mechanochemical synthesis of carbon and carbon …
TRANSCRIPT
The Pennsylvania State University
The Graduate School
Department of Chemistry
MECHANOCHEMICAL SYNTHESIS OF CARBON AND CARBON NITRIDE
NANOTHREAD SINGLE CRYSTALS
A Dissertation in
Chemistry
by
Xiang Li
2018 Xiang Li
Submitted in Partial Fulfillment
of the Requirements
for the Degree of
Doctor of Philosophy
August 2018
The dissertation of Xiang Li was reviewed and approved* by the following:
John V. Badding
Professor of Chemistry, Physics and Materials Science and Engineering
Dissertation Advisor
Chair of Committee
Vincent H. Crespi
Professor of Physics, Materials Science and Engineering, and Chemistry
Paul S. Cremer
Professor of Chemistry and Biochemistry and Molecular Biology
Mauricio Terrones
Professor of Physics, Chemistry and Materials Science and Engineering
Thomas E. Mallouk
Professor of Chemistry, Biochemistry and Molecular Biology, Physics, and
Engineering Science and Mechanics
Head of the Department of Chemistry
*Signatures are on file in the Graduate School
iii
ABSTRACT
Carbon nanomaterials such as fullerenes, nanotubes, and graphene have been widely
studied in recent decades. Benefitting from their unique bonding, they possess extraordinary
physical and chemical properties. Compared with sp2 hybridized carbon allotropes, there are
significantly fewer new carbon materials dominated by sp3 bonding that have been developed.
Adamantane and graphane represent the smallest unit and thinnest sheet of diamond possible,
respectively. One-dimensional, mostly sp3 hybridized nanocarbon, did not yet exist in 2013, when
the first synthesis of carbon nanothreads finally filled up the last remaining entry in the matrix of
dimensionality and hybridization of carbon nanomaterials that year.
Carbon nanothread was first made by compressing benzene to ~25 GPa in a large-volume
anvil cell and slowly decompressing back to ambient pressure by an alumnus of the Badding
group. Background about high-pressure chemistry will be introduced in Chapter 1. An overview
of basic principles and the core instrumental techniques employed in this dissertation will be
provided in Chapter 2. In Chapter 3, I will present the progress of carbon nanothread synthesis
since 2013. Before my thesis work, only polycrystalline quality carbon nanothreads had been
made. With my optimized synthetic protocol, a single crystal carbon nanothread has been
successfully synthesized both in large-scale and in standard high-pressure apparatuses. High-
pressure x-ray diffraction illustrating the first direct in situ observation of nanothread formation
during compression will be presented in this chapter as well. The result of this experiment
demonstrates that the transformation from benzene to carbon nanothread is a unique non-
topochemical solid state reaction.
In Chapter 4, I will report the synthesis and structural characterization of the second
member in the nanothread family. Carbon nitride nanothread has been obtained by compressing
pyridine with the same slow compression/decompression method, suggesting that this
iv
mechanochemical synthetic approach is possibly quite general. The shift of the fluorescence
emission wavelength compared with carbon nanothread indicates that tuning the physical
properties of nanothreads can be realized by introducing heteroatoms or functional groups to the
benzene precursor.
A new high-pressure phase of pyridine has been discovered from the in situ diffraction
study of the carbon nitride nanothread reaction pathway. Preliminary analysis and provisional
crystal structures will be presented in Chapter 5.
Chapter 6 includes a concluding summary as well as an outlook, with a broader picture
than the insight sections at the end of Chapters 3, 4 and 5 provided.
v
TABLE OF CONTENTS
List of Figures .......................................................................................................................... vii
List of Tables ........................................................................................................................... xiv
Acknowledgements .................................................................................................................. xv
Chapter 1 Carbon Materials and High-Pressure Chemistry ..................................................... 1
1.1 Natural Carbon Allotropes: Graphite and Diamond .................................................. 2 1.1.1 Bonding Environments and Crystal Structures ............................................... 3 1.1.2 Structural Properties and Applications ............................................................ 5
1.2 Era of Synthetic Carbon Allotropes: Nanocarbons .................................................... 8 1.3 Molecular Crystals under High Pressure .................................................................... 11
1.3.1 Introduction of Pressure Effects ...................................................................... 11 1.3.2 Pressure Dependence of Chemical Equilibrium and Reaction Rates .............. 15 1.3.3 Pressure Effects on Electronic Structure ......................................................... 17
1.4 Goals of the Dissertation ............................................................................................ 20 References ........................................................................................................................ 22
Chapter 2 Instrumentation and Characterization Techniques .................................................. 26
2.1 High Pressure Cells .................................................................................................... 26 2.1.1 Diamond Anvil Cell ........................................................................................ 26 2.1.2 Paris-Edinburgh Cell ....................................................................................... 28
2.2 Pressure Control System ............................................................................................ 30 2.3 High Pressure X-ray Diffraction ................................................................................ 33
2.3.1 Single Crystal Crystallography under Pressure ............................................... 34 2.3.2 Determine the Pressure in Diamond Anvil Cell .............................................. 36 2.3.3 High Pressure Single Crystal Diffraction at Synchrotron Beamline ............... 38
2.4 Vibrational Spectroscopy ........................................................................................... 40 2.4.1 Principles of Infrared Spectroscopy ................................................................ 40 2.4.2 Principles of Raman Spectroscopy .................................................................. 42 2.4.3 Application of Polarized Raman Spectroscopy ............................................... 45 2.4.4 Ultra-low Frequency Raman Spectroscopy with Volume Bragg Gratings
as Optical Filter ................................................................................................ 47 References ........................................................................................................................ 52
Chapter 3 Mechanochemical Synthesis of Carbon Nanothread Single Crystals ..................... 55
3.1 Introduction ................................................................................................................ 55 3.2 Materials and Methods ............................................................................................... 58 3.3 Results and Discussion ............................................................................................... 60 3.4 Conclusions ................................................................................................................ 70 3.5 Insights and Future Consideration ............................................................................. 71
3.5.1 Sign of Reaction .............................................................................................. 71
vi
3.5.2 Polarized Raman Spectroscopy of Carbon Nanothread .................................. 73 3.5.3 Preliminary Ultra-Low Frequency Raman Study ............................................ 75
References ........................................................................................................................ 77
Chapter 4 Carbon Nitride Nanothread Crystals Derived from Pyridine .................................. 79
4.1 Introduction ................................................................................................................ 79 4.2 Materials and Methods ............................................................................................... 83
4.2.1 Synthesis ......................................................................................................... 83 4.2.2 X-ray Diffraction ............................................................................................. 84 4.2.3 Infrared Spectroscopy (IR) .............................................................................. 84 4.2.4 X-ray Photoelectron Spectroscopy (XPS) ....................................................... 85 4.2.5 Nuclear Magnetic Resonance (NMR) Spectroscopy ....................................... 85 4.2.6 Fluorescence Microscopy ................................................................................ 85 4.2.7 Photoluminescence Spectroscopy ................................................................... 86 4.2.8 Combustion Elemental Analysis ..................................................................... 86
4.3 Results and Discussion ............................................................................................... 87 4.4 Insights and Future Considerations ............................................................................ 98 References ........................................................................................................................ 101
Chapter 5 Structure Exploration of Solid Pyridine under Pressure .......................................... 103
5.1 Motivation and Background ....................................................................................... 103 5.2 Revisit of High Pressure Phase I ................................................................................ 107 5.3 Unknown High pressure phase II’ .............................................................................. 109
5.3.1 Powder X-ray diffraction ................................................................................ 110 5.3.2 Single Crystal X-ray Diffraction ..................................................................... 112
5.4 Insights for Reaction Mechanism and Synthesis Optimization.................................. 116 5.4.1 Compression Process ....................................................................................... 116 5.4.2 Decompression Process ................................................................................... 118
References ........................................................................................................................ 121
Chapter 6 Concluding Remarks and Future Outlook ............................................................... 122
References ........................................................................................................................ 126
vii
LIST OF FIGURES
Figure 1-1. Phase diagram of carbon.15 .................................................................................... 3
Figure 1-2. Demonstration of sp2 hybridization.17 ................................................................... 4
Figure 1-3. Scheme of sp3 hybrid orbitals formation.18 ........................................................... 4
Figure 1-4. Lattice of graphite and diamond. ........................................................................... 5
Figure 1-5. Expanding application field of single crystal diamond.33 ..................................... 7
Figure 1-6. Timeline of the inventions of new nanophase carbon materials.41 (Copyright
© 2010, Springer Nature) ................................................................................................. 8
Figure 1-7. Trends in carbon nanotube research and commercial production.61 (Copyright
© 2013, American Association for the Advancement of Science) .................................. 10
Figure 1-8. Lattice parameters of cubic C60 as a function of temperature (left)79 and
pressure (right).80 (Copyright © 2006, Taylor & Francis) .............................................. 13
Figure 1-9. Free energy variation of CsI crystal under pressure change (solid dots) and
temperature change (hollow circles).81 (Copyright © 1989, Annual Reviews Inc.) ........ 14
Figure 1-10. Schematic configuration of coordinate diagram showing the evolution of the
electronic structure at initial volume (left) and at reduced volume (right).86
(Copyright © 1972, National Academy of Sciences)....................................................... 14
Figure 1-11. Calculated HOMO-LUMO gap as a function of pressure of [email protected] .......... 15
Figure 1-12. Micrograph of ε-oxygen crystal in helium at 17.6 GPa.93 (Copyright © 2006,
Springer Nature) ............................................................................................................... 18
Figure 1-14. Matrix of carbon nanomaterial dimensionality and bonding geometry. ............. 20
Figure 2-1. Top: Assembling of symmetric diamond anvil cell.7 (Image courtesy of Dr.
Barbara Lavina at UNLV.) Bottom: configuration of opposed diamond anvils (left)
and indented gasket as sample chamber (right).8 ............................................................. 27
Figure 2-2. Left: actual photo of Paris-Edinburgh (PE) cell at SNAP beamline, SNS.
Right: schematic figure of the parts used for PE cell.15 (Copyright © 2012 American
Institute of Physics) .......................................................................................................... 29
Figure 2-3. Schematic pressure profile assumed for the sample and gasket assembly in PE
cell and ρ0 is the maximum pressure.17 (Copyright © 2010, American Physical
Society) ............................................................................................................................ 30
Figure 2-4. Left: the maximum principal stress (solid curve) and shear stress (dotted
curve) on the surface of double-toroidal PE cell anvil as a function of distance from
viii
anvil center.15 (Copyright © 2012, American Institute of Physics) Right: pressure
distributions in a beveled DAC.5 (Copyright © 2018, National Academy of
Sciences) .......................................................................................................................... 30
Figure 2-5. Left: decompression attachment of double-diaphragm design. Right: Dual
double-diaphragm setup for compression and decompression experiment.27
(Copyright © 2015, AIP Publishing LLC.) ...................................................................... 32
Figure 2-6. Ewald construction and reciprocal lattice.9 (Copyright © 2018
GeoScienceWorld) ........................................................................................................... 35
Figure 2-7. Diffraction geometry of transmission mode (left) and diffraction conditions in
reciprocal space (right). Dashed lines are the boundary circles which limit the area
for accessible reciprocal lattice vectors k.9 (Copyright © 2018 GeoScienceWorld) ........ 36
Figure 2-8. Scheme of energy levels demonstrate ruby fluorescence.34 .................................. 37
Figure 2-9. Opposite oscillated molecular dipole generated by oscillating electric field of
photon.41 (Copyright © 2011 Elsevier Inc.) ..................................................................... 41
Figure 2-10. Induced dipole moment of a homonuclear diatomic molecule originates from
the oscillating electric filed (represented by a charged plates capacitor on the left) of
the incident radiation.41 (Copyright © 2011 Elsevier Inc.) .............................................. 42
Figure 2-11. Rayleigh scattering and Stokes and anti-Stokes Raman scattering.41
(Copyright © 2011 Elsevier Inc.) ..................................................................................... 43
Figure 2-12. Raman tensor ellipsoids and different polarizations. The arrows indicate the
polarization direction of the incident beam and scattered lights.56(Copyright © 2006,
Springer Science Business Media, Inc.) ........................................................................... 46
Figure 2-13. Polarized Raman of isolated single wall carbon nanotube (SWNT). For αi=0˚
and 180˚, the polarization of the incident beam is parallel to tube axis.58 (Copyright
© 2000, American Physical Society) ............................................................................... 47
Figure 2-14. Schematics of reflecting volume Bragg grating (RBG). RBG only allows
selection of a narrow spectral region of light within an acceptance angle determined
by the Bragg conditions.61 (Copyright © 2012, SPIE) ..................................................... 48
Figure 2-15. Demonstration of a home-built Raman system equipped with three
BragGrateTM notch filters (BNF) as single monochromator.71 (Copyright © 2012,
Springer Nature) ............................................................................................................... 49
Figure 2-16. Left: Raman spectra for C peak which corresponds to the doubly degenerate
rigid layer shear mode and G peak of graphene.71(Copyright © 2012, Springer
Nature) Right: Raman spectra of ZnTe nanoribbons at various temperatures. The
sharp peaks are the multi-order LO modes.72(Copyright © 2016, Springer Nature) ....... 50
ix
Figure 3-1. Predicted and Observed Nanothread Crystal Diffraction Patterns. a,
Polytwistane crystal taken as representative of nanothread packing, viewed down the
hexagonal c-axis (threads are parallel to it) with a 6.5 Å spacing and (100) planes 5.6
Å apart, and a side view down the b-axis. b, X-ray diffraction (300μm CuKα1 beam)
for a nanothread crystal synthesized from polycrystalline, multiphase benzene shows
a hexagonal pattern that matches c, that predicted for the c-axis of the polytwistane
crystal. d, Diffraction after 90° rotation of the nanothread crystal, i.e. approximately
along the b-axis, matches the pattern predicted in e, Diffraction features marked with
an asterisk in d are from the polymer loop mount. ........................................................... 56
Figure 3-2. Polarization Analysis of Nanothread Samples from Paris-Edinburgh Press .
(Top) Mass of nanothreads synthesized from polycrystalline, multiphase benzene
between crossed polars. (Bottom) Nanothread fiber between cross polars. After
rotation by 45°, transmission increases greatly in the region that is thin and
appropriately oriented, demonstrating the presence of strong birefringence. Rotation
to 90° reestablishes extinction. ......................................................................................... 58
Figure 3-3. (Left) Neutron Diffraction Pattern of Benzene Phase I and Phase II Mixture at
2 GPa in a Paris-Edinburgh Press upon Increasing Pressure. (Right) Synchrotron X-
ray Diffraction Pattern of Benzene Phase I and Benzene Phase II Mixture at 14 GPa
upon Increasing Pressure in a Diamond Anvil Cell. The red curves in both patterns
are Pawley fit and the blue curves are the residual. ......................................................... 62
Figure 3-4. In situ Nanothread Diffraction at High Pressure and Deduced Spatial
Relationships a, Indexed diffraction pattern collected upon increasing pressure to 3.3
GPa for a phase-II single crystal. Large dark spots are diamond anvil reflections. b,
Expanded view of diffraction pattern upon increasing pressure to 23 GPa for phase-
II single crystal showing psuedohexagonal carbon nanothread {100} peaks
beginning to appear. The inset shows the benzene (100) peak neighboring the
nanothread (100) peak. (010) is >30 counts above background, though difficult to
see in the figure. c, a, b, and b-c columns of benzene molecules along which reaction
to form nanothreads is considered; unit cell in blue. d, Stereographic projection of
benzene phase-II diffraction planes (hkl) and zone axes [uvw], showing the angular
relationship at 23 GPa of the nanothread crystal [001] c-axis to the benzene crystal
[100] a, [010] b, and [011] bc axes. Only the directions and planes on the “north”
side of the projection are shown (e.g., [100] is on the south side, 180 degrees from
[1 ̅00]). e, The monoclinic benzene crystal viewed down the unique b-axis, showing
diffraction planes and their interplanar spacings at 23 GPa with the orientation and
spacing of diffraction spots in the inset. Diffraction planes for the grey spots are not
shown. f, A monoclinic (pseudohexagonal, γ=117°) nanothread crystal viewed down
the c axis with interplanar spacings at 23 GPa. Diffraction planes have the same
scale as d. Note the slight expansion required for the benzene (100) planes to form
nanothread (100) planes and the larger shifts required for planes at angles to (100). ...... 65
Figure 3-5. Monoclinic Phase II Benzene Crystal Viewed Down The [011] b-c Axis.
Some diffraction planes and their interplanar spacings at 23 GPa are shown, with the
orientation and spacing of the corresponding diffraction spots in the inset.
Diffraction planes for the grey spots are not shown. The smallest amount of
expansion is needed for the (100) planes to form the psuedohexagonal or hexagonal
x
nanothread structure (Figure 3-7d). Other sets of planes such as the (11-1)and (01-1)
must expand to a much larger extent. The b-c stacks have the columns with
alternating orientations of benzene molecules different from the benzene columns
along the b axis, which have uniform orientation. Both types of stacks are shown in
Figure 3-7b. In contrast to the b stacks, the b-c stacks are not suited for [4+2]
cycloaddition, but “para polymerization” starting with a diradical may be possible. 22 .. 67
Figure 3-6. Diffraction Patterns of Nanothreads Synthesized from Single Crystals upon
Decreasing Pressure Along [001] Zone Axis. (a) Five of six nanothread reflections
can be observed in a diffraction pattern at 23 GPa, along with the benzene (100)
neighboring the nanothread (100) and the benzene (-100) neighboring the
nanothread (-100). (b) Diffraction pattern at 19 GPa after reducing the pressure from
23 GPa. All six nanothread reflections can be observed and are somewhat more
intense. The relative lower intensity observed for the two rightmost nanothread
reflections is likely associated with a tilt away from the [001] zone axis, as these
diffraction images were collected at the maximum ω angle of 15°. The increase of
the reflection intensities as the pressure is reduced is likely associated either with
more nanothread formation or better ordering of threads that have already formed.
(c) Evolution of Benzene and Nanothread Diffraction Patterns with Pressure. One-
dimensional diffraction patterns vs. pressure for the nanothread sample synthesized
from pure single-crystal phase II benzene as a function of pressure. Down-arrows
indicate decompression. ................................................................................................... 69
Figure 3-7. Constraining Nanothread Structure with Experimental and Calculated {100}
Interplananar Spacings. a, Interplanar spacings calculated for candidate achiral and
chiral nanothread structures at T=0, including both ordered and disordered
axial/azimuthal arrangements at empirical and first-principles levels, to show which
ones are consistent with the measured interplanar spacings in the nanothreads
synthesized from mixed phase I/Phase II benzene. The anticipated shifts in the
experimental data due to thermal contraction to T=0 are depicted by orange bars,
and the full-width-at-half-maximum over 10 instances of calculated disorder are
marked by vertical blue bars. Experimental nanothread {100} interplanar spacings at
ambient pressure are from sample synthesized in a diamond cell from
polycrystalline, multiphase benzene. b, Nanothread (100) interplanar spacings
during decompression. The threads again formed from a polycrystalline mixture of
phase I and phase II in a diamond cell. Calculated first principles {100} interplanar
spacings derived from lattice parameters are also shown for two representative
nanothread crystals. .......................................................................................................... 69
Figure 3-8. In situ X-ray diffraction and Raman spectra of benzene in diamond anvil cell. ... 72
Figure 3-9. Raman spectra of benzene nanothread measured under different polarization. .... 74
Figure 3-10. Ultra-low frequency Raman spectrum of benzene nanothread synthesized in
PE cell. ............................................................................................................................. 76
Figure 4-1. Carbon Nitride Nanothread Structures (a) View down the hexagonal c axis
and (b) view perpendicular to the c axis of a degree-6 carbon nitride nanothread tube
xi
(3,0)_123456. (c) Degree-4 carbon nitride nanothread structure (IV-7) with C-N
double bonds in red. ......................................................................................................... 81
Figure 4-2. (a) Few examples of carbon nitride nanothread structures from theoretical
enumeration and used for diffraction pattern comparison. (b) Ring numbering for
four types of benzene nanothreads. The orange-numbered ring denotes the first ring;
the blue-numbered ring below is the second ring and so on. The third ring in tube
(3,0) is the same as the first ring because of symmetry and is numbered in orange.
The nomenclature tube (3,0)_1245 means this pyridine nanothread is based on the
structure of the tube (3,0) benzene nanothread, substituting N atoms for the C-H
bonds at position 1 in 1st ring, position 2 of 2nd ring, position 4 of the 3rd ring and
position 5 of the 4th ring; the repeat unit contains four rings.15 ....................................... 82
Figure 4-3. Diffraction experiment and modeling. (a) Synchrotron diffraction pattern
(left) collected down the hexagonal c axis of the nanothread crystal showing its six-
fold symmetry. Simulated diffraction pattern (right) of carbon nitride
tube(3,0)_123456. (b) Experimental and simulated (100) interplanar spacings for
different carbon nitride nanothread structures. ................................................................ 88
Figure 4-4. Experimental infrared spectrum. The major radial breathing mode (top left)
and the inter-ring C-N stretch mode (bottom right) in tube (3,0)_123456 are shown.
Peaks with asterisks may be associated with pyridinic derivatives either as
amorphous carbon or as substituted pyridine linking nanothreads. The intensities of
these peaks are significantly lower than in the IR spectra of amorphous recovered
samples.17-18 ...................................................................................................................... 90
Figure 4-5. High Resolution XPS Scan of Carbon Nitride Nanothreads. C1s (left) and
N1s (right) XPS spectra of the raw surface of the recovered sample are presented as
experimental data (open circles) and fit curves (solid lines). All spectra were
obtained after exposure to air. .......................................................................................... 93
Figure 4-6. NMR Spectra. Solid-state (a) 15N and (b) 13C NMR spectra of nanothreads
made from 15N-enriched pyridine. Peak areas are nearly quantitative. Selective
spectra of NH and of nonprotonated C are shown by colored thin lines in (a) and (b),
respectively. The original signal positions of pyridine are marked by dashed vertical
blue lines. The natural-abundance 13C spectrum shows the characteristic chemical-
shift increases due to N-bonding of nearly half of all C. ................................................. 95
Figure 4-7. Confocal fluorescence microscopy images of (a) blue, green, and red
emission from carbon nanothreads and (b) blue, green and red emission from carbon
nitride nanothreads. Details of excitation and emission are pro-vided in the
supporting information. .................................................................................................... 96
Figure 4-8. Photoluminescence Spectra of Carbon Nanothread and Carbon Nitride
Nanothread. Three lasers with different wavelength were employed as excitation
sources. Both samples were measured under the same conditions for experiment at
each wavelength. Carbon nanothreads have stronger photoluminescence than carbon
nitride nanothreads when excited by 364 nm laser. On the other hand, when excited
by longer wavelength lasers (488 nm and 514 nm), carbon nitride nanothreads show
xii
more intense photoluminescence than carbon nanothreads. The results are in accord
with the observations in Figure 4-7a that blue light (excited by 405 nm) is the major
fluorescence emission of carbon nanothreads while red light emission (excited by
546 nm) is very weak. The green light emission (excited by 488 nm) of carbon
nitride nanothread is much stronger than its blue light emission, while red light
emission is very weak. ..................................................................................................... 98
Figure 4-9. Left: Experimental variant of optical gap vs. sp2 fraction for a-C:H, ta-C:H
and ta-C.48 (Copyright © 2002 Elsevier Science B.V.) Right: Tauc optical gap of ta-
CN and a-CN films as a function of the N/C ratio.49 (Copyright © 2003 Elsevier
Science B.V.) ................................................................................................................... 99
Figure 4-10. Tauc plots of same carbon nanothread sample derived from absorption
spectrum (red) and diffuse reflectance spectrum (green). ................................................ 100
Figure 5-1. Isothermal growth of pyridine single crystal phase I. (a) liquid-solid
equilibrium at 1.00 GPa/295 K; (b-i) crystal growth following sample chamber
volume decrease; (j) one single crystal fully filled the chamber at 1.23 GPa/295 K.9
(Copyright © 2010, Royal Society of Chemistry) ........................................................... 105
Figure 5-2. Failed isochoric growth of pyridine single crystal phase II. (a,b)
polycrystalline phase II at 1.20 GPa/295 K; (c) one crystal seed at 350 K; (d-i) single
crystal growth during cooling; (i) phase II single crystal before phase transition at
310 K; (j) phase I powder obtained at 0.9 GPa/295 K due to the destructive II/I
phase transition.9 (Copyright © 2010, Royal Society of Chemistry) ............................... 105
Figure 5-3. Schematic demonstration of phase transitions of pyridine under pressure.12 ........ 106
Figure 5-4. Le Bail fit of power diffraction pattern collected at 1.13 GPa. Blue crosses are
the observed data points and fitting line is in green. Pyridine phase I crystal structure
solved at 1.23 GPa9 was employed in the fitting. ............................................................. 108
Figure 5-5. Raman spectra of the lattice phonon region measured in this dissertation work
(Left) and by Fanetti et al. (Right). The 1.0 GPa and 1.7 GPa curves are considered
to be the characteristic spectra of phase I and phase II respectively. ............................... 108
Figure 5-6. Raman spectra of two pyridine crystals formed from direct compression of
fluid in different experiments. .......................................................................................... 110
Figure 5-7. Integrated one-dimensional powder diffraction pattern collected at 1.23 GPa
in this work (black curve) compares with reported two phases at similar pressure.
The unfitted peaks are labeled with asterisks. .................................................................. 111
Figure 5-8. Le Bail fit of power diffraction pattern collected at 2.4 GPa. Blue crosses are
the observed data points and fitting line is in green. The tetragonal unit cell obtained
from autoindexing (a=5.3100 Å, c=13.3169 Å) was employed in the fitting. The
inset figure is the two dimensional pattern illustrating the powder quality, .................... 112
xiii
Figure 5-9. Omega scan of pyridine single crystal obtained at 3.0 GPa. The black boxes
are predicted reflections from the tetragonal structure (a=5.2852 Å, c= 13.2495 Å)
indexed by this dataset. .................................................................................................... 113
Figure 5-10. Comparison of experimental powder diffraction measured at 2.4 GPa with
three candidate structures predicted by DFTB method. ................................................... 114
Figure 5-11. Rietveld refinement of pyridine powder diffraction collected at 2.4 GPa by
tetragonal unit cell predicted from DFTB. ....................................................................... 115
Figure 5-12. In situ pyridine powder diffraction patterns collected during compression.
The dotted arrow is the guidance of the new peak appearance and growth located at
2θ=8˚. ............................................................................................................................... 117
Figure 5-13. Two-dimensional patterns of pyridine powder during compression. Pyridine
solid became less powdery and with stronger preferred orientations after pressure
beyond 8 GPa. .................................................................................................................. 117
Figure 5-14. Two-dimensional patterns of pyridine single crystal during compression.
Circles are used for guiding the observation of new reflections. ..................................... 118
Figure 5-15. Two-dimensional patterns of pyridine single crystal during decompression.
Carbon nitride nanothread (100) reflection arcs are pointed out by green arrows.
Pyridine crystal melted when pressure released to ~0.64 GPa thus only nanothread
reelections remained. ....................................................................................................... 120
Figure 6-1. Theoretical modelling of two-dimensional stacked 3,4-connected carbon nets
from plane nets.8 (Copyright © 1987, American Chemical Society) ............................... 124
Figure 6-2. Future prospect on nanothread functionalization and application. ........................ 125
xiv
LIST OF TABLES
Table 1-1. Examples of free energy increase per unit cell in molecular crystal system
under pressure .................................................................................................................. 13
Table 1-2. Activation volume values of various reactions91 .................................................... 17
Table 4-1. Comparison of IR Peak Positions in Experiment and Simulation .......................... 92
Table 5-1. Decompression rates applied in diamond anvil cell synthesis ................................ 119
xv
ACKNOWLEDGEMENTS
I would like to first thank my adviser Prof. John Badding for his great support, guidance
and encouragement during my graduate program. John provided me with great freedom and trust
to plan and conduct the experiments and research that I am interested in. Besides the advising in
science, John is very supportive in setting up instrumentations especially unconventional ones.
This thesis work was supported as part of the Energy Frontier Research in Extreme Environments
(EFree) Center, an Energy Frontier Research Center funded by the U.S. Department of Energy,
Office of Science.
This thesis work requires significant amount of efforts from theory and modeling. I am
very appreciate that I can collaborate with two excellent theory teams. Prof. Vincent Crespi from
department of physics in Penn State and his graduate students Tao Wang and Enshi Xu are
experts in density function theory calculations and the pioneers in nanothread structure
enumeration and modeling. One of the most memorable scenes in graduate school is that Tao and
I sitting side by side by our laptops and working on papers from afternoons to the next mornings.
I feel truly blessed to have such a nice collaborator and also as a friend. I am also very grateful to
get support from Prof. Roald Hoffmann and his postdoc Dr. Bo Chen from Cornell University.
Their precious inspirations and insights in chemical reaction mechanism greatly promoted the
progress in reaction pathway study.
A special thanks goes to Dr. Maria Baldini who is the X-ray coordinator of EFree. She is
the mentor of me for conducting high pressure experiment and synchrotron X-ray diffraction.
