Spatially Resolved Correlation of Active and Total Doping

Subscriber access provided by University of Sydney Library

Communication

Spatially resolved correlation of active and total doping concentrations in VLS grown nanowires Iddo Amit, Uri Givan, Justin G. Connell, Dennis F Paul, John Hammond, Lincoln J. Lauhon, and Yossi Rosenwaks Nano Lett., Just Accepted Manuscript • DOI: 10.1021/nl4007062 • Publication Date (Web): 13 May 2013 Downloaded from http://pubs.acs.org on May 19, 2013

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Nano Letters is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Spatially resolved correlation of active and total doping concentrations in VLS grown nanowires

Iddo Amit1†, Uri Givan2,3**,†, Justin G. Connell2, Dennis F. Paul4, John S. Hammond4, Lincoln J. Lauhon2 and Yossi Rosenwaks1,*

1

Department of Physical Electronics, School of Electrical Engineering, Tel-Aviv University, Ramat-Aviv 69978, Israel.

2

Department of Materials Science and Engineering, Northwestern University, Evanston, IL, USA. 3

Max Plank Institute of Microstructure Physics, Weinberg 2, 06120, Halle, Germany. 4

Physical Electronics Inc., 18725 Lake Drive East, Chanhassen, MN, USA, 55317 †

These authors contributed equally to this work *

**

Corresponding author: [email protected]

Corresponding author: [email protected]

ACS Paragon Plus Environment

1

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 27

ABSTRACT

Controlling axial and radial dopant profiles in nanowires is of utmost importance for NW-based devices, as the formation of tightly controlled electrical junctions is crucial for optimization of device performance. Recently, inhomogeneous dopant profiles have been observed in vaporliquid-solid grown nanowires, but the underlying mechanisms

that produce these

inhomogeneities have not been completely characterized. In this work, P doping profiles of axially modulation-doped Si nanowires were studied using nanoprobe scanning Auger microscopy and Kelvin probe force microscopy in order to distinguish between vapor-liquidsolid doping and the vapor-solid doping. We find that both mechanisms result in radially inhomogeneous doping, specifically, a lightly doped core surrounded by a heavily doped shell structure. Careful design of dopant modulation enables the contributions of the two mechanisms to be distinguished, revealing a surprisingly strong reservoir effect that significantly broadens the axial doping junctions.

KEYWORDS: Nanowires; VLS; Doping; nano-probe Scanning Auger microscopy; Kelvin probe force microscopy


2

Page 3 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

Vapor-liquid-solid (VLS) grown nanowires (NWs) constitute a family of nanostructured materials that enable the exploration of nanoscale phenomena1 and the development of novel devices2-7. The performance of such devices depends in many instances on the ability to achieve tight control over the density and spatial distribution of active impurity atoms within the NW during its growth and throughout the device fabrication process8,9. Specifically, determination of intentional doping profiles (both axial and radial), probing of unintentional dopant distributions, and characterizing electrical junction regions are key features in chemical and electrical characterization of a NW-based device9,10. The presently incomplete understanding of the growth and doping mechanisms at the nanoscale regime poses fundamental challenges for the fabrication of well-defined structures with controlled atomic composition1,11,12. For example, it is not always clear whether dopant incorporation is dominated by the VLS9,13-15 process through the catalyst, by a vapor-solid (VS) process through the wire surface, or by a combination of both16-19. The VLS technique, first described by Wagner and Ellis in 196413, is the most commonly employed method of bottom-up semiconductor NW fabrication of all compositions (groups II-VI, III-V and IV), enabling versatile fabrication of various structures as homojunctions, heterojunctions and superlattices9,2022

. Control of doping employing the catalytic VLS growth mechanism is, therefore, of utmost

importance for future nanoscale applications. In many NW systems the majority carrier type is controlled by intentional doping, which is commonly achieved by adding a gas-phase doping precursor to the growth system22-25. For example, the addition of a phosphine (PH3) precursor to the growth system produces an n-type silicon NW (SiNW), while the addition of diborane (B2H6) or trimethylborane (TMB) will produce a p-type SiNW9 ,26-30. In 2009, Schmid et al. showed, using four-probe measurements,


3

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

that the maximum attainable doping concentration using a PH3 precursor is 1.5×1020 cm-3

Page 4 of 27

31

.

