Epitaxial NixPd1âx (111) Alloy Substrates with Continuously Tunable

Subscriber access provided by University of Newcastle, Australia

Article x

1-x

Epitaxial NiPd (111) Alloy Substrates with Continuously Tunable Lattice Constants for 2D Materials Growth Gregory S. Hutchings, Jin-Hao Jhang, Chao Zhou, David Hynek, Udo D. Schwarz, and Eric I. Altman ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.7b01369 • Publication Date (Web): 10 Mar 2017 Downloaded from http://pubs.acs.org on March 12, 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.

ACS Applied Materials & Interfaces 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 20

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 Applied Materials & Interfaces

Epitaxial NixPd1-x (111) Alloy Substrates with Continuously Tunable Lattice Constants for 2D Materials Growth Gregory S. Hutchings,a,b Jin-Hao Jhang,a,b Chao Zhou,a,c David Hynek,a,b Udo D. Schwarz,a,b,c and Eric I. Altman*,a,b a

Center for Research on Interface Structures and Phenomena (CRISP), Yale University, New

Haven, Connecticut 06520, USA b

Department of Chemical and Environmental Engineering, Yale University, New Haven,

Connecticut 06520, USA c

Department of Mechanical Engineering and Materials Science, Yale University, New Haven,

Connecticut 06520, USA

ABSTRACT Epitaxial strain can be a powerful parameter for directing the growth of thin films. Unfortunately, conventional materials only offer discrete choices for setting the lattice strain. In this work, it is demonstrated that epitaxial growth of transition metal alloy solid solutions can provide thermally stable, high quality growth substrates with continuously tunable lattice

ACS Paragon Plus Environment

1


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 20

constants. Molecular beam epitaxy was used to grow NixPd1-x(111) alloy films with lattice constants between 3.61 and 3.89 Å on the hexagonal (0001) basal planes of α-Al2O3 and Cr2O3 (grown as epitaxial films on α-Al2O3 (0001)). The Cr2O3 acted as an adhesion layer, which not only improved the high temperature stability of the films but also produced single domain films with NixPd1-x 112 parallel to Cr2O3 [1120], in contrast to growth on α-Al2O3 that yielded twinned films.

Surface characterization by electron diffraction and scanning tunneling

microscopy (STM) as well as bulk x-ray diffraction analysis indicated that the films are suitable as inexpensive and stable substrates for thin film growth and for surface science studies. To demonstrate this suitability, bilayer SiO2, a two-dimensional van der Waals material, was grown on a NixPd1-x(111) film tuned to closely match the film’s lattice constant, with STM and electron diffraction results revealing a highly ordered, single phase crystalline state.

KEYWORDS molecular beam epitaxy, thin film growth, alloy, nickel, palladium, bilayer silica, twodimensional materials


2

Page 3 of 20

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. Introduction The careful selection of materials with specific lattice constants is critical for low-strain epitaxial growth of thin films. In particular, finding systems with a range of lattice constants suitable for novel two-dimensional (2D) van der Waals materials growth and surface science studies has been the subject of considerable recent work,1-6 but requires renewed attention as emerging materials are discovered. Where the intrinsic chemical reactivity of the surface is less important than the lattice strain, as is frequently the case for the growth of van der Waals materials,5, 7 the ability to tune the lattice constant over a continuous range is attractive to achieve large, defectfree domains over wide surface areas, or to tune the strain to drive the layers into metastable structures with unique properties.8-9 In this paper it will be shown that epitaxially-grown metal alloy solid-solution thin films can be engineered to meet this requirement while offering the advantages of low cost compared to single crystals and orientation control compared to polycrystalline surfaces. The NixPd1-x alloy system can be best described as a solid solution binary alloy over all temperatures below the melting point for Pd concentrations above 30%, with the remaining range accessible through precise quenching.10 Individually, both Ni and Pd (111) films have been synthesized on hexagonal substrates, and films with a narrower continuous lattice constant range have been synthesized with Ni/Cu alloys (3.52 to 3.61 Å); however, these surfaces tend to be significantly twinned, and the Ni/Cu system tends to phase separate throughout much of the composition range at moderate temperatures.11-15 With the Ni-Pd system, face-centered cubic unit cells with side lengths a between 3.52 Å and 3.89 Å may be accessed continuously, yielding (111) films with threefold symmetry and nearestneighbor distances between 2.49 Å and 2.75 Å. These ranges are ideal for matching either


