Enhanced Electron Mobility in Nonplanar Tensile Strained Si

Subscriber access provided by READING UNIV

Communication

Enhanced Electron Mobility in Non-Planar Tensile Strained Si Epitaxially Grown on SixGe1-x Nanowires Feng Wen, and Emanuel Tutuc Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.7b03450 • Publication Date (Web): 29 Nov 2017 Downloaded from http://pubs.acs.org on December 5, 2017

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 22

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

Enhanced Electron Mobility in Non-Planar Tensile Strained Si Epitaxially Grown on SixGe1-x Nanowires Feng Wen and Emanuel Tutuc* Microelectronics Research Center, The University of Texas at Austin, 10100 Burnet Road, Bldg. 160, Austin, Texas 78758, USA KEYWORDS: Nanowire, core-shell, elastic strain, TEM, Raman spectroscopy, field-effect transistor, silicon, silicon-germanium.

ABSTRACT: We report the growth and characterization of epitaxial, coherently strained SixGe1x-Si

core-shell nanowire heterostructure through vapor-liquid-solid growth mechanism for the

SixGe1-x core, followed by an in-situ ultra-high-vacuum chemical vapor deposition for the Si shell. Raman spectra acquired from individual nanowire reveal the Si-Si, Si-Ge and Ge-Ge modes of the SixGe1-x core, and the Si-Si mode of the shell. Due to the compressive (tensile) strain induced by lattice mismatch, the core (shell) Raman modes are blue (red) shifted compared to those of unstrained bare SixGe1-x (Si) nanowires, in good agreement with values calculated using continuum elasticity model coupled with lattice dynamic theory. A large tensile strain of up to 2.3% is achieved in the Si shell, which is expected to provide quantum confinement for electrons due to a positive core-to-shell conduction band offset. We demonstrate n-type metal-

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 22

oxide-semiconductor field-effect transistors using SixGe1-x-Si core-shell nanowires as channel, and observe a 40% enhancement of the average electron mobility compared to control devices using Si nanowires, due to an increased electron mobility in the tensile-strained Si shell.

Group IV semiconductor nanowires are potential candidates for ultimately scaled transistor devices.1-4 Nanowire devices are closely related to non-planar complementary metal-oxidesemiconductor (MOS) technologies, and allow the realization of Ω-gate or gate-all-around (GAA) MOS transistors with superior electrostatic control of the channel compared to planar MOS devices, which mitigates short channel effect.5 Strain engineering is a key technological advance to address some of limitations associated with the device scaling,6-8 and can be implemented in nanowire heterostructures. Both axial9, 10 and radial11, 12 heterostructures have been demonstrated in semiconductor nanowires grown by the vapor-liquid-solid (VLS) mechanism. Moreover, one-dimensional carrier quantum confinement can be realized in radial, core-shell nanowire heterostructures due to the radial band offset among different group IV materials.11, 13-17 Hole confinement has been realized in Ge-Si and Ge-SixGe1-x core-shell nanowires,14, 17 due to a large valence band offset between Ge and Si. Electron confinement in group IV core-shell heterostructures is, however, more challenging to achieve due to smaller conduction band offset between Si and Ge. Furthermore, the relative position of the conduction band edge depends on the strain in a Si/Ge system.18,

19

In epitaxial, coherently strained core-shell nanowire

heterostructures the lattice mismatch-induced elastic strain changes both the band structure and the band alignment between core and shell. Previous studies have reported electron confinement in coherently tensile-strained Si layers on relaxed20, 21 and compressive-strained22 SixGe1-x planar


2

Page 3 of 22

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

substrates. Moreover, those heterostructures also show enhanced electron mobility with respect to unstrained silicon, due to a reduced electron effective mass in the transport direction, and a reduction of intervalley phonon scattering.23 Various techniques have been adopted to induce tensile strain in non-planar Si field-effect transistors (FETs), including nano-patterning of strained Si on insulating substrates,24, 25 the use of SiC as source/drain and/or SiN liner,26 strain response due to gate electrode27 or oxide,28 and the use of strain-relaxed SiGe buffer as virtual substrate.29 The one-dimensional counterpart of strained Si on SixGe1-x, namely coherently strained SixGe1-x-Si core-shell nanowires represent a promising platform to study the impact of strain on electron transport, and has relevance for high mobility n-type MOSFETs (nMOSFETs), as they combine quantum confinement and enhanced electron mobility in the strained Si shell, with the electrostatic control of the Ω-gate or GAA geometry. In this study, we present the growth, structural and electrical characterization of epitaxial, coherently strained SixGe1-x-Si core-shell nanowires. We use Raman spectroscopy to probe the strain-induced shift of optical phonon frequencies in both core and shell regions, and compare the experimental data with calculations using the calculated strain simulated assuming a continuum elasticity model combined with lattice dynamic theory. We provide a scaling study of SixGe1-x-Si core-shell nanowire nMOSFETs with highly doped source and drain, and show that the tensile strain in the Si shell leads to a 40% electron mobility enhancement compared to bare Si nanowire FETs. Figure 1a depicts the growth of our SixGe1-x-Si core-shell nanowires, consisting of a sequence of VLS SixGe1-x core growth, followed by an in-situ Si shell growth using ultra-high-vacuum chemical vapor deposition (CVD). We start with a Si(111) wafer, using diluted hydrofluoric acid to remove the native oxide and then evaporating an 8 Å thick Au film. The substrate is then