During my graduate career, especially year 3 and 4, I have spent enormous amount of time with
Maria in Argonne National labs. Her patience and approachable character make this experience
much enjoyable.
xvi
A recent collaboration with Prof. Klaus Schmidt-Rohr and his graduate student Pu Duan
from Brandeis University brought in essential insights in chemical bonding of nanothread
according to their advanced NMR technique. I would like to thank Prof. Schmidt-Rohr for his
patient discussion with me via emails.
I would also acknowledge Dr. Malcolm Guthrie who is a neutron expert and
crystallographer. Even though he is far away in Sweden, he is always willing to provide
informative suggestions and share his own practical experience through emails and skype meeting
whenever I encountered with difficulties.
Due to the special requirement of experimental conditions, I need to travel to national
labs quite often. I am so grateful that I obtained so much help and support from the beamline
scientists and staffs. I would like to dedicate my gratefulness to Chris Tulk, Jamie Molaison and
Antonio Moreira dos Santos from SNAP (SNS), Stanislav Sinogeikin, Rich Ferry, Curtis Kenney-
Benson, Ross Hrubiak, Jesse Smith and Changyong Park from HPCAT (APS), A.J.Ramirez-
Cuesta and Luke Daemen from VISION (SNS), Takanori Hattori and Asami Sano from PLANET
(J-PARC), Yu-Sheng Chen from ChemMatCARS (APS), Ayman Said from HERIX (APS),
Zhenxian Liu (BNL) and also Chen Li (neutron coordinator of Efree). I would like to sincerely
thank Dr. Tim Strobel from Geophysical Laboratory of the Carnegie Institution of science for his
generosity of providing us the gas controllers. Hemant P Yennawar and Nicole Wondering also
helped me a lot in X-ray diffraction at Penn State.
I would like to thank my amiable lab mates in John’s group. I am appreciate that Tom
Fitzgibbons start the work of nanothread so that I can have the chance to take over this interesting
project. A lot of thank to Paramita Ray, Steve Juhl, Haw-tying Huang and Arani Biswas being so
helpful and teach me lot of knowledge and techniques. Those long-hour discussions with them
are very impressive. I would also like to thank Yan, Mike Coco, Alex, Yunzhi, Brianna, Sub and
xvii
Steve Aro for being such supportive and friendly lab mates. Thanks for our group assistant staff,
Jessica, who always work with patience and high efficiency to help me scheduling my trips.
Many thanks to my friends in State College, they make my graduate school life joyful.
Last but not the least, I will appreciate my parents and grandparents for their unconditional love
and support.
1
Chapter 1
Carbon Materials and High-Pressure Chemistry
The element carbon is detected in abundance in the universe, in the sun, stars, comets,
and in the atmosphere of the planets. It is the fourth most abundant element (with mass fraction
~4600 ppm) in the solar system, after hydrogen, helium, and oxygen.1 Even though the earth is
missing a great deal of carbon which has been lost as volatile hydrocarbons, the element carbon is
widely distributed in nature in the ratio of 180 ppm in the earth’s crust.2 Many of these natural
compounds are essential to the production of synthetic carbon materials and include various coals
(bituminous and anthracite), hydrocarbons complexes (petroleum, tar, and asphalt) and the
gaseous hydrocarbons (methane and others).3 Only two polymorphs of carbon are found on earth
as minerals: graphite and diamond. Their crystal structure, chemical bonding and application
which will be briefly introduced in the first section of this chapter. Carbon owns the fascinating
ability to form many allotropes by binding to itself and nearly all elements in almost limitless
varieties due to its valency. Carbon allotropes-based nanomaterials, such as fullerenes, carbon
nanotubes and graphene have been invented in recent decades after the inception of
nanotechnology. Their unique and superior chemical and physical properties attract enormous
research interests in developing new carbon materials with enhanced performance. On the other
hand, great efforts have been taken in scaling up the production for commercial applications such
as rechargeable batteries, automotive parts, printed electronics and water filters.
This dissertation will focus on a novel carbon nanomaterial, carbon nanothread, which is
synthesized under high pressure from simple small organic molecules.4 This new material
promises to have superlative physical and chemical properties.5-7 Detailed discussions of
2
nanothread synthesis, characterization and reaction mechanism studies will be presented in
subsequent chapters. Chapter 1 is the introduction of carbon allotropes in both bulk and
nanoscale. Background and recent advances of high pressure chemistry will be introduced in this
chapter as well.
1.1 Natural Carbon Allotropes: Graphite and Diamond
Graphite occurs naturally in metamorphic rocks as a result of the reduction of
sedimentary carbon compounds during metamorphism. It also occurs in igneous rocks and in
meteoritesis.8 World production of natural graphite in 2016 was 1,200,000 tons.9 Unlike graphite,
diamonds are extremely rare, with concentrations of at most parts per billion in source rock. Most
diamond were formed at depths of 150 to 250 kilometers in the Earth's mantle, although a few
have come from as deep as 800 kilometers.10 At room temperature and pressure, graphite is the
most stable form while diamond is considered to be metastable since it’s kinetically stable, not
thermodynamically stable.11
The transformation from graphite to diamond is not easy due to the very large activation
barrier. High temperature and pressure is required as shown in the carbon phase diagram. Even
though diamonds can be artificially synthesized by high pressure, high temperature method
(HPHT) which approximately simulates the conditions in the Earth's mantle or by chemical vapor
deposition (CVD) technique, the transition mechanism is not as straight forward as it’s shown in
the phase diagram (Figure 1-1). It undergoes a typical diffusionless transition,12 which governed
by nucleation–growth mechanism under low pressures (<20 GPa).13-14 The rapid and local atomic
restructuring happened when reaction conditions reached makes it very difficult to characterize
the morphology and interfacial structure of nucleation core. Herein, continuous studies in both
theory and experiments have been conducted.
3
Figure 1-1. Phase diagram of carbon.15
1.1.1 Bonding Environments and Crystal Structures
Graphite and diamond consist of extended networks of sp2- and sp3 -hybridized carbon
atoms, respectively. The sp2-hybridization is the combination of one s-orbitals with only two p
orbitals, namely px and py (Figure 1-2). They contribute together to a planar assembly with a
characteristic angle of 120° between hybrid orbitals forming a σ-bond. The additional pz-orbital is
perpendicular to the sp2-hybrid orbitals and forms a π-bond.16 In the case of sp3-hybridization
(Figure 1-3), the 2s orbital is mixed with all three of the 2p orbitals, creating four hybridized sp3
orbitals. Each of these has 25% s and 75% p character. A tetrahedral shape is favored due to
electron repulsion, therefore the orbitals are 109.5° apart from each other.
4
Figure 1-2. Demonstration of sp2 hybridization.17
Figure 1-3. Scheme of sp3 hybrid orbitals formation.18
Graphite has a layered, planar structure. The carbon atoms are arranged in a honeycomb
lattice in each layer with separation of 0.142 nm, and the inter-layer distance is 0.335 nm.19
Graphite crystalizes in hexagonal lattice with space group of P63mc.20 Although hexagonal
diamonds have been found in meteorites and synthesized by shock wave compression of graphite,
cubic diamond is more common form. Its cubic structure is in the Fd-3m space group which
5
follows the face-centered cubic Bravais lattice.21 Crystal structures of graphite and cubic diamond
are shown in Figure 1-4. The differences of the C-C interatomic distances also indicate the
different bonding types and resonance in the two carbon allotropes. Their characteristic bond
lengths, 1.54 Å for diamond (single bond) and 1.42 Å for graphite (one-third double bond) are
used for constructing empirical curve of the dependence of the C-C distance and relative degree
of single-bond-double-bond character by Pauling in 1930’s.22
Figure 1-4. Lattice of graphite and diamond.
1.1.2 Structural Properties and Applications
“Structure Determines Properties” is a powerful concept has long guided the discovery
and optimization of novel materials. Both graphite and diamond consist of carbon atoms only, but
diamond is renowned as the hardest bulk material while graphite is considered as a soft material.
As shown in Figure 1-4, graphite has highly anisotropic structure. Carbon atoms form tight
covalent bonds within the same layer. The bonding between the layers are governed by weak van
der Waals bonds, making graphite sheets easy to be separated, or to slide past each other.
Graphite can be ultimately exfoliated into single layer, which later known as graphene. Diamond
6
has a different story that all the carbon atoms bond into an inflexible three-dimensional lattice
leading to its superlative mechanical properties.
Pure diamond is transparent and colorless while pristine graphite is opaque and black.
The reason that causes the different appearances is closely related to the band structure which is
also associated with their chemical bonding and crystal structures. The electrons in diamond are
highly localized in between adjacent carbon atoms. In addition, only σ→σ* electrical transition
exists in this material because all the carbon atoms are sp3-hybridized. The tight-binding fit to the
band structure of diamond by Painter, et al.23 gives the pz-s interaction to be 5.45 eV. The
absolute values of s-s, and s-p interactions are 4.55 and 5.2 eV, respectively. These results
suggest the C-C bond strength to be on the order of 5 e V. Since visible light has energies of
between 1.65 and 3.1 eV, only UV photons have enough energy to be absorbed by pure diamond
that therefore caused the transparency. Owing to the large bandgap, diamond is an excellent
electrical insulator. In the case of graphite, each carbon atom contributes one electron to a
delocalized system. The delocalized electrons are free to move throughout the plane acting as a
“sea of electrons” like those of a metal.24-25 As a result, graphite is able to absorb all wavelength
of light which makes it black. Further, with its metallic nature, the optical constants of graphite
cause it to be highly reflective to a large range of wavelengths.26-28
Although diamond is a poor electrical conductor, natural single-crystal diamond is known
to feature the highest thermal conductivity of all the bulk materials studied thus far, as high as
2200 W/mK at room temperature.29 The numbers for aluminum and copper are about 250 W/mK
and 400 W/mK, respectively. Similar to its mechanical properties, thermal properties of graphite
are highly anisotropic, since phonons propagate quickly along the tightly-bonded planes, travel
much slower from one plane to another. Herein the thermal conductivity of highly crystalline
graphite in the basal plane is comparable to diamond (1900 W/mK), but the thermal conductivity
through the thickness of the sheet, along c axis, is 4 orders of magnitude smaller.30 Benefit from
7
this anisotropy, graphite can function as both a heat spreader and an insulator in electronic
components to eliminate localized hot spots.31
As described above, diamond possesses superior mechanical, thermal and optical
properties. After being treasured as gemstones since 6000 years ago,32 diamond is now employed
in various fields both in industry and applied science with proper preparation and processing.
Figure 1-5 summarizes some of the major applications.
Figure 1-5. Expanding application field of single crystal diamond.33
There are much less review articles focus on applications of graphite than of graphene,
since the latter is known as “wonder material”. However, both natural and synthetic graphite, has
been playing indispensable roles in everyday life and frontier research. Natural graphite is widely
used in steel making,34 lubricant,35 lithium-ion battery36-37 etc. Synthetic graphite, such as highly
oriented pyrolytic graphite (HOPG), is used as monochromator in x-ray optics and Gilsocarbon,
special grades of synthetic graphite is used as neutron moderator within nuclear reactors.38
8
1.2 Era of Synthetic Carbon Allotropes: Nanocarbons
Under the concept that carbon allotropes are possibly constructed by altering the periodic
binding motif in multiple-dimensional networks and the hybridizations of carbon atoms, new
target systems can be designed by controlled combination of structural and functional building
blocks through tailored synthetic chemistry. 39-40 Fullerene, carbon nanotube and graphene are the
representative materials that have been created and rapidly developed in this new era of synthetic
carbon allotropes (Figure 1-6).
Figure 1-6. Timeline of the inventions of new nanophase carbon materials.41 (Copyright © 2010,
Springer Nature)
In 1985, Smalley, Curl and Kroto et al, discovered a new type of carbon allotrope, that
later inspired intense research on itself and derivatives due to their fascinating chemical
properties and technological applications. At first, Kroto studied the carbon chains near various
red giant stars. After collaborating with Smalley and Curl to recreate the conditions in lab by laser
vaporization of graphite within gas medium, they found this unexpected product who has exactly
60 carbon atoms.42-45 They proposed that this new molecule owning sphere-like structure and
9
therefore name it as buckminsterfullerene. After 5 years, large scale synthesis was developed by
Krätschmer et al,46 and one year after the soccer ball framework was confirmed by Hawkins et. al,
via X- ray diffraction.47
In the same year, another remarkable discovery was made by Sumio Iijima from the
materials grown on the negative end of the carbon electrode used in the d.c. arc-discharge
evaporation of carbon in an argon vessel. He observed thin graphitic carbon needles ranging from
3~40 nm in diameter under electron microscope and named them as carbon nanotube (CNT).48 In
1993, two years after the first discovery of multiwall CNT, he reported the synthesis of single
wall CNT together with Bethune.49-50 By now, CNT is categorized in the fourth allotrope of
carbon, following diamond, graphite and fullerenes.
Graphene (GR), a two-dimensional material that representing a single layer of graphene
sheet is a much younger synthetic carbon allotrope. It has long been considered as an exclusively
theoretical material as it is the ultimate example of expanded aromatic carbon.41 In 2004, Geim
and Novoselov managed to extract pure single atomic thick crystal lattice from bulk graphite by
the famous scotch tape method.51 This finding opened up a new branch of quantum physics as
there were no actual 2D system for the physical manifestation of quantum mechanics.52
All these three revolutionary new synthetic carbon allotropes not only possess
aesthetically pleasing architectures but also their outstanding materials properties make them
applicable in varieties of fields. Fullerene has compelling applications in medicinal chemistry.
For example C60 and their derivatives have potential antiviral activity, which has strong
implications on the treatment of HIV-infection.53 They can also be used as photosensitizers, such
as for tumor uptake, in photodynamic therapy.54-55 Individual carbon nanotubes have appealing
mechanical properties with tensile strength of 100 GPa has been measured experimentally.56
Individual single wall carbon nanotubes show superior thermal conductivity (3500 W/mK)57 than
diamond (2200 W/mK) which has long been believed to be the best thermal conductor as
10
aforementioned. Most of the commercial applications of CNT nowadays refer to composite
materials or thin films without further fabricating the tubes in certain order. Relative products are
playing actively roles in coating/painting of fuel lines and ship hull,58 strong while lightweight
wind turbine blades59 and sporting goods.60 Highly ordered CNTs possess other functionalities,
such as high damping, terahertz polarization, large-stroke actuation, near-ideal black-body
absorption.61 Great number of researches have been done in the CNT applications in
microelectronics and energy storage. Graphene is a well-known gapless semiconductor62-63
because its band structure consist two groups of conical valleys and they touch at two points
which are known as Dirac points.63 Owning to this merit, graphene shows remarkable charge
carrier mobility of 2000-5000 cm-2/Vs64 and suspended graphene solutions at low temperature
exhibit charge carrier mobilities in excess of 200000 cm-2/Vs.65-66 Such high electron transport
properties render graphene exceptional performance in high-speed electronic applications such as
graphene transistors.67 Graphene also offers a tremendously high optical transparency of up to
97.7%68 that would great benefit for transparent electrode in solar cell application69 and
holographic data storage.70
Figure 1-7. Trends in carbon nanotube research and commercial production.61 (Copyright © 2013,
American Association for the Advancement of Science)
11
Fullerenes, carbon nanotubes and graphene all have realized large scale synthesis from
bench work in lab research to industry production line in factory (Figure 1-7). In 2003, Frontier
Carbon Corporation started fullerene production of 40 tons per year.71 In 2009, the confirmed
yearly production capability of carbon nanotube reached 1000 tons.61 Between 2009 and 2013,
global production of graphene rose from 12 to 205 tons,72 and the number jumped to 1000 tons in
2015.73 Having new synthetic nanocarbon allotropes hitting the market is an auspicious sign of a
new generation of devices and materials that will change our life.
1.3 Molecular Crystals under High Pressure
1.3.1 Introduction of Pressure Effects
Temperature, pressure and chemical potential are the thermodynamic variables that
control the physical transformations and chemical reactions. Tuning the temperature is a more
accustomed method for chemists to tune the reactivity than applying pressure. High pressure is
usually referred to the ranges of thousands (kilobars) or millions (megabars) of times atmospheric
pressure which requires specific equipment. Percy W. Bridgman, one of the pioneers of this field
won the 1946’s Nobel Prize in Physics for his work started from early 20th.
Temperature can affect a system in two ways. Firstly, internal energy of the system will
be changed with the excess energy redistributed among the internal vibrations such as
translational and rotational motions of the molecules which are corresponded to acoustic and
librational phonons of crystalline systems.74 Secondly, temperature will cause volume change of
the system by the effect of the anharmonicity of the interaction potential. Lowering the
temperature and increase the pressure will both result in shortening the interatomic and inter-
molecular distances.
12
In molecular crystals, which this thesis focus on, the attractive part of the interaction
potential mainly arises from London dispersion forces while the repulsive part generally derives
from electron overlap and electron-electron correlation. The form of Lennard-Jones potential75-78
summarizes both parts as in Equation 1-1, where inverse of 6th power of intermolecular distance
term represents the attractive part and the inverse of 12th power term represents the repulsive part.
6 124 [ ( ) ( ) ]r r
E
(Equation 1-1)
Although temperature and pressure both are primary parameters control the volume
dilatation or compression of a system, they have significant differences in their effects. For
example, Figure 1-8 (left) plots the unit cell parameter change of C60 fullerene crystal (with cubic
structure) as a function of temperature in the range from 0K to 360K.79 Because C60 crystal and
most other molecular crystals are built by rigid molecular units which holding by weak van der
Waals interactions , they are expected to be very “soft”. However, the variation is only 3~4% in
terms of volume within this temperature interval for such molecular crystals (considering melting
or decomposition). On the other hand, the unit cell length a, easily shrunk below 14 Å under
compression with little pressure applied (~2 GPa) at room T and finally reached to 13 Å before
turning to amorphous phase (Figure 1-8, right).80 Thus, changing pressure is usually a more
efficient way to induce the volume change comparing to temperature. The CsI crystal, even
though not a molecular crystal, is a good example to demonstrate this statement in a more
comprehensive way.81 As shown in Figure 1-9, tuning the pressure enabled the change of free
energy surface in much more extended region. Table 1-1 summarized few examples of energy
increase per unit cell of some molecular crystals.82-85 Most bond dissociation energy at 298K is
below 10 eV/bond and therefore the increased amount of free energy as listed in the table would
bring out reorganization of electronic distribution and/or even potential chemical reaction under
proper pressure.
13
Figure 1-8. Lattice parameters of cubic C60 as a function of temperature (left)79 and pressure
(right).80 (Copyright © 2006, Taylor & Francis)
Table 1-1. Examples of free energy increase per unit cell in molecular crystal system under
pressure
Crystal P0 (GPa) P(GPa) Δ(PV)(eV) Cell occupancy
Methane82 12 202 222 4
Benzene83 1 27 22 2
Nitrobenzene84 2 27.5 27 2
Acetaminophen85 1 4 11 2
14
Figure 1-9. Free energy variation of CsI crystal under pressure change (solid dots) and
temperature change (hollow circles).81 (Copyright © 1989, Annual Reviews Inc.)
In the in the schematic configuration coordinate diagram from Drickamer (Figure 1-10),
potential energies of the ground and excited states are plotted against some characteristic
displacement of the system.86 It’s clear to see the activation energy (E≠) reduced after
pressurization which indicates lower barrier is needed to overcome reaching the transition state.
At the same time, the Eth, which is the over-all energy E0 in Hush’s original work,87 also
decreased at higher pressure. The relation between E≠ and Eth is described in Equation 1-2.
Figure 1-10. Schematic configuration of coordinate diagram showing the evolution of the
electronic structure at initial volume (left) and at reduced volume (right).86 (Copyright © 1972,
National Academy of Sciences)
2
max
max
( )
4( )th
hE
h E
(Equation 1-2)
The changes of the electronic structure due to reduced volumes in a two-level system can
be stated by the terms of HOMO (Highest Occupied Molecular Orbital) and LUMO (Lowest
Unoccupied Molecular Orbital) in molecular system. The HOMO-LUMO gap would be narrowed
at high pressure because the Eth is much lower, and a significant population of the excited state
can occur at room temperature.74 Figure 1-11 presents the calculated HOMO-LUMO gap
evolving with the pressure change in Sm@C88 system as a demonstration.88 Further, the
configuration coordinate diagram shows the structures of the ground and excited state became
very similar after volume reducing. Consequently, some energy barriers will be reduced as well.
15
Figure 1-11. Calculated HOMO-LUMO gap as a function of pressure of [email protected]
1.3.2 Pressure Dependence of Chemical Equilibrium and Reaction Rates
Under isothermal condition, equilibrium constant K increases exponentially with pressure
when the process involving volume contraction as demonstrated in Equation 1-3.
0ln( )T
KRT V
P
(Equation 1-3)
Therefore, the tuning of chemical equilibrium by pressure is only associated with a volume
variation but not controlled by dynamic factors. When there is a large pressure change,
compressibility of the system has to be taken into account. There has been much less attention to
the chemical equilibrium of condensed phases than liquid solution due to technical challenges.
The molecular mobility in solid state is considerably slower than in liquid phases due to the
density increase. Meanwhile the steep increase of the energy barrier associate with structural
constraints further reduced the mobility. The effect of steric hindrance in condensed phase may
make the system remain at a secondary minimum, i.e., a metastable state.
16
Arrhenius Equation 1-4 is the empirical law to describe the temperature dependence of
rate constant in many reactions.
2
ln aEd k
dT RT (Equation 1-4)
The activation energy Ea is usually obtained by measuring the rate constants k at different
temperature under isobaric condition and derived from the lnk vs 1/T slope. Analysis on pressure
effects on reaction rate is less common and even so most reported studies are focused on the low-
pressure region (less than 0.2 GPa).89 Kinetic studies beyond 1 GPa is extremely rare while the
threshold polymerization pressures of most organic molecules are higher than 1 GPa.
As discussion above the significant reduction of intermolecular distances will greatly
slow down or even inhibit the reaction process and if the reacting species are not in favorable
orientation, a rearrangement will be required. Herein, the impact of intermolecular interactions is
not negligible but decisive in the kinetic terms of pressure induced transformation and reaction.
Combining Arrhenius Equation with Equation 1-3, the dependence on pressure of the reaction can
be written as Equation 1-5.
ln k V
P RT
(Equation 1-5)
ΔV≠ is the pressure-independent activation volume that extracted by using isothermal rate
constant measurement at different pressures. However, the practical measurement is not as
straightforward as the equation. The technical complications impede the measurements of those
reactions that have large ΔV≠ (for example more than -10 cm3/mol). The normal range of volumes
of activation for organic reactions is -30 to +30 cm3/mol as listed in Table 1-2. Moreover, it has
been found that the linear evolution of lnk with pressure can only maintain in very limited
pressure range. To depict the reaction trajectory, more complex volume terms need to be
introduced and varies from reaction to reaction. 90
17
Table 1-2. Activation volume values of various reactions91
Reaction ΔV≠ (cm3/mol)
Homolysis 5 to 20
Polymerization (radical propagation) ~-20
Cycloaddition Diels-Alder -25 to -40
Cycloaddition intramolecular -25 to -30
Cycloaddition dipolar -40 to -50
Cycloaddition (2+2) -40 to -55
Epoxide-ring opening -15 to -20
Even though people have gained good knowledge of the chemical kinetics in solution
phase and gas phase, but they are not immediately applicable to crystal phase. For example
butadiene dimerization follows a second-order kinetic law in solution whereas it performs as first
order in solid.92 And it’s worth to note that the rate-determining step in these two cases are
completely different indicating reaction kinetics are strongly related to the molecular
reorganization in the transition state. The constraints of geometry and volume could determine the
feasibility of possible reaction pathways. Therefore, in order to trigger the reactions that require
high mobility, thermal treatment is necessary or else this pathway would be retarded or even
prevented.74 However, on the good aspect, the confinement of the molecule, sharing the same
spirit of crystal engineering, will realize some particular reaction processes that not possible at
ambient pressure.
1.3.3 Pressure Effects on Electronic Structure
Many molecualr systmes exhibit color changes under pressure such as solid oxyen turned
to red color when the pressure is close to 20 GPa (Figure 1-12).93 This observation is closely
related to perssure shift fo the abosroption edge. Beside direct observation, electronic absorption
and emission spectra are commonly employed to understand the change electrnic properties by
18
pressure tuning. The increase of dielectric constants of hydrocarbons is more than 10% within
few kilobars change of pressure.94
Figure 1-12. Micrograph of ε-oxygen crystal in helium at 17.6 GPa.93 (Copyright © 2006,
Springer Nature)
Figure 1-10 in section 1.3.1 explains the two major contribution of increased pressure:
shift of absorption edge and the other is the broadening of the absorption bands.86 The first
contribution is related to the vertical shift of the ground and excited states as briefly discussed
earlier. The changes of electronic distribution could sufficiently shrinkage of the energy gap and
populated the excited state and therefore triggering a chemical reaction. The second contribution
is corresponded to the lateral shift of the relative position of two minima regarding to possible
different compressibility of the two electronic states along specific configuration coordinates.
Frequency shift of both absorption and emission bands will be observed. These two contributions
are not necessary to be in the same trend and sometimes they have opposing effects.74
In molecular systems, π-π* transitions are generally most pressure sensitive because π
orbitals have lower overlap than sigma orbitals at ambient conditions. As a consequence,
unsaturated organic compounds are particularly reactive under high pressure. It’s a common
phenomenon to see energy shift during in situ vibrational spectroscopy measurements. If a
significant increase of the dipole moment happened upon excitation, a red shift of the transition
energy will be observed along with the increase of pressure. On the other hand, blue shift is
19
corresponded to decreased dipole moment when the configuration achieved. Recent studies of
absorption spectra of aromatic compounds, such as benzene,95-96 anthracene,97-98 pyridine,99
provide more insight on optical gap dependence of pressure.
More recently, some works on two-photon absorption spectroscopy have been reported.95,
99 The advantages comparing to conventional one-photon absorption techniques includes
overcoming the limitation brought by diamond absorption edge and by the small cross section
that minimizes the production of excimers, potential nuclei of a reaction. This new technique also
benefits the studies of fluorescence and photo chemical reactive. For example, in 2011, Bini et al.,
reported an intriguing work on fluorescence of pyridine under pressure.99 Pyridine is expected to
have low fluorescence quantum yield since its lowest excited state has nπ* character. However,
solid pyridine presented remarkable increase of fluorescence (six orders of magnitude higher than
liquid) after pressurized to 3 GPa. The large intensification was explained by the inversion of the
lowest nπ*(1B1) and ππ* (1B2) excited states (Figure 1-13) due to the stronger hydrogen bond
network with increasing pressure.
Figure 1-13. Schematic energy diagram of two-photon induced fluorescence of pyridine
due to compression. The inversion of the lowest nπ* and ππ* excited states are demonstrated.99
(Copyright © 2011, American Chemical Society)
20
1.4 Goals of the Dissertation
The motivation of this project is to stabilize and characterize new forms of carbon
through tailored synthetic processes for the development of new structural materials for energy
application. Nanothread is a novel one-dimensional sp3 hybridized carbon nanomaterial derived
from compressing benzene to 25 GPa It was first synthesis by Dr. Fitzgibbons from Prof.
Badding’s group in 2013.4 This material is predicted to possess superlative mechanical properties
and fills the last remaining entries of the dimensionality- hybridizations matrix of carbon material
(Figure 1-14).
Figure 1-14. Matrix of carbon nanomaterial dimensionality and bonding geometry.
As this new material is in its infancy, there is broad range of research topics remain to be
explored. The goals of this dissertation include: (1) improve the crystallinity of carbon nanothread
via optimizing synthesis method for further structural identification; (2) realize nanothread
formation in diamond anvil cell which is a more common high pressure apparatus which will
21
allow varies in situ studies to understand reaction pathway; (3) expand nanothread family by
developing new types of nanothreads that owing different physical and chemical properties and
therefore enable varieties of future application.
Chapter 3 covers the progress that has been made for goal (1) and (2). Synthesis of
benzene nanothread single crystal will be presented and a unique non-topochemical solid state
reaction will be discussed. Chapter 4 demonstrates the synthesis and structural characterization of
a new member of nanothread family, carbon nitride nanothread that derived from pyridine to
achieve goal (3). The crystal structure of pyridine under high pressure was found different from
literature reports when studying the formation of carbon nitride nanothread. Primary
crystallography analysis and discussion of potential high-pressure phase transitions are embodied
in Chapter 5.
22
References
(1) Croswell, K. The Alchemy of the Heavens, Anchor Books: New York, 1995.
(2) Crosbie, T. M.; Eathington, S. R.; Johnson, G. R.; Edwards, M.; Reiter, R.; Stark, S.;
Mohanty, R. G.; Oyervides, M.; Buehler, R. E.; Walker, A. K.; Dobert, R.; Delannay, X.;
Pershing, J. C.; Hall, M. A.; Lamkey, K. R. Plant Breeding: Arnel R Hallauer International
Symposium 2006, 3-50.