3


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 20

directly or with compact supercells to several 2D materials under active development, such as bilayer SiO2 (5.30 Å repeat length),16-21 BN (2.50 Å),22 and WS2 (3.18 Å).23 Besides merely providing a surface lattice match to 2D materials, the continuously-tunable nature of the NixPd1-x alloy provides the opportunity to engineer the strain and potentially introduce new phases with interesting structural, electronic and chemical properties.8-9, 24 Here, we demonstrate the molecular beam epitaxy (MBE) growth of high quality, thermally stable, single domain, epitaxial NixPd1-x(111) alloy surfaces spanning a wide composition range by employing a Cr2O3(0001) adhesion layer on α-Al2O3(0001). Chromium oxide has previously been used to promote oxide epitaxy on sapphire and metal adhesion to nitride layers.25-26 It will be shown that the inclusion of a 150-200 Å Cr2O3(0001) layer allows for improved alignment and adhesion of the co-deposited alloy, eliminating twinning and decreasing the surface roughness. Through both UHV and ex situ characterization techniques, we have confirmed that the final films are pure, single-phase, and suitable for use as growth substrates. Furthermore, we have demonstrated the growth of highly crystalline bilayer SiO2 on a NixPd1-x alloy film tailored to the unstrained lattice constant of this 2D van der Waals material, providing an immediate application for the alloyed films and paving the way for precisely and systematically strain engineering 2D van der Waals materials.

2. Results and Discussion Synthesis and Characterization of NixPd1-x Alloy Films The NixPd1-x(111) films and Cr2O3(0001) adhesion layers were deposited through MBE on c-cut, polished α-Al2O3(0001) single crystals, as outlined in the Experimental Section. For verification of the final NixPd1-x alloy phases on both the α-Al2O3(0001) and Cr2O3(0001)/α-Al2O3(0001)


4

Page 5 of 20

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


substrates, X-ray diffraction (XRD) data were recorded ex situ and are presented in Figure 1a and 1b for a wide range of compositions. The alloy lattice constants were determined from the peak position of the NixPd1-x (111) reflection; (0006) reflections for the α-Al2O3(0001) substrate (41.69°) and Cr2O3(0001) layer (39.78°) overlap within the range of the NixPd1-x reflections, but did not interfere with the determination of alloy lattice constants. While the (111) peaks for films on α-Al2O3(0001) are noticeably broadened and suggest small amounts of phase segregation (Figure 1a), the films grown on Cr2O3(0001)/α-Al2O3(0001) (Figure 1b) are consistent with a single NixPd1-x alloy phase. Despite the inclusion of a 10 Å thick Ni layer at the interface for most of the films, there is no separate peak for Ni and it is assumed that the Ni incorporated into the co-deposited alloy. Figure 1c shows X-ray reflectivity (XRR) data for the films grown on Cr2O3(0001)/α-Al2O3(0001); the results are consistent with alloy film thicknesses between 400 and 600 Å and Cr2O3 between 100 and 200 Å thick as expected from quartz crystal microbalance (QCM) measurement. Moderate attenuation of the higher-frequency oscillations is consistent with low interfacial roughness (~5 Å from fits of the decay rates for both the Cr2O3(0001)/αAl2O3(0001) and NixPd1-x (111)/Cr2O3(0001) interfaces). The XRR of the Cr2O3(0001)/αAl2O3(0001) substrate prior to deposition of the a = 3.72 Å sample is also presented for comparison, further verifying the thickness range and smoothness of the growth substrate prior to metal deposition.