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

Page 4 of 22

transferred to a cold wall ultra-high-vacuum CVD growth chamber, annealed in a H2 ambient at 370 °C for 15 min, which leads to the formation of Au nanoparticles as catalysts for the VLS growth. The SixGe1-x nanowire cores are grown at a temperature of 305 °C and pressure of 10 Torr using a combination of SiH4 (100%, 100 sccm) and GeH4 (20% diluted in He, 20 sccm) as precursors. The epitaxial Si shell growth is then performed in-situ at a temperature of 460 °C, pressure of 40 mTorr, using SiH4 as gas source. While the shell growth will result in additional VLS growth, based on the cross-sectional scanning electron micrograph (SEM) of the grown nanowires, we find that the axial growth during the Si shell growth is less than 100 nm for a 10 µm long nanowire, because of the lower precursor pressure. We also grow SixGe1-x nanowires using the same VLS growth recipe as the SixGe1-x cores of SixGe1-x-Si core-shell nanowires, as well as Si nanowires at a temperature of 420 °C and pressure of 10 Torr using SiH4 as precursor. Bare SixGe1-x and Si nanowires serve as baselines to extract the unstrained optical phonon frequencies in the core and shell, respectively, while bare Si nanowires are also used to fabricate nMOSFET control devices. We employ transmission electron microscopy (TEM) to study the morphology and crystal structure of the SixGe1-x-Si core-shell nanowires. Figure 1b illustrates a single crystal nanowire heterostructure with shell grown epitaxially on core, where the sidewall demonstrates no obvious saw-tooth facets30, 31 or strain-induced surface roughening.32, 33 The Fourier transform of panel 1b data shown in the inset reveals the growth direction is along in the 20 – 40 nm diameter range. We note it is also the case for bare Si and SixGe1-x nanowires. Figure 1c shows the crosssectional TEM image of a SixGe1-x-Si core-shell nanowire. Facets for the core and shell are not well discernible, except for {100} planes of the shell. We measure the Si shell thickness (tsh) as 4.2 nm, corresponding to a CVD growth rate of 0.47 Å/min.


4

Page 5 of 22

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

Figure 1. (a) Schematic of the SixGe1-x-Si core-shell nanowire growth, showing Au catalyst nanoparticles, the SixGe1-x core growth by VLS, and the Si shell growth by CVD. Arrows indicate the growth direction in the corresponding regime. (b) TEM image of a oriented SixGe1-x-Si core-shell nanowire, where interfaces are marked for clarity. Inset: FFT of the main panel data. (c) Cross-sectional transmission electron micrograph of a oriented SixGe1-x-Si core-shell nanowire. The core-shell interface is marked for clarity. Inset: FFT of the main panel data. The zone axis is assumed along [110] direction. One main attribute of such lattice-mismatched heterostructure is the associated elastic strain. The tensile strain in the Si shell induces a conduction band energy splitting associated with the crystal asymmetry, resulting in reduced electron effective mass and enhanced electron mobility. The strain will also shift the optical phonon frequencies, and therefore change the Raman spectrum of the heterostructure. We use Raman spectroscopy on individual SixGe1-x-Si core-shell nanowire to probe the shifted optical phonon frequencies (see Methods). Figure 2 shows the Raman spectra of a 28 nm diameter SixGe1-x-Si core-shell nanowire [panel (a)], a 40 nm diameter SixGe1-x-Si core-shell nanowire [panel (b)], a bare SixGe1-x nanowire [panel (c)] and a bare Si nanowire [panel (d)]. The Raman spectrum of the Si nanowire reveals a single peak at 520.6 cm-