(3) Pierson, H. O. Handbook of carbon, graphite, diamonds and fullerenes: processing,
properties and applications, Noyes Publications: Park Ridge, 1995.
(4) Fitzgibbons, T. C.; Guthrie, M.; Xu, E. S.; Crespi, V. H.; Davidowski, S. K.; Cody, G. D.;
Alem, N.; Badding, J. V. Nat Mater 2015, 14, 43-7.
(5) Roman, R. E.; Kwan, K.; Cranford, S. W. Nano Lett 2015, 15, 1585-1590.
(6) Zhan, H. F.; Zhang, G.; Tan, V. B. C.; Cheng, Y.; Bell, J. M.; Zhang, Y. W.; Gu, Y. T. Adv
Funct Mater 2016, 26, 5279-5283.
(7) Zhan, H. F.; Zhang, G.; Tan, V. B. C.; Gu, Y. T. Nat Commun 2017, 8.
(8) Rubin, A. E. Mineral Mag 1997, 61, 699-703.
(9) Ober, J. A. Mineral commodity summaries 2017; US Geological Survey: 2017.
(10) Litvin, Y. A. Mantle Rocks and Diamond-Associated Phases: Role in Diamond Origin,
Springer, 2017.
(11) West, A. R. Solid state chemistry and its applications, John Wiley & Sons: Chichester, 2014.
(12) Erskine, D. J.; Nellis, W. J. Nature 1991, 349, 317-319.
(13) Khaliullin, R. Z.; Eshet, H.; Kühne, T. D.; Behler, J.; Parrinello, M. Nature materials 2011,
10, 693.
(14) Britun, V. F.; Kurdyumov, A. V.; Petrusha, I. A. Powder Metall Met C+ 2004, 43, 87-93.
(15) https://electronics-cooling.com/wp-content/uploads/2002/02/cphased.gif.
(16) Peschel, G. Obtained online from: http://www. physik. fu-berlin. de/einrichtungen/ag/ag-
reich/lehre/Archiv/ss2011/docs/Gina_Peschel-Handout. pdf, published on 2011, 5.
(17)
http://wps.prenhall.com/wps/media/objects/340/348272/Instructor_Resources/Chapter_02/Text_I
mages/FG02_14.JPG.
(18) http://wps.prenhall.com/wps/media/objects/3081/3155729/blb0905.html.
(19) Delhaes, P. Graphite and precursors, CRC Press: 2014.
(20) Hassel, O.; Mark, H. Zeitschrift für Physik 1924, 25, 317-337.
(21) Riley, D. P. Nature 1944, 153, 587-588.
(22) Pauling, L.; Brockway, L. O. Journal of the American Chemical Society 1937, 59, 1223-
1236.
(23) Painter, G.; Ellis, D.; Lubinsky, A. Physical Review B 1971, 4, 3610.
(24) Primak, W.; Fuchs, L. Phys Rev 1952, 86, 651-651.
(25) Dutta, A. K. Phys Rev 1953, 90, 187-192.
(26) Rupprecht, G.; Smakula, P. Carbon 1964, 1, 393-393.
(27) Carter, J. G.; Huebner, R. H.; Hamm, R. N.; Birkhoff, R. D. Phys Rev 1965, 137, A639-&.
(28) Taft, E. A.; Philipp, H. R. Phys Rev 1965, 138, A197-&.
(29) Kidalov, S. V.; Shakhov, F. M. Materials 2009, 2, 2467-2495.
(30) Sun, K.; Stroscio, M. A.; Dutta, M. Superlattice Microst 2009, 45, 60-64.
(31) Smalc, M.; Shives, G.; Chen, G.; Guggari, S.; Norley, J.; Reynolds, R. A. Advances in
Electronic Packaging 2005, Pts a-C 2005, 79-89.
(32) Kaehr, R. The book of Diamonds. Glasgow 2007.
(33) http://www.d-edp.jp/en/technology-3.html.
23
(34) Hazen, R. M. The diamond makers, Cambridge University Press: 1999.
(35) Savage, R. H. J Appl Phys 1948, 19, 1-10.
(36) Ohzuku, T.; Iwakoshi, Y.; Sawai, K. Journal of The Electrochemical Society 1993, 140,
2490-2498.
(37) Peled, E.; Menachem, C.; Bar‐Tow, D.; Melman, A. Journal of The Electrochemical
Society 1996, 143, L4-L7.
(38) Arregui-Mena, J. D.; Bodel, W.; Worth, R. N.; Margetts, L.; Mummery, P. M. Carbon 2016,
110, 497-517.
(39) Karfunkel, H. R.; Dressler, T. Journal of the American Chemical Society 1992, 114, 2285-
2288.
(40) Diederich, F.; Rubin, Y. Angew Chem Int Edit 1992, 31, 1101-1123.
(41) Hirsch, A. Nature Materials 2010, 9, 868-871.
(42) Smalley, R. E. Rev Mod Phys 1997, 69, 723-730.
(43) Kademani, B.; Kalyane, V.; Kumar, V. SRELS Journal of Information Management 2002,
39, 409-434.
(44) Ahmad, S. Iete Tech Rev 1999, 16, 297-310.
(45) Dresselhaus, M. S.; Dresselhaus, G.; Eklund, P. C. Science of fullerenes and carbon
nanotubes: their properties and applications, Elsevier: 1996.
(46) Kratschmer, W.; Lamb, L. D.; Fostiropoulos, K.; Huffman, D. R. Nature 1990, 347, 354-
358.
(47) Hawkins, J. M.; Meyer, A.; Lewis, T. A.; Loren, S.; Hollander, F. J. Science 1991, 252, 312-
313.
(48) Iijima, S. Nature 1991, 354, 56-58.
(49) Iijima, S.; Ichihashi, T. Nature 1993, 363, 603-605.
(50) Bethune, D. S.; Kiang, C. H.; Devries, M. S.; Gorman, G.; Savoy, R.; Vazquez, J.; Beyers,
R. Nature 1993, 363, 605-607.
(51) Novoselov, K. S.; Geim, A. K.; Morozov, S. V.; Jiang, D.; Zhang, Y.; Dubonos, S. V.;
Grigorieva, I. V.; Firsov, A. A. Science 2004, 306, 666-669.
(52) Geim, A. K.; Novoselov, K. S. Nature Materials 2007, 6, 183-191.
(53) Friedman, S. H.; Decamp, D. L.; Sijbesma, R. P.; Srdanov, G.; Wudl, F.; Kenyon, G. L.
Journal of the American Chemical Society 1993, 115, 6506-6509.
(54) Bosi, S.; Da Ros, T.; Spalluto, G.; Balzarini, J.; Prato, M. Bioorg Med Chem Lett 2003, 13,
4437-4440.
(55) Ji, Z. Q.; Sun, H. F.; Wang, H. F.; Xie, Q. Y.; Liu, Y. F.; Wang, Z. J Nanopart Res 2006, 8,
53-63.
(56) Peng, B.; Locascio, M.; Zapol, P.; Li, S. Y.; Mielke, S. L.; Schatz, G. C.; Espinosa, H. D.
Nat Nanotechnol 2008, 3, 626-631.
(57) Pop, E.; Mann, D.; Wang, Q.; Goodson, K. E.; Dai, H. J. Nano Lett 2006, 6, 96-100.
(58) Beigbeder, A.; Degee, P.; Conlan, S. L.; Mutton, R. J.; Clare, A. S.; Pettitt, M. E.; Callow,
M. E.; Callow, J. A.; Dubois, P. Biofouling 2008, 24, 291-302.
(59) Yang, J. Carbon Nanotubes Reinforced Composites for Wind Turbine Blades. Case Western
Reserve University, 2012.
(60) Plunkett, J. W. Plunkett's Sports Industry Almanac 2009, Plunkett Research, Ltd.: 2008.
(61) De Volder, M. F. L.; Tawfick, S. H.; Baughman, R. H.; Hart, A. J. Science 2013, 339, 535-
539.
(62) Klintenberg, M.; Lebegue, S.; Katsnelson, M. I.; Eriksson, O. Physical Review B 2010, 81.
(63) Sahoo, S. Indian J Pure Ap Phy 2011, 49, 367-371.
(64) Novoselov, K. S.; Jiang, D.; Schedin, F.; Booth, T. J.; Khotkevich, V. V.; Morozov, S. V.;
Geim, A. K. P Natl Acad Sci USA 2005, 102, 10451-10453.
24
(65) Soldano, C.; Mahmood, A.; Dujardin, E. Carbon 2010, 48, 2127-2150.
(66) Bolotin, K. I.; Sikes, K. J.; Hone, J.; Stormer, H. L.; Kim, P. Phys Rev Lett 2008, 101.
(67) Lin, Y. M.; Dimitrakopoulos, C.; Jenkins, K. A.; Farmer, D. B.; Chiu, H. Y.; Grill, A.;
Avouris, P. Science 2010, 327, 662-662.
(68) Brownson, D. A. C.; Kampouris, D. K.; Banks, C. E. Chem Soc Rev 2012, 41, 6944-6976.
(69) Kim, K. S.; Zhao, Y.; Jang, H.; Lee, S. Y.; Kim, J. M.; Kim, K. S.; Ahn, J. H.; Kim, P.;
Choi, J. Y.; Hong, B. H. Nature 2009, 457, 706-710.
(70) Li, X. P.; Zhang, Q. M.; Chen, X.; Gu, M. Sci Rep-Uk 2013, 3.
(71) Murayama, H.; Tomonoh, S.; Alford, J. M.; Karpuk, M. E. Fuller Nanotub Car N 2004, 12,
1-9.
(72) http://www.physics.manchester.ac.uk/our-research/research-impact/graphene/.
(73) https://www.nextbigfuture.com/2015/06/about-1000-tons-per-year-of-graphene.html.
(74) Vincenzo, S.; Roberto, B. Materials under extreme conditions: molecular crystals at high
pressure, World Scientific: 2013.
(75) Kitaigorodsky, A. Molecular crystals and molecules, Elsevier: 2012.
(76) Califano, S.; Schettino, V.; Neto, N. Lattice dynamics of molecular crystals, Springer
Science & Business Media: 2012.
(77) Jones, W. Organic molecular solids: properties and applications, CRC Press: 1997.
(78) Bernstein, J.; Bernstein, J. M. Polymorphism in molecular crystals, Oxford University Press:
2002.
(79) David, W. I. F.; Ibberson, R. M.; Dennis, T. J. S.; Hare, J. P.; Prassides, K. Europhys Lett
1992, 18, 219-225.
(80) Qiu, W.; Chowdhury, S.; Hammer, R.; Velisavljevic, N.; Baker, P.; Vohra, Y. K. High
Pressure Res 2006, 26, 175-183.
(81) Jeanloz, R. Annu Rev Phys Chem 1989, 40, 237-259.
(82) Sun, L. L.; Yi, W.; Wang, L.; Shu, J. F.; Sinogeikin, S.; Meng, Y.; Shen, G.; Bai, L. G.; Li,
Y. C.; Liu, J.; Mao, H. K.; Mao, W. L. Chem Phys Lett 2009, 473, 72-74.
(83) Ciabini, L.; Santoro, M.; Bini, R.; Schettino, V. J Chem Phys 2001, 115, 3742-3749.
(84) Citroni, M.; Datchi, F.; Bini, R.; Di Vaira, M.; Pruzan, P.; Canny, B.; Schettino, V. J Phys
Chem B 2008, 112, 1095-1103.
(85) Boldyreva, E. V.; Shakhtschneider, T. P.; Vasilchenko, M. A.; Ahsbahs, H.; Uchtmann, H.
Acta Crystallogr B 2000, 56, 299-309.
(86) Drickame.Hg; Slichter, C. P.; Frank, C. W. P Natl Acad Sci USA 1972, 69, 933-937.
(87) Hush, N. S. Electrochim Acta 1968, 13, 1005.
(88) Cui, J.; Yao, M.; Yang, H.; Liu, Z.; Ma, F.; Li, Q.; Liu, R.; Zou, B.; Cui, T.; Liu, Z. Sci Rep-
Uk 2015, 5, 13398.
(89) Vaneldik, R.; Asano, T.; Lenoble, W. J. Chem Rev 1989, 89, 549-688.
(90) Winter, R.; Jonas, J. High pressure molecular science, Springer Science & Business Media:
2012.
(91) Isaacs, N. S. Liquid phase high pressure chemistry, John Wiley & Sons Incorporated: New
York, 1981.
(92) Citroni, M.; Ceppatelli, M.; Bini, R.; Schettino, V. J Chem Phys 2003, 118, 1815-1820.
(93) Lundegaard, L. F.; Weck, G.; McMahon, M. I.; Desgreniers, S.; Loubeyre, P. Nature 2006,
443, 201-204.
(94) Brazier, D.; Freeman, G. Canadian Journal of Chemistry 1969, 47, 893-899.
(95) Citroni, M.; Bini, R.; Foggi, P.; Schettino, V. P Natl Acad Sci USA 2008, 105, 7658-7663.
(96) Ciabini, L.; Santoro, M.; Bini, R.; Schettino, V. Phys Rev Lett 2002, 88.
(97) Aust, R. B.; Drickamer, H. G.; Bentley, W. H. J Chem Phys 1964, 41, 1856.
25
(98) Sonnenschein, R.; Syassen, K.; Otto, A. The Journal of Chemical Physics 1981, 74, 4315-
4319.
(99) Fanetti, S.; Citroni, M.; Bini, R. J Phys Chem B 2011, 115, 12051-12058.
26
Chapter 2
Instrumentation and Characterization Techniques
Polymerization of benzene happened ~20 GPa is far beyond the pressure range that
organic or polymer chemistry typically handling with. High pressure cells, which are more
commonly used in geoscience and physics, have been employed as the reactor chamber in
nanothread synthesis. A special pressure control system is required for this kinetically controlled
reaction. At the same time, one of topic that this thesis focused on is to explore the formation
mechanisms thus in situ characterization ability is quite essential. In the end of this chapter, basic
background of vibrational spectroscopy will be presented along with two featured Raman
techniques that capable in our lab.
2.1 High Pressure Cells
The synthesis of carbon nanothread was first performed in Paris-Edinburgh (PE) cell by
compressing benzene to 25 GPa.1 Optimized PE cell synthesis enables the high pressure recovery
possesses single crystal quality. Progress has also been made in enabling nanothread synthesis in
diamond anvil cell (DAC) which makes more in situ characterization approaches available. Both
two works will be presented in Chapter 3.
2.1.1 Diamond Anvil Cell
The diamond anvil cell (DAC) is a commonly used high pressure device because it can
generate high pressure by simply squeezing samples in between two opposing diamonds. The first
27
DAC was made by Charlie Weir in 1958.2 After 60 year’s development, DACs are capable in a
wide range of temperature conditions, from mili-Kelvin3 to 6000 Kelvin4 and the accepted static
pressure limit has been pushed to >400 GPa.5 Because diamond is chemically inert, it’s suitable
for studies of both organic and inorganic systems. These versatilities make DACs become popular
in the researches on geoscience, physics, chemistry and even bioscience6 to explore novel
materials or new phenomena at extreme conditions.
Figure 2-1. Top: Assembling of symmetric diamond anvil cell.7 (Image courtesy of Dr. Barbara
Lavina at UNLV.) Bottom: configuration of opposed diamond anvils (left) and indented gasket as
sample chamber (right).8
Figure 2-1 demonstrates the assembling of a symmetric DAC which is the type of cell
used in our lab. A thin metal foil, typically around 0.3mm in thickness, works as gasket for DAC.
The gasket need to be preindented by the pair of diamonds inside the cell and a hole is drilled in
the center that later used as sample chamber.9 Rhenium is a desired gasket material since it’s
strong and stiff. Stainless steel is a cheaper alternative and usable for low pressure experiments.
Both types of gaskets were used in this dissertation work. Pressure adjustment can be simply
28
achieved by turning the screws. Motorized mechanical pressure control or membrane control is
employed in the circumstances to fulfil the special requirements such as, remote control in
synchrotron or neutron radiation hutch, low/high temperature conditions and very rapid or slow
compression/decompression. During nanothread synthesis, pressure rates in both uploading and
downloading process have to be delicately controlled. Therefore a double-stage membrane system
is required. Detailed discussion of this technique is in section 2.2.
2.1.2 Paris-Edinburgh Cell
The motivation of building PE cell is to benefit neutron diffraction which requires much
larger (104~106 orders) sample volumes than X-ray diffraction.10 Before the invention of PE cell,
Bridgeman anvils have been used for creating pressure on large samples for a long time.
Bridgman anvils typically have flat faces and the deformation is restricted by the internal friction
in the gasket and the frictional forces at the anvil-gasket interface only.11-12 Therefore the total
sample volume is determined by the diameter of the sample and initial gasket thickness.
However, the larger the gasket initially is, the more readily the gasket will be deformed under
applied load. One way to reduce the outward low and thinning of the gasket is by introducing a
profile onto the face of the anvil such as have recesses machined onto the pressure-generating
surfaces.13-14 The recesses form toroidal volumes which support the central part of gasket to
against the hoop stresses within it and greatly restrict the outward radial flow of the gasket.
Herein the sample volume becomes more pressure-attainable.
The anvils applied in this thesis work have the double-toroidal design which has a central
recess for accommodating sample and another two concentric toroidal recesses at surrounding.
Figure 2-2 shows the actual set-up and schematic drawing.15 The center section A1 an A2 are
made of polycrystalline diamond (PCD). B1 and B2 are tungsten carbide supporting rings
29
providing reinforcement for the PCD dies. The toroidal profile is spark-eroded in the blank
sintered diamond section of the die before final polishing of the surface. The gasket has five parts
(D1, D2, E, F and G) and the sample volume is ~17 mm3. In neutron diffraction experiment, TiZr
(67:33 molar ratio alloy)16 gasket is required. Since the TiZr alloy is very expansive, stainless
steel gasket and Zr gasket are also used for off-line synthesis.
Figure 2-2. Left: actual photo of Paris-Edinburgh (PE) cell at SNAP beamline, SNS. Right:
schematic figure of the parts used for PE cell.15 (Copyright © 2012 American Institute of Physics)
It’s noteworthy that the pressure distribution in the sample chamber of PE cell and DAC
is different. And it has been found that this difference may influence nanothread formation
(Chapter 3). A parabolic pressure profile was first assumed for the pressure on the sample in PE
cell (Figure 2-3).17 Later, a finite element method study combined with experimental neutron
scattering results15, 18 indicates the pressure distribution in PE cell (Figure 2-4, left) is less
uniaxial than in DAC (Figure 2-4, right ).
30
Figure 2-3. Schematic pressure profile assumed for the sample and gasket assembly in PE cell
and ρ0 is the maximum pressure.17 (Copyright © 2010, American Physical Society)
Figure 2-4. Left: the maximum principal stress (solid curve) and shear stress (dotted curve) on the
surface of double-toroidal PE cell anvil as a function of distance from anvil center.15 (Copyright
© 2012, American Institute of Physics) Right: pressure distributions in a beveled DAC.5
(Copyright © 2018, National Academy of Sciences)
2.2 Pressure Control System
The pressure rates have strong impacts on the high pressure polymerization of benzene.
Fast or moderate compression and decompression usually leads to amorphous solid recovery. The
formation of nanothread requires much slower pressuring and releasing rates than hand tightening
31
screws can provide. At BL-3 (Spallation Neutrons and Pressure Diffractometer, SNAP), where
the large-scale nanothread synthesis performed, PE cell anvils are pushed by a hydraulic oil pump
and the pressure rates generated on the sample can be tuned by the flow rates of the pump. In
order to realize synthesis of nanothread in DAC, delicate control of pressure rates on both
uploading and downloading processes is the first and most critical issue to overcome. The double-
diaphragm design of pneumatic remote pressure control system developed by High Pressure
Collaborative Access Team (HPCAT) is the key to solve this problem.
The pneumatically controlled DACs19-23 are first applied in cryogenic studies and later
become popular in a wide range of pressure-temperature conditions24-25 by proving remote and
automated static and dynamic pressure control. The double-diaphragms, which is more
commonly called as double-membranes, are two identical thin steel sheets that deform plastically
and elastically during expansion.26 The membranes provided by HPCAT are manufactured by
laser welded 304 type stainless steel with 250 μm in thickness. Typically, the membranes are
inflated to 1-1.5 mm to ensure their re-usability although the limit of inflation is up to 2-4 mm.
They are quite resilient and can serve near a hundred of cycles without failing if used with care.27
The commercially available automated pressure controller (GE PACE 5000) is able to generate
pressure in the range of 0-210 bar (0-3000 Psi) with excellent precision (0.02%) and
extraordinary stability (0.003%). It greatly fits the need of nanothread synthesis that the force
applied to DAC is no more than ~700 psi to reach ~25 GPa.
In practical, most of experiments only need one membrane and the mechanism of
increasing pressure is more straightforward. However, the decreasing pressure process has always
been problematic due to the plastic deformation of the gasket and frication of the DAC between
piston and cylinder.27 Especially when the pressure media is condensed gases or liquid who are
more compressible, the gasket need to be plastically deformed to generate appreciable pressure on
sample. As a result, with single membrane system, the DACs are readily to get stuck at high
32
pressure while the pressure inside the membrane has already been released completely. In order
to push back DAC pressure to ambient by one membrane, one has to reduce the pressure across
the elastic limit of the gasket by releasing some pressure medium from sample chamber. This
procedure has very little control of the rates and has the risks of losing sample unexpectedly.
Thus double-membrane is necessarily needed to decrease the pressure in a controllable
fashion. Figure 2-5 (left) shows a symmetric DAC with a pneumatic decompression attachment.27
A and C are the double-can parts. F is the membrane while G is the pusher piston. The assembly
is attached by two screws (H) to the piston part of the DAC (E). The two setscrews (B) and two
pins (D) play very important roles in pressure control. The setscrews need to be carefully adjusted
to the position where when the pushing pins touch the cylinder side of DAC there is a gap about
few hundred μm between the can (C) and pusher piston (G). Therefore, the distance between G
and cylinder part of DAC is constrained by the pins. When inflated membrane pushes G toward
the DAC, the pin will push the cylinder of the cell and the net result is the cell piston get pulled
back and thus make the cell open and pressure decrease. Figure 2-5 (right) demonstrates the
complete “push-pull” double-membrane setup.
Figure 2-5. Left: decompression attachment of double-diaphragm design. Right: Dual double-
diaphragm setup for compression and decompression experiment.27 (Copyright © 2015, AIP
Publishing LLC.)
33
Besides providing very slow compression/decompression (the slowest rate has been
applied is 0.01 GPa/min) that benefit for nanothread formation, this configuration has also been
applied in pressure quench experiments for exploring metastable phases which requires very rapid
pressure releasing rates. For example, a high purity metallic silicon phase was quenched from 20
GPa to ambient pressure at the rate of ~0.6-1 TPa/s.28
2.3 High Pressure X-ray Diffraction
X-ray diffraction (XRD) is a powerful and indispensable technique to obtain structural
information such as unit cell parameters, atomic positions, thermal parameter, internal strain, site
occupancy and electron density distributions.29 Like diffraction experiments done at ambient
pressure, there are two categories of high pressure diffraction techniques differentiated by sample
character. In high pressure crystallography works, powder XRD is commonly used in studying
phase identification, pressure-induced chemical reaction, preferred orientation, particle size,
texture and strain effects. Rietveld refinement30 is the critical step to extract these useful
information after experimental measurement but it won’t work without an initial model. As
discussed in Chapter 1, high pressure phase transitions and reactions will bring up new materials
that not available at ambient pressure and as a result most of the structures are unknown.
Therefore, before performing Rietveld refinement, traditional methods such as Patterson and
direct methods, or more modern first principle calculations31-32 are needed to establish the basic
features of the structure. When the sample is a single crystal, both intensities of diffracted beams
and the orientation of the crystal will be measured. In this way, all the basic crystallographic
information (lattice parameters, atomic positions, thermal displacement parameter and electron
density distribution) could be solved solely by the diffraction without necessarily knowing the
34
initial model. High pressure single crystal diffraction played important roles in the research works
in Chapter 3 and 5 of this dissertation. More background knowledge and basic principles of high
pressure diffraction will be introduced in this section (with more focus on single crystal
diffraction though).
2.3.1 Single Crystal Crystallography under Pressure
A we all know that water transforms to solid at low temperature and varies of studies
have shown that there are more than one phase of ice existed. By employing cryogrinding
techniques, powdered ice can be obtained with reduced crystallite size and random orientations
that allows for structure solution and refinements by powder diffraction methods. However, this
sample preparation technique is not readily used in high pressure researches of small organic
molecules those are liquid at ambient pressure at room temperature. In practical experiments, the
samples are more likely to be highly textured that offer poor powder average which greatly
impede accurate structure identification and sometimes make structure determination impossible.
There are two modes of diffraction geometry are available for high pressure diffraction,
transmission mode and transverse mode.9 In this dissertation, only transmission mode was
employed as it’s more commonly used in most designs of diamond anvil cell by having the
incident X-ray beam pass through one diamond, the sample, and then the opposing diamond.
Sample need to be rotated during diffraction image collecting at synchrotron beamline. Therefore,
high pressure single crystal diffraction has several important differences compared with ambient
condition experiments due to the configuration of diamond anvil cell.33 First, limited X-ray
angular access from cell body and back plates significantly constrains the access of reciprocal
spaces. Second, diamond absorption correction is needed to correct the diffraction signal in order
to get reliable structure factors. Third, as the volume of sample chamber is very small, any scatter
35
or diffractions from gasket, pressure calibrant, pressure medium and diamond have to be clearly
intensified. Last but not the least, the rotation center of the cell has to be guaranteed being
corrected especially when there is large change of pressure.
Figure 2-6. Ewald construction and reciprocal lattice.9 (Copyright © 2018 GeoScienceWorld)
Ewald construction is a useful tool to elucidate the relationship between DAC symmetry
and its accessibility in reciprocal space. As shown in Figure 2-6,9 S is the origin of crystal, O* is
the origin of the reciprocal lattice and Phkl is a reciprocal lattice point. Rotating crystal will lead to
equivalent rotation of reciprocal lattice. Coherent scattering occurs when reciprocal space vector
k (λ/dhkl) between O* and Phlk lies on the surface of Ewald sphere. The incident beam is
represented by vector SPhkl and thus the diffraction angle 2θ is given by the angle between SPhkl
and SO*. In the case of single crystal diffraction performed at ambient conditions, sample crystal
has 360 degrees of rotation freedom and all k vectors within this spherically shaped reciprocal
space (with radius 2) will successively pass through the boundary of the Ewald sphere. However,
in DAC experiment, opening angle 2α (Figure 2-7) is the critical parameter that limit the
accessible region in reciprocal space. In transmission mode, 2θ is restricted in the range smaller
than (α+ψi), where ψi is the angle of incident beam relative to DAC. Therefore, the maximum
length of the limiting reciprocal lattice vector can be described as Equation 2-1 which is also the
boundary condition for diffraction:
36
max 2 / cos( )2
Ik
(Equation 2-
1)
Figure 2-7. Diffraction geometry of transmission mode (left) and diffraction conditions in
reciprocal space (right). Dashed lines are the boundary circles which limit the area for accessible
reciprocal lattice vectors k.9 (Copyright © 2018 GeoScienceWorld)
And the trace of the segments is kmax=2/ (λ cosα) for (ψi= α). The dumbbell shape
portions with grey shades in Figure 2-7 (right) represent the accessible reciprocal space. It’s good
to note that different form transverse mode, in transmission geometry, with conical access
windows for both incident and diffracted beams, opening angle 2α will not change by lateral
inclination. To get larger accessible reciprocal space in high pressure single crystal diffraction,
DACs with larger 2α are strongly preferred.
2.3.2 Determine the Pressure in Diamond Anvil Cell
When pressure is used as the parameter to induce phase transitions and chemical
transformations, precise measurements of pressure become essential. Direct calculation from
applied load to the anvil size can only provide approximate sample pressure but not fulfill the
requirements for any standard analysis and research. There are more factors need to take careful
37
consideration such as the distribution of the load over the anvils, the internal frication and plastic
or elastic deformation due to gaskets. Thus the use of secondary standard material as pressure
calibrant will overcome these issues.
There are two major approaches to determine the pressure in diamond anvil cell: the first
one is by monitoring the pressure induced wavelength shift of the calibrant via spectroscopic
measurement; the other is to add an internal diffraction standards with known equation of state.