However, phosphorous impurities tend to accumulate at the surface, forming a thin layer of highly doped silicon32,33. This tendency, together with the pentavalent nature of P, causes a termination of Si (001) surface dangling bonds, thus decreasing the growth rate of the NWs and also of the fraction of electrically active dopants due to dielectric mismatch resulting in an increase of their ionization energy8,17. Hence, an additional challenge is to distinguish between chemical and active dopant concentrations. In 2009, Koren et al. showed, employing KPFM, that in-situ P doping of Si results in a nonuniform distribution of electrically active dopants along the growth axis19, which has been attributed to the exposure of the SiNW surface to PH3 throughout the growth9,17,19,34-36. Similar results were obtained by Perea et al. employing atom probe tomography (ATP) to study P incorporation into GeNWs37. The VS mechanism, also referred to as surface deposition, generates a dopant-rich area at the base of the NW with a gradual decrease in dopant concentration towards the wire tip. This structure is a result of the fact that the wire base is exposed to the precursor atmosphere for longer time than the tip. Moreover, since the material deposited at the surface is richer in dopants as compared to the material deposited through the catalyst for hydride growth chemistries, in-situ doping usually results in a pseudo-core-shell structure, with a few nanometer-thick, heavily doped shell covering a lightly doped core18,37,38. Recently, Connell et al. used atom probe tomography to show that for both B-doped SiNWs and P-doped GeNWs, the dopants incorporated through the liquid catalyst are not uniformly distributed, with up to 100-fold enhancements in concentration near the VLS tri-junction39. Here we demonstrate the effects of the two different dopant incorporation mechanisms on both chemical and active dopant profiles considering both unintentional radial inhomogeneities and


4

Page 5 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

intentional axial junctions. The chemical composition and electronic properties of modulation doped wires and the junctions thus formed are correlated by combining nano-probe scanning Auger microscopy (np-SAM), current-voltage measurements, and Kelvin probe force microscopy (KPFM) measurements of segmentally P doped SiNWs (Fig. 1(a))40,41. The jointly active VLS and VS incorporation mechanisms lead to notable differences in properties between an i-n junction (i.e. a junction where the i region is grown first) and an n-i junction (where the intrinsic region is grown after the doped region). Spatially resolved electrical characterization provides evidence of significant dopant incorporation into a nominally intrinsic segment through the so called "reservoir effect"17,42-47 which has been previously demonstrated for NWs heterostructures and uniformly grown NWs. The combined physical and electronic characterization presented here provides new insights into how to engineer the properties of semiconductor nanowires and dopant junctions. The segmented SiNWs produced for this study were grown in a hot-wall low-pressure chemical vapor deposition (LPCVD) setup. 80 nm Au nano-particles (AuNPs) were used as catalysts, high-purity SiH4 as the precursor gas, He as the carrier gas, and PH3 (1000 ppm in H2 carrier) served as the n-doping precursor gas. Growth conditions were chosen to be comparable to those conventionally used for both intrinsic and highly P doped SiNWs growth17,18,38 while enabling taper-free growth for a relatively long growth period. The NWs used in this study are nominally 80 nm in diameter and ~20 µm long with no observable tapering (see SEM image in figure 1(b) and topographic AFM image in figure 3(b)), and composed of three distinct segments shown in figure 1(a): the first segment, 4-5 µm in length, was grown without phosphine; the second segment, 10 µm in length, was grown at a phosphine to silane precursor ratio of 1:500; and the last segment, 5 µm in length, was grown


5

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 27

without phosphine. The segmented structure was achieved by independently scaling the growth times of both intrinsic and highly doped SiNWs grown under the same conditions: P=20torr, T=4600C, and He, SiH4 and PH3 (200 ppm in He) flow rates of 30, 2, and 20 SCCM, respectively. A detailed description of the basic growth procedure was previously reported1719,38,40

. The segmented SiNWs contain two junctions (transitions between segments).

