University of Groningen
Growth and nanostructure of tellurides for optoelectronic, thermoelectric and phase-change
applications
Vermeulen, Paul Alexander
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date: 2019
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Vermeulen, P. A. (2019). Growth and nanostructure of tellurides for optoelectronic, thermoelectric and phase-change applications.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.
Growth and nanostructure of tellurides
for optoelectronic, thermoelectric and phase-change applications
Zernike Institute PhD thesis series
2019-08
ISSN:1570-1530
ISBN (printed):
978-94-034-1422-5
ISBN (electronic):
978-94-034-1421-8
The work presented in this thesis was performed in the
Nanostructured Materials and Interfaces group at the Zernike
Institute for Advanced Materials of the University of Groningen,
The Netherlands.
Cover Design by Paul Vermeulen
Printed by Gildeprint
© Paul Vermeulen 2019
Cover Images:Front: Electron diffraction pattern of a mica surface, obtained using Reflective High-Energy Electron Diffraction (RHEED).
Back: Crystal structure models of Bi2Te3, WTe2 and Sb2Te3, transmission electron microscopy image of herringbone structure of GeTe.
Growth and nanostructure of tellurides
for optoelectronic, thermoelectric and
phase-change applications
PhD thesis
to obtain the degree of PhD at the
University of Groningen
on the authority of the
Rector Magnificus prof. E. Sterken
and in accordance with
the decision by the College of Deans.
This thesis will be defended in public on
Friday 26 April 2019 at 11.00 hours
by
Paul Alexander Vermeulen
born on 4 September 1991
in Pasadena, United States of America
Supervisor
Prof. B.J. Kooi
Co-supervisor
Dr. G.R. Blake
Assessment Committee
Prof. B. Noheda Pinuaga
Prof. G. Koster
Prof. F. Rao
*Chapters 4 through 9 have been (submitted for) publication in peer-reviewed journals.
Table of contents*
Chapter 1. Introduction – Why study tellurides? ... 1
Chapter 2. Experimental Methods. ... 13
Chapter 3. Pulsed laser deposition of tellurides. ... 31
Chapter 4. Pulsed laser deposition of epitaxial tungsten-telluride heterostructures. ... 45
Chapter 5. Strain engineering of van der Waals heterostructures. ... 59
Chapter 6. Multilevel reflectivity switching of ultrathin phase change films. ... 83
Chapter 7. Unravelling the domain structures in GeTe and LaAlO3. ... 103
Chapter 8. Reversible amorphous-crystalline phase changes in a wide range of Se1-xTex alloys studied using ultrafast differential scanning calorimetry. .... 127
Chapter 9. Combining ultrafast calorimetry and electron microscopy: Reversible phase transformations in SeTeAs alloys. ... 165
Chapter 10. English summary ... 185
Chapter 11.
Nederlandse samenvatting ... 189
Chapter 12. Nederlandse samenvatting voor niet-natuurkundigen ... 195
List of publications ... 199
1
Image: Mount Everest (8848 m, Black pyramid), Nuptse (7861 m, right), and the Khumbu Icefall (center) taken from Kala Patthar (5645 m) by the author.
Chapter 1. Introduction – Why study
tellurides?
Crystal structures, open questions and applications of bulk and thin-film tellurides are discussed. The quest for artificial nanostructuring in thin films is introduced. Finally, an outline of the thesis is given.
1.1
Why?
Any question of ‘why’ may be asked and answered on several levels. This chapter will attempt to answer the titular question in a few more and less profound ways. Obviously we describe the wide range of promising applications of tellurides, which may positively affect the people and world we live in. We introduce some mysteries any curious scientist would like to start solving immediately. We also show the fascinating structures they form, as well as synthesis equipment allowing us to play with their building blocks and satisfy our inner child. These answers however, while valid and true, seem less appropriate than the answer George Mallory gave in 1923, when asked why he would lead an expedition up the thus far unclimbed Mount Everest: “Because it’s there”.
Chapter 1. Introduction – Why study tellurides?
1.2
Crystal Structure and epitaxy
The crystal structure of the group V2-VI3 tellurides (e.g. Bi2Te3) is highly
anisotropic, as can be seen in figure 1. In fact, while the bonding within a so-called quintuple-layer block (|Te-Bi-Te-Bi-Te|) is relatively strong since it is based on convalent bonds adjacent Te layers are only weakly bound by molecular or van der Waals (vdW) forces. Recently it has been argued that these tellurides contain a special type of bonding called ‘metavalent’.1 Although not as anisotropic as V2-VI3
materials, IV-VI materials (e.g. GeTe) tend towards a Peierls-like distortion from the rocksalt structure which gives rise to the formation of IV-VI bilayers (|Ge-Te|) along the hexagonal [0001] (rhombohedral/cubic [111]) axis.2 Furthermore, due to
the thermodynamically favourable mixing and identical (0001) surface symmetry of GeTe and Sb2Te3/Bi2Te3, IVx-V2y-VIx+1.5y structures containing vdWaals gaps can
be grown, of which, as an example, Ge2Sb2Te5 is depicted in the Kooi stacking in
figure 1.3,4 These GST/GBT-materials are widely used in several technologies, which
will be discussed in the following sections. All the above mentioned telluride crystals in their stable states belong to the trigonal crystal system and can be described by rhombohedral or hexagonal lattices.
Tungsten-telluride belong to a large class of materials called the Transition-Metal Di-Chalcogenides (TMDCs), which consist of triplet-layers |Te-W-Te| separated by vdWaals gaps. Several structural modifications of WTe2 may be
stabilized, but when viewed along the 0001 (monoclinic 001)axis, all possess a (distorted) trigonal surface symmetry similar to that of the V2-VI3 materials. The
use of crystalline TMDCs is currently mostly limited to fundamental research applications.
This anisotropy and weak interlayer bonding of these compounds is favourable for the synthesis, or growth of these ‘2D’ structures. Koma first introduced the concept of vdW epitaxy in 1991.5 Due to the weak interlayer bonding, he showed
highly dissimilar materials could be stacked into a heterostructure with atomically sharp interfaces. Even more conveniently, the materials prefer to grow epitaxially, without generating lattice-matching strain. In 2014 Novoselov and Geim6,7 added
to this idea by envisioning heterostructures of graphene, TMDCs and the aforementioned group V-VI materials to tune functional properties.6,7
As we will show in this thesis, the concept of vdW epitaxy is a very useful one, allowing facile growth of heterostructures of various (quasi-)2D-bonded materials. This strong/weak bonding anisotropy also seems to give rise to several extremely useful functional properties, which we will discuss in the following sections.8
1.2 Crystal Structure and epitaxy
3
Figure 1. The crystal structures of several tellurides: Bi2Te3, Sb2Te3 (R-3m), GeTe
(R3m) and WTe2 (P21m). While Bi2Te3, Sb2Te3 and WTe2 consist of quintuple or
triplet layers separated by vdW gaps, a more subtle bond anisotropy is present in GeTe which causes alternatingly short and long Ge-Te bonds. This will generally lead to domain formation for films thicker than a few monolayers.9 The surfaces of all these tellurides can generally be treated as having hexagonal (sixfold) symmetry, which becomes apparent from the Bi2Te3 and WTe2 structures viewed
along the [0001] crystal axis. The golden Te atoms reside at the vdWaals gap and form a hexagonally symmetric shape. Each Te atom is bound to three bismuth/tungsten-atoms however, reducing the symmetry to threefold. For 1T´-WTe2 the symmetry is further reduced by a distortion along the [100] axis,
creating alternating long- and short Te distances. GeTe and Sb2Te3 easily mix and
form Ge2Sb2Te5, hereshown in the Kooi-structure,3,10 where the material forms
nontuple blocks, with a preference for Sb to occupy the outer sites and Ge the inner sites.3,11,12
Chapter 1. Introduction – Why study tellurides?