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

1

Page 6 of 22

, associated with Si-Si Raman mode, while that of the SixGe1-x nanowire reveals three major

peaks at 295, 407 and 478 cm-1, associated with the Ge-Ge, Si-Ge and Si-Si Raman modes, respectively. Compared to the SixGe1-x nanowire, Raman spectra of the SixGe1-x-Si core-shell nanowires show one additional peak near 510 cm-1, associated with the shell Si-Si mode. Furthermore, the three Raman modes of the SixGe1-x core of SixGe1-x-Si core-shell nanowires are blue-shifted compared to those of bare SixGe1-x nanowire, because of the compressive strain induced by the epitaxial Si shell. Conversely, the shell Si-Si Raman modes in the same SixGe1-xSi core-shell nanowires are red-shifted with respect to the unstrained Si-Si mode, indicating a tensile strain. A comparison of Fig. 2a and 2b data shows that the three core Raman modes of the 28 nm diameter SixGe1-x-Si core-shell nanowire have larger blue shifts, while the shell Si-Si mode has a smaller red shift, compared to those of the 40 nm diameter one. This finding can be explained by an increasing compressive strain in the core, and a decreasing tensile strain in the shell with reducing the nanowire diameter at a constant tsh. The average Si:Ge core content value, x = 0.41 is determined from the relative intensities among the Si-Si, Si-Ge, and Ge-Ge core Raman modes,34,

35

in good agreement with the value measured from bare SixGe1-x

nanowires using energy-dispersive X-ray spectroscopy with scanning TEM. We find the composition depends weakly on diameter, and varies by less than 3% in the diameter range probed here.36 To better understand the dependence of Raman modes on elastic strain and further substantiate the coherently strained nature of our nanowire heterostructures, we calculate the strain and the strain-induced shift of Si-Si Raman modes in both core and shell regions as a function of diameter, and compare it with the experimental data. We calculate the elastic strain using a finite element modelling15, 37 approach (Abaqus®, Dassault Systemes), which takes into account the


6

Page 7 of 22

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

crystal elastic anisotropy. For simplicity, we assume all SixGe1-x-Si core-shell nanowires have cylindrical cross-sections. Figure 2 also presents the simulated two-dimensional contour plots of the strain tensor [panel (e-h)], along with the hydrostatic strain = ( + + )/3 [panel (i)] and the corresponding conduction band energy (EC) calculated under strain38,

39

for a

Si0.41Ge0.59-Si core-shell nanowire [panel (j)]. In order to employ a complete stiffness tensor including three independent components the calculations are done in Cartesian coordinates with the x, y and z-axes of strain tensors as the principal crystallographic axes [100], [010] and [001]. On the other hand, the nanowire growth, and cross section axes are different as indicated in panel (e). We find the distribution of individual strain tensor component in the core is uniform compared to that in the shell,40 while is uniform in both regions. The average values,

closely related to the volume change,41 is -0.65% in the core and 0.56% in the shell in this example. We note the magnitude is large compared to values reported in strained Si n-type

FinFETs.42 We calculate a core-to-shell conduction band offset of 270 meV in this example based on the mean values of the calculated conduction band energies in both regions, indicating electron confinement in the Si shell. This offset ranges from 265 to 280 meV for nanowire with diameter from 25 to 40 nm.


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 22

Figure 2. Comparison of Raman spectra acquired from an individual (a) 28 nm diameter Si0.41Ge0.59-Si core-shell nanowire, (b) 40 nm diameter Si0.41Ge0.59-Si core-shell nanowire, (c) Si0.41Ge0.59 nanowire, and (d) Si nanowire. The Si-Si in bulk Si, and Ge-Ge, Si-Ge, and Si-Si in bulk Si0.41Ge0.59 Raman modes are marked with vertical dash-dotted lines. Two-dimensional contour plots of the simulated (e-h) strain tensor components, (i) and (j) calculated EC for a Si0.41Ge0.59-Si core-shell nanowire, with tsh = 4.2 nm. The crystal directions are indicated in panel

(e). Next we evaluate the impact of elastic strain on the optical phonon frequencies using lattice dynamic theory. The presence of elastic strain in a cubic crystal shifts and splits the optical phonon modes initially triply degenerated at the zone center. The strain-induced shift of each Raman mode can be calculated using a linear deformation potential approach, by applying the secular equation of lattice dynamic theory:43, 44 + + 2 2

