REVISITING EFFECTS OF NITROGEN INCORPORATION AND GRAPHITIZATION ON

CONDUCTIVITY OF ULTRA-NANO-CRYSTALLINE DIAMOND FILMS

By

Tanvi Nikhar

A THESIS

Michigan State University

in partial fulﬁllment of the requirements

Submitted to

for the degree of

Electrical Engineering – Master of Science

2019

ABSTRACT

REVISITING EFFECTS OF NITROGEN INCORPORATION AND GRAPHITIZATION ON

CONDUCTIVITY OF ULTRA-NANO-CRYSTALLINE DIAMOND FILMS

By

Tanvi Nikhar

Diamond has many superior electric and electronic properties over other materials, yet its appli-
cation in electronic devices is severely limited due to diﬃculties in producing n-type conductivity
in diamond as available dopants all have activation energy much larger than the thermal energy,
kT, at room temperature. The substitutional doping of nitrogen in grain boundaries of granular
graphitic-rich ultra-nanocrystalline diamond (UNCD) ﬁlms provides for high conductivities even
at room temperature with an apparent activation energy to be much lower than kT. Such low activa-
tion energy is a consequence of the formation of new electronic states associated with carbon and
nitrogen near and above the Fermi level. However, the relative contribution of sp2 graphitic phase
and incorporated nitrogen to the conduction of the ﬁlms remains unclear.

In the present work, structural and electrical properties of nitrogen-incorporated (N)UNCD
ﬁlms are studied as a function of deposition temperature and nitrogen concentration in the precursor
synthesis gas mixture consisting of H2, CH4 and N2. Four sets of (N)UNCD ﬁlms with 0%, 5%,
10% and 20% nitrogen concentration in the synthesis gas were produced by microwave assisted
chemical vapor deposition on intrinsic (100)-oriented Si substrates.

Resistance tunability over a range of about 4 orders of magnitude was achieved by varying the
growth parameters. Raman spectroscopy and scanning electron microscopy (SEM) results show
that the UNCD ﬁlms undergo a progressive and highly reproducible material phase transformation,
from ultra-nano-diamond to nanocrystalline graphite as deposition temperature increases. Addition
of nitrogen increases the amount of sp2 bonded carbon in the ﬁlms thus enhancing the physical
connectivity in the grain boundary network that have high electronic state density and leads to
improved conductivity. However, crystallinity of sp2 carbon phase remains another (previously
underestimated) major factor in determining the conductivity of (N)UNCD ﬁlms.

ACKNOWLEDGEMENTS

Firstly, I would like to express my sincere gratitude to my advisor, Dr. Baryshev, for his constant
support and guidance throughout my Master’s study, from encouraging me to work with my creative
freedom to keeping me motivated during any setbacks.

This work would not have been possible without the resources made available to us at Fraunhofer
USA Center for Coatings and Diamond Technologies and the technical expertise of Mr. Robert
Rechenberg and other people at the facility.

I would also like to thank Dr. Hogan and Dr. Grotjohn for their encouragement and insights
towards improving this work and willingly addressing any and everything that was needed during
the process.

Finally, a special gratitude to my friends and family for keeping me sane in my moments of

jitters and for their constant help and support throughout the duration of this project.

iii

TABLE OF CONTENTS

LIST OF TABLES .
LIST OF FIGURES .
CHAPTER 1

.
.

.
.

.
.

.
.

.
.

.
.

.
.
.
.
INTRODUCTION .
.
.
.
.

.
.

.
.

1.1 Literature Review .
1.2 Thesis Objectives .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii
1
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7

vi

CHAPTER 2 CHEMICAL VAPOR DEPOSITION . . . . . . . . . . . . . . . . . . . . .

9
2.1 Heteroepitaxial Film Growth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.2 Growth Chemistry .
2.3 Microwave Plasma CVD System . . . . . . . . . . . . . . . . . . . . . . . . . .
. 13

.

.

.

.

.

.

.

.

3.1 Sample Growth .

3.2 Sample Analysis .

CHAPTER 3 UNCD GROWTH AND ANALYSIS . . . . . . . . . . . . . . . . . . . . . 15
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.1.1 Deposition Chamber
3.1.2
Substrate Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.1.3 Growth Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.2.1 Weight measurement
. . . . . . . . . . . . . . . . . . 17
3.2.2
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2.3
3.2.4 Raman Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.2.5 UV-vis Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2.6
Scanning Electron Microscopy . . . . . . . . . . . . . . . . . . . . . . . . 20

Four-point probe resistivity measurement
Proﬁlometry .

.

.

.

.

.

.

.

.
.

.
.

.
.

.
.

CHAPTER 4 RESULTS .
. .
.

4.1 Resistivity . .
4.2 Analysis of Raman Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
. 26
4.2.1 Raman spectra as a function of Td . . . . . . . . . . . . . . . . . . . . . . 27
4.2.1.1 Graphitization . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.2.2 Raman spectra as a function of N2 concentration in the feed gas
. . . . . . 33
4.2.3 Resistivity as a function of G band peak position . . . . . . . . . . . . . . 35
. 37
Surface Morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
4.3.1.1 Changes in surface morphology with varying Td . . . . . . . . . 37
37
4.3.1.2 Changes in surface morphology with increasing N2 concentration
4.3.2 Bulk Morphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.4 Reﬂectance Spectrum .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.5 Thickness .
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4.6 Deﬁning UNCD based on Raman spectra . . . . . . . . . . . . . . . . . . . . . . . 42

. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

4.3 Analysis of SEM images

4.3.1

.
.

.

.

.

.

.

.

iv

CONCLUSION .
.
APPENDIX .
.
.
.
BIBLIOGRAPHY .

.
.

.

.
.

.

.
.

.

.
.

.

.
.

.

.
.

.

.
.

.

.
.

.

.
.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
. 47
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v

LIST OF TABLES

Table 4.1: Series SA ﬁlms grown with 0% N2 concentration . . . . . . . . . . . . . . . . . 23

Table 4.2: Series SB ﬁlms grown with 5% N2 concentration . . . . . . . . . . . . . . . . . 23

Table 4.3: Series SC ﬁlms grown with 10% N2 concentration . . . . . . . . . . . . . . . . 24

Table 4.4: Series SD ﬁlms grown with 20% N2 concentration . . . . . . . . . . . . . . . . 24

vi

LIST OF FIGURES

Figure 2.1: Phase diagram of carbon. Regions of metastability of diamond and graphite
are bounded by extensions of the melting curves of diamond and graphite, re-
spectively as shown by dashed lines. Approximate regions for high-pressure,
high-temperature (HPHT) and chemical vapor deposition (CVD) synthesis of
diamond are shown [1].

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

Figure 2.2: Diamond growth on a nondiamond substrate in sequential steps shown through
SEM images: (a) nucleation; (b and c) growth of individual crystallites; (d)
faceting and coalescence of crystallites; (e and f ) continuous ﬁlm growth [2].

. 11

Figure 2.3: Hydrogen abstraction reactions leading to the incorporation of carbon in
the diamond lattice during the growth process of poly crystalline diamond.
(Courtesy: Fraunhofer CCD)

. . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Figure 2.4: Schematic of the MPCVD system used for the growth of UNCD ﬁlms in

present work.

. .

. .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Figure 4.1: Resistivity of all the ﬁlms of series SA, SB, SC and SD as a function of

deposition temperature

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

Figure 4.2: Raman spectra for 5% N2 samples as a function of Td . . . . . . . . . . . . . . 27

Figure 4.3: Raman spectra for 0% N2 samples as a function of Td . . . . . . . . . . . . . . 29

Figure 4.4: Raman spectra for 10% N2 samples as a function of Td

. . . . . . . . . . . . . 30

Figure 4.5: Raman spectra for 20% N2 samples as a function of Td

. . . . . . . . . . . . . 30

Figure 4.6: Comparison of Raman spectra of highly graphitized ﬁlms with single crytsal

diamond and highly oriented pyrolytic graphite . . . . . . . . . . . . . . . . . . 32

Figure 4.7: Raman spectra for graphitic and UNCD sites on the damaged (a) and pre-

damaged (b) cathodes due to electron ﬁeld emission. . . . . . . . . . . . . . . . 33

Figure 4.8: Raman spectra for samples deposited at 1098 K with 0, 5 and 10% N2

concentration .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

.

.

.

.

.

.

