Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Controllable and scalable thermal sublimation thinning of black phosphorus Luo, Weijun 2017

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2017_may_luo_weijun.pdf [ 16.3MB ]
Metadata
JSON: 24-1.0345629.json
JSON-LD: 24-1.0345629-ld.json
RDF/XML (Pretty): 24-1.0345629-rdf.xml
RDF/JSON: 24-1.0345629-rdf.json
Turtle: 24-1.0345629-turtle.txt
N-Triples: 24-1.0345629-rdf-ntriples.txt
Original Record: 24-1.0345629-source.json
Full Text
24-1.0345629-fulltext.txt
Citation
24-1.0345629.ris

Full Text

Controllable and Scalable Thermal Sublimation Thinningof Black PhosphorusbyWeijun LuoB. Sc, Fudan University, Shanghai, China, 2013A THESIS SUBMITTED IN PARTIAL FULFILLMENTOF THE REQUIREMENTS FOR THE DEGREE OFMASTER OF APPLIED SCIENCEinTHE FACULTY OF GRADUATE AND POSTDOCTORALSTUDIES(Materials Engineering)The University of British Columbia(Vancouver)April 2017c©Weijun Luo, 2017AbstractTwo-dimensional lamellar black phosphorus (BP) has emerged as a promisingsemiconductor for next generation integrated circuits (IC) and photonics, especiallyin flexible and ultra-thin electronic and photonic devices. With layer numbers of>20 to 1, the electronic energy band gap of BP covers the range from 0.3 (bulk) to2 eV (single-layer), which can fill the gap between graphene and transition metaldi-chalcogenides (TMDCS). It is necessary to prepare uniform, large scale and crys-talline few-layer BP for industry applications.We investigated thinning rates of BP at different temperatures so that the userscan control the time of heating and have the ability to monitor the thickness ofBP during heating processes. Identification of crystallographic orientation (CO)of BP by Raman Spectroscopy is applied, which enables the Raman intensity ra-tios between BP and substrates to be only thickness-dependent. This ratio can beused as a non-contact optical method to determine the actual thickness of BP dur-ing preparation, which is crucial to determine the end point of the thinning process.In this thesis work, we first reported the layer-by-layer sublimation of BP below600 K, which was observed by optical color changes; secondly, we investigated thethinning rates of BP at 500 K and 550 K to be ∼ 0.2 nm / min 500 K and ∼ 1.5nm / min at 550 K; thirdly, we investigated the effective determination of CO of BPand underlying Si by polarized Raman Spectroscopy with excitation wavelength of441.6 nm; fourthly, we investigated the thickness-dependent Raman peak intensityratio SiA2g at a fixed CO, which can be used as an indicator of the thickness of BP;lastly, we presented the successful and repeatable preparation of large crystalline 2iito 4 -layer BP.This work is the first study available to use the sublimation thinning as a con-trollable method to prepare large, uniform and crystalline BP down to 2-4 atomiclayers. This work is the first study available on developing an all-Raman methodin identifying the CO of BP, determining the in-situ and ex-situ thickness and con-firming the crystallinity and uniformity of prepared BP.iiiPrefaceAll of the work was done in the Department of Materials Engineering at UBCduring the past two years, including the designs of experiments such as in-situRaman Spectroscopy, Raman imaging and Angle-resolved Polarized Raman spec-troscopy (ARPRS). The Atomic Force Microscopy was performed in 4D LABS,Simon Fraser University, Burnaby, B.C., Canada.The author prepared most of the original BP samples and developed subli-mation method in preparing BP with different thickness; Mr. Jialun Liu and Prof.Wenjuan Zhu from the University of Illinois at Urbana-Champaign (U.S.A.) pro-vided the sample used in Figure 2.3 and 2.4. The author conceived and conductedmost of the Raman and AFM measurements, and did all the data analyses in thiswork; Mr. Rui Yang from Prof. Joshua Folk’s group at the University of BritishColumbia (Canada) helped with the AFM measurements illustrated in Figure 2.4.Mr. David Tuschel from Horiba Scientific in NJ, U.S.A inspired the authorboth in comprehension of Raman theory and experiment design.One conference presentation was made at Material Research Society 2016Fall meeting on this work in Chapter 2.1. MRS 2016 Fall, NM2.2.08, Boston, M.A, U.S.A.Weijun Luo, Rui Yang, Jialun Liu, Wenjuan Zhu, Guangrui (Maggie) Xia.“Thermal Sublimation as a Scalable and Controllable Thinning Method for theFabrication of Few-Layer Black Phosphorus.”ivIn addition, one journal article has been submitted for review based on thework in Section 2; one journal article is in preparation for review based on thework in Chapter 3.1. Luo, W., Yang, R., Liu, J., Zhu, W. and Xia, G., 2016. Thermal Subli-mation: a Scalable and Controllable Thinning Method for the Fabrication of Few-Layer Black Phosphorus. arXiv preprint arXiv:1601.04103.2. Luo, W., Song, Q., Zhou, G., Tuschel, D. and Xia, G., 2016. Study ofBlack Phosphorus Using Angle-Resolved Polarized Raman Spectroscopy with 442nm Excitation. arXiv preprint arXiv:1610.03382.vTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ixList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xGlossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xixAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xx1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Background And Motivations . . . . . . . . . . . . . . . . . . . . 11.1.1 BP Structure: Layer-stacked 2D Semiconductor Material . 21.1.2 Layer-dependent Tunable Band Gap, High Carrier Mobil-ity, Fast And High Photo-response . . . . . . . . . . . . . 41.1.3 Strong In-plane Anisotropies Of Black Phosphorus . . . . 71.2 Current Challenges In BP Study . . . . . . . . . . . . . . . . . . 131.2.1 Degradation Of BP In Air . . . . . . . . . . . . . . . . . . 131.2.2 Fabrication Of Large-scale Few-layer Crystalline BP . . . 131.2.3 Debate On The CO Determination Of BP By ARPRS . . . . 241.3 Basic Knowledge Of This Thesis Work . . . . . . . . . . . . . . 281.3.1 Basic Knowledge of Raman Scattering . . . . . . . . . . 28vi1.3.2 Basic Knowledge Of Polarized Raman Spectroscopy . . . 301.3.3 Basic Knowledge Of Phonon Modes Of Black Phosphorus 341.3.4 Crystallographic Orientation Of Si (100) Wafer And ItsOrientation Dependent Raman Response . . . . . . . . . 371.3.5 Thickness Determination By The N-dependent Raman In-tensity Ratio Of ISiI2DM . . . . . . . . . . . . . . . . . . . . 381.3.6 Thermal Sublimation For Preparation Of Few-layer 2D Ma-terials . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391.3.7 Previous Studies On Thermal Decomposition And Subli-mation Of Black Phosphorus . . . . . . . . . . . . . . . . 411.4 Problem Definition And Thesis Goal . . . . . . . . . . . . . . . . 421.5 Scope Of The Thesis . . . . . . . . . . . . . . . . . . . . . . . . 421.6 The Organization Of This Thesis . . . . . . . . . . . . . . . . . . 432 Thermal Sublimation of Black Phosphorus . . . . . . . . . . . . . . 452.1 General Procedure Of The Experimental Work Of This Chapter . . 452.1.1 Thermal Sublimation . . . . . . . . . . . . . . . . . . . . 462.1.2 Raman Measurement . . . . . . . . . . . . . . . . . . . . 462.1.3 Determination Of The Crystallographic Orientations (CO)Of BP And underlying Si (100) Substrate . . . . . . . . . 482.1.4 Atomic Force Microscopy (AFM) . . . . . . . . . . . . . 482.2 Results And Discussions . . . . . . . . . . . . . . . . . . . . . . 492.2.1 Investigations Of Thermal Sublimation And Its Mechanism 492.2.2 Investigation of The Thinning Rates . . . . . . . . . . . . 662.2.3 Preparation Of The few-layer BP . . . . . . . . . . . . . . 812.2.4 Comparison Of BP Products With Previous Techniques . . 882.3 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 883 Determination of Crystallographic Orientations For BP ThicknessDetermination And Control By Raman Spectroscopy With 442 nmExcitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 903.1 Determination Of BP CO By ARPRS . . . . . . . . . . . . . . . . 903.2 Experiment Methods . . . . . . . . . . . . . . . . . . . . . . . . 91vii3.2.1 Sample Preparation . . . . . . . . . . . . . . . . . . . . . 923.2.2 Angle-resolved Polarized Raman Spectroscopy (ARPRS) . 933.2.3 Atomic Force Microscopy (AFM) . . . . . . . . . . . . . 943.3 Results And Discussion . . . . . . . . . . . . . . . . . . . . . . . 943.3.1 ARPRS Study In The Determination Of CrystallographicOrientation (CO) Of BP . . . . . . . . . . . . . . . . . . 943.3.2 Raman Peak Intensity Ratios: ISiIA2gAs A Function Of BPThickness . . . . . . . . . . . . . . . . . . . . . . . . . . 1203.3.3 Conclusion: . . . . . . . . . . . . . . . . . . . . . . . . . 1294 Conclusion and outlook . . . . . . . . . . . . . . . . . . . . . . . . . 1304.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1304.2 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133viiiList of TablesTable 1.1 Previous studies on liquid exfoliation of BP [1]. ?aq.: deoxy-genated water. . . . . . . . . . . . . . . . . . . . . . . . . . . 16Table 1.2 The energy balance sheet for Raman process [2]. . . . . . . . 28Table 1.3 Raman intensity of BP Ag and B2g modes under “parallel” con-figuration. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37Table 3.1 Polar diagrams of peak intensity ratios of A1g over A2g of 11 BPsamples on SiO2 / Si substrates acquired by Angle-resolved Po-larized Raman Spectroscopy (ARPRS). . . . . . . . . . . . . 102Table 3.2 Raman spectra: with maximum A1g / A2g Raman intensities. . . 106Table 3.3 Polar diagrams of 16 BP samples on polyimide substrates ac-quired by Angle-resolved Polarized Raman Spectroscopy (ARPRS).110Table 3.4 Raman spectra: with maximum A1g / A2g Raman intensities. . . 115Table 3.5 Optical images, AFM thickness profiles of 11 BP Samples onSiO2/Si substrate. No.1-5 were samples prepared by sublima-tion thinning; No.6-12 were samples prepared by exfoliation. . 117Table 3.6 Optical images, AFM images and thickness profiles of 16 BPSamples on polyimide substrate. No.1-5, 8, 10 and 12 weresamples prepared by sublimation thinning; No.6-7, 9, 11 and13-16 were samples prepared by exfoliation. . . . . . . . . . . 120ixList of FiguresFigure 1.1 The atomistic ball-stick models of few-layer black phosphorus. 2Figure 1.2 The first BP based transistor which was fabricated on a pieceof 6.5-nm-thick mechanically exfoliated BP sample. Figure isreprinted from [3] with permission. . . . . . . . . . . . . . . 3Figure 1.3 The band gap range of 2D and conventional bulk semiconduc-tors. Figure is reprinted from [4] with permission. . . . . . . 5Figure 1.4 (A): mobility and current on/off ratio of BP; (B): the photo-responsivity and response time of BP. Figure is reprinted from[4] with permission. . . . . . . . . . . . . . . . . . . . . . . 6Figure 1.5 (A): Material parameters of single-layer BP; (B): the currentcharacteristics of the n-type (solid) and p-type (dash) BP field-effect transistors at the x- (“zigzag”) and y-direction (“arm-chair”). m0 is the free electron mass which has a value of 9.1× 10−31 kg. Figure is reprinted from [5] with permission. . . 8Figure 1.6 (A): Light absorption of BP with different thickness (Figure isreprinted from [6] with permission) , here the “x” denotes forthe “armchair” direction and the “y” denotes for the “zigzag”direction; (B): a black phosphorus vertical p-n junction (Figureis reprinted from [7] with permission); (C): the photorespon-sivity of the BP p-n junction shown in (B), here 0◦ representsthe “armchair” direction and 90◦ represents the “zigzag” direc-tion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10xFigure 1.7 The strain anisotropy of BP (Source: Ref [8]. Copyright 2016American Chemical Society Reprinted with permission.). (A):Illustration of applying uniaxial tensile strain on the BP sample.(B): Uniaxial “AC” (“armchair”) and “ZZ” (“zigzag”) tensilestrain-induced Raman shifts of the A1g, B2g and A2g modes. (C):Definitions of the P-P bond lengths of “R1”, “R2” and “R3”.(D): Strain-induced changes of the bond lengths. . . . . . . . 12Figure 1.8 Mechanical exfoliation of black phosphorus. (A): A 7.5-nm-thick BP sample based field-effect-transistors (Figure is reprintedfrom [3] with permission); (B): Raman spectra of single-layerand bilayer phosphorene and bulk black phosphorus films. (Fig-ure is reprinted from [9] with permission). . . . . . . . . . . . 14Figure 1.9 (A): Illustration of liquid exfoliation method (Figure is reprintedfrom [10] with permission). (B): typical BP products preparedby liquid exfoliation method (Figure is reprinted from [11]with permission). . . . . . . . . . . . . . . . . . . . . . . . . 15Figure 1.10 The few-layer BP prepared by plasma thinning. (Figure isreprinted from [12] with permission) . . . . . . . . . . . . . 17Figure 1.11 The illustration of the layer-by-layer nano-patterning and thin-ning (by removing the oxidized byproduct through water rins-ing) of BP. Figure is reprinted from [13] with permission. . . 18Figure 1.12 The few-layer BP prepared by using chemical vapor deposition(CVD) method. (A): the morphology and thickness profile ofthe final product; (B): the comparison of the Raman spectra ofthe exfoliated BP, prepared amorphous BP and prepared crys-talline BP. Figure is reprinted from [14] with permission. . . 19Figure 1.13 The illustration of the few-layer BP prepared by using pulsedlaser deposition (PLD) method. Figure is reprinted from [15]with permission . . . . . . . . . . . . . . . . . . . . . . . . 21Figure 1.14 (1) - (4) represent the Raman spectra of amorphous BP, redphosphorus, BP prepared by PLD (Pulsed Laser Deposition)deposition and exfoliated bulk BP. Figure is reprinted from[15] with permission. . . . . . . . . . . . . . . . . . . . . . . 23xiFigure 1.15 The progress of ARPRS study using different excitation wave-lengths in determining CO of BP: (1) [16]; (2) [17]; (3) [18];(4) [19]; (5) this work. . . . . . . . . . . . . . . . . . . . . . 25Figure 1.16 ARPRS with different excitation wavelengths performed on a20-nm-thick BP sample on SiO2 / Si substrate. Figure is reprintedfrom [18] with permission. . . . . . . . . . . . . . . . . . . . 26Figure 1.17 ARPRS with different excitation wavelengths performed on BPsamples with various thickness on SiO2 / Si substrate. Figureis reprinted from [19] with permission. . . . . . . . . . . . . 27Figure 1.18 The energy-level diagram showing the states involved in Ra-man spectra. Figure is reprinted from Wikipedia under theGNU Free Documentation License. . . . . . . . . . . . . . . 29Figure 1.19 The energy-level diagram of the Stokes band of Raman scat-tering. Figure is reprinted from [2] with permission. . . . . . 29Figure 1.20 Figures (A) - (C) are reprinted from [20] with permission. (A):Illustration of the polarized Raman scattering light for the backscattering geometry. (B): Illustration of the polarization of thescattered light being “parallel” to that of the incident light. (C):Illustration of the polarization of the scattered light being “per-pendicular” to that of the incident light. . . . . . . . . . . . . 31Figure 1.21 Raman-active vibrational symmetries and Raman tensors forthe crystal symmetry classes. Figure is reprinted from [21]with permission. . . . . . . . . . . . . . . . . . . . . . . . . 33Figure 1.22 Illustration of the “θ” angle, the “zigzag” direction of the BPsample and the reading angle “α” of the rotation stage. Thered arrows indicate the polarization of the incident laser. . . . 36Figure 1.23 Illustration of Si (100) substrate. . . . . . . . . . . . . . . . . 37Figure 1.24 The ARPRS results of the polarized Raman measurements of aSi (100) wafer measured under under the “parallel” and “cross”configuration. Image courtesy of David Tuschel, Horiba Sci-entific, USA. . . . . . . . . . . . . . . . . . . . . . . . . . . 38xiiFigure 1.25 (A): Sublimation thinning of a MoS2 flake: mechanically ex-foliated bilayer MoS2 flake was thinned to single layer afterannealed at 330 C for 15 h. (Figure is reprinted from [22]with permission); (B) sublimation thinning of Bi2Se3: Bi2Se3plates prepared by deposition with an original thickness of 3.5µm was heated at 510 ◦ for 10 minutes. (Figure is reprintedfrom [23] with permission); (C): sublimation thinning of WS2:mechanically exfoliated WS2 flakes listed in the upper rowwere held at 600 ◦ for 45 minutes, and the Ar gas with a flowrate of 300 cc / min was purged in to protect the sample. Atlast, the figures in the lower row demonstrated that those flakeswere thinned down to 8 to 45 nm from left to right. Figure isreprinted from [24] with permission. . . . . . . . . . . . . . . 40Figure 1.26 (A): Observation of decomposition of BP at 400 ◦C by usingtransmission electron microscopy (TEM) (Figure is reprintedfrom [25] with permission); (B): the illustration of the de-composition mechanism of BP: detachments with phosphorusdimers (Figure is reprinted from [26] with permission). . . . 41Figure 2.1 Illustration of the sublimation process of BP. . . . . . . . . . 47Figure 2.2 (A) - (D): Observation of the layer-layer-layer sublimation pro-cess of BP at 600 K. The different colors in the area 1 and area2 indicated different thicknesses. . . . . . . . . . . . . . . . 49Figure 2.3 AFM and optical images of a trapezoid BP sample. (A) and(B): before heating; (C) and (D): after heating in nitrogen at550 K for 195 mins. The dispersed white pits on the samplesurface were attributed to the fast surface reactions with oxy-gen and moisture in the air [21] during sample transportationto the AFM and during the AFM measurements in the air. Thered dashed lines indicated the edges of the samples and thenumbers next to them are edge dimensions. . . . . . . . . . . 51xiiiFigure 2.4 (A):the optical image of this BP flake before 500 K heating pro-cess, and the inset AFM line-scan profile indicated its thick-ness to be 36 nm; (B) - (C): Spatial Raman maps of A1g andA2g peak intensity acquired on a 36-nm-thick BP flake; (D): theoptical image of this BP flake after 500 K heating process for120 minutes, , and the inset AFM line-scan profile indicatedits thickness to be 13 nm; (E) - (F): Raman images (A1g and A2gmodes) of this BP flake after 500 K heating process; . . . . . . 52Figure 2.5 (A): the optical image of this BP flake before 550 K heatingprocess, and the inset AFM line-scan profile indicated its thick-ness to be 85 nm; (B) - (C): Spatial Raman maps of A1g and A2gpeak intensity acquired on an 85-nm-thick BP flake; (D): theoptical image of this BP flake after 550 K heating process for25 minutes, and the inset AFM line-scan profile indicated itsthickness to be 49 nm; (E) - (F): Raman images (A1g and A2gmodes) of this BP flake after 550 K heating process; . . . . . . 53Figure 2.6 Sublimation study of BP at 450 K (∼ 180 ◦C). (A): the BPsamples heated at 450 K for 1 minute; (B): the BP samplesheated at 450 K for 61 minutes; (C): zoom in of the sample inthe center of (A); (D): zoom in of the sample in the center of (B). 55Figure 2.7 Sublimation study of BP at 500 K (∼ 230 ◦C). (A): the BPsamples heated at 500 K for 1 minute; (B): the BP samplesheated at 500 K for 61 minutes; (C): zoom in of the sample inthe center of (A); (D): zoom in of the sample in the center of (B). 56Figure 2.8 Sublimation study of BP at 550 K (∼ 280 ◦C). (A): the BPsamples heated at 550 K for 1 minute; (B): the BP samplesheated at 550 K for 61 minutes. . . . . . . . . . . . . . . . . 57Figure 2.9 Sublimation study of BP at 600 K (∼ 330 ◦C). (A): the BPsamples heated at 600 K for 1 minute; (B): the BP samplesheated at 600 K for 41 minutes; (C)-(K): the zoom-in opticalimages of the sample in the center at (A) heated at 600 K for1, 21, 31 and 41 minutes, respectively. . . . . . . . . . . . . . 58xivFigure 2.10 Sublimation study of BP samples on another piece of Si waferat 600 K (∼ 330 ◦C). (A): the BP samples heated at 600 K for1 minute; (B): the BP samples heated at 600 K for 46 minutes;(C)-(K): the zoom-in optical images of the sample in the centerat (A) heated at 600 K for 1, 11, 20, 30, 40, 42, 43, 44 and 46minutes, respectively. . . . . . . . . . . . . . . . . . . . . . . 59Figure 2.11 Sublimation study of BP at 620 K (∼ 350 ◦C). (A): the BPsamples heated at 620 K for 1 minute; (B): the BP samplesheated at 620 K for 9 minutes; (C)-(H): the zoom-in opticalimages of the marked sample in (A) heated at 620 K for 1, 2,3, 4, 5, 6, 7, 8 and 9 minutes, respectively. . . . . . . . . . . . 60Figure 2.12 Sublimation study of BP at 670 K (∼ 400 ◦C). (A): the BPsamples heated at 670 K for 1 minute; (B): the BP samplesheated at 670 K for 22 minutes; (C)-(G): the zoom-in opticalimages of the marked sample in (A) heated at 670 K for 1, 5,10, 15 and 22 minutes, respectively. . . . . . . . . . . . . . . 61Figure 2.13 (A): Optical image of the 36-nm-thick sample used in Section2.2.1.3 (P 52); (B): 3D AFM images of the 36-nm-thick sampleshown in (A); (C): Optical image of this sample after thinningdown to 18-nm-thick; (D): 3D AFM images of the 18-nm-thickshown in (C). . . . . . . . . . . . . . . . . . . . . . . . . . . 