2 + ( + ) 2

2 2

+ +

=0


8

Page 9 of 22

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

where εij are the strain tensor components, p, q, and r are the material’s normalized phonon deformation potential values, given in Table 1. The strain tensor components are referenced to the principal crystallographic axes. The eigenvalue λi = ωi2 − ωi02 is the strain-induced shift of

mode i, where ωi0 is the unstrained frequency. The associated intensity of mode i is ∝

| ∙ !" ∙ #$% |&, where !" is the strained Raman tensor for mode i,45 and #$% are the

incident and scattered light polarization, respectively.12, 46 The light polarizations are assumed

parallel to the nanowire axis, consistent with the “antenna effect” for coupling with the nanowire structure.47, 48 We have experimentally verified this assumption by a cos2θ dependence of the scattered beam intensity if the incident light polarization is aligned at an angle θ with respect to the nanowire main axis, and also polarizing the scattered light. Table 1. Normalized Phonon Deformation Potentials for Si and Ge: Material

p/ω02

q/ω02

r/ω02

Si

-1.8449

-2.3549

-0.7150

Ge

-1.6649

-2.1949

-1.1151

Normalized phonon deformation potential values of Si0.41Ge0.59 used are the linear interpolations between those values of Si and Ge. Figure 3a summarizes the diameter dependence of the calculated (solid lines), and experimentally acquired (symbols) Si-Si Raman modes for Si0.41Ge0.59-Si core-shell nanowires. The unstrained Si-Si Raman modes of core and shell are marked with red and black dashed lines, at 478 cm-1 and 520.6 cm-1, respectively. The experimental data and calculations are in very good agreement, and the diameter dependence is explained by an increasing elastic strain in the


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 22

core at reduced nanowire diameters, and a larger strain-induced shift of the Raman mode. Figures. 3b-d show two-dimensional contour plots of the calculated Si-Si Raman mode shift, and Figs. 3e-g show the associated intensities in both core and shell regions. The three-fold degeneracy of the Si-Si Raman mode in either core and shell can be lifted by the elastic strain, with values for individual modes determined by the secular equation. The intensity of each Raman mode is also calculated to identify the active modes. In the Si0.41Ge0.59 core, we find only one active Si-Si Raman mode in the core, labelled mode 1. On the other hand, two modes in the Si shell, labelled 1 and 2 have comparable intensity, while a third mode 3 has a weaker intensity. Experimentally only one peak associated with the shell can be resolved in the Raman spectra, albeit with a larger full width at half-maximum (FWHM) of 11 - 14 cm-1, compared to 4 cm-1 measured in bare Si nanowires (Figs. 2a-d). Consequently, in Fig. 3a we include the calculated values for the three shell Si-Si Raman mode along with the position of the weighted average using the calculated relative intensities.52 The calculated individual shell Si-Si Raman modes are presented in a corridor of values to indicate the variation across the shell. The total broadening of the shell Si-Si Raman mode ranges from 12.5 to 13.6 cm-1 according to Fig. 3a, explaining the increased FWHM observed experimentally. Overall, we find the experimental data agrees well with the calculated Si-Si Raman modes in both core and shell regions, consistent with a coherently strained heterostructure. We also find that the normal strain shifts the shell Si-Si Raman mode from the unstrained value at 520.6 cm-1,53 while the shear strain only lifts the threefold degeneracy but does not change the weighted average. Therefore, based on the linear deformation potential model applied in this study, the calculated shell Si-Si Raman mode shift from 520.6 cm-1 is linearly proportional to . To help visdualize the relation between strain and


10

Page 11 of 22

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 Si-Si Raman mode shift in the Si shell, the right y-axis in Fig. 3a shows mapped from the Raman shift (left y-axis).

Figure 3. (a) Diameter dependence of the core and shell Si-Si Raman modes for Si0.41Ge0.59-Si core-shell nanowires. The Si-Si Raman modes for core and shell are shown in black and red, respectively. Symbols represent experimental data, and solid lines represent calculations. The black and red dashed lines mark the position of the unstrained Si-Si Raman modes in shell and core, respectively. Experimental data of Si-Si Raman modes acquired from bare Si nanowire are included for comparison. The calculation results of three different shell Si-Si modes are presented with three corridors to indicate their variations across the shell. Calculated (b-d) Si-Si modes Raman shift, and (e-g) corresponding intensities for a coherently-strained Si0.41Ge0.59-Si core-shell nanowire, assuming the incident and scattered light are parallel to the nanowire growth axis. Next, we turn to electron transport in the -oriented Si0.41Ge0.59-Si core-shell nanowires using Ω-gated nMOSFETs with highly doped source and drain. The details of the fabrication process can be found in the Method section. Figures 4a-b show the SEM images of a Si0.41Ge0.59-