1.3
Phase change materials
With the continuing exponential increase in data storage and computer memory requirements, ever smaller and more energy-efficient devices are needed. The GeSbTealloys constitute the functional component in rewritable optical discs (CD, DVD, Blu-Ray).13 More recently they have been incorporated as on-chip memory
elements which unlike some other popular memory elements, like DRAM and SRAM, persist after the power is removed.14,15 These cells have been manufactured
to sizes around 50 nm, with their smallest dimension approaching several nanometers.16,17 The working principle of these memory cells, or bits, is equivalent
in both optical and electrical applications. The phase-change material (PCM) is switched between a higher- and lower- electrically conducting or reflecting states, which represent the ‘0’ and ‘1’ bit states. This switch is accomplished using a melt-quench procedure. Using a weak light- (optical disk) or electric- (memory) heating pulse, the sample will switch from the amorphous to the crystalline phase. By applying a stronger pulse, the crystal film melts, and due to the high cooling rate within the cold surrounding, returns to the amorphous (glassy) state. The material can be cycled electrically about a billion times before degradation causes malfunction.18
The classical PCMs consisted of ‘bulk-like’ alloys. However using nanostructuring to engineer so-called interfacial phase-change materials (iPCM), it was shown that switching energy and time could be significantly reduced.19 These
materials consist of a superlattice of epitaxially grown Ge2Sb2Te5 and Sb2Te3 unit
cell blocks, and seem to achieve both the high-and low resistance states within the crystalline phase. The exact nature of the switching mechanism and chemical bonding within these superlattices is still hotly debated.3,20
It seems clear from many reports that engineering the exact texture and structure of a thin nanostructured film may provide tremendous improvement of functional properties. The use of a crystalline template to enhance crystallization rate,21 as well as the use of strained multilayers,22,23 where a GeTe layer is
sandwiched between two Sb2Te3 layers, yield improved switching characteristics.
1.4
Thermoelectrics
The field of thermoelectrics (TEs) aims to optimize the Seebeck effect to harvest electrical current from heat gradients or, vice versa, using the Peltier effect to cool an area using electrical current.24,25 The energy-generating Seebeck effect has
applications in any situation where regular power sources are unavailable, or no moving parts can be used, such as military or aerospace applications.26 Waste heat
sources such as steel manufacturing plants and solar arrays may also be sources of energy by using TE devices.27 Peltier coolers may be applied for on-chip or in
1.5 Optoelectronics
5
or effectively.28 While TE is an extremely large field, the Bi2Te3 and GeTe based
TAGS (TeAgGeSb) families are among the most successful TE materials, and are still studied extensively.29,30
The efficiency of a TE device is given by its figure of merit:
2
S
ZT
T
(1.1)
Where σ is the electrical conductivity, S the seebeck coefficient (in V/K) and
the thermal conductivity in (W/mK).25 The figure of merit immediately shows thedilemma of TE: optimum ZT is reached for a phonon glass (low
) and electron crystal (high
) or PGEC. In other words, one has to find a structure which conducts electrons extremely well, but blocks phonons (and high-energy electrons).25 The Seebeck coefficient may be optimized by increasing the amount ofadditional electrons that become available when the temperature is increased: the Fermi level should be properly situated with respect to the band gap: a metal is therefore unfavorable.
As we discussed for PCMs, in TEs nanostructuring provides ways for improvement. Promising material systems are prepared using various dopants to improve the band structure, different heat-treatments are used to obtain various crystal structures and grain sizes in an attempt to reduce thermal conductivity.31,32
An early report on heterostructures of Sb2Te3 and Bi2Te3 claiming a strong
reduction in
was the inspiration for much more research into ordered multilayers continuing to this day.33The similarities between using tellurides for TEs and PCMs have been pointed out in the literature.34 Both profit from low thermal conductivity (increasing energy
efficiency), and both applications seem to show high performance when the materials involved possess a so-called metavalent bonding state, which can be broken by the thermal gradient or amorphisation pulse in the system.1,35
Furthermore, the expected lower thermal conductivity along the [0001] axis may be exploited in nanostructures and superlattices.
1.5
Optoelectronics
Bulk optical fibers
Optical sensor and communication fibers require the use of highly homogeneous bulk glassy materials. Traditionally selenium-based materials have been used due to the ease of obtaining the glassy state. More recently, attention has shifted to include the heavier tellurium which has a broader frequency transmission window but is harder to amorphise.36,37 Tellurium alloys can be
Chapter 1. Introduction – Why study tellurides?
increases the glass forming ability. Although the optical and thermal properties of these alloys have been investigated frequently,38 little is known concerning the
actual microstructure and crystallization mechanism of these alloys, even in bulk.39
Thin-film sensor and display
While phase-change optical disc storage constitutes a mature technology, and memory devices are industrially produced, active nanostructured optical systems such as displays and communication elements based on GST-like materials are still in the research phase. Due to the switchable nature of these materials, they are excellent candidates to be used as optical sensors, phase array antennas, display pixels, sub-wavelength optical lenses, and (anti-)reflective coatings. This requires development of tunable small-scale elements which possess high contrast, low noise, and are structurally stable under repeated switching. 40–43
TMDC-based optoelectronics
TMDC mono- or bilayers can show a variety of interesting behavior and properties: several of them are direct bandgap semiconductors, making them more suitable than graphene to use as transistors or sensors.44,45 They show high optical
absorbance even at monolayer thickness, which is extremely suitable for incorporation within photovoltaic applications. This interaction is improved using graphene/TMDC heterostructures.46,47 Furthermore, the thin-film structure is
stable against bending, and strain may even improve functional properties.48 Major
challenges include large-area deposition and long-term stability.
1.6
Pure nanoscale effects
Many nanometer scale research efforts, both theory and experimental work, focus on effects only observable in thin films approaching monolayer thickness. Topological Insulators (TI),49–51 Weyl semimetals,52 and effects like quantum
spin-Hall effect,53,54 and superconductivity55 are still poorly understood, but have been
found for both the discussed group IV-VI and TMDC materials. Root cause of many of these effects is the discrepancy between electric potential at the material’s edge and within the bulk: hence the need for extremely thin films or even monolayers. By isolating an extremely thin sample, edge states dominate properties. Especially for TMDCs, a strong discrepancy between mono-and bilayer films is reported due to a change in structure. Also for GeTe, a similar instability seems present.56 The
investigation of these effects and states requires extremely high-quality samples in terms of stoichiometry and thickness uniformity: these nanoscale phenomena rely on the absence of doping and concomitant impurity carriers. For this reason, the use of bulk or exfoliated crystals was usually preferred, but thin-film deposition techniques are rapidly improving.57
1.7 Synthesis
7
1.7
Synthesis
All mentioned fields of application for tellurides require thin films to investigate nanoscale effects. As will become apparent in chapters two and three, this challenge is twofold: (1) one has to synthesize these films, and (2) then efficiently characterize them, neither of which is trivial at the nanoscale. For proper optimization, a quick back-and-forth interaction between both steps is essential.
The aim of this thesis was to grow telluride films using Pulsed Laser Deposition (PLD). While the technique is quite mature for the growth of oxides, other material classes have not been studied extensively. Specifically within the tellurides, the mature deposition methods include Molecular Beam Epitaxy (MBE) and sputtering. PLD promises to have several advantages over those techniques. Firstly, films generally adopt the target stoichiometry. Opposed to MBE, switching and adding ablation targets is much easier than exchanging or adding Knudsen cells. Furthermore, large-area PLD systems are already commercially available.58 Finally,
due to the PLD material versatility, targets can be oxide capped within the system. Compared to sputtering, the pulsed laser system and in-situ control through RHEED allow for more precise (monolayer) thickness control and diagnostics. The system therefore represents an ideal middle ground between control, accuracy, speed and versatility.
1.8
Summary
We have introduced the crystal structure of several telluride vdWaals materials, which will feature in the following chapters. Furthermore, we shortly introduced the synthesis technique to make thin films. Several promising areas of investigation for telluride thin films, with a strong tie to functional applications have been introduced. These applications served as our guide in optimizing our thin film growth. Finally, we shortly outlined the deposition technique used to grow telluride thin films.