63Figure 2.14 (A): Optical image of a prepared 10-nm-thick sample on poly-imide substrate; (B): 3D AFM images of the 10-nm-thick sam-ple shown in (A); (C): Optical image of a prepared 20-nm-thicksample on polyimide substrate: (D): 3D AFM images of the 20-nm-thick sample shown in (C). . . . . . . . . . . . . . . . . . 64Figure 2.15 (A): Optical image of sample A; (B): the Raman spectrum withthe maximum intensity of A1g mode and the minimum intensityof A2g mode; . . . . . . . . . . . . . . . . . . . . . . . . . . . 67Figure 2.16 (A): Optical image of sample B; (B): the Raman spectrum withthe maximum intensity of A1g mode and the minimum intensityof A2g mode; . . . . . . . . . . . . . . . . . . . . . . . . . . . 69xvFigure 2.17 (A): Optical image of sample A to be heated at 500 K; (B)the illustration of the interrrupted heating process (5 cycles)to be performed at sample A; (C): Optical image of sample Bto be heated at 500 K. (D): the illustration of periodic heatingprocess (6 cycles) to be performed on sample B. . . . . . . . . 70Figure 2.18 (A): Thickness in different areas of sample A as a function oftime, and (B): the thickness in different areas of sample B asa function of time. At each red dot, a Raman spectrum wascollected (analysis of those collected Raman spectra would bediscussed in Section 3.3.2 (P 118) of Chapter 3). . . . . . . . 71Figure 2.19 (A): Optical image of area 3 in Sample A, and the inset AFMprofile indicated it was thinned down to 5-nm-thick; (B): Ra-man spectrum of the 5-nm-thick BP sample showing the 3 char-acteristic peaks of BP. . . . . . . . . . . . . . . . . . . . . . 72Figure 2.20 Thickness-color map: the colors were extracted from thosesample of which the thicknesses had been measured by AFM; . 73Figure 2.21 (A) & (B): Optical images of Sample C before and after contin-uous heating process at 500 K, the heating time was 221 min-utes; (C) & (D): Optical images of Sample D before and aftercontinuous heating process at 550 K, the heating time was 51minutes. The inset profile in each picture is the line-scan AFMprofile indicating the thickness of this sample. . . . . . . . . 75Figure 2.22 Sample C was annealed at 500 K for 221 minutes. . . . . . . . 76Figure 2.23 In-situ Raman spectra of the sample C at 500 K. . . . . . . . 76Figure 2.24 Sample D was annealed at 550 K for 51 minutes. . . . . . . . 77Figure 2.25 In-situ Raman spectra of the sample D at 550 K. . . . . . . . 77Figure 2.26 Comparison of thinning rates of sample A and C. . . . . . . . 78Figure 2.27 Comparison of thinning rates of sample B and D. . . . . . . . 79Figure 2.28 (A): Optical image the BP sample thinned down to 2-nm-thick;(B): Raman spectrum collected at the circled area in (A); (C)& (D): Spatial Raman maps of sample shown in (A). . . . . . 80xviFigure 2.29 (A): Optical image the BP sample thinned down to 2-nm-thick,and the inset figures represent the spatial AFM image and thick-ness profile of circled area. (B) - (D): Optical contrast analysisof Fig 2.28 (A), which was splitted into R, G and B channels,and the G channel figure in (C) was selected for analysis. . . 81Figure 2.30 (A): Optical image a BP sample thinned down to 3 and 4 -layers, and the inset figure represents the original BP sample.(B); Raman spectra collected in circled area in (A); (C) - (E):Spatial Raman maps of the BP sample shown in (A); (F) - (H):Optical contrast analysis of (A), which was splitted into R, Gand B channels, and the G channel figure in (C) was selectedfor analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . 83Figure 2.31 (A): Optical image of a BP sample thinned down to 4 layers,and the inset figure represents the original BP sample. (B); Ra-man spectra collected in circled area in (A). (C) - (E): SpatialRaman maps of the BP sample shown in (A). (F) - (H): Opti-cal contrast analysis of (A), which was splitted into R, G andB channels, and the G channel figure in (C) was selected foranalysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85Figure 2.32 (A): Optical image of a BP sample thinned down to 4 and dou-ble layers, and the inset figure represents the original BP sam-ple. (B); Raman spectra collected in circled area in (A). (C) -(E): Spatial Raman maps of the BP sample shown in (A). (F) -(H): Optical contrast analysis of (A), which was splitted into R,G and B channels, and the G channel figure in (C) was selectedfor analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . 86Figure 2.33 (A): Optical image of a BP sample thinned down to singlelayer. (B): Raman spectra of this sample collected at 3 regions. 87Figure 3.1 (A): AFM profile of a blank Si wafer, which indicates its rough-ness to be ∼ 0.6 nm; (B): AFM profile of a blank polyimidewafer, which indicates its roughness to be ∼ 1 nm; . . . . . . 92xviiFigure 3.2 Illustration of ARPRS experiment setup: the optical path of theRaman system under “parallel” configuration. . . . . . . . . 93Figure 3.3 (A): Optical image of the BP sample on SiO2 / Si substraterotated at the position with maximum Raman intensity of A1gmode and (B): at the position with the maximum Raman inten-sity of A2g mode. The circled area indicates where the Ramanspectra were collected; . . . . . . . . . . . . . . . . . . . . . 95Figure 3.4 (A) & (C): Raman spectra of the BP sample on SiO2 / Si sub-strate shown in Fig 3.3 (A), in which the A1g mode showedmaximum Raman intensity; (B) & (D): Raman spectra of theBP sample shown in Fig 3.3 (B), in which the A1g mode showedmaximum Raman intensity. Raman measurements shown in(A) & (B) were performed using the excitation wavelength of442 nm while those in (C) & (D) were by 633 nm. . . . . . . 96Figure 3.5 ARPRS results with 442 nm excitation: (A): the polar diagramof A1g mode; (B): the polar diagram of B2g mode; (C): the polardiagram of A2g mode. ARPRS results with 633 nm excitation:(D): the polar diagram of A1g mode; (E): the polar diagram ofB2g mode; (F): the polar diagrams of A2g mode. The arrows infigures indicate the two axes of BP (zigzag and armchair). . . 98Figure 3.6 “Sample A”. Raman intensities of A1g and A2g modes and un-derlying (100) Si wafer as functions of the rotation angle. . . . 123Figure 3.7 “Sample B”. Raman intensities of BP A1g and A2g modes andunderlying Si(100) as functions of the rotation angle. . . . . . 124Figure 3.8 Room temperature Raman intensity ratios of SiA2g as a functionof thickness measured by AFM, here the blue line representsthe sample A and the red line represents sample B. . . . . . . 126Figure 3.9 High temperature Raman intensity ratios of SiA2g as a functionof measured thickness, here the blue line represents sample Atreated at 500 K and the red line represents sample B at 550 K. 127Figure 3.10 High temperature Raman intensity ratios of SiA2g as a functionof measured thickness, here the blue line represents sample Atreated at 500 K and the red line represents sample B at 550 K. 128xviiiGlossaryAFM Atomic Force MicroscopyARPRS Angle-resolved Polarized Raman SpectroscopyBP Black PhosphorusCO Crystallographic OrientationIR Infra-redIC integrated circuitTMDCS Transition Metal Dichalcogenides2D 2-Dimensional3D 3-DimensionalCVD Chemical Vapor DepositionPLD Plasma Laser DepositionTEM Transmission electron microscopyxixAcknowledgments”To explain all nature is too difficult a task for any one man or even for anyone age. ’Tis much better to do a little with certainty and leave the rest for othersthat come after you.”— Isaac NewtonDuring my master’s study, I am very happy to have many great people stand-ing on my side all the time.I would like to express my gratitude to my supervisor, Dr. Guangrui Xia. Ithas been my honor to be her fourth M.A.Sc. student. Under her supervision, Ilearned how to conduct good research. Her guidance, patience, trust and time en-abled my research to be productive and very enjoyable. I would like to thank Mr.David Tuschel from Horiba Scientific (U.S.A), for offering me generous advice onboth Raman theory and experimental operations of Raman spectrometer. I wouldlike to thank Texas Instruments (TI, U.S.A), for sponsoring me with the 2-year re-search fellowship during my M.A.Sc study. I would like to thank Professor JoshuaFolk from UBC Physics & Astronomy and Professor Chad Sinclair from UBC Ma-terials Engineering to sit on my thesis committee. I would like to thank ProfessorXi Ling from the Department of Chemistry at Boston University (U.S.A) to giveme much advice in research.I would like to thank Dr. Yiheng Lin, my former colleague and friend, forhis kind and patient instructions. I felt very thankful for countless constructivediscussions in all kinds of topics with him. I would like to thank Dr. S. ArashxxSheikholeslam, Mr. Rui Yang, Dr. Zihe Ren, Ms. Zenan Jiang, Mr. Joshua Cantinfor their kind help in my research and life. I feel surely thankful to my colleagues,Mr. Jiaxin Ke, Mr. Guangnan Zhou, Mr. Yunlong Zhao and Mr. Qian Song forhelpful discussions.I would give my sincere thanks to Professor Olav Slaymaker from UBC Ge-ography, for his kind guidance in helping me to keep in faith and go through thehardest time in my master’s study. I would thank to Gordon and Ute Carkner, sin-cerely for praying and sharing the love of god, for teaching me how to be a kindperson. I would like to thank those whose names haven’t been mentioned here, Iknow many people helped me more or less in my life. I will contribute my best tothe society.Last, but the most importantly, I would like to thank my family, especially myparents, thank you for your strong and all-round support in my whole life. Loveyou!xxiChapter 1Introduction“I don’t know how to do this on a small scale in a practical way, but I do knowthat computing machines are very large; they fill rooms. Why cant we make themvery small, make them of little wires, little elements, and by little, I mean little?”— Richard Feynman1.1 Background And MotivationsSince the physical limitation of the dimensionality of bulk Si impedes theshrinkage of transistors, and hurdles the promotion of device performance accord-ing to Moore’s law. According to the 2016 International Technology Road mapfor Semiconductors (ITRS) [27], the adoption of low-dimensional new materialsbecomes a critical issue.Two-dimensional (2D) materials, such as graphene, black phosphorus, molyb-denum disulfide (MoS2) and other members of the transition metal dichalcogenides(TMDCS) family, represents the potential solutions of materials beyond conven-tional silicon for the further development of scaling-down-dimensions in “Morethan Moores Law” [28][29].2D BP was first isolated from bulk BP in 2013 [3] by the mechanical exfoli-ation method [30], and it has been shown to have many unique and useful properties1in flexible and ultra-thin electronics [31][32][33][34] and photonics [35][36][37][38][39][40].1.1.1 BP Structure: Layer-stacked 2D Semiconductor MaterialSemiconducting orthorhombic black phosphorus (BP), like graphene, is a layer-stacked 2-dimensional (2D) nanomaterial with sp3-hybridized phosphorus atomscovalently bonded to a puckered structure via weak van der Waals forces betweenlayers [41]. The atomic structure of BP is shown in Fig 1.1. Bulk BP was first syn-thesized in 1914 by applying high hydrostatic pressure of 1.2 GPa at an elevatedtemperature of 200 ◦C on white phosphorus [42].Figure 1.1: The atomistic ball-stick models of few-layer black phosphorus.2In 2013, Li et al. [3] first reported the successful mechanical exfoliation offew-layer black phosphorus and preparation of a BP based field-effect transistor(See Fig 1.2). After that work, the atomically thin BP, which is also termed as“few-layer phosphorene”, attracted great attentions in studying it as a new 2D semi-conductor material.Figure 1.2: The first BP based transistor which was fabricated on a piece of6.5-nm-thick mechanically exfoliated BP sample. Figure is reprintedfrom [3] with permission.31.1.2 Layer-dependent Tunable Band Gap, High Carrier Mobility,Fast And High Photo-responseBP always has a direct band gap at the Γ point of the Brillouin zone regard-less of layer numbers, which is different from the single-layer-only direct bandgapof semiconducting transition metal dichalcogenides (TMDCS) such as MoS2, WS2and WSe2 [43][44][45], and the zero band gap of graphene. In Fig 1.3, the bandgap range [46] of BP is compared with other major 2D and bulk semiconductors.The energy band gap of single layer BP of 2.05 eV has been measured by scan-ning tunneling spectroscopy (STS) [47]. With the increasing number of layers, theband gap decreases and finally reaches 0.3 eV for bulk BP (thickness > 10 nm)[3][48][49], which covers the gap between graphene and TMDCS for 2D materials.In the range of 0.3 - 1 eV, BP is the only 2D semiconductor material that is suitablefor thermal imaging, thermoelectrics [50][51], infra-red detection [52], telecom[35] and photovoltaics [35][53] applications. In addition, Fig 1.4 (A) compares 2major figures of merits for materials for transistor applications: carrier mobilityand “current on/off ratio”. It can be seen that BP not only has a carrier mobilityclose to silicon but also has a reasonable current on/off ratio; in Fig 1.4 (B), itshows that the photo-response of BP has high responsivity of photo-current likegraphene while also has a fast response time like TMDCS. Overall, those propertiesenable BP to be a promising candidate for applications in field-effect transistors[3][54][55][56][57][58] and photo-detectors [59][60].4Figure 1.3: The band gap range of 2D and conventional bulk semiconductors.Figure is reprinted from [4] with permission.5Figure 1.4: (A): mobility and current on/off ratio of BP; (B): the photo-responsivity and response time of BP. Figure is reprinted from [4] withpermission.61.1.3 Strong In-plane Anisotropies Of Black Phosphorus1.1.3.1 Anisotropic Transport PropertyBecause of the puckered honeycomb orthorhombic structure, many basic phys-ical properties of BP show strong in-plane anisotropies. For instance, by applyingthe same biasing voltage, the “armchair” direction provides higher drive currentthan the “zigzag” direction, since the effective carrier masses in the “armchair”direction are significantly lighter than those in the “zigzag” direction [5]. Thedefinition of “armchair” direction can be seen in Fig 1.1, which is the “z axis” per-pendicular to the ridge direction, and the “zigzag” one is the “x axis” parallel withthe ridge direction.Fig 1.5 (A) illustrates that, in the “armchair” direction (the “x” direction), dueto similar carrier masses of electron (0.17 m0) and holes (0.16 m0), the ION (see“blue” and “red” circles in the upper part of the diagram) of n- and p-type deviceare almost the same; but in the “zigzag” direction (the “y” direction), because ofthe heavier effective masses of holes, the p-type device has notably lower ION thanthat of the n-type device.7Figure 1.5: (A): Material parameters of single-layer BP; (B): the current char-acteristics of the n-type (solid) and p-type (dash) BP field-effect transis-tors at the x- (“zigzag”) and y-direction (“armchair”). m0 is the freeelectron mass which has a value of 9.1 × 10−31 kg. Figure is reprintedfrom [5] with permission.8Despite of this high-degree anisotropy of the transport property of BP, the rel-atively high carrier mobility (see Fig 1.3 (A)), and the large on / off ratio makesencapsulated few-layer BP promising for complementary metal-oxide semiconduc-tor (CMOS) devices [61], especially in ultra-thin and wearable electronics.1.1.3.2 Anisotropic Optical PropertyBP is a linear dichroism crystal [6][59]. In other words, it shows a stronganisotropy of the light absorption. For example, in Fig 1.6 (A), Qiao et al. [6]studied the light absorption of the “x” (“armchair”) and “y” (“zigzag”) direction ofBP with different thickness, and indicated that absorption edges were from 1.55 eV(monolayer BP) to 0.46 eV (bulk BP) for the “x” direction; while those for the “y”direction were from 3.14 eV (monolayer BP) to 2.76 eV (bulk BP). Moreover, in2015, Yuan et al. [7] first investigated the different broadband photo-responses atthe “zigzag” and “armchair” directions by preparing and measuring a black phos-phorus vertical p-n junction (see in Fig 1.6 (B)), and their results (see in Fig 1.6(C)) indicated that the photoresponsivity of the “armchair” direction was notably3.5 times greater than that of the “zigzag” direction.9Figure 1.6: (A): Light absorption of BP with different thickness (Figure isreprinted from [6] with permission) , here the “x” denotes for the “arm-chair” direction and the “y” denotes for the “zigzag” direction; (B): ablack phosphorus vertical p-n junction (Figure is reprinted from [7] withpermission); (C): the photoresponsivity of the BP p-n junction shown in(B), here 0◦ represents the “armchair” direction and 90◦ represents the“zigzag” direction.101.1.3.3 Anisotropic Thermal ConductivityMoreover, the thermal conductivity of BP also shows a strong in-plane anisotropy.For instance, Luo et al. [62] investigated the “zigzag” and “armchair” thermal con-ductivities of different thickness of BP, which showed that, for BP thicker than 15nm, the “armchair” thermal conductivity has a value of ∼ 20 Wm−1K−1 while the“zigzag” one is ∼ 40 Wm−1K−1. However, for BP thinner than 15 nm, the “arm-chair” and “zigzag” thermal conductivities decrease to ∼ 10 Wm−1K−1 and ∼ 20Wm−1K−1, respectively.1.1.3.4 Anisotropic Strain-induced Raman ResponseRecently, Li et al. [8] investigated the anisotropic Raman shifts of the vibra-tional modes of BP by applying a uniaxial strain along either the “zigzag” directionor the armchair direction. In Fig 1.7 (A), it illustrates the uniaxial strain is appliedto the specific “zigzag” or “armchair” direction of the BP sample through bendingthe flexible polyethylene terephthalate (PET) substrate after determining the CO. InFig 1.7 (B), their experiment results of the strain-induced Raman spectra indicatedthat, when the uniaxial tensile strain is applied along the “AC” (“armchair”) direc-tion, the Raman peak positions of in-plane vibrational A2g and B2g modes can beshifted by -3 and -11 cm−1/%, while the out-of-plane vibrational A1g mode almostremains at its original peak position. On the other side, when the uniaxial tensilestrain is applied along the “ZZ” (“zigzag”) direction, the A1g mode can be shiftedby -3 cm−1/% while A2g and B2g modes don’t undergo obvious shifts of the peakpositions. This anisotropy of strain-induced Raman shifts of the 3 Raman modesof BP can be explained by the bond lengths of the P atoms. In Fig 1.7 (C), the“R1” and “R2” represent the in-plane atomic bonding while the “R3” representsthe atomic bonding of the interlayer. In Fig 1.7 (D), it shows that, except for “R2”,the bond lengths of “R1” and “R3” differ under uniaxial “ZZ” (“zigzag”) strainand “AC” (“armchair”) strain, and obviously the bond length of “R1 (the in-planeatomic bond) is affected by the uniaxial tensile strain “ZZ (“zigzag”), correspond-ing to the relative large red shifts of the in-plane B2g and A2g modes shown in Fig1.7 (B), but for the bond length of “R3”, the “AC” (“armchair”) strain affects more11than the “ZZ” strain does, and considering the “AC” and “ZZ” strain affect it muchless than the in-plane “R1, the relative weak red-shifts of the A2g mode under “ZZ”strain shown in Fig 1.7 (B) can be explained.Figure 1.7: The strain anisotropy of BP (Source: Ref [8]. Copyright 2016American Chemical Society Reprinted with permission.). (A): Illustra-tion of applying uniaxial tensile strain on the BP sample. (B): Uniaxial“AC” (“armchair”) and “ZZ” (“zigzag”) tensile strain-induced Ramanshifts of the A1g, B2g and A2g modes. (C): Definitions of the P-P bondlengths of “R1”, “R2” and “R3”. (D): Strain-induced changes of thebond lengths.121.2 Current Challenges In BP Study1.2.1 Degradation Of BP In AirSo far, there remain 3 major challenges for BP research: the air-stabilizationand the controllable fabrication of atomically thin BP layers [4]. The oxidationconditions of BP have been widely studied before: Favron et al. [63] first indicatedthat light, oxygen and water were three factors cause the oxidation of BP; Luo etal. [64] reported that with 5% oxygen/Ar or 2.3% H2O/Ar, the oxidation rate was< 5A˚ for 5h from XPS results; Wang et al. [65] reported BP could remain stablein pure water without the presence of oxygen molecules from nuclear magneticresonance (NMR) spectroscopy results; Li et al. [66] reported that heating couldremove the metastable oxygen adsorbed on the surface of BP. In order to solve thisproblem, the passivation of BP was successfully achieved via both chemical mod-ification method like covalent aryl diazonium functionalization [67] and physicalencapsulation methods such as atomic layer deposited dielectric passivation [68]or h-BN passivation [64].1.2.2 Fabrication Of Large-scale Few-layer Crystalline BPLet us look at the current fabrication methods in preparing few-layer blackphosphorus thin films. The top-down and bottom-up methods are two approachestowards few-layer BP. Current top-down methods towards ultra-thin BP such asmechanical exfoliation by ScotchTM tapes [3][59], shear exfoliation [69][70][71]in liquids, plasma thinning [12][72][73], anodic oxidation and water rinsing [13]still couldn’t be used