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 22

Si core-shell and a bare Si nanowire nMOSFET, respectively. The devices have a same gate length LG of 840 nm and diameter d of 28 and 25 nm. To realize Ohmic contact between metal contacts and the nanowires we perform a low energy (5 keV) ion implantion of phosphorous in source and drain regions using gates as self-aligned masks. This approach yields low (< 0.3 kΩ) metal-to-semiconductor resistance values, and an extension resistance of 34 ± 4 kΩ/µm in the nanowire diameter range investigated. We estimate the total source/drain external resistance to be 50 ± 6 kΩ, for nanowire devices with a 1.5 µm source and drain extension length. Figures 4c-d show the drain current (ID) vs. drain voltage (VD) measured at various gate voltages (VG) for a Si0.41Ge0.59-Si core-shell (Fig. 4c), and a bare Si nanowire device (Fig. 4d). The corresponding ID − VG characteristics at different VD for the same devices are shown in Fig. 4e-f. Consistent with the source and drain doping, the devices show n-type enhancement mode MOSFET behavior with an on/off current ratio at a drain bias VD = 50 mV larger than 106 in Si0.41Ge0.59-Si core-shell nanowire devices. The devices possess small drain induced barrier lowering, of a few tens of mV threshold voltage (VT) shift between VD = 50 mV and VD = 1 V. The subthreshold swing is ~100 mV/dec. In addition, we notice a reduced average VT of 1.2 V for Si0.41Ge0.59-Si core-shell nanowires to 1.4 V for Si nanowires, which we attribute to a smaller bandgap, and lower conduction band for the tensile-strained Si.54 To decouple the contact and channel instrinsic resistance in Figs. 4g-h we plot the total device resistance (RTotal) vs. LG at different gate overdrive, VG − VT values. The intersection point of the linear fit to RTotal vs. LG data at each gate overdrive yield the total source/drain external resistance (RC) of 45 kΩ.


12

Page 13 of 22

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

Figure 4. Electrical characteristics of Si0.41Ge0.59-Si core-shell and bare Si nanowire nMOSFETs. SEM images of (a) a Si0.41Ge0.59-Si core-shell and (b) a bare Si nanowire nMOSFETs. The source, gate, and drain contacts are labeled, and the phosphorus implant region are shaded in red. (c-d) Output and (e-f) transfer characteristics of the same devices shown in panels (a), and (b), respectively. The VG (VD) value is indicated in each trace for the output (transfer) curves. RTotal versus LG at different overdrive voltages (VG − VT) for (g) Si0.41Ge0.59-Si core-shell and (h) bare Si nanowire nMOSFETs, respectively. Linear fits (solid lines) to the experimental data (symbols) are also included. To compare the electron mobility ('( ) in Si and Si0.41Ge0.59-Si core-shell nanowires, we extract

the '( values in multiple devices from the VG dependence of the intrinsic channel channel

conductance ) = '( *+ (,+ , )/-+ , where CG is the Ω-gate capacitance per unit length, and

) = (!.%$/ !0 )12. The CG values are extracted from finite-element simulations (Sentaurus

TCAD, Synopsys). We deduce the electron mobility using '( = (-+ /*+ ) ∙ (3) /3,+ ). Figure 5 histograms summarize the '( values extracted from over eighty Si and Si0.41Ge0.59-Si core-shell


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 22

nanowire nMOSFET devices. The lines are Gaussian distributions with the mean and variance of the experimental data. We find the mean value of the electron mobilities of Si0.41Ge0.59-Si core-shell nanowire devices shows a 40% increase compared to that of Si nanowire devices. We note that the non-optimized gate stack reduces the mobilties for both bare Si and the Si0.41Ge0.59Si core-shell nanowire nMOSFETs, and offsets in part the mobility enhancement associated with tensile strain in Si. Indeed, we expect that the use of an optimized gate stack to lead to a higher mobility in bare Si nMOSFETs, and a larger mobility enhancement for Si0.41Ge0.59-Si core-shell nanowire nMOSFETs. It is particularly noteworthy that oriented tensile-strained Si channel has also been found to be most desirable for trigate nMOSFETs,55 coinciding with the growth direction of small diameter VLS Si (Ge) nanowires.56, 57

Figure 5. Histograms of '( values in (a) 31 Si nanowire nMOSFETs and (b) 52 Si0.41Ge0.59-Si core-shell nanowire nMOSFETs. Solid lines mark the Gaussian distributions corresponding to the experimental data.