1.9
Outline of this thesis
The thesis can be divided into two parts: artificial nanostructures grown using PLD (chapters 3-6) and natural nanostructures (chapters 7-9).
Chapter 2 introduces the experimental techniques used in this thesis. Special
attention is given to the pulsed laser deposition (PLD) technique and the attached reflective high-energy electron diffraction (RHEED) system.
Chapter 1. Introduction – Why study tellurides?
Chapter 3 is devoted to experimental results concerning the growth of
tellurides using PLD. The results mainly show the effect of varying different process parameters on the layer growth, and may be taken as a ‘tuning in’ of the system to optimum conditions.
Chapter 4 describes the growth optimization of WTe2 using PLD. This
challenging growth required single crystal targets and a thin seed layer of Bi2Te3.
We achieved low-temperature single-crystal-like growth of epitaxial heterostructures, and monolayer thickness control.
Chapter 5 elaborates on the growth of multilayered (Sb/Bi)2Te3 – GeTe films.
Using RHEED we observed an unexpected and persistent long-range strain gradient within the films, which denies the paradigm of true vdWaals epitaxy.
Chapter 6 shows the optical properties of PLD grown films. We present
multilayered devices which can be switched to obtain various reflectivity profiles.
Chapter 7 gives an analysis of the domain structures observed in GeTe and
LaAlO3. They possess a similar crystallographic symmetry: a rhombohedrally
distorted cubic structure. This gives rise to a herringbone domain structure, which alters the local crystal domain orientation, and might contribute to the high thermoelectric performance of GeTe. We proposed a crystallographic domain model for this herringbone structure, which is supported by TEM, EBSD, and optical microscopy, and is expected to be present in all structures based on a rhombohedral distortion of an initial cubic crystal structure.
Chapter 8 switches topics to the thermal analysis of reversible crystallization
within the SeTe alloy system. We reveal an extended phase diagram for Se-Te as derived from Ultrafast Differential Scanning Calorimetry (UFDSC) analysis, and show that the alloys behave non-Arrhenian as a fragile (undercooled) liquid at high heating rates.
Chapter 9 deals with the analysis of the Se1-xTexAs10 alloys, which form an
intricate two-phase structure upon crystallization. The alloy was thermally analyzed similarly to the SeTe alloy in chapter 7. We also report a novel method developed to transfer samples from UFDSC to TEM, which allows microstructural analysis. We show that the combination of both methods is a powerful tool for structural and thermal analyses.
When a chapter has been published in a peer-reviewed journal, this is mentioned on the first page of the chapter. A full list of publications can be found at the end of this thesis.
1.10 Literature
9
1.10
Literature
1. WUTTIG,M.,DERINGER,V.L.,GONZE,X.,BICHARA,C.&RATY,J.INCIPIENT METALS :FUNCTIONAL MATERIALS WITH A UNIQUE BONDING MECHANISM.
2. LITTLEWOOD,P.B.THE CRYSTAL STRUCTURE OF IV-VI COMPOUNDS.I.CLASSIFICATION AND DESCRIPTION.J.PHYS.CSOLID STATE PHYS.13,4855(1980).
3. MOMAND,J. ET AL.INTERFACE FORMATION OF TWO- AND THREE-DIMENSIONALLY BONDED MATERIALS IN THE CASE OF GETE–SB 2TE 3 SUPERLATTICES.NANOSCALE (2015). DOI:10.1039/C5NR04530D
4. MOMAND,J.STRUCTURE AND RECONFIGURATION OF EPITAXIAL GETE/SB2TE3 SUPERLATTICES. (2017).
5. KOMA,A.,UENO,K.&SAIKI,K.HETEROEPITAXIAL GROWTH BY VAN DER WAALS INTERACTION IN ONE-, TWO- AND THREE-DIMENSIONAL MATERIALS.J.CRYST.GROWTH 111,1029–1032(1991). 6. GEIM, A K.&GRIGORIEVA,I.V.VAN DER WAALS HETEROSTRUCTURES.NATURE 499,419–25
(2013).
7. NOVOSELOV,K.S.,MISHCHENKO,A.,CARVALHO,A.,NETO,A.H.C.&ROAD,O.2D MATERIALS AND VAN DER WAALS HETEROSTRUCTURES.SCIENCE (80-.).353,461(2016).
8. SHPORTKO,K. ET AL.RESONANT BONDING IN CRYSTALLINE PHASE-CHANGE MATERIALS.NAT. MATER.7,653–658(2008).
9. WANG,R. ET AL.FORMATION OF RESONANT BONDING DURING GROWTH OF ULTRATHIN GETE FILMS. NPGASIA MATER.9, E396(2017).
10. KOOI,B.J.&DE HOSSON,J.T.M.ELECTRON DIFFRACTION AND HIGH-RESOLUTION TRANSMISSION ELECTRON MICROSCOPY OF THE HIGH TEMPERATURE CRYSTAL STRUCTURES OF GE[SUB X]SB[SUB 2]TE[SUB 3+X](X=1,2,3) PHASE CHANGE MATERIAL.J.APPL.PHYS.92,3584(2002). 11. WANG,R.,BRAGAGLIA,V.,BOSCHKER,J.E.&CALARCO,R.INTERMIXING DURING EPITAXIAL
GROWTH OF VAN DER WAALS BONDED NOMINAL GETE/SB2TE3SUPERLATTICES.CRYST.GROWTH DES.16,3596–3601(2016).
12. MOMAND,J. ET AL.ATOMIC STACKING AND VAN-DER-WAALS BONDING IN GETE–SB2TE3 SUPERLATTICES.J.MATER.RES.1–10(2016). DOI:10.1557/JMR.2016.334
13. WUTTIG,M.&YAMADA,N.PHASE-CHANGE MATERIALS FOR REWRITEABLE DATA STORAGE.NAT. MATER.6,824–32(2007).
14. RAOUX,S.,WEŁNIC,W.&IELMINI,D.PHASE CHANGE MATERIALS AND THEIR APPLICATION TO NONVOLATILE MEMORIES.CHEM.REV.110,240–267(2010).
15. BURR,G.W. ET AL.PHASE CHANGE MEMORY TECHNOLOGY.223,(2010).
16. SIMPSON,R.E. ET AL.TOWARD THE ULTIMATE LIMIT OF PHASE CHANGE IN GE2SB2TE5.NANO LETT. 10,414–419(2010).
17. SALINGA,M. ET AL.MONATOMIC PHASE CHANGE MEMORY.NAT.MATER.17,681–685(2018). 18. OOSTHOEK,J.L.M. ET AL.EVOLUTION OF CELL RESISTANCE, THRESHOLD VOLTAGE AND
CRYSTALLIZATION TEMPERATURE DURING CYCLING OF LINE-CELL PHASE-CHANGE RANDOM ACCESS MEMORY.J.APPL.PHYS.110,(2011).
19. SIMPSON,R.E. ET AL.INTERFACIAL PHASE-CHANGE MEMORY.NAT.NANOTECHNOL.6,501–505 (2011).
20. YU,X.&ROBERTSON,J.MODELING OF SWITCHING MECHANISM IN GESBTE CHALCOGENIDE SUPERLATTICES.SCI.REP.5,1–8(2015).
21. SIMPSON,R.E.,FONS,P.,KOLOBOV,A.V.,KRBAL,M.&TOMINAGA,J.ENHANCED
CRYSTALLIZATION OF GETE FROM AN SB 2TE 3 TEMPLATE.APPL.PHYS.LETT.100,3–7(2012). 22. KALIKKA,J. ET AL.STRAIN-ENGINEERED DIFFUSIVE ATOMIC SWITCHING IN TWO-DIMENSIONAL
CRYSTALS.NAT.COMMUN.7,11983(2016).