5


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 20

Figure 1. XRD and XRR data for the as-deposited NixPd1-x alloy films, plotted on a logarithmic scale. (a) XRD for films grown on α-Al2O3(0001), (b) XRD for films grown on Cr2O3(0001)/αAl2O3(0001), and (c) XRR for films grown on Cr2O3(0001)/α-Al2O3(0001) and the Cr2O3 layer prior to metal deposition. *: (0006) reflection of α-Al2O3; #: (0006) reflection of Cr2O3.

By fitting a collection of bulk NixPd1-x alloy data from previous work,27 we find that the lattice constant of the alloy may be represented with a quadratic approximation of Vegard’s law, 1 1 , with = 0.1251 Å (R2 = 0.9995). This expression is plotted as the solid curve in Figure 2a, alongside the experimental Pd concentrations as determined by peak-to-peak Auger electron spectroscopy (AES) intensity ratio measurements and tabulated sensitivity factors (1.5 for Pd MNN, 0.28 for Ni LMM)28 vs. the lattice constant from XRD. Interestingly, we find that the surface Pd concentration is consistently higher than that expected from the lattice constant. A likely explanation is surface enrichment in Pd, which may reflect preferential Ni sputtering during the sample cleaning process before the


6

Page 7 of 20

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


AES measurement or a thermodynamic preference for Pd segregation to the surface due to lower surface energy.29 The Ni concentration measured by AES was typically within 5% of the deposition target, which may be improved through a more rigorous, iterative QCM calibration. Full AES spectra are also provided for the films grown on Cr2O3(0001)/α-Al2O3(0001) in Figure 2b; pure Ni and Pd are observed with no substrate contributions of Al, Cr, or O.

Figure 2. (a) Film lattice constant as determined through XRD vs. calculated Pd concentration from AES peak-to-peak ratios. Fit to bulk data is the quadratic approximation of Vegard’s law based on prior bulk data.27 (b) Full AES derivative spectra for the NixPd1-x films grown on Cr2O3(0001)/α-Al2O3(0001). Data above 350 eV is multiplied by a factor of 3.

The high quality of the NixPd1-x alloy surfaces is apparent in the low-energy electron diffraction (LEED) patterns shown in Figure 3a, with sharp spots and low background comparable to single crystal data. There is no evidence of small domains or a distribution of lattice constants at the surface, which would cause broadening of the spots. Patterns recorded for the films grown on α-Al2O3(0001) show rings of 12 spots indicating twinning; this phenomenon


7


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 20

has been observed in previous studies of epitaxial Ni(111) growth on α-Al2O3(0001) when no adhesion layer was employed.11 With the Cr2O3 layer, only a single domain is observed across the entire sample area. Furthermore, the observed single-domain LEED remains consistent to at least 975 K in UHV over multiple annealing cycles; therefore, the films are stable to high temperatures, which is critical for further use as tunable-strain substrates in epitaxial growth. As shown in Figure 3b, reflection high-energy electron diffraction (RHEED) patterns recorded at endpoints during the deposition of the a = 3.72 Å NixPd1-x film on Cr2O3(0001)/α-Al2O3(0001) also indicate high surface ordering. While a clear RHEED pattern with long streaks is observed after 10 Å Ni growth, the pattern at the final thickness is significantly attenuated, suggesting a rougher, more poorly ordered surface. The sharpness of the LEED patterns in Figure 3a indicates that annealing eliminates this roughness.

Figure 3. (a) Representative LEED patterns for annealed NixPd1-x films grown on both αAl2O3(0001) and Cr2O3(0001)/α-Al2O3(0001) substrates. (b) RHEED patterns along the 0110


8

Page 9 of 20

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


direction recorded at deposition endpoints for a NixPd1-x film with a = 3.72 Å grown on Cr2O3(0001)/α-Al2O3(0001).