14

Page 15 of 22

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 conclusion, we demonstrate the growth, structural and electrical characterization of epitaxial, coherently strained SixGe1-x-Si core-shell nanowires. Conduction band edge energy calculations indicate a positive core-to-shell conduction band offset of 280 meV due to a large tensile strain in the Si shell. Raman spectroscopy reveals peaks associated with Si-Si, Si-Ge, Ge-Ge modes in the core and Si-Si mode in the shell. The core (shell) peaks are blue-shifted (red-shifted) from their unstrained values, consistent with a compressive (tensile) strain in the core (shell) region, and the Raman shift values agree well with calculations using lattice dynamic theory combined with finite-element strain simulation. Enhancement mode nMOSFETs in SixGe1-x-Si core-shell nanowires show a significant increase in mobility by comparison with control Si nanowire devices. Methods: Raman Spectroscopic Measurements: Raman measurement were conducted using a Renishaw InVia µ-Raman spectrometer with backscattering geometry, 532 nm incident laser with around 1µm2 focused spot size and 13kW·cm−2 power density, and 100× objective lens. To eliminate Raman signals originating from the Si substrate the nanowires are dispersed on 20/60 nm Ti/Au films. Nanowire nMOSFETs Fabrication: The nMOSFETs are prepared using undoped Si and SixGe1-x-Si core-shell nanowires. The nanowires are first transferred to a degenerately doped Si substrate with 64 nm thermal oxide. The sample is then etched in dilute hydrofluoric acid to remove the native oxide on the nanowire surface. A 7.5 nm Al2O3 top-gate dielectric is grown by atomic layer deposition at 250 °C, using trimethylaluminum as precursor, and O2 plasma as oxidizer. The dielectric constant of the deposited Al2O3 film is 7.6, extracted from MOS capacitor devices. The gate is patterned using electron-beam lithography (EBL), TaN sputtering,


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 22

and liftoff. The samples are then implanted with phosphorus ions at an energy of 5 keV with a total dose of 5×1014 cm-2. During implant the sample is tilted by 32° in four quadrants to get uniform coverage. Phosphorus dopants are activated using rapid thermal annealing process at 550 °C for 10 min in a N2 ambient. The source and drain contacts are defined by EBL, Ni deposition, and liftoff. AUTHOR INFORMATION Corresponding Author *Emanuel Tutuc. E-mail: [email protected] Author Contributions F.W. performed the nanowire growth, TEM, Raman measurements, simulations, calculations, FET fabrication and electrical characterization. F.W. and E.T. analyzed the data and co-wrote the manuscript. Both authors have contributed to and approved the final version of the manuscript. Funding Sources This work was supported by the National Science Foundation, grant DMR-1507654. Notes The authors declare no competing financial interest. TOC


16

Page 17 of 22

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.

Wang, D.; Wang, Q.; Javey, A.; Tu, R.; Dai, H.; Kim, H.; McIntyre, P. C.; Krishnamohan,

T.; Saraswat, K. C. Applied Physics Letters 2003, 83, (12), 2432. 2.

Schmidt, V.; Riel, H.; Senz, S.; Karg, S.; Riess, W.; Gosele, U. Small 2006, 2, (1), 85-88.

3.

Liang, G.; Xiang, J.; Kharche, N.; Klimeck, G.; Lieber, C. M.; Lundstrom, M. Nano Lett

2007, 7, (3), 642-646. 4.

Dillen, D. C.; Kim, K.; Liu, E.; Tutuc, E. Nat Nanotechnol 2014, 9, (2), 116-120.

5.

Jae Young, S.; Woo Young, C.; Ju Hee, P.; Jong Duk, L.; Byung-Gook, P. IEEE

Transactions On Nanotechnology 2006, 5, (3), 186-191. 6.

Uchida, K.; Krishnamohan, T.; Saraswat, K. C.; Nishi, Y., In Physical mechanisms of

electron mobility enhancement in uniaxial stressed MOSFETs and impact of uniaxial stress engineering in ballistic regime. Electron Devices Meeting, 2005. IEDM Technical Digest. IEEE International, 129 - 132. 7.

Hoyt, J. L.; Nayfeh, H. M.; Eguchi, S.; Aberg, I.; Xia, G.; Drake, T.; Fitzgerald, E. A.;

Antoniadis, D. A., In Strained silicon MOSFET technology. Electron Devices Meeting, 2002. IEDM Technical Digest. IEEE International, 23 - 26.


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

8.

Page 18 of 22

Rim, K. et al., In Fabrication and Mobility Characteristics of Ultra-thin Strained Si

Directly on Insulator (SSDOI) MOSFETs. Electron Devices Meeting, 2003. IEDM Technical Digest. IEEE International, 3.1.1 - 3.1.4. 9.