23. ZHOU,X. ET AL.PHASE-CHANGE MEMORY MATERIALS BY DESIGN:ASTRAIN ENGINEERING APPROACH.ADV.MATER.28,3007–3016(2016).
24. BELL,L.E.COOLING,HEATING,GENERATING POWER, AND RECOVERING WASTE HEAT WITH THERMOELECTRIC SYSTEMS.SCIENCE (80-.).321,1457–1461(2008).
25. ROWE,D.CRC HANDBOOK OF THERMOELECTRICS.(1995). AT <HTTP://BOOKS.GOOGLE.COM/BOOKS?HL=EN&LR=&ID=CRTJC
-LUHLEC&OI=FND&PG=PA1&DQ=CRC+HANDBOOK+ON+THERMOELECTRICS&OTS =2I-4H6XL7A&SIG=CRPGO1U-VX6_QG4FPJOYPXQY6X4>
26. RINEHART,G.H.DESIGN CHARACTERISTICS AND FABRICATION OF RADIOISOTOPE HEAT SOURCES FOR SPACE MISSIONS.PROG.NUCL.ENERGY 39,305–319(2001).
Chapter 1. Introduction – Why study tellurides?
MANAGE.REV.55,149–160(2012).
28. JENKINSON,I.R.&WYATT,T.DOES BLOOM PHYTOPLANKTON MANAGE THE PHYSICAL
OCEANOGRAPHIC ENVIRONMENT?LASSUS,P.,ARZUL,G.,ERARD,E.,GENTIEN,P.,MARCAILLOU,C. HARMFUL MAR.ALGAL BLOOM.LAVOISIER SCI.PUBL.PARIS 0,603–608(1993).
29. SNYDER,G.&TOBERER,E.COMPLEX THERMOELECTRIC MATERIALS.NAT.MATER.7,105–114 (2008).
30. KUMAR,A. ET AL.PHASE TRANSITIONS OF THERMOELECTRIC TAGS-85.INORG.CHEM.(2017). DOI:10.1021/ACS.INORGCHEM.7B02433
31. ZHAO,L.-D.,DRAVID,V.P.&KANATZIDIS,M.G.THE PANOSCOPIC APPROACH TO HIGH PERFORMANCE THERMOELECTRICS.ENERGY ENVIRON.SCI.7,251(2014).
32. HEREMANS,J.&DRESSELHAUS,M.WHEN THERMOELECTRICS REACHED THE NANOSCALE.NAT.…8, 471–473(2013).
33. VENKATASUBRAMANIAN,R.,SIIVOLA,E.,COLPITTS,T.&O’QUINN,B.THIN-FILM THERMOELECTRIC DEVICES WITH HIGH ROOM-TEMPERATURE FIGURES OF MERIT.NATURE 413,597–602(2001). 34. SCHNEIDER,M.N.,ROSENTHAL,T.,STIEWE,C.&OECKLER,O.FROM PHASE-CHANGE MATERIALS TO
THERMOELECTRICS?ZEITSCHRIFT FUR KRIST.225,463–470(2010).
35. ZHU,M. ET AL.UNIQUE BOND BREAKING IN CRYSTALLINE PHASE CHANGE MATERIALS AND THE QUEST FOR METAVALENT BONDING.ADV.MATER.1706735,1706735(2018).
36. CUI,S. ET AL.FROM SELENIUM- TO TELLURIUM-BASED GLASS OPTICAL FIBERS FOR INFRARED SPECTROSCOPIES.MOLECULES 18,5373–5388(2013).
37. HOUIZOT,P. ET AL.INFRARED SINGLE MODE CHALCOGENIDE GLASS FIBER FOR SPACE.OPT.EXPRESS 15,12529–12538(2007).
38. CHURBANOV,M.F. ET AL.HIGH-PURITY AS-S-SE AND AS-SE-TE GLASSES AND OPTICAL FIBERS. INORG.MATER.43,441–447(2007).
39. LIGERO,R. A.,VAZQUEZ,J.,VILLARES,P.&JIMENEZ-GARAY,R.CRYSTALLIZATION KINETICS IN THE AS-SE-TE SYSTEM.THERMOCHIM.ACTA 162,427–434(1990).
40. KATS,M.A.,BLANCHARD,R.,GENEVET,P.&CAPASSO,F.NANOMETRE OPTICAL COATINGS BASED ON STRONG INTERFERENCE EFFECTS IN HIGHLY ABSORBING MEDIA.NAT.MATER.12,20–24(2013). 41. HOSSEINI,P.,WRIGHT,C.D.&BHASKARAN,H.AN OPTOELECTRONIC FRAMEWORK ENABLED BY
LOW-DIMENSIONAL PHASE-CHANGE FILMS.NATURE 511,206–211(2014).
42. WUTTIG,M.,BHASKARAN,H.&TAUBNER,T.PHASE-CHANGE MATERIALS FOR NON-VOLATILE PHOTONIC APPLICATIONS.NAT.PHOTONICS 11,465–476(2017).
43. MKHITARYAN,V.K. ET AL.TUNABLE COMPLETE OPTICAL ABSORPTION IN MULTILAYER STRUCTURES INCLUDING GE2SB2TE5 WITHOUT LITHOGRAPHIC PATTERNS.ADV.OPT.MATER.5,1–7(2017). 44. JARIWALA,D.,SANGWAN,V.K.,LAUHON,L.J.,MARKS,T.J.&HERSAM,M.C.EMERGING DEVICE
APPLICATIONS FOR SEMICONDUCTING TWO-DIMENSIONAL TRANSITION METAL DICHALCOGENIDES. ACSNANO 8,1102–1120(2014).
45. JARIWALA,D.,MARKS,T.J.&HERSAM,M.C.MIXED-DIMENSIONAL VAN DER WAALS HETEROSTRUCTURES.NAT.MATER.16,170–181(2016).
46. BRITNELL,L. ET AL.STRONG LIGHT-MATTER INTERACTIONS THIN FILMS.SCIENCE (80-.).340, 1311--1314(2013).
47. LI,Y. ET AL.MEASUREMENT OF THE OPTICAL DIELECTRIC FUNCTION OF MONOLAYER TRANSITION -METAL DICHALCOGENIDES:MOS2,MO S E2,WS2, AND WS E2.PHYS.REV.B-CONDENS.MATTER MATER.PHYS.90,1–6(2014).
48. AHN,G.H. ET AL.STRAIN-ENGINEERED GROWTH OF TWO-DIMENSIONAL MATERIALS.NAT.COMMUN. 8,1–7(2017).
49. DE VRIES,E.K. ET AL.TOWARDS THE UNDERSTANDING OF THE ORIGIN OF CHARGE-CURRENT -INDUCED SPIN VOLTAGE SIGNALS IN THE TOPOLOGICAL INSULATOR BI2SE3.PHYS.REV.B-CONDENS. MATTER MATER.PHYS.92,1–5(2015).
50. ZHANG,H. ET AL.TOPOLOGICAL INSULATORS IN BI2SE3,BI2TE3 AND SB2TE3 WITH A SINGLE DIRAC CONE ON THE SURFACE.NAT.PHYS.5,438–442(2009).
51. MIWA,J.A. ET AL.VAN DER WAALS EPITAXY OF TWO-DIMENSIONAL MOS<INF>2</INF >-GRAPHENE HETEROSTRUCTURES IN ULTRAHIGH VACUUM.ACSNANO 9,6502–6510(2015). 52. SOLUYANOV,A.A. ET AL.TYPE-IIWEYL SEMIMETALS.NATURE 527,495–498(2015).
53. QIAN,X.,LIU,J.,FU,L.&LI,J.QUANTUM SPIN HALL EFFECT IN TWO -DIMENSIONAL TRANSITION METAL DICHALCOGENIDES.SCIENCE (80-.).346,1344–1347(2014).
