar behavior was also observed for NWs measured from samples with lower Si dopant cell fluxes (see Supporting Information). The fact that the enriched Si layer is confined within the surface oxide suggests that only minimal carriers would be activated from this highly doped region and that the free carrier concentration is determined by the Si dopant concentration in the bulk part of the NW. Indeed, this correlation can be directly seen when comparing the Si dopant concentrations in the bulk of the NW (APT data) with the activated bulk carrier concentrations extracted from Seebeck effect measurements as shown in Figure 5. The results clearly show that the active carrier concentration matches well with the incorporated dopant concentration in the center of the NW within the experimental error for each of the three investigated Si-doped InAs NW samples. This means that carrier compensation effects are negligible for the range of Si dopant concentrations explored here. Figure 5 also plots the respective carrier concentrations estimated from NW-FET measurements. The fact that these are higher than the incorporated dopant concentration within the NW (especially for high dopant fluxes) confirms that the NW-FET measurements are impacted by surface/interface effects and the presence of the enriched Si layer. We speculate that the Si1608


Article

ACS Nano

DMR-1611341. M.O.H. acknowledges support of the NSF GRFP.

segregation to the lateral sidewall surfaces of NWs seems to be also explained by recent first-principles calculations of the segregation energies of various different dopants in III−V based NWs.49 In the framework of these calculations the segregation effects were also predicted for Si-doped InAs NWs, where dopant atoms prefer to substitute for surface atoms with extra dangling bonds or may be trapped at subsurface sites close to atoms with additional dangling bonds.

REFERENCES (1) Perea, D. E.; Hemesath, E. R.; Schwalbach, E. J.; Lensch-Falk, J. L.; Voorhees, P. W.; Lauhon, L. J. Direct Measurement of Dopant Distribution in an Individual Vapour-liquid-solid Nanowire. Nat. Nanotechnol. 2009, 4, 315. (2) Wallentin, J.; Borgström, M. T. Doping of Semiconductor Nanowires. J. Mater. Res. 2011, 26, 2142. (3) Fang, Z.; Robin, E.; Rozas-Jimenez, E.; Cros, A.; Donatini, F.; Mollard, N.; Pernot, J.; Daudin, B. Silicon Donor Incorporation in GaN Nanowires. Nano Lett. 2015, 15, 6794. (4) Yang, C.; Zhong, Z.; Lieber, C. M. Encoding Electronic Properties by Synthesis of Axail Modulation-doped Silicon Nanowires. Science 2005, 310, 1304. (5) Patsha, A.; Amirthapandian, S.; Pandian, R.; Bera, S.; Bhattacharya, A.; Dhara, S. Direct Evidence of Mg Incorporation Pathway in Vapor-Liquid-Solid Grown P-type Nonpolar GaN Nanowires. J. Phys. Chem. C 2014, 118, 24165. (6) Casadei, A.; Krogstrup, P.; Heiss, M.; Röhr, J. A.; Colombo, C.; Ruelle, T.; Upadhyah, S.; Sorensen, C. B.; Nygard, J.; Fontcuberta i Morral, A. Doping Incorporation Paths in Catalyst-free Be-doped GaAs Nanowires. Appl. Phys. Lett. 2013, 102, 013117. (7) Amit, I.; Givan, U.; Connell, J. G.; Paul, D. F.; Hammond, J. S.; Lauhon, L. J.; Rosenwaks, Y. Spatially Resolved Correlation of Active and Total Doping Concentrations in VLS Grown Nanowires. Nano Lett. 2013, 13, 2598. (8) Dufouleur, J.; Colombo, C.; Garma, T.; Ketterer, B.; Uccelli, E.; Nicotra, M.; Fontcuberta i Morral, A. P-Doping Mechanisms in Catalyst-free Gallium Arsenide Nanowires. Nano Lett. 2010, 10, 1734. (9) Piccin, M.; Bais, B.; Grillo, V.; Jabeen, F.; De Franceschi, S.; Carlino, E.; Lazzarino, M.; Romanato, F.; Businaro, L.; Rubini, S.; Martelli, F.; Franciosi, A. Growth by Molecular Beam Epitaxy and Electrical Characterization of GaAs Nanowires. Phys. E 2007, 37, 134. (10) Ghaderi, N.; Peressi, M.; Binggeli, N.; Akbarzadeh, H. Structural Properties and Energetics of Intrinsic and Si-doped GaAs Nanowires: First-principles Pseudopotential Calculations. Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 81, 155311. (11) Algra, R. E.; Verheijen, M. A.; Borgström, M. T.; Feiner, L.-F.; Immink, G.; van Enckevort, W. J. P.; Vlieg, E.; Bakkers, E. P. A. M. Twinning Superlattices in Indium Phosphide Nanowires. Nature 2008, 456, 369. (12) Wallentin, J.; Mergenthaler, K.; Ek, M.; Wallenberg, L. R.; Samuelson, L.; Deppert, K.; Pistol, M.-E.; Borgström, M. T. Probing the Wurtzite Conduction Band Structure Using State Filling in Highly Doped InP Nanowires. Nano Lett. 2011, 11, 2286. (13) Hemesath, E. R.; Schreiber, D. K.; Gulsoy, E. B.; Kisielowski, C. F.; Petford-Long, A. K.; Voorhees, P. W.; Lauhon, L. J. Catalyst Incorporation at Defects during Nanowire Growth. Nano Lett. 2012, 12, 167. (14) Hersee, S. D.; Sun, X.; Wang, X. The Controlled Growth of GaN Nanowires. Nano Lett. 2006, 6, 1808. (15) Fernandez-Garrido, S.; Kong, X.; Gotschke, T.; Calarco, R.; Geelhaar, L.; Trampert, A.; Brandt, O. Spontaneous Nucleation and Growth of GaN Nanowires: The Fundamental Role of Crystal Polarity. Nano Lett. 2012, 12, 6119. (16) Noborisaka, J.; Motohisa, J.; Fukui, T. Catalyst-free Growth of GaAs Nanowires by Selective Area Metalorganic Vapor-Phase Epitaxy. Appl. Phys. Lett. 2005, 86, 213102. (17) Ghoneim, H.; Mensch, P.; Schmid, H.; Bessire, C. D.; Rhyner, R.; Schenk, A.; Rettner, C.; Karg, S.; Moselund, K. E.; Riel, H.; Björk, M. T. In Situ Doping of Catalyst-free InAs Nanowires. Nanotechnology 2012, 23, 505708. (18) Hertenberger, S.; Rudolph, D.; Bichler, M.; Finley, J. J.; Abstreiter, G.; Koblmüller, G. Growth Kinetics in Position-controlled and Catalyst-free InAs Nanowire Arrays on Si(111) grown by Selective Area Molecular Beam Epitaxy. J. Appl. Phys. 2010, 108, 114316.