for scalable and controllable production of large uniform crys-talline few-layer BP thin films. In addition, the quality (uniformity, scalability andcrystallinity) of few-layer BP made by pulsed laser deposition [14][15] and chemi-cal vapor deposition [14] (CVD) have not been sufficient for industry applications.Hence, developing a scalable and controllable massive fabrication method of few-layer BP thin films with good uniformity and crystallinity will greatly promote thewide application of BP. The details of the preparation methods of BP mentionedabove were discussed below.13Mechanical ExfoliationMechanical exfoliation of 2D materials such as graphene by ScotchTM tapes,was first demonstrated by Novoselov et al. [30], which is the most conventionalmethod in preparing few-layer BP samples. In Fig 1.8 (A), the first BP-based field-effect transistor was fabricated on an 6.5-nm-thick exfoliated black phosphorusflake with an area around 50 µm2; and Fig 1.8 (B) shows the comparison amongthe exfoliated single-layer, bilayer and bulk BP. However, the poor scalability anduniformity of prepared samples via the ScotchTM tapes based exfoliation methodmade it unsuitable for industrial applications.Figure 1.8: Mechanical exfoliation of black phosphorus. (A): A 7.5-nm-thickBP sample based field-effect-transistors (Figure is reprinted from [3]with permission); (B): Raman spectra of single-layer and bilayer phos-phorene and bulk black phosphorus films. (Figure is reprinted from [9]with permission).14Sonication Of BP In Liquid SolutionsFigure 1.9: (A): Illustration of liquid exfoliation method (Figure is reprintedfrom [10] with permission). (B): typical BP products prepared by liquidexfoliation method (Figure is reprinted from [11] with permission).Sonication of BP in liquid solutions [69][11][74][75] is another mechanicalexfoliation method using liquid solutions for the isolation of BP. The exfoliationmechanism can be illustrated as seen in Fig 1.9 (A): the solvent molecules act aswedges and peel a piece of monolayer BP from bulk BP. Fig 1.9 (B) shows thetypical BP products prepared by liquid exfoliation method which have their areasless than 1 µm2. In addition, previous studies on liquid exfoliation of BP weresummarized in Table 1.1, which indicates this method is incapable of preparingsingle-layer BP which has a thickness of 0.53 nm [49][63]. Though the liquid ex-foliation method could be used for massive and economic production of few-layercrystalline BP, the lack of uniformity and scalability of final products still hurdlesits future prospection for applications.15Table 1.1: Previous studies on liquid exfoliation of BP [1]. ?aq.: deoxy-genated water.Solutions Thickness of prepared BP (nm) Reference numberOrganic solvantsDMSO 15 - 20 [11]NMP 1 - 5 [69]NMP < 10 [74]NMP 17.6 [76]Saturated NaOH NMP solution 2.8 - 5.3 [75]CHP 9.4 ± 1.3 [70]AqueousDistilled water 2 [77]?aq. with 1% w/v Triton X-100 solution < 20 [78],[79]?aq. with 2% (wt/vol) SDS 4.5 [76]Ionic liquids[BMIM][TfO] 8.5 - 12.8 [80][HOEMIM][TfO] 3.6 - 8.9 [80]16Plasma ThinningIn 2014, Lu et al. [12] reported the successful preparation of monolayer BPfilm. In Fig 1.10, it shows the BP products prepared by the plasma thinning method:the total area of the products were less than 300 µm2 and the stacked flakes indi-cated the non-uniformity of BP products. Some other works on plasma thinningof black phosphorus were also reported [81]. The fabrication process can be de-scribed as: Ar+ plasma was used to etch away the surface phosphorus atoms suchthat a thick flake could be thinned down. However, the size of the plasma beamand defects introduced by Ar+ plasma bombardment cause the non-uniformity andlow crystallinity of products.Figure 1.10: The few-layer BP prepared by plasma thinning. (Figure isreprinted from [12] with permission)17Local Anodic Oxidation And Water RinsingRecently, Liu et al. [13] reported the layer-by-layer thinning of black phos-phorus via anodic oxidation and water rinsing method. In Fig 1.11, it shows thatthe simultaneous thinning of BP layers in each patterned area which has an area of1 µm2. However, they didn’t thin the sample down to a few layers. The fabricationprocess can be described as follows: By scanning a BP sample with an AFM tip,the oxidation of the patterned area was realized by applying DC (direct current)bias voltages (the “anodic oxidation”) through a tip in the ambient environment;the generated P-O product absorbed water from the ambient environment, and thenthe liquid-phase patterning byproduct could be easily removed by water rinsing.This method could provide precise control of thinning rates while the uniformityand crystallinity of BP sample are retained. Though this method has its advantagesin thickness controllability of bulk BP (thickness more than 10 nm [49]) , which aresuitable for sub-micro and nano -scale fabrication of prototype devices, the limita-tions of its scalability, high consumption of expensive AFM tips and introduction ofoxygen impurities make it unable to be used for massive and economic productionof few-layer crystalline BP thin film.Figure 1.11: The illustration of the layer-by-layer nano-patterning and thin-ning (by removing the oxidized byproduct through water rinsing) ofBP. Figure is reprinted from [13] with permission.18Chemical Vapor Deposition (CVD)In 2016, few-layer BP prepared by the chemical vapor deposition (CVD)method was reported by Smith et al. [14]. In Fig 1.12 (A), it shows that theBP products has an area of less than 1 µm2 and the thickness of the sample isnot uniform. The fabrication process can be described as follows: an amorphousred phosphorus film was first grown on a clean surface in vacuum environment andthen the Sn/SnI4 mixture was used as catalyst to turn the amorphous red phosphorusinto black phosphorus under different temperatures and pressures. It is importantto note that, the red phosphorus is made of an amorphous network of phosphorusatoms linked with covalent bonds [82], and BP has a puckered honeycomb struc-ture composed of covalently bonded phosphorus atoms. However, in Fig 1.12 (B),the Raman spectra indicated the qualities of the final BP such as the uniformity andcrystallinity still were not as good as the exfoliated BP.Figure 1.12: The few-layer BP prepared by using chemical vapor deposition(CVD) method. (A): the morphology and thickness profile of the finalproduct; (B): the comparison of the Raman spectra of the exfoliatedBP, prepared amorphous BP and prepared crystalline BP. Figure isreprinted from [14] with permission.19Pulsed Laser Deposition (PLD)A pulsed laser deposition (PLD) method for preparing few-layer BP was alsoreported [15]. In Fig 1.13, it shows that the BP product has an area of less than 30µm2 and the surface of it is non-uniform. The fabrication process can be describedas follows: the amorphous few-layer BP film was prepared by ablating a bulk BPcrystal as a target material by a KrF pulsed laser ( λ = 248 nm ) in vacuum envi-ronment. It is important to note that black phosphorus has three crystalline phaseswhich are orthorhombic [83], rhombohedral [83], and simple cubic [84], and oneamorphous form [85]. Since the final products were in the form amorphous BP inthat work, which indicated that this method was incapable of preparing crystallinefew-layer BP film.20Figure 1.13: The illustration of the few-layer BP prepared by using pulsedlaser deposition (PLD) method. Figure is reprinted from [15] withpermission21Fig 1.14 shows a comparison of the Raman spectra of BP prepared by me-chanical exfoliation, and the black phosphorus, red phosphorus and amorphousphosphorus films prepared by PLD deposition [15]. First of all, the Raman spec-trum of the mechanically exfoliated BP showed three strong characteristic peaks atabout 360, 438 and 468 cm−1, respectively; secondly, the Raman spectra of the BPfilm grown also showed 3 weaker and broader characteristic peaks, and all peakpositions red-shifted; thirdly, the Raman spectra of the red phosphorus preparedshowed three weak characteristic peaks at aorund 350, 380 and 440 cm−1, respec-tively; fourthly, the Raman spectra of the amorphous phosphorus (see the purpleprofile) showed no characteristic peaks, and signals shown in profile (1) were fromthe Si substrate, as amorphous P has no characteristic peaks. The significant differ-ence between the Raman spectra of crystalline BP and amorphous phosphorus ofthe same chemical composition is primarily because of the presence or absence ofspatial order and long-range translational symmetry, respectively [86]. Therefore,Raman spectroscopy can be used for identifying the crystallinity of prepared BP inthis work.22Figure 1.14: (1) - (4) represent the Raman spectra of amorphous BP, redphosphorus, BP prepared by PLD (Pulsed Laser Deposition) depositionand exfoliated bulk BP. Figure is reprinted from [15] with permission.231.2.3 Debate On The CO Determination Of BP By ARPRSBesides the fabrication method, the fast and non-destructive determination ofthe BP crystallographic orientations still remained unsolved. To date, BP CO couldbe determined by angle-resolved conductance of BP sheet [59][62][16], diffrac-tion pattern by high-resolution transmission electron microscope [19][46], opti-cal absorption [59][19] and ARPRS. Among them, only ARPRS is a fast, conve-nient and non-destructive method. However, because the ARPRS response of BPis quite complicated, which depends on the excitation wavelength and thickness[16][17][18][19], there is a debate on whether or not the CO of BP could be simplydetermined by ARPRS.Fig 1.15 summarized the 4 previous works, which studied ARPRS of BP be-fore this work. In 2014, Wu et al. [16] first investigated the ARPRS of BP by usingthe excitation wavelength of 514.5 nm, and summarized that the CO could be de-termined according to the polar diagrams of the A1g and A2g modes of BP under“parallel” configuration: a relatively larger local maximum peak intensity of A2gmode indicates the “armchair” direction, and a relatively smaller local maximumpeak intensity of A2g mode shows at the “zigzag” direction. In 2015, Ribeiro et al.[17] reported the opposite results: when using the excitation wavelength of 532 nmunder “parallel” configuration, a relatively larger local maximum peak intensity ofA2g mode showed “zigzag” direction and a relatively smaller local maximum peakintensity indicated the “armchair” direction. In addition, they also first indicatedthat the ARPRS results of BP varied by the excitation wavelength.In 2015, Kim. et al. [18] investigated the ARPRS of 4 BP samples (5, 65, 70and 90 -nm-thick) by using the excitation wavelength from 442 to 633 nm underthe “parallel” configuration (see Fig 1.16). They first indicated that the CO of BPcould be determined by ARPRS with the laser line of 442 nm according to: 1. bothA1g and A2g modes show clear bow-tie shapes, and the maximum / minimum in-tensities of A1g and A2g modes are orthogonal to each other; and 2. the maximumpeak intensity of A2g indicates the “armchair” direction, and the maximum peakintensity of A1g mode shows at the “zigzag” direction. However, for longer wave-24Figure 1.15: The progress of ARPRS study using different excitation wave-lengths in determining CO of BP: (1) [16]; (2) [17]; (3) [18]; (4) [19];(5) this work.lengths, the shape of A1g mode showed an oval shape (488 nm), a circular shape(514.5 nm & 532 nm) and a peanut shape; for A2g mode, the polar diagrams (488- 633 nm) showed two different local maximum Raman intensities at orthogonaldirections instead of the bow-tie shape (442 nm), and the relatively larger / smallerlocal maximum Raman intensity can appear at the angle either perpendicular (488- 532 nm) or parallel (633 nm) to the angle of the maximum A1g Raman intensity.Therefore, this complexity of ARPRS using wavelength longer than 442 nm makesit hard to identify the CO of BP.Moreover, in 2016, Ling et al. [19] investigated the ARPRS of 13 BP sam-ples (10 to 200-nm-thick) by using excitation wavelengths of 532, 633 and 785 nm(see Fig 1.17). They indicated that the polar diagrams of A1g and A2g of BP werenot only wavelength-dependent but also thickness-dependent. For instance, in Fig1.17, under the 633 nm laser, the A1g mode has a peanut shape (F1, ∼ 5 nm), anoval shape (F7,∼ 40 nm) and a circular shape (F 10,∼ 200 nm) for different thick-ness. What’s more, the relative larger / smaller local maximum Raman intensityof A2g mode also appeared at the angle either perpendicular (F7, ∼ 40 nm) or par-25Figure 1.16: ARPRS with different excitation wavelengths performed on a 20-nm-thick BP sample on SiO2 / Si substrate. Figure is reprinted from[18] with permission.allel (F1, ∼ 5 nm) to the angle of the maximum A1g Raman intensity. They alsoindicated that without the explicit consideration of excitation wavelength and flakethickness, ARPRS could not be used to determine the CO as commonly used pre-viously. However, they didn’t investigate the thickness-dependent ARPRS resultswith the excitation wavelength of 442 nm.Considering the strong thickness dependence of the ARPRS results of BP andthe limited data in Kim et al’s work [18], it is necessary to perform more ARPRSstudies on BP with wider thickness range to confirm the effectiveness of the 442nm laser in CO determination, and provide some theoretical explanations.26Figure 1.17: ARPRS with different excitation wavelengths performed on BPsamples with various thickness on SiO2 / Si substrate. Figure isreprinted from [19] with permission.27Table 1.2: The energy balance sheet for Raman process [2].Initial state energy Final state energy Scattered frequency ωsdetermined by energyconservationPhotons molecule Photons moleculen1h¯ω1 Ei (n-1)h¯ω1+ h¯ωs E f = Ei + h¯ω f i ωs = ω1−ω f i1.3 Basic Knowledge Of This Thesis Work1.3.1 Basic Knowledge of Raman ScatteringRaman Spectroscopy is a very useful tool in studying 2D materials, and willbe used extensively in this work. The Raman effect is an inelastic light scatteringphenomenon first discovered by C. V. Raman [87]. Raman spectroscopy is increas-ingly becoming standard analytical techniques in academic and industrial studies.The whole process can be described as follows: with a laser light of specific wave-length comes into materials; then, most of the laser light is transmitted without anychanges, some scattering happens because of the interactions between photons andatoms / molecules, which can be in either ground or excited rotational-vibrationalstates. As Fig 1.18 shows, the scattering without a change of frequency is calledRayleigh scattering, and that with change of frequency is called Raman scattering.The change of frequency gives information about the rotational and vibrationalmodes of the system. What’s more, the final frequencies can be either ω1−ωi orω1 +ωi according to whether the original states are in ground or excited levels,which can be termed as the Stokes or anti-Stokes bands. In Table 2.1, the energybalance sheet of the Stokes band of Raman scattering following the conservation ofenergy is listed. Stokes Raman scattering is normally much stronger because mostatoms / molecules of the systems are usually in ground states [88]. Therefore, inmost cases, only the Stokes Raman scattering part will be illustrated in a typicalRaman spectrum.28Figure 1.18: The energy-level diagram showing the states involved in Ramanspectra. Figure is reprinted from Wikipedia under the GNU Free Doc-umentation License.Figure 1.19: The energy-level diagram of the Stokes band of Raman scatter-ing. Figure is reprinted from [2] with permission.29As is shown in Fig 1.19, the normal Raman scattering process starts and endsat two explicit stationary states i and f. However the transition state m during theprocess doesn’t have a well-defined value of energy, thus the absorption part hasno energy conservation and is called virtual absorption. The intermediate statem is a virtual state [2]. Thus the energy-level diagram is just for simple illustra-tion of electron-photon and electron-phonon interactions, and is not able to explainthe complicated photon energies and energy states of the atom / molecule that areimplicated in the scattering process. Raman Spectroscopy measures the phononenergies, which depend on many materials properties and conditions such as thebonding strength, crystallinity, temperature, strain, phase, etc. For a material withknown chemical concentrations and a good crystallinity at room temperature andfree of strain, its Raman spectra can be used to identify the allotrope of this mate-rial and corresponding phase.1.3.2 Basic Knowledge Of Polarized Raman SpectroscopyIn this thesis, the Raman system was configured with the backscattering ge-ometry. Fig 1.20 (A) illustrate that: the incident light is linear polarized and thescattered light is circular-polarized after Raman scattering. By adding an analyzerwith an angle parallel or perpendicular to the incident light polarization before thedetector, only the scattered light with a polarization which is parallel (B) or per-pendicular (C) to that of the incident light can be collected.After discussing the Raman geometry and polarization, let us look at the in-tensity of the Raman-scattered light. In general, the Raman intensity of a specificRaman mode I can be described by the form [2]:I ∝ |ei ·α ′ · es|2; (1.1)α ′ is the Raman tensor of a given mode. In an X-Y-Z Cartesian coordinatesystem, the Raman tensor can be written as:30Figure 1.20: Figures (A) - (C) are reprinted from [20] with permission. (A):Illustration of the polarized Raman scattering light for the back scat-tering geometry. (B): Illustration of the polarization of the scatteredlight being “parallel” to that of the incident light. (C): Illustration ofthe polarization of the scattered light being “perpendicular” to that ofthe incident light.[α ′] = α′xx α ′xy α ′xzα ′yx α ′yy α ′yzα ′zx α ′zy α ′zz (1.2)For a single crystal, for each of the 32 crystal classes (symmetry point groups),the form of Raman tensors is summarized in Fig 1.21 [21]. What’s more, the unitvectors ei and es are the light polarization vectors of the incident and scattered light,respectively. Therefore, with proper selected Raman geometries and polarization,the polarized Raman Spectroscopy can be used to study the Raman crystallography.Due to the scattered Raman intensity depends on the CO of the crystal under aspecific configuration of the analyzer, ARPRS was developed to measure the Raman31response of crystal at different orientations in order to determine its CO.32Figure 1.21: Raman-active vibrational symmetries and Raman tensors for thecrystal symmetry classes. Figure is reprinted from [21] with permis-sion.331.3.3 Basic Knowledge Of Phonon Modes Of Black PhosphorusBlack phosphorus has an orthorhombic (D2h, Cmca, #64) structure, and itsphonon modes can be expressed as [16]:Γ= 2Ag+B1g+B2g+2B3g+A1u+2B1u+2B2u+B3u (1.3)The first 4 terms at the right side of Eq. 3.3 are the Raman active modes. Fromgroup theory [21], the Raman tensor for Ag modes of BP in the backscatteringgeometry is [89]:R˜(Ag) = a bc (1.4)R˜(B1g) = dd (1.5)R˜(B2g) = ee (1.6)R˜(B3g) = ff (1.7)Here a, b, c, d and f are Raman tensor elements [19][89], in which:a = |a|eiφa = ∂εxx∂qAg=∂ε ′xx∂qAg+∂ε ′′xx∂qAg(1.8)b = |b|eiφb = ∂εyy∂qAg=∂ε ′yy∂qAg+∂ε ′′yy∂qAg(1.9)c = |c|eiφc = ∂εzz∂qAg=∂ε ′zz∂qAg+∂ε ′′zz∂qAg(1.10)34f = | f |eiφ f = ∂εxz∂qB2g=∂ε ′xz∂qB2g+∂ε ′′xz∂qB2g(1.11)Here the ε ′ii and ε ′′ii (i=x, y, z) are the real and imaginary parts of the dielectricconstant along different crystalline orientations (armchair and zigzag directions),and qAg and qB2g are the normal coordinates of the Raman modes. The complexvalues of the Raman tensor elements are due to the light absorption of BP. Due tothe back-scattering configuration, the direction of incident laser is perpendicular tothe stacked layer plane of the sample, so that B1g and B3g modes are forbidden andAg and B2g modes can be observed during Raman measurements.After that, according to the interpretation of crystallography of BP shown inFig 1.21, the Raman tensors can be then expressed as:R˜(Ag) = asin2θ + ccos2θ 0 12(a− c)sin2θ0 b 012(a− c)sin2θ 0 acos2θ + csin2θ (1.12)R˜(B2g) = −esin2θ 0 ecos2θ0 b 0ecos2θ 0 esin2θ (1.13)In addition, ei and es in Eq. 3.1 can be written as:ei = es = (sinθ , 0, cosθ );As demonstrated in Fig 1.22, the θ here is defined as the angle between thezigzag direction and the polarization ei of the incident light (here the polarizationof the incident light is defined at 0 ◦ and 180 ◦). Moreover, in Table 3.1, the general-ized form of Raman intensities of Ag and B2g modes under “parallel” configurationare shown in Table 3.1. Here the term “Parallel” direction means: the scattered lightis parallel to the incident laser because of the polarization by the inserted analyzer[90]. When θ = 0◦ or 180◦, in other words, the incident light is parallel to zigzagdirection of BP, I‖Ag,0◦/180◦ = a2 and I‖B2g = 0; When θ = 90◦ or 270◦, namely the in-35Figure 1.22: Illustration of the “θ” angle, the “zigzag” direction of the BPsample and the reading angle “α” of the rotation stage. The