54. TANG,S. ET AL.QUANTUM SPIN HALL STATE IN MONOLAYER 1T’-WTE2.NAT.PHYS.13,683–687 (2017).
55. ASABA,T. ET AL.MAGNETIC FIELD ENHANCED SUPERCONDUCTIVITY IN EPITAXIAL THIN FILM WTE 2.1–7(2018). DOI:10.1038/S41598-018-24736-X
1.10 Literature
11
56. WANG,R. ET AL.ORDERED PEIERLS DISTORTION PREVENTED AT GROWTH ONSET OF GETE ULTRA -THIN FILMS.SCI.REP.6,32895(2016).
57. GINLEY,T.,WANG,Y.&LAW,S.TOPOLOGICAL INSULATOR FILM GROWTH BY MOLECULAR BEAM EPITAXY:AREVIEW.CRYSTALS 6,154(2016).
58. BLANK,D.H. A,DEKKERS,M.&RIJNDERS,G.PULSED LASER DEPOSITION IN TWENTE: FROM RESEARCH TOOL TOWARDS INDUSTRIAL DEPOSITION.J.PHYS.D.APPL.PHYS.47,034006(2014).
13
Image: the inside of the Pulsed Laser Deposition vacuum chamber. The bright plasma is visible, as well as the substrate directly opposite.
Chapter 2. Experimental Methods.
All experimental techniques are introduced, with emphasis on PLD.
2.1
Abstract
In this chapter the experimental techniques employed in the other chapters will be briefly introduced. Since the Pulsed Laser Deposition and RHEED technique were used most extensively, and were a new addition to our research group, this system is described in more detail to provide an introduction that can be used by future operators.
Chapter 2. Experimental Methods.
2.2
Differential Scanning Calorimetry
Calorimetry is a well-established technique in many (chemical) labs. A traditional Differential Scanning Calorimeter (DSC) heats two closed sample pans or ‘crucibles’ simultaneously. One pan contains the specimen, the other is empty. The measurement signal is simply the difference in energy being supplied to both pans to maintain a given temperature profile. Since the specimen can undergo both exothermic and endothermic reactions, the observed energy difference might be either positive or negative. Furthermore, since the specimen has a certain heat capacity the heat flow to the specimen pan will in general be higher, which can also be measured. A downside of traditional DSC is the mass of the sample pans and the heater elements, which impose limits on the attainable heating and cooling rates.
By using a microchip-based DSC, (Ultrafast DSC-1, Mettler Toledo), heating and cooling rates of 10 000 K/s can be achieved. Since all components (heaters/sample pans/thermocouples) are incorporated within a 500 μm diameter chip membrane area, the energy requirements and heat gradients can be reduced tremendously.1–4
The downside of using this small-scale system is that complications arise due to excessive evaporation: the samples are not enclosed, but merely held under nitrogen atmosphere on the chip. Furthermore, sample preparation and structural analysis after thermal treatment are more challenging. Specimen have to be applied to the chip area using a hair. The specimen sticks to the hair using static forces and the sample transfer takes considerable dexterity. In general the chip sensor area will be too dirty for reuse, therefore a clean sensor chip is installed and calibrated before each new sample is applied.
Figure 1. a.) the Mettler Toledo Ultrafast DSC 1. The microscope attachment is essential to the application of the micron-sized samples using a hair. The device is closed to allow for a controlled nitrogen flow and to prevent the precipitation of ice on the cold-sink (not shown), which is kept at -90 °C. b) Shows the sensor chip and the sample and reference ‘crucibles’ in the center. c) Shows the active area of the chip and a small sample slightly off-center. A white (reflective) halo of redeposited evaporated material is visible. Figure adapted from 5.
2.3 X-ray diffraction
15
2.3
X-ray diffraction
Standard laboratory X-ray diffractometers are quite common and can yield information from a wide range of samples and on a variety of crystallographic questions. In this thesis we mainly present data obtained using thin-film X-ray diffraction (XRD), obtained using the varyingly named Bragg-Brentano, θ-2θ, or
2θ-ω geometry. A sketch of the setup is shown in figure 2. An X-ray source Cu Kα (λ
= 1.54Å) is directed at the plane of the film, with incidence angle θ. A detector is placed directly opposite, at the same angle, which allows detection of the diffracted beam intensity. This geometry allows for the measurement of crystal plane spacing parallel to the surface of the film/substrate, according to Bragg’s law. When operated at small incidence angles (< 10°) the reflected intensity oscillates, regardless of crystal structure, with a period corresponding to the film thickness. Such a thickness measurement using the Kiessig oscillations is called an X-Ray Reflectivity (XRR) measurement. The XRD and XRR analysis in this thesis is performed using a Philips X’PERT MRD system.
Figure 2. Several common modes of thin-film X-ray measurements are shown. Source (Src) and detector (Det) as well as thin film (parallel lines) can move and rotate along a number of axes.
Chapter 2. Experimental Methods.
2.4
Atomic Force Microscopy
Atomic force microscopy (AFM) in tapping mode (TM) was extensively used as a characterization tool for thin films grown using PLD. The technique is easy to use and provides information on the local height variations of the film, with sub-nm accuracy, while scanning surface area of many square μm. Next to a visually interpretable height map, which will reveal the geometry of surface features, such as crystallites or amorphous globular features, the room-mean-square (RMS) roughness may be determined. Additionally, the technique was often used to determine the thickness of a grown film by making a sharp cut, which terminated at the much harder (silicon) substrate. By scanning the AFM across the cut, thicknesses in the 10-100nm range could be accurately obtained. An AFM is operated in tapping mode by using a cantilever that resonantly oscillates at high frequency above the surface of the sample. The oscillation frequency is kept on resonance by a feedback loop which adjusts the height of the cantilever when the resonance frequency changed due to the proximity of a surface feature. The tapping is extremely gently and does not deform the surface of the films, although when the surface consists of loosely bound particulates, the cantilever may pick up a particle, which will give repeated feature artifacts in the scan image. The AFM analysis in this thesis is performed using a Bruker Veeco Multimode 8 AFM.
2.5
Scanning Electron Microscopy
A scanning electron microscope (SEM) can be used with a large number of detectors and geometries, yielding various kinds of information. A highly focused beam of accelerated electrons (usually 20-30 kV) is scanned across the sample surface. Several detectors provide imaging: detecting Secondary Electron (SE) generally provides the highest resolution, while backscattered electrons (BSE), which have been elastically scattered by the sample, provide element-specific contrast. Energy Dispersive X-ray Spectroscopy (EDS) is based on the specific spectral lines within the X-ray regime emitted by materials excited by an electron beam. A spectrum obtained from a material irradiated by the electron beam can be fitted using standard materials spectra to obtain the elemental composition of the sample. While the accuracy of these measurements is highly dependent on the sample geometry, and is in principle not defined for thin films since the interaction volume of the electron beam is roughly a μm, the precision of the method is quite high. We therefore usually restrict ourselves to the comparison of highly similar samples when performing EDS analysis. Finally, crystal structure and texture can be obtained in SEM using Electron Back-Scatter Diffraction (EBSD). The sample is tilted to an angle of 70° with respect to the incoming beam. Part of the beam now diffracts off the surface, somewhat analogously to RHEED (introduced later). The diffracted pattern shows Kikuchi-lines and is captured using a phosphor screen and
2.6 Transmission Electron Microscopy
17
CCD camera. An automated analysis program (EDAX-OIM software v. 7.2.1) was used to fit the Kikuchi patterns with a known crystal structure, to obtain the local crystal orientation to within one degree. Due to the high incidence angle, the spot size (and therefore resolution limit) of the SEM deteriorates. In chapter 7 we show that the ~100nm domains could just be resolved. No attempt was made to image the thin films grown using PLD, since their grain size is around 80 nm. In this thesis, SEM analysis is performed using Philips XL30S, FEI Nova NanoSEM, and FEI Helios Dual Beam systems.