Approaching the 1st junction from the bottom of Figure 1(a), the intrinsic core region is covered by a phosphorus-rich surface layer that was deposited from the vapor (VS doping) during the second (middle) segment growth, while the second n-type segment was doped via both VLS and VS dopant incorporation mechanisms. The 2nd junction is formed between the phosphorus-doped second (central) segment and the third (top) segment, which was grown without phosphine. As we will show, this unintentionally doped segment contains dopants that are incorporated solely from the catalyst reservoir. Fig. 1a illustrates schematically the segmented SiNW’s configuration. The NWs were mechanically transferred onto a highly oriented pyrolytic graphite (HOPG) film supported by conductive copper tape (to avoid charging effects) for npSAM measurements. The use of HOPG reduced the Auger spectral background, thus lowering the P detectability limit. The Auger data were acquired by a Physical Electronics PHI 700Xi Scanning Auger equipped with a Schottky field emission electron gun and a coaxial Cylindrical Mirror Analyzer (CMA) featuring a probe diameter down to ∼6 nm. The Auger data for elemental identification and for calculations of surface atomic concentrations were obtained using the dN(E)/dE peak to peak heights of the P LMM and Si KLL spectra and PHI supplied sensitivity factors. The P detectability limit was determined by the signal-to-noise limit of the observable Auger P LMM peak and was found to be 1/6000 (8×1018 cm-3) for the HOPG substrate. Depth profiling was


6

Page 7 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

performed by Argon ion sputtering and calibrated employing atomic force microscopy (AFM). [See Supplementary Material Online for the Auger technique technical details and data analysis.] Single NW devices, used for KPFM measurements, were fabricated by electron beam lithography and nickel evaporation to form symmetric contacts. The wafer on which the NWs were dispersed forms the backgate terminal, and was grounded throughout the measurements. Prior to the electrode evaporation step, the contact area was treated with a 24 sec oxygen plasma to remove residual photoresist and organic contamination, and with a 3 sec wet etch in buffered hydrofluoric acid to remove the native oxide layer. Current-voltage measurements were conducted using a semiconductor parameter analyzer (Agilent B1500A). The KPFM measurements were conducted using a Dimension 3100 AFM system (Bruker AXS) and an Au coated tip, in a controlled nitrogen environment glove box (less than 5ppm H2O). Measurements were performed in the ‘lift mode,’ where the topography is first scanned in the ‘tapping mode’. Next, the tip is raised to a constant 8 nm height above the topography trajectory and the CPD is measured. The limited characteristic depth from which Auger electrons are emitted (escape depth) is the basis for the surface sensitivity of Auger electron spectroscopy (AES). A sampling depth d of 2.6 nm is used in this study, ensuring that measured surface concentration is independent of the concentration measured after ~3 nm of material is removed by chemical etching [see Supplementary Material Online for further details]. We have previously shown how sequential etching is useful for radial dopant profiling18,38. It should be noted that np-SAM measures the total dopant concentration, whereas KPFM is sensitive to the carriers generated by active dopants. Thus the combination of these two techniques is particularly useful.


7

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 27

Fig. 2 shows the axial dependence of the surface P distribution along segmented NWs versus the distance from the catalyst averaged over 3 NWs (green points), one of which was remeasured (red points) after 3 nm etching with an Ar ion gun. The first (nominally undoped) segment has a surface P concentration of about 0.65% throughout its whole 5 µm length, which is attributed to surface deposition of P during growth of the second (highly doped) segment. Segment 2 exhibits decreasing P surface doping from 0.65% (3.25×1020cm-3) to slightly less than 0.1% (5×1019cm-3) towards the tip due to the accumulation of P on the surface during its growth. npSAM measurements of P doped NWs grown under similar conditions as the second segments yielded a similar decrease of the P surface concentration along the NW40. The final 5 µm segment exhibits a lower surface P concentration of about 0.1%. The detection of P in this third segment, despite the fact that no PH3 was flowing during its growth, indicates that residual P in the catalyst serves as the doping source. The last 200 nm of the third segment (in close vicinity of the Au tip) shows a rapid increase in P concentration reaching up to 4% ~25 nm from the Au tip. The rather high P concentration in vicinity of the Au tip also implies that traces of P remaining in the Au catalyst were expelled during cooling below the eutectic temperature. Etching 3 nm of the NW surface, thus removing the VS surface-doped layer, drastically changes the axial P concentration profile (Figure 2, red points). The first (VS doped) segment showed no detectable P concentration. In contrast, a measurable P concentration of about 0.1% was obtained for the second segment, which is attributed to incorporation from the catalyst into the NW core. The initial gradual increase, if considered statistically significant, may be due to a buildup of P in the catalyst upon initial exposure to phosphine. The third segment, which was doped by the residual P in the catalyst, did not exhibit any P doping above the detectability limit of 8×1018 cm-3 following the surface etching.