The RHEED data in Figure 3b were recorded with the incident beam parallel to the α-Al2O3 0110 direction. The number of spots and streaks did not change as the different films were added; however, their spacing slightly increased when Cr2O3 was added, then decreased after adding Ni, then increased again after growing the alloy. This sequence is consistent with NixPd1 x112

∥ Ni112 ∥ Cr2O31120 ∥ α-Al2O31120 and the lattice constants of these species.

This relationship aligns the close-packed direction of the alloy surface with the close-packed direction of the corundum oxygen sublattice, yielding a lattice mismatch between 14.9% to 4% tensile for NixPd1-x (O-O distance on Cr2O3 is 2.86 Å, and 2.75 Å on α-Al2O3). A schematic representation of the epitaxial alignment is given in Figure 4a, where the positions of Ni (or Pd) atoms are rotated 30° relative to the (0001) substrate unit cell. Note that this orientation can place the metal atoms at the interface above similar high symmetry oxygen sites, assumed to be threefold hollows in the schematic. The 30° separation observed between twins on α-Al2O3(0001) implies the existence of domains with NixPd1-x110 ∥ α-Al2O31120 (0° rotation between the film and substrate unit cells), as indicated in Figure 4b. By necessity, such a relationship puts every other metal atom in very different sites with respect to the oxygen sublattice, threefold hollow and roughly on-top in Figure 4b. Interestingly, this alignment reduces the lattice mismatch between twice the Ni lattice constant and α-Al2O3(0001) to 3.9% (tensile) and to less than 0.5% for Cr2O3(0001). The preference to uniquely align the Ni close-packed direction parallel to the close-packed direction of the oxygen sublattice in Cr2O3 despite this orientation creating a much larger mismatch


9


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 20

implies O-Ni bonding at the interface plays a decisive role in setting the alignment for Cr2O3. Thus, the more reactive Cr2O3 surface both eliminates twins and improves adhesion.

Figure 4.

Representation of the epitaxial relationship between the α-Al2O3(0001) or

Cr2O3(0001) surfaces and deposited NixPd1-x (111) with rotations relative to the substrate unit cell of (a) 30° (112 ∥ 1120, green dotted lines) or (b) 0° (110 ∥ 1120, yellow dotted lines). Black dotted lines indicate the surface unit cell for the substrate. The red balls represent O, and the blue balls either Al or Cr.

Scanning tunneling microscopy (STM) images representative of films on α-Al2O3 and Cr2O3 substrates are presented in Figure 5. Steps separating terraces on the order of hundreds to thousands of Å wide are clearly visible, further indicating the suitability of these films for epitaxial growth and surface science studies. The terraces appear somewhat rougher for the film grown on α-Al2O3(0001) (Figure 5a), which may reflect adsorbate accumulation on the surface. Nonetheless, concurrent atomic force microscopy (AFM) measurements confirm that the RMS roughness is substantially reduced when the Cr2O3 layer is included, e.g., from 15.40 Å to 4.53 Å over a 10 µm by 10 µm scan window on the two films with a = 3.72 Å.


10

Page 11 of 20

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


Figure 5.

Representative STM images of (a) NixPd1-x alloy film with a = 3.68 Å on α-

Al2O3(0001), and (b) NixPd1-x alloy film with a = 3.80 Å on Cr2O3(0001)/α-Al2O3(0001). Both images were recorded at 0.75 V sample bias.

Example - Growth of Zero-Strain 2D Bilayer SiO2 To demonstrate the suitability of these alloy substrates for 2D materials growth, bilayer SiO2 was grown on a NixPd1-x alloy film on Cr2O3(0001)/α-Al2O3(0001) following previously published procedures.20 Prior work on bilayer silica has revealed coexistence of amorphous and crystalline phases, with the distribution between the two phases and the defect density in the crystalline phase both sensitive to the lattice constant of the metal substrate.17-21,