Dayeh, S. A.; Wang, J.; Li, N.; Huang, J. Y.; Gin, A. V.; Picraux, S. T. Nano Lett 2011,

11, (10), 4200-4206. 10.

Clark, T. E.; Nimmatoori, P.; Lew, K.-K.; Pan, L.; Redwing, J. M.; Dickey, E. C. Nano

Lett 2008, 8, (4), 1246-1252. 11.

Lauhon, L. J.; Gudiksen, M. S.; Wang, D.; Lieber, C. M. Nature 2002, 420, 57-61.

12.

Dillen, D. C.; Varahramyan, K. M.; Corbet, C. M.; Tutuc, E. Physical Review B 2012, 86,

(4). 13.

Lu, W.; Xiang, J.; Timko, B. P.; Wu, Y.; Lieber, C. M. Proc Natl Acad Sci U S A 2005,

102, (29), 10046-51. 14.

Nah, J.; Dillen, D. C.; Varahramyan, K. M.; Banerjee, S. K.; Tutuc, E. Nano Lett 2012,

12, (1), 108-112. 15.

Dillen, D. C.; Wen, F.; Kim, K.; Tutuc, E. Nano Lett 2016, 16, (1), 392-398.

16.

Huang, S.; Yang, L. Applied Physics Letters 2011, 98, (9), 093114.

17.

Xiang, J.; Lu, W.; Hu, Y.; Wu, Y.; Yan, H.; Lieber, C. M. Nature 2006, 441, (7092), 489-

493. 18.

Yang, L.; Watling, J. R.; Wilkins, R. C. W.; Boriçi, M.; Barker, J. R.; Asenov, A.; Roy, S.

Semiconductor Science and Technology 2004, 19, (10), 1174-1182. 19.

Rieger, M. M.; Vogl, P. Physical Review B 1993, 48, (19), 14276-14287.

20.

Mii, Y. J.; Xie, Y. H.; Fitzgerald, E. A.; Monroe, D.; Thiel, F. A.; Weir, B. E.; Feldman,

L. C. Applied Physics Letters 1991, 59, (13), 1611.


18

Page 19 of 22

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

21.

Gámiz, F.; Cartujo-Cassinello, P.; Roldán, J. B.; Jiménez-Molinos, F. Journal of Applied

Physics 2002, 92, (1), 288-295. 22.

Abstreiter, G.; Brugger, H.; Wolf, T.; Jorke, H.; Herzog, H. J. Phys Rev Lett 1985, 54,

(22), 2441-2444. 23.

Welser, J.; Hoyt, J. L.; Takagi, S.; Gibbons, J. F., In Strain dependence of the

performance enhancement in Strained-Si n-MOSFETs. Electron Devices Meeting, 1994. IEDM '94. Technical Digest., International, 373 - 376. 24.

Yang, F.-L. et al., In Strained FIP-SOI (FinFET/FD/PD-SOI) for Sub-65 nm CMOS

Scaling. VLSI Technology, 2003. Digest of Technical Papers. 2003 Symposium on, 137 - 138. 25.

Hashemi, P.; Gomez, L.; Canonico, M.; Hoyt, J. L., In Electron Transport in Gate-All-

Around Uniaxial Tensile Strained-Si Nanowire n-MOSFETs. Electron Devices Meeting, 2008. IEDM 2008. IEEE International, 1 - 4. 26.

Ang, K.-W. et al., In 50 nm Silicon-On-Insulator N-MOSFET Featuring Multiple

Stressors: Silicon-Carbon Source/Drain Regions and Tensile Stress Silicon Nitride Liner. VLSI Technology, 2006. Digest of Technical Papers. 2006 Symposium on, 66 - 67. 27.

Kang, C. Y. et al., In A Novel Electrode-Induced Strain Engineering for High

Performance SOI FinFET utilizing Si (110) Channel for Both N and PMOSFETs. Electron Devices Meeting, 2006. IEDM '06. International, 1 - 4. 28.

Sekaric, L.; Gunawan, O.; Majumdar, A.; Liu, X. H.; Weinstein, D.; Sleight, J. W.

Applied Physics Letters 2009, 95, (2), 023113. 29.

Xie, R. et al., In A 7nm FinFET Technology Featuring EUV Patterning and Dual

Strained High Mobility Channels. Electron Devices Meeting (IEDM), 2016 IEEE International, 2.7.1 - 2.7.4.


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 22

30.

Ross, F. M.; Tersoff, J.; Reuter, M. C. Phys Rev Lett 2005, 95, (14), 146104.

31.