Figure 2-8. Scheme of energy levels demonstrate ruby fluorescence.34
Ruby fluorescence is one of the most popular method for pressure calibration. Suitable
ruby materials contain 3000 to 5500 ppm Cr3+ and the sizes of chips or spheres are in the order of
10 microns.9 Ruby fluorescence is characterized by the doublelet with sharp band components
centered at 14402 cm-1 (R1 line at 694.2 nm) and 14432 cm-1 (R2 line at 692.8 nm) due to the 2E
4A2 electronic transition of Cr3+ in a distorted octahedral crystal field. The mechanism is
illustrated in Figure 2-8. Both lines exhibit pronounced red shift with increasing pressure and thus
make ruby fluorescence suitable as pressure calibrant. Varies of studies on pressure (to mega bar
range) and temperature (both high T and low T) of wavenumber shift have been performed and
accuracy is found to be within 2% up to 55 GPa.35
38
The internal diffraction standards usually provide more accurate pressure measurements
than fluorescence methods. This method can overcome the systematic errors such as differential
thermal expansion of the components of the DAC and the retarded mechanical relaxation within
the gasket. The pressure determination is from the measurement of unit cell volume changes of
the standard material and then derived from its equation of state curve. In order to obtain high
precision volume changes, materials with high compressibility are required to guarantee a high
significance pressure determination.9 On the other hand, the internal standards should have high
symmetry (ideally cubic) and small unit cell volumes so that minimized number of reflections
will appear along with sample material’s pattern. Au, Pt, Mo, MgO and NaCl are commonly used
internal pressure standard and their performance had been check to 2.5 Mbar.36
Ruby fluorescence method was employed in all the diamond anvil cell work presented in
this dissertation, including in situ X-ray diffraction and Raman spectroscopy experiments because
it enables rapid pressure determination (~1 second) with acceptable accuracy and precision.
Unlike DACs that allowing optical access, Paris-Edinburgh cell has to utilize internal standards
for in situ measurements. Lead (Pb) was loaded as pressure calibrant in neutron diffraction works
that will be presented in subsequent chapters. The equation of state and phase transitions of Pb
had been studied to 238 GPa.37
2.3.3 High Pressure Single Crystal Diffraction at Synchrotron Beamline
There are excellent works on small molecules crystal structures determined in diamond
anvil cell by in-house X-ray facilities. One of the classical example is that benzene phase II
structure solved by Piermarini at 2.5 GPa using lab-source X-ray diffractometer in 1969 as a
demonstration of progress in high pressure diffraction technique.38 Large single crystals usually
recrystallize via heating across melting point and then slowly cooled down to room temperatures.
39
The beam sizes of lab source X-ray are in the range of 100~300 micron which means a single
crystal in hundreds micron scale is needed. Large sample volume will decrease the maxima
pressure the DAC could reach and the crystal may break due to phase transitions or just because
pure compression since they are soft materials.39 Due to these limitation, most crystal structures
reported by in house x-ray work are below 2 GPa. Herein, the develop of synchrotron based X-
ray with smaller beam size but higher brightness is the key to realize single crystal diffraction at
much higher pressure range to probe abundant physical properties changes and chemical
reactions. The advantages of synchrotron radiation include29: high intensity over wide energy
range; small angular divergence; tunable energy and bandwidth; pulsed time structure; polarized
radiation and coherence of the X-ray wave. Great efforts have been make in cooperating the
merits of SR with HP research since 1980s. It took about three decades to reduce the X-ray beam
of synchrotron radiation sized from 30-50 micron to currently 2-3 micron.
The synchrotron X-ray diffraction works covered in this dissertation were conducted in
sector 16 which established by high pressure collaborative access team (HPCAT) of the
Advanced Photon Source (APS) of Argonne National Laboratory. In situ high pressure diffraction
experiments had been performed at beamline 16-ID-B and 16-BM-D (ID stands for “insertion
device” and BM shorts for “bending magnet”).40 APS type-A undulator is used at 16-ID-B which
provides high flux of >51011 photons/second at 30 keV and >31012 photons/second at 20 keV
on sample. Beam sizes are controlled as ~36 μm2 or ~12 μm2 by Kirkpatrick-Baez (KB)
mirrors in different dimensions. Beamline 16-BM-D allows for both angular-dispersive XRD and
energy-dispersive XRD (Laue diffraction). Bending magnet monochromatic beam provides flux
of >5108 photons/second at 30 keV and >3109 photons/second at 15 keV with 4 μm μm beam
size. The stage assembly for sample positioning is designed to have excellent stability and good
resolution of motion. Motorized x-y-z translation stages have a position resolution capable of
0.1μm and the angular resolution of rotation stage is 0.001 degree.
40
2.4 Vibrational Spectroscopy
Infrared (IR) and Raman spectroscopy both are common and important techniques used
for vibrational modes studies of materials which is quite useful for structural identification,
chemical component testing and analysis. Since these two forms of spectroscopy arise from
different processes and follow different selections rules, they are usually considered as
complementary techniques. The frequency, intensity and shape of vibration band measured by IR
and Raman provide informative knowledge such as energy, polar character or polarizability and
bonding environment. Therefore, some unique features shown in the vibrational spectra can be
used as “fingerprints” to identify particular functional groups or structures.
This section will first explain the basic principles of IR absorption and Raman scattering.
Two specific Raman techniques that have been performed other than standard Raman
measurement in this dissertation work will be presented in 2.4.3 (polarized Raman spectroscopy)
and 2.4.4 (ultra-low frequency Raman spectroscopy).
2.4.1 Principles of Infrared Spectroscopy
Typical IR spectrometer covers broad range from 14000 to 4000 cm-1. Mid-IR (4000-400
cm-1) spectroscopy provides characteristic fundamental vibrations including the “fingerprint” of
molecular structures. Near-IR (beyond 4000 cm-1) spectroscopy measures the overtone and
combination bands and has outstanding performance in rapid and accurate quantitation.41 Mid-IR
spectrometer was employed in all the IR spectroscopy works demonstrated in this thesis.
IR spectroscopy measures the transitions between molecular vibrational energy levels as
the result of absorption. Radiation frequency and dipole moment are the important components in
this process. The interaction of the radiation with matters is regarding to a resonance condition
41
where the specific oscillating radiation frequency matches the intrinsic frequency of a particular
normal mode of vibration.41 Transferring the energy from IR photon to the molecule requires a
change of dipole moment between molecular vibrational energy levels. This is well-known as the
selection rule for IR spectroscopy.
Figure 2-9. Opposite oscillated molecular dipole generated by oscillating electric field of
photon.41 (Copyright © 2011 Elsevier Inc.)
Figure 2-9 depicts this process: the oscillating electric field of IR radiation generates
forces on the dipole of the molecule. The oppositely directed forces on the positive and negative
charges of molecular dipole moment and make the dipole moment oscillating and the dipole
spacing increasing and decreasing alternatively. Here the electric field is assumed to be uniform
over the entire molecule as the wavelength is much larger than most molecules. From quantum
mechanics, it’s learnt that the measured IR mode intensity is proportional to the square of the
change in the diploe moment. This knowledge will benefit in practical spectrum analysis.
42
2.4.2 Principles of Raman Spectroscopy
Deformability of the electron clouds about the molecule, namely polarizability, can bring
out induced dipole moment of the molecule by applying an external electric field. Figure 2-10
describes the response of a non-polar diatomic molecule placed in an oscillating electric field.
The negatively charged side attracts the nuclei and the positively charged side attracts the least
tightly bound outer electrons and herein an induced dipole moment is created. It’s notable that the
interaction between light and matter of IR absorption is a resonant condition involving electric
dipole mediated transition between vibration energy levels while the induced dipole moment is an
off-resonance interaction mediated by an oscillating electric field.41
Figure 2-10. Induced dipole moment of a homonuclear diatomic molecule originates from the
oscillating electric filed (represented by a charged plates capacitor on the left) of the incident
radiation.41 (Copyright © 2011 Elsevier Inc.)
Both Rayleigh and Raman are two photon processes. The incident photon is momentarily
absorbed by a transition from the ground state into a virtual state and a new photon is created and
scattered by a transition from this virtual state. Rayleigh scattering is an elastic scattering with no
energy change as shown in Figure 2-11 that the photon transits from virtual state back the ground
state. Raman scattering has a different story that the energy of scattered photon is different from
43
original photon because the transition happens from the virtual state to the first excited state due
to the inelastic collision between photon and molecule. Rayleigh scattering occurs at the laser
frequency and is the most probable event with 10-3 less intensity than the original radiation.
Raman scattering is much less probable because only ~1 in 10 millions of photon scattered
inelastically by an excitation.42
Figure 2-11. Rayleigh scattering and Stokes and anti-Stokes Raman scattering.41 (Copyright ©
2011 Elsevier Inc.)
Considering about the different initial states, Raman scattering is divided to two types:
Stokes and anti-Stokes. The former has the initial state at ground state and the latter has excited
state as the initial state. They can be clearly distinguished by the plot in Figure 2-11. The intensity
of Raman scattering is described in the Equation 2-2,41 where I0 is the incident laser intensity, N
is the number of scattering molecules in a given state, ʋ is the excitation frequency of the laser, α
is the polarizability and Q is the vibrational amplitude. This equation can explain why the anti-
Stokes Raman lines are weaker than Stokes lines. From thermal equilibrium Boltzmann’s law we
44
know that ground state is the most populated state at ambient temperature which means N is
larger in Stokes scattering than in anti-Stokes scattering.
4 2
0 ( )RI I NQ
(Equation 2-2)
This equation also implies that in order to enhance the intensity of Raman peak, applying
the laser with shorter excitation wavelength can be a solution. When characterize the
hydrogenated amorphous carbon (a-C:H) materials by visible laser, the weak peaks usually
superimposed by very strong photoluminescence (PL) background that makes the spectrum
analysis very challenging or even impossible. In order to solve this problem, UV Raman
(λ<300nm) have been employed resulting in stronger Raman signals due to enlarged Raman cross
section and resonance-enhancement.43 However, using UV laser is risky since UV photons may
photodissociate the molecules. Another way to tackle this difficulty is switching to near-IR laser
source such as 785 or 1064 nm. The photoluminescence occurs in a-C:H is due to the tail states
that formed from clusters of sp2 sites and they recombine within each cluster.44 The logic of using
longer wavelength laser is to make the particular electronic transitions unavailable since the
photons do not have high enough energy and therefore the PL background is suppressed. The
shortcoming of this method is that longer acquisition need to be compensated and potential
thermal decomposition by heat effect need to be considered especially when 1064 nm laser is
employed. To obtain decent Raman spectra of benzene nanothread, which only contains hydrogen
and carbon, 633nm Raman was used in both previous group member’s work1 and this dissertation
work.
45
2.4.3 Application of Polarized Raman Spectroscopy
Polarized Raman spectroscopy has been applied in the studies of molecular orientation,
crystalline and amorphous phases in polymers for more than four decades. Knowledge from these
studies provide insight of the relationships between microstructure and macroscopic physical
properties in polymers that benefit future designs of new materials. More recently, π-conjugated
polymers and oligomers brings a lot of interests due to their application in organic electronic
devices such as light-emitting diodes,45-48 field effect transistors49 and photovoltaic cells.50-51 The
crystallinity and the orientation of the crystalline regions in polycrystalline films highly influence
their performances when applied to the device. Therefore, polarized Raman becomes popular to
characterize the morphologies of different polymer materials as a fast and accurate method.
Beside polymers, polarized Raman has been performed in characterization of the structure
disorder,52-54 stress, compositional effects and phono confinement effects in Cu2ZnSnS4 (CZTS)
semiconductor.55
Raman tensor is an important concept in terms of polarized Raman as it represents the
differential polarizability of the molecular bond. Raman tensor is often illustrated as an ellipsoid
and larger differential polarizability will be demonstrated by shorter axis of the ellipsoid. The
shape and orientation of the ellipsoids can be determined by the measurement of angular
distribution of the polarized Raman scattering intensity.56 In order to achieve accurate polarized
Raman scattering intensity measurement, highly polarized incident laser beam is required.
Polarization filter and half-wave plate are used to control the laser polarization in nowadays
experiments. Depolarization ratio (ρ) is an important parameter to depict the shape and
orientation of Raman tensor ellipsoid and can be measured experimentally by Equation 2-3.
I
I (Equation 2-3)
46
I⊥ and I|| correspond to the Raman intensities when the polarization direction of analyzer is
normal and parallel to the incident laser beam respectively. The Raman tensor ellipsoid and
scattering geometry are illustrated in Figure 2-12.
Figure 2-12. Raman tensor ellipsoids and different polarizations. The arrows indicate the
polarization direction of the incident beam and scattered lights.56(Copyright © 2006, Springer
Science Business Media, Inc.)
The value of depolarization ratio of a Raman band is related to the symmetry of the
molecule and normal vibrational mode. Under Placzek’s polarizability approximation, totally
symmetric vibrational mode (A1 symmetry) is less than 0.75.57 This is an important piece of
information in assigning the radial breathing mode (RBM) of benzene nanothread in the first
nanothread paper.1 However, in some circumstances there are more than one factor can affect the
value of depolarization ratio therefore careful and comprehensive analysis is needed before
making any further conclusion. In the fundamental work of polarized Raman study on isolated
single wall carbon nanotube,58 orientation-dependent measurements reveals that when the
nanotubes are aligned parallel to the polarization of the incident laser beam (αi=0˚) all the Raman
modes are in the maximum intensity (Figure 2-13). This finding is significantly deviated from the
selection predicted by theory (αi=45˚). This discrepancy corresponds to two effects. One is the
extremely anisotropic property of carbon nanotube with length to diameter ration on the order of
1000. The other one is the unexpected observation of strong RBM at αi=0˚ with 633nm excitation
47
indicating the electronic resonance effects that lead to the breakdown of the Raman selection
rules.
Figure 2-13. Polarized Raman of isolated single wall carbon nanotube (SWNT). For αi=0˚ and
180˚, the polarization of the incident beam is parallel to tube axis.58 (Copyright © 2000,
American Physical Society)
The motivations of applying polarized Raman in this dissertation work include: (1) assign
and distinguish the Raman modes observed in nanothread spectrum via depolarization ratio and
selection rules; (2) observe and measure anisotropic properties of nanothread.
2.4.4 Ultra-low Frequency Raman Spectroscopy with Volume Bragg Gratings as Optical
Filter
Volume Bragg grating (VBG) is a diffractive grating produced by refractive index
modulation in the volume of photosensitive material, such as photo-thermo-reflective (PTR)
glass.59-60 Unlike silicate glass matrix which is susceptible to humidity, VBGs in PTR glass
present superior stability. More importantly, the production of refractive index modulation in
48
PTR glass is by thermal precipitation other than photo-induced process makes it stable under
whole range of irradiation.61 VBGs have been used in various application including solid-state
laser resonators,62-63 stretchers and compressors for picosecond and femtosecond lasers,64-65
mirrors for high brightness dens spectral beam combining66-67 etc. The applications of VBGs
using as optical filters emerged since 2010.68-70
There are mainly three types of Bragg gratings: transmitting Bragg grating (TBG),
reflecting Bragg grating (RBG) and chirped Bragg grating (CBG). They are different in several
aspects, such as the diffraction angle, orientation of a grating in the plate and the period
modulation. The optical filters equipped on ultra-low frequency Raman system in Badding’s
group is of the RBG type. As shown in Figure 2-14, RBGs function as a narrow band spectral
filter and angular filter at same time.61 Thus, only a narrow spectral range of the incident beam
can be reflected from the front surface and has to meet Bragg conditions, in another word, the
acceptance angle is very narrow as well. The advantage of having RBGs formed in bulk glass
makes it possible to fabricate very large numbers of grating planes with unmatched ultra-narrow
linewidth. The limit of numbers of planes in typical thin film filter is around 100 while RBG can
have more than 10000 planes in one element.
Figure 2-14. Schematics of reflecting volume Bragg grating (RBG). RBG only allows selection of
a narrow spectral region of light within an acceptance angle determined by the Bragg
conditions.61 (Copyright © 2012, SPIE)
49
The studies of vibrations in terahertz range (5~100 cm-1) have attracted broad interests in
physics and material science.70-71 As a consequent, the development of ultra-low frequency (UFL)
Raman with lower cut-off limit, cleaner spectral background is in great need. As discussed in
2.4.2, Rayleigh light is the dominant scattering that occurs at laser frequency. Therefore, it has to
be rejected as closer to the laser line as possible to gain access to range that interested vibrations
may appear. Also learnt from previous discussion, the Rayleigh scattering has to be suppressed at
least six orders of magnitude to make the weak Raman modes detectable. Currently there are
three types of filters on market: (1) typical notch filters which have cut-off frequency around 200
cm-1 and able to measure both Stokes and anti-Stokes modes in one system; (2) Edge filters that
have steeper slope than the first one which can go down to 50~100 cm-1. But it can only one side
of the laser line can be measured; (3) VBG based BragGrateTM notch filters (BNFs) which is the
filters used in our instrument can reach to ~5 cm-1 from each side of the laser line.61
Figure 2-15. Demonstration of a home-built Raman system equipped with three BragGrateTM
notch filters (BNF) as single monochromator.71 (Copyright © 2012, Springer Nature)
Figure 2-15 demonstrates a configuration of having three BNFs in a single stage
monochromatic Raman system.71 The first BNF filter has two functions including cleaning up the
incident laser beam that inject to the microscope chamber and meanwhile rejecting the Rayleigh
light reflected form the sample as notch filter. The other two BNFs behind are just used as regular
50
notch filters to further clean up the Rayleigh light before sending to spectrometer. The
combination of three BNFs can reject and more than six orders of magnitude of Rayleigh light
and enable cut-off frequency going down to 5 cm-1 from laser line.
To further reduce the spectral noise that caused by amplified spontaneous emission and
plasma line, adding a bandpass filter is preferred. Thin film filter satisfies the requirement by
cleaning up any noise stronger than -60 dB but the high cut-off (100-200 cm-1) constraints its
application in UFL Raman. In our instrument, we use one BragGrateTM Bandpass Filter (BPF) to
remove the noise down to -60~70 dB while keeping the same cut-off frequency of the BNFs by
taking the advantage that BPF has same linewidth as the notch filter. Actually BPF separates the
signal and noise by reflecting the signal instead of transmitting.61
Figure 2-16. Left: Raman spectra for C peak which corresponds to the doubly degenerate rigid
layer shear mode and G peak of graphene.71(Copyright © 2012, Springer Nature) Right: Raman
spectra of ZnTe nanoribbons at various temperatures. The sharp peaks are the multi-order LO
modes.72(Copyright © 2016, Springer Nature)
Figure 2-16 illustrates two examples of recent UFL Raman researches. The left one is the
representative work on shear mode of multilayer graphene.71 Ultra-low frequency Raman
51
technique has been largely applied in graphene than any other materials. It’s known that
longitudinal acoustic mode of polymers has been studied by low frequency Raman since late
1970s.73-76 However, the longitudinal optical (LO) mode study by Raman is relatively new. The
second example in Figure 2-16 is the UFL Raman work on resolved-sideband cooling of a LO
phonon in polar semiconductor ZnTe nanobelts.72 Resolved-sideband cooling plays an important
role in quantum ground-state preparation and coherent quantum- state manipulation.77-80
52
References
(1) Fitzgibbons, T. C.; Guthrie, M.; Xu, E. S.; Crespi, V. H.; Davidowski, S. K.; Cody, G. D.;
Alem, N.; Badding, J. V. Nat Mater 2015, 14, 43-47.
(2) Bassett, W. A. High Pressure Res 2009, 29, Cp5-186.
(3) Palmer, A.; Silevitch, D.; Feng, Y.; Wang, Y.; Jaramillo, R.; Banerjee, A.; Ren, Y.;
Rosenbaum, T. F. Review of Scientific Instruments 2015, 86, 093901.
(4) Meng, Y.; Hrubiak, R.; Rod, E.; Boehler, R.; Shen, G. Review of Scientific Instruments 2015,
86, 072201.
(5) Li, B.; Ji, C.; Yang, W. G.; Wang, J. Y.; Yang, K.; Xu, R. Q.; Liu, W. J.; Cai, Z. H.; Chen, J.
H.; Mao, H. K. P Natl Acad Sci USA 2018, 115, 1713-1717.
(6) Sharma, A.; Scott, J. H.; Cody, G. D.; Fogel, M. L.; Hazen, R. M.; Hemley, R. J.; Huntress,
W. T. Science 2002, 295, 1514-1516.
(7) https://serc.carleton.edu/NAGTWorkshops/mineralogy/mineral_physics/diamond_anvil.html.
(8) https://www.washingtonpost.com/news/speaking-of-science/wp/2017/02/13/forget-the-ring-
lab-grown-diamonds-are-a-scientists-best-friend/?utm_term=.d8365411feec.
(9) Miletich, R.; Allan, D. R.; Kuhs, W. F. Rev Mineral Geochem 2000, 41, 445-519.
(10) Besson, J. M.; Nelmes, R. J.; Hamel, G.; Loveday, J. S.; Weill, G.; Hull, S. Physica B 1992,
180, 907-910.
(11) Levitas, V.; Dushinskaya, G. Sverkhtverdye materialy 1983, 5, 7-11.
(12) Taniguchi, T.; Shimomura, O.; Yamaoka, S. Japanese journal of applied physics 1993, 32,
4012.
(13) Khvostantsev, L.; Vereshchagin, L.; Novikov, A. High Temp.-High Press 1977, 9, 637-640.
(14) Khvostantsev, L.; Slesarev, V.; Brazhkin, V. High Pressure Res 2004, 24, 371-383.
(15) Fang, J.; Bull, C. L.; Loveday, J. S.; Nelmes, R. J.; Kamenev, K. V. Review of Scientific
Instruments 2012, 83.
(16) Marshall, W. G.; Francis, D. J. Journal of Applied Crystallography 2002, 35, 122-125.
(17) Drewitt, J. W. E.; Salmon, P. S.; Barnes, A. C.; Klotz, S.; Fischer, H. E.; Crichton, W. A.
Phys Rev B 2010, 81.
(18) Loveday, J.; Nelmes, R. Physical Chemistry Chemical Physics 2008, 10, 937-950.
(19) Ward, J. H. R. Review of Scientific Instruments 1965, 36, 1376-&.
(20) Endo, S.; Mitsui, T. Review of Scientific Instruments 1976, 47, 1275-1278.
(21) Webb, A. W.; Gubser, D. U.; Towle, L. C. Review of Scientific Instruments 1976, 47, 59-62.
(22) Sakai, N.; Kajiwara, T.; Tsuji, K.; Minomura, S. Review of Scientific Instruments 1982, 53,
499-502.
(23) Feng, Y. J.; Silevitch, D. M.; Rosenbaum, T. F. Review of Scientific Instruments 2014, 85.
(24) Letoullec, R.; Pinceaux, J.; Loubeyre, P. International Journal of High Pressure Research
1988, 1, 77-90.
(25) Le Toullec, R.; Loubeyre, P.; Pinceaux, J.; Mao, H.; Hu, J. International Journal of High
Pressure Research 1992, 8, 691-696.
(26) Boehler, R.; Guthrie, M.; Molaison, J. J.; dos Santos, A. M.; Sinogeikin, S.; Machida, S.-i.;
Pradhan, N.; Tulk, C. A. High Pressure Res 2013, 33, 546-554.
(27) Sinogeikin, S. V.; Smith, J. S.; Rod, E.; Lin, C. L.; Kenney-Benson, C.; Shen, G. Y. Review
of Scientific Instruments 2015, 86.
(28) Smith, J. S.; Sinogeikin, S. V.; Lin, C.; Rod, E.; Bai, L.; Shen, G. Review of Scientific
Instruments 2015, 86, 072208.
(29) Shen, G. Y.; Mao, H. K. Rep Prog Phys 2017, 80, 1-53.
(30) Rietveld, H. M. Journal of Applied Crystallography 1969, 2, 65-&.
53
(31) Lyakhov, A. O.; Oganov, A. R.; Valle, M. Modern methods of crystal structure prediction
2010, 147-180.
(32) Wang, Y. C.; Ma, Y. M. J Chem Phys 2014, 140.
(33) Zhang, D.; Dera, P. K.; Eng, P. J.; Stubbs, J. E.; Zhang, J. S.; Prakapenka, V. B.; Rivers, M.
L. Journal of visualized experiments: JoVE 2017.
(34) https://duffy.princeton.edu/laboratory/ruby-fluorescence.
(35) Mao, H. K.; Xu, J.; Bell, P. M. J Geophys Res-Solid 1986, 91, 4673-4676.
(36) Dorfman, S.; Prakapenka, V.; Meng, Y.; Duffy, T. Journal of Geophysical Research: Solid
Earth 2012, 117.
(37) Mao, H.; Wu, Y.; Shu, J.; Hu, J.; Hemley, R.; Cox, D. Solid State Communications 1990, 74,
1027-1029.
(38) Piermarini, G.; Mighell, A.; Weir, C.; Block, S. Science 1969, 165, 1250-1255.
(39) Podsiadło, M.; Jakóbek, K.; Katrusiak, A. CrystEngComm 2010, 12, 2561-2567.
(40) Shen, G.; Chow, P.; Xiao, Y.; Sinogeikin, S.; Meng, Y.; Yang, W.; Liermann, H.-P.;
Shebanova, O.; Rod, E.; Bommannavar, A. High Pressure Res 2008, 28, 145-162.
(41) Larkin, P. Infrared and Raman spectroscopy: principles and spectral interpretation,
Elsevier: 2011.
(42) Harris, D. C.; Bertolucci, M. D. Symmetry and spectroscopy: an introduction to vibrational
and electronic spectroscopy, Courier Corporation: 1978.
(43) Liu, C.; Berg, R. W. Implementation of deep ultraviolet Raman spectroscopy. 2011.
(44) Robertson, J. Phys Rev B 1996, 53, 16302-16305.
(45) Berggren, M.; Inganas, O.; Gustafsson, G.; Andersson, M. R.; Hjertberg, T.; Wennerstrom,
O. Synthetic Met 1995, 71, 2185-2186.
(46) Kaneto, K.; Kohno, Y.; Yoshino, K.; Inuishi, Y. J Chem Soc Chem Comm 1983, 382-383.
(47) Letizia, J. A.; Facchetti, A.; Stern, C. L.; Ratner, M. A.; Marks, T. J. Journal of the American
Chemical Society 2005, 127, 13476-13477.
(48) Braun, D.; Gustafsson, G.; McBranch, D.; Heeger, A. Journal of applied physics 1992, 72,
564-568.
(49) Tsumura, A.; Koezuka, H.; Ando, T. Applied Physics Letters 1986, 49, 1210-1212.
(50) Glenis, S.; Horowitz, G.; Tourillon, G.; Garnier, F. Thin Solid Films 1984, 111, 93-103.
(51) Sicot, L.; Fiorini, C.; Lorin, A.; Raimond, P.; Sentein, C.; Nunzi, J.-M. Improvement of the
photovoltaic properties of polythiophene-based cells. Elsevier 2000.
(52) Grossberg, M.; Krustok, J.; Raudoja, J.; Raadik, T. Applied Physics Letters 2012, 101,
102102.
(53) Scragg, J. J.; Choubrac, L.; Lafond, A.; Ericson, T.; Platzer-Björkman, C. Applied Physics
Letters 2014, 104, 041911.
(54) Valakh, M. Y.; Kolomys, O.; Ponomaryov, S.; Yukhymchuk, V.; Babichuk, I.;
Izquierdo‐ Roca, V.; Saucedo, E.; Perez‐ Rodriguez, A.; Morante, J.; Schorr, S. physica status
solidi (RRL)-Rapid Research Letters 2013, 7, 258-261.
(55) Guc, M.; Levcenko, S.; Bodnar, I. V.; Izquierdo-Roca, V.; Fontane, X.; Volkova, L. V.;
Arushanov, E.; Pérez-Rodríguez, A. Scientific reports 2016, 6, 19414.
(56) Tanaka, M.; Young, R. Journal of Materials Science 2006, 41, 963-991.
(57) Banwell, C. N.; McCash, E. M. Fundamentals of molecular spectroscopy, McGraw-Hill
New York: 1994.
(58) Duesberg, G. S.; Loa, I.; Burghard, M.; Syassen, K.; Roth, S. Phys Rev Lett 2000, 85, 5436-
5439.
(59) Glebov, L. In Fluorinated silicate glass for conventional and holographic optical elements,
Window and Dome Technologies and Materials X, International Society for Optics and Photonics
2007; p 654507.
54
(60) Ciapurin, I. V.; Glebov, L. B.; Smirnov, V. I. In Modeling of Gaussian beam diffraction on
volume Bragg gratings in PTR glass, Practical Holography XIX: Materials and Applications,
International Society for Optics and Photonics 2005; pp 183-195.
(61) Glebov, A. L.; Mokhun, O.; Rapaport, A.; Vergnole, S.; Smirnov, V.; Glebov, L. B. In
Volume Bragg gratings as ultra-narrow and multiband optical filters, Micro-Optics 2012,
International Society for Optics and Photonics 2012; p 84280C.
(62) Chung, T.-y.; Rapaport, A.; Smirnov, V.; Glebov, L. B.; Richardson, M. C.; Bass, M. Optics
letters 2006, 31, 229-231.
(63) Vorobiev, N.; Glebov, L.; Smirnov, V. Optics express 2008, 16, 9199-9204.
(64) Liao, K.-H.; Cheng, M.-Y.; Flecher, E.; Smirnov, V. I.; Glebov, L. B.; Galvanauskas, A.
Optics express 2007, 15, 4876-4882.
(65) Moulton, P.; Slobodchikov, E. In 1-GW-peak-power, Cr: ZnSe laser, CLEO: Applications
and Technology, Optical Society of America 2011; p PDPA10.