2.6
Transmission Electron Microscopy
While also an electron-microscopy technique, the sample requirements, contrast mechanism, and achievable resolution are vastly different for TEM than for SEM. The sample has to be polished to be electron transparent, using progressively more gentle methods to obtain a sample ‘film’ with a thickness of less than 100 nm. The details of TEM sample preparation of thin-films are provided in
6. The TEM operates by directing a high–energy (200 kV) beam of electrons onto
the specimen. These electrons form a coherent wavefront, which scatters on the specimen and is transmitted through onto a detector. In Bright-Field (BF) mode, darker areas indicate a strong scattering contrast: either the material is thick, or the atomic planes are in a scattering condition which diverts the incoming wave away from the optical axis. When operated in Scanning TEM mode, the beam is focused into a small spot on the specimen and rastered across the area of interest. The non-scattered beam is captured. When high angle annular dark field (HAADF) STEM is used, only electrons scattered to high angles are detected, which leads to incoherent imaging. Then, complicating wave interference effects are lost enabling a much more straightforward and simple image interpretation. Atomic columns are always imaged as bright spots in a dark surrounding and the intensity of the bright spots is directly related to the average atomic number Z in the atomic column (typically intensity is proportional to Z to the power 1.6-2.0).7
Similarly to SEM, TEM offers the option of detecting elemental composition by analyzing the x-ray radiation due to electron-beam excitation of the specimen. When a STEM-mode is used to raster across the image, elemental mappings can be performed. In principle the spatial resolution of TEM-EDS can in principle go down to atomic resolution, whereas this resolution cannot be better than a few hundred nm with SEM-EDS on bulk specimen. Since the TEM is thus intrinsically a low-sample volume technique, the elemental composition of a thin film can be measured more accurately than in SEM.
By modifying the active lens configuration of the TEM behind the sample (mainly the strength of the intermediate lens), the specimen can be imaged in reciprocal (diffraction) space. The beam illumination is spread to obtain a planar wavefront at the specimen. By tilting the specimen across multiple axes, the
Chapter 2. Experimental Methods.
specimen will diffract along various zone axes, revealing the symmetry of the crystal, as well as twinning, exact lattice parameters, and possible presence of multiple phases.
The TEM analysis in this thesis is performed using JEOL 2010 and JEOL 2010F systems. Images were analyzed using the Digital Micrograph software (Gatan) The compositions of the samples were investigated using a SiLi EDS detector. Accurate composition information was obtained by Cliff-Lorimer w/o absorbance fitting in the NSS 2.3 software (Thermo Scientific).
2.7
Ellipsometry
To investigate the optical properties of the thin films grown using PLD, ellipsometry was performed. Linearly polarized light is reflected off the sample surface, scanning through a range of wavelengths (300-1800nm). After reflection a second polarizer and detector obtain the ratio of reflected p- and s- polarized light. The system allows heating during measurement (dynamic ellipsometry) as well as measuring absolute wavelength dependent reflectivity as well.
The obtained values of the reflected wave can be parametrized by ψ (phase rotation) and Δ (phase difference) values, and these are fitted to a model which describes the dielectric properties of the film, as well as its thickness. Many texts explain the intricacies of fitting various models, so we will not give an extensive description here.8,9 Generally, a material is modeled using a simple (nonphysical)
description to obtain layer thicknesses, then a real fit using one or more oscillators (resonant frequencies) is performed. For good electrical conductors, a Drude contribution is also included. The ellipsometry analysis in this thesis is perfomed using a J. Woollam UV-VIS spectroscopic ellipsometer and the VASE software as well as a custom Matlab code.
In order to produce actual optical devices, it is useful to model reflection profiles and use a dedicated angle-resolved reflectometer to experimentally verify this profile. Both a freely available script based on the TransferMatrix algorithm and a self-written angle-resolved reflectivity script were used. The reflectometer was a home-built system with a light source, polarizer, and prism analyzer within 350-800nm range.
2.8 Pulsed Laser Deposition
19
Figure 3. Measurement principle of ellipsometry. A Polarized beam is reflected off the sample surface and detected. The intensity and phase rotation is measured. The complex dielectric constant can be fitted. Adapted from. 8
2.8
Pulsed Laser Deposition
Pulsed Laser Deposition (PLD) is the current term used for a deposition process which has been experimented with for half a century, going back at least to 1965.10
The technique was more commonly referred to as Laser Ablation and Deposition (LAD), which in fact reveals more about the actual deposition mechanism. Interestingly, many older reports describe experiments using metals and semiconductors, while the mature technique known as PLD is mostly employed for the growth of oxides.11 The PLD setup consists of a high-vacuum chamber (10-8
mbar) where the deposition takes place, and a laser and optics table. Deposition of a thin film is achieved by irradiation and ablation of a target. Material leaves the target as a plume of plasma, and is deposited onto a substrate. The substrate can optionally be heated to promote surface diffusion. The process is moderated using a background gas such as oxygen or argon. A Reflective-High-Energy-Electron-Diffraction (RHEED) system, which consists of an electron gun and a phosphor screen, is used to monitor the growth process.
While many reports and even books exist on the deposition of a plethora materials, no general consensus can be found on the exact mechanics of many of the physical phenomena taking place during deposition. We will therefore also not attempt to describe this in too much detail. In the following section the components and capabilities of the PLD setup are described, and in chapter 3 the result of tuning various process parameters is shown.
Chapter 2. Experimental Methods.
Figure 4. The components of the PLD system’s main chamber are shown schematically. Adapted from the TSST PLD system user manual.
Vacuum System
The system consists of a main chamber and an attached loadlock for quick substrate and target exchange. The chamber, loadlock, and RHEED system are pumped by three turbofan pumps backed by two roughing pumps. The chamber pressure is monitored using Ion Gauge for high-vacuum and baratron gauge for low-vacuum.
Excimer Laser
A KrF excimer laser emitting light at 248 nm is used to ablate target materials. The high-energy UV-wavelength is chosen since it is well absorbed into many materials. The laser works by discharging a ~25 kV potential between 2 parallel plates within a 3 Bar KrF environment. The Kr+ and F- components are ionized, and their recombination emits 248 nm photons. The beam is not as coherent or parallel as a traditional laser. The beam path (several meters in length) is long enough to significantly reduce the observed laser power density at the beam center. The laser energy output reduces over time (weeks), due to the degeneration of the excimer gas. The laser pulse time is roughly 20 ns, the repetition rate can be varied from 0.1 Hz to 10 Hz, and typical pulse energies are in the 10 mJ range.
Beam Path
The optical bench contains only a few elements. Starting from the laser aperture, the light is passed through a mask, which takes the central area of the beam, and has the rectangular shape of the space between the discharge plates.
2.8 Pulsed Laser Deposition
21
This part of the beam contains the most homogeneous energy distribution. The mask further serves as the “object” for the lens. The lens is positioned just outside the vacuum chamber, and projects a focused “image” of the mask called the spot onto the ablation target. The position of the target is fixed but the lens and mask can be moved, allowing the spot size to be changed at will by using the lens equation for object, image and lens focal distance. Due to the divergence of the laser beam, as well as the KrF gas degradation, laser energy has to be carefully tuned before deposition using an energy meter. While advanced meters will map the entire beam profile, a basic model to measure total spot energy suffices. The meter is placed after the lens, just before the laser window of the vacuum chamber.
Laser Window
The laser enters through a window into the deposition chamber. Since the window transparency is reduced due to material deposition within the chamber, periodic cleaning has to be performed. The energy loss due to the window glass is 8%, and cleaning is recommended to keep the spot energy well-defined, as well as prevent the spot from burning into the window.
Targets
PLD-targets are bonded using heat-conductive glue to steel stubs. A total of five target stubs can be mounted on a carrousel, which can be rotated during a deposition, to make multi-material films. The targets usually contain a stoichiometric mixture of sinter-pressed powders of the deposition material, but single crystals can also be used and are in fact preferable. A high density and flat surface are critical to obtaining a clean deposition. Targets are periodically grazed using consecutively finer-grained paper to obtain a smooth surface.