Figure 4.9: Raman spectra for samples deposited at 1126 K with 0, 5 and 20% N2

concentration .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

Figure 4.10: Resistivity as a function of G peak position . . . . . . . . . . . . . . . . . . . . 36

vii

Figure 4.11: SEM images of samples grown with 0% N2 concentration . . . . . . . . . . . . 38

Figure 4.12: SEM images of samples grown with 5% N2 concentration . . . . . . . . . . . . 38

Figure 4.13: SEM images of samples grown with 10% N2 concentration . . . . . . . . . . . 38

Figure 4.14: SEM images of samples grown with 20% N2 concentration . . . . . . . . . . . 38

Figure 4.15: SEM and Raman spectrum of sample: SB4, grown with 5% N2 . . . . . . . .

. 39

Figure 4.16: SEM and Raman spectrum of sample: SD1, grown with 20% N2

. . . . . . . . 39

Figure 4.17: SEM and Raman spectrum of sample: SB9, grown at high Td . . . . . . . . . . 39

Figure 4.18: Comparison of reﬂectance spectra of UNCD and highly graphitized ﬁlm . . . . 41

Figure 4.19: Conﬁrmation of graphitization through thickness measurement

. . . . . . . . . 42

viii

CHAPTER 1

INTRODUCTION

Diamond exhibits a unique combination of superior material properties including extreme hardness,
high thermal conductivity, low coeﬃcient of friction, high degree of chemical inertness [3]; that
make it suitable for a diverse range of applications [4, 5, 6]. The potential of diamond to be
used as electron emitter or ampliﬁer, high ﬂux particle and UV or X-ray detector, electrodes
for electrochemistry and in active electronic devices has attracted the interest of the research and
technology-oriented communities. More speciﬁcally, it is the low, and sometimes negative, electron
aﬃnity that makes diamond a good electron source. The high breakdown ﬁeld in comparison with
both silicon and gallium arsenide, large band gap, large thermal conductivity and large momentum
relaxation time indicate a potential for diamond in high-frequency, high-temperature, high-power
applications [7, 8].

Many proof-of principle two-terminal and three-terminal devices have been made and tested
[9, 10, 11, 12]. However, the development of diamond devices poses great challenges due to problem
with a large carrier activation energy of available dopants which necessitates specialized concepts
(e.g. surface transfer doping) to allow room temperature operation [13, 14, 15]. p-type doping
can be achieved with addition of boron, yielding an activation energy of 0.37 eV [16]. Along with
hydrogen-induced surface conductivity (surface transfer doping), high performance devices have
been reported with extremely low activation energy [17]. However, to date, problems with n-type
doping of diamond [18, 19], in part due to its high atomic density, have hindered its development
as a ubiquitous device material. Using nitrogen as a dopant is diﬃcult as doping of bulk diamond
with nitrogen produces a deep donor level of 1.7 eV below the conduction band. Other n-type
dopants have been attempted, such as sulfur and phosphorus, with only limited success as the
room-temperature conductivities are too low for the application of these materials in conventional
electronic device architectures. A shallow donor species is still sought for diamond.

Another relevant issue related to the development of diamond ﬁlm-based electronics, is that

1

currently single crystal diamond (SCD) ﬁlms have been used to produce high power diamond-based
electronic devices with promising performance [20, 21, 8]. However, SCD ﬁlms can be grown only
on SCD diamond and other relatively expensive substrates with very small areas, that is available
for large scale integration. On the other hand, polycrystalline diamond (PCD) ﬁlms have been
demonstrated to grow on large areas and relatively inexpensive substrates with high uniformity
[3], which with proper optimization have the potential to provide the base for the development of
transformational diamond-based electronics.

Ultra-nanocrystalline diamond (UNCD) is another form of diamond which is shown to ex-
hibit outstanding and tunable n-type conductivity, and excellent biocompatibility [22, 23, 24, 25],
thus enabling a broad range of applications to a new generation of electronic devices including
bio-medical devices. The unconventional properties of UNCD ﬁlms are due to its microstruc-
ture characterized by randomly oriented nanosized diamond crystallites bonded together by grain
boundaries [26, 27]. The complex bonding structure at these grain boundaries [28, 29] imparts a
striking set of attributes to UNCD ﬁlms, viz. dramatic diﬀerence in electron and phonon proper-
ties of UNCD (resulting in dramatic electrical and thermal conductivities) exists as compared to
intrinsic SCD. The key diﬀerence is that new electronic state bands are introduced into the 5.5 eV
band gap of bulk diamond due to dangling bonds and π-bonded carbon atoms present in the grain
boundaries [29], which is further enhanced by the preferential incorporation of nitrogen in the
GBs [30]. The incorporation of nitrogen via the addition of N2 gas to the feed gas mixture brings
about changes in morphology [31] and electronic structure within the grain boundaries that leads
to orders of magnitude of change to the electrical conductivity [32]. The discussion on the exact
mechanisms of the eﬀects of N2 on electronic properties of UNCD remains open.

1.1 Literature Review

The growth of nanocrystalline diamond by CVD process under hydrogen-poor and argon/carbon-
rich conditions was ﬁrst reported by Gruen et al.
in 1994 [33, 34, 35]. The control of diamond
microstructure through the crystallite size was further analyzed in 1999 [27] by replacing hydrogen

2

with argon in the growth process. Conventional diamond ﬁlm CVD typically uses a gas mixture
of 1% CH4 in 99% H2 where the incorporation of carbon in the diamond lattice is made possible
through a series of hydrogen abstraction reactions on CH4. Replacement of hydrogen with argon
was done by using fullerene C60 as a precursor molecule which can dissociate in argon plasma to
give carbon dimer C2 from which diamond can nucleate and grow in the absence of hydrogen.

As argon replaces hydrogen in the initially 1% CH4/H2 balanced plasma (eventually yielding 1%
CH4/Ar balanced plasma), the morphology of the ﬁlm transitions from well faceted microcrystalline
diamond (MCD) with grain sizes of 0.5-2 µm and columnar growth structure to nanocrystalline
diamond (NCD) with no observable morphological growth features due to the nano-scale size of
the crystals. These NCD ﬁlms were found to be composed of two components: nanodiamond
crystallites (>95 vol.%) and amorphous carbon grain boundaries (<5 vol.%), made of either glassy
carbon or tetrahedrally-bonded diamond-like amorphous carbon. The label ultra-nanocrystalline
diamond (UNCD) was used to deﬁne a class of NCD having phase-pure diamond crystallites of
size of 3-5 nm and 0.3 to 0.5 nm wide GB. The GBs retain ∼10 vol.% of all the carbon atoms of the
ﬁlm having mixed sp2 and sp3 hybridization and unpaired electrons forming dangling bonds [36].
The synthesis and characterization of nitrogen-doped UNCD ﬁlms (N-UNCD) was ﬁrst reported
by Bhattacharyya et al. in 2001 [32]. They examined the eﬀect of changes in plasma chemistry
on the ﬁlm morphology and electrical transport properties as nitrogen was incorporated into the
ﬁlms. The ﬁlms were grown in CH4(1%)/Ar/N2 plasma varying N2 from 1% to 20% with all other
parameters ﬁxed. Dramatic increase in room-temperature conductivity was observed, increasing
-1cm-1 (for 20% N2), which represents an increase by
from 0.016 Ω
roughly ﬁve orders of magnitude over undoped UNCD ﬁlms. The latter value of conductivity
was much higher than any other previously reported for n-type diamond and was comparable to
heavily boron-doped p-type diamond [32]. Hall and Seebeck eﬀect measurements conﬁrmed the
n-type nature of the conductivity for the nitrogen containing samples, as well as high and nearly
temperature-independent values for the Hall mobility, with room temperature values in the range of
1 to 10 cm2V-1s-1 in the case of highly conductive samples. All the reports that followed, conﬁrmed

-1cm-1 (for 1% N2) to 143 Ω

3

that addition of nitrogen to plasma yields greatly enhanced conductivity (N)UNCD; grown either
by RF or HF CVD [37, 38].

Even so, how much nitrogen is required to bring the conductivity to the champion metrics of
∼100 S/cm is still under debate. The same Argonne group reported that N2 plasma content can lie
within a very large range, between 8 to 20 %; as the conductivity increases as nitrogen is increased
from 0 to 8% and saturates thereafter for all the higher nitrogen concentrations in the plasma upto
20% [37]. The main reason brought up to explain this eﬀect was saturating nitrogen incorporation.
Secondary ion mass spectroscopy (SIMS) showed a continuous increase in the incorporated nitrogen
in the ﬁlm as nitrogen was increased in the plasma with a maximum of 2.2×1020 cm-3 for 18% N2
in the plasma [39]. This is in contradiction with other reports where the incorporated nitrogen was
shown to initially increase but then saturate at ∼2×1020 atoms cm-3 for somewhere between 5-10%
N2 in the plasma which might also be a cause of the observed saturation in conductivity [32, 37].
Where does nitrogen go in the ﬁlm? Originally, based on the tight binding molecular dynamics
simulation results, it was determined that nitrogen is much more favored to be incorporated into
the grain boundaries rather than grains. Indeed, high-resolution transmission electron microscopy
(HRTEM) found that as the nitrogen content in the plasma is changed from 0% to 20%, the grain-
boundary width increases from about 0.5 to 2.2 nm. Simultaneous grain growth from 4 to 16
nm was explained by the presence of CN at enhanced nitrogen content that does not itself lead to
diamond growth but rather reduces the renucleation rate that results from C2 dimer based diamond
growth [31]. HRTEM also demonstrated that the GBs, being less dense than the grains, are built
of sp2 hybridized carbon [32].