(66) Sevian, A.; Andrusyak, O.; Ciapurin, I.; Smirnov, V.; Venus, G.; Glebov, L. Optics letters
2008, 33, 384-386.
(67) Andrusyak, O.; Smirnov, V.; Venus, G.; Rotar, V.; Glebov, L. IEEE journal of selected
topics in quantum electronics 2009, 15, 344-353.
(68) Glebov, A.; Mokhun, O.; Smirnov, V.; Glebov, L.; Rapaport, A.; Roussel, B.; Reich, H.-J.;
Adar, F. In Novel volume Bragg grating notch filters for ultralow-frequency Raman
measurements," The 3rd Scientific EOS Annual Meeting (EOSAM 2010), paper, 2010.
(69) Rapaport, A.; Roussel, B.; Reich, H. J.; Adar, F.; Glebov, A.; Mokhun, O.; Smirnov, V.;
Glebov, L. In Very Low Frequency Stokes and Anti‐ Stokes Raman Spectra Accessible with a
Single Multichannel Spectrograph and Volume Bragg Grating Optical Filters, AIP Conference
Proceedings, AIP 2010; pp 808-809.
(70) Ibanez, J.; Rapaport, A.; Boney, C.; Oliva, R.; Cuscó, R.; Bensaoula, A.; Artús, L. Journal of
Raman Spectroscopy 2012, 43, 237-240.
(71) Tan, P. H.; Han, W. P.; Zhao, W. J.; Wu, Z. H.; Chang, K.; Wang, H.; Wang, Y. F.; Bonini,
N.; Marzari, N.; Pugno, N.; Savini, G.; Lombardo, A.; Ferrari, A. C. Nat Mater 2012, 11, 294-
300.
(72) Zhang, J.; Zhang, Q.; Wang, X.; Kwek, L. C.; Xiong, Q. Nature Photonics 2016, 10, 600.
(73) Hsu, S.; Krimm, S. Journal of Applied Physics 1976, 47, 4265-4270.
(74) Rabolt, J.; Fanconi, B. Journal of Polymer Science Part C: Polymer Letters 1977, 15, 121-
127.
(75) Chang, C.; Krimm, S. Journal of Polymer Science Part B: Polymer Physics 1979, 17, 2163-
2170.
(76) Olf, H.; Peterlin, A.; Peticolas, W. Journal of Polymer Science Part B: Polymer Physics
1974, 12, 359-384.
(77) Diedrich, F.; Bergquist, J.; Itano, W. M.; Wineland, D. Phys Rev Lett 1989, 62, 403.
(78) Leibfried, D.; Blatt, R.; Monroe, C.; Wineland, D. Reviews of Modern Physics 2003, 75, 281.
(79) Verhagen, E.; Deléglise, S.; Weis, S.; Schliesser, A.; Kippenberg, T. J. Nature 2012, 482, 63.
(80) Spillane, S.; Kippenberg, T.; Vahala, K. Nature 2002, 415, 621.
55
Chapter 3
Mechanochemical Synthesis of Carbon Nanothread Single Crystals
Synthesis of well-ordered reduced-dimensional carbon solids with extended bonding
remains a challenge. For example, few single-crystal organic monomers react under topochemical
control to produce single-crystal extended solids. We report a mechanochemical synthesis in
which slow compression at room temperature under uniaxial stress can convert polycrystalline or
single-crystal benzene monomer into single-crystalline packings of carbon nanothreads, a one-
dimensional sp3 carbon nanomaterial. The long-range order over hundreds of microns of these
crystals allows them to readily exfoliate into fibers. The mechanochemical reaction produces
macroscopic single crystals despite large dimensional changes caused by the formation of
multiple strong, covalent C-C bonds to each monomer and a lack of reactant single-crystal order.
Therefore, it appears not to follow a topochemical pathway, but rather one guided by uniaxial
stress, to which the nanothreads consistently align. Slow-compression room-temperature
synthesis may allow diverse molecular monomers to form single-crystalline packings of
polymers, threads, and higher dimensional carbon networks.
3.1 Introduction
Although the profound kinetic stability of carbon covalent bonding yields an extraordinary
diversity of well-ordered molecular states, this versatility is not reflected in a similar abundance
of well-ordered extended carbon-based solids. Diamond, graphite, certain fullerenes, and some
polymers form macroscopic single crystals;1-3 carbon nanotubes are well-ordered along their
length, but do not form large-scale crystalline packings.4 Most other solid-state carbons are either
56
highly disordered like amorphous or nanoporous carbons or derive from known periodic systems
like graphite fluoride and intercalation compounds. Excepting the exquisitely tuned geometries of
a limited number of topochemically polymerized systems,1-2 the near-equilibrium conditions
typically used for production of compact, well-ordered solids tend to defeat the metastable
configurational diversity of the carbon-carbon bond that is so evident in organic chemistry.
Materials properties across many domains – mechanical, electrical, optical, chemical – often
depend critically on the length scale and character of structural order. For example, amorphous
silicon as a material is as distinct from the single-crystalline form as single-crystal silicon is from
germanium. Macroscopic single crystals can enable advanced applications and an understanding
of structure property relations not possible for polycrystalline or amorphous solid-state carbons.1
Here we describe a dense, one-dimensional sp3 carbon-based material with near-single-crystal
character over macroscopic length scales.
Figure 3-1. Predicted and Observed Nanothread Crystal Diffraction Patterns. a, Polytwistane
crystal taken as representative of nanothread packing, viewed down the hexagonal c-axis (threads
are parallel to it) with a 6.5 Å spacing and (100) planes 5.6 Å apart, and a side view down the b-
axis. b, X-ray diffraction (300μm CuKα1 beam) for a nanothread crystal synthesized from
polycrystalline, multiphase benzene shows a hexagonal pattern that matches c, that predicted for
the c-axis of the polytwistane crystal. d, Diffraction after 90° rotation of the nanothread crystal,
57
i.e. approximately along the b-axis, matches the pattern predicted in e, Diffraction features
marked with an asterisk in d are from the polymer loop mount.
Benzene exemplifies a large class of multiply unsaturated molecules that can polymerize
under pressure or by other means to form multiply connected, extended carbon structures.5
Unfortunately, the six new, strong, covalent C-C bonds emanating from each monomer shrink
intermolecular separations so dramatically that the products are typically amorphous6-7 and bear
no simple relation to the reactant benzene phase II molecular crystal.8 In contrast, the unit cell of
C60 molecules, which are much larger than benzene, shrinks relatively little upon the formation of
covalent intermolecular bonds, facilitating a rare topochemical reaction with minimal molecular
motion to form a single crystalline extended solid.3 The irreversibility of C-C bond formation
between molecules such as benzene provides an additional challenge in realizing crystalline
extended solids through non-topochemical routes, one that has been addressed in lower density,
porous frameworks by incorporating reversible boroxine bonds.9 However, if solid benzene is
compressed slowly at room temperature, a one-dimensional sp3-bonded product results:10 these
“carbon nanothreads,” depicted in Figures 3-1,11-14 may combine the highest specific strength
known with flexibility, insensitivity to defects, and resilience.15-16 However, the nanothread
samples reported to date have been polycrystalline, composed of multiple crystallites of different
orientations.
Here we report mechanochemical synthesis of macroscopic single crystals of nanothreads
with order over much larger dimensions, hundreds of microns, in a consistent, controlled
crystallographic orientation. A slower rate of ambient-temperature compression of crystalline
benzene monomer than previously employed (see Materials and Methods) reproducibly produces
single crystals of nanothreads aligned along a near-hexagonal c axis, also the axis of uniaxial
stress (Figure 3-1). They exhibit macroscopic striations consistent with one-dimensional character
(Figure 3-2 top) and exfoliate into fibers (Figure 3-2 bottom). Uniaxial stress thus appears to
58
select a direction of reaction within the three-dimensional benzene phase II crystal, although it is
likely that monomer crystallographic orientation and geometrical packing effects also play a
role.17 The tendency of the stress to guide a uniform direction of reaction is so strong that oriented
nanothread crystals form not only from appropriately oriented single-crystals of benzene phase II,
but from unoriented, polycrystalline mixtures of phase I and phase II as well, and in different
types of opposed-anvil pressure apparatus.
Figure 3-2. Polarization Analysis of Nanothread Samples from Paris-Edinburgh Press . (Top)
Mass of nanothreads synthesized from polycrystalline, multiphase benzene between crossed
polars. (Bottom) Nanothread fiber between cross polars. After rotation by 45°, transmission
increases greatly in the region that is thin and appropriately oriented, demonstrating the presence
of strong birefringence. Rotation to 90° reestablishes extinction.
3.2 Materials and Methods
We used both benzene and perdeuterated benzene as reactants for syntheses in the Paris-
Edinburgh press at the SNAP instrument at the Spallation Neutron Source. Benzene was frozen
59
using liquid nitrogen to avoid evaporation and loaded into encapsulated TiZr gaskets; the dry
nitrogen atmosphere from the liquid prevented condensation of water. The samples were placed
into a VX3 Paris-Edinburgh press 18 equipped with double-toroid polycrystalline diamond anvils
19 and compressed. The system was driven by an automatic hydraulic oil syringe pump, allowing
controlled pressure ramps. The pressure-load calibration curve was obtained from in situ neutron
diffraction. The sample pressure was approximately 23 GPa at the maximum oil pressure of 1,650
bar. The average compression rate was 3 GPa/hr (versus 5–6 GPa/hr in previous syntheses 10).
The sample was maintained at the maximum pressure for one hour before decompression to
ambient pressure at 4 GPa/hr.
A symmetric diamond anvil cell with a 300 μm culet was used for in situ characterization at
high pressure. Rhenium or stainless steel gaskets were indented to 40–50 μm thickness and a
100–110 μm diameter sample hole was drilled.20 Pressure was determined in situ by ruby
fluorescence. Liquid benzene was loaded into the gasket hole at room temperature. The cell was
quickly closed to avoid evaporation and water contamination. Pure phase II single crystals were
successfully synthesized in situ using resistive heating.8 Two sets of gas membranes and digital
gas controllers were employed to control the pressure rates precisely during both compression
and decompression to reproduce PE cell synthesis conditions.
In-house X-ray diffraction data on the sample recovered from the Paris-Edinburgh press were
collected with a 300 μm diameter, monochromatic copper Kα1 X-ray beam and a CCD area
detector. In situ high–resolution high pressure X-ray diffraction measurements were conducted at
beamline 16-ID-B at the Advanced Photon Source. A 30 keV 55 μm beam and both Pilatus 1M-
F and Mar CCD area detectors were used for the powder and the single-crystal diffraction
experiments. These two-dimensional diffraction patterns were reduced to one-dimensional
patterns using Dioptas21 to determine interplanar spacings vs. pressure. The only extraneous
60
reflections observed in any of the in situ diffraction patterns were due to the diamond anvils; no
diffraction from either the ruby pressure sensors or metal gaskets was observed.
Optical microscopy images were collected with an Olympus BX61 optical microscope with
crossed polarizers and a sample rotation stage. Ex situ Raman spectra were measured using a
Renishaw inVia microscope with 633 nm excitation. In situ high pressure Raman spectra were
collected in a backscattering geometry using a custom confocal spectrometer with 660 nm laser
excitation and type IIA ultrapure diamond anvils.
3.3 Results and Discussion
We compressed polycrystalline mixtures6 of benzene phases I and II (Figure 3-3) to 23 GPa
in an Paris-Edinburgh press over 8 hours, held them at pressure for 1 hour, and released them to
ambient pressure over 6 hours. Diffraction patterns collected after sample recovery to 0.1 MPa
with an X-ray beam incident along the prior compression axis reveal six-fold arcs (Figure 3-1b)
in close agreement with the prediction (Figure 3-1c) for the c axis zone [001] of a representative
array of nanothreads with 6.5 Å lattice constant; this confirms the hexagonal unit cell obtained by
tentative indexing of a sparse one-dimensional powder diffraction pattern 10 (For reference, [uvw]
indicates a real-space diffraction zone axis, while (hkl) indicates a reciprocal space diffraction
plane and {hkl} a family thereof). Multiple crystals produced in separate synthesis runs all show
a six-fold pattern (Figure 3-1b) under illumination with a 300 µm-diameter laboratory X-ray
beam, indicating that crystalline order is macroscopic. The variation in arc intensities in Figure 3-
1b suggests that the crystal tilts slightly off [001]. The form factor for a cylindrical shell of charge
as wide as a nanothread drops off rapidly with increasing scattering angle, thus the intensity of
the {100} reflections greatly exceeds that of all others in both simulation and experiment (very
weak {200} reflections are observed with a synchrotron X-ray source). Although these data
61
provide compelling evidence for a hexagonal close-packed lattice of cylinders of the same
diameter as nanothreads, they cannot discern which of the many enumerated nanothread
structures is formed, because the [001] zone does not support l≠0 reflections.14 Rotation of the
crystal by 90° from the c-axis (Figure 3-1b) reveals two-fold arcs (Figure 3-1d) that match the
{100} reflections predicted for a beam incident along the [010] axis (Figure 3-1e), but alas no
reflections with l≠0: the nanothreads are thus likely axially aperiodic (i.e. either helical, short, or
disordered along their length, and possibly lacking thread-thread registry) – long-range axial
periodicity likely awaits further advances in synthetic protocols.
The presence of short six-fold arcs rather than a uniform diffraction ring shows that the
nanothreads also order in the hexagonal a-b plane (Figure 3-1b). Polymer fibers oriented solely
by stretching/extension or shear typically exhibit fiber diffraction and thus lack this order, and in
any case the shear-driven flow in the Paris-Edinburgh press is perpendicular to the thread axis.
The 15–20° azimuthal full-width at half-maximum of the six-fold arcs (Figure 3-1b) and the
width of the two-fold arcs (Figure 3-1d) are characteristic of a crystal with a relatively large
mosaic spread. We observe single crystals with considerably smaller spreads in diamond cell
syntheses, as discussed next.
62
Figure 3-3. (Left) Neutron Diffraction Pattern of Benzene Phase I and Phase II Mixture at 2 GPa
in a Paris-Edinburgh Press upon Increasing Pressure. (Right) Synchrotron X-ray Diffraction
Pattern of Benzene Phase I and Benzene Phase II Mixture at 14 GPa upon Increasing Pressure in
a Diamond Anvil Cell. The red curves in both patterns are Pawley fit and the blue curves are the
residual.
To gain insight into reaction conditions, synchrotron diffraction data in situ in a diamond
anvil cell were acquired on polycrystalline mixtures of benzene phase I and phase II using a 5×5
μm beam of 0.4067 Å wavelength (Figure 3-3).8 Reaction begins at 16–19 GPa; i.e. we recovered
no nanothreads from samples compressed to only 16 GPa. Compression to 23 GPa reproducibly
yields six-fold diffraction patterns which persist upon release of pressure. These patterns can be
indexed to a monoclinic cell with a unique crystallographic c-axis (the single cell axis with
highest rotational symmetry) angle of 117°, i.e. very close to hexagonal. For convenience, we
refer to this cell as “pseudohexagonal” and the six narrow arcs as a symmetry-related set {100}.
The threads again align closely to the compression axis, which is again parallel to the x-ray beam.
These arcs (8–12° FWHM) are narrower in azimuth than those of the Paris-Edinburgh samples,
indicating a smaller mosaic spread in the a-b plane across the sampling volume of the 5×5 μm x-
ray beam. They also have narrower 2θ diffraction linewidths. The improved order may be due to
larger uniaxial stress in the diamond anvil cell.
Modeling suggests that nanothreads are much stiffer than polymers,22 with persistence lengths
of ~100 nm.15 Consistent with an ordered packing of aligned, stiff threads bound by van der
Waals forces (and unlike conventional hydrocarbon polymers or polycrystalline materials),
nanothread crystals can be mechanically exfoliated into fibers (Figure 3-2 Bottom) just as
graphite can be exfoliated ultimately to graphene. Exfoliated fibers are birefringent, consistent
with the optical anisotropy expected from such a structure. Exfoliation into single threads may
provide nanoscale building blocks with tunable strength and ductility that can be re-assembled
and/or functionalized to control properties and processability.16
63
It is remarkable that macroscopic aligned single crystals of nanothreads emerge from
polycrystalline mixtures of benzene phases I and II compressed in both PE and diamond cells in
which each grain has a complex three-dimensional crystal structure (the diamond cell diffraction
patterns confirm that multiple phase I and phase II crystallites are in the x-ray beam before
reaction; similarly, neutron powder diffraction from two phase benzene samples compressed in
PE cells that indicate they are polycrystalline). The reverse of this process – i.e. single crystals
breaking up into polycrystals due to reaction-induced stresses – is much more common.7 Benzene
phase II hosts three crystallographically distinct nearest-neighbor stacks of molecules (reaction
occurs in Phase II – see below). One or more of these stacks may have a geometry that allows
collapse into threads by a radical or cycloaddition mechanism.13 In appropriately aligned
crystallites, uniaxial stress will preferentially bring the molecules in these stacks closer together.
Once initiated, reaction would most easily propagate along the uniaxial stress axis, aided by
packing constraints similar to those that operate in nematic liquid crystals. The benzene stacks
must contract dramatically along the thread axis (by 40–50%) as nanothreads form, indicating
that the reaction cannot be topochemical and further underlining the remarkable fact that well-
ordered one-dimensional crystals can be obtained from an uncatalyzed room-temperature reaction
involving carbon-carbon bonds. Any crystallites not oriented to react would simply melt upon
pressure release; we observe benzene liquid coexisting with the solid nanothreads upon recovery
before it evaporates. Remarkably, the order observed in the a-b plane (Figure 3-1b) appears to
propagate through the multiple crystallites selected for reaction. Single-crystal PbS nanosheets
have been synthesized by uniaxial stress applied to unaligned packings of nanoparticles,
suggesting similar guiding roles for uniaxial stress in the two cases.23 Speaking more
speculatively, the strong kinetic preference for the single lowest-barrier pathway that is implied
by our slow compression at low (i.e. room) temperature may also, ironically, facilitate formation
64
of well-ordered nanothread crystals by suppressing less favorable reactions that e.g. cross-link the
nascent threads.
To further probe the importance of crystal alignment in nanothread formation, we produced
crystals of pure phase-II benzene inside a diamond anvil cell by annealing and analyzed the
resulting system in situ under pressure by the rotating crystal method.8, 24 Step rotation and
continuous wide-angle rotation two-dimensional synchrotron x-ray diffraction patterns were
collected on crystals that were larger than the 5×5 μm beam between –15° and 15° of ω rotation
angle about a vertical axis perpendicular to the horizontal synchrotron beam (the range being
limited by the diamond cell geometry).24 At 3.3 GPa, we observed a phase II single crystal
(Figure 3-4a).
Upon compression to 23 GPa, five new diffraction spots characteristic of a nanothread crystal
were observed at ω=15°, reasonably close to the axis of compression (Figure 3-4b); we index
these spots with a pseudohexagonal cell and a unique c-axis angle of 117°.25 These results show
that nanothread synthesis does not require Phase I, which in any case lacks suitable benzene
contacts. The “slipped” molecular stacks along the a, b, and b-c axes of benzene phase II (shown
in Figure 3-4c) are instead the most likely candidates for intra-stack reaction.13 To assess the
spatial relationships between the benzene phase II and nanothread crystals, we determined their
orientation matrices at 23 GPa.24 The nanothread crystal that forms has its pseudohexagonal c
axis oriented 15° from the axis of compression and close to midway between the b and bc stacks
of the Phase II crystal (about 27° from either), as shown in Figure 3-4d.
Reaction along the a axis can thus be ruled out, since it is nearly orthogonal to the
nanothread c axis, while the b and b-c stacks of benzene phase II remain viable candidates for
nanothread formation.13 In accord with this conclusion, we observe no nanothread diffraction
65
peaks for a different phase II single-crystal sample that had its c axis, which has no stacks suitable
for reaction,13 parallel to the axis of compression and the x-ray beam.
Figure 3-4. In situ Nanothread Diffraction at High Pressure and Deduced Spatial Relationships a,
Indexed diffraction pattern collected upon increasing pressure to 3.3 GPa for a phase-II single
crystal. Large dark spots are diamond anvil reflections. b, Expanded view of diffraction pattern
upon increasing pressure to 23 GPa for phase-II single crystal showing psuedohexagonal carbon
nanothread {100} peaks beginning to appear. The inset shows the benzene (100) peak
neighboring the nanothread (100) peak. (010) is >30 counts above background, though difficult to
see in the figure. c, a, b, and b-c columns of benzene molecules along which reaction to form
nanothreads is considered; unit cell in blue. d, Stereographic projection of benzene phase-II
66
diffraction planes (hkl) and zone axes [uvw], showing the angular relationship at 23 GPa of the
nanothread crystal [001] c-axis to the benzene crystal [100] a, [010] b, and [011] bc axes. Only
the directions and planes on the “north” side of the projection are shown (e.g., [100] is on the
south side, 180 degrees from [1 ̅00]). e, The monoclinic benzene crystal viewed down the unique
b-axis, showing diffraction planes and their interplanar spacings at 23 GPa with the orientation
and spacing of diffraction spots in the inset. Diffraction planes for the grey spots are not shown. f,
A monoclinic (pseudohexagonal, γ=117°) nanothread crystal viewed down the c axis with
interplanar spacings at 23 GPa. Diffraction planes have the same scale as d. Note the slight
expansion required for the benzene (100) planes to form nanothread (100) planes and the larger
shifts required for planes at angles to (100).
As the nanothreads emerge by reaction along a suitable benzene column, a plausible origin of
the increase in symmetry from monoclinic benzene to the pseudohexagonal nanothread crystal is
the close-packing of extended objects with near-cylindrical symmetry in two dimensions. The
candidate b-axis benzene stacks must undergo a large rearrangement to form hexagonally-packed
nanothreads: for example, the cross-section perpendicular to b must expand by ~45% along its
short axis to match hexagonal packing, as shown in Figure 3-4 e,f. The spacing of the candidate
b-c stacks must change greatly as well (Figure 3-5) as must any other (less likely) reaction stack.
Such large geometric changes are not typical of topochemical reactions.26
Upon lowering the pressure on the single-crystal sample, the sixth remaining nanothread
reflection appears (it was likely not within the range of available ω angles at 23 GPa), and all the
reflection intensities increase considerably (Figure 3-6). As the pressure drops, either further
reaction to form more threads occurs or the nanothreads that have already formed optimize their
hexagonal packing, or both. These diffraction arcs are quite narrow (Figure 3-6).
Although the existing diffraction data cannot uniquely determine the atomic-scale structure
along the thread axis, the high level of precision in the observed pseudohexagonal lattice
parameters can be combined with careful first-principles density functional theory and empirical
molecular dynamics to significantly constrain which of the candidate nanothread bond topologies
are consistent with the experimental packing.27 As shown in Figure 3-7a, simulations reveal that
many of these candidates are not consistent with the experimental {100} interplanar spacings, in
67
this case those measured at 0.1 MPa for the sample that was synthesized from polycrystalline
benzene (when shrunk by 1–3% to account for thermal contraction to the T=0 calculations28-29).
Of those thread structures that are consistent with the experimental lattice parameters,
polytwistane (143652) and tube (3,0) (123456) both have a single symmetry-distinct carbon site,
although on-thread structural disorder could increase this number; others such as rotationally
ordered achiral-3 or achiral-5 have multiple such sites, and expansion of the analysis to include
degree 4 threads could yield additional options. Several other thread structures examined are too
loosely packed along at least one dimension.
Figure 3-5. Monoclinic Phase II Benzene Crystal Viewed Down The [011] b-c Axis. Some
diffraction planes and their interplanar spacings at 23 GPa are shown, with the orientation and
spacing of the corresponding diffraction spots in the inset. Diffraction planes for the grey spots
are not shown. The smallest amount of expansion is needed for the (100) planes to form the
psuedohexagonal or hexagonal nanothread structure (Figure 3-7d). Other sets of planes such as
the (11-1)and (01-1) must expand to a much larger extent. The b-c stacks have the columns with
alternating orientations of benzene molecules different from the benzene columns along the b
axis, which have uniform orientation. Both types of stacks are shown in Figure 3-7b. In contrast
to the b stacks, the b-c stacks are not suited for [4+2] cycloaddition, but “para polymerization”
starting with a diradical may be possible. 22
68
Based on the optimized packing geometries, we calculated the lattice parameters of tube (3,0)
and polytwistane within density functional theory from 0 to 24 GPa and then back to zero, with
one thread per near-hexagonal unit cell (Figure 3-7b). The compressibilities of both candidate
threads match reasonably well to the experimental decompression curves, keeping in mind that
the experimental data includes an estimated 1–3% thermal expansion from 0 to 300K. Modest on-
thread structural disorder that does not substantially change the local diameter of the thread is
unlikely to significantly degrade this agreement. One possible discrepancy is a steeper upslope of
the experimental curves below 3 GPa, suggesting that slight imperfections in the experimental
nanothread packing may be squeezed out by moderate pressure; taking this into account then
favors the more densely packed options in Figure 3-7a. The weaker nature of this discrepancy in
the single-phase sample of is consistent with this scenario.
69
Figure 3-6. Diffraction Patterns of Nanothreads Synthesized from Single Crystals upon
Decreasing Pressure Along [001] Zone Axis. (a) Five of six nanothread reflections can be
observed in a diffraction pattern at 23 GPa, along with the benzene (100) neighboring the
nanothread (100) and the benzene (-100) neighboring the nanothread (-100). (b) Diffraction
pattern at 19 GPa after reducing the pressure from 23 GPa. All six nanothread reflections can be
observed and are somewhat more intense. The relative lower intensity observed for the two
rightmost nanothread reflections is likely associated with a tilt away from the [001] zone axis, as
these diffraction images were collected at the maximum ω angle of 15°. The increase of the
reflection intensities as the pressure is reduced is likely associated either with more nanothread
formation or better ordering of threads that have already formed. (c) Evolution of Benzene and
Nanothread Diffraction Patterns with Pressure. One-dimensional diffraction patterns vs. pressure
for the nanothread sample synthesized from pure single-crystal phase II benzene as a function of
pressure. Down-arrows indicate decompression.
Figure 3-7. Constraining Nanothread Structure with Experimental and Calculated {100}
Interplananar Spacings. a, Interplanar spacings calculated for candidate achiral and chiral
nanothread structures at T=0, including both ordered and disordered axial/azimuthal
arrangements at empirical and first-principles levels, to show which ones are consistent with the
measured interplanar spacings in the nanothreads synthesized from mixed phase I/Phase II
benzene. The anticipated shifts in the experimental data due to thermal contraction to T=0 are
depicted by orange bars, and the full-width-at-half-maximum over 10 instances of calculated
disorder are marked by vertical blue bars. Experimental nanothread {100} interplanar spacings at
70
ambient pressure are from sample synthesized in a diamond cell from polycrystalline, multiphase
benzene. b, Nanothread (100) interplanar spacings during decompression. The threads again
formed from a polycrystalline mixture of phase I and phase II in a diamond cell. Calculated first
principles {100} interplanar spacings derived from lattice parameters are also shown for two
representative nanothread crystals.
3.4 Conclusions
The robust, reproducible formation of nanothread crystals from benzene suggests that other
unsaturated molecular crystals that are unsuitable for topochemical reaction may nevertheless
form single-crystalline packings of polymers or nanothreads upon slow uniaxial compression. For
example, we have recently synthesized single crystals of carbon/nitrogen nanothreads from
pyridine, which has a different crystal structure from benzene.30 Slow, cold compression may thus
provide a general route to synthesizing single-crystalline packings of low-dimensional fully or
partially saturated carbon-based nanomaterials with extended bonding. Uniaxial stress applied
sequentially along different directions, with each step performing only partial saturation, could
create higher-dimensional networks. The diversity of fully (degree 6) and partially (degree 4)
saturated nanothread structures enumerated theoretically for benzene further suggests that
multiple structures might be accessible from a single precursor molecule through appropriate
control of pressure, stress, temperature, photochemistry, catalysts, and chemical initiators. These
same parameters could also be exploited to reduce reaction pressure,10, 31-33, as could precursors
with reduced aromaticity. If the synthesis pressure could be reduced to ~6 GPa (at which >106
kg/yr of diamond abrasive is manufactured),34 production could reach an industrial scale.
Above is adapted with permission from J. Am. Chem. Soc. 2017, 139, 16343–16349.
Copyright (2017) American Chemical Society.
71
3.5 Insights and Future Consideration
3.5.1 Sign of Reaction
When there is a novel material being successfully synthesized with promising physical
and chemical properties, it always attracts people to reveal the reaction mechanism. Because a
good understanding of reaction will help to further optimize the synthesis protocol thus increase
the reaction yield and sample quality. On the other hand, if the mechanism turns to be universal, it
will give birth to series of new materials with tunable functionalization. The current work is
standing the very early stage of learning the mechanism but there are few insights that would be
practically useful in conducting in situ experiments in the future.