CONCLUSIONS In summary, we demonstrated a direct correlation between the chemical and electrically active dopant concentrations in Sidoped InAs nanowires grown in a completely catalyst-free growth mode. By employing atom probe tomography, Seebeck effect measurements, and NW-FET characterization, we found a characteristic dopant incorporation limit at concentrations around mid-1018 cm−3. Up to these dopant concentrations, the free carrier concentration follows nicely the Si dopant densities incorporated in the bulk of the NW, confirming that compensation effects are negligible. For comparison, undoped InAs NWs exhibit free carrier concentrations of ∼5 × 1016−1 × 1017 cm−3 which are among the lowest ever reported values for n-type InAs NWs. The results reveal also that under the employed growth and Si doping conditions, a thin layer of segregated Si forms on the NW sidewall surfaces with concentrations of >1020 cm−3, which confirms recent firstprinciples calculations of doped III−V NWs. ASSOCIATED CONTENT S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsnano.7b08197. Details on SEM images of the entire Si-doped InAs NW series, calculations of n-type carrier concentration dependence of the Seebeck coefficient, field-effect mobility extracted from NW-FET measurements, and additional atom probe tomography data and analysis (PDF)

AUTHOR INFORMATION Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. ORCID

Megan O. Hill: 0000-0002-7663-7986 Lincoln Lauhon: 0000-0001-6046-3304 Gregor Koblmüller: 0000-0002-7228-0158 Author Contributions ∥

These authors contributed equally.

Notes

The authors declare no competing financial interest.