Adding to that, electron energy-loss spectroscopy (EELS) spectra for the grains show phase pure
diamond grains with sp3 bonded carbon which does not change as nitrogen is added to the plasma.
The grain boundary spectra indicate that the amount of sp2 bonding increases only slightly while
the sp3 peak remains unchanged with nitrogen addition, keeping the ratio of sp2 to sp3 bonded
carbon relatively unchanged. It is therefore concluded that the increase in sp2 content in the ﬁlm
as a whole is derived from a slight increase in the grain-boundary material in the ﬁlm, and not a

4

change in the local bonding structure within the grains or grain boundaries [31].

Near edge x-ray absorption ﬁne structure (NEXAFS) results show that all ﬁlms, even those
deposited using high amounts of nitrogen in the growth plasma, have only a small fraction of sp2-
bonded carbon in the ﬁlm, being remarkably similar to phase pure diamond and that as nitrogen
is added to the ﬁlm, the sp2 bonding increases only slightly. The fraction of sp2 to sp3-bonded
carbon is shown to increase from about 6.7% in undoped UNCD to 13.5% in 20% N2 UNCD. This
increase is proposed to be due to a larger total volume of grain boundaries with the same sp2/sp3
ratio based on previous HRTEM results [39].

Theoretical models predict that UNCD grain boundary defects produce a number of states
in the forbidden band of diamond: π and π∗ states for the sp2-bonded atoms and σ∗ states for
the sp3-coordinated atoms forming a continuous band of delocalized energy states, across which
conduction takes place. When nitrogen is incorporated at the grain boundaries, new electronic
states are further introduced into the band gap. At the same time, a strong morphology change
resulting from more than 10% addition of nitrogen to the plasma could lead to an increase in the
robustly physically connected grain boundaries aiding conductivity.

There is therefore a contradiction whether nitrogen doping takes place or whether nitrogen
mediates enhanced conductivity by promoting grain growth (and their connectivity). The con-
tradiction is further introduced by the results of Ikeda et al. who found the strong dependence
of conductivity on the deposition temperature (Td) which was due to the changes induced in the
structure of nanodiamond while addition of nitrogen did not play much role [40]. More speciﬁcally,
the room temperature resistivity decreases exponentially by more than three orders of magnitude
when Td rises to 1200 K, and then saturates at ∼10-2
Ωcm for further increase in Td. The structure
of GB which is mainly composed of amorphous sp2-bonded carbon is expected to vary signiﬁcantly
with Td because Td promotes a transformation from sp3- to sp2-bonded carbon, thus increasing
conduction.

Room temperature resistivity recorded for ﬁlms grown at Td=1100 and 1200 K as a function
of gas-phase N2 concentration show that for both Td, the resistivity decreases by two to three

5

orders of magnitude upon 3%N2 addition and saturates for further N2 up to 30% [40]. This is in
accordance with previously observed saturation in conductivity for higher nitrogen content in the
plasma [37]. The minimum resistivity in the saturation region is lower for higher Td ﬁlm. Since
the incorporated nitrogen concentration measured by XPS is constant at about 0.3 at. % at all gas
phase N2 concentrations and also for both Td (saturation eﬀect seen before [32, 37]), it is concluded
that sp2 phase (dependent upon Td in this case) is more responsible for increasing conduction than
the nitrogen concentration itself. It also agrees with previous observations in that, it is not as much
the incorporated nitrogen as is the relative increase in the sp2 content induced by nitrogen that
ultimately aﬀects conductivity.

The Td studies done by Ikeda et al. though, for diﬀerent N2 concentrations is only limited to
two Td values with no analysis of the structure and various phases present in the ﬁlm and how their
change is related to conductivity. Exceptional conductivity tunability down to 100 S/cm based
on only varying Td was recently further discussed in detail by J.J. Alcantar-Pena et al.. [41] for
undoped (no N2 in plasma) UNCD.

The same study emphasized the critical importance of Raman spectroscopy in studying UNCD
even though it had been a premier method in characterizing diamond. The Raman of (N)UNCD
features two major optical bands called D and G. Monitoring I(D)/I(G) ratio as well as the D and G
band peak positions reveal particularly that the material is highly "dynamic", i.e it undergoes many
simultaneous modiﬁcations as nitrogen and/or sp2 phase is injected via adjusting plasma N2 content
or substrate temperature during growth. The D and G bands not only monitor diamond-to-graphite
ratio but also trace down structural disorder. For example, a clear shift of the G band peak towards
a higher wavenumber for higher Td, which has never been addressed in literature. The general
serious problem exists as there is vast variation of Raman spectra all interpreted as UNCD form
of diamond, thus raising a challenging question as to what UNCD is and what is the range of
application of this term as the Raman spectrum modiﬁes from nearly NCD to nearly graphite type
of ﬁngerprint spectra. UNCD deﬁnition further fades away as UNCD was introduced to only grow
in Ar rich or balanced plasmas [33, 34, 35, 42] while UNCD was shown to grow in H2 balanced

6

plasmas as well [43].

To conclude, there is no consistent and complete understanding of the factors responsible for
conduction in (N)UNCD ﬁlms. Following details about these factors are still not clear: i) the phase
of the sp2 carbon and ii) the changes in energy space, physical space or both due to N2 incorporation.
Electron transport can be explained based on two competitive processes. First, an increase in the
spatial connectivity of sp2-bonded carbon is expected from the rise in the relative amount of sp2-
bonded carbon. This causes a better overlap of the electron orbitals among the sp2 clusters and
enhances the delocalization of the π and π* states making electron transport more probable and
increasing the electron mobility. Second, the decrease in the relative amount of diamond, increases
the mean distance between diamond grains forming a spatially heterogeneous band structure. The
system then behaves as grains and grain boundaries embedded in amorphous matrix. The relative
contribution of these factors is also not clear. From the literature overview in the past 18 years, the
community has moved from nitrogen doping [32] to geometrical and graphitization eﬀects [43, 41].
The issue can be reduced to a very simple question – is it a conventional doping or just size eﬀect,
or both, that promotes high conductivity?

1.2 Thesis Objectives

1) The clear feasibility of engineering grain boundary structure to obtain tailored resistivity of
UNCD ﬁlms paves a way to the development of diamond based electronic power and bio-medical
devices. Resistivity tunability has been studied either through the variation of plasma nitrogen
concentration or deposition temperature, but never both. Both of these two synthesis variations
have been shown to alter the physical and electronic density of state structure of UNCD being
intertwined. The objective of this work is therefore to systematically vary both N2 content and
deposition temperature to better assess the physical mechanisms transforming UNCD from high to
low resistive state.

2) The deﬁnition of UNCD based on Raman spectra is still vague as there is no deﬁnite set of
features that can be conclusively identiﬁed as representative of UNCD. In addition to this, these

7

features are not analyzed in detail to completely interpret the changes in the position, shape and
intensity of the peaks with changing growth conditions. For instance, the signiﬁcant shift in the G
band for both varying crystallite sizes and for varying N2 concentration in the plasma as seen in the
visible Raman spectra is either not addressed at all [44, 40, 42] or not understood [39]. Another
objective of this work is to revisit interpretation of Raman spectra that contain information of all
carbon phases present in UNCD and their crystallinity to shed light onto conductivity in UNCD
and the basic deﬁnition of the term UNCD.

8

CHAPTER 2

CHEMICAL VAPOR DEPOSITION

Synthesis of diamond [1] using high-temperature, high pressure processes, in which diamond is the
stable form of carbon, as can be seen in Fig. 2.1, has proved to be commercially successful but is not
amenable to direct deposition of diamond on substrates or to large-area deposition. A distinctive
method of diamond synthesis at low pressures, in which diamond is metastable with respect to
graphite has oﬀered the capability of direct deposition of diamond on a variety of substrates and
has opened the possibility of new applications of diamond. In low-pressure diamond synthesis,
common sources of carbon, eg. methane, are transformed by chemical vapor deposition (CVD)
to diamond at atmospheric or sub atmospheric pressures and at temperatures that are common to
materials processing.