Figure 3-8 (left) shows the in situ X-ray Diffraction of benzene phase II (100) and
nanothread (100) interplanar spacings vs. pressure in increasing and decreasing pressure
directions in a diamond anvil cell. These diffraction planes are nearly parallel to both the incident
x-ray beam and the axis of compression. They are thus impacted by the stress perpendicular to
the axis of compression in the plane of the diamond culet. After the pressure is raised above ~1
GPa, the sample is a polycrystalline mixture of benzene phase I and phase II. In the increasing
pressure direction benzene phase II (100) diffraction arcs (black open circles) exhibit a small
discontinuity at 18 GPa, decreasing in interplanar spacing. New diffraction arcs at a slightly
larger interplanar spacing than the benzene (100) ones also appear at 10 GPa in the increasing
pressure direction (red open circles). It is not yet certain whether these arcs are associated with a
nanothread reaction intermediate or nanothreads themselves. At 18 GPa there is a sudden
expansion of these new arcs, which appears to be correlated with the reaction to form
nanothreads. After this expansion at 18 GPa upon increasing pressure, the diffraction peaks can
be traced up to 23 GPa and back to ambient pressure (solid red triangles) in the decreasing
72
pressure direction. Upon release of pressure and melting of the solid benzene a six-fold
symmetric diffraction pattern characteristic of nanothreads remains. The benzene phase II (100)
diffraction arcs in the decreasing pressure direction (solid black triangles) diverge at about 14
GPa from those in the increasing pressure direction such that there is hysteresis. It may be that
after benzene has collapsed into nanothreads, there is slightly less stress in the plane
perpendicular to the x-ray beam, which causes this hysteresis. Figure 3-8 (right) is the Raman
spectra of benzene vs. pressure during Compression. An increase in background
photoluminescence was observed in the Raman spectrum of polycrystalline, multiphase benzene
when the pressure reached 19 GPa. The pressure at which the background increases correlates
with the onset of new diffraction peaks associated with nanothread formation in the
polycrystalline phase I/II samples. Prior investigations reported that oligomers of benzene formed
after benzene was compressed to 16 GPa. Thus, the background here may arise from these
relatively low molecular weight benzene oligomers. Benzene scattering and/or background
photoluminescence make it difficult to observe nanothread Raman in situ at high pressure.
Figure 3-8. In situ X-ray diffraction and Raman spectra of benzene in diamond anvil cell.
73
From the observations above, there are two important pressure regions during
compression need to be take special care. One is the 10 GPa which the arcs first appear that later
turned to nanothread (100) and the other is the ~18 GPa when the usual expansion of (100)
interplanar spacing and the significant photoluminescence background happen. Applying slow
compression rates above ~18 GPa or slightly lower will favor the formation of nanothread instead
of amorphous products. Even slower rates within this pressure range will help single crystalline
nanothread growth. Therefore, when dealing with a new small molecule, finding the onset
pressure of strong photoluminescence background will be essential and slower rates should be
applied at least from this point. If the conditions allow, in situ X-ray diffraction experiments are
strongly preferred and special attention should be taken below that pressure.
3.5.2 Polarized Raman Spectroscopy of Carbon Nanothread
By now, the structural identification is still under thread level. More efforts are needed
such as high resolution TEM study on isolated nanothread, more comprehensive diffraction
measurements of recovered sample in multiple rotation geometries and further interpretation of
vibration spectra to nail down the atomic configuration of the threads. The most recent solid state
nuclear magnetic resonance (NMR) work35 proof the existence of degree-4 nanothread13 by
having one unsaturated double bond in each benzene ring. This result will provide a new clue for
Raman spectra interpretation especially for sp2 carbons. The optimized benzene nanothread
sample shows less photoluminescence background compared with the originally synthesized
samples as presented. Therefore, there are more peaks that allow for polarized Raman
measurement and analysis as shown in Figure 3-9.
74
Figure 3-9. Raman spectra of benzene nanothread measured under different polarization.
It’s worth noting that there are more than one peak near 800 cm-1 and they all have very
low depolarization ratio which indicate to be polarized modes. The peak at ~800 cm-1 was
identified as radial breathing mode in the first paper. Since we now know that degree-4
nanothread also present within the sample, more simulated vibrational spectra should be
compared beside fully saturated structures. Another peak that could be important is the sharp
peak positioned at 1002 cm-1 as it may reveal some information of backbones as this range
usually associated with the skeletal C-C stretch modes in polymers. Moreover, the Raman
features of nanothread in the range of 1500 cm-1 to 1700 cm-1 and 2800 cm-1 to 3000 cm-1 are
more complicated than diamond-like carbons thus they cannot be simply assigned by G peak and
one sp3 C-H stretching mode as a whole. More detailed analysis of experimental Raman modes
requires more comprehensive modeling for comparison and applying the knowledge from NMR
as reference. An angular dependent polarized Raman spectroscopy will benefit the understating of
depolarization ratio as the example of single wall nanotube discussed in Chapter 2. Nanothread
75
has one dimensional structure as well and this anisotropy character will influence the values
depolarization ratio so that they cannot be directly used for vibrational modes symmetry
determination. Because there are small portions of unknown oligomers detected in NMR. Another
procedure may help to simplify the spectra is to rinse the oligomer with organic solvent. Utilizing
isotope labeling is another way to identify the vibrations. Limited to experiment condition, only
deuterated benzene nanothread had been synthesized. For a more completed screening, 13C
labeled sample is needed.
3.5.3 Preliminary Ultra-Low Frequency Raman Study
In Chapter 2, an ultra-low frequency Raman spectrometer equipped with Bragg Grating
optical filters was introduced. This home-made Raman enable the measurement down to ~5 cm-1.
Only preliminary result has been obtained as presented below (Figure 3-10). There is a pair of
Raman peaks show up around ± 20 cm-1. It’s the first spectrum that ultra-low frequency Raman
modes observed in nanothread sample. Even though the assignment of this mode is unknown, it’s
still encouraging as amorphous polymer samples do not have distinct peaks in this region, but
only crystalline samples do.36-37 The vibrational modes at such low frequency are associated with
the larger-scale movement such as shear modes between graphene layers, characteristic mode of
“opened” structures of metal organic framework.38 Therefore this mode could be related to
relative interaction between threads or longitudinal acoustic mode of the threads. Learning the
behavior of this mode by cooling, heating even by applying some pressure would be helpful.
76
Figure 3-10. Ultra-low frequency Raman spectrum of benzene nanothread synthesized in PE cell.
There are another two features positioned at -4.6 cm-1 and 5.6 cm-1 respectively. It’s
difficult to make a solid claim at this point if they are real Raman modes or due to artificial effect
since they are very close to the laser line. More delicate alignment of this set-up is necessary for
future studies including polarized Raman measurement of this range.
77
References
(1) Dou, L. T.; Zheng, Y. H.; Shen, X. Q.; Wu, G.; Fields, K.; Hsu, W. C.; Zhou, H. P.; Yang, Y.;
Wudl, F. Science 2014, 343, 272-277.
(2) Lauher, J. W.; Fowler, F. W.; Goroff, N. S. Accounts Chem. Res. 2008, 41, 1215-1229.
(3) Yamanaka, S.; Kini, N. S.; Kubo, A.; Jida, S.; Kuramoto, H. J. Am. Chem. Soc. 2008, 130,
4303-4309.
(4) Thess, A.; Lee, R.; Nikolaev, P.; Dai, H. J.; Petit, P.; Robert, J.; Xu, C. H.; Lee, Y. H.; Kim,
S. G.; Rinzler, A. G.; Colbert, D. T.; Scuseria, G. E.; Tomanek, D.; Fischer, J. E.; Smalley, R. E.
Science 1996, 273, 483–487.
(5) Bini, R.; Schettino, V. Materials under extreme conditions : molecular crystals at high
pressure, Imperial College Press: London, 2014.
(6) Ciabini, L.; Santoro, M.; Gorelli, F. A.; Bini, R.; Schettino, V.; Raugei, S. Nat. Mater. 2007,
6, 39–43.
(7) Mcbride, J. M.; Segmuller, B. E.; Hollingsworth, M. D.; Mills, D. E.; Weber, B. A. Science
1986, 234, 830-835.
(8) Piermarini, G. J.; Mighell, A. D.; Weir, C. E.; Block, S. Science 1969, 165, 1250–1255.
(9) Diercks, C. S.; Yaghi, O. M. Science 2017, 355, 923-931.
(10) Fitzgibbons, T. C.; Guthrie, M.; Xu, E. S.; Crespi, V. H.; Davidowski, S. K.; Cody, G. D.;
Alem, N.; Badding, J. V. Nat. Mater. 2015, 14, 43–47.
(11) Maryasin, B.; Olbrich, M.; Trauner, D.; Ochsenfeld, C. J. Chem. Theory Comput. 2015, 11,
1020–1026.
(12) Olbrich, M.; Mayer, P.; Trauner, D. J Org Chem 2015, 80, 2042–2055.
(13) Chen, B.; Hoffmann, R.; Ashcroft, N. W.; Badding, J.; Xu, E. S.; Crespi, V. J Am Chem Soc
2015, 137, 14373–14386.
(14) Xu, E. S.; Lammert, P. E.; Crespi, V. H. Nano Lett 2015, 15, 5124–5130.
(15) Roman, R. E.; Kwan, K.; Cranford, S. W. Nano Lett. 2015, 15, 1585–1590.
(16) Zhan, H.; Zhang, G.; Tan, V. B. C.; Cheng, Y.; Bell, J. M.; Zhang, Y.-W.; Gu, Y. Nanoscale
2016, 8, 11177–11184.
(17) Halasz, I. Cryst. Growth Des. 2010, 10, 2817-2823.
(18) Klotz, S.; Hamel, G.; Frelat, J. High Pressure Res 2004, 24, 219–223.
(19) Fang, J.; Bull, C. L.; Loveday, J. S.; Nelmes, R. J.; Kamenev, K. V. Rev Sci Instrum 2012,
83, 093902.
(20) Hrubiak, R.; Sinogeikin, S.; Rod, E.; Shen, G. Y. Rev Sci Instrum 2015, 86, 072202.
(21) Prescher, C.; Prakapenka, V. B. High Pressure Res 2015, 35, 223–230.
(22) Stojkovic, D.; Zhang, P. H.; Crespi, V. H. Phys Rev Lett 2001, 87, 125502.
(23) Wang, Z. W.; Schliehe, C.; Wang, T.; Nagaoka, Y.; Cao, Y. C.; Bassett, W. A.; Wu, H. M.;
Fan, H. Y.; Weller, H. J. Am. Chem. Soc. 2011, 133, 14484-14487.
(24) Boldyreva, E.; Dera, P. High-Pressure Crystallography From Fundamental Phenomena to
Technological Applications, Springer: 2010.
(25) Hahn, T. International Tables for Crystallography Vol. A: Space Group Symmetry, 5th
ed.;Wiley: Hoboken, 2005.
(26) Scheffer, J. R.; Scott, C. Science 2001, 291, 1712–1713.
(27) Plimpton, S. J. Comput. Phys. 1995, 117, 1–19.
(28) Breitling, S. M.; Jones, A. D.; Boyd, R. H. J Chem Phys 1971, 54, 3959–3964.
(29) Lee, W. Y.; Slutsky, L. J. J Phys Chem-Us 1975, 79, 2602–2604.
(30) Li, X.; Wang, T.; Duan, P.; Baldini, M.; Huang, H.-T.; Chen, B.; Juhl, S. J.; Koeplinger, D.;
Crespi, V. H.; Schmidt-Rohr, K. J Am Chem Soc 2018, 140, 4969-4972.
78
(31) Bini, R. Accounts Chem. Res. 2004, 37, 95–101.
(32) Ciabini, L.; Santoro, M.; Bini, R.; Schettino, V. Phys. Rev. Lett. 2002, 88, 085505.
(33) Ciabini, L.; Gorelli, F. A.; Santoro, M.; Bini, R.; Schettino, V.; Mezouar, M. Phys. Rev. B
2005, 72, 094108.
(34) Sparks, J. R.; He, R. R.; Healy, N.; Chaudhuri, S.; Fitzgibbons, T. C.; Peacock, A. C.; Sazio,
P. J. A.; Badding, J. V. Adv. Funct. Mater. 2013, 23, 1647-1654.
(35) Duan, P.; Li, X.; Wang, T.; Chen, B.; Juhl, S. J.; Koeplinger, D.; Crespi, V. H.; Badding, J.;
Schmidt-Rohr, K. J. Am. Chem. Soc. 2018.
(36) Smith, G. P.; Huff, G. S.; Gordon, K. C. Spectroscopy 2016, 31, 42-50.
(37) Spells, S.; Shepherd, I. The Journal of Chemical Physics 1977, 66, 1427-1433.
(38) Krylov, A.; Vtyurin, A.; Petkov, P.; Senkovska, I.; Maliuta, M.; Bon, V.; Heine, T.; Kaskel,
S.; Slyusareva, E. Physical Chemistry Chemical Physics 2017, 19, 32099-32104.
79
Chapter 4
Carbon Nitride Nanothread Crystals Derived from Pyridine
Carbon nanothreads are a new one-dimensional sp3 carbon nanomaterial. They assemble
into hexagonal crystals in a room temperature solid-state reaction induced by slow compression
of benzene to 23 GPa, followed by slow decompression. These crystals form despite the large
changes in structural geometry that occur when three covalent bonds per benzene monomer form;
such non-topochemical transformations typically do not yield crystalline or-der. Here we show
that pyridine also reacts under compression to form a well-ordered sp3 product: C5NH5 carbon
nitride nanothreads, in both µg and mg scale. Solid pyridine has a different crystal structure from
solid benzene, suggesting that the non-topochemical formation of low-dimensional crystalline
solids by slow compression of small aromatics is a general phenomenon and that tuning of
nanothread properties through organic chemistry is possible. The nitrogen in the carbon nitride
nanothreads may improve their processability, alters their photoluminescence, and is predicted to
reduce their bandgap.
4.1 Introduction
Relative to the vast corpus of organic molecules, we have few examples of extended
carbon networks: diamond, graphite, graphene, and carbon nanotubes. Nevertheless, they have
attracted much interest due to their superlative thermal, electronic, and mechanical properties and
symmetric geometries. Theory reveals an unrealized structural diversity for extended carbon
solids: more than 500 new carbon networks have been predicted.1 Unfortunately, the well-
developed methods of synthetic organic chemistry generally cannot be applied to extended carbon
80
solids because they are insoluble. Recently we showed that polycrystalline, solid benzene
transforms under pressure into single-crystalline packings of carbon nanothreads.2 These
nanothreads are sp3-bonded and one-dimensional3-4 and thus occupy a distinct position in a matrix
of hybridization (sp2/sp3) and dimensionality (0D/1D/2D/3D) for carbon nanomaterials.5 Fully
saturated degree-6 nanothreads6-7 could exhibit a unique combination of strength, flexibility, and
resilience,8-9 while partially saturated degree-4 threads may act as novel organic conductors.2-3
Moreover, the transformation from polycrystalline benzene to nanothread crystal cannot be
topochemical10 because it involves large changes in volume and symmetry as multiple short,
strong carbon-carbon bonds form between each benzene pair.2 Such non-topochemical reactions
usually disrupt single-crystalline order and often yield amorphous solid products, but the benzene
nanothread reaction instead creates single-crystalline order. Single crystals free from grain
boundaries are very desirable for many applications and can facilitate understanding of structure-
property relations.2 The requirement for commensuration between reactant and product structures
for topochemical reactions10 severely constrains the number of suitable monomers; freedom from
this requirement could allow for chemical design through solid-state organic synthesis of
crystalline extended networks from new types of monomers.
81
Figure 4-1. Carbon Nitride Nanothread Structures (a) View down the hexagonal c axis and (b)
view perpendicular to the c axis of a degree-6 carbon nitride nanothread tube (3,0)_123456. (c)
Degree-4 carbon nitride nanothread structure (IV-7) with C-N double bonds in red.
Here we show that slow compression and decompression of polycrystalline pyridine
allows for the formation of single-crystal packings of carbon nitride nanothreads (Figure 4-1 a,b)
with empirical formula C5NH5. Notably, the crystal structure of the molecular pyridine reactant,
although not yet fully elucidated,11 is different from that of the benzene reactant,2 demonstrating
the independence of the synthesis from the precise crystalline packing of the precursor. Chemical
substitution of heteroatoms such as nitrogen into carbon extended networks is generally difficult
because nitrogen substituting carbon is not isoelectronic; carbon nitride nanothreads, in contrast,
are isoelectronic to carbon nanothreads, just as pyridine is isoelectronic to benzene. (This
isoelectronicity also suggests that nanothreads with different C:N stoichiometries should be
possible, in contrast to the more fixed 3:4 ratio of β-C3N412-14). Room-temperature reaction such
as used for nanothreads is more likely to allow for heteroatom incorporation than are the closer-
to-equilibrium high-temperature conditions often used for carbon networks, since C-N single
bonds are weaker than C- C single bonds.14
The isomeric possibilities and nomenclature for substituted nanothreads with chemical
formula ((CH)5X)n (X=heteroatom or -CR substituent) have been explored systematically with a
symmetry-conditioned permutational constraint.15 For the isomers of carbon nitride nanothreads
with chemical formula ((CH)5N)n we adopt the naming convention of the enumerated carbon
nanothreads4 followed by an integer string that indicates the positions of nitrogen atoms in every
six-membered progenitor (i.e. molecular) ring. For example, isomer tube (3,0)_123456 can be
generated from the tube (3,0) carbon nanothread where the C-H pairs on positions 1, 2, 3, 4, 5 and
6 of six successive rings in the stack are substituted with N atoms, as shown in Figure 4-1a,b.
(See Figure 4-2 for ring position numbering). Degree-4 unsaturated nanothreads have one double
bond per pyridine formula (Figure 4-1c) while degree-2 threads have two.6
82
Figure 4-2. (a) Few examples of carbon nitride nanothread structures from theoretical
enumeration and used for diffraction pattern comparison. (b) Ring numbering for four types of
benzene nanothreads. The orange-numbered ring denotes the first ring; the blue-numbered ring
below is the second ring and so on. The third ring in tube (3,0) is the same as the first ring
because of symmetry and is numbered in orange. The nomenclature tube (3,0)_1245 means this
pyridine nanothread is based on the structure of the tube (3,0) benzene nanothread, substituting N
atoms for the C-H bonds at position 1 in 1st ring, position 2 of 2nd ring, position 4 of the 3rd ring
and position 5 of the 4th ring; the repeat unit contains four rings.15
83
4.2 Materials and Methods
4.2.1 Synthesis
A symmetric diamond anvil cell (DAC) with a 300 μm culet was used for synthesis of
carbon nitride nanothreads. Stainless steel gaskets were indented to 40–50 μm thickness and a
100–110 μm diameter sample hole was drilled.16 Pressure was determined in situ by ruby
fluorescence. Liquid pyridine (anhydrous, 99.8% from Sigma-Aldrich) was loaded into the gasket
hole at room temperature. After compression to ~1 GPa, pyridine froze to a solid as previously
reported.17-18 Two sets of gas membranes and digital gas controllers were employed to control the
rates of compression and decompression precisely. Both compression and decompression
proceeded over 8–10 hours between near-ambient pressure and 23 GPa. The compression rate
was reduced to 2–3 GPa/h above 14 GPa and slowed again to 0.6–1.2 GPa/h above 19 GPa, with
similar rates for decompression.
We used 15N pyridine (C515NH5) and deuterated pyridine (C5ND5) as reactants for larger-
scale syntheses in a Paris-Edinburgh press. Pyridine was loaded into encapsulated stainless-steel
gaskets and TiZr null-scattering alloy gasket (for the deuterated pyridine synthesis run that was
loaded with glass wool). Liquid nitrogen was used to freeze the liquid pyridine into a solid to
ensure that the gaskets were fully filled without trapped atmosphere; the evaporated nitrogen gas
further helped to exclude oxygen and water from the atmosphere. The samples were then placed
into a VX3 Paris-Edinburgh press19 equipped with double-toroid polycrystalline diamond anvils.20
The system was driven by an automatic hydraulic oil syringe pump, allowing for controlled
pressure ramps. The pressure-load calibration curve was obtained from in situ neutron diffraction.
Both compression and decompression took ~8 hours, as for the diamond anvil cell experiments.
The sample pressure was approximately 23 GPa at the maximum oil pressure of 1,650 bar. When
84
the oil pressure reached 1500 bar, a slow rate of increase (1bar/min) was employed in both
compression and decompression.
4.2.2 X-ray Diffraction
The sample recovered to ambient from the DAC was investigated at the 16-BM-D X-Ray
beamline at the Advanced Photon Source with a 20 keV 5×5 μm beam and Mar 345 image plate
area detector. Both powder and single-crystal high-resolution high-pressure x-ray diffraction
measurements were conducted in situ in the DAC, again at the same beamline with a 20 keV 5×5
μm beam and Mar 345 image plate area detector. Omega rotation (±14°) was used for single-
crystal diffraction. Two-dimensional diffraction patterns were reduced to one-dimensional
patterns using the software program Dioptas21 to determine interplanar spacings.
In-house x-ray diffraction data on the sample recovered from the Paris-Edinburgh press
were collected with a 300 μm diameter, monochromatic copper Kα1 x-ray beam and a CCD area
detector.
4.2.3 Infrared Spectroscopy (IR)
We performed mid-IR absorption measurement on the DAC-recovered sample with a
Bruker Vertex V70 spectrometer equipped with a Hyperion 3000 FT-IR microscope. Background
spectra were collected under the same experimental conditions and subtracted using OPUS
software.
85
4.2.4 X-ray Photoelectron Spectroscopy (XPS)
XPS experiments were performed using a Physical Electronics VersaProbe II equipped
with a monochromatic Al Kα x-ray source (hν = 1,486.7 eV) and a concentric hemispherical
analyzer. Charge neutralization was performed using both low-energy electrons (<5 eV) and
argon ions. The binding energy axis was calibrated using sputter-cleaned Cu foil (Cu 2p3/2 =
932.7 eV, Cu 3p3/2 = 75.1 eV). Corrections for charging were done using a Au referencing
standard cold pressed onto the sample surface. The Au 4f 7/2 line was assigned to a binding energy
of 84.0 eV. Measurements were made at a takeoff angle of 45° with respect to the sample surface
plane, resulting in a typical sampling depth of 3–6 nm. Quantification was done using locally-
derived sensitivity factors from a pure polyvinylpyrrolidone reference sample. All the high-
resolution XPS analysis presented here were performed on the unsputtered PE cell sample made
from deuterated pyridine (C5ND5).
4.2.5 Nuclear Magnetic Resonance (NMR) Spectroscopy
Solid-state NMR spectroscopy was performed on a Bruker DSX 400 Avance
spectrometer at 40 MHz 15N and 100 MHz 13C resonance frequencies. Six mg of nanothreads
made from 15N-pyridine synthesized in 6 synthesis runs were center-packed into a 4-mm magic-
angle spinning (MAS) rotor. Pulse lengths for 90° flip angles were 4.2 μs for 1H and 13C, and 7 μs
for 15N.
4.2.6 Fluorescence Microscopy
Confocal images were obtained using an Olympus FluoView™ FV1000 confocal
microscope. The blue channel was visualized with a 405 nm violet laser as excitation and 425 nm
86
with a band filter of 50 nm as emission. The green channel was visualized with a 488 nm blue
laser as excitation and 520 nm with a band filter of 30 nm as emission. The red channel was
visualized with a 546 nm green laser as excitation and >572 nm from a long-pass filter as
emission. In all imaging experiments, the laser transmissivity (50% of 5 mW) and PMT voltage
(625) were identical for blue, green, and red channels. All images were collected with a 20×
optical lens with NA of 0.50 and with image size of 317×317 µm.
4.2.7 Photoluminescence Spectroscopy
Photoluminescence (PL) spectra were collected with excitation from a 488 nm laser
through a 50× objective with numerical aperture (NA) of 0.5 with a Horiba LabRAM Raman
spectrometer. The power was reduced to 5µW by neutral-density filters. We used a Renishaw
inVia Raman microscope to collect PL spectra with both 364 nm and 514 nm excitation through a
20× objective with NA=0.4. The laser power was 5µW.
4.2.8 Combustion Elemental Analysis
We employed a CE Instruments EA 1110 CHNS-O elemental analyzer in the CHN
measurement mode. The CHNS(O) Analyzer determines the percentages of C, H, N, S & O of
organic compounds, based on the "Dumas method" which involves the complete and
instantaneous oxidation of the sample by "flash combustion”.22 The combustion products are
separated by a chromatographic column and detected by a thermal conductivity detector (T.C.D.),
which gives an output signal (mVolts) proportional to the concentration (%) of the individual
components of the mixture. The carbon nitride nanothread sample used for this measurement is
the same one that was characterized by x-ray diffraction and XPS. It was synthesized in a Paris-
87
Edinburgh press from deuterated pyridine (C5ND5) during an in-situ neutron powder diffraction
experiment. Glass wool was added to it in an attempt to break up the solidified pyridine into
smaller and more random crystalliates to improve the powder diffraction statistics. This glass
wool should not have a deleterious effect on the combustion analysis, as it is routinely carried out
in the presence of silicates. The weight of sample was 0.702mg, which includes the mass of the
glass wool.
4.3 Results and Discussion
Liquid pyridine was loaded into a double-stage membrane diamond anvil cell23 together
with a ruby sphere for pressure calibration. As we slowly compressed the pyridine over 8–10
hours, it first froze to solid pyridine at 1–2 GPa17-18 and then reacted at ~18 GPa . After holding
the sample at 23 GPa for one hour, the pressure was released to ambient over 8–10 hours. The
recovered sample is a yellow/orange translucent solid. The diamond cell induces an anisotropic
(uniaxial) stress parallel to the force applied by the diamonds.24 Synchrotron x-ray diffraction
patterns collected with an area detector (Figure 4-3a, left) and the beam parallel to the uniaxial
stress axis are characteristic of a hexagonal single crystal with its c-axis parallel to the x-ray
beam, suggesting a role for uniaxial stress in guiding the growth of the crystal.2 There is good
agreement with the predicted six-fold single-crystal diffraction pattern of a representative tube
(3,0)_123456 carbon nitride nanothread with structural parameters determined by modeling
(Figure 4-3a, right). The ability of the uniaxial stress to align nanothreads is remarkable: even
when the pyridine reactant is loaded into a large volume pressure cell with glass wool in an
attempt to disrupt single-crystal order, a crystal embedded in this wool is still obtained.
88
Figure 4-3. Diffraction experiment and modeling. (a) Synchrotron diffraction pattern (left)
collected down the hexagonal c axis of the nanothread crystal showing its six-fold symmetry.
Simulated diffraction pattern (right) of carbon nitride tube(3,0)_123456. (b) Experimental and
simulated (100) interplanar spacings for different carbon nitride nanothread structures.
These single-crystal diffraction patterns provide compelling evidence for a hexagonally
packed array of nanothreads of diameter ~6.5 Å, but do not yet constrain which of several
possible pyridine nanothread atomic structures15 are formed. In contrast, previous investigations
of the reaction products from the compression of pyridine reported them to be amorphous,17-18
possibly because faster compression/decompression rates led to a more highly branched set of
reaction pathways for intermolecular bonding.2-3
The diffraction patterns are pseudohexagonal,2 i.e. with small deviations from perfect
hexagonal symmetry. The interplanar spacings for the three Friedel pairs of arcs from experiment
can be compared with those from simulations (Figure 4-3b) that were performed at 0 K for six
optimized pseudo-hexagonal packed carbon nitride nanothreads. Each of those six threads is
89
energetically most stable among their isomeric structures. The optimized packing geometries
were obtained from energy minimization of the originally ordered and perfectly hexagonally
packed threads. After shrinking the experimental spacings by 1–3% to account for thermal
contraction to 0 K, there is reasonable agreement with the predicted spacings for tube
(3,0)_123456. Polytwistane_153 and zipper polymer_2415 are also consistent with experiment,
but with a slightly denser packing along one dimension. Polymer I_3-3_25 and Polymer I_4-
2_3615 seem less likely candidates, with significant differences of spacings when ordered packed,
but azimuthally and/or axially disordered packing will narrow the differences among the
interplanar spacings.
Amorphous carbon nitride (a-C:H:N) films deposited by plasma-enhanced chemical
vapor deposition have broad infrared (IR) absorption spectra.25 In contrast, carbon nitride
nanothreads exhibit more and narrower absorption features (Figure 4-4) suggestive of greater
structural order not only of the thread packing (Figure 4-3), but also of the atomic structure along
the length of the threads. We calculated the IR absorption spectra of the six candidate structures
mentioned above. Two notable characteristic vibrations of a 1D thread or tube are radial
breathing and axial elongation-compression modes.26 The former is Raman (but not IR) active in
benzene nanothreads and is observed in experimental Raman spectra at 805 cm–1,3 much higher in
frequency than corresponding modes for (larger-diameter) sp2 carbon nanotubes.26 However,
symmetry breaking due to the nitrogen atoms in pyridine threads splits this mode and makes it IR
active. In all of the simulated spectra, there are one or several abs absorption peaks in the range
from 575 cm–1 to 650 cm–1 associated with radial breathing.