ACKNOWLEDGMENTS The authors acknowledge financial support by the German Excellence Initiative via the program “Nanosystems Initiative Munich (NIM)”, the German Research Foundation (DFG) via DFG-projects KO-4005/2-1 and KO-4005/5-1, the Technical University of Munich (TUM), Institute for Advanced Study (IAS-TUM), and the International Graduate School for Science and Engineering (IGSSE-TUM) via Project 9.04-NW-Electronics. M.O.H. and L.J.L. further acknowledge support by NSF 1609


Article

ACS Nano

Quantum Transport and Sub-Band Structure of Modulation-Doped GaAs/AlAs Core-Superlattice Nanowires. Nano Lett. 2017, 17, 4886. (38) Wunnicke, O. Gate Capacitance of Back-gated Nanowire FieldEffect Transistors. Appl. Phys. Lett. 2006, 89, 083102. (39) Speckbacher, M.; Treu, J.; Whittles, T. J.; Linhart, W. M.; Xu, X.; Saller, K.; Dhanak, V. R.; Abstreiter, G.; Finley, J. J.; Veal, T. D.; Koblmüller, G. Direct Measurements of Fermi Level Pinning at the Surface of Intrinsically N-type InGaAs Nanowires. Nano Lett. 2016, 16, 5135. (40) Blömers, C.; Grap, T.; Lepsa, M. I.; Moers, J.; Trellenkamp, S.; Grützmacher, D.; Lüth, H.; SchäPers, T. Hall Effect Measurements on InAs Nanowires. Appl. Phys. Lett. 2012, 101, 152106. (41) Karataev, V. V.; Milvidskii, M. G.; Rytova, N. S.; Fistul, V. I. Sov. Phys. Semicond. 1977, 1, 1718. (42) Jeon, N.; Loitsch, B.; Morkötter, S.; Abstreiter, G.; Finley, J.; Krenner, H. J.; Koblmüller, G.; Lauhon, L. J. Alloy Fluctuations Act as Quantum Dot-like Emitters in GaAs-AlGaAs Core-Shell Nanowires. ACS Nano 2015, 9, 8335. (43) Petrovykh, D. Y.; Yang, M. J.; Whitman, L. J. Chemical and Electronic Properties of Sulfur-Passivated InAs Surfaces. Surf. Sci. 2003, 523, 231. (44) Sumner, J.; Oliver, R. A.; Kappers, M. J.; Humphreys, C. J. Scanning Capacitance Microscopy Studies of Unintentional Doping in Epitaxial Lateral Overgrowth GaN. J. Appl. Phys. 2009, 106, 104503. (45) Cho, A. Y.; Hayashi, I. P-N Junction Formation during Molecular-Beam Epitaxy of Ge-Doped GaAs. Metall. Trans. A 1971, 2, 777. (46) Schubert, E. F. Doping in III-V Semiconductors, Cambridge Studies in Semiconductor Physics and Microelectronics Engineering; Cambridge University Press: Cambridge, 1993. (47) Wassermeier, M.; Behrend, J.; Däweritz, L.; Ploog, K. Reconstruction of the GaAs (001) Surface induced by Submonolayer Si Deposition. Phys. Rev. B: Condens. Matter Mater. Phys. 1995, 52, R2269. (48) Guha, S.; Bojarczuk, N. A.; Cardone, F. Mg in GaN: Incorporation of a Volatile Species at High Temperatures during Molecular Beam Epitaxy. Appl. Phys. Lett. 1997, 71, 1685. (49) Galicka, M.; Buczko, R.; Kacman, P. Segregation of Impurities in GaAs and InAs Nanowires. J. Phys. Chem. C 2013, 117, 20361.