Formation of diamond from nascent carbon containing species under metastable conditions is
both thermodynamically allowed and readily achieved under proper deposition conditions which
can be achieved by a variety of methods. These methods share certain generic features which
include:

(i) Gaseous Feedstock- The gaseous feedstock includes at least one source of carbon-containing
species, like methane; and a diluting gas such as hydrogen, as used by most methods. The
energy provided to the system breaks down the carbon-containing gas into a variety of
subspecies which are involved in donating a carbon atom to the growing diamond ﬁlm on the
substrate. It also breaks down the hydrogen gas, producing large amounts of atomic hydrogen
which play which assist the growth process.

(ii) Energy Source- A means to provide energy to the deposition chamber is required to break
down the input gases. Various CVD methods have been developed based on the source
of energy such as thermally assisted hot-ﬁlament and plasma torch deposition, combustion

9

Figure 2.1: Phase diagram of carbon. Regions of metastability of diamond and graphite are bounded
by extensions of the melting curves of diamond and graphite, respectively as shown by dashed lines.
Approximate regions for high-pressure, high-temperature (HPHT) and chemical vapor deposition
(CVD) synthesis of diamond are shown [1].

synthesis, laser-assisted and optically pumped techniques, and dc arc jet, rf and microwave
plasma-assisted deposition based on electrical energy inputs.

(iii) Substrate on which the ﬁlm grows.

2.1 Heteroepitaxial Film Growth

Diamond ﬁlms can be grown on non-diamond substrates using CVD processes by treating or seeding
the substrate to initiate and enhance the nucleation of diamond. There are several substrate surface
pre-treatment methods like polishing with diamond powder, ultrasonic treatment in a slurry of micro
or nano-sized diamond powder, bias enhanced nucleation, pulsed-laser irradiation, carburization,

10

Figure 2.2: Diamond growth on a nondiamond substrate in sequential steps shown through SEM
images: (a) nucleation; (b and c) growth of individual crystallites; (d) faceting and coalescence of
crystallites; (e and f ) continuous ﬁlm growth [2].

chemical etching, coating with carbon materials. These seeds/defects introduced on the substrate,
when exposed to the diamond growth environment, convert to nuclei of crystalline diamond or serve
as nucleation sites via a heterogeneous nucleation mechanism. High surface nucleation density
leads to better diamond-substrate adhesion, improves the homogeneity of the ﬁlm and reduces the
formation of voids at the substrate interface, thus reducing surface roughness.

Conventional growth of polycrystalline diamond generally consists of up to ﬁve distinguishable
stages [1]:
(1) The system undergoes an incubation period followed by (2) three-dimensional
nucleation of individual crystallites on the substrate surface which ceases when all the activated
nucleation active sites have been occupied or when the discrete crystallites grow together and
completely cover the surface. (3) Subsequently, the individual crystallites grow three-dimensionally
and develop faceting due to the relatively high rate of surface carbon diﬀusion from the surrounding
substrate surface.
(4) Once the crystals grow large enough to impinge upon one another, they
coalesce, forming grain boundaries and then (5) continue growing homoepitaxially as a continuous

11

ﬁlm, exhibiting competitive crystal growth. These diﬀerent stages are illustrated by the micrographs
shown in Fig. 2.2.

Often some degree of non-diamond carbon is incorporated during this nucleation and coales-
cence stage. The growth chemistry (gaseous carbon species, atomic hydrogen, dopant/impurity
concentration etc.) and substrate temperature determine the morphology, quality, doping and
degree of bulk defects in the ﬁlm.

2.2 Growth Chemistry

The initiation of the gaseous chemistry is dominated by the dissociation of 1-40 % of the molecular
hydrogen into atomic hydrogen. The subsequent gaseous chemistry is driven by reactions of
atomic hydrogen with hydrocarbon species and reactions among the hydrocarbon species. With a
hydrocarbon precursor such as CH4, gas-phase hydrogen abstraction reactions lead to the generation
of the methyl radical CH·
3, which adsorbs on a carbon radical site also created by hydrogen
abstraction from the hydrogen-terminated growing diamond surface.

Figure 2.3: Hydrogen abstraction reactions leading to the incorporation of carbon in the diamond
lattice during the growth process of poly crystalline diamond. (Courtesy: Fraunhofer CCD)

Additional hydrogen abstraction reactions (shown in Fig. 2.3) allow the carbon in the absorbed
methyl radical to form carbon-carbon bonds and thus be incorporated into the diamond lattice

12

[27, 45]. Because graphite is thermodynamically more stable than diamond in the given growth
environment, the growth of metastable diamond is stabilized by the presence of atomic hydrogen
which removes graphitic nuclei when they do form because of the preferential etching of graphite
over diamond.

2.3 Microwave Plasma CVD System

Figure 2.4: Schematic of the MPCVD system used for the growth of UNCD ﬁlms in present work.

The role of the microwave plasma-assisted machine in the deposition of diamond is to create the
chemical/thermal environment needed for diamond deposition. The microwave plasma-assisted
diamond CVD (MPCVD) system used for the growth of UNCD ﬁlms in this work is shown in

13

Fig. 2.4. It consists of various subsystems, that work in unison, under computer control, to provide
the required deposition environment.

A magnetron power supply generates the microwave power which is supplied to the discharge
chamber via the microwave energy coupling system consisting of a waveguide, cavity resonator
and antenna to guide the microwave energy required to create and maintain the plasma discharge.
The tuning elements for impedance matching of the waveguide to the reactor load inﬂuence the
plasma discharge shape and size, directly impacting the uniformity of diamond deposition across
the deposition surface.

The discharge chamber conﬁnes the plasma discharge and positions the substrate surface for
optimum deposition. The substrate is placed on a stage which can be used in actively cooled or
ﬂoating conﬁguration to control the substrate temperature. A microwave transparent dome couples
the microwave power while providing a gastight chamber. The ﬂow rate of the feed gases to be
delivered to the discharge chamber is controlled via mass ﬂow controllers (MFC). The pressure
control system pumps the reacted gas from the deposition chamber in a continuous manner and
maintains the desired pressure using an adjustable (throttle) valve located between the pump and
the deposition chamber.

14

CHAPTER 3

UNCD GROWTH AND ANALYSIS

3.1 Sample Growth

3.1.1 Deposition Chamber

Ultra-nanocrystalline diamond ﬁlms were grown using MPCVD technique in a microwave cavity
plasma reactor (MCPR) developed at Michigan State University and Wavemat/Norton in 1986-1995
[1]. The microwave discharge is produced inside a quartz dome which is located at one end of a
microwave cavity. The cavity is formed by a cylindrical wall and a movable top short deﬁning the
top end of the cavity. The microwave energy is transmitted from a 2.45-GHz microwave generator
through a rectangular waveguide to a transition into a coaxial waveguide that ends as a length
adjustable excitation probe in the cylindrical cavity.

A hemispherical plasma discharge is created and is placed in direct contact of the substrate by
adjusting the length of the probe and the location of the sliding short (Ls). The diameter of the
microwave cavity is 17.8 cm with the substrate holder being able to accommodate substrate sizes
up to 7.6 cm in diameter. The sliding short height Ls is adjusted to a length of approximately 21
cm. This cavity height produces a TM013 resonant microwave mode.

3.1.2 Substrate Preparation

The substrates used for growth were intrinsic (100)-oriented Si wafers of 0.6mm thickness and
approximately 10×10 mm2 cross-section having sheet resistance of 5MΩ/(cid:3). These substrates
were ultrasonically cleaned in acetone to remove oil and organic residues, followed by methanol
to dissolve any acetone residues left for 2 minutes each. They were then rinsed with deionized
(DI) water and blow dried with nitrogen. The substrates were prepared for deposition by ultrasonic
seeding in Carbodeon µ Diamond hydrogen terminated nanodiamond aqueous dispersion (2.5 wt.%)

15

with nanodiamond crystal size ranging from 4.0 to 6.0 nm for 1 minute, immediately followed by
DI rinse and blow drying with nitrogen.

3.1.3 Growth Conditions

The seeded samples were loaded in the CVD system on a molybdenum holder mounted on a
ﬂoating stage. A base pressure of 10-7 torr was achieved using a roughening and a turbo pump.
Four diﬀerent series of samples were grown from a hydrogen rich H2/N2/CH4 feed gas mixture
with 0, 5, 10, and 20% of nitrogen (volume %) in the gas phase which corresponds to ﬂow rates
of 0, 10, 20 and 40 standard cubic centimeters per min (sccm) of N2. The CH4 ﬂow rate was kept
constant at 10 sccm (5% by volume), while the ﬂow rate of N2 and H2 was varied to maintain a
total ﬂow rate of 200 sccm. Diﬀerent deposition temperatures were achieved ranging from 1043
K-1295 K by varying the total pressure of the vacuum chamber (between 35-60 torr) and the input
microwave power (between 2500-3000 W). The substrate temperature was measured at the interval
of every 10 minutes during a growth process of 60 minutes, using an infrared pyrometer with the
sensor calibrated at an emissivity value of 0.6.