red arrowsindicate the polarization of the incident laser.cident light is parallel to the armchair direction of BP, I‖Ag,90◦/270◦ = c2 and I‖B2g = 0.Therefore, Raman intensities of Ag and B2g modes under “parallel” configurationhave periodicities of 180◦ and 90◦, respectively. In addition, when incident laser isparallel to either zigzag or armchair direction, B2g mode is always forbidden.Therefore, with a specific excitation wavelength and for all thickness range, ifI‖Ag ,0◦/180◦I‖Ag ,90◦/270◦ 1 or 1, then ARPRS can provide unambiguous determination of theCO of BP.36Table 1.3: Raman intensity of BP Ag and B2g modes under “parallel” config-uration.Ag mode B2g mode(acos2θ + csin2θ )2 e2sin22θFigure 1.23: Illustration of Si (100) substrate.1.3.4 Crystallographic Orientation Of Si (100) Wafer And ItsOrientation Dependent Raman ResponseLike BP, the Raman intensity of Si (100) substrate also changes periodicallywith its CO and the laser polarization [91][92]. For instance, under parallel config-uration, when the incident polarization is parallel to Si {100} planes (see Fig 1.23),the Si Raman mode has a maximum / minimum Raman intensity [92].For example, Fig 1.24 illustrates the ARPRS results on a Si (100) wafer underthe “parallel” and “cross” configuration. Under both configurations, the Raman in-tensity of the Si (100) wafer shows a sinusoidal function of the rotation angle “α”(the reading from the rotation stage) with a periodicity of 90◦.37Figure 1.24: The ARPRS results of the polarized Raman measurements of aSi (100) wafer measured under under the “parallel” and “cross” con-figuration. Image courtesy of David Tuschel, Horiba Scientific, USA.1.3.5 Thickness Determination By The N-dependent RamanIntensity Ratio Of ISiI2DMIt is important to note that, Raman spectroscopy can also serve as a fast andnon-destructive method in identifying thickness ( or the layer number N ) of 2Dmaterial flakes [93]. SiO2 / Si (Si (100) wafer with an oxide layer) has been widelyused as the substrates of 2D materials. For instance, Li et al. [94] reported thedetermination of the layer number of graphene up to 100 layers according to theRaman peak intensity ratio of I(G)I(Si) ; similarly, the layer number of MoS2 can alsobe determined byI(E12g)I(Si) andI(A1g)I(Si) [95].38So far, for black phosphorus,IA1gISi[46] and ISiIA2g[12] as functions of layer num-bers N have been reported. However, because of the orientation-dependent Ramanresponse of BP, the CO of the BP sample needs to be known first for this methodto work. In this work, the investigations of the one-to-one relationship betweenthe thickness (the layer number N) and the Raman intensity ratios of ISiIA2gwere bepresented.1.3.6 Thermal Sublimation For Preparation Of Few-layer 2DMaterialsSo far, very few studies on the successful fabrication of few-layer 2D materi-als via thermal thinning were reported. In Fig 1.25 (A), the exfoliated 2 to 4 -layerMoS2 flakes were reported [22][96] to be reduced to single-layer via thermal thin-ning method at 330 ◦C; in Fig 1.25 (B), Huang et al. [23] reported that single andbilayer Sb2Te3 and Bi2Se3 flakes with lateral dimensions exceeding 10 µm couldbe prepared from thick flakes by performing thermal thinning at 490 ◦C and 510◦C respectively; in Fig 1.25 (C), the thermal thinning study of WSe2 from a recentwork [24] was shown, which demonstrated that the surface defects of a pristinesample could degenerate the quality of final products. Therefore, thermal thinninghas been studied as an effective method for reducing bulk 2D materials into a fewlayers. However, so far no studies have been reported using sublimation as a top-down thinning method to prepare few-layer black phosphorus.39Figure 1.25: (A): Sublimation thinning of a MoS2 flake: mechanically exfo-liated bilayer MoS2 flake was thinned to single layer after annealed at330 C for 15 h. (Figure is reprinted from [22] with permission); (B)sublimation thinning of Bi2Se3: Bi2Se3 plates prepared by depositionwith an original thickness of 3.5 µm was heated at 510 ◦ for 10 min-utes. (Figure is reprinted from [23] with permission); (C): sublimationthinning of WS2: mechanically exfoliated WS2 flakes listed in the up-per row were held at 600 ◦ for 45 minutes, and the Ar gas with a flowrate of 300 cc / min was purged in to protect the sample. At last, thefigures in the lower row demonstrated that those flakes were thinneddown to 8 to 45 nm from left to right. Figure is reprinted from [24]with permission.401.3.7 Previous Studies On Thermal Decomposition And SublimationOf Black PhosphorusAccording to our best knowledge, only 2 studies reported thermal decompo-sition and sublimation of black phosphorus. In 2015, by using in-situ transmissionelectron microscopy (TEM), Liu et al. [25] first observed sublimation of BP at 400◦C (See Fig 1.26 (A)) starting at flake edges and defects, and then forming eye-shaped cracks along <001> directions. Su et al. [97] also observed the decompo-sition and sublimation of BP at 350◦C. In 2016, another work [26] indicated thatthe sublimation of BP took place at 375◦C and the process involved detachmentsof pairs of P atoms (See Fig 1.26 (B)) by using low energy electron diffraction(LEED).Figure 1.26: (A): Observation of decomposition of BP at 400 ◦C by usingtransmission electron microscopy (TEM) (Figure is reprinted from[25] with permission); (B): the illustration of the decomposition mech-anism of BP: detachments with phosphorus dimers (Figure is reprintedfrom [26] with permission).411.4 Problem Definition And Thesis GoalFirst, so far none of the 6 available fabrication methods for few-layer BP thinfilm fabrication could satisfy all the 5 critical requirements: scalability, uniformity,crystallinity, thickness controllability and economic fabrication.It is a must to develop a new fabrication method which would meet thoseessential requirements. To date, thermal thinning has been extensively used forthinning down thick 2D material flake to a few layers, and so far no works havebeen reported on successful thinning of BP. Therefore, two major objectives havebeen proposed in this work on thermal thinning of BP:1. The thermal thinning method should be capable of reducing thick BP flakesto a few layers with good crystallinity, uniformity and scalability.2. The final thickness of BP products should be tunable, in other words, thethermal thinning process should be controllable.Second, up to now, ARPRS has not been confirmed as an effective methodin BP CO determination because of the complicated thickness and wavelength -dependent ( wavelength > 442 nm) ARPRS response. In the other side, limited datafrom previous report were also insufficient to confirm ARPRS with 442 nm exci-tation to be effective for CO determination. Therefore, the third major objectivehas been proposed in this work on BP CO determination by ARPRS with 442 nmexcitation:3. Investigations should be conducted on more samples to confirm there is nothickness-dependency of ARPRS results.1.5 Scope Of The ThesisIn this thesis, optical microscopy, atomic force microscopy (AFM) and Ra-man spectroscopy were used to study layer-by-layer thermal sublimation of black42phosphorus. AFM and Raman mapping measurements indicated the repeatable andsuccessful preparation of few-layer black phosphorus. Through the interrupted andcontinuous heating study of BP, the thinning rates of BP at 500 K and 550 K weremeasured.The effectiveness of ARPRS with 442 nm excitation in fast and non-destructivedetermination of BP CO was confirmed by investigating the ARPRS results of 27 BPsamples on 2 different substrates. After that, an in-situ Raman method in monitor-ing thickness was realized by using the thickness ( the layer number N ) -dependentRaman peak intensity ratio of I(Si) over I(A2g):ISiIA2g.In general, the scope of this thesis work can be demonstrated that:1. Thermal thinning method can used for controllably preparing large scale,uniform and crystalline few-layer BP.2. Raman spectroscopy could be used as a versatile tool to determine the crys-tallographic orientation, thickness, uniformity and crystallinity of BP.1.6 The Organization Of This ThesisAfter this chapter of introduction, in Chapter 2, the thermal sublimation of BPwas studied in this work. This is the first work to use the thermal sublimation thin-ning to controllably prepare large, crystalline few-layer BP. The experiments ofsublimation thinning of BP were presented: BP flakes were heated at high temper-atures and color changes of samples were recorded; AFM measured the changes ofthickness; spatial Raman mapping measurements were directed on samples beforeand after several thinning processes, and the results indicated the integrity, unifor-mity and crystallinity of BP sample remained. After that, by performing interruptedand continuous heating, the sublimation rates of BP were measured to be 0.2 nm/min at 500 K and 1.5 nm / min at 550 K from AFM results. At last, by using AFM,optical contrast and Raman mapping, the successful and repeatable preparation of43large and crystalline few-layer black phosphorus via the simple sublimation thin-ning method were demonstrated.In Chapter 3, the experiments of Raman study on BP samples with thick-nesses ranging from 10 to 200 nm on SiO2 / Si and polyimide substrates werepresented: the orientation-dependent Raman responses of 11 BP flakes on Si sub-strates were studied by angle-resolved polarized Raman spectroscopy (ARPRS).The results revealed that, with the excitation wavelength of 442 nm and under par-allel configuration, the maximum and minimum Raman peak intensities of A1g andA2g modes are orthogonal; in addition, when the maximum Raman intensity of A2gmode appears, it will be always 10 - 100 times greater than its minimum Ramanintensity at ± 90◦ orthogonal to its current position. Such unambiguous Ramanresponses didn’t appear in previous ARPRS studies using other excitation wave-lengths. Therefore, ARPRS with the excitation wavelength of 442 nm and underparallel configuration were confirmed to unambiguously determine the CO of BP.After that, it was demonstrated that, with a BP flake rotated to generate maximumA1g Raman intensity, a one-to-one relationship between the Raman intensity ratioof ISiIA2gand BP thickness could be established. What’s more, the ISiIA2gincreases withdecreasing thickness monotonically. Therefore, first developed in this work, thisall-Raman method provided the CO identification as well as the capability of in-situdetermining the thickness of BP.In summary, Chapter 4 summarizes the results of this work and suggests thatthe sublimation thinning method is a promising method of preparing few-layer BPand the Raman spectroscopy is a powerful tool in determining the CO, thicknessand crystallinity of BP prepared by thermal thinning method. Potential directionsfor future work are also proposed.44Chapter 2Thermal Sublimation of BlackPhosphorus“Everything should be made as simple as possible, but not simpler.”— Albert Einstein2.1 General Procedure Of The Experimental Work OfThis ChapterIn this section, uniform layer-by-layer sublimation of the black phosphorus at≤ 600 K ( ∼ 330 ◦C ) were reported.Raman imaging confirmed the uniformity and crystallinity of the BP samplesafter processing. The sublimation rates of BP were demonstrated to be 0.18 nm /min at 500 K ( ∼ 230 ◦C ) and 1.5 nm / min at 550 K ( ∼ 280 ◦C ). Moreover,the successful and repeatable preparation of large and few-layer crystalline BP in-dicated the thermal sublimation could be used as a simple method for controllablypreparing crystalline few-layer BP.452.1.1 Thermal SublimationThe BP flakes were mechanically exfoliated from bulk BP crystals (99.998%, Smart Elements, Vienna, Austria) in a glove box using Nitto SPV 224 R tapeand transferred to a Si (100) wafer (4 inch, 0.56 mm thick) with a 300-nm-thickthermaly-grown SiO2 layer) or a polyimide substrate. Ultrasonication in acetoneand rinsing process in deionized water were preformed to pre-clean this wafer-/polyimide. The wafer was then cut into small pieces (∼ 4 mm×4 mm) by a glasscutter. After that, a small piece of wafer was loaded into a Linkam TS1200 heatingstage.The heating stage was then mounted on the motorized XY stage of a HoribaScientific LabRAM HR800 con-focal Raman microscope. After closing the lid,the valves on both sides of the heating stage (see Fig 2.1) were opened to purgethe high purity nitrogen gas (Praxair, industrial grade, purity: 99.995%) flowingcontinuously through the heating stage. This step was in order to exhaust the airin the heating stage and build up an inert environment for following heating pro-cess. After purging N2 gas for 30 minutes, a programed isothermal joule heatingprocess was performed on the wafer placed on a sapphire plate in the chamber ofthe heating stage with nitrogen gas flowing through the chamber. The ramping-upand cooling-down rates were both programed at 100 K / min. The continuous N2flow would be kept until the sample was taken out for AFM measurements.2.1.2 Raman MeasurementIn section 2.4.1 to 2.4.3, all Raman measurements were performed with aHoriba LabRam HR800 Raman system in a backscattering configuration with 441.6nm excitation. The Raman system has two modes: video and Raman. the videomode was used to record the optical images. The background light source for thevideo mode was provided by an integrated EUROMEXT M IlUMINATOR EK-5light, and the degree of light for all optical observations in this thesis was fixed at“5”. The Raman spectra were collected through an Olympus 50X (NA = 0.55) ob-jective lens and recorded with the grating of 2400 lines mm−1 which has a spectralresolution of 0.27 cm−1. When performing Raman measurement at 500 K / 550K, the laser power was reduced to be less than 0.2 mW/µm2 by adding a filter in46Figure 2.1: Illustration of the sublimation process of BP.order to minimize the laser heating effect. The acquisition time and accumulationtimes were optimized to be 6 sec and 2 times, respectively, in order to get a signal-to-noise ratio of 100 approximately as well as to avoid long time laser damage toBP samples.Furthermore, in this section, all two-dimensional Raman mapping imageswere acquired using the following configurations: the increments of X and Y di-rection were 1.6 and 1.5 µm /step, respectively. It is worth mentioning that, thelaser spot size was ∼ 2 µm2. In addition, during Raman mapping measurements,all BP flakes were still kept in the heating stage protected in the inert environmentof high purity nitrogen gas.In section 2.4.3, all Raman measurements were performed with a HoribaLabRam HR800 Raman system in a backscattering configuration with 632.8 nmexcitation. The Raman spectra were collected through an Olympus 50X (NA =0.55) objective lens and recorded with the grating of 600 lines mm−1 which has47a spectral resolution of 0.63 cm−1. The acquisition time and accumulation timeswere optimized with 3 sec and 2 times, respectively. Moreover, in this section,all two-dimensional Raman mapping images were acquired using the followingconfigurations: the increments of X and Y direction were 1.2 and 1.1 µm /step,respectively. During the Raman mapping measurements, all BP flakes were placedin the heating stage with high purity nitrogen gas purging in to avoid oxidation.2.1.3 Determination Of The Crystallographic Orientations (CO) OfBP And underlying Si (100) SubstrateThe 441.6 nm laser beam was polarized, and an analyzer was placed in theparallel configuration before the entrance of spectrometer in the same manner asprevious studies [18][19][98]. BP samples were placed on a rotation stage and ro-tated 180◦ about the microscope optical axis in 12 steps ( 15 ◦/ step ). Consideringthe periodicity of A1g and A2g modes were 180◦, and the maximum Raman inten-sity of A1g and A2g modes were orthogonal, the maximum Raman intensity of A1gor A2g mode could be easily measured with rotating only 90◦ ( 6 steps ). In thiswork, the sample was rotated at an angle with the maximum Raman intensity ofA1g mode. What’s more, since the rotation angle acquired from the previous 6-stepoperation might still differ from the real angle with the maximum Raman intensityof A1g mode, half-interval search method was used for further refinement of rota-tion angle. Similar operations were directed for underlying Si (100) wafer to findthe maximum and minimum Si Raman intensity. In addition, during every step, thelaser spot was focused at the same point on the sample to ensure the consistency ofresults.2.1.4 Atomic Force Microscopy (AFM)AFM measurements were performed in contact mode by an Asylum ResearchMolecular Force Probe 3D atomic force microscope and a Bruker Atomic ForceMicroscopy System. Samples were stored in an enclosed container with desiccantsbefore transferring to AFM measurements in order to minimize the oxidation.482.2 Results And Discussions2.2.1 Investigations Of Thermal Sublimation And Its Mechanism2.2.1.1 Sublimation of BP At 600 K: Uniform Color Changes With NoObservable CracksFigure 2.2: (A) - (D): Observation of the layer-layer-layer sublimation pro-cess of BP at 600 K. The different colors in the area 1 and area 2 indi-cated different thicknesses.First, uniform color changes with no area shrinkages observed during theisothermal joule heating under the protection of high purity nitrogen gas was re-49ported. Fig 2.2 (A)−(D) show the optical images of a BP sample on SiO2 / Sisubstrate at four different times during 600 K heating from 0 to 51 minutes.Obvious changes of color, which are directly related to the thickness, wereobserved during the isothermal heating. Fig 2.2 (A) - (D) showed optical imagesof BP Sample A at four different times during 600 K heating from 0 to 51 mins.0 min indicated the time point where the temperature just reached the isothermalholding temperature. In Fig 2.2 (A), the pink zone (area 2) was thinner than thegreen zone (area 1). Fig 2.2 (B) showed area 1 turned to red and area 2 to peachafter heating for 21 minutes. In Fig 2.2 (C), area 2 turned to gray at 41 minutes,which was the characteristic color of few-layer BP as reported previously [46][12].In Fig 2.2 (D), at 51 minutes, area 2 was fully disappeared. In addition, no micron-scale area shrinkages were observed during the whole heating process. The colorsof each region were uniformly changed, which suggested that the sublimation pro-cess was uniform and happened layer by layer.2.2.1.2 Sublimation Of BP At 500 K: AFM StudyTo capture the thickness changes, a large trapezoid thick BP flake, with an areaover 1,200 µm2, was heated at 500 K for 195 minutes. AFM measurements wereperformed immediately before and after the heating to acquire the thickness andmorphology and to avoid sample degradation in the air. In Fig 2.3 (A) and (B), themorphology and thickness results of the original sample before heating, showedtwo regions with different colors (light green for Zone 1 and peach for Zone 2)and thickness (35 and 45 nm for Zone 1 and 2 respectively). Fig 2.3 (C) and (D)showed the thickness of the sample after heating was reduced to 15 and 25 nm forZone 1 and 2 respectively. Therefore, both Zone 1 and Zone 2 were thermally re-duced by the same thickness of 20 nm, and the step height between them remainedunchanged after the heating process. This can be explained by that the sublimationof Zone 1 took place along the edges which is the step. Therefore, the thicknessmeasurements by AFM confirmed that the BP flake underwent an layer-by-layeruniform sublimation, and the sublimation rate didn’t depend on the original thick-50Figure 2.3: AFM and optical images of a trapezoid BP sample. (A) and (B):before heating; (C) and (D): after heating in nitrogen at 550 K for 195mins. The dispersed white pits on the sample surface were attributed tothe fast surface reactions with oxygen and moisture in the air [21] duringsample transportation to the AFM and during the AFM measurementsin the air. The red dashed lines indicated the edges of the samples andthe numbers next to them are edge dimensions.ness. No obvious cracks were found on the sample in all the above experiments.2.2.1.3 Sublimation Of BP At 500 K And 550 K: AFM And 2D RamanMappingAdditionally, in Fig 2.4 and 2.5, high temperature heating on 2 BP flakes wereperformed at 500 and 550 K, respectively. Spatial Raman maps and thicknessmeasurements by atomic force microscopy (AFM) of both BP flakes were taken51Figure 2.4: (A):the optical image of this BP flake before 500 K heating pro-cess, and the inset AFM line-scan profile indicated its thickness to be 36nm; (B) - (C): Spatial Raman maps of A1g and A2g peak intensity acquiredon a 36-nm-thick BP flake; (D): the optical image of this BP flake after500 K heating process for 120 minutes, , and the inset AFM line-scanprofile indicated its thickness to be 13 nm; (E) - (F): Raman images (A1gand A2g modes) of this BP flake after 500 K heating process;before and after high temperature heating. Fig 2.4 (A) and (D) showed a 36-nm-thick BP flake was reduced to 13-nm-thick after heating at 500 K for 120 minutes.In comparison with Raman maps of A1g and A2g modes of the pristine