The targets are aligned parallel to the substrate surface, since the ablation plume always erupts perpendicularly to the surface of the target. This means the targets are mounted at a 45° angle to the incoming laser beam, since it needs to pass the substrate holder. The spot is therefore smeared out due to this angled incidence. The target is scanned in the plane parallel to its surface to optimize material use and to prevent excessive wear within one deposition. Generally, the scan speed is chosen such that successive spots overlap significantly, which yields a homogeneous ablation track. Usually scanning is performed in a row-by-row pattern.
Substrate heater
The substrate is mounted on a heater holder at a distance of 6 cm from the target, with the laser spot location directly opposite to the substrate center. The substrate may either be glued onto the heater to optimize heat contact using silver glue, or clamped for faster loading. The substrate itself is generally either one cm by one cm or smaller. An attachment to load 3 mm diameter TEM grids was also designed. When a deposition is performed, the entire holder is taken out of the
Chapter 2. Experimental Methods.
vacuum system, the sample is mounted, and the heater is inserted back into the chamber.
The heater is capable of employing controlled heating rates up to 20 °C/min and maintaining temperatures up to 850 °C. The heater holder can be rotated along an axis perpendicular to the substrate surface (azimuth), and an axis within the plane of the sample (tilt). This is necessary for alignment with respect to the RHEED system.
A well-controlled temperature is essential to many depositions. However, the heater cannot be controlled by a feedback loop during PLD deposition, since this would affect the RHEED system due to the proximity of the heater current loop to the electron beam. Therefore, before starting a deposition, it is essential to let the system equilibrate itself at the deposition temperature before switching off the control loop and fixing the output voltage.
Gas inlet system
PLD is generally performed in a gas environment, which can have two reasons: (1) Inert gases (such as argon) are used to confine the ablated material
into a forward-moving plume, optimizing material deposition onto the substrate
(2) Reactive gas (oxygen) is used for oxide depositions, since vacuum depositions generally yield oxygen-poor films.
The gas pressure is a tuning parameter to tune deposition rate. Since the different elements will scatter differently, stoichiometry might also be affected, and these interactions can be quite complex. For the depositions in this thesis, generally a pressure of 0.12 mBar Ar was used. The gas pressure is controlled both upstream (inlet) and downstream (outlet). The inlet valve contains a flow controller, where for normal operation a flow of one sccm is used. By changing the aperture of the outlet valve, the pressure in the main chamber is controlled.
RHEED system
The RHEED system is an extremely useful tool in structural sample analysis and is much more than a simple deposition monitoring system. The system consist of several components: a 30kV electron gun (tungsten filament) emits electrons, which are focused within the RHEED tube by several lenses and apertures. This section of the RHEED system is differentially pumped using a separate turbo-pump to high-vacuum pressures. When the electron beam enters the main chamber, it is scattered by the process gas, reducing the signal quality. The beam is aligned to hit the substrate at a glancing angle (~3°). This ensures that the electron beam diffracts off the surface, with an estimated penetration depth of about 1 nm. The diffracted beam is imaged on a phosphor screen and captured on a CCD camera.
2.9 RHEED Analysis
23
2.9
RHEED Analysis
When one is familiar with electron-diffraction mode in TEM, the RHEED seems quite similar at first glance. The imaged pattern represents reciprocal space, where distanced relate inversely to plane spacing in the crystalline film. Due to the surface sensitivity however, not all symmetry features of a crystal are observed (see e.g. paragraph 5.10). Similarly, the grazing incidence of the beam breaks the symmetry of the diffraction pattern. While one axis parallel to the substrate on the phosphor screen will show diffraction spots with spacings relatable to the surface lattice spacing, the axis perpendicular to the substrate contains different information as will be explained below.
Pattern diagnostics
When the surface of a single-crystalline atomically flat surface is imaged, the resulting spot pattern features a set of spots on a ring, corresponding to the intersection of the crystal lattice points with the Ewald sphere. By increasing the tilt angle (i.e. the incidence angle), the whole pattern moves, since the diffraction is still subject to the angle of incidence = angle of reflection law. The spot intensity might change however, allowing for optimization of the pattern intensity. In single crystals, Kikuchi-lines and bands are observed as well, allowing for identification and alignment of the zone axis. The overlap of a band and a diffraction spot might produce a more intense spot. This is undesirable when one wants to observe spot intensity oscillations.
When the single-crystal surface contains roughness on the atomic scale, the spots become elongated into streaks. This can be explained by considering again the Ewald sphere. The relrods now penetrate the Ewald sphere along lines with length according to the height difference in the sample. When the coherence length of the electron beam is large, this roughness might be minor and still show a streaky pattern.
Chapter 2. Experimental Methods.
Figure 5. The electron beam is incident on the substrate at a glancing angle. The diffraction pattern can be calculated using the Ewald sphere construction. Assuming the substrate to be perfectly flat, relrods extend upward to intersect the sphere at well-defined points: these are the diffraction spots. Adapted from 12.
An untextured, but atomically flat film will intersect the Ewald sphere along sharply defined rings, since the crystalline spacing perpendicular to the direction of the beam is a continuous distribution. An amorphous material resembles this untextured but flat film: due to the more poorly defined lattice parameter however, the rings become blurred, analogously to those observed in TEM. When the roughness is increased more dramatically, a so-called 3D-pattern is observed, the electron beam penetrates through rough surfaces, yielding a transmission-condition, like in TEM. A symmetric spot pattern is visible along both axes.12,13
The RHEED patterns shown usually also contain a spot called the direct beam. This is part of the unscattered beam which passed by the sample. This is due to the grazing incidence. The amount of diffracted intensity is usually highest when some of the direct beam is still visible. The comparison of the direct beam intensity and the diffracted intensity can be used as a qualitative measure of surface quality (assuming proper alignment). Figure 6 shows several examples of substrates and samples with different crystalline states.
2.9 RHEED Analysis
25
Figure 6. a) amorphous SiO2 substrate. Next to the direct spot, only a featureless
reflected intensity is visible. b) amorphous or nanocrystalline film (WTe2) gives
blurred rings. c) polycrystalline film without texture (pure Te) gives sharp rings. d) rough film (GeTe) 3D-pattern of spots reminiscent of electron diffraction in TEM. Can be determined to have out-of plane texture but random in-plane. e) A smooth film (Bi2Te3) gives streaks. The Bi2Te3 has out-of plane but no in-plane
texture, as can be determined from the streak spacings. f) Single-crystalline, atomically smooth mica substrate shows a spotty pattern with spots on the first Laue circle. Kikuchi bands (straight lines fanning out from center) are also a good indicator of high-quality substrates.
RHEED scale calibration
Since RHEED allows the determination of crystal symmetry, it seems useful to calibrate the detection camera, to obtain a pixel – nm conversion. In practice however, this is not possible since the actual spot separation on the screen is affected by a number of factors, including the exact position where the substrate intercepts the RHEED beam, the position of metal components (such as the shutter), and most crucially, the electric current in the heater holder. This means it is impossible to obtain a universal calibration. The best practice is therefore to use a well-characterized substrate/film, from which the RHEED can be internally calibrated.
Chapter 2. Experimental Methods. Lattice spacing evolution using RHEED
Since the spot spacing in the RHEED patterns is directly related to the film lattice parameter, this can be monitored during film growth. Typically, the RHEED pattern was captured every 200 ms, to allow for a variable intensity in-between laser pulses. A custom script was developed to assist with analyzing the large amount of generated data (figure 7). The script takes an intensity profile across the streaky diffraction pattern, then sums the intensity per row, creating an intensity profile. The user can select the three peaks and the limits used to subtract a baseline. The script will fit this baseline using a linear function. The peak itself is fitted using a single Gaussian peak profile. The three peak locations are extracted. A linear function is fitted through the three peak locations to obtain the peak spacing. This spacing (in pixels) is plotted as a function of time. This data is calibrated using a known lattice parameter. We observe that a sub-pixel resolution can easily be obtained, yielding a real-space resolution of ±0.01 Å.