After each deposition run, the samples were left to cool in nitrogen at ∼500 torr for about 10

minutes and then in vacuum for about an hour.

3.2 Sample Analysis

3.2.1 Weight measurement

The seeded Si substrates were weighed before and after deposition using Mettler Toledo XS 105
Scale with 10-5 g readability to ﬁnd the weight of the deposited material. Assuming most of this
deposited material to be sp3 bonded carbon forming diamond grains (the rest being the sp2 bonded
carbon and other defects in the grain boundaries), the density of pure diamond, 3.51g/cm3 was
used to get a rough estimate of the thickness of the material grown using the following relation-

16

T hickness(µm) =

∆w · 104
A · d

(3.1)

where,
∆w = wt. after deposition-wt. before deposition (g)
d = 3.51 g/cm3
A = surface area of the ﬁlm (cm2)

3.2.2 Four-point probe resistivity measurement

Four-point probe method is an electrical impedance measuring technique that uses separate pairs
of current-carrying and voltage-sensing electrodes to make more accurate measurements than
the simpler and more usual two-terminal sensing. Separation of current and voltage electrodes
eliminates the lead and contact resistance from the measurement. This is an advantage for precise
measurement of low resistance values.

Current applied via a pair of current or force leads generates a voltage drop across the impedance
to be measured according to Ohm’s law V=IR. Sense connections are made immediately adjacent
to the target impedance using a pair of voltage leads, so that they do not include the voltage drop in
the force leads or contacts. Since almost no current ﬂows to the measuring instrument, the voltage
drop in the sense leads is negligible. Usually, the sense wires are arranged as the inside pair, while
the force wires are the outside pair, otherwise, the accuracy can be aﬀected, because more of the
lead resistance is included in the measurement.

Signatone SP4 four-point probe head and HP 4145B Semiconductor Parameter Analyzer with
a source voltage of 0.5 V was used to measure and analyze the sheet resistance of the thin ﬁlms
grown. The resistivity was then calculated using the following relation:

ρ = Rs · t · 10−4

(3.2)

where,
ρ = Resistivity (Ωcm),

17

Rs = Sheet Resistance (Ω/(cid:3)),
t = ﬁlm thickness (µm)

3.2.3 Proﬁlometry

Proﬁlometry is a technique used to extract topographical data from a surface in order to get surface
morphology, step heights and surface roughness. Critical dimensions such as step height, curvature,
ﬂatness can be computed from the surface topography. There are two types of proﬁlometers: stylus
and optical. Stylus proﬁlometers use a probe to detect the surface, physically moving a probe along
the surface in order to trace the surface proﬁle. This is done mechanically with a feedback loop
that monitors the force from the sample pushing up against the probe as it scans along the surface.
Optical proﬁlometry uses light instead of a physical probe to measure the topography.

Dektak 6M proﬁlometer was used to record the surface roughness of the ﬁlms electro-
mechanically by moving the sample beneath a diamond tipped stylus with 5 µm tip radius and
a constant stylus force of 15 mg. The stage was set to move the sample at a rate of 6.86 µm/s over
a scan length of 480 µm. The roughness of the sample surface, which translates to the vertical
motion of the stylus, was measured in terms of arithmetic average deviation from the mean line
within the assessment length (Ra) and average height diﬀerence between the ﬁve highest peaks and
ﬁve lowest valleys of the scan (Rz).

3.2.4 Raman Spectroscopy

Raman spectroscopy is a technique used to observe vibrational, rotational, and other low-frequency
transitions in molecules upon interaction with photons which provide a structural ﬁngerprint by
which molecules can be identiﬁed. It is a relatively easy, non-destructive, non-contact and quick
measurement method and can be performed at room temperature and ambient pressure.

When light is incident upon molecules, most of the photons are scattered elastically, also called
Rayleigh scattering such that the scattered photons have the same energy. However, a small fraction
of the photons is scattered inelastically, where the energy of the incident photons is either gained or

18

lost so that the scattered photons are shifted in frequency. Such inelastic scattering is called Raman
scattering. Because this frequency modulation is speciﬁc to molecular vibration and phonon in
crystal, it is possible to analyze composition of material or crystal lattice information by analyzing
the spectrum of Raman scattered light.

Typically, a sample is illuminated with a laser beam serving as a source of monochromatic
light. Electromagnetic radiation from the illuminated spot is collected with a lens and sent through
a monochromator. Elastic scattered radiation corresponding to the Rayleigh scattered photons is
ﬁltered out, while the rest of the collected light is dispersed onto a detector. Raman scattered
light detected on the shorter wavelength side than Rayleigh scattered light is called anti-Stokes line
resulting from an increment in the energy of the incident photon, while that detected on the longer
wavelength side is called Stokes line which is due to the loss of energy of the incident photon. In
general, Stokes lines with higher intensity are used for analysis.

The information of each dispersed wavelength is converted to wavenumber (1/wavelength) and
the diﬀerence with the wave number of incident light is plotted on the horizontal axis providing
vibrational information of the molecule and the vertical axis represents the intensity of activity.
Furthermore, to analyze microscopic areas of the sample, Micro Raman Spectroscopy is used which
combines aspects of both the microscope and the Raman spectrometer.

Micro Raman spectra of UNCD ﬁlms were obtained by using a Horiba Raman spectrometer
with 532 nm laser and a grating with a groove density of 1800 lines/mm. A 50× objective lens
was used with the optical microscope. The spectra were collected on at least 2 diﬀerent locations
(mostly mid and edge) on the sample from 400 to 3000 cm-1 for 30 s accumulated over 1 cycle.

3.2.5 UV-vis Spectroscopy

Ultraviolet-visible spectroscopy refers to absorption or reﬂectance spectroscopy in part of the
ultraviolet and the full, adjacent visible spectral regions, and is routinely used in analytical chemistry
for the quantitative determination of diﬀerent analytes. In this work, relative specular reﬂectance
measurements, where reﬂectance is calculated from the strength ratio after comparing the light

19

reﬂected from the reference sample with the light reﬂected from the measurement sample, were
taken to calculate the thickness of the ﬁlms. This method of determining thickness is based on the
interference patterns observed in N-UNCD submicron ﬁlms. The maxima and minima positions in
the interference patterns detected are shown to be substrate independent, representing the bulk of
N-UNCD ﬁlm and are used in thickness calculations as shown in Eq. (3.3) and (3.4) .

d =

λ
4nk

k = −

1

1 − λk
λk+1

(3.3)

(3.4)

where,
d = thickness (nm)
λ = maxima or minima position
λ k,λ k+1 = neighboring maximum and minimum positions
n’ = refractive index spectrally dependent on the wavelength [46]
k = integer number of a peak/valley, even values of k describe the maxima, and odd the minima

Reﬂectance measurements were performed using Shimadzu UV-vis 2600 spectrometer in the
range 200 to 1200 nm at an incidence angle of 8°. The above relations were used to calculate
thickness of each ﬁlm based on their respective interference patterns.

3.2.6 Scanning Electron Microscopy

Electron microscopy is a method of using an electron beam to form magniﬁed images of specimens.
Since the wavelength of electrons is much smaller than the wavelength of light, electrons provide
as much as thousand-fold increase in resolving power, i.e. the ability to distinguish between ﬁne
details. A scanning electron microscope (SEM) is a type of electron microscope that produces
images of a sample by scanning the surface with a focused beam of electrons. The electrons
interact with atoms in the sample, producing various signals that contain information about the

20

surface topography and composition of the sample. The electron beam is scanned in a raster scan
pattern, and the position of the beam is combined with the intensity of the detected signal to produce
an image.

The interaction of electrons with a sample can result in the generation of many diﬀerent types of
electrons, photons or irradiations. In the case of SEM, the two types of electrons used for imaging
are the backscattered (BSE) and the secondary electrons (SE). Backscattered electrons belong to
the primary electron beam and are reﬂected after elastic interactions between the beam and the
sample. On the other hand, secondary electrons originate from the atoms of the sample resulting
from inelastic interactions between the electron beam and the sample. BSE come from deeper
regions of the sample, showing high sensitivity to diﬀerences in atomic number, while SE originate
from surface regions providing more detailed surface information.

High resolution scanning electron microscopy was performed using a JEOL JSM 7500F with
0.9 nm resolution. Imaging was done at a working distance of 4.5 mm with an accelerating voltage
of 5 kV, emission current of 20 µA for the high conductivity samples and 10 µA for the low
conductivity samples to minimize charging eﬀects. A secondary electron detector was used to
image the surface and cross-section of the ﬁlms grown.