8

Page 9 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

The contributions of different dopant incorporation mechanisms to the surface dopant concentration can be identified by comparing the distinct doping profiles in each of the three segments. By integrating the P concentration throughout the third segment, including the region of high P concentration in the vicinity of the tip, the (previously unknown) P concentration in the liquid catalyst (ClP) during growth of the last segment is determined to be in the range of 1-10 at%. In the absence of a flux of P across the vapor-liquid interface, ClP must decrease throughout the third segment37. However, the rate of decrease is not resolved by the npSAM measurement as the dopant concentration lies at or below the detectability limit. The nearly constant near-surface dopant concentration throughout most of segment three is consistent with a very small distribution coefficient, leading to a very diffuse junction when considering primarily near surface (but not fully active) dopants. Comparison of the dopant profiles in the last two segments obtained after the 3 nm etch shows a much higher P incorporation into the NW core in the second segment (app. 7×1019 cm-3) than in the third segment, which was below the detectability limit of 8×1018 cm-3. We are therefore unable to determine the precise abruptness of the axial junction within the core. We note however that reduction in P concentration below the detection limit upon etching segment 3 suggests that most of the P was incorporated near the VLS trijunction, in agreement with recent APT observations39. Previous analyses of modulation doped structures including intrinsic-highly P doped segments in SiNWs in ref. 31 and axial p-n junctions in ref. 48 appear to show transitions that are more abrupt than those measured here, but the prior measurements are sensitive only to active dopant concentration. The surface dopants, to which the npSAM measurement is especially sensitive, are not expected to be fully active8. Under certain conditions, the properties of the electrical junction may be dominated by axial concentration variations within the core, rather than within


9

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 27

the shell. We investigate this possibility in some detail in the next section, but it is clear from the discussion above that one should consider both the radial and axial dopant distribution in order to accurately describe the behavior of a nominally axial junction. With the np-SAM results in mind, the NWs should show electrical characteristics similar to those of a single n+-n diode with a graded junction. Due to the heavily doped surface layer surrounding the interface between the 1st and the 2nd segments, both sides of this junction are expected to be degenerately doped. A sizable depletion region is therefore likely to form only in the second junction between the heavily doped 2nd segment and the lightly doped 3rd segment. Fig. 3 (a) shows a typical current-voltage (I-V) curve of a segmented SiNW device. Several deviations from the ideal diode characteristics are observed in Fig. 3 (a). First, the slope of the forward bias is significantly lower than that expected for an ideal diode. Such a phenomenon can be modeled by two resistors in series: (1) the internal resistance of the NW and (2) an ohmic contact resistance. Second, an offset bias is clearly seen in the graph. This offset bias is usually attributed to the built-in potential across the junction (Vbi); however, an accurate determination of the built-in potential from the I-V curve is impossible as the diode resistance in the offset region changes drastically over a small bias interval. Third, a "knee" in the reverse current, typically indicative of a generation-recombination current, is indicated at bias values lower than −0.5V. An equivalent circuit diagram accounting for these deviations from ideality is shown in the inset of Fig. 3 (a). Fig. 3 shows topography (b) and CPD (c) images of a segmented SiNW. The base of the NW is on the left side and the Au catalyst is on the right side. Both terminals of the device were grounded, as was the bottom gate terminal. The two line profiles shown in Fig. 3 (d) are line scans of the two surface measurements at the center of the wire, respectively. An abrupt change


10

Page 11 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

in the surface potential is observed at a distance of 15µm, which is the expected location for the junction between the 2nd and the 3rd segments. A typical series of CPD line-profiles taken at different reverse-biases from a single NW junction is shown in Fig. 4 (a). The measurement was performed by applying negative biases to the top (catalyst side) electrode, which is connected to the 3rd segment of the wire, while keeping the bottom (n+) electrode grounded. The bias applied to the top electrode was varied between 0V and −2V with 0.25V steps, which is a reverse biasing of the n+-n junction. The junction-potential drop under reverse bias, shown in Fig. 4 (a), follows the expected Vbi+VR behavior of a typical diode, where VR is the applied reverse bias. The figure demonstrates a considerable non-linear voltage drop on the n side of the wire with a built-in potential of Vbi = 0.15 V. The cylindrical n+-n NW system cannot be modeled using a simple analytical expression. A two dimensional model was therefore implemented using Sentaurus TCAD Device Simulator (Synopsis Inc.). The program numerically solves the Poisson equation coupled to the electron and hole continuity equation to simulate the electrical behavior of a single isolated semiconductor device, or several devices combined in a circuit. Using the one-dimensional builtin voltage for an n+-n junction:

Vbi =

k BT  n +  ln   q  n 

(1)

and using the measured Vbi of 0.15V, the ratio between the electron densities in the second and third segments, n+/n, is calculated to be 312.5. Initially, the device was modeled using a n+-n diode model (Fig. 4 (b)). The model is of a semiconductor NW composed of two 10µm long and 80nm wide segments with different P doping concentrations, placed over a 200nm oxide layer and a thick silicon substrate, representing the complete device geometry. The calculation was repeated several times using different doping profiles, ranging from n+=1017 cm-3 to n+=1020


11

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 27

cm-3. The n concentration was changed accordingly to keep the n+/n ratio constant at 312.5, the value that was extracted from the n+-n diode equation. The results of this calculation with the highest and lowest concentrations used are shown in Fig. 4 (c) and (d), respectively, overlayed with the experimental surface potential profiles. Fig. 4 shows that these simulations do not reproduce the measured potential profiles under reverse bias. Also, the simulated voltage drop over the n segment is approximately linear, whereas the measured curves show a significant nonlinear behavior. The measurements show that the band bending changes from that of a heavily doped wire at a small reverse bias to that of an intrinsic wire for larger values of reverse bias, suggesting that a thin doped layer dominates the device behavior under low bias but becomes fully depleted under higher bias. To improve the agreement with the measurements, a modified model, shown in Fig. 5 (a) was employed. This model includes a core-shell structure in the n segment with a P doped shell of thickness t and an intrinsic core. The simulation parameters n+, n and t were optimized to fit the measured results and were found to be: n+=5×1018 cm-3, n=7×1016 cm-3 in the shell, and 1014 cm-3 in the core. The shell thickness was found to be 10nm. The optimized results of the modified model, overlaying the measured results, are shown in Fig. 5 (b). The simulations show that the electrostatic characteristic of the junction is drastically affected by the changes in chemical composition along the NW radius. A recent study of ultrashallow doped p-n junction by Popadić et al.49 showed that in such depletable junctions, the current-voltage characteristics approach those of Schottky junctions, with significantly larger reverse (leakage) currents and ideality factors. This correlated study has two main advantages. First, validation of the lightly doped coreheavily doped shell structure of the VLS-doped third segment was achieved independently by


12

Page 13 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

both npSAM and KPFM. These findings are consistent with a recent report that VLS doping is concentrated near the tri-junction39. Second, correlating the chemical (npSAM) and active (KPFM) dopant concentrations enables us to determine the fraction of near-surface dopants that are active. The sampling depth of the npSAM is ~2.6 nm, so we assume in this simple analysis that the measured near-surface (chemical) dopant concentration of 5×1019 cm-3 is the average within that depth. For comparison, the 2-D device simulation of the KFPM data yielded an active dopant concentration of 7×1016 cm-3 within a shell thickness of 10 nm within a nanowire of 40 nm radius. The ratio of the total number of dopants to the number of the active dopants is in this case ~200:1. Hence, the chemical dopant concentration is two orders of magnitude higher than the active surface dopant concentration. Clearly some of the P atoms, such as those consumed during formation of the native oxide, are not in bonding configurations that lead to electron donation. Additionally, surface segregation of P may lead to an increase in ionization energy and subsequent donor deactivation50, and it is known that donors become deactivated due to the dielectric mismatch between the semiconductor NW and its surroundings8. The measurements and simulations presented here consistently show that regardless of the incorporation mechanism, the doped NWs have highly doped shell enclosing a lightly doped core with an active surface dopant concentration approximately two orders of magnitude smaller than the actual chemical composition. This observation, and the distinctions between the different dopant incorporation mechanisms discussed above, are in agreement with the recent finding that the NW’s perimeter is the preferred site for dopant incorporation through the VLS catalyst39. Furthermore, the high P concentration in or at the surface of the Au-Si liquid catalyst coupled with the small distribution coefficient account for the very pronounced reservoir effect observed in the n+-i junctions. Control of doping profiles in VLS grown NWs has been a major


13

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 27

challenge due to the multiple doping pathways, the sometimes subtle effects of these pathways, and the considerable challenges in metrology. The results obtained in this study, which advance a fundamental understanding of the dopant incorporation mechanisms, will be useful for the design and fabrication of many device structures, including doped semiconductor NWs with better control over their electrical properties and heterostructures with better-defined junctions.