(19) Rudolph, D.; Hertenberger, S.; Bolte, S.; Paosangthong, W.; Spirkoska, D.; Döblinger, M.; Bichler, M.; Finley, J. J.; Abstreiter, G.; Koblmüller, G. Direct Observation of a Noncatalytic Growth Regime for GaAs Nanowires. Nano Lett. 2011, 11, 3848. (20) Furtmayr, F.; Vielemeyer, M.; Stutzmann, M.; Arbiol, J.; Estrade, S.; Peiro, G.; Morante, J. R.; Eickhoff, M. Nucleation and Growth of GaN Nanorods on Si(111) Surfaces by Plasma-assisted Molecular Beam Epitaxy: The Influence of Si- and Mg- Doping. J. J. Appl. Phys. 2008, 104, 034309. (21) Li, S.; Waag, A. GaN based Nanorods for Solid State Lighting. J. Appl. Phys. 2012, 111, 071101. (22) Wirths, S.; Weis, K.; Sladek, K.; Volk, C.; Alagha, S.; Weirich, T. E.; von der Ahe, M.; Hardtdegen, H.; Lüth, H.; Grützmacher, D.; SchäPers, T. Effect of Si-doping on InAs Nanowire Transport and Morphology. J. Appl. Phys. 2011, 110, 053709. (23) Dimakis, E.; Ramsteiner, M.; Huang, C.-N.; Trampert, A.; Davydok, A.; Biermanns, A.; Pietsch, U.; Riechert, H.; Geelhaar, L. In Situ Doping of Catalyst-free InAs Nanowires with Si: Growth, Polytypism, and Local Vibrational Modes of Si. Appl. Phys. Lett. 2013, 103, 143121. (24) Hertenberger, S.; Rudolph, D.; Bolte, S.; Döblinger, M.; Bichler, M.; Spirkoska, D.; Finley, J. J.; Abstreiter, G.; Koblmüller, G. Absence of Vapor-Liquid-Solid Growth during Molecular Beam Epitaxy of Selfinduced InAs Nanowires on Si. Appl. Phys. Lett. 2011, 98, 123114. (25) Hertenberger, S.; Rudolph, D.; Becker, J.; Bichler, M.; Finley, J. J.; Abstreiter, G.; Koblmüller, G. Rate Limiting Mechanisms in Hightemperature Growth of Catalyst-free InAs Nanowires with Large Thermal Stability. Nanotechnology 2012, 23, 235602. (26) Chai, Y.; Chow, R.; Wood, C. The Effect of Growth Conditions on Si Incorporation in Molecular Beam Epitaxial GaAs. Appl. Phys. Lett. 1981, 39, 800. (27) O’Hagan, S.; Missous, M. Effect of As4/Ga Flux Ratio on Electrical and Optical Properties of Low-temperature GaAs grown by Molecular Beam Epitaxy. J. Appl. Phys. 1994, 75, 7835. (28) Koblmüller, G.; Vizbaras, K.; Hertenberger, S.; Bolte, S.; Rudolph, D.; Becker, J.; Döblinger, M.; Amann, M.-C.; Finley, J. J.; Abstreiter, G. Diameter Dependent Optical Emission Properties of InAs Nanowires grown on Si. Appl. Phys. Lett. 2012, 101, 053103. (29) Morkötter, S.; Funk, S.; Liang, M.; Döblinger, M.; Hertenberger, S.; Treu, J.; Rudolph, D.; Yadav, A.; Becker, J.; Bichler, M.; Scarpa, G.; Lugli, P.; Zardo, I.; Finley, J. J.; Abstreiter, G.; Koblmüller, G. Role of Microstructure on Optical Properties in High-uniformity InGaAs Nanowire Arrays: Evidence of a Wider Wurtzite Band Gap. Phys. Rev. B: Condens. Matter Mater. Phys. 2013, 87, 205303. (30) Schmidt, V.; Mensch, P. F. J.; Karg, S. F.; Gotsmann, B.; Das Kanungo, P.; Schmid, H.; Riel, H. Using the Seebeck Coefficient to Determine Charge Carrier Concentration, Mobility, and Relaxation Time in InAs Nanowires. Appl. Phys. Lett. 2014, 104, 012113. (31) Nolas, G. S.; Sharp, J.; Goldsmid, H. J. Thermoelectrics − Basic Principles and New Materials Developments; Springer: Berlin, 2001. (32) Rhode, D. L. Electron Transport in InSb, InAs, and InP. Phys. Rev. B 1971, 3, 3287. (33) Blömers, C.; Rieger, T.; Grap, T.; Raux, M.; Lepsa, M. I.; Lüth, H.; Grützmacher, D.; SchäPers, T. Gate-induced Transition between Metal-type and Thermally Activated Transport in Self-catalyzed MBEgrown InAs Nanowires. Nanotechnology 2013, 24, 325201. (34) Thelander, C.; Caroff, P.; Plissard, S.; Dey, A. W.; Dick, K. A. Effects of Crystal Phase Mixing on the Electrical Properties of InAs Nanowires. Nano Lett. 2011, 11, 2424. (35) Dayeh, S. A.; Susac, D.; Kavanagh, K. L.; Yu, E. T.; Wang, D. Structural and Room-Temperature Transport Properties of Zinc Blende and Wurtzite InAs Nanowires. Adv. Funct. Mater. 2009, 19, 2102. (36) Chuang, S.; Gao, Q.; Kapadia, R.; Ford, A. C.; Guo, J.; Javey, A. Ballistic InAs Nanowire Transistors. Nano Lett. 2013, 13, 555. (37) Irber, D. M.; Seidl, J.; Carrad, D. J.; Becker, J.; Jeon, N.; Loitsch, B.; Winnerl, J.; Matich, S.; Döblinger, M.; Tang, Y.; Morkötter, S.; Abstreiter, G.; Finley, J. J.; Grayson, M.; Lauhon, L. J.; Koblmüller, G. 1610


Correlated Chemical and Electrically Active Dopant Analysis in

Recommend Documents