21

CHAPTER 4

RESULTS

Ultra-nanocrystalline diamond (UNCD) ﬁlms were grown at various deposition temperatures for
four diﬀerent nitrogen concentrations in the feed gas mixture resulting in four series of samples
referred here as SA, SB, SC and SD for 0, 5, 10 and 20% feed gas N2 respectively. The following
tables (Table 4.1, 4.2, 4.3 and 4.4 for series SA, SB, SC and SD respectively) summarize the
resultant properties of each series of ﬁlms and the values of input parameters used to tune the
growth conditions to obtain these ﬁlms.

The thickness of the ﬁlms grown is measured by the weight method (tweight) for all the ﬁlms
and UV-vis spectroscopy (tUV-vis) for the ﬁlms with a measurable interference spectrum. Since
tUV-vis values are more accurate, they are used for resistivity calculations while tweight is used for
ﬁlms for which tUV-vis could not be calculated. The percentage values of CH4 and N2 in the tables
refers to the volume percentage of these gases in the feed gas mixture to obtain a total 200 sccm
ﬂow rate with the rest of the gas being H2.

Control of deposition temperature for growth was achieved by varying the input parameters
viz. the forward microwave power and total pressure in the chamber. An increase in the substrate
temperature was observed for a set value of all the input parameters, as nitrogen was increased in
the plasma by replacing hydrogen to keep the total gas ﬂow rate constant. Since the only change
in the input parameters is the change in partial pressure of nitrogen and hydrogen, the increase in
the substrate temperature might be explained in terms of how the energy is transferred from the
gas molecules/ions to the substrate. Since N2 is a heavier molecule compared to H2, the energy
transfer resulting from the bombardment of the molecule with the substrate is more eﬃcient leading
to higher substrate temperature.

22

Sample

SA1
SA2
SA3
SA4
SA5
SA6

tweight
(nm)
463
659
657
532
525
687

tUV-vis
(nm)
499.6
550.4
NA
NA
NA
NA

Sheet
Resistance
Rs (kΩ/(cid:3))
Too high
3.5x103
30.8
6.4
0.297
0.128

Resistivity
(Ωcm)

Avg.
Roughness
Ra (nm)
1.44
2.82
1.66
1.5
2.11
3.19

Peak to peak
Roughness
Rz (nm)
8.85
20.85
9.22
8.14
10.61
18.02

Deposition
Temperature
(Kelvin)
1048
1098
1126
1156
1213
1248

192.64
2.025
0.341
0.015
0.009
Table 4.1: Series SA ﬁlms grown with 0% N2 concentration

Deposition
Time (hrs)
1
1
1
1
1
1

Total
Pressure
(sccm)
40
45
50
55
57
60

CH4 N2

Forward
Power
(W)
5% 0% 2500
5% 0% 3000
5% 0% 3000
5% 0% 3000
5% 0% 3000
5% 0% 3000

Sample

SB1
SB2
SB3
SB4
SB5
SB6
SB7
SB8
SB9

tweight (nm)
456
541
498
636
595
479
481
453
532

tUV-vis (nm)
393.52
528.9
518.4
584.37
539.5
NA
NA
NA
NA

Sheet
Resistance
Rs (kΩ/(cid:3))
15x103
5x103
12.56
20.63
20.5
3.76
5.29
2.74
0.17

Resistivity
(Ωcm)
590.2
264.4
0.65
1.2055
1.106
0.1803
0.2544
0.124
0.009

Avg.
Roughness
Ra (nm)
4.2
4.47
4.83
1.88
1.7
1.83
4.51
2.09
2.76

Peak to peak
Roughness
Rz (nm)
19.34
24.61
18.78
13.55
12.02
14.9
21.6
9.93
14.79

Deposition
Temperature
(Kelvin)
1043
1085
1103
1118
1125
1148
1151
1163
1229

Deposition
Time (hrs)
1
1
1
1
1
1
1
1
1

Total
Pressure
(sccm)
35
40
40
45
45
43
45
50
60

CH4 N2

Forward
Power
(W)
5% 5% 2500
5% 5% 2500
5% 5% 2700
5% 5% 2500
5% 5% 2700
5% 5% 3000
5% 5% 3000
5% 5% 3000
5% 5% 3000

Table 4.2: Series SB ﬁlms grown with 5% N2 concentration

23

Sample

SC1
SC2
SC3
SC4
SC5

tweight (nm)
391
500
627
497
517

tUV-vis (nm)
344.69
484.13
595.27
513.58
NA

Sheet
Resistance
Rs (kΩ/(cid:3))
4x103
42.9
50.38
13.5
0.66

Resistivity
(Ωcm)
137.87
2.077
2.999
0.69
0.034

Avg.
Roughness
Ra (nm)
2.13
2.68
2.27
1.88
2.39

Peak to peak
Roughness
Rz (nm)
19.02
20.51
22.7
16.72
12.17

Deposition
Temperature
(Kelvin)
1068
1101
1145
1163
1223

Deposition
Time (hrs)
1
1
1
1
1

Total
Pressure
(sccm)
35
40
45
45
55

CH4 N2

Forward
Power
(W)
5% 10% 2500
5% 10% 2500
5% 10% 2700
5% 10% 3000
5% 10% 3000

Table 4.3: Series SC ﬁlms grown with 10% N2 concentration

Sample

SD1
SD2
SD3
SD4
SD5
SD6
SD7

tweight (nm)
194
279
371
367
552
479
365

tUV-vis (nm)
223.16
239.66
325.31
398.96
NA
NA
NA

Sheet
Resistance
Rs (kΩ/(cid:3))
30x103
12x103
2x103
19.7
8.36
4.7
1.45

Resistivity
(Ωcm)
669.48
287.6
65.06
0.786
0.46
0.225
0.053

Avg.
Roughness
Ra (nm)
2.44
2.13
2.81
3.53
3.15
2.79
2.56

Peak to peak
Roughness
Rz (nm)
21.51
18.61
21.01
27.86
18.5
18.9
14.73

Deposition
Temperature
(Kelvin)
1115
1138
1195
1220
1259
1263
1295

Deposition
Time (hrs)
1
1
1
1
1
1
1

Total
Pressure
(sccm)
35
40
45
50
55
60
65

CH4 N2

Forward
Power
(W)
5% 20% 2500
5% 20% 2500
5% 20% 3000
5% 20% 3000
5% 20% 3000
5% 20% 3000
5% 20% 3000

Table 4.4: Series SD ﬁlms grown with 20% N2 concentration

24

4.1 Resistivity

The room temperature resistivity of the ﬁlms as a function of deposition temperature is shown in
Fig. 4.1. Deposition temperature (Td) here refers to the substrate temperature measured with an
optical pyrometer. A variation of about +10K (e1) was observed based on diﬀerent viewing angles
along with a temperature variation of about +5K (e2) during the one hour growth of each ﬁlm. A
2)1/2 for the x-axis is thus calculated taking both these
resultant error bar of +11.18K (e1
variations into account.

2 + e2

Figure 4.1: Resistivity of all the ﬁlms of series SA, SB, SC and SD as a function of deposition
temperature

Fig. 4.1 presents a direct comparison of resistance tunability for undoped and nitrogen incor-
porated UNCD ﬁlms resulting from a range of Td. The resistivity decreases exponentially by more
than four orders of magnitude from 102 to 10-2
Ωcm as Td increases over a range of 250 K for
various concentrations of nitrogen in the plasma, including the undoped samples. The exponential
-1cm-1
decay and the lowest resistivity measured corresponding to a conductivity value of 100 Ω

25

is in accordance with previously reported values for N(UNCD) where the trend was seen indepen-
dently for two diﬀerent deposition conﬁgurations- 1) for increasing nitrogen concentration (0 to
20%) in the plasma with a constant deposition temperature [32, 38] and, 2) for increasing deposition
temperature with constant nitrogen concentration in the plasma [41, 40].

While the trend and resistivity values recorded here are in accordance with the increasing Td
conﬁguration for a constant amount of nitrogen in the plasma, it is not the same for the increasing
nitrogen conﬁguration. It can be seen from the graph in Fig. 4.1 that for a given Td, the resistivity
value does not change as much with nitrogen as it changes with Td and it in fact increases with
increasing nitrogen content as opposed to previous ﬁndings [32, 38].

To understand the factors responsible for the achieved resistance tunability, the changes in the
structure of the ﬁlm based on diﬀerent growth parameters are discussed in the following sections
based on the results obtained from Raman spectroscopy and SEM images of the ﬁlm.