30

While entirely

amorphous films have been prepared on Pt(111) where the lattice mismatch is 4.7% (tensile),30 it has proven difficult to prepare solely crystalline films with long-range order.17-21 On hexagonal substrates, crystalline SiO2 bilayers form (2 × 2) overlayers.31 The repeat length of the unstrained bilayer is 5.30 Å16,

24

and so a (111) substrate lattice constant of 2.65 Å was targeted,

corresponding to 52% Pd based on the fit to Vegard’s law. The film produced had a slightly


11


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 20

smaller lattice constant of a = 3.72 Å, which results in a nearest-neighbor distance (on the (111) plane of 2.63 Å, thus putting the SiO2 bilayer under slight compression. After SiO2 deposition and annealing to 1025 K, the quality of the bilayer film was assessed with LEED (Figures 6a and 6b). The simple (2 × 2) pattern is consistent with growth of a single domain crystalline bilayer on a hexagonal metal surface.31 The patterns show no evidence of a ring associated with the amorphous phase, and the sharpness of the SiO2 spots indicates long-range order and large domains. Additionally, there is no evidence of secondary phases, suggesting that the alloy surface was not significantly damaged during high-temperature annealing under O2. Concurrent measurement with AES showed clear Si and O peaks, along with attenuation of Pd and Ni signals as expected for uniform bilayer deposition.

Figure 6. Characterization data for a SiO2 bilayer film grown on a NixPd1-x alloy film with a = 3.72 Å on Cr2O3(0001)/α-Al2O3(0001). LEED recorded at (a) 60 eV and (b) 130 eV. (c) Representative STM image at 2.35 V sample bias (unoccupied states). The gray scale has been cycled through twice so that atomic scale features could be seen on the two terraces separated by the step edge that runs vertically through the image.


12

Page 13 of 20

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


Further confirmation of the quality of the crystalline 2D layer is provided by STM (Figure 6c). The orientation of the bilayer relative to the step edge (steps run predominantly along [110] on fcc (111) surfaces) is as expected for commensurate growth of the film. The honeycomb appearance is in accord with prior STM measurements on crystalline 2D SiO2 bilayers19-20, 31-32 and the structure of the crystalline phase (mirror image planes of six-membered rings of cornersharing tetrahedra).16,

18-20, 31

While point defects are still visible, only a single phase and

orientation can be seen across the image.

3. Conclusions We have demonstrated the successful growth of epitaxial NixPd1-x(111) alloy films with continuously tunable lattice constants. A collection of UHV and ex situ surface, bulk and interface analysis techniques revealed the high quality of these films and surface characteristics at least comparable to the surfaces of bulk single crystals. Through use of a Cr2O3(0001) adhesion layer, single domain films with high temperature stability were synthesized with lattice constants ranging from 3.61 to 3.89 Å. The improvement engendered through use of a more reactive oxide adhesion layer highlights the role of interfacial interactions in directing and stabilizing epitaxial metal growth. Although the films were deposited using MBE, both the Cr2O3 and alloy films could be readily deposited using more cost-effective sputtering methods onto commercially available sapphire wafers. The resulting inexpensive single crystal alloy films hold the potential to be ideal, stable substrates for 2D van der Waals materials growth and surface science studies; the former was demonstrated through the growth of bilayer SiO2. By tuning the alloy surface lattice constant to match that of the freestanding bilayer, a highly crystalline 2D material was obtained. The lack of strain help eliminate the amorphous 2D silica phase often


13


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 20

seen when the material is grown on pure metal substrates that impart tensile strain, showing the potential of using substrate alloying to manipulate 2D materials growth. 4. Experimental Section For all films, polished α-Al2O3(0001) growth substrates (>99.99%, +/- 0.5° tolerance, MTI) were first ultrasonically cleaned in isopropanol and deionized water, then annealed in air at 1075 K for 12 h. The substrates were then loaded into a UHV chamber configured for MBE with

Epitaxial NixPd1âx (111) Alloy Substrates with Continuously Tunable

Recommend Documents

Epitaxial NixPd1âx (111) Alloy Substrates with Continuously Tunable