90
Figure 4-4. Experimental infrared spectrum. The major radial breathing mode (top left) and the
inter-ring C-N stretch mode (bottom right) in tube (3,0)_123456 are shown. Peaks with asterisks
may be associated with pyridinic derivatives either as amorphous carbon or as substituted
pyridine linking nanothreads. The intensities of these peaks are significantly lower than in the IR
spectra of amorphous recovered samples.17-18
In contrast to the corresponding higher-symmetry pure-carbon threads, mode eigenvectors for the
radial breathing modes of carbon nitride threads include non-radial (axial, circumferential)
motions, the magnitudes of which depend on the precise nitrogen placements on successive rings.
Considering the complexity of the vibration modes in this region and possibility of having more
than one type of nanothread formed during reaction, these calculated modes are consistent with
the broad peaks observed in experiment at 580 cm–1 and 645 cm–1. For example,
polytwistane_153 and zipper_24 threads have multiple modes below 700 cm-1 (as listed in Table
4-1) that are associated with mixed radial and non-radial motions.
Axial elongation and compression in carbon nitride nanothread include stretching of both
C–C and C–N single bonds formed between pyridine rings upon reaction. A peak in the
experimental IR spectrum at 1117 cm–1 is identified as inter- ring C-N stretching because all the
91
calculated spectra show this mode near 1120 cm–1 with moderate-to-strong intensity. Unlike the
peak at 1178 cm–1 that is associated with intra-ring C–C and C–N stretching, neither the 1117 cm–
1 inter-ring peak nor the radial breathing modes near 600 cm–1 appear in prior samples derives
from pyridine at high pressure17 – i.e. they appear to be diagnostic of a thread-like structure,
although the current IR spectra cannot yet uniquely determine which thread structure is formed.15
As all six structures are degree-6 nanothreads (i.e. fully sp3-hybridized), there are no IR
modes above 1400 cm–1 other than C–H stretching modes. By comparing the other modes in the
range from 600 cm–1 to 1400 cm–1 aside from radial breathing and inter-ring C–N stretching, the
IR peaks observed in experiment are reasonably consistent with the calculated frequencies
obtained from modeled structures. In contrast, the IR spectra of two samples synthesized by
compression of pyridine17-18 (both are claimed to be amorphous) exhibit several strong absorption
peaks (highlighted with grey background in Table 4-1) that are not observed or have very low
intensities in both the calculated and experimental carbon nitride nanothread spectra. The peaks at
708 cm–1 and 767 cm–1 are very likely coming from di-substituted pyridine and/or of tri-
substituted pyridine27 and the peaks at 1438 cm–1 and 1472 cm–1 can be assigned to methylene (–
CH2–) C–H bend28 which does not appear in ideal nanothread structures, but may appear in
defective regions. We are not yet able to assign a specific carbon nitride nanothread structure with
the present spectra and modeling, but the reasonable accordance with a set of different calculated
structural candidates and the obvious discrepancy with amorphous polymerized pyridine
recovered from high pressure provide support for the claim of carbon nitride nanothread
synthesis.
92
Table 4-1. Comparison of IR Peak Positions in Experiment and Simulation
Exp freq (cm-1) Cal freq (cm-1) Assignment
this
work ref 2 ref 3
tube 30 tube 30 polytwistane polymer I polymer I zipper
123456 1245 153 33_25 42_36 24
580 - - 576 600.91 583/592
both vw 589
592
571 vw
radial
breathing,
flexture
645 - - 636 625 656 vw 682 625/646/659
/683 all vw
radial
breathing,
flexture
845 - 843 - - - 836 829 816 C-H wag
(in plane)
998 987 997
985
998
987/997 990
- -
C-N stretch,
C-H wag
1012 1008 1008 C-H wag
(in plane)
1043 1050 - 1051 1042 1040/1049 1030 w - - C-H bend (out
of plane)
1089 - - 1083 1082/1097 1074 s/1084
w - 1072 1098
C-H bend (out
of plane)
1117 - - 1121 1126 1118 1121 1135 1132 C-N stretch
(inter-ring)
1178 1177 1163
1169 - 1168
1158 1168
1161 C-C stretch
(intra-ring)
1198 1190 1183 w 1181 C-N stretch
(intra-ring)
1345 1338 1330 1341 1335 w/
1317 s 1350 1317 1307 1345
C-H bend (out
of plane)
1376 - - - 1359 - 1374 1366 1367 C-H bend (out
of plane)
717* 706 707 - - - 718 - 730 stronger in
amorphous a
samples
771* 758 753 - 767 - 770 759 -
928* 903 - - 917 932 w 930 905 911
- 1218 - 1219 w - 1211 w 1200 w - -
only in
amorphous a
samples
The small (µg) size of the samples made chemical analysis challenging. Therefore, we
prepared larger (mg) samples in a Paris-Edinburgh pressure apparatus. Only C, N and O were
found in x-ray photoelectron spectroscopy survey scans. The observed N:C ratio, 1 : 5.08, also
confirmed by combustion experiment, is very close to that of pyridine, which contrasts with
93
chemical vapor deposition approaches for sp2 systems, which form low nitrogen content (x =
0.01~0.02) CNx nanotubes when using pyridine as a precursor.29
Figure 4-5. High Resolution XPS Scan of Carbon Nitride Nanothreads. C1s (left) and N1s (right)
XPS spectra of the raw surface of the recovered sample are presented as experimental data (open
circles) and fit curves (solid lines). All spectra were obtained after exposure to air.
We fit the C1s XPS spectrum by four Gaussian/Lorentzian mixed function components
with the same full width at half maximum (FWHM) as shown in Figure 4-5 (left). The two peaks
located at 284.9 eV and 286.0 eV correspond to carbon atoms (either sp2 or sp3) connected to
either no or one nitrogens, respectively.30 The component at 287.3 eV corresponds to sp3-carbon
bonded to two nitrogen neighbors.31-34 Exposure of the sample to air leads to some surface
oxidization, which is evident as carboxylic acid groups (COOH) in the spectrum.35 We
deconvolute the high-resolution N1s spectrum into three components with equal FWHM (Figure
4-5, right). The peak at 399.20 eV corresponds to nitrogen atoms connected to three sp3-carbon
atoms. A smaller component locating at 400.5 eV is attributed to sp2-nitrogen having at least one
sp2-carbon as a nearest neighbor.30 Most XPS studies assign the component with the highest
binding energy (>402 eV) to N–O bonding,36-39 but Kennou and Bertoti considered the
94
component at 402.4 eV to be a combination of N–O bonds at 402.8 eV and –N=C< bonds at 402
eV with a delocalized non-bonding pair on the nitrogen.36, 40 In a degree-4 carbon nitride
nanothread, the unsaturated component could be C=C or N=C; thus is it reasonable to obtain –
N=C< bonding in the XPS spectrum, which is also consistent with the NMR results that will
discussed later.
Solid-state nuclear magnetic resonance (NMR) spectroscopy of carbon nitride
nanothreads made from 15N-enriched pyridine provides a quantitative characterization of their
chemical composition and local chemical structure. We note that the NMR samples were
assembled from 6 synthesis runs and thus there may be more variation in bonding and impurities
in these mg size samples than for the diffraction and spectroscopy samples of μg size. The 15N
and 13C NMR spectra (Figure 4-6) reveal nearly identical sp2/sp3 ratios for nitrogen and carbon
(22–23% sp2, 77–78% sp3), indicating significant conversion from sp2- to sp3-bonding during
polymerization, as expected, with similar outcomes for C and N. About 45% of the sp3 carbons
have chemical shifts between 50 and 95 ppm, indicating that they are bonded to N. This exceeds
the C–N bond fraction of 2/5 = 40% in pyridine and thus indicates C–N bond formation between
pyridine rings, which is consistent with the observation of inter-ring C–N stretching in IR. While
the 15N chemical shift of ~310 ppm is similar for pyridine and imine N=CH, the pronounced CH
resonances near 170 ppm in the 13C spectrum, distinct from pyridine signals at ≤152 ppm,
document the presence of imine N=CH as found in degree-4 nanothreads. Further, the relative 13C
NMR peak intensities indicate that in degree-4 threads, N=CH is strongly favored over N–
CH=CH, which would produce a CH peak near 106 ppm that is not significantly observed in
Figure 4-6b. The estimated fraction of degree-6 carbon nitride nanothreads is ca. 40%, after
taking into account the sp3-hybridized carbons associated with degree-4 nanothreads. These two
different types of threads still appear to pack into ordered crystals (Figure 4-3a). Their
proportions might be altered by further refinement of the synthesis protocol (e.g., duration, stress,
95
photochemistry, initiators, temperature etc.). Figure 4a also shows some amides and N–H
moieties, which may be due to reaction with O2 and/or H2O from air.
Figure 4-6. NMR Spectra. Solid-state (a) 15N and (b) 13C NMR spectra of nanothreads made from 15N-enriched pyridine. Peak areas are nearly quantitative. Selective spectra of NH and of
nonprotonated C are shown by colored thin lines in (a) and (b), respectively. The original signal
positions of pyridine are marked by dashed vertical blue lines. The natural-abundance 13C
spectrum shows the characteristic chemical-shift increases due to N-bonding of nearly half of all
C.
The success in producing nanothreads from benzene2-3 and pyridine, both flat rings but
with a distinct molecular crystal structures, suggests that general principles govern nanothread
formation. Under compression, an efficient packing to reduce the volume of the solid is to form
stacks of parallel aligned flat molecules. Such stacks, usually slipped and along which π orbitals
of adjacent molecules interact, provide a natural direction for a polymerization that yields 1D
structures, i.e. nanothreads. The small size of the rings impedes π overlap between molecules in
adjacent stacks, disfavoring crosslink polymerization to form 2D networks. The uniaxial
component of the applied stress defines the thread axis, and the enthalpic driving force for
efficient packing under pressure strongly favors a dense near-hexagonal packing of the threads.
96
Low (i.e. room) reaction temperatures favors the single lowest-barrier kinetic pathway towards
thread formation and requires slow compression and decompression to afford the reaction
sufficient time to proceed.
Figure 4-7. Confocal fluorescence microscopy images of (a) blue, green, and red emission from
carbon nanothreads and (b) blue, green and red emission from carbon nitride nanothreads. Details
of excitation and emission are pro-vided in the supporting information.
Nanothreads exhibit photoluminescence, likely due to defects or sp2 carbon functions
along their length (Figure 4-7 a,b) or bandgap emission.41-42 There has been much interest in
carbon nanomaterials such as nanodiamond and nanotubes for luminescent biological imaging
applications as well as nanomedicinal therapy.42 The blue, green, and red emission properties of
carbon nanothreads differ from those of carbon nitride nanothreads, with the latter having
stronger emission at longer wavelengths when excited in a confocal fluorescence microscope
(Figure 4-8). Detailed investigation of the luminescence properties of nanothreads is beyond the
scope of the present work, but it is reasonable to expect the shift to longer emission wavelengths
97
to be associated with the decrease in bandgap predicted for carbon nitride nanothreads43 or
nitrogen-related defects. With further synthesis work the defects, bandgaps, and surface
functionalization of nanothreads might be designed for specific biological applications.
Since pyridine incorporates nitrogen isoelectronically, it potentially provides a means to
n-type dope nanothreads in two steps: isoelectronic nitrogen incorporation followed by activation
through covalent attachment to the nitrogen site; in contrast, n-doped diamond requires the
bandgap penalty of n-type doping to be paid simultaneous with nitrogen incorporation into the 3D
lattice. Carbon nitride nanothreads might exhibit the photocatalytic properties that have attracted
interest in other carbon nitride solids,44 especially as more nitrogen is added to their backbone and
when they have partial sp2 bonding, with photochemistry and electron transfer controlled by
reactant design. There has been much interest in the mechanical and other properties of sp3-
bonded three-dimensional carbon nitrides, but synthesis has been challenging.12-14 Similar to
carbon nanonthreads, carbon nitride nanothreads are predicted to have remarkable mechanical
properties, such as ~1 TPa Young’s moduli.43 The chemical synthesis of crystalline sp3-bonded
carbon nitride nanothreads with extended bonding and a well-defined geometry suggests more
generally an opportunity to develop and design nitrogen-vacancy centers with specific geometries
of interest for quantum computing and magnetometry.45 The amine functions in carbon nitride
nanothreads might allow for improved solubility or dispersion in protic (especially acidic)
solvents and facilitate solution-based processing. Most importantly, the observation of nanothread
formation from multiple aromatic precursors suggests that many other aromatics may react to
form nanothreads; some of them might react at pressures (~5 GPa) practical for synthesis of >106
kg/yr.46
Above is adapted with permission from J. Am. Chem. Soc. 2018, 140, 4969–4972.
Copyright (2018) American Chemical Society.
98
4.4 Insights and Future Considerations
The fluorescence properties of carbon nitride nanothread raise up few questions
including: what are they derive from, how to do quantify the intensity and efficiency. Limited to
the experimental condition that there are no proper solid fluorescence spectrometer available, a
photoluminescence (PL) measurement had been performed instead as mentioned in last section.
The spectra are presented in Figure 4-8. It’s clearly showing that the peak position shifted with
different excitation wavelength. This shifting indicates the luminescence is not only due to the
band emission but also some surface defects. The defects in nanothread sample are not
necessarily to be amorphous carbon but could be a second phase since there are degree-6 threads,
degree-4 threads and some oligomers co-existed. However, in the view of potential application
such as in bio-imaging, PL spectrum cannot replace the measurements of fluorescence quantum
yields, fluorescence lifetime correlation spectroscopy etc.
Figure 4-8. Photoluminescence Spectra of Carbon Nanothread and Carbon Nitride Nanothread.
Three lasers with different wavelength were employed as excitation sources. Both samples were
measured under the same conditions for experiment at each wavelength. Carbon nanothreads
have stronger photoluminescence than carbon nitride nanothreads when excited by 364 nm laser.
On the other hand, when excited by longer wavelength lasers (488 nm and 514 nm), carbon
nitride nanothreads show more intense photoluminescence than carbon nanothreads. The results
are in accord with the observations in Figure 4-7a that blue light (excited by 405 nm) is the major
fluorescence emission of carbon nanothreads while red light emission (excited by 546 nm) is very
weak. The green light emission (excited by 488 nm) of carbon nitride nanothread is much
stronger than its blue light emission, while red light emission is very weak.
99
It’s common to see that the optical bandgap of semiconductors, for example ZnO thin
film, are determined by tauc plots via UV-Vis measurements.47 There are considerable large
number of literatures reported the band gap of carbon materials obtained by the same methods.
It’s known that the increase of sp2 fraction with decrease the gap even if the sp3 carbon is the
majority within the material (Figure 4-9, left). Moreover, the gap is not only depend on the
amount of sp2 contents but also the configuration of sp2 sites. In another word, the bandgap
doesn’t vary with sp2 content in every situation indicating special consideration is required for
each individual case.48 Carbon nitride materials share the same consideration and with one more
factor (nitrogen content) need to take care (Figure 4-9, right).49 For carbon nanothread and carbon
nitride nanothread, the optical gaps derived from tauc plots are below 3 eV which is lower than
fully saturated threads. The reason of smaller gap could due to the existence of degree-4
nanothread or some other sp2 components, such as the aromatic “linker” attached with degree-6
nanothread.50 To reconcile this question, more explicitly interpretation of Raman and IR spectra
will be of great help.
Figure 4-9. Left: Experimental variant of optical gap vs. sp2 fraction for a-C:H, ta-C:H and ta-C.48
(Copyright © 2002 Elsevier Science B.V.) Right: Tauc optical gap of ta-CN and a-CN films as a
function of the N/C ratio.49 (Copyright © 2003 Elsevier Science B.V.)
Another consideration about tauc plot is that the transmission mode and diffuse reflection
mode of UV-Vis measurements lead to different results. This concern came up when attampting
100
to compare the tauc gap of pyridine anothread and benzene nanothread. Figure 4-10 is the overlay
of two curves obtained from absorption spectrum (transmission mode) and diffuse reflectance
spectrum (reflection mode) of a same PE cell benzene nanothread sample respectively. The y axis
in the former case is (αhʋ)1/2and the latter is (F(R) hʋ)1/2 for indirect band gap, where α is the
extinction coefficient and F(R)=(1-R)2/2R is Kubelka-Munk function that is proportional to α.51-52
In practical data process, absorbance (A) is used as α because the precise thickness of sample is
difficult to measure. The discrepancy of these two curves is significant. One reason could be that
the absorption spectrum becomes very noisy after 3.2 eV and the lack of larger range data may
mislead the tauc gap determination. The other reason is that there are different species of
substance dominating the detection in each mode. The predicted bandgap values for degree-6
nanothread are above 4 eV which quite beyond the tauc gaps in either mode. Therefore, a more
reliable method is needed for band gap study of this type of material. If degree-6 thread and
degree-4 thread can be separated in recovered sample or selectively synthesized under pressure, it
will be largely benefit the studies on physical properties.
Figure 4-10. Tauc plots of same carbon nanothread sample derived from absorption spectrum
(red) and diffuse reflectance spectrum (green).
101
References
(1) Hoffmann, R.; Kabanov, A. A.; Golov, A. A.; Proserpio, D. M. Angew. Chem. Int. Edit. 2016,
55, 10962-10976.
(2) Li, X.; Baldini, M.; Wang, T.; Chen, B.; Xu, E.-s.; Vermilyea, B.; Crespi, V. H.; Hoffmann,
R.; Molaison, J. J.; Tulk, C. A.; Malcolm; Guthrie; Sinogeikin, S.; Badding, J. V. J. Am. Chem.
Soc. 2017, 139, 16343–16349.
(3) Fitzgibbons, T. C.; Guthrie, M.; Xu, E. S.; Crespi, V. H.; Davidowski, S. K.; Cody, G. D.;
Alem, N.; Badding, J. V. Nat. Mater. 2015, 14, 43–47.
(4) Xu, E. S.; Lammert, P. E.; Crespi, V. H. Nano Lett 2015, 15, 5124–5130.
(5) Badding, J. V.; Crespi, V. H., Synthesizing Carbon Nanothreads from Benzene. SPIE
Newsroom 2015.
(6) Chen, B.; Hoffmann, R.; Ashcroft, N. W.; Badding, J.; Xu, E. S.; Crespi, V. J Am Chem Soc
2015, 137, 14373–14386.
(7) Olbrich, M.; Mayer, P.; Trauner, D. J. Org. Chem 2015, 80, 2042–2055.
(8) Roman, R. E.; Kwan, K.; Cranford, S. W. Nano Lett. 2015, 15, 1585–1590.
(9) Zhan, H.; Zhang, G.; Tan, V. B. C.; Cheng, Y.; Bell, J. M.; Zhang, Y.-W.; Gu, Y. Nanoscale
2016, 8, 11177–11184.
(10) Toda, F. Organic Solid-State Reactions, Springer: Berlin, 2005.
(11) Podsiadlo, M.; Jakobek, K.; Katrusiak, A. CrystEngComm 2010, 12, 2561-2567.
(12) Liu, A. Y.; Cohen, M. L. Science 1989, 245, 841-842.
(13) Dong, H. F.; Oganov, A. R.; Zhu, Q.; Qian, G. R. Sci Rep-Uk 2015, 5, 9870.
(14) Badding, J. V. Adv Mater 1997, 9, 877-886.
(15) Chen, B.; Wang, T.; Crespi, V. H.; Li, X.; Badding, J.; Hoffmann, R. J. Chem. Theory
Comput. 2018, 14, 1131-1140.
(16) Hrubiak, R.; Sinogeikin, S.; Rod, E.; Shen, G. Y. Rev. Sci. Instrum. 2015, 86, 072202.
(17) Fanetti, S.; Citroni, M.; Bini, R. J. Chem. Phys. 2011, 134, 204504.
(18) Zhuravlev, K. K.; Traikov, K.; Dong, Z.; Xie, S.; Song, Y.; Liu, Z. Phys. Rev. B 2010, 82,
064116.
(19) Klotz, S.; Hamel, G.; Frelat, J. High Pressure Res. 2004, 24, 219–223.
(20) Fang, J.; Bull, C. L.; Loveday, J. S.; Nelmes, R. J.; Kamenev, K. V. Rev. Sci. Instrum. 2012,
83, 093902.
(21) Prescher, C.; Prakapenka, V. B. High Pressure Res. 2015, 35, 223–230.
(22) Matejovic, I. Rostlinna Vyroba-UZPI (Czech Republic) 1996.
(23) Sinogeikin, S. V.; Smith, J. S.; Rod, E.; Lin, C.; Kenney-Benson, C.; Shen, G. Rev. Sci.
Instrum. 2015, 86, 072209.
(24) Yao, M. G.; Fan, X. H.; Zhang, W. W.; Bao, Y. J.; Liu, R.; Sundqvist, B.; Liu, B. B. Appl
Phys Lett 2017, 111, 101901.
(25) Ferrari, A. C.; Rodil, S. E.; Robertson, J. Phys. Rev. B 2003, 67, 155306.
(26) Dresselhaus, M. S.; Dresselhaus, G.; Eklund, P. Science of Fullerenes and Carbon
Nanotubes: Their properties and Applications, Elsevier: New York, 1996.
(27) Abramovitch, R. A.; Giam, C. S.; Notation, A. D. Can. J. Chem. 1960, 38, 624-626.
(28) Coates, J. Interpretation of Infrared Spectra, A Practical Approach, Meyers, R. A., Ed. John
Wiley & Sons, Ltd: Chichester, 2006.
(29) Liu, J.; Czerw, R.; Carroll, D. L. J. Mater. Res. 2005, 20, 538-543.
(30) Titantah, J. T.; Lamoen, D. Diamond Relat. Mater. 2007, 16, 581-588.
(31) Hammer, P.; Lacerda, R. G.; Droppa, R.; Alvarez, F. Diamond Relat. Mater. 2000, 9, 577-
581.
102
(32) Kusunoki, I.; Sakai, M.; Igari, Y.; Ishidzuka, S.; Takami, T.; Takaoka, T.; Nishitani-Gamo,
M.; Ando, T. Surf. Sci 2001, 492, 315-328.
(33) Ripalda, J. M.; Montero, I.; Galan, L. Diamond Relat. Mater. 1998, 7, 402-406.
(34) Le Normand, F.; Hommet, J.; Szörényi, T.; Fuchs, C.; Fogarassy, E. Phys. Rev. B 2001, 64.
(35) Ray, K. R.; McCreery, R. L. J. Electroanal. Chem. 1999, 469, 150-158.
(36) Kennou, S.; Logothetidis, S.; Sygellou, L.; Laskarakis, A.; Sotiropoulou, D.; Panayiotatos,
Y. Diamond Relat. Mater. 2002, 11, 1183-1187.
(37) Marton, D.; Boyd, K. J.; Albayati, A. H.; Todorov, S. S.; Rabalais, J. W. Phys. Rev. Lett.
1994, 73, 118-121.
(38) Zou, Y. S.; Wang, Q. M.; Du, H.; Song, G. H.; Xiao, J. Q.; Gong, J.; Sun, C.; Wen, L. S.
Appl. Surf. Sci. 2005, 241, 295-302.
(39) Bongiorno, G.; Blomqvist, M.; Piseri, P.; Milani, P.; Lenardi, C.; Ducati, C.; Caruso, T.;
Rudolf, P.; Wachtmeister, S.; Csillag, S.; Coronel, E. Carbon 2005, 43, 1460-1469.
(40) Bertoti, I.; Toth, A.; Mohai, M.; Ujvari, T. Surf. Interface Anal. 2000, 30, 538-543.
(41) Rusli; Robertson, J.; Amaratunga, G. A. J. J. Appl. Phys. 1996, 80, 2998–3003.
(42) Hong, G. S.; Diao, S. O.; Antaris, A. L.; Dai, H. J. Chem. Rev. 2015, 115, 10816-10906.
(43) Silveira, J. F. R. V.; Muniz, A. R. Carbon 2017, 113, 260–265.
(44) Miller, T. S.; Suter, T. M.; Telford, A. M.; Picco, L.; Payton, O. D.; Russell-Pavier, F.;
Cullen, P. L.; Sella, A.; Shaffer, M. S. P.; Nelson, J.; Tileli, V.; McMillan, P. F.; Howard, C. A.
Nano Lett. 2017, 17, 5891-5896.
(45) Schirhagl, R.; Chang, K.; Loretz, M.; Degen, C. L. Annu Rev Phys Chem 2014, 65, 83-105.
(46) US Geological Survey Diamond Report.
https://minerals.usgs.gov/minerals/pubs/commodity/diamond/mcs-2016-diamo.pdf.
(47) Viezbicke, B. D.; Patel, S.; Davis, B. E.; Birnie, D. P. physica status solidi (b) 2015, 252,
1700-1710.
(48) Robertson, J. Materials Science and Engineering: R: Reports 2002, 37, 129-281.
(49) Rodil, S.; Muhl, S.; Maca, S.; Ferrari, A. Thin Solid Films 2003, 433, 119-125.
(50) Duan, P.; Li, X.; Wang, T.; Chen, B.; Juhl, S. J.; Koeplinger, D.; Crespi, V. H.; Badding, J.;
Schmidt-Rohr, K. J. Am. Chem. Soc. 2018.
(51) Kubelka, P.; Munk, F. Z. Tech. Phys 1931, 12.
(52) Morales, A. E.; Mora, E. S.; Pal, U. Revista mexicana de física 2007, 53, 18-22.
103
Chapter 5
Structure Exploration of Solid Pyridine under Pressure
5.1 Motivation and Background
The transformation from benzene to crystalline carbon nanothread is considered as a
unique non-topochemical solid state reaction.1 Pyridine is a prototype of aromatic 6-membered
heterocycle. The carbon nitride nanothreads derived from pyridine under high pressure share
much similarities with benzene nanothread.2 Herein, it’s interesting and essential to learn if
carbon nitride nanothread formation also undergoes the same type of transformation that lead to
crystalline product while freedom from topochemical constraints. In order to do so, knowing the
structure of parent crystal, pyridine, is critical.
The first crystal structure of solid benzene was solved in 1964 at low temperature3 and
the two high pressure phases were identified 5 years later.4 However, the structural study on solid
pyridine is quite scarce. The low temperature phase I was determined in 1981.5 Surprisingly,
although pyridine is one of the simplest heteroaromatic molecule, it’s crystal structure is
unusually complicate, with 4 independent molecules in the asymmetric unit. Since the rapid
development of ab initio calculation of twentieth century, especially the excellent performances
on simulation of relative stabilities of isomers of gas-phase molecules, research interests in solid-
state structures have been stimulated as well. Pyridine was one of the systems to demonstrate this
approach. In 2008, by learning that there are over a dozen of crystal structures were energetically
competitive with phase I,6 experimental searches were conducted at low temperature. Proteated
C5NH5 remains as phase I (h5-I_LT) down to 5 K without any phase transition. Unexpectedly, a
deuterated pyridine (C5ND5) solid crystallized from pentane at 188 K was yielded into a new
104
orthorhombic phase but in a different space group from h5-I. This new phase (d5-II_LT) was then
confirmed by a pure temperature driven experiment.7 Neat d5-pyridine was first frozen at 77 K
followed by rapid cooling to 2 K and obtained d5-I which is same as h5-I_LT. A sluggish phase
transition of d5-I_LTd5-II_LT happened when temperature increased to 170 K. The d5-II
phase transformed back to d5-I phase at 215 K. Isotopic polymorphism have been observed in
other systems such as potassium dihydrogen phosphate, chromium hydroxide oxide and
ammonium dihydrogen arsenate.8 But this phenomenon is very rare for organic molecules. The
origin of isotopic polymorphism of pyridine is still unclear.
This new low temperature phase inspired the first experimental attempt of structure
search under pressure. Crystal structure of high pressure phase II of proteated pyridine (h5-II_HP)
was determined by single crystal diffraction at 1.1 GPa.7 The protocol of achieving this crystal is
very ambiguous though, as below:
“High-pressure single-crystal X-ray study of h5-II: Phase h5-II was obtained at
high pressure by loading [H5] pyridine into a Merrill-Bassett diamond anvil cell,
and then applying enough pressure to solidify the sample into a polycrystalline
mass. The pressure was reduced to just above the melting pressure, and the cell
was heated until just one small seed crystal remained, and the apparatus then
allowed to cool to room temperature.”
A phase transition of d5-I_HP d5-II_HP was claimed at 1.1 GPa in the same paper but there is
no published data available.
In 2010, another research group managed to grow a high quality C5NH5 single crystal in
diamond anvil cell isothermally as shown in Figure 5-1. Crystal structure of h5-I_HP was then
solved at 1.23 GPa, 295K9 and it resembles to h5-I_LT.5 After pressure increased to 2.00 GPa, the
single crystal was damaged due to a solid-solid phase transition. It’s interesting to note that
isochoric growth of pyridine single crystal was failed as illustrated in Figure 5-2 because of the
destructive phase I and II transition which is very different from the behavior of solid benzene
105
under pressure. Limited to experimental condition, the crystal structure of phase II was unable to
determine.