BP flakeshown in the Fig 2.4 (B) and (C), those shown in Fig 2.4 (E) and (F) indicated that,these Raman peak intensities were uniform across the sample, showing that the BPsample kept its integrity and uniformity after the heating process, which confirmedthe layer-by-layer sublimation of BP.Similarly, the comparison between Fig 2.5 (A)-(C) and Fig 2.5 (D)-(F) of an-other flake also indicated that the layer-by-layer sublimation of heating process op-erated at 550 K for 25 minutes. Moreover, since the conditions of the backgroundlight for all images taken in this work were consistant (see methods in Section 2.1.1in P 46), the color changes with different thickness can be attributed to the lightwave interference effect between the incident, transmitted and reflected light.52Figure 2.5: (A): the optical image of this BP flake before 550 K heating pro-cess, and the inset AFM line-scan profile indicated its thickness to be 85nm; (B) - (C): Spatial Raman maps of A1g and A2g peak intensity acquiredon an 85-nm-thick BP flake; (D): the optical image of this BP flake af-ter 550 K heating process for 25 minutes, and the inset AFM line-scanprofile indicated its thickness to be 49 nm; (E) - (F): Raman images (A1gand A2g modes) of this BP flake after 550 K heating process;2.2.1.4 Thermal Thinning Of BP At More TemperaturesIn contrary, previous studies [25][97][26] all reported that the sublimationand decomposition of BP were above 350 ◦C (623 K), which was more than 50◦C higher than the temperatures studied here. In Liu. et al’s [25] work, they per-formed transmission electron microscopy (TEM) observation after heating for lessthan 20 minutes and in another work [26], real-time observations at different hightemperatures by low energy electron microscopy (LEEM) on a heating ramp withless than 20 seconds were performed. Both of the two previous works didn’t inves-tigate the relative long term thermal stability of BP at the temperatures they studied.In addition, some other experimental details of those two reports were dif-ferent from this study as well: 1. in Liu. et al’s [25] report, they prepared theirsamples by liquid exfoliation method and the substrate was copper TEM grids, andin Fortin. et al’s work [26], they prepared samples also by mechanical exfoliationmethod and used Si (100) substrates without oxide layer; 2. the energies of TEM53and LEEM electron beams for imaging used in those 2 works were 200 KeV and30-35 eV, respectively. According to Vierimaa. et al’s report [99], if the energy ofelectron beam is larger than 80 KeV, it would immediately rapidly lead to a consid-erable amount of damage to BP sample. Therefore, our work cannot be compareddirectly with those 2 studies directly without considering these experimental de-tails.So far, no comprehensive investigations have been conducted on the thermalstability of BP below 375 ◦C (650 K) under stable heating and without the introduc-tion of electron beams. To address, 6 independent investigations were conductedto observe the sublimation process of BP heated for 1 hour at 450 K (∼ 180 ◦C),500 K (∼ 230 ◦C), 550 K (∼ 270 ◦C), 600 K (∼ 330 ◦C), 620 K (∼ 350 ◦C), and670 K (∼ 400 ◦C), respectively.First, Fig 2.6 (A) - (D) illustrated the comparison of BP flakes before and af-ter heating for 1 hour at 450 K. There was a large BP flake with an area over than200 µm2 in the center of (A), with many small BP flakes dispersed around it. Sig-nificantly, in (B) and (D), which showed that after heating for 1 h, no observablechanges took place. Therefore, it could be concluded that the exfoliated BP flakesremained stable during one-hour heating under 450 K (∼ 180 ◦C).Secondly, Fig 2.7 (A) - (D) illustrate the comparison of BP flakes before andafter heating for 1 hour at 500 K. Obviously, after heating for 1 hour, the compar-ison between (A) and (B) shows that many small flakes disappeared, which wasdifferent from Fig 2.6. In addition, for the large BP flake in the center of (A), thecolor changes in (D) indicated the attenuation of thickness also took place. More-over, no observable changes of area of the flake were observed. Then it could beconcluded that, the sublimation of the exfoliated BP flakes took places at 500 K (∼230◦C).54Figure 2.6: Sublimation study of BP at 450 K (∼ 180 ◦C). (A): the BP sam-ples heated at 450 K for 1 minute; (B): the BP samples heated at 450K for 61 minutes; (C): zoom in of the sample in the center of (A); (D):zoom in of the sample in the center of (B).55Figure 2.7: Sublimation study of BP at 500 K (∼ 230 ◦C). (A): the BP sam-ples heated at 500 K for 1 minute; (B): the BP samples heated at 500K for 61 minutes; (C): zoom in of the sample in the center of (A); (D):zoom in of the sample in the center of (B).56In our study, besides uniform thinning without the generation of observabledefects, the propagation of micron-scale cracks and non-uniform thinning was alsoobserved, which showed a strong dependence on the initial sample quality. A fewexamples are shown below.Thirdly, the thermal stability of the exfoliated BP flakes at 550 K (∼ 280 ◦C)was studied: in Fig 2.8, in comparison with the original sample shown in (A), thesample in (B) after heating for 1 hour showed non-uniform colors, indicating differ-ent thickness, and that was in agreement with (A). In addition, in (B), an eye-shapecrack generated in the circled area. The original sub-micro scale defects, surfacecontaminants and cracks might have existed in the original sample, and those re-sulted in the eye-shape crack under high temperature.Figure 2.8: Sublimation study of BP at 550 K (∼ 280 ◦C). (A): the BP sam-ples heated at 550 K for 1 minute; (B): the BP samples heated at 550 Kfor 61 minutes.57Fourthly, Fig 2.9 (A) - (B) show the comparison of BP flakes before and afterheating for 46 minutes at 600 K (∼ 330 ◦C). Obviously, the comparison between(A) and (B) demonstrated that after heating for 46 minutes, most of the small flakesdisappeared, which was similar to Fig 2.7. Moreover, for the large BP flake in thecenter of (A), the optical images taken at different time intervals revealed the dy-namic sublimation process of this BP flake: apparently, this flake started with dif-ferent original thicknesses (see the non-uniform colors) and many visible defectsintroduced during mechanical exfoliation (see the rough surface and cracks alongthe edges). With increasing time, in the circled area, more severe sublimation tookplace than the other area, and eventually in (F), the eye-shape crack propagated.The arrows in (C) - (F) indicated the formation of cracks. Therefore, surface con-taminants and non-uniform starting thickness resulted in the generation of cracksand fractures of BP flakes during sublimation process.Figure 2.9: Sublimation study of BP at 600 K (∼ 330 ◦C). (A): the BP sam-ples heated at 600 K for 1 minute; (B): the BP samples heated at 600 Kfor 41 minutes; (C)-(K): the zoom-in optical images of the sample in thecenter at (A) heated at 600 K for 1, 21, 31 and 41 minutes, respectively.58After that, via heating BP samples on another piece of Si wafer at 600 K for 46minutes, another observation was performed. The comparison between (A) and (B)indicated an eye-shaped hole appeared in the center of the large BP flake; and theoptical images of (C) - (K) taken at different time intervals revealed the dynamicformation and expansion of the hole which initiated from relative thinner centerregion (see (G)), then grew concentrically and eventually generated the large hole(see (K)).Figure 2.10: Sublimation study of BP samples on another piece of Si wafer at600 K (∼ 330 ◦C). (A): the BP samples heated at 600 K for 1 minute;(B): the BP samples heated at 600 K for 46 minutes; (C)-(K): thezoom-in optical images of the sample in the center at (A) heated at600 K for 1, 11, 20, 30, 40, 42, 43, 44 and 46 minutes, respectively.59Fifthly, Fig 2.11 shows the comparison of BP flakes before and after heatingfor 9 minutes at 620 K (∼ 350 ◦C). Notably, the comparison between (A) and (B)indicated that, both the left parallelogram-shape BP flake and right trapezoid-shapeone retained their shapes except for shrinkage of area. Moreover, the optical im-ages recorded at different time segments from (C) - (H) revealed the dynamic areashrinkage process. Based on the observation, it could be suggested that the thick-ness at edges were thinner than that of the center region, and due to the layer-by-layer sublimation, the edge regions were thinned down to 0 layer, in other words,fully sublimated first, which appeared as an area shrinkage.Figure 2.11: Sublimation study of BP at 620 K (∼ 350 ◦C). (A): the BP sam-ples heated at 620 K for 1 minute; (B): the BP samples heated at 620K for 9 minutes; (C)-(H): the zoom-in optical images of the markedsample in (A) heated at 620 K for 1, 2, 3, 4, 5, 6, 7, 8 and 9 minutes,respectively.60Sixthly, in Fig 2.12, it showed the comparison of BP flakes before and afterheating for 22 minutes at 670 K (∼ 400 ◦C). Similar to Fig 2.7, Fig 2.9 - Fig 2.11,in (B), most of the small flakes disappeared after heating for 22 minutes. Then,from (C) to (G), the clear, uniform and regular color changes in separated parts ofthis large BP flake indicated the dynamic sublimation process to be layer-by-layer.In addition, the well-preserved integrity of this flake during the progress of heatingindicates that the surface of the original sample was relatively free of contaminantsand defects.Figure 2.12: Sublimation study of BP at 670 K (∼ 400 ◦C). (A): the BP sam-ples heated at 670 K for 1 minute; (B): the BP samples heated at 670K for 22 minutes; (C)-(G): the zoom-in optical images of the markedsample in (A) heated at 670 K for 1, 5, 10, 15 and 22 minutes, respec-tively.612.2.1.5 Requirements For Sample SelectionTo summarize, during the thermal thinning, both uniform and non-uniformthinning were observed. Some BP samples had crack generated, which in mostcases can be related to the quality of the pristine BP flakes. Through optical mi-croscopy, it was observed that large area samples free of micron-scale defects onthe surface had more chances to be thinned down uniformly and without visiblemicron-scale defect generation.As optical images and 2D Raman mapping are limited to micron-scale, fora BP flake which satisfies this method for sample selection, there are still pos-sibilities that this flake has sub-micron-scale and nanoscale defects not observedby optical microscopy or 2D Raman mapping. Ideally, AFM/TEM should be usedfor that purpose. As this work was the first study on thermal thinning of BP, andalso long-time air-exposure in AFM generate surface defects, thermal thinning wasperformed immediately after mechanical exfoliation and optical inspection. Thisempirical method for sample selection worked well. It was estimated that about70 % of BP flakes selected by optical inspection could be thinned down withoutobservable defect generation.Let us look at some 3D AFM images of the samples which were thinned downsuccessfully with their integrities kept, which can give more insights to the sampleselection. In Fig 2.13 (A), the BP flake was the sample used for the successfullayer-by-layer sublimation study for 500 K heating in section 2.2.1.3 (P 52), whichhad been heated for 2 cycles and showed a thickness of 25 nm, and clearly, thissample had no visible micron-scale defects. In (B), obviously, the 3D image in-dicated the steep sidewall, and the surface of the sample was flat which indicatethe thickness to be uniform. In addition, significantly, despite the lower-left partof this flake, both the optical and the 3D AFM image confirmed the uniform edgesof this sample. Moreover, after the third cycle of heating process, in (C), the op-tical image reflected the uniform color changes of this sample, and essentially thearea of the sample remained same as that shown in (A). In (D), the 3D AFM imagedemonstrated not only the vertical side wall of the sample kept but also the flatness62Figure 2.13: (A): Optical image of the 36-nm-thick sample used in Section2.2.1.3 (P 52); (B): 3D AFM images of the 36-nm-thick sample shownin (A); (C): Optical image of this sample after thinning down to 18-nm-thick; (D): 3D AFM images of the 18-nm-thick shown in (C).of the sample remained.After that, in Fig 2.14, in (A) and (C), two BP flakes were successfully thinneddown to 10-nm-thick and 20-nm-thick with well-kept integrity on polyimide sub-strate. Their 3D AFM images shown in the (B) and (D) also revealed the samplehad uniform vertical side wall and flat surface (despite the minor oxidation) innanoscale. Therefore, for successful thinning of a thick BP to few-layer, the re-quirements of selection of pristine sample could be summarized as:1. The surface of sample should be clean, with no visible cracks and contam-63Figure 2.14: (A): Optical image of a prepared 10-nm-thick sample on poly-imide substrate; (B): 3D AFM images of the 10-nm-thick sample shownin (A); (C): Optical image of a prepared 20-nm-thick sample on poly-imide substrate: (D): 3D AFM images of the 20-nm-thick sampleshown in (C).inants dispersed on the surface of sample.2. The sample should have uniform edges and vertical side walls. Opticalimaging, although not sufficient in resolution, can be used to do a coarse sampleselection fast and non-destructively. Future work needs to be conducted to find abetter solution to check nm scale defects before thinning.642.2.1.6 Thermal Stability Of BP & Sublimation MechanismAfter the 6 independent studies of BP sublimation at different temperatures, itcould be summarized that:1. Thermal Stability of BPThe exfoliated BP flakes have no observable changes after one hour heatingwith a temperature less than 450 K (∼ 180◦C), which is typically lower than thereported temperature of 350 ◦C [25]; when the temperature is above 500 K (∼230◦C), the sublimation of BP takes place if given enough time. This was still inagreement with the two previous reports on thermal stability of BP: the anneal-ing time in Liu. et al.’s [25] work was less than 20 minutes and they indicatedthe decomposition temperature to be 400 ◦C; while another work [26] reported thedecomposition temperature to be 375 ◦C ± 25 ◦C from an real-time measurementwith increasing temperature. Both of the two previous works used far less heatingtime during the high temperature study of BP than this work. Therefore, not onlythe heating temperature but also the heating time will affect the thermal stability ofBP. On the other side, the sublimation rates increase with increasing temperature.In other words, the sublimation rate is tunable by controlling the temperature.2. Sublimation mechanism of BPFrom above micron-scale observations by optical microscopy, it can be seenthat, the samples (samples for 620 K and 670 K heating) which had relative cleanand smooth surfaces as well as uniform edges underwent layer-by-layer attenuationof thickness with their uniformity well maintained, but those samples which hadcontaminated and rough surfaces as well as nonuniform edges (samples for 550 Kand 600 K heating) were destructed by the formation of cracks.The reported sublimation mechanism [26] can be used to explain the results:since the sublimation is initiated by the detachments of phosphorus dimers at theedges and vacancies , and the detachment rate along [100] direction (armchair) is65faster than that of [001] direction (zigzag), the eye-shaped cracks propagate alongthe [100] direction, enlarge and cause the full layer to be sublimated.Therefore, the surface impurities and non-uniform edges are the main reasonsthat cause the decomposition or unsuccessful thinning of BP. The selection of sam-ples then becomes critical for successful thinning.2.2.2 Investigation of The Thinning RatesAfter the discussion on the observation and mechanism of sublimation thin-ning, the thinning rates at different temperatures were investigated by two experi-ments termed “interrupted heating” and “continuous heating”, respectively.Due to the anisotropic BP and Si (100) in-plane structures, the Raman peakintensities depend on the crystallographic orientation of the BP samples in respectto the polarization direction of the incident laser. The effectiveness of ARPRS mea-surements in identification of CO of BP was also a very important part of this workand the details would be discussed in section 3.3.2 (P 118) of Chapter 3. In short,from the ARPRS measurements, the axes of BP could be determined by the orthog-onal directions of its maximum A1g (zigzag) and A2g (armchair) Raman intensitiesappeared [18][98]. Thus, in this work, the CO of samples was first determined byARPRS and then transferred to the heating stage carefully with their positions un-changed.In Fig 2.15 (A), a single flake (named “Sample A”) has been rotated to havemaximum A1g Raman intensity and the minimum A2g intensity (see the Raman spec-trum shown in Fig 2.15 (B)). There were 3 areas in the “Sample A” showing var-ious colors indicating different thicknesses of those 3 areas. A Raman spectrumwas taken from area 1. Similarly, in Fig 2.16 (A), another flake (named “sampleB”, with two areas in different colors) was prepared by exfoliation. The Ramanspectrum taken at area 2 also showed it has maximum A1g Raman intensity and theminimum A2g intensity.66Figure 2.15: (A): Optical image of sample A; (B): the Raman spectrum withthe maximum intensity of A1g mode and the minimum intensity of A2gmode;67“Interrupted Heating” At 500 K And 550 KThe “interrupted heating” was operated as follows: 5 cycles of the heatingprocess were directed on single BP flake at 500 K (or 550 K). The time of eachcycle was illustrated in Fig 2.17 (B). AFM measured the thickness before and af-ter each heating process, and the results were demonstrated in Fig 2.18 (A). The“interrupted heating” can be described as as follows: one cycle of heating wasperformed on single BP flake at 500 K (or 550 K), thicknesses before and afterheating were acquired by AFM measurements. For each cycle, 2 Raman spectrawere taken before and after heating at room temperature respectively; similarly, 2Raman spectra were collected at the beginning and ending of the heating processat high temperature, respectively. The Raman results would then be discussed insection 3.3.2 (P 118) of Chapter 3.In Fig 2.18 (A), it illustrates that, for each heating cycle, each area of sam-ple A was reduced by the same thickness, revealing that different areas underwentidentical sublimation rates. Similarly, in Fig 2.17 (C), sample B with 2 differentareas was heated at 550 K for 6 cycles and the heating time for each cycle waslisted in Fig 2.17 (D), the recorded thickness changes were plotted in Fig 2.18(B), revealing that, at the heating temperature of 550 K, the sublimation took placeisotropically in different areas.Moreover, in Fig 2.19 (A), after heating for 3 cycles, the thickness of the area3 of the “sample A” was reduced to ∼ 5 nm (see inset AFM profile in (A)). Sincethe critical thickness of BP from bulk to 2-dimensional was predicted as ∼ 10 nm[49] (or 20 layers), area 3 has a thickness of 10 layers. In Fig 2.19, the Ramanspectrum indicated the uniformity and crystallinity of area 3. Significantly, this 5-nm-thick flake has a large area of around 180 µm2. In addition, Raman intensitiesof A1g and A2g remained strong while the Si Raman peak intensity was enhanced incomparison with the Raman spectrum of the original sample shown in Fig 2.15 (B).68Figure 2.16: (A): Optical image of sample B; (B): the Raman spectrum withthe maximum intensity of A1g mode and the minimum intensity of A2gmode;69Figure 2.17: (A): Optical image of sample A to be heated at 500 K; (B) the il-lustration of the interrrupted heating process (5 cycles) to be performedat sample A; (C): Optical image of sample B to be heated at 500 K. (D):the illustration of periodic heating process (6 cycles) to be performedon sample B.As mentioned in the observations of BP sublimation, the colors of BP changedwith decreasing thickness. Therefore, if knowing the characteristic color of BPat each thickness, a thickness-color can be established which would be useful forthickness determination. This was realized after the “interrupted heating experi-ments. The characteristic colors extracted from those recorded optical images andtheir related thickness from AFM measurements were plotted as the thickness-colormap shown in Fig 2.20.70Figure 2.18: (A): Thickness in different areas of sample A as a function oftime, and (B): the thickness in different areas of sample B as a functionof time. At each red dot, a Raman spectrum was collected (analysis ofthose collected Raman spectra would be discussed in Section 3.3.2 (P118) of Chapter 3).71Figure 2.19: (A): Optical image of area 3 in Sample A, and the inset AFMprofile indicated it was thinned down to 5-nm-thick; (B): Raman spec-trum of the 5-nm-thick BP sample showing the 3 characteristic peaksof BP.72Figure 2.20: Thickness-color map: the colors were extracted from those sam-ple of which the thicknesses had been measured by AFM;73“Continuous Heating” At 500 K And 550 KIn order to get accurate thinning rates, the “Continuous