Acknowledgements: Y.R. and I.A. would like to acknowledge the support of the Israel Binational Science Foundation (BSF) and the Israel Science foundation (ISF grant 498/11), Ministry of Science (MOS, grant 3-66799). L.L. would like to acknowledge the BSF for travel funds and the support of NSF under DMR-1006069. The atom-probe tomographic measurements were performed at the Northwestern University Center for Atom-Probe Tomography (NUCAPT). The local-electrode atom-probe (LEAP) tomograph was purchased and upgraded with funding from NSF-MRI (DMR-0420532) and ONR-DURIP (N00014-0400798, N000140610539, N00014-0910781) grants. Instrumentation at NUCAPT was supported by the Initiative for Sustainability and Energy at Northwestern (ISEN). NUCAPT is a Shared Facility at the Materials Research Center of Northwestern University, supported by the National Science Foundation's MRSEC program (DMR-1121262).


14

Page 15 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

References (1) Lu, W.; Lieber, C. M. J. Phys. D: Appl. Phys. 2006, 39, R387. (2) Garnett, E. C.; Yang, P. J. Am. Chem. Soc. 2008, 130, 9224-9225. (3) Czaban, J. A.; Thompson, D. A.; LaPierre, R. R. Nano Lett. 2009, 9, 148-154. (4) Yang, C.; Barrelet, C. J.; Capasso, F.; Lieber, C. M. Nano Lett. 2006, 6, 2929-2934. (5) Minot, E. D.; Kelkensberg, F.; van Kouwen, M.; van Dam, J. A.; Kouwenhoven, L. P.; Zwiller, V.; Borgström, M. T.; Wunnicke, O.; Verheijen, M. A.; Bakkers, E. P. A. M. Nano Lett. 2007, 7, 367-371. (6) Qian, F.; Gradečak, S.; Li, Y.; Wen, C.-Y.; Lieber, C. M. Nano Lett. 2005, 5, 2287-2291. (7) Patolsky, F.; Zheng, G.; Hayden, O.; Lakadamyali, M.; Zhuang, X.; Lieber, C. M. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 14017-14022. (8) Björk, M. T.; Schmid, H.; Knoch, J.; Riel, H.; Riess, W. Nature Nanotech. 2009, 4, 103107. (9) Yang, C.; Zhong, Z.; Lieber, C.M. Science 2005, 310, 1304-1307. (10) Smith, J.; Sandow, C.; Das, S.; Minamisawa, R.; Mantl, S.; Appenzeller, J. IEEE T. Electron. Dev. 2011, 58, 1822 -1829. (11) Ho, J.; Yerushalmi, R.; Jacobson, Z.; Fan, Z.; Alley, R.; Javey, A. Nat. Mater. 2008, 7, 62-67.


15

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 27

(12) Lee, S. T.; Zhang, Y. F.; Wang, N.; Tang, Y. H.; Bello, I.; Lee, C. S.; Chung, Y. W. J. Mater. Res. 1999, 14, 4503-4507. (13) Wagner, R. S.; Ellis, W. C. Appl. Phys. Lett. 1964, 4, 89-90. (14) Duan, X.; Lieber, C. M. Advanced Materials 2000, 12, 298-302. (15) Duan, X.; Wang, J.; Lieber, C. M. Appl. Phys. Lett. 2000, 76, 1116-1118. (16) Perea, D.; Wijaya, E.; Lensch-Falk, J.; Hemesath, E.; Lauhon, L. J. Solid State Chem. 2008, 181, 1642 - 1649. (17) Allen, J. E.; Perea, D. E.; Hemesath, E. R.; Lauhon, L. J. Advanced Materials 2009, 21, 3067-3072. (18) Koren, E.; Berkovitch, N.; Rosenwaks, Y. Nano Lett. 2010, 10, 1163-1167. (19) Koren, E.; Rosenwaks, Y.; Allen, J. E.; Hemesath, E. R.; Lauhon, L. J. Appl. Phys. Lett. 2009, 95, 092105. (20) Wu, Y.; Yang, P. J. Am. Chem. Soc. 2001, 123, 3165-3166. (21) Schmidt, V.; Wittemann, J.; Senz, S.; Gösele, U. Advanced Materials 2009, 21, 26812702. (22) Givargizov, E. J. Cryst. Growth 1975, 31, 20 - 30. (23) Haraguchi, K.; Katsuyama, T.; Hiruma, K.; Ogawa, K. Appl. Phys. Lett. 1992, 60, 745747. (24) Gudiksen, M.; Lauhon, L.; Wang, J.; Smith, D.; Lieber, C. Nature 2002, 415, 617-620.