Figure 7. a. The RHEED pattern during deposition. b. The row-summed intensity profile is shown on the right. c. Every row represents one intensity profile taken during deposition. With the naked eye one cannot distinguish the spacing changes. d. The spacing between streaks is plotted versus time/frame number. The pixel spacing clearly varies.
2.9 RHEED Analysis
27 Live monitoring
As was discussed earlier, an atomically flat surface yields intense diffraction spots. When a material is grown however, this smoothness is (temporarily) destroyed. Depending on the growth mode, the layer can either recover this atomically smooth surface (layer by layer), or will retain some roughness (islanded). When the material grows in a layer-by-layer mode, the resulting oscillation of RHEED intensity can be captured as a function of time, and directly interpreted as the growth rate in monolayers per pulse. An example of such an oscillatory intensity is shown in figure 8.
Figure 8. RHEED streak intensity oscillations of Bi2Te3 being grown on mica at
Chapter 2. Experimental Methods.
2.10
Literature
1. VERMEULEN,P.A.,MOMAND,J.&KOOI,B.J.REVERSIBLE AMORPHOUS-CRYSTALLINE PHASE CHANGES IN A WIDE RANGE OF SE1-XTEX ALLOYS STUDIED USING ULTRAFAST DIFFERENTIAL SCANNING CALORIMETRY.J.CHEM.PHYS.024502,(2014).
2. IERVOLINO,E. ET AL.TEMPERATURE CALIBRATION AND ELECTRICAL CHARACTERIZATION OF THE DIFFERENTIAL SCANNING CALORIMETER CHIP UFS1 FOR THE METTLER-TOLEDO FLASH DSC1. THERMOCHIM.ACTA 522,53–59(2011).
3. POEL,G.VANDEN,ISTRATE,D.,MAGON,A.&MATHOT,V.PERFORMANCE AND CALIBRATION OF THE FLASH DSC1, A NEW,MEMS-BASED FAST SCANNING CALORIMETER.J.THERM.ANAL.CALORIM.110, 1533–1546(2012).
4. VAN HERWAARDEN,S. ET AL.DESIGN, PERFORMANCE AND ANALYSIS OF THERMAL LAG OF THE UFS1 TWIN-CALORIMETER CHIP FOR FAST SCANNING CALORIMETRY USING THE METTLER-TOLEDO FLASH DSC1.THERMOCHIM.ACTA 522,46–52(2011).
5. CHEN,B.GE-SB-TE BASED PHASE-CHANGE NANOPARTICLES.(2017).
6. MOMAND,J.STRUCTURE AND RECONFIGURATION OF EPITAXIAL GETE/SB2TE3 SUPERLATTICES. (2017).
7. WILLIAMS,D.&CARTER,C.THE TRANSMISSION ELECTRON MICROSCOPE.(1996). AT <HTTP://LINK.SPRINGER.COM/CHAPTER/10.1007/978-1-4757-2519-3_1>
8. FUJIWARA,H.SPECTROSCOPIC ELLIPSOMETRY PRINCIPLES AND APPLICATIONS.SPECTROSCOPIC ELLIPSOMETRY PRINCIPLES AND APPLICATIONS (2007). DOI:10.1002/9780470060193
9. TOMPKINS,H.G.AUSER’S GUIDE TO ELLIPSOMETRY.AUSER’S GUIDE TO ELLIPSOMETRY (1993). DOI:10.1016/B978-0-12-693950-7.50008-1
10. SMITH,H.M.&TURNER,A.F.VACUUM DEPOSITED THIN FILMS USING A RUBY LASER.APPL.OPT. 4,147–148(1965).
11. DOESWIJK,L.M.,RIJNDERS,G.&BLANK,D.H.A.PULSED LASER DEPOSITION:METAL VERSUS OXIDE ABLATION.APPL.PHYS.AMATER.SCI.PROCESS.78,263–268(2004).
12. RIJNDERS,G.&BLANK,D.H.A.IN SITU DIAGNOSTICS BY HIGH-PRESSURE RHEEDDURING PLD. PULSED LASER DEPOS.THIN FILM.APPL.GROWTH FUNCT.MATER.85–97(2006).
DOI:10.1002/9780470052129.CH4
13. TANG,F.,PARKER,T.,WANG,G.-C.&LU,T.-M.SURFACE TEXTURE EVOLUTION OF
POLYCRYSTALLINE AND NANOSTRUCTURED FILMS:RHEED SURFACE POLE FIGURE ANALYSIS.J. PHYS.D.APPL.PHYS.40,R427–R439(2007).
31
Image: A series of Bi2Te3 thin films on Si-SiO2 thermal oxide wafers.
Chapter 3. Pulsed laser deposition of
tellurides.
Basic principles and results for the growth optimization of telluride thin films are presented.
3.1
Introduction
While extensive research has been performed on oxide growth using PLD (where it is in fact the preferred tool), the growth of (telluride) chalcogenides has received far less attention. Consequently, the literature is relatively scarce, and reports on process parameters and/or achieved layer quality were incomplete or conflicting. The aim of the present thesis has been to grow textured (at least out-of plane), flat, stoichiometric, and thickness controlled films of single-and multi-component films. To this end, a large number of basic experiments was carried out to get a basic understanding of the PLD process and the growth of these materials. Three key deposition stages can be identified:
1. Ablation
2. Transport (to the substrate) 3. Growth
All steps allow interdependent tuning of parameters, and therefore some rudimentary understanding (if not full characterization) of all three steps is needed to attain control of layer growth.
Chapter 3. Pulsed laser deposition of tellurides.
3.2
Ablation
Target surface analysis
The source material for PLD comes from targets which were obtained commercially (Ktech). The targets are obtained using a powder sintering process providing a highly dense target which is able to withstand repeated laser pulses without cracking. A target is generally chosen to have the stoichiometry which is desired in the film, but sometimes one might deviate from this, when the film turns out to be off-stoichiometric. Since a target is hit repeatedly (thousands of pulses per deposition), it is moved to prevent excessive erosion of one particular spot. This erosion usually alters the local composition, and may also result in an increase of particulates on the film (that is the result of a process called ‘splashing’). To study this erosion phenomenon, as well as determine the spot size on the target, several lens positions were used to make spots of various sizes. Similarly, different numbers of pulses were used to obtain a qualitative understanding of the speed of degradation. Figure 1 shows the Bi2Te3 target with many different spot marks. Note
that in this case target scanning was intentionally not used.
Figure 1. Optical image of the Bi2Te3 target (diameter 2.54 cm). Different spot
sizes can be observed, as well as progressively eroded spots due to increasing pulse number (bottom row).
The spots were studied using SEM and are shown in figure 2. Although the number of pulses might seem low compared to an actual deposition, one has to keep in mind the target is generally scanned, so we expect a single area to be hit less than 20 times during a deposition of 4000 pulses. As the same area is hit consecutively, the area is roughened considerably, and edge features start to develop. While edges are not a problem by itself, this is indicative of partial melting and/or element specific evaporation. When these effects can accumulate over many depositions, film stoichiometry will degrade.
3.2 Ablation
33
Figure 2. Spots made with the same spotsize and laser energy ( 1 Jcm-2) on a
Bi2Te3 target. The number gives the amount of pulses. The spot morphology
clearly changes, large-scale structures seem to emerge, and the edges of the spot become slanted, which is related to the 45° incidence angle of the laser beam.
Stoichiometric ablation
In a regular deposition, the spot is continuously scanned, which yields an ablation track as shown in figure 3. The surface looks smooth (except for some small edge effects), indicative of proper ablation and local melting, but no element specific evaporation or pillar formation. Using EDS, the ablation track is confirmed to still possess the Bi2Te3 stoichiometry (Bi 41.3 at.%, Te 58.7 at.%) within the
measurement accuracy, which means the initial ablation plume contains the correct stoichiometry as well. Similar results were obtained for Sb2Te3.