Heating” was per-formed on two thick BP flakes labeled as sample C and D at 500 K and 550 K,respectively. In Fig 2.21 (A) and (B), the thickness of sample C was reduced from85 to 45 nm after heating process at 500 K for 221 minutes; in Fig 2.21 (C) and (D),the thickness of sample D was reduced from 105 to 45 nm after heating process at550 K for 51 minutes.At different time intervals, the optical images were taken at the sample in or-der to record the gradual color changes (see in Fig 2.22 and 2.24). In addition,in-situ Raman spectra at each time interval were collected before taking images,and the frequent laser exposure caused the black spots on sample C and D shownin Fig 2.21 (B) and (D). The in-situ Raman spectra at 500 K and 550 K were shownin Fig 2.23 and Fig 2.25. Obviously, with increasing time during thinning, the Sipeak intensity was enhanced while all 3 BP peaks remained unchanged.74Figure 2.21: (A) & (B): Optical images of Sample C before and after continu-ous heating process at 500 K, the heating time was 221 minutes; (C) &(D): Optical images of Sample D before and after continuous heatingprocess at 550 K, the heating time was 51 minutes. The inset profilein each picture is the line-scan AFM profile indicating the thickness ofthis sample.75Figure 2.22: Sample C was annealed at 500 K for 221 minutes.Figure 2.23: In-situ Raman spectra of the sample C at 500 K.76Figure 2.24: Sample D was annealed at 550 K for 51 minutes.Figure 2.25: In-situ Raman spectra of the sample D at 550 K.77Figure 2.26: Comparison of thinning rates of sample A and C.Thinning ratesThe thicknesses of sample C and D shown in those pictures of Fig 2.22 and2.23 were read from the color-thickness map in Fig 2.20 and then interpolated intoFig 2.26 and Fig 2.27 ( the red line ). In addition, as shown in Fig 2.26 and Fig2.27, the thickness-heating time profiles of sample A and C ( the blue lines ) forinterrupted heating were also integrated. The thickness-heating time profiles of in-terrupted and continuous heating at 500 and 550 K matched. The thinning rateswere ∼ 0.18 nm/min at 500 K and ∼ 1.5 nm/min at 550 K.78Figure 2.27: Comparison of thinning rates of sample B and D.79Figure 2.28: (A): Optical image the BP sample thinned down to 2-nm-thick;(B): Raman spectrum collected at the circled area in (A); (C) & (D):Spatial Raman maps of sample shown in (A).80Figure 2.29: (A): Optical image the BP sample thinned down to 2-nm-thick,and the inset figures represent the spatial AFM image and thicknessprofile of circled area. (B) - (D): Optical contrast analysis of Fig 2.28(A), which was splitted into R, G and B channels, and the G channelfigure in (C) was selected for analysis.2.2.3 Preparation Of The few-layer BPAn additional cycle of heating was then performed with sample A. In Fig 2.27(A), area 3 was entirely sublimated. In Fig 2.27 (B), the Raman spectrum at the cir-cled position of Fig 2.27 (A) showed 3 typical BP peaks of A1g, B2g and A2g modesat 361.95, 439.57 and 466.81 cm−1, respectively. What’s more, in Fig 2.27 (C) &(D), the spatial Raman maps of A1g and A2g modes showed strong contrast betweenthe sample and substrate, which indicated the crystallinity of whole BP sample was81still retained. In Fig 2.28 (A), the thickness of the circled area was measured: theinset AFM image and profile were shown in Fig 2.28 (A), which indicated that thethickness of area 1 was reduced to ∼2-nm-thick. According to previous report, asingle layer BP sample has a theoretical thickness of 0.53 nm [49][63], so that thatprepared BP sample has 4 layers.For thickness measurement of 2D material less than 15 layers, AFM measure-ments can bear significant errors due to the surface absorption layers (absorptionof water and oxygen in air) [100]. For example, Liu. et al. [9] reported that thick-ness of a single layer was measured to be 0.85 nm by AFM, which was far greaterthan 0.53 nm. That should be attributed to the fast degradation of BP during mea-surements. Therefore, it is necessary to find an effective non-contact method toquantify the layer number of BP when it is less than 15 layers.As the average optical contrast has been demonstrated to be a useful tool inindicating 2D materials with a thickness below a few nanometers [101][100], Fig2.29 (A) was splitted into monochromic red (R), green (G), and blue (B) channelimages via Image J [102] shown in Fig 2.29 (C) - (D). Here the highest-contrastgray-scale image in the G channel was selected and then shown in Fig 2.29 (C).It is important to note that, the average optical contrast between the sample andthe substrate in the G channel image increases by the layer number. Accordingto Gomez et al. [46]’s report, monolayer BP showed an optical contrast of 2.8 %,while the bilayer BP flake showed an optical contrast of 5.5 ± 0.2 %. The contrastof the sample in (A) had a mean value of 10.89%. As mentioned above, this sam-ple was measured to be 2-nm-thick, corresponding to ∼ 4 layers. Therefore, theoptical contrast can be regarded as a linear function of layer number when it is lessthan 15 layers and it could be used for determination of layer number.After that, repeatable preparation of large few-layer BP was performed. In Fig2.30 (A), successful preparation of a thin BP flake with an area over ∼ 1,000 µm2by continuously heating a thick bulk BP flake (see the inset image of Fig 2.30 (A))at 300 ◦ C under N2 protection was demonstrated. Despite the cracks generatedin the circled area, the entire sample retained its integrity after the process. In Fig82Figure 2.30: (A): Optical image a BP sample thinned down to 3 and 4 -layers,and the inset figure represents the original BP sample. (B); Ramanspectra collected in circled area in (A); (C) - (E): Spatial Raman mapsof the BP sample shown in (A); (F) - (H): Optical contrast analysis of(A), which was splitted into R, G and B channels, and the G channelfigure in (C) was selected for analysis.2.30 (B), the point Raman spectrum collected circled area in Fig 2.30 (A) showed3 typical BP Raman peaks at 355.414, 431.21 and 459.24 cm−1, respectively. Inaddition, Raman mapping at this sample was performed. In Fig 2.30 (C) - (E), the3 BP vibrational modes A1g, B2g and A2g showed a strong intensity contrast in com-parison with the substrate, which indicated the crystallinity of the prepared sample.Moreover, in Fig 2.30 (F) - (H), the average optical contrast results in Fig 2.30 (A)have 2 mean values of 7.21% and 10.96%, corresponding to 3 and 4 layers.83Then, repeatedly, in Fig 2.31 (A) and Fig 2.32 (A), another 2 independentfew-layer BP flakes were prepared: the sample in Fig 2.31 (A) has an area of∼ 250µm2, and the average optical contrast shows a mean value of 9.20%, correspondingto 4 layers; the sample in Fig 2.32 (A) has two areas, and the average optical con-trast results show two means values of 6.52% and 7.68%, corresponding to bilayerand 3 layers.After that, in Fig 2.31 (B), the Raman spectrum collected at the sample shownin Fig 2.31 (A) showed 3 strong BP peaks at 356.05, 432.48 and 459.87, respec-tively. Moreover, in Fig 2.31 (C) - (E), Raman mapping results also indicated thecrystallinity of the entire sample retained after thermal thinning process.Similarly, in Fig 2.32 (B) & (C), the Raman spectra of the 4-layer / bilayer re-gion showed 3 typical BP peaks at 361.03 / 361.39 cm−1, 437.75 / 438.19 cm−1 and465.19 / 465.55 cm−1, respectively. Additionally, in Fig 2.32 (D) to (E), the spatialRaman maps of the BP sample also confirmed the crystallinity of the whole sample.Overall, our results showed that the sublimation thinning method was an ef-fective method in preparing large few-layer BP.84Figure 2.31: (A): Optical image of a BP sample thinned down to 4 layers, andthe inset figure represents the original BP sample. (B); Raman spectracollected in circled area in (A). (C) - (E): Spatial Raman maps of theBP sample shown in (A). (F) - (H): Optical contrast analysis of (A),which was splitted into R, G and B channels, and the G channel figurein (C) was selected for analysis.85Figure 2.32: (A): Optical image of a BP sample thinned down to 4 and dou-ble layers, and the inset figure represents the original BP sample. (B);Raman spectra collected in circled area in (A). (C) - (E): Spatial Ra-man maps of the BP sample shown in (A). (F) - (H): Optical contrastanalysis of (A), which was splitted into R, G and B channels, and theG channel figure in (C) was selected for analysis.86Furthermore, in Fig 2.33 (A), in order to prepare monolayer BP, a thick BPflake was carefully thinned down to have an average optical contrast of 4.88%,which corresponds to the layer number of 1 - 2 layers. Then Raman measurementsof this flake were performed: in Fig 2.33 (B), the Raman spectrum (the red profile1) collected at the “area 1” in (A) showed only the A2g mode of BP at 468 cm−1,and no A1g and B2g peaks were observed; the Raman spectrum (the black profile2) collected at the “area 2” in (A) showed the same profile as the amorphous Pshown in Fig 1.4 (A) of Chapter 1, which indicated this area to be amorphous andthe Raman signals were from the Si substrate; the blue profile 3 was the Ramanspectrum of the Si substrate (the “area 3”).Overall, the preparation of single-layer BP wasn’t successful in this work.The reason lies in: when the BP flake is thinned down to single-layer, if the heatingcontinuous, the crystallinity of BP will degrade, in other words, the phase transfor-mation will take place from crystalline BP to amorphous phosphorus. Therefore,in order to prepare crystalline single-layer BP, more works should be conducted onthe thermodynamics of phase change, which can be used for optimizing the heatingtime at specific temperatures.Figure 2.33: (A): Optical image of a BP sample thinned down to single layer.(B): Raman spectra of this sample collected at 3 regions.872.2.4 Comparison Of BP Products With Previous TechniquesIn comparison with the final BP products prepared by previous techniques, thethermal thinning method has several advantages:1. Thinning limit: The thinning limit in this work is double layers, corre-sponding to 1 nm, which is close to the mechanical exfoliation and plasma thinningmethods, identical to liquid exfoliation method, and better than the anodic oxida-tion thinning, PLD and CVD methods.2. Scalability of products: The BP products prepared in this work have areasall greater than 250 µm2, which are 10 to 100 times larger than the BP productsprepared by the previous 6 techniques.3. Uniformity 2D Raman mapping results indicate the uniformity of the BPproducts prepared in this work kept well after thinning process, which is far betterthan the BP products prepared by previous 6 techniques. The uniformity of the finalproducts is only decided by that of the pristine sample.4. Crystallinity The crystallinity of the prepared products remains after thin-ning process, which is identical to those prepared by the mechanical exfoliationmethod, and better than the other methods.5. Controllability of thickness: In next chapter, the controllability of thick-ness by Raman intensity ratios of ISiIA2gwill be discussed. However, there are nodiscussions on thickness control in the 6 previous techniques.2.3 ConclusionsIn summary, the observation of the layer-by-layer sublimation of black phos-phorus at 500 - 600 K was first reported. The thinning rates for “continuous heat-ing” were ∼ 0.18 nm/min and 1.5 nm/min at 500 and 550 K, respectively. Large(with areas > 200 µm2 ) and few-layer (2 to 4 layers) BP flakes with good integrity,uniformity and crystallinity were prepared successfully and repeatedly. No micron88scale defects were observed. Overall, the sublimation thinning method provides anew approach to reduce BP thickness to few layers while keeping its crystallinityand integrity and with a good thickness controllability. This method is also ex-pected to work for deposited BP besides exfoliated BP. The sublimation thinningmethod is promising in further fabrication of high quality few-layer BP in largescale.89Chapter 3Determination ofCrystallographic OrientationsFor BP Thickness DeterminationAnd Control By RamanSpectroscopy With 442 nmExcitation“Have not the small Particles of Bodies certain Powers, Virtues or Forces,by which they act at a distance, not only upon the Rays of Light for reflecting,refracting and reflecting them, but also upon one another for producing a greatpart of the Phenomena of Nature?”— Sir Isaac Newton3.1 Determination Of BP CO By ARPRSThe objective of this chapter is to investigated the thickness-dependent Ramanintensity ratios between the characteristic peak of BP and that of the underlying Si90(100) substrate, which can then be used as a universal reference for further thick-ness determination.However, since the anisotropic in-plane structure of BP will result in the orientation-dependent Raman response, the crystallographic orientations of the BP sampleneeds to be known first before measuring the Raman intensity ratios.For the identifications of the crystallographic orientation (CO) of BP , angle-resolved conductance of BP sheet [16][59][62], diffraction pattern by high-resolutiontransmission electron microscope [19][46], optical absorption [59][19] and Angle-resolved Polarized Raman Spectroscopy (ARPRS) have been used to determine thein-plane CO of BP [16][18][62].In consideration of the in-plane anisotropy and fast degradation of BP in air,developing a fast, convenient and non-contact method to determine the CO andthickness of BP has become necessary. ARPRS can serve as a solution to solvethese two important problems. In addition, it has been widely used for studyingthe crystallography of many materials such as GaN [103], carbon nanotube [104],strained-graphene [105], etc.3.2 Experiment MethodsIn this work, first, the ARPRS results of a BP sample on Si (100) substrateby using the excitation wavelengths of 442 nm and 633 nm were compared; thenby using the excitation wavelength of 442 nm under “parallel” configuration, theARPRS measurements of 27 BP samples on 300-nm-thick SiO2 / Si substrates andon 100-µm-thick polyimide substrates prepared by sublimation thinning and me-chanical exfoliation were performed, with the thickness range of 15 - 195 nm ( 10to 200 nm ) for samples on SiO2 / Si (polyimide) substrate.91Figure 3.1: (A): AFM profile of a blank Si wafer, which indicates its rough-ness to be∼ 0.6 nm; (B): AFM profile of a blank polyimide wafer, whichindicates its roughness to be ∼ 1 nm;3.2.1 Sample PreparationBP flakes were first mechanically exfoliated from bulk BP crystals (99.998 %,Smart Elements, Vienna, Austria) in a glove box and transferred to a pre-cleaned Si(100) wafer (4 inch, 0.56 mm thick) with a 300-nm-thick SiO2 layer and a piece of100-µm-thick polyimide film (DuPontT M Kapton HN general-purpose film. In Fig3.1 (A) and (B), the surface roughness of SiO2 / Si (polyimide) from the line-scanAFM results was around 0.3 nm (1 nm). The wafer and polyimide were then cutinto small pieces (3 mm × 3 mm). In order to prepare thin BP flakes with variousthickness less than 50 nm, thermal thinning (the samples would be heated at 230◦C, see the methods in Chapter 2) were performed on some exfoliated BP flakes.In this study, those BP samples on SiO2 / Si with thickness less than 45 nm wereprepared by thermal sublimation. The BP samples on polyimide substrate with thethickness of: 10 - 40 nm, 58 - 70 nm and 85 nm were prepared by the thermalsublimation method.92Figure 3.2: Illustration of ARPRS experiment setup: the optical path of theRaman system under “parallel” configuration.3.2.2 Angle-resolved Polarized Raman Spectroscopy (ARPRS)In Fig 3.3 (A) & (B), the experimental setup was illustrated: typically, the BPwas placed in the center of a rotation stage mounted on the motorized stage un-der the objective lens; the Raman spectra were collected by using a backscatteredHoriba Jobin Yvon HR800 Raman system with the excitation wavelength of 442/ 633 nm (2.81 eV / 1.95 eV ) from the He - Cd / He - Ne laser. The 442 / 633nm incident laser beam was polarized, and a polarization analyzer was placed inparallel “configuration” (incidence and scattering light are parallel) between thenotch filter and the entrance of the CCD detector [16]. In this work, all ARPRSexperiments were conducted under the “parallel” configuration following previousworks [18][16][106]. BP samples were rotated clockwisely for 360◦ in 24 steps(15◦/step). The angle values of the polar diagrams were angle readings of the ro-tation stage. During every step, Raman measurement was performed at the samepoint on each sample to ensure the consistency of results. The grating number of93detector was set as 2400 and the spectral range from 250 / 300 to 600 / 600 cm−1for the excitation wavelength of 442 / 633 nm. The spectral resolution before fittingwas ≈ 0.27 / 0.09 cm−1 for the excitation wavelength of 442 / 633 nm. Hereafter,considering that polyimide substrate was damaged by 442 laser without reducingits power (around 2 mW / µm2), filter D = 1 was then chosen in all ARPRS exper-iments in this work. This lowered the laser power by 10 times (around 0.2 mW /µm2) in order to avoid damage on both the samples and polyimide substrate. Forexcitation wavelength of 633 nm, filter D = 0.3 was selected. The acquisition timeand accumulation times were optimized at 3 sec and ×2 for both 442 and 633 nmlasers in order to enhance the Raman signals of BP samples as well as minimizelaser damage to BP samples.3.2.3 Atomic Force Microscopy (AFM)AFM measurements were performed using the contact mode by an Asylum Re-search Molecular Force Probe 3D atomic force microscope and a Bruker AtomicForce Microscopy System.3.3 Results And Discussion3.3.1 ARPRS Study In The Determination Of CrystallographicOrientation (CO) Of BPIntensity Diagrams Of A1g And A2g ModesFirst, ARPRS measurements with excitation wavelengths of 442 nm and 633nm were performed on a thick BP flake on SiO2 / Si substrate. In Fig 3.3, opticalimages showed the positions of the BP sample where A1g mode has maximum (Fig3.3 (A)) and A2g (Fig 3.3 (B)) Raman intensities.Let us look at the Raman spectra of the BP sample shown in Fig 3.4. In Fig3.4, (A) / (C) and (B) / (D) are the Raman spectra of the sample shown in Fig 3.394Figure 3.3: (A): Optical image of the BP sample on SiO2 / Si substrate rotatedat the position with maximum Raman intensity of A1g mode and (B):at the position with the maximum Raman intensity of A2g mode. Thecircled area indicates where the Raman spectra were collected;(A) and (B), respectively, which were measured by using 442 / 633 nm excitation.In addition, (A) / (C) and (B) / (D) have maximum A1g and A2g Raman intensities.In Fig 3.4 (A), the peak positions of the A1g and the A2g showed at 362.22 cm−1and 466.81 cm−1, respectively; A1g reached its maximum Raman intensity while A2ggot a minimum Raman intensity; B2g disappeared and A2g reached its minimum Ra-man intensity. The intensity counts of the A1g and A2g mode were ∼ 130 and ∼ 50,respectively. In Fig 3.4 (B), A2g mode showed at 466.81 cm−1 while A1g became in-visible, and notably, the intensity counts of the A2g mode was increased to ∼ 1100,which was more than 20 times of that shown in Fig 3.4 (A). However, for the ex-citation wavelength of 633 nm, in Fig 3.4 (C) and (D), maximum and minimumRaman intensities of A1g (A2g) modes were ∼ 70 (∼ 75) and ∼ 55 ( ∼ 50), whichshowed no apparent differences in comparison with those from Raman spectra ac-quired by the 442 nm laser.95Figure 3.4: (A) & (C): Raman spectra of the BP sample on SiO2 / Si substrateshown in Fig 3.3 (A), in which the A1g mode showed maximum Ramanintensity; (B) & (D): Raman spectra of the BP sample shown in Fig 3.3(B), in which the A1g mode showed maximum Raman intensity. Ramanmeasurements shown in (A) & (B) were performed using the excitationwavelength of 442 nm while those in (C) & (D) were by 633 nm.96Then in Fig 3.5 (A) to (C) / (D) to (F), the intensity polar diagrams were usedto demonstrate the ARPRS results acquired by using 442 / 633 nm laser. For the ex-citation wavelength of 442 nm, in Fig 3.5 (A), the Raman intensity polar diagramof A1g mode showed its maximum Raman intensity at 165◦ and 345◦ of samplerotation angle (labelled by double-head arrows); in Fig 3.5 (C), maximum Ramanintensities of A2g appeared at 75◦ and 255◦ of the rotation angle. For the excitationwavelength of 633 nm, though the variation of A1g mode was similar to that fromthe results of 442 nm excitation, but the change of A2g mode was different fromthat measured by 442 nm laser, in that when A1g reached its maximum Raman in-tensity, A2g mode reached a relative smaller maximum Raman intensity instead ofa minimum shown in Fig 3.5 (C). In addition, for both 442 and 633 nm excitation,the B2g mode was invisible when the A1g or A2g mode reached its maximum Ramanintensity.As discussed above in section 1.3.3 (Page 35), under parallel configuration,B2g mode was always forbidden when the incident laser was parallel to eitherzigzag or armchair direction. Therefore, according to Fig 3.5, the rotation anglesα showing at 60 ◦ / 240 ◦ and 150 ◦ / 330 ◦ indicated