Xu, T.; Nys, J. P.; Addad, A.; Lebedev, O. I.; Urbieta, A.; Salhi, B.; Berthe, M.;

Grandidier, B.; Stiévenard, D. Physical Review B 2010, 81, (11). 32.

Schmidt, V.; McIntyre, P. C.; Gösele, U. Physical Review B 2008, 77, (23).

33.

Day, R. W.; Mankin, M. N.; Gao, R.; No, Y. S.; Kim, S. K.; Bell, D. C.; Park, H. G.;

Lieber, C. M. Nat Nanotechnol 2015, 10, (4), 345-352. 34.

Mooney, P. M.; Dacol, F. H.; Tsang, J. C.; Chu, J. O. Applied Physics Letters 1993, 62,

(17), 2069. 35.

Tsang, J. C.; Mooney, P. M.; Dacol, F.; Chu, J. O. Journal of Applied Physics 1994, 75,

(12), 8098. 36.

Zhang, X.; Lew, K.-K.; Nimmatoori, P.; Redwing, J. M.; Dickey, E. C. Nano Lett 2007, 7,

(10), 3241-3245. 37.

Jones, E. J.; Ermez, S.; Gradecak, S. Nano Lett 2015, 15, (12), 7873-7879.

38.

Van de Walle, C. G. Physical Review B 1989, 39, (3), 1871-1883.

39.

Friedel, P.; Hybertsen, M. S.; Schlüter, M. Physical Review B 1989, 39, (11), 7974-7977.

40.

Singh, R.; Dailey, E. J.; Drucker, J.; Menéndez, J. Journal of Applied Physics 2011, 110,

(12), 124305. 41.

Birch, F. Physical Review 1947, 71, (11), 809-824.

42.

Saitoh, M. et al., In Three-Dimensional Stress Engineering in FinFETs for Mobility/On-

Current Enhancement and Gate Current Reduction. VLSI Technology, 2008 Symposium on, 18 - 19. 43.

Ganesan, S.; Maradudin, A. A.; Oitmaa, J. Annals of Physics 1970, 56, (2), 556-594.


20

Page 21 of 22

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

44.

Anastassakis, E.; Pinczuk, A.; Burstein, E.; Pollak, F. H.; Cardona, M. Solid State

Communications 1970, 8, (2), 133-138. 45.

Loudon, R. Advances in Physics 2001, 50, (7), 813-864.

46.

Menéndez, J.; Singh, R.; Drucker, J. Annalen der Physik 2011, 523, (1-2), 145-156.

47.

Wang, J.; Gudiksen, M. S.; Duan, X.; Cui, Y.; Lieber, C. M. Science 2001, 293, (5534),

1455-1457. 48.

Qi, J.; Belcher, A. M.; White, J. M. Applied Physics Letters 2003, 82, (16), 2616.

49.

Reparaz, J. S.; Bernardi, A.; Goñi, A. R.; Alonso, M. I.; Garriga, M. Applied Physics

Letters 2008, 92, (8), 081909. 50.

Anastassakis, E.; Cantarero, A.; Cardona, M. Physical Review B 1990, 41, (11), 7529-

7535. 51.

Cerdeira, F.; Buchenauer, C. J.; Pollak, F. H.; Cardona, M. Physical Review B 1972, 5,

(2), 580-593. 52.

Etzelstorfer, T.; Wyss, A.; Süess, M. J.; Schlich, F. F.; Geiger, R.; Frigerio, J.; Stangl, J.

Measurement Science and Technology 2017, 28, (2), 025501. 53.

Sui, Z.; Burke, H. H.; Herman, I. P. Physical Review B 1993, 48, (4), 2162-2168.

54.

Rim, K. et al., In Strained Si NMOSFETs for high performance CMOS technology. VLSI

Technology, 2001. Digest of Technical Papers. 2001 Symposium on, 59 - 60. 55.

Irisawa, T.; Numata, T.; Tezuka, T.; Usuda, K.; Sugiyama, N.; Takagi, S.-I. IEEE

Transactions on Electron Devices 2008, 55, (2), 649-654. 56.

Cui, Y.; Lauhon, L. J.; Gudiksen, M. S.; Wang, J.; Lieber, C. M. Applied Physics Letters

2001, 78, (15), 2214. 57.

Schmidt, V.; Senz, S.; Gosele, U. Nano Lett 2005, 5, (5), 931-935.


21

Nano Letters

Page 22 of 22

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 ACS Paragon Plus Environment

22

Enhanced Electron Mobility in Nonplanar Tensile Strained Si

Recommend Documents