4.2 Analysis of Raman Spectra

The Raman spectra collected for all the four series of ﬁlms grown with various nitrogen concen-
trations exhibit some prominent features described as follows- 1) A band with the peak position
ranging between 1333 cm-1 and 1350 cm-1. The 1333 cm-1 peak is the characteristic diamond
peak attributed to sp3 bonded carbon. The 1350 cm-1 feature appears for graphitic materials in the
form of a band induced by the defect or disorder in the sp2 bonded carbon. Polycrystalline graphite
exhibits D band in its spectrum due to the outer rim of graphite crystallites which tends to have
more defects as opposed to the inside of the graphite crystal for which the band is not present [47].
In this discussion, this band will be referred as the D band as has been used traditionally for

systems containing mixed phases of sp2 and sp3 bonded carbon.
2) A band with the peak position ranging between 1545 cm-1 and 1590 cm-1 which can be recognized
as the G band. In addition to the D band, graphitic materials also exhibit a band at 1588 cm-1 called
the G band attributed to the crystalline graphitic structure with sp2 bonded carbon. The integrated
intensity ratio of the D and G bands, I(D)/I(G) is directly proportional to the defect quantity in

26

Figure 4.2: Raman spectra for 5% N2 samples as a function of Td

graphitic materials and inversely proportional to the in-plane crystallite size.
3) A shoulder at 1140 cm-1 and a weak peak at 1480 cm-1 representing ω1 and ω3 modes of
trans-polyacetylene (t-PA) chains respectively.

The Raman spectra studied both as a function of Td and N2 concentration in the feed gas mixture

show various trends in the position, intensity and shape of the peaks discussed as follows.

4.2.1 Raman spectra as a function of Td

Fig. 4.2 shows the Raman spectra of ﬁlms from SB series, grown with 5% N2 in the feed gas
mixture for various Td. As the temperature increases, the G band peak shifts from 1550 cm-1 to
1590 cm-1 with decreasing intensity which shows a reduction of the amorphous carbon content and

27

a conversion from short-range to long range order of sp2 bonded carbon [48]. The peak shift is also
accompanied by its narrowing indicating an increase in the crystallinity of the material. Thus, with
the increase in Td, the sp2 bonded carbon changes from amorphous to polycrystalline graphite.
The high intensity of defect induced D band with respect to G band can be attributed due to a large
volume of crystallite edges resulting from small crystallite size. The size of these crystallites is in
nanometer scale as conﬁrmed by SEM images in the next section.

For lower Td, the D band is quite broad with its peak at 1333 cm-1. As Td increases, the band
gets narrower and the peak position stays constant up to 1148 K. The ﬁlm grown has a mixture of
sp2 and sp3 bonded carbon forming a matrix of diamond grains surrounded by grain boundaries,
identiﬁed as ultra-nano crystalline diamond (UNCD) based on the SEM images (shown in the next
section). As Td is increased beyond 1148 K, the D band continues to get narrower and the peak at
1333 cm-1 jumps to 1350 cm-1, indicating that there is very less or no amount of sp3 bonded carbon
to be detected by Raman spectroscopy which is more sensitive towards the detection of sp2 bonded
carbon. A complete shift of the 1333 cm-1 peak to 1350 cm-1 peak at higher temperatures showing a
material change of diamond to graphite is normal expect due to the existence of fundamental phase
transition at 1500 K from metastable diamond structure to its fundamental ground state which is
graphite. The phenomenon of diamond being damaged or changed into other carbon structures
when exposed to high temperatures, is also in agreement with previous results [41, 49].

The features at 1140 cm-1 and 1480 cm-1 de-intensify as Td increases and completely disappear
after 1103 K due to the release of hydrogen at increased temperatures. The two peaks were
recognized as characteristic signatures of nanocrystalline diamond until it was realized that they
resemble the behavior of the ω1 (in-plane C-H bending combined with C-C stretching) and the
ω3 (C=C stretching) vibration modes of trans-polyacetylene (t-PA). Accordingly, these features
were associated with fragments of t-PA-like species also present in the grain boundaries. The
disappearance of these features due to annealing also conﬁrms that these modes of t-PA are not
related to diamond grains [50].

Fig.4.3, 4.4 and 4.5 show the Raman spectra as a function of Td for SA, SC and SD series of

28

Figure 4.3: Raman spectra for 0% N2 samples as a function of Td

ﬁlms grown with 0, 10, 20% feed gas N2 respectively. The trends in the features of Raman spectra
are very consistent with increasing Td irrespective of the concentrations of nitrogen in the plasma.
The trends common to all the four series of ﬁlms can be summarized as follows-
1) The shift in the D band peak from 1333 cm-1 to 1350 cm-1 along with the narrowing of the
band with increasing Td.
2) The decrease in the intensity and the shift in the G band peak from 1550 cm-1 to 1590 cm-1
along with the narrowing of the band with increasing Td.
3) The decrease in the intensity of the t-PA features at 1140 cm-1 and 1480 cm-1 till they completely
disappear as Td increases. These features disappear after about 1100 K for all the four series of
ﬁlms which again conﬁrms that these are not related to diamond. It also conﬁrms the consistency
of the pyrometer temperature measurements. The removal of t-PA chains present in the grain
boundaries also enables the formation of better sp2 bonded carbon phase in the grain boundaries
in all four series of ﬁlms.

29

Figure 4.4: Raman spectra for 10% N2 samples as a function of Td

Figure 4.5: Raman spectra for 20% N2 samples as a function of Td

30

Thus, in all 4 series of experiments, there was a consistent and reproducible material transfor-

mation from UNCD to nanocrystalline graphite.

The ﬁlms of series SD grown with 20% feed gas N2 show an additional feature in the D band as
can be seen in Fig. (20% N2). The D band has two peaks at 1333 cm-1 due to sp3 bonded carbon
and at 1364 cm-1 due to disordered sp2 bonded carbon. The 1333 cm-1 peak keeps intensifying
with respect to the 1364 cm-1 peak as Td increases upto 1259 K after which the peaks merge
together and shift completely to 1350 cm-1 at 1295 K. The appearance of the 1333 cm-1 peak
separately conﬁrms that the ﬁlm consists of a mixture of sp3 bonded crystalline carbon and sp2
bonded disordered carbon.

4.2.1.1 Graphitization

The change in the peak positions and width of the D and G bands has been noted for all the four
series of ﬁlms which indicate a material transition from diamond grains embedded in amorphous
carbon matrix to nanocrystalline graphite. To conﬁrm the process of graphitization in the ﬁlms,
Td was increased further for the 0% N2 ﬁlms. Fig. 4.6 shows the comparison between the Raman
spectra of 0% N2 ﬁlms grown at high Td, single crystal diamond (SCD) sample with peak at 1333
cm-1 and highly ordered pyrolytic graphite (HOPG) sample with peak at 1582 cm-1. The intensity
of G band peak at 1550 cm-1 of the 0% N2 ﬁlms keeps decreasing owing to the decrease in the
amorphous content in the ﬁlm along with the shift of the peak position from 1550 cm-1 to 1588
cm-1 as evident from Fig. 4.3.

When the peak has completely shifted to 1588 cm-1, it doesn’t change its position any further
with an increase in Td beyond 1156 K and starts intensifying as shown in Fig. 4.6. The increase
in the 1588 cm-1 peak shows an increase in the amount of crystalline form of sp2 bonded carbon.
The absence of the diamond peak and the presence of both D and G bands in the spectra of 0%
N2 ﬁlms as opposed to just a G band from HOPG sample conﬁrms that now the material grows
completely as nanocrystalline graphite. The increase in the crystalline form of sp2 bonded carbon
is also evident from Fig. 4.2 for the 5% N2 ﬁlms where G band peak completely shifts to 1590 cm-1

31

Figure 4.6: Comparison of Raman spectra of highly graphitized ﬁlms with single crytsal diamond
and highly oriented pyrolytic graphite

at 1163 K and starts intensifying for higher Td.

A similar material transition was also observed in our study of glow plasma discharge during
ﬁeld emission from UNCD cathodes used as emitters [51]. Data from SEM and Raman show the
conversion of nano diamond to nano graphite caused by the electron emission self-induced heating
of the emitter to 3000-4000 K. Fig. 4.7 compares the Raman spectra of the cathode before and after
the damage due to heating. It shows that the damaged areas, denoted by the dark spots in the inset
image, exhibit the Raman characteristics of nano graphite.

32

Figure 4.7: Raman spectra for graphitic and UNCD sites on the damaged (a) and pre-damaged (b)
cathodes due to electron ﬁeld emission.