the two axes of BP (the zigzagand armchair directions, and in our work, the definition of zigzag direction as the“main axis” [19]) was followed.It is suggested that the ARPRS results acquired by 633 nm laser could notprovide simple and fast distinguishment of the zigzag and armchair directions fromthese two axes. The reason could be explained that according to the ARPRS of thiswork shown in Fig 3.5 (D) - (F), and previous works [18][19], the relative larger/ smaller local maximum Raman intensity of A2g mode could appear at the angleeither perpendicular or parallel to the angle of the maximum A1g Raman intensity,which made it infeasible for the determination of the CO of BP simply from atwhich angle the A1g and A2g modes had maximum and minimum Raman intensities.97Figure 3.5: ARPRS results with 442 nm excitation: (A): the polar diagram ofA1g mode; (B): the polar diagram of B2g mode; (C): the polar diagram ofA2g mode. ARPRS results with 633 nm excitation: (D): the polar diagramof A1g mode; (E): the polar diagram of B2g mode; (F): the polar diagramsof A2g mode. The arrows in figures indicate the two axes of BP (zigzagand armchair).98But for 442 nm excitation, according to the ARPRS results shown in Fig 3.4(A) & (B) and Fig 3.5 (A) - (C) of this work and previous work [18], ARPRS with442 nm excitation could be able to tell the CO of BP in that: 1. both intensity polardiagrams of A1g and A2g modes showed clear bow - tie shapes instead of elliptical/ circular / peanut shapes, which made the anlges of maximum and minimum Ra-man intensities more readable; 2. the maximum and minimum Raman intensitiesof A1g and A2g modes always appeared in orthogonal directions explicitly; 3. whenincident polarization was parallel to armchair direction, the Raman intensity of A2gmode was dozens of times of that measured when incident polarization was paral-lel to zigzag direction.Though Kim et al. [18] proposed the effectiveness of ARPRS with 442 nmexcitation in CO determination of BP, the data were limited for only 4 samples;in addition, though Ling et al. [19] systematically studied the thickness-dependentARPRS of BP by using 532, 633 and 785 nm excitation and pointed out their ARPRSresults couldn’t well indicate the CO of BP, they didn’t investigate the thickness-dependent ARPRS of BP using 442 nm excitation. Overall, using ARPRS with 442nm excitation for CO determination of BP has not been widely accepted.In order to reveal this effectiveness of CO determination of BP by ARPRS withexcitation wavelength of 442 nm, investigations of BP in a wide range of thicknesswere conducted: 11 and 16 BP samples with thickness ranging from 10 to 200 nmon SiO2 / Si and polyimide substrates were prepared. ARPRS studies by 442 nmlaser were performed on those samples. All the results (See Table 3.2 to Table 3.5)were consistent with the Raman spectra shown in Fig 3.4 and the polar diagramsshown in Fig 3.5.Therefore, with expanded thickness ranges and the addition of polyimide sub-strates, it could be concluded that, with the excitation wavelength of 442 nm,the CO identification (zigzag and armchair directions) agrees with previous work[16][18][62]: in a polar diagram, the direction along which the maximum Ramanpeak intensity of A1g was obtained is always perpendicular to the direction for themaximum A2g peak intensity. No thickness dependence was obtained, which meant99that the CO determination by ARPRS using 442 nm excitation could be conductedregardless of the BP thickness. Additionally, the fact that A2g Raman intensity mea-sured at armchair direction are enhanced to be 10 - 100 times of that measuredat zigzag direction, could be attributed to the resonance Raman effect. Therefore,more investigations should be conducted to study the conditions of this orientation-dependent resonance Raman effect of BP in the future.• 1. Polar diagrams of ARPRS results: 11 samples on SiO2 / Si substratesNo. Thickness (nm) Polar diagram of A1g mode Polar diagram of A2g mode1. 152. 203. 251004. 305. 456. 507. 608. 651019. 7510. 11511. 195Table 3.1: Polar diagrams of peak intensity ratios of A1g over A2g of 11 BPsamples on SiO2 / Si substrates acquired by Angle-resolved PolarizedRaman Spectroscopy (ARPRS).102• 2. Raman spectra of ARPRS results: 11 samples on SiO2 / Si substratesNo. Thickness (nm) Raman profile: maximumA1g mode intensityRaman profile: maximumA2g mode intensity1. 152. 203. 251034. 305. 456. 501047. 608. 659. 7510510. 11511. 195Table 3.2: Raman spectra: with maximum A1g / A2g Raman intensities.106• 3. Polar diagrams of ARPRS results: 16 samples on polyimide sub-stratesNo. Thickness (nm) Polar diagram of A1g mode Polar diagram of A2g mode1. 102. 203. 254. 301075. 406. 457. 558. 589. 6010810. 7011. 8312. 8513. 9514. 13510915. 15016. 200Table 3.3: Polar diagrams of 16 BP samples on polyimide substrates ac-quired by Angle-resolved Polarized Raman Spectroscopy (ARPRS).• 4. Raman spectra of ARPRS results: 16 samples on polyimide substratesNo. Thickness (nm) Raman profile: maximumA1g mode intensityRaman profile: maximumA2g mode intensity1. 101102. 203. 254. 301115. 406. 457. 551128. 589. 6010. 7011311. 8312. 8513. 9511414. 13515. 15016. 200Table 3.4: Raman spectra: with maximum A1g / A2g Raman intensities.115• 5. AFM data of BP samples on Si (100) waferNo. Thickness Optical image AFM profile1. 15 nm2. 20 nm3. 25 nm4. 30 nm5. 45 nm6. 50 nm1167. 60 nm8. 65 nm9. 75 nm10. 115 nm11. 195 nmTable 3.5: Optical images, AFM thickness profiles of 11 BP Samples onSiO2/Si substrate. No.1-5 were samples prepared by sublimation thin-ning; No.6-12 were samples prepared by exfoliation.117• 6. AFM data of BP samples on polyimide substrateNo. Thickness Optical image AFM profile1. 10 nm2. 20 nm3. 25 nm4. 30 nm5. 40 nm6. 45 nm7. 55 nm1188. 58 nm9. 60 nm10. 70 nm11. 83 nm12. 85 nm13. 95 nm14. 135 nm15. 150 nm11916. 200 nmTable 3.6: Optical images, AFM images and thickness profiles of 16 BPSamples on polyimide substrate. No.1-5, 8, 10 and 12 were samplesprepared by sublimation thinning; No.6-7, 9, 11 and 13-16 were samplesprepared by exfoliation.3.3.2 Raman Peak Intensity Ratios: ISiIA2gAs A Function Of BPThickness3.3.2.1 Definitions Of “Off-axis” And “On-axis” Raman Intensity Ratios ISiIA2gAfter confirming the effectiveness of CO determination of BP by ARPRS with442 nm excitation, let us focus on the Raman intensity ratios of ISiIA2gas a universalmethod for BP thickness determination.The ARPRS results in the last section revealed the fact that if the incident po-larization is parallel to the “zigzag direction”, A1g mode can reach its maximum,and both A1g and Si peak can be seen, and if the incident polarization is parallel tothe “armchair” direction, A2g mode can reach its maximum but the Si peak cannotbe seen because of the enhancement of A2g Raman intensity.Therefore, for investigations of thickness-dependent Raman intensity ratios,the Raman intensity ratio of ISiIA2gwas selected. Therefore, the samples should berotated to have their “zigzag” directions to be parallel to the incident polarization.Furthermore, for more accurate determination of the layer number N by usingthe Raman intensity ratio of ISiIA2g, it is expected to have as large Raman intensity ofunderlying Si (100) substrate as possible. Thus, it is always expected to rotate theSi (100) wafer to make the (100) plane also be parallel to the incident polarization120to have maximum Si Raman intensity ISi. This situation is termed as “on-axis”which is very seldom.It will be more often that the “zigzag” direction of BP and (100) plane of theunderlying Si substrate are not both parallel to the incident polarization at the sametime. This situation is termed as “off-axis” in this work.When meeting the case of “off-axis”, the first step is to let the “zigzag” di-rection of BP be parallel to the incident polarization; then the measured Ramanpeak intensity from the underlying Si substrate and the rotation angle “α” at thismoment are also recorded.In the other side, the orientations of that Si substrate can be also measuredby ARPRS directly at the blank area of this wafer. Therefore, the smallest angledifference between the rotation angle “β” where the maximum blank Si Ramanintensity appears and the zigzag direction of BP can be known. After that, becausethe orientation-dependent Raman intensity of Si (100) wafer is a sinusoidal func-tion with a periodicity of 90◦ [91][92], the measured underlying Si Raman intensitycan be then converted to that at the angle where the maximum blank Si Raman in-tensity appears according that known smallest angle difference.Through this method, the “off-axis” underlying Si Raman intensity can bethen converted to “on-axis” underlying Si peak intensities. The conversion willhelp to realize the expected ”on-axis” situation.According to the in-situ high temperature Raman spectra in Fig 2.23 and Fig2.25 of Chapter 2, the Si Raman intensity increased monotonically with decreasingBP thickness. Therefore, after the BP sample being rotated at the angle (θ = 0 ◦)where the incident polarization was parallel to its main axis of zigzag direction, foreach BP thickness, there could be always a corresponding Raman intensity ratio ISiIA2gbetween the A2g mode and underlying Si (100) substrate.121What’s more, meeting the case of “off-axis” between the BP and underlyingSi substrate, the ratio ofISi,θ ′ 6=0IA2gcould then be converted to the “on-axis”ISi,θ ′=0IA2gac-cording to the smallest angle difference θ ′. As mentioned in Section 2.2.2 (P 68)of Chapter 2, Raman spectra were collected for each cycle of the periodic heat-ing process on the sample A and B, and both “sample A” and “sample B” weremeasured to have a difference of 15◦ between the rotation angles of maximum A1gand Si intensity (Page 45). Therefore, the ”on-axis” underlying Si Raman intensityshould twice of the measured underlying Si Raman intensity.3.3.2.1 Results Of “Off-axis” And “On-axis” Raman Intensity Ratios ISiIA2gIn this thesis work, the ARPRS measurements were performed on “SampleA” and “Sample B” before used for the interrupted heating experiments in Section2.2.2 (P 65) of Chapter 2.In Fig 3.6, the intensities of BP A1g and A2g modes as well as underlying Si(100) substrate of “Sample A” were plotted as functions of rotation angles. The ro-tation angles of maximum A1g and Si peak intensity showed that the smallest angledifference between the maximum Raman intensity of Si and that of BP A2g modewas 15◦. Similarly, in Fig 3.7, those of “sample B” also showed a smallest angledifference of 15◦. Thus, the two samples “Sample A” and “Sample B” met the“off-axis” case.Since the smallest angle differences of “Sample A” and “Sample B” were both15◦, if converting this “off-axis” case to the ”on-axis” case, the measured under-lying Si Raman intensity should be 50% of the maximum underlying Si Ramanintensity because of the sinusoidal function with a periodicity of 90◦ mentionedbefore.122Figure 3.6: “Sample A”. Raman intensities of A1g and A2g modes and under-lying (100) Si wafer as functions of the rotation angle.123Figure 3.7: “Sample B”. Raman intensities of BP A1g and A2g modes and un-derlying Si(100) as functions of the rotation angle.124As mentioned before at the beginning of the “Interrupted Heating” experi-ments of Section 2.2.2 (P 65) in Chapter 2: for each heating cycle, 2 Raman spectrawere taken before and after heating at room temperature respectively and 2 Ramanspectra were collected at the beginning and ending of the heating process at hightemperature; in the other side, the thicknesses of BP before and after each cyclewere measured by AFM. Then the thickness-dependent Raman intensity ratios ofISiIA2gfrom those measurements were plotted in Fig 3.9 and Fig 3.10.In Fig 3.9, overall, the Raman intensity ratio of ( ISiIA2g) increased monotonicallywith decreasing thickness, and the results from measurements at room temper-ature show that the two Raman intensity ratio -thickness curves of “Sample A”and “Sample B” agreed well, despite a minor disagreement at the data points be-tween the 32-nm-sample of “Sample A” (blue solid circle) and the 32-nm-sampleof “Sample B” (red solid cubic). This can be explained by 2 reasons:1. The experimental errors might be introduced by AFM measurements. Inother words, the surface oxidation took place because of the frequent exposure inair ambient during AFM measurements, and with decreasing thickness, the oxida-tion proceeded faster [63], and then the measured thickness might be greater thanthe real thickness of prepared fresh sample. Therefore, the thickness of that 32-nm-thick data point of “Sample B” (red solid cubic) should be less than 32 nm.2. Before this cycle of heating, the experimental errors might be caused bythat the minor vibrations during the placement of “Sample A”, and then make theBP sample be a bit off the angle where its A1g mode have the maximum Ramanintensity.So far, the general agreement of those two profiles indicates that, for “off-axis” case between the BP sample and underlying Si substrate, with each definitesmallest angle difference, there is always only one Raman intensity ratio of ISiIA2gforeach thickness.125Figure 3.8: Room temperature Raman intensity ratios of SiA2g as a function ofthickness measured by AFM, here the blue line represents the sample Aand the red line represents sample B.126Figure 3.9: High temperature Raman intensity ratios of SiA2g as a function ofmeasured thickness, here the blue line represents sample A treated at500 K and the red line represents sample B at 550 K.127Figure 3.10: High temperature Raman intensity ratios of SiA2g as a function ofmeasured thickness, here the blue line represents sample A treated at500 K and the red line represents sample B at 550 K.In addition, the high temperature thickness-dependent Raman intensity ratiosISiIA2gof “Sample A” (measured at 500 K) and “Sample B” (measured at 550 K) Wereplotted in Fig 3.9. It is important to note that the thicknesses were from the AFMmeasurements at room temperature and the consideration of no thickness attenua-tion during the ramping-up and cooling-down process. These two profiles providesthe capability of in-situ determining BP thickness at 500 K and 550 K respectively.After that, the “off-axis” Raman intensity ratios ISiIA2gwere then converted to the “on-axis” ones according to the smallest angle differences of 15◦ which were shown inFig 3.10. Therefore, with known smallest angle difference between the maximumRaman intensity of A1g and that of the underlying Si substrate. The “on-axis” Ra-man intensity ratios of ISiIA2gcan always be achieved.In general, for future purpose of non-contact, in and ex -situ, instant determi-nation of the thickness during fabrication, the thickness-dependent Raman intensityratios ( ISiIA2g) can be served as a solution as long as enough references have been pro-128vided. That is to say, if the thickness-dependent Raman intensity ratios of ( ISiIA2gatmany smallest angle differences are provided, this method can satisfy the purpose.3.3.3 Conclusion:To summarize, ARPRS measurements of 10 to 200 nm thick black phosphorusflakes on SiO2 / Si and polyimide substrates by angle-resolved polarized RamanSpectroscopy were performed. The results revealed that ARPRS with 442 nm ex-citation and under parallel configuration can provide unambiguous, convenient,non-destructive and fast determination of the crystallographic orientation of BP.In addition, the Raman intensity ratio SiA2g increases monotonically with decreasingthickness, which can serve as a non-contact optical method for determining thick-ness of BP both in-situ and ex-situ.129Chapter 4Conclusion and outlook4.1 ConclusionAs stated in Chapter 1, BP is promising for wide applications including ther-mal imaging, thermoelectronics, fibre-optic communications, photo-voltaics anddigital / high frequency electronics. The controllable isolation of BP to atomicallythin layers remains as one of the key challenges in current research. To developa method for economic controllable growth of large scale, uniform and crystallinefew-layer BP film is the main goal of this thesis.This thesis present the experimental study in preparation and characterizationof few-layer BP: 1. we achieved the successful and repeatable isolation of largefew-layer crystalline BP by thermal thinning of thick BP flakes; 2. We revelaed thatRaman spectroscopy could be used as a powerful and versatile tool for: 1. clearlyindentifying the CO of BP; 2. determining the instant thickness of BP in and ex-situ; 3. providing the crystallinity and uniformity information non-destructively.We observed micro-scale thickness attenuation of BP at the high temperatureby the color changes under the optical microscope. After the heating process, Ra-man mapping results demonstrate the size, uniformity and crystallinity of the BPsamples retained, and indicate the thinning process of BP to be isotropic within theplane. Interrupted and continuous heating processes at 500 K and 550 K were per-130formed at BP samples. The AFM and Raman spectra results measured the thicknessof BP before and after the heating process. The results indicate the thinning ratesof BP at 500 K and 550 K were measured to be 0.15 nm / min and 2 nm / minfrom AFM results of BP samples before and after heating process. After that, AFM,Raman mapping and optical contrast demonstrated the successful and repeatablepreparation of the large area ( > 200 µm2 ), few-layer ( < 5 layers ), crystalline BPby the thermal thinning method.The Angle-resolved Polarized Raman Spectroscopy with the excitation wave-length of 442 nm measured the CO of BP: 1. both intensity polar diagrams of A1gand A2g modes showed clear bow - tie shapes instead of elliptical / circular / peanutshapes, which made the anlges of maximum and minimum Raman intensities morereadable; 2. the maximum and minimum Raman intensities of A1g and A2g modesalways appeared in orthogonal directions explicitly; 3. when incident polarizationwas parallel to armchair direction, the Raman intensity of A2g mode was dozens oftimes of that measured when incident polarization was parallel to zigzag direction.After finishing CO identification, the sample was aligned to have maximum A1g Ra-man intensity, and then the Raman intensity ratio A2gSi and thickness of BP showedone-to-one correspondence, which indicated that, Raman spectroscopy can be usedas a non-contact tool for in and ex -situ thickness measurement.Overall, this work present the effectiveness of the thermal thinning methodin preparing large, uniform, crystalline few-layer BP. This work also contributedto the development of the all-Raman method in controllable sublimation thinningof BP. This all-Raman method integrated the 3 major functions: 1. identificationof CO of BP; 2. determination of the thickness of BP; 3. confirmation of thecrystallinity and uniformity, which are the 3 criticle steps of controllable thinningof BP.4.2 OutlookAs an outlook for future studies, and by taking advantage of this work, 3 po-tential directions are suggested here in order to improve the fabrication technique,131study the basic properties like the electronic and optical properties of preparedfew-layer BP, and explore its wide applications: 1. controllable preparation oflarge, uniform, crystalline few-layer BP on other substrates; 2. measurements ofthe angle or layer dependent transport properties such as the carrier mobility, theon-off ratio, resistivity and etc of BP-based devices; 3. by stacking with other 2Dmaterials to form 2D van der Waals heterostructure, IR absorption then might gainsignificant enhancement, which might be promising for making wearable electron-ics and solar cells. In addition, more investigations also need to be conducted to:1. the control of sublimation and the critical temperature of phase change and 2:study the conditions of resonance Raman effect of BP in the future.132Bibliography[1] Lewis, E. A., Brent, J. R., Derby, B., Haigh, S. J. & Lewis, D. J. Solutionprocessing of two-dimensional black phosphorus. ChemicalCommunications (2017). → pages ix, 16[2] Long, D. The raman effect: a unified treatment of the theory of ramanscattering by molecules. 2002. West Sussex, England: John Wiley & SonsLtd . → pages ix, xii, 28, 29, 30[3] Li, L., Yu, Y., Ye, G. J., Ge, Q., Ou, X., Wu, H., Feng, D., Chen, X. H. &Zhang, Y. Black phosphorus field-effect transistors. Naturenanotechnology 9, 372–377 (2014). → pages x, xi, 1, 3, 4, 13, 14[4] Castellanos-Gomez, A. Black phosphorus: narrow gap, wide applications.The journal of physical chemistry letters 6, 4280–4291 (2015). → pages x,5, 6, 13[5] Lam, K.-T., Dong, Z. & Guo, J. Performance limits projection of blackphosphorous field-effect transistors. IEEE Electron Device Letters 35,963–965 (2014). → pages x, 7, 8[6] Qiao, J., Kong, X., Hu, Z.