Figure 5-1. Isothermal growth of pyridine single crystal phase I. (a) liquid-solid equilibrium at
1.00 GPa/295 K; (b-i) crystal growth following sample chamber volume decrease; (j) one single
crystal fully filled the chamber at 1.23 GPa/295 K.9 (Copyright © 2010, Royal Society of
Chemistry)
Figure 5-2. Failed isochoric growth of pyridine single crystal phase II. (a,b) polycrystalline phase
II at 1.20 GPa/295 K; (c) one crystal seed at 350 K; (d-i) single crystal growth during cooling; (i)
phase II single crystal before phase transition at 310 K; (j) phase I powder obtained at 0.9
GPa/295 K due to the destructive II/I phase transition.9 (Copyright © 2010, Royal Society of
Chemistry)
From 2010 to 2011, there are three other papers on high pressure phase transitions and
potential chemical transformations and all of them are spectroscopic studies. Few possible phase
106
transitions were reported by Zhuravlev et. al10 and Yasuzuka et.al11 at 2, 5, 8 ,11 and 16 GPa
based on slope changes of Raman shift versus pressure or appearance of new lattice mode.
However, Fanetti et.al12 believe that there is no phase transition at least till 12 GPa by performing
series of convolution of minimum number of peaks for lattice modes. All these research works
reported irreversible chemical transformation when pyridine compressed above ~20 GPa. Fanetti
et.al also concluded the crystallization directly occurred in phase II and never observed the direct
crystallization of phase I (Figure 5-3). In fact phase I can only obtained by decompressing phase
II to around 1 GPa as demonstrated in the single crystal diffraction work of Podsiadlo et.al9
discussed above.
Figure 5-3. Schematic demonstration of phase transitions of pyridine under pressure.12
After 2011 to now, there is only 1 more paper focus on solid pyridine by studying the
distributed intermolecular force-fields via modelling that published in 2017. The only two
determined high pressure phases are h5-I at 1.23 GPa9 and h5-II at 1.11 GPa.7 This chapter will
first revisit high pressure phase I of proteated pyridine by correlating the experimental results
obtained from Raman spectroscopy and X-ray diffraction. These different characterizations were
only presented and discussed individually in separate literatures. Second, a new unknown high-
107
pressure phase observed at 2 GPa will be discussed. Last but not the least, some insights of
possible reaction mechanism and enlightening of possible optimization of synthesis will be
addressed.
5.2 Revisit of High Pressure Phase I
As introduce in Background section, Fanetti et.al12 found that high pressure phase I
cannot form directly by compressing fluid but by decompressing phase II down to ~1 GPa
through Raman spectroscopy studies. Podsiadlo et.al9 managed to grow a phase I single crystal in
the scale of few hundreds of µm via isothermal method (described in 5.1) and solved its structure
at 1.23 GPa. However, there is a discrepancy of the transition pressure from liquid to phase I
among different works. Podsiadlo et.al reported the freezing pressure of pyridine is 0.53 GPa.
However, in the experiments done between 80’s to 90’s13-14 and two more recent experiments10-11
in 2011, pyridine was observed solidifying ~1GPa or above. This discrepancy is later explained
as the supercompression behavior of liquid and studied by Raman susceptibility.12
108
Figure 5-4. Le Bail fit of power diffraction pattern collected at 1.13 GPa. Blue crosses are the
observed data points and fitting line is in green. Pyridine phase I crystal structure solved at 1.23
GPa9 was employed in the fitting.
In this dissertation work, the freezing pressure by direct compressing liquid pyridine is
higher than 0.53 GPa and slightly above 1 GPa as well. By reproducing the procedures applied in
the literatures above, pyridine is found keeping as solid at 0.6 GPa when released from higher
pressure (above 2 GPa). Large crystallites recrystallized from melt after pressure released from
2.40 GPa down to 0.6 GPa and then recompressed to 1.13 GPa for high pressure X-ray diffraction
experiment. The 1-D integrated diffraction pattern (Figure 5-4) can be reasonably fitted by using
Le Bail methods with pyridine high pressure phase I structure that has been solved by Podsiadlo
et.al.9 Therefore, the phase I that determined by X-ray diffraction is confirmed to be reproducible.
Figure 5-5. Raman spectra of the lattice phonon region measured in this dissertation work (Left)
and by Fanetti et al. (Right). The 1.0 GPa and 1.7 GPa curves are considered to be the
characteristic spectra of phase I and phase II respectively.
This sample was also measured by Raman spectroscopy at 1.13 GPa as shown on the left
of Figure 5-5 (left). It’s not consistent with either spectra of two different phases of pyridine
obtained under pressure by Fanetti et.al (Figure 5-5, right). The Raman spectrum took at 0.6 GPa
after recrystallized from liquid is very similar to phase I spectrum (at 1.0PGa) reported in the
literature beside moderate blue shift. The phase transition boundary depicted in Podsiadlo et.al
109
work by combining diffraction and optical observation is ~1.25 GPa. Considering about the
precision of pressure determined by Ruby fluorescence method, the different features shown in
1.13 GPa spectrum may be explained as the co-existence of phase I and next high-pressure phase.
At this point, it would be a bit arbitrary to claim the phase I that identified by Raman and
diffraction works are actually equivalent. To be more rigorous, a combination of Raman and
diffraction study on high purity pyridine crystal within 0.6~1.0 GPa is required.
5.3 Unknown High pressure phase II’
Benzene undergoes two phase transitions before polymerization under pressure. The
transition from orthorhombic phase I to monoclinic phase II is very sluggish unless proper
annealing is employed. Phase II has been confirmed being able to polymerize into carbon
nanothread via a unique non-topochemical reaction based on the work in Chapter 3. Herein,
learning the transformation pathway from solid pyridine to carbon nitride is naturally motivated.
From earlier Raman studies, pyridine was observed crystallizing into high pressure phase II at~1
GPa directly from liquid and kept in this phase until reaching polymerization threshold pressure
(17~18 GPa).12 This spectroscopic work identified the phase II by lattice modes and accounted it
as the same phase II that determined by X-ray diffraction at 1.23 GPa.9 This chapter will present
the discovery of a high pressure phase obtained from same compression approach as described in
the Raman work but its diffraction patterns do not fit any known crystal structure of pyridine. By
the end of this chapter, some insights for potential reaction mechanism and optimization of
synthesis protocol will be discussed according to most-recent X-ray diffraction collected.
110
5.3.1 Powder X-ray diffraction
The procedure of synthesis carbon nitride nanothread is compressing liquid pyridine to
1~2 GPa without any extra recrystallization such as annealing or decompression. Therefore,
Raman spectra were taken at this state and compared with previous literature. As shown in Figure
5-6 the direct compression to 1.4 GPa and 2.4 GPa in two separate experiments exhibit same
Raman modes in low frequency region beside red shift of the latter due to higher pressure. Both
of the spectra are very nicely consistent with phase I reported by Fanetti et.al.
Figure 5-6. Raman spectra of two pyridine crystals formed from direct compression of fluid in
different experiments.
A powder X-ray diffraction was conducted on the 2.4 GPa sample first. In order to make
it easier for direct comparison with phase I and phase II structures solved by Podsiadlo9 and
Crawfold,7 the pressure of this sample was later released to 1.23 GPa. From Figure 5-7, one can
see there is a major discrepancy which is the refection observed at lowest two theta angle couldn’t
fit to either of the two phases. This reflection appeared repeatedly in multiple crystals among
different x-ray diffraction experiments. Beside this, there are other two peaks with considerable
111
intensities lack of fitting as well. Therefore, it’s reasonable to believe that this phase which is the
starting structure for nanothread synthesis is the same phase II observed in previous Raman study
but not the same as the phase II determined in diffraction work.
Figure 5-7. Integrated one-dimensional powder diffraction pattern collected at 1.23 GPa in this
work (black curve) compares with reported two phases at similar pressure. The unfitted peaks are
labeled with asterisks.
An autoindexing was perform in GSAS-II15 on the powder patterns obtained at 1.23 GPa
and 2.4 GPa. Orthorhombic unit cell with both a and b close to 5.3 Å, c close to 13.3 Å is
favored. Within monoclinic unit cells searched out by autoindexing those who have similar lattice
parameters as the orthorhombic cell and β nearly 90˚ are more favored than the others. It’s very
encouraging to see that tetragonal unit cell with the lattice parameter as described above can fit
the pattern reasonably well with least reflections comparing to orthorhombic and monoclinic
cells. Figure 5-8 shows the Le Bail fit16 of the 2.4 GPa powder pattern with tetragonal cell
(a=5.3100 Å, c=13.3169 Å).
112
Figure 5-8. Le Bail fit of power diffraction pattern collected at 2.4 GPa. Blue crosses are the
observed data points and fitting line is in green. The tetragonal unit cell obtained from
autoindexing (a=5.3100 Å, c=13.3169 Å) was employed in the fitting. The inset figure is the two
dimensional pattern illustrating the powder quality,
The space group and atomic positions are not able to be extracted from these current data though.
This work indicates that the mutual examination of spectroscopic and diffractometric
measurements are essential and necessary for determination of new phases.
5.3.2 Single Crystal X-ray Diffraction
In order to solve the crystal of the unknown phase discussed in 5.3.1, high pressure single
crystal diffraction was performed. To get single crystal, annealing had been tried but failed
because the destructive phase transition as described in Podsiadlo’s work.9 However, single
crystal with tens of micron scale can be obtained by slow compression from pyridine liquid
113
Figure 5-9. Omega scan of pyridine single crystal obtained at 3.0 GPa. The black boxes are
predicted reflections from the tetragonal structure (a=5.2852 Å, c= 13.2495 Å) indexed by this
dataset.
directly at room temperate. Benefit from the small beam size (~5 μm) at 16-ID-B, these crystals
are large enough to perform single crystal studies. The limitation of this work is the opening
angle of the diamond anvils and their seats. As introduced in Chapter 2, the opening angle will
decide the accessible range of the reciprocal lattice. Normally, to solve an unknown high pressure
phase, ±40 degrees opening is highly desirable. In this work, only ±10 degrees are allowed which
makes the direct indexing extremely difficult. Fortunately, with limited access to the reciprocal
space, a reasonable unit cell is searched out by Cell Now17 and its lattice parameters are quite
similar to the autoindexing results from powder data. With further refinement, a tetragonal unit
cell is obtained. Figure 5-9 presents the omega scan pattern of a pyridine single crystal collected
114
at 3.0 GPa. The boxed spots are the predicted reflection from the refined tetragonal cell. It’s
worth to note that lattice of parameters of this 3 GPa cell derived from single crystal dataset is:
a=5.2852 Å, c= 13.2495 Å while for the 2.4 GPa powder dataset they are: a=5.3100 Å,
c=13.3169 Å. This consistency further supports that the unknown phase is very possibly to a
tetragonal cell with the cell dimension listed above.
Figure 5-10. Comparison of experimental powder diffraction measured at 2.4 GPa with three
candidate structures predicted by DFTB method.
Because of the instrumental constraints, it’s very challenge to use direct method or
Patterson method to solve the atomic positions. Under this circumstance, theoretical calculation
and prediction become extremely useful. At beginning of this chapter, the motivation of structure
search of solid pyridine is to demonstrate the development of ab initio method. By collaboration
with Dr. Amsler, 60 different structures of pyridine under ~2 GPa with similar lattice parameters
indexed from experiment were obtained by Density Functional based Tight Binding (DFTB)
method.18 The preliminary comparison of 1-D patterns filters out 10 structures and 7 of them are
115
with low energy. With further filtering, it’s surprising to learn that the structure sharing most
similarities of the experimental integrated pattern is the one belong to tetragonal lattice.
Meanwhile this structure is the third lowest energetic one among the 60. There are three
structures not in the space group P1. As shown in Figure 5-10, the other two structures are
monoclinic and orthorhombic cell. They both possess higher energy than tetragonal unit cell and
fit less nicely.
Figure 5-11. Rietveld refinement of pyridine powder diffraction collected at 2.4 GPa by
tetragonal unit cell predicted from DFTB.
An attempt of Rietveld refinement19 on 2.4 GPa powder diffraction was conducted and
shown in Figure 5-11. The lattice parameters after refinement are a= 5.3078 Å and c= 13.3235 Å.
A more delicate calculation based on the filtered results have been employed. Further analysis
will be performed in the near future to fully elucidate the space group and detailed atomic
positions.
116
5.4 Insights for Reaction Mechanism and Synthesis Optimization
5.4.1 Compression Process
The onset of photoluminescence background is commonly treated as the sign of
polymerization. In the case of pyridine, the reaction threshold is considered around 17~18 GPa.
Since only Raman and IR studies have been done on pyridine at high pressure beyond 2 GPa, the
existence of possible phase transitions before chemical transformation were speculated by the
appearance/disappearance or certain modes or discontinuity slope of peak position vs pressure.
Potential phase transitions at 2,5, 8, 11 and 16 GPa were claimed by Zhuravlev et al10 and
Yasuzuka et al.11 However, the phase transition at 8 GPa was later opposed by Fanetti et al12 and
explained that the new lattice mode observed at this pressure belongs to the low frequency pattern
that the previous authors didn’t have access to experimental limitations. In fact, they believed
there is no phase transitions at least to till 12 GPa by performing convolution of minimum
number of peaks. Therefore, most of the new peaks shown up are explained as different crystal
orientations. Another supporting evidence is that the spectrum appearance is reversible in
decompression.
However, a new reflection around 2θ=8˚ is observed in the high pressure powder X-ray
diffraction powder as shown in Figure 5-12 when pressure reached 8 GPa and became more
distinct at higher pressure. The quality of the powder patterns also changed significantly during
compression (Figure 5-13). Both of these observations indicating the crystals are undergoing
certain changes. In Fanetti’s work, the lattice mode region was fitted with 15 peaks when pressure
below 9 GPa but 18 peaks afterwards. The new added peaks are all away from the cutoff
frequency so their absences at lower pressure are not due to instrumental limitation. But the
authors didn’t provide any explanation about this different treatment.
117
Figure 5-12. In situ pyridine powder diffraction patterns collected during compression. The dotted
arrow is the guidance of the new peak appearance and growth located at 2θ=8˚.
Figure 5-13. Two-dimensional patterns of pyridine powder during compression. Pyridine solid
became less powdery and with stronger preferred orientations after pressure beyond 8 GPa.
Another high pressure single crystal diffraction on pyridine was performed as well. It’s
very clear to see that more reflection spots showed up when pressure increased to 9 GPa (Figure
5-14). They cannot fit as the same domain that derived from the crystal measured at low pressure.
118
The six inner most reflections turned to nanothread (100) after tracking back to ambient pressure
while the other new reflections disappeared after solid pyridine melted.
Figure 5-14. Two-dimensional patterns of pyridine single crystal during compression. Circles are
used for guiding the observation of new reflections.
Consequently, the observations from this experiment raise up two questions: one is
whether the nanothread formation starting from 10 GPa instead of 17 GPa; and the second on is if
there is any phase transition happening around 10 GPa. In order to answer these questions, more
studies are needed including accurate understanding of the crystal structure of pyridine below this
pressure. It will also be interesting to associate the potential existence of benzene phase III’ that
few new reflections observed along with the appearance of nanothread (100) during compression
but disappeared after the melting of solid benzene. Even though there are many details haven’t
been figured out yet, the appearance of 6-fold pattern is unambiguously suggesting that slower
compression rates should be applied at least from 10 GPa.
5.4.2 Decompression Process
In the original benzene nanothread work,20 slower decompression rate is emphasized as
the key step in the synthesis. When looking into the detailed protocol, much slower rates near
maximum pressure are applied than the rates used closed to ambient. Table 5-1 demonstrates the
119
decompression rates used for diamond anvil cell synthesis of benzene nanothread and pyridine
nanothread. Meanwhile, from multiple in situ high pressure diffraction experiments on both
carbon nanothread and carbon nitride nanothread, the diffraction arcs present in the patterns of
recovered sample at ambient pressure are found always being broader than the ones observed at
high pressure. Therefore, it's necessary to figure out which process lead to the increase of disorder
thus corresponded synthesis procedure can be improved.
Table 5-1. Decompression rates applied in diamond anvil cell synthesis
P1 P0 Pressure rate (GPa/min)
23 19 0.02
19 13 0.03~0.04
13 10 0.06~0.07
10 4 0.1
4 0 0.2
Here presents the first exploration of how downloading rates near ambient pressure effect
crystallinity of recovered sample. Instead of 0.2 GPa/min, 0.05 GPa/min was applied after sample
pressure decrease to 4 GPa. As shown in Figure 5-15, the nanothread (100) arcs at 2.59 GPa
remain considerably narrow width. When reaching melting pressure, these arcs became broader
but still narrower than the ones observed from recovered sample in previous work. The next step
is to further inflate the decompression membrane so that the sample pressure can be pushed to
zero. A significant arc broadening was observed after the two anvils separated.
From the observations above, two pressure points should be noted that maybe closely
related to increase of disorder during nanothread synthesis. The first one is the melting pressure
of precursor crystal and the second one is the when sample pressure released back to ambient.
The former indicates the unreacted precursor solid acting as the template of nanothread and
holding the thread columns from collapse. The latter implies the uniaxial stress playing an
important role in guiding the formation of crystalline and order-packed reaction products not only
120
in high pressure range that close to polymerization threshold but also when the pressure is very
low.
Figure 5-15. Two-dimensional patterns of pyridine single crystal during decompression. Carbon
nitride nanothread (100) reflection arcs are pointed out by green arrows. Pyridine crystal melted
when pressure released to ~0.64 GPa thus only nanothread reelections remained.
Herein, in order to improve crystallinity of recovered sample in future synthesis, slower
decompression near the solid liquid transition boundary will be performed to reduce the
disorder in the first step. Other strategies are needed to avoid sudden loss of uniaxial stress during
cell opening.
121
References
(1) Li, X.; Baldini, M.; Wang, T.; Chen, B.; Xu, E.-s.; Vermilyea, B.; Crespi, V. H.; Hoffmann,
R.; Molaison, J. J.; Tulk, C. A. Journal of the American Chemical Society 2017, 139, 16343-
16349.
(2) Li, X.; Wang, T.; Duan, P.; Baldini, M.; Huang, H.-T.; Chen, B.; Juhl, S. J.; Koeplinger, D.;
Crespi, V. H.; Schmidt-Rohr, K. Journal of the American Chemical Society 2018, 140, 4969-
4972.
(3) Bacon, G.; Curry, N. t.; Wilson, S. Proc. R. Soc. Lond. A 1964, 279, 98-110.
(4) Piermarini, G.; Mighell, A.; Weir, C.; Block, S. Science 1969, 165, 1250-1255.
(5) Mootz, D.; Wussow, H. G. The Journal of Chemical Physics 1981, 75, 1517-1522.
(6) Anghel, A.; Day, G.; Price, S. CrystEngComm 2002, 4, 348-355.
(7) Crawford, S.; Kirchner, M. T.; Bläser, D.; Boese, R.; David, W. I.; Dawson, A.; Gehrke, A.;
Ibberson, R. M.; Marshall, W. G.; Parsons, S. Angewandte Chemie International Edition 2009,
48, 755-757.
(8) Zhou, J.; Kye, Y.-S.; Kolesnikov, A. I.; Harbison, G. S. Isotopes in environmental and health
studies 2006, 42, 271-277.
(9) Podsiadło, M.; Jakóbek, K.; Katrusiak, A. CrystEngComm 2010, 12, 2561-2567.
(10) Zhuravlev, K. K.; Traikov, K.; Dong, Z.; Xie, S.; Song, Y.; Liu, Z. Physical Review B 2010,
82, 064116.
(11) Yasuzuka, T.; Komatsu, K.; Kagi, H. Chemistry Letters 2011, 40, 733-735.
(12) Fanetti, S.; Citroni, M.; Bini, R. The Journal of chemical physics 2011, 134, 204504.
(13) Heyns, A.; Venter, M. The Journal of Physical Chemistry 1985, 89, 4546-4549.
(14) Heyns, A.; Venter, M. The Journal of chemical physics 1990, 93, 7581-7591.
(15) Toby, B. H.; Von Dreele, R. B. Journal of Applied Crystallography 2013, 46, 544-549.
(16) Le Bail, A. Accuracy in powder diffraction II, Special publication 1992, 846, 213.
(17) Dera, P.; Zhuravlev, K.; Prakapenka, V.; Rivers, M. L.; Finkelstein, G. J.; Grubor-Urosevic,
O.; Tschauner, O.; Clark, S. M.; Downs, R. T. High Pressure Research 2013, 33, 466-484.
(18) Elstner, M.; Seifert, G. Phil. Trans. R. Soc. A 2014, 372, 20120483.
(19) Rietveld, H. Journal of applied Crystallography 1969, 2, 65-71.
(20) Fitzgibbons, T. C.; Guthrie, M.; Xu, E. S.; Crespi, V. H.; Davidowski, S. K.; Cody, G. D.;
Alem, N.; Badding, J. V. Nat Mater 2015, 14, 43-47.
122
Chapter 6
Concluding Remarks and Future Outlook
This thesis work is supported by Energy Frontier Research in Extreme Environments
(EFree). The mission of this project is to accelerate the discovery and synthesis of energy
materials using extreme conditions. By taking one of the thrive which is to stabilize and
characterize new forms of carbon through tailored synthetic processes for the development of
new structural materials for energy application, I demonstrated the progress in optimizing and
developing new carbon nanomaterials in both large-volume press (Paris-Edinburgh cell) and
standard high pressure apparatus (diamond anvil cell). Benefit from the optimized protocol,
carbon nanothread single crystals derived from benzene have been synthesized. Moreover, the
new synthetic method turns out to be quite robust and it enables the invention of carbon nitride
nanothread single crystals obtained by compressing pyridine molecules under pressure and makes
it the second member of nanothread family. The synthesis of carbon nitride nanothread strongly
indicates that slow/compression approach could be a universal means to recover crystalline
product from pressure-driven polymerization instead of only amorphous samples that have been
made for decades. On the other side, the capable synthesis in diamond anvil cell provides various
in situ characterization approaches for the exploration of reaction mechasinms. As demonstrated
in Chapter 3, in situ high pressure diffraction experiment of benzene reveals that the reaction
from solid benzene to carbon nanothread is a unique non-topochemical solid state transformation.
Freedom of the structural constraints between precursor and product, it saves the complicated
procedures of crystal engineering. More importantly, small organic molecules, such as benzene,
who lack of commensurate structure and make them almost impossible to undergo topochemical
123
pathway thus lead to amorphous hydrocarbon materials. In the work illustrated Chapter 5, a new
pyridine high pressure phase has been discovered when studying the formation of carbon nitride
nanothread from pyridine. Even though the data are not ideal, some provisional structures have
been searched out and indexed that can fit both powder and single crystal diffraction data
reasonably. This work suggesting the necessity and advantage of diffraction–spectroscopy
combined research mode.
It’s a natural path that after the prototype of a new class of material being developed, the
next step it to expand the family and therefore people can tune their physical and chemical
properties for various applications. Looking into the development of carbon nanotube and
graphene, functionalization through either switching precursor materials or post-modification has
become individual sub-field. As a result, nanotube-based and graphene-based materials have
exhibited outstanding performances in more than one discipline.
In the case of nanothread, the difference in fluorescence properties between benzene
nanothread and pyridine nanothread is a good example showing that simply changing of
precursor molecules makes the physical properties of reaction products tunable. As mentioned
above, the optimized synthesis protocol is very robust. Very recently, Teflon nanothreads have
been successfully developed in Badding’s group applying the same recipe. The Teflon nanothread
sample shows different fluorescence emission compared to carbon nanothread and carbon nitride
nanothread. This result is quite encouraging. Because the molecules used for nanothread synthesis
all have been studied under high pressure before and ended up with amorphous recovery. The
considerations for future synthesis work would include: (1) introducing other heteroatoms into the
system, such as boron, oxygen and phosphorus; (2) elucidating the effect of aromaticity by
changing the symmetry of substitution group; (3) exploring the new systems beside six-
membered ring. For example five-membered rings and polycyclic aromatic rings.
124
The second concern about nanothread synthesis is to lower the reaction threshold
pressure. Both property characterization and device fabrication need large amount of sample. A
possible starting point is to try with the previously studied molecules with lower polymerization
pressure, such as 1,3,5-triazene1-4 and furan.5 Another approach is to work with co-crystal. For
example, the C6F6-C6H6 co-crystal shows strong quadrupole interactions6-7 at low temperature and
supposed to do the same at high pressure. In addition to pick particular reactants, exterior
conditions also effect the chemical transformation at high pressure. Heating-assisted and photo-
induced processes are the promising solutions.
Figure 6-1. Theoretical modelling of two-dimensional stacked 3,4-connected carbon nets from
plane nets.8 (Copyright © 1987, American Chemical Society)
Beyond the scope of one-dimensional material, this uniaxial stress guided synthesis
method could also applied in higher order of dimension. Figure 6-1, shows a theoretically model
of 3,-4 connected carbon net predicted in 1987.8 In 2018, two-layer graphene on SiC exhibits a
transverse stiffness and hardness comparable to diamond after nano-indentation.9 The pressure
reversibly transformed to 2-layer epitaxial graphene is in the order of 1-10 GPa. Density
functional theory suggested the compression produce both elastic deformation and sp2 to sp3
transformation. Inspired by these works, compressing graphene via similar method as nanothread
synthesis is worth to try. With higher pressure regime that diamond anvil cell can reach than
nano-indentaion technique, the chemical changes may be able to overcome the limitation that the
125
transformation described in 2018’s work can only work for 2-layer graphene but not applicable
for multilayers.
At last, I will use the figure below as the epilogue to this dissertation.
Figure 6-2. Future prospect on nanothread functionalization and application.
126
References
(1) Citroni, M.; Fanetti, S.; Bini, R. The Journal of Physical Chemistry C 2014, 118, 10284-
10290.
(2) Fanetti, S.; Citroni, M.; Bini, R. The Journal of Physical Chemistry C 2014, 118, 13764-
13768.
(3) Citroni, M.; Fanetti, S.; Bazzicalupi, C.; Dziubek, K.; Pagliai, M.; Nobrega, M. M.; Mezouar,
M.; Bini, R. The Journal of Physical Chemistry C 2015, 119, 28560-28569.
(4) Li, S.; Li, Q.; Xiong, L.; Li, X.; Li, W.; Cui, W.; Liu, R.; Liu, J.; Yang, K.; Liu, B. The
Journal of Chemical Physics 2014, 141, 114902.
(5) Ceppatelli, M.; Santoro, M.; Bini, R.; Schettino, V. The Journal of chemical physics 2003,
118, 1499-1506.
(6) Williams, J. H. Accounts of Chemical Research 1993, 26, 593-598.
(7) Battaglia, M.; Buckingham, A.; Williams, J. Chemical Physics Letters 1981, 78, 421-423.
(8) Merz Jr, K. M.; Hoffmann, R.; Balaban, A. T. Journal of the American Chemical Society
1987, 109, 6742-6751.
(9) Gao, Y.; Cao, T.; Cellini, F.; Berger, C.; de Heer, W. A.; Tosatti, E.; Riedo, E.; Bongiorno, A.
Nature nanotechnology 2018, 13, 133.
127
VITA
Xiang Li
EDUCATION
The Pennsylvania State University, University Park, PA, US
Ph.D., Chemistry 2013/8-2018/7
Nanjing University, Nanjing, Jiangsu, China
B.S., Chemistry 2009/9-2013/7
PUBLICATION
1. Wang, T., Duan, P., Xu, E., Vermilyea, B., Chen, B., Li, X., Badding, J. V., Schmidt-Rohr,
K., Crespi, V. H. Constraining Carbon Nanothread Structures by Experimental and
Calculated Nuclear Magnetic Resonance Spectra. Nano Lett., DOI:
10.1021/acs.nanolett.8b01736, (2018).
2. Duan, P., Li, X., Wang, T., Chen, B., Juhl, S. J., Koeplinger, D., Crespi, V. H., Badding, J. V.,
Schmidt-Rohr, K. The Chemical Structure of Carbon Nanothreads Analyzed by Advanced
Solid-State NMR. J. Am, Chem. Soc., 140, 7658 (2018).
3. Li, X., Wang, T., Duan, P., Baldini, M., Huang, H., Chen, B., Koeplinger, D., Crespi, V.,
Schmidt-Rohr, K., Hoffmann, R., Guthrie, M., Badding, J., Carbon Nitride Nanothread Single
Crystal Derived from Pyridine, J. Am, Chem. Soc., 140, 4969 (2018).
4. Chen, B., Wang, T., Crespi, V., Li, X., Badding, J., Hoffmann, R., All the Ways to Have
Substituted Nanothreads, J. Chem. Theory Comput.,14, 1131 (2018).
5. Li, X., Baldini, M., Wang, T., Chen, B., Xu, E.-S., Vermilyea, B., Crespi, V., Hoffmann,R.,
Molaison, J., Tulk, C., Guthrie, M., Sinogeikin, S., Badding, J.V., Mechanochemical
Synthesis of Carbon Nanothread Single Crystals, J. Am. Chem. Soc., 139, 16343 (2017).