16

Page 17 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

(25) Cui, Y.; Duan, X.; Hu, J.; Lieber, C. M. J. Phys. Chem. B 2000, 104, 5213-5216. (26) Wang, Y.; Lew, K.-K.; Ho, T.-T.; Pan, L.; Novak, S. W.; Dickey, E. C.; Redwing, J. M.; Mayer, T. S. Nano Lett. 2005, 5, 2139-2143. (27) Chung, S.-W.; Yu, J.-Y.; Heath, J. R. Appl. Phys. Lett. 2000, 76, 2068-2070. (28) Cui, Y.; Zhong, Z.; Wang, D.; Wang, W. U.; Lieber, C. M. Nano Lett. 2003, 3, 149152. (29) Lew, K.-K.; Pan, L.; Bogart, T. E.; Dilts, S. M.; Dickey, E. C.; Redwing, J. M.; Wang, Y.; Cabassi, M.; Mayer, T. S.; Novak, S. W. Appl. Phys. Lett. 2004, 85, 3101-3103. (30) Huang, Y.; Duan, X.; Cui, Y.; Lauhon, L. J.; Kim, K.-H.; Lieber, C. M. Science 2001, 294, 1313-1317. (31) Schmid, H.; Björk, M. T.; Knoch, J.; Karg, S.; Riel, H.; Riess, W. Nano Lett. 2009, 9, 173-177. (32) Fukata, N.; Ishida, S.; Yokono, S.; Takiguchi, R.; Chen, J.; Sekiguchi, T.; Murakami, K. Nano Lett. 2011, 11, 651-656. (33) Oberbeck, L.; Curson, N. J.; Hallam, T.; Simmons, M. Y.; Bilger, G.; Clark, R. G. Appl. Phys. Lett. 2004, 85, 1359-1361. (34) Tutuc, E.; Chu, J. O.; Ott, J. A.; Guha, S. Appl. Phys. Lett. 2006, 89, 263101. (35) Xie, P.; Hu, Y.; Fang, Y.; J., H.; Lieber, C. Proc. Natl. Acad. Sci. USA 2009, 106, 15254-15258.


17

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 27

(36) Garnett, E.; Tseng, Y.; Khanal, D.; Wu, J.; Bokor, J.; Yang, P. Nat. Nanotech. 2009, 4, 311-314. (37) Perea, D.; Hernesath, E.; Schwalbach, E.; Lensch-Falk, J.; Voorhees, P.; Lauhon, L. J. Nat. Nanotech. 2009, 4, 315-319. (38) Koren, E.; Hyun, J.K.; Givan, U.; Hemesath, E.R.; Lauhon, L.J.; Rosenwaks, Y. Nano Lett. 2011, 11, 183-187. (39) Connell, J. G.; Yoon, K.; Perea, D. E.; Schwalbach, E. J.; Voorhees, P. W.; Lauhon, L. J. Nano Lett. 2013, 13, 199-206. (40) Givan, U.; Hyun, J.; Koren, E.; Hammond, J.; Paul, D.; Lauhon, L.; Rosenwaks, Y. MRS 2011 spring meeting proceedings 2011, 1349. (41) Rosenwaks, Y.; Shikler, R.; Glatzel, T.; Sadewasser, S. Phys. Rev. B 2004, 70, 085320. (42) Clark, T. E.; Nimmatoori, P.; Lew, K.-K.; Pan, L.; Redwing, J. M.; Dickey, E. C. Nano Lett. 2008, 8, 1246-1252. (43) Li, N.; Tan, T. Y.; Gösele, U. Applied Physics A: Materials Science & Processing 2008, 90, 591-596. (44) Kodambaka, S.; Tersoff, J.; Reuter, M. C.; Ross, F. M. Science 2007, 316, 729-732. (45) Wen, C.-Y.; Reuter, M. C.; Bruley, J.; Tersoff, J.; Kodambaka, S.; Stach, E. A.; Ross, F. M. Science 2009, 326, 1247-1250. (46) Dick, K. A.; Bolinsson, J.; Borg, B. M.; Johansson, J. Nano Lett. 2012, 12, 3200-3206.