4.2.2 Raman spectra as a function of N2 concentration in the feed gas

Fig. 4.8 and 4.9 show a comparison of Raman spectra of ﬁlms with varying nitrogen concentration
deposited at 1098 K and 1126 K respectively. A comparison of the peak position and intensity
of the G band shows that the amount of amorphous sp2 bonded carbon increases with increasing
feed gas N2 concentration for ﬁlms deposited at a constant Td. This can be explained by the
amount of energy required in the formation of increasing amount of CN bonds in the ﬁlms with
increasing nitrogen content in the plasma. The positive enthalpy of formation of CN bonds makes
the incorporation of nitrogen in the ﬁlms endothermic. Thus, the available thermal energy gets
distributed between the incorporation of nitrogen through the formation of CN bonds in the grain
boundary of the ﬁlms and the conversion of amorphous sp2 carbon to crystalline form. This accounts
for the increased amount of amorphous sp2 carbon detected in high nitrogen concentration ﬁlms.

33

Figure 4.8: Raman spectra for samples deposited at 1098 K with 0, 5 and 10% N2 concentration

Figure 4.9: Raman spectra for samples deposited at 1126 K with 0, 5 and 20% N2 concentration

It also explains the increase in resistivity of the ﬁlms with increasing feed gas nitrogen concen-

34

tration for a given Td, as shown in Fig. 4.1. An upward shift in the required deposition temperature
is observed to get UNCD ﬁlms with similar resistivities for increasing nitrogen content in the
plasma. For instance, the resistivity value of 0.6 Ωcm was achieved at 1103 K with 5% N2 as
opposed to 1163 K for 10% N2 and 1221 K for 20% N2. Increasing CN bond formation for the
incorporation of nitrogen leads to an overall increase in the required thermal energy to achieve the
same level of crystallinity of sp2 bonded carbon which is the major contributing factor towards the
conductivity of the ﬁlms.

The incorporation of nitrogen in the ﬁlms with increasing feed gas nitrogen concentration has
been reported to saturate at 2x1020 cm-3 but the value of feed gas nitrogen concentration at which
saturation occurs is not consistent in literature [32, 39, 40]. The values were calculated from
SIMS measurements which does not accurately resolve changes in value smaller than an order of
magnitude. XPS spectra for N-UNCD ﬁlms shows a slight increase in the N 1s peak but due to low
signal to noise ratio, this change is not properly resolved [40]. These reports make the argument
of saturation of incorporated nitrogen debatable. Here, we assume that the nitrogen incorporation
does not saturate and the increase is too small to be detected by traditional methods.

4.2.3 Resistivity as a function of G band peak position

The mechanism of resistance tunability with varying nitrogen concentrations can be explained
using Fig. 4.10 which shows resistivity as a function of G peak position between 1550 cm-1 to
1590 cm-1, for various nitrogen concentrations in the feed gas. As the G peak position at 1550
cm-1 corresponds to amorphous phase of sp2 bonded carbon, a lower value of G peak position
corresponds to a lower amount of crystalline sp2 carbon phase which is responsible for conduction.
Films corresponding to higher feed gas N2 content can obtain a particular value of resistivity with
a lower G peak position, i.e., with lower crystallinity of the sp2 bonded carbon, as compared to
the ﬁlms corresponding to lower feed gas N2 content with higher crystallinity of the sp2 bonded
carbon. Thus the amount of feed gas N2 is another factor that aﬀects conductivity apart from the
extent of crystallinity of the sp2 bonded carbon.

35

Figure 4.10: Resistivity as a function of G peak position

Addition of nitrogen in the feed gas leads to the incorporation of nitrogen in the grain boundaries
which increases the amount of sp2 bonded carbon due to the increase in the volume of grain
boundaries. This results to better spatial connectivity of sp2 bonded carbon, thus, enhancing the
conductivity of the ﬁlm. This leads to increased conductance of the higher nitrogen concentration
ﬁlms grown at low Td with lower crystallinity of the sp2 bonded carbon. This provides us with the
ability to tune the resistivity of the UNCD ﬁlms without transforming it a more graphitic material.
As Td increases, the crystallinity of sp2 carbon increases along with increased graphitization of
the ﬁlm, which leads all the four series of ﬁlms to move towards lower resistivity value and converge
at the same G peak position at 1590 cm-1, irrespective of the feed gas nitrogen concentration.

36

4.3 Analysis of SEM images

Electron imaging of the surface of the ﬁlms shows the growth of ultra-nanocrystalline diamond
with strong morphology changes for varying deposition temperatures and nitrogen concentrations.

4.3.1 Surface Morphology

Fig. 4.11, 4.12, 4.13 and 4.14 show the surface morphology of ﬁlms with increasing Td for 0, 5, 10
and 20% N2 respectively.

4.3.1.1 Changes in surface morphology with varying Td

Comparison of the morphology of ﬁlms grown at a particular nitrogen concentration shows a
transition from UNCD to nanocrystalline graphite with increasing Td. Films grown at lower Td as
shown in Fig. 4.11(a), 4.12(a), 4.13(a) and 4.14(a) exhibit densely packed nanometer sized grains
forming a continuous ﬁlm and thus, conﬁrming the growth of UNCD ﬁlms. At higher deposition
temperatures, it is seen in the SEM images, and as already indicated by the Raman spectroscopy,
that the amount of deposited graphite increases. Fig. 4.11(b), 4.12(b), 4.13(b) and 4.14(b) show
the appearance of small voids as Td increases. These voids grow wider with further increase in
Td and the morphology now looks like a layer of graphite ﬂakes as seen in Fig. 4.11(c), 4.12(c),
4.13(c) and 4.14(c). This is in accordance with the transition in morphology from densely packed
nano-diamond ﬁlm to ﬂakes of graphite with increasing Td reported earlier [49].

4.3.1.2 Changes in surface morphology with increasing N2 concentration

Films grown at lower Td show a change in the grain size as nitrogen concentration is varied. For
low N2 concentration (Fig. 4.11(a), 4.12(a) for 0%, 5% nitrogen respectively), the morphology of
the ﬁlms remains largely unchanged. Fig. 4.13(a) shows a slight increase in the grain size for 10%
N2 ﬁlm. However, grain size increases signiﬁcantly and form clusters for the 20% N2 ﬁlm as seen
in Fig. 4.14(a) while still lying in the ultra-nano scale range.

37

(a) SA1: Td=1048 K

(b) SA3: Td=1126 K

(c) SA4: Td=1156 K

Figure 4.11: SEM images of samples grown with 0% N2 concentration

(a) SB1: Td=1043 K

(b) SB7: Td=1080 K

(c) SB9: Td=1229 K

Figure 4.12: SEM images of samples grown with 5% N2 concentration

(a) SC2: Td=1101 K

(b) SC4: Td=1163 K

(c) SC5: Td=1223 K

Figure 4.13: SEM images of samples grown with 10% N2 concentration

(a) SD1: Td=1115 K

(b) SD4: Td=1120 k

(c) SD7: Td=1295 K

Figure 4.14: SEM images of samples grown with 20% N2 concentration

38

These results are in accordance with previously reported changes in grain size with varying nitrogen
concentration [32, 31].

4.3.2 Bulk Morphology

(a) Top view
Figure 4.15: SEM and Raman spectrum of sample: SB4, grown with 5% N2

(b) Cross-section

(c) Raman spectrum

(a) Top view
Figure 4.16: SEM and Raman spectrum of sample: SD1, grown with 20% N2

(b) Cross-section

(c) Raman spectrum

(a) Top view
Figure 4.17: SEM and Raman spectrum of sample: SB9, grown at high Td

(b) Cross-section

(c) Raman Spectrum

39

The cross-sectional view of the ﬁlms grown at lower Td shows an absence in an identiﬁable
columnar structure for both low and high nitrogen concentration samples, however, there is signif-
icant diﬀerence between the low and high nitrogen concentration samples owing to varying grain
sizes.

The cross-section of 5% N2 ﬁlms in Fig. 4.15(b) shows smooth fracture surfaces which suggest
that the nanocrystalline diamond does not grow from the initial nulclei at the substrate-ﬁlm interface
but is the result of very high renucleation rates [27]. This results in a highly dense continuous ﬁlm
with no discernable growth morphology as the crystallite sizes are too small as seen in Fig. 4.15(a)
to be resolved in the cross-sectional view. The cross-section of 20% N2 ﬁlms in Fig. 4.16(b) also
shows a highly dense ﬁlm resulting from high renuclation rate along with discernable grains due to
larger crystallite size as seen in Fig. 4.16(a). Fig. 4.17(b) shows the cross-sectional view of highly
graphitized ﬁlm grown at high Td in which the crystallites of graphite are loosely packed in the
form of ﬂakes.

4.4 Reﬂectance Spectrum

A comparison of the reﬂectance spectrum obtained from UV-vis spectroscopy of samples grown at
low and high Td