-X., Yang, F. & Ji, W. High-mobility transportanisotropy and linear dichroism in few-layer black phosphorus. Naturecommunications 5 (2014). → pages x, 9, 10[7] Yuan, H., Liu, X., Afshinmanesh, F., Li, W., Xu, G., Sun, J., Lian, B.,Curto, A. G., Ye, G., Hikita, Y. et al. Polarization-sensitive broadbandphotodetector using a black phosphorus vertical p–n junction. Naturenanotechnology 10, 707–713 (2015). → pages x, 9, 10[8] Li, Y., Hu, Z., Lin, S., Lai, S. K., Ji, W. & Lau, S. P. Giant anisotropicraman response of encapsulated ultrathin black phosphorus by uniaxialstrain. Advanced Functional Materials (2016). → pages xi, 11, 12133[9] Liu, H., Neal, A. T., Zhu, Z., Luo, Z., Xu, X., Toma´nek, D. & Ye, P. D.Phosphorene: an unexplored 2d semiconductor with a high hole mobility.ACS nano 8, 4033–4041 (2014). → pages xi, 14, 82[10] Sresht, V., Pa´dua, A. A. & Blankschtein, D. Liquid-phase exfoliation ofphosphorene: design rules from molecular dynamics simulations. ACSnano 9, 8255–8268 (2015). → pages xi, 15[11] Yasaei, P., Kumar, B., Foroozan, T., Wang, C., Asadi, M., Tuschel, D.,Indacochea, J. E., Klie, R. F. & Salehi-Khojin, A. High-quality blackphosphorus atomic layers by liquid-phase exfoliation. Advanced Materials27, 1887–1892 (2015). → pages xi, 15, 16[12] Lu, W., Nan, H., Hong, J., Chen, Y., Zhu, C., Liang, Z., Ma, X., Ni, Z., Jin,C. & Zhang, Z. Plasma-assisted fabrication of monolayer phosphorene andits raman characterization. Nano Research 7, 853–859 (2014). → pages xi,13, 17, 39, 50[13] Liu, X., Chen, K.-S., Wells, S. A., Balla, I., Zhu, J., Wood, J. D. & Hersam,M. C. Scanning probe nanopatterning and layer-by-layer thinning of blackphosphorus. Advanced Materials (2016). → pages xi, 13, 18[14] Smith, J. B., Hagaman, D. & Ji, H.-F. Growth of 2d black phosphorus filmfrom chemical vapor deposition. Nanotechnology 27, 215602 (2016). →pages xi, 13, 19[15] Yang, Z., Hao, J., Yuan, S., Lin, S., Yau, H. M., Dai, J. & Lau, S. P.Field-effect transistors based on amorphous black phosphorus ultrathinfilms by pulsed laser deposition. Advanced Materials 27, 3748–3754(2015). → pages xi, 13, 20, 21, 22, 23[16] Wu, J., Mao, N., Xie, L., Xu, H. & Zhang, J. Identifying the crystallineorientation of black phosphorus using angle-resolved polarized ramanspectroscopy. Angewandte Chemie 127, 2396–2399 (2015). → pages xii,24, 25, 34, 91, 93, 99[17] Ribeiro, H. B., Pimenta, M. A., de Matos, C. J., Moreira, R. L., Rodin,A. S., Zapata, J. D., de Souza, E. A. & Castro Neto, A. H. Unusual angulardependence of the raman response in black phosphorus. ACS nano 9,4270–4276 (2015). → pages xii, 24, 25[18] Kim, J., Lee, J.-U., Lee, J., Park, H. J., Lee, Z., Lee, C. & Cheong, H.Anomalous polarization dependence of raman scattering and134crystallographic orientation of black phosphorus. Nanoscale 7,18708–18715 (2015). → pages xii, 24, 25, 26, 48, 66, 91, 93, 97, 99[19] Ling, X., Huang, S., Hasdeo, E. H., Liang, L., Parkin, W. M., Tatsumi, Y.,Nugraha, A. R., Puretzky, A. A., Das, P. M., Sumpter, B. G. et al.Anisotropic electron-photon and electron-phonon interactions in blackphosphorus. Nano letters 16, 2260–2267 (2016). → pages xii, 24, 25, 27,34, 48, 91, 97, 99[20] Tanaka, M. & Young, R. Review polarised raman spectroscopy for thestudy of molecular orientation distributions in polymers. Journal ofMaterials Science 41, 963–991 (2006). → pages xii, 31[21] Graves, P., Gardiner, D. & Graves, P. Practical raman spectroscopy (1989).→ pages xii, 31, 33, 34[22] Wu, J., Li, H., Yin, Z., Li, H., Liu, J., Cao, X., Zhang, Q. & Zhang, H.Layer thinning and etching of mechanically exfoliated mos2 nanosheets bythermal annealing in air. Small 9, 3314–3319 (2013). → pages xiii, 39, 40[23] Huang, Y.-K., Cain, J. D., Peng, L., Hao, S., Chasapis, T., Kanatzidis,M. G., Wolverton, C., Grayson, M. & Dravid, V. P. Evaporative thinning: Afacile synthesis method for high quality ultrathin layers of 2d crystals. ACSnano 8, 10851–10857 (2014). → pages xiii, 39, 40[24] Lu, W., Zhang, Y., Zhu, Z., Lai, J., Zhao, C., Liu, X., Liu, J. & Sun, D.Thin tungsten telluride layer preparation by thermal annealing.Nanotechnology 27, 414006 (2016). → pages xiii, 39, 40[25] Liu, X., Wood, J. D., Chen, K.-S., Cho, E. & Hersam, M. C. In situ thermaldecomposition of exfoliated two-dimensional black phosphorus. TheJournal of Physical Chemistry Letters 6, 773–778 (2015). → pages xiii, 41,53, 65[26] Fortin-Descheenes, M., Levesque, P. L., Martel, R. & Moutanabbir, O.Dynamics and mechanisms of exfoliated black phosphorus sublimation.The journal of physical chemistry letters 7, 1667–1674 (2016). → pagesxiii, 41, 53, 65[27] Rosso, D. 2015 international technology roadmap for semiconductors(itrs). www.semiconductors.org 3 (2017). → pages 1[28] Ornes, S. News feature: Beyond graphene. Proceedings of the NationalAcademy of Sciences 112, 13128–13130 (2015). → pages 1135[29] Schwierz, F., Pezoldt, J. & Granzner, R. Two-dimensional materials andtheir prospects in transistor electronics. Nanoscale 7, 8261–8283 (2015).→ pages 1[30] Novoselov, K. S., Geim, A. K., Morozov, S. V., Jiang, D., Zhang, Y.,Dubonos, S. V., Grigorieva, I. V. & Firsov, A. A. Electric field effect inatomically thin carbon films. science 306, 666–669 (2004). → pages 1, 14[31] Zhu, W., Yogeesh, M. N., Yang, S., Aldave, S. H., Kim, J.-S., Sonde, S.,Tao, L., Lu, N. & Akinwande, D. Flexible black phosphorus ambipolartransistors, circuits and am demodulator. Nano letters 15, 1883–1890(2015). → pages 2[32] Zhu, W., Park, S., Yogeesh, M. N., McNicholas, K. M., Bank, S. R. &Akinwande, D. Black phosphorus flexible thin film transistors at gighertzfrequencies. Nano letters 16, 2301–2306 (2016). → pages 2[33] Kang, J., Jariwala, D., Ryder, C. R., Wells, S. A., Choi, Y., Hwang, E., Cho,J. H., Marks, T. J. & Hersam, M. C. Probing out-of-plane charge transportin black phosphorus with graphene-contacted vertical field-effecttransistors. Nano letters 16, 2580–2585 (2016). → pages 2[34] Jeon, P. J., Lee, Y. T., Lim, J. Y., Kim, J. S., Hwang, D. K. & Im, S. Blackphosphorus–zinc oxide nanomaterial heterojunction for p–n diode andjunction field-effect transistor. Nano letters 16, 1293–1298 (2016). →pages 2[35] Youngblood, N., Chen, C., Koester, S. J. & Li, M. Waveguide-integratedblack phosphorus photodetector with high responsivity and low darkcurrent. Nature Photonics; doi:10.1038/nphoton.2015.23 (2015). → pages2, 4[36] Chen, C., Youngblood, N., Peng, R., Yoo, D., Mohr, D. A., Johnson, T. W.,Oh, S.-H. & Li, M. Three-dimensional integration of black phosphorusphotodetector with silicon photonics and nanoplasmonics. Nano Letters(2017). → pages 2[37] Lin, C., Grassi, R., Low, T. & Helmy, A. S. Multilayer black phosphorus asa versatile mid-infrared electro-optic material. Nano letters 16, 1683–1689(2016). → pages 2[38] Zhao, R., Li, J., Zhang, B., Li, X., Su, X., Wang, Y., Lou, F., Zhang, H. &He, J. Triwavelength synchronously mode-locked fiber laser based on136few-layered black phosphorus. Applied Physics Express 9, 092701 (2016).→ pages 2[39] Yang, Y., Gao, J., Zhang, Z., Xiao, S., Xie, H.-H., Sun, Z.-B., Wang, J.-H.,Zhou, C.-H., Wang, Y.-W., Guo, X.-Y. et al. Photocathodes: Blackphosphorus based photocathodes in wideband bifacial dye-sensitized solarcells (adv. mater. 40/2016). Advanced Materials 28, 8787–8787 (2016). →pages 2[40] Wang, T., Hu, S., Chamlagain, B., Hong, T., Zhou, Z., Weiss, S. M. & Xu,Y.-Q. Visualizing light scattering in silicon waveguides with blackphosphorus photodetectors. Advanced Materials 28, 7162–7166 (2016). →pages 2[41] Ling, X., Wang, H., Huang, S., Xia, F. & Dresselhaus, M. S. Therenaissance of black phosphorus. Proceedings of the National Academy ofSciences 112, 4523–4530 (2015). → pages 2[42] Bridgman, P. Two new modifications of phosphorus. Journal of theAmerican Chemical Society 36, 1344–1363 (1914). → pages 2[43] Splendiani, A., Sun, L., Zhang, Y., Li, T., Kim, J., Chim, C.-Y., Galli, G. &Wang, F. Emerging photoluminescence in monolayer mos2. Nano letters10, 1271–1275 (2010). → pages 4[44] Mak, K. F., Lee, C., Hone, J., Shan, J. & Heinz, T. F. Atomically thin mos2: a new direct-gap semiconductor. Physical Review Letters 105, 136805(2010). → pages 4[45] Zhao, W., Ghorannevis, Z., Chu, L., Toh, M., Kloc, C., Tan, P.-H. & Eda,G. Evolution of electronic structure in atomically thin sheets of ws2 andwse2. ACS nano 7, 791–797 (2012). → pages 4[46] Castellanos-Gomez, A., Vicarelli, L., Prada, E., Island, J. O.,Narasimha-Acharya, K., Blanter, S. I., Groenendijk, D. J., Buscema, M.,Steele, G. A., Alvarez, J. et al. Isolation and characterization of few-layerblack phosphorus. 2D Materials 1, 025001 (2014). → pages 4, 24, 39, 50,82, 91[47] Liang, L., Wang, J., Lin, W., Sumpter, B. G., Meunier, V. & Pan, M.Electronic bandgap and edge reconstruction in phosphorene materials.Nano letters 14, 6400–6406 (2014). → pages 4137[48] Takao, Y., Asahina, H. & Morita, A. Electronic structure of blackphosphorus in tight binding approach. Journal of the Physical Society ofJapan 50, 3362–3369 (1981). → pages 4[49] Tran, V., Soklaski, R., Liang, Y. & Yang, L. Layer-controlled band gap andanisotropic excitons in few-layer black phosphorus. Physical Review B 89,235319 (2014). → pages 4, 15, 18, 68, 82[50] Qin, G., Yan, Q.-B., Qin, Z., Yue, S.-Y., Cui, H.-J., Zheng, Q.-R. & Su, G.Hinge-like structure induced unusual properties of black phosphorus andnew strategies to improve the thermoelectric performance. Scientificreports 4, 6496 (2014). → pages 4[51] Lv, H., Lu, W., Shao, D. & Sun, Y. Large thermoelectric power factors inblack phosphorus and phosphorene. arXiv preprint arXiv:1404.5171(2014). → pages 4[52] Lu, S., Miao, L., Guo, Z., Qi, X., Zhao, C., Zhang, H., Wen, S., Tang, D. &Fan, D. Broadband nonlinear optical response in multi-layer blackphosphorus: an emerging infrared and mid-infrared optical material. Opticsexpress 23, 11183–11194 (2015). → pages 4[53] Hong, T., Chamlagain, B., Lin, W., Chuang, H.-J., Pan, M., Zhou, Z. & Xu,Y.-Q. Polarized photocurrent response in black phosphorus field-effecttransistors. Nanoscale 6, 8978–8983 (2014). → pages 4[54] Koenig, S. P., Doganov, R. A., Schmidt, H., Neto, A. C. & Oezyilmaz, B.Electric field effect in ultrathin black phosphorus. Applied Physics Letters104, 103106 (2014). → pages 4[55] Das, S., Demarteau, M. & Roelofs, A. Ambipolar phosphorene field effecttransistor. ACS nano 8, 11730–11738 (2014). → pages 4[56] Du, Y., Liu, H., Deng, Y. & Ye, P. D. Device perspective for blackphosphorus field-effect transistors: Contact resistance, ambipolar behavior,and scaling. ACS nano 8, 10035–10042 (2014). → pages 4[57] Kim, J. S., Jeon, P. J., Lee, J., Choi, K., Lee, H. S., Cho, Y., Lee, Y. T.,Hwang, D. K. & Im, S. Dual gate black phosphorus field effect transistorson glass for nor logic and organic light emitting diode switching. Nanoletters 15, 5778–5783 (2015). → pages 4138[58] Yin, D. & Yoon, Y. Design strategy of two-dimensional materialfield-effect transistors: Engineering the number of layers in phosphorenefets. Journal of Applied Physics 119, 214312 (2016). → pages 4[59] Xia, F., Wang, H. & Jia, Y. Rediscovering black phosphorus as ananisotropic layered material for optoelectronics and electronics. Naturecommunications 5 (2014). → pages 4, 9, 13, 24, 91[60] Chen, H., Liu, H., Zhang, Z., Hu, K. & Fang, X. Nanostructuredphotodetectors: From ultraviolet to terahertz. Advanced Materials 28,403–433 (2016). → pages 4[61] Doganov, R. A., OFarrell, E. C., Koenig, S. P., Yeo, Y., Ziletti, A.,Carvalho, A., Campbell, D. K., Coker, D. F., Watanabe, K., Taniguchi, T.et al. Transport properties of pristine few-layer black phosphorus by vander waals passivation in an inert atmosphere. Nature communications 6(2015). → pages 9[62] Luo, Z., Maassen, J., Deng, Y., Du, Y., Garrelts, R. P., Lundstrom, M. S.,Peide, D. Y. & Xu, X. Anisotropic in-plane thermal conductivity observedin few-layer black phosphorus. Nature communications 6 (2015). → pages11, 24, 91, 99[63] Favron, A., Gaufre`s, E., Fossard, F., Phaneuf-LHeureux, A.-L., Tang, N. Y.,Le´vesque, P. L., Loiseau, A., Leonelli, R., Francoeur, S. & Martel, R.Photooxidation and quantum confinement effects in exfoliated blackphosphorus. Nature materials 14, 826–832 (2015). → pages 13, 15, 82,125[64] Luo, W., Zemlyanov, D. Y., Milligan, C. A., Du, Y., Yang, L., Wu, Y. &Peide, D. Y. Surface chemistry of black phosphorus under a controlledoxidative environment. Nanotechnology 27, 434002 (2016). → pages 13[65] Wang, Y., Yang, B., Wan, B., Xi, X., Zeng, Z., Liu, E., Wu, G., Liu, Z. &Wang, W. Degradation of black phosphorus: a real-time 31p nmr study. 2DMaterials 3, 035025 (2016). → pages 13[66] Li, Q., Chen, J., Feng, Z., Feng, L., Yao, D. & Wang, S. The role of airadsorption in inverted ultrathin black phosphorus field-effect transistors.Nanoscale Research Letters 11, 521 (2016). → pages 13[67] Ryder, C. R., Wood, J. D., Wells, S. A., Yang, Y., Jariwala, D., Marks, T. J.,Schatz, G. C. & Hersam, M. C. Covalent functionalization and passivation139of exfoliated black phosphorus via aryl diazonium chemistry. Naturechemistry (2016). → pages 13[68] Wood, J. D., Wells, S. A., Jariwala, D., Chen, K.-S., Cho, E., Sangwan,V. K., Liu, X., Lauhon, L. J., Marks, T. J. & Hersam, M. C. Effectivepassivation of exfoliated black phosphorus transistors against ambientdegradation. Nano letters 14, 6964–6970 (2014). → pages 13[69] Brent, J. R., Savjani, N., Lewis, E. A., Haigh, S. J., Lewis, D. J. & O’Brien,P. Production of few-layer phosphorene by liquid exfoliation of blackphosphorus. Chemical Communications 50, 13338–13341 (2014). →pages 13, 15, 16[70] Hanlon, D., Backes, C., Doherty, E., Cucinotta, C. S., Berner, N. C.,Boland, C., Lee, K., Harvey, A., Lynch, P., Gholamvand, Z. et al. Liquidexfoliation of solvent-stabilized few-layer black phosphorus forapplications beyond electronics. Nature communications 6 (2015). →pages 13, 16[71] Seo, S., Lee, H. U., Lee, S. C., Kim, Y., Kim, H., Bang, J., Won, J., Kim,Y., Park, B. & Lee, J. Triangular black phosphorus atomic layers by liquidexfoliation. Scientific reports 6 (2016). → pages 13[72] Lee, G., Lee, J.-Y., Lee, G.-H. & Kim, J. Tuning the thickness of blackphosphorus via ion bombardment-free plasma etching for deviceperformance improvement. Journal of Materials Chemistry C (2016). →pages 13[73] Pei, J., Gai, X., Yang, J., Wang, X., Yu, Z., Choi, D.-Y., Luther-Davies, B.& Lu, Y. Producing air-stable monolayers of phosphorene and their defectengineering. Nature communications 7 (2016). → pages 13[74] Kang, J., Wood, J. D., Wells, S. A., Lee, J.-H., Liu, X., Chen, K.-S. &Hersam, M. C. Solvent exfoliation of electronic-grade, two-dimensionalblack phosphorus. ACS nano 9, 3596–3604 (2015). → pages 15, 16[75] Guo, Z., Zhang, H., Lu, S., Wang, Z., Tang, S., Shao, J., Sun, Z., Xie, H.,Wang, H., Yu, X.-F. et al. From black phosphorus to phosphorene: basicsolvent exfoliation, evolution of raman scattering, and applications toultrafast photonics. Advanced Functional Materials 25, 6996–7002 (2015).→ pages 15, 16140[76] Kang, J., Wells, S. A., Wood, J. D., Lee, J.-H., Liu, X., Ryder, C. R., Zhu,J., Guest, J. R., Husko, C. A. & Hersam, M. C. Stable aqueous dispersionsof optically and electronically active phosphorene. Proceedings of theNational Academy of Sciences 201602215 (2016). → pages 16[77] Wang, H., Yang, X., Shao, W., Chen, S., Xie, J., Zhang, X., Wang, J. &Xie, Y. Ultrathin black phosphorus nanosheets for efficient singlet oxygengeneration. Journal of the American Chemical Society 137, 11376–11382(2015). → pages 16[78] Kumar, V., Brent, J. R., Shorie, M., Kaur, H., Chadha, G., Thomas, A. G.,Lewis, E. A., Rooney, A. P., Nguyen, L., Zhong, X. L. et al.Nanostructured aptamer-functionalized black phosphorus sensing platformfor label-free detection of myoglobin, a cardiovascular disease biomarker.ACS Applied Materials & Interfaces 8, 22860–22868 (2016). → pages 16[79] Brent, J. R., Ganguli, A. K., Kumar, V., Lewis, D. J., McNaughter, P. D.,O’Brien, P., Sabherwal, P. & Tedstone, A. A. On the stability ofsurfactant-stabilised few-layer black phosphorus in aqueous media. RSCAdvances 6, 86955–86958 (2016). → pages 16[80] Zhao, W., Xue, Z., Wang, J., Jiang, J., Zhao, X. & Mu, T. Large-scale,highly efficient, and green liquid-exfoliation of black phosphorus in ionicliquids. ACS applied materials & interfaces 7, 27608–27612 (2015). →pages 16[81] Jia, J., Jang, S. K., Lai, S., Xu, J., Choi, Y. J., Park, J.-H. & Lee, S.Plasma-treated thickness-controlled two-dimensional black phosphorus andits electronic transport properties. ACS nano 9, 8729–8736 (2015). →pages 17[82] Bachhuber, F., von Appen, J., Dronskowski, R., Schmidt, P., Nilges, T.,Pfitzner, A. & Weihrich, R. The extended stability range of phosphorusallotropes. Angewandte Chemie International Edition 53, 11629–11633(2014). → pages 19[83] Sugai, S., Ueda, T. & Murase, K. Pressure dependence of the latticevibration in the orthorhombic and rhombohedral structures of blackphosphorus. Journal of the Physical Society of Japan 50, 3356–3361(1981). → pages 20141[84] Kawamura, H., Shirotani, I. & Tachikawa, K. Anomaloussuperconductivity in black phosphorus under high pressures. Solid statecommunications 49, 879–881 (1984). → pages 20[85] Rissi, E. N., Soignard, E., McKiernan, K. A., Benmore, C. J. & Yarger,J. L. Pressure-induced crystallization of amorphous red phosphorus. SolidState Communications 152, 390–394 (2012). → pages 20[86] Tuschel, D. Why are the raman spectra of crystalline and amorphous solidsdifferent? (2017). → pages 22[87] Raman, C. V. & Krishnan, K. S. A new type of secondary radiation. Nature121, 501–502 (1928). → pages 28[88] Venkateswarlu, K. Relative intensities of stokes and anti-stokes raman linesin crystals. Proceedings Mathematical Sciences 13, 64–67 (1941). →pages 28[89] Bowley, H. J., Gerrard, D. L., Louden, J. D., Turrell, G., Gardiner, D. J. &Graves, P. R. Practical raman spectroscopy (Springer Science & BusinessMedia, 2012). → pages 34[90] Lu, W., Ma, X., Fei, Z., Zhou, J., Zhang, Z., Jin, C. & Zhang, Z. Probingthe anisotropic behaviors of black phosphorus by transmission electronmicroscopy, angular-dependent raman spectra, and electronic transportmeasurements. Applied Physics Letters 107, 021906 (2015). → pages 35[91] Tuschel, D. Raman crystallography in theory and in practice. Spectroscopy27, 22–27 (2012). → pages 37, 121[92] Yoo, W. S., Harima, H. & Yoshimoto, M. Polarized raman signals from siwafers: Dependence of in-plane incident orientation of probing light. ECSJournal of Solid State Science and Technology 4, P356–P363 (2015). →pages 37, 121[93] Li, X.-L., Han, W.-P., Wu, J.-B., Qiao, X.-F., Zhang, J. & Tan, P.-H.Layer-number dependent optical properties of 2d materials and theirapplication for thickness determination. Advanced Functional Materials(2017). → pages 38[94] Li, X.-L., Qiao, X.-F., Han, W.-P., Lu, Y., Tan, Q.-H., Liu, X.-L. & Tan,P.-H. Layer number identification of intrinsic and defective multilayeredgraphenes up to 100 layers by the raman mode intensity from substrates.Nanoscale 7, 8135–8141 (2015). → pages 38142[95] Li, X.-L., Qiao, X.-F., Han, W.-P., Zhang, X., Tan, Q.-H., Chen, T. & Tan,P.-H. Determining layer number of two-dimensional flakes oftransition-metal dichalcogenides by the raman intensity from substrates.Nanotechnology 27, 145704 (2016). → pages 38[96] Lu, X., Utama, M. I. B., Zhang, J., Zhao, Y. & Xiong, Q. Layer-by-layerthinning of mos 2 by thermal annealing. Nanoscale 5, 8904–8908 (2013).→ pages 39[97] Su, L. & Zhang, Y. Temperature coefficients of phonon frequencies andthermal conductivity in thin black phosphorus layers. Applied PhysicsLetters 107, 071905 (2015). → pages 41, 53[98] Luo, W., Song, Q., Zhou, G., Tuschel, D. & Xia, G. Study of blackphosphorus using angle-resolved polarized raman spectroscopy with 442nm excitation. arXiv preprint arXiv:1610.03382 (2016). → pages 48, 66[99] Vierimaa, V., Krasheninnikov, A. V. & Komsa, H.-P. Phosphorene underelectron beam: from monolayer to one-dimensional chains. Nanoscale 8,7949–7957 (2016). → pages 54[100] Li, H., Wu, J., Huang, X., Lu, G., Yang, J., Lu, X., Xiong, Q. & Zhang, H.Rapid and reliable thickness identification of two-dimensional nanosheetsusing optical microscopy. ACS nano 7, 10344–10353 (2013). → pages 82[101] Li, H., Lu, G., Yin, Z., He, Q., Li, H., Zhang, Q. & Zhang, H. Opticalidentification of single-and few-layer mos2 sheets. Small 8, 682–686(2012). → pages 82[102] Schneider, C. A., Rasband, W. S. & Eliceiri, K. W. Nih image to imagej:25 years of image analysis. Nature methods 9, 671 (2012). → pages 82[103] Azuhata, T., Sota, T., Suzuki, K. & Nakamura, S. Polarized raman spectrain gan. Journal of Physics: Condensed Matter 7, L129 (1995). → pages 91[104] Duesberg, G., Loa, I., Burghard, M., Syassen, K. & Roth, S. Polarizedraman spectroscopy on isolated single-wall carbon nanotubes. Physicalreview letters 85, 5436 (2000). → pages 91[105] Huang, M., Yan, H., Chen, C., Song, D., Heinz, T. F. & Hone, J. Phononsoftening and crystallographic orientation of strained graphene studied byraman spectroscopy. Proceedings of the National Academy of Sciences106, 7304–7308 (2009). → pages 91143[106] Ribeiro-Soares, J., Almeida, R., Cancado, L., Dresselhaus, M. & Jorio, A.Group theory for structural analysis and lattice vibrations in phosphorenesystems. Physical Review B 91, 205421 (2015). → pages 93144

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0345629/manifest

Comment

Related Items