18

Page 19 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

(47) Perea, D. E.; Li, N.; Dickerson, R. M.; Misra, A.; Picraux, S. T. Nano Lett. 2011, 11, 3117-3122. (48) Vallett, A. L.; Minassian, S.; Kaszuba, P.; Datta, S.; Redwing, J. M.; Mayer, T. S. Nano Lett. 2010, 10, 4813-4818. (49) Popadic, M.; Lorito, G.; Nanver, L. IEEE T. Electron. Dev. 2009, 56, 116 -125. (50) Fernández-Serra, M. V.; Adessi, C.; Blase, X. Phys. Rev. Lett. 2006, 96, 166805.


19

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 27

Figure Captions

Figure 1: (a): Schematic illustration of the segmented wire growth and predicted dopant profile. Step 1- growth of the 1st, 5µm long, intrinsic segment. Step 2- growth of highly P doped (1/500), 10µm, segment on top of the intrinsic segment as well as the VS growth of a heavily doped shell around both segments. Step 3- growth of the 3rd ,5µm long, intrinsic segment in which the catalyst is the only dopant source. (b): SEM image of a 20µm long, 80nm diameter, segmented SiNW with no observable morphological signs of the different growth conditions of the three segments.

Figure 2: The segmented NW’s P profile obtained by np-SAM as measured at the NWs surface and averaged over 3 NWs as grown (green) and after 3nm etching (red). Error bars were defined according to the variation between the 3 different NWs measured, or, in the case of the etched NW (red), according to the instrument systematic error (the maximum variation of repeated measurements taken at the same location).

Figure 3: (a) A typical current - voltage (I-V) curve showing the junction-like behavior of the segmented NW. An equivalent circuit that accounts for the offset bias and the series resistance, as well as the diode’s rectifying behavior, is also shown. The inset shows the semi-log scale I-V characteristics where the device resistance is clearly evident. (b) A topography image showing the NW’s uniform diameter and (c) CPD image of a segmented NW. The electrodes are shown on the sides, the gold catalyst (not shown) is in the right hand side of the figure. The dashed


20

Page 21 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

(green) line indicates the step in surface potential at the 2nd junction. (d) Line-profiles showing the segmented NW’s topography (bottom) and CPD (top). The step in the CPD is attributed to the diode built-in potential. Figure 4: (a) CPD line scan profiles of a segmented NW under reverse bias. The dashed line separates the n segment from the n+ segment, for convenience. The (red) shaded area indicates the position of the left hand side biased electrode. The right hand side grounded electrode is not shown. (b) The initial model used to simulate the device electric behavior. The model consists of two 80nm thick silicon bars representing the 2nd and the 3rd segments of the wire, positioned on top of a highly doped thermally oxidized silicon wafer that acts as a bottom gate electrode. (c+d) An overlay of the results of the electrostatic simulations on the measured results showing the heaviest (n+ = 1020 cm-3, n = 3.2×1017 cm-3) (c) and lightest (n+ = 1017 cm-3, n = 3.2×1014 cm-3) (d) doping profiles used in the simulations.

Figure 5: (a) The modified model used to simulate the segmented NW under reverse bias. The n segment is modeled with a core-shell configuration. Contacts are shown in thick (purple) lines (b) A comparison of the optimized results of the modified simulation (with n+=5×1018 cm-3 and n=7×1016cm-3) and a typical measurement result.


21

Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

254x128mm (150 x 150 DPI)


Page 22 of 27

Page 23 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

254x179mm (149 x 147 DPI)


Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

134x190mm (150 x 150 DPI)


Page 24 of 27

Page 25 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

261x195mm (150 x 150 DPI)


Nano Letters

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

154x190mm (150 x 150 DPI)


Page 26 of 27

Page 27 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Nano Letters

221x95mm (96 x 96 DPI)


Spatially Resolved Correlation of Active and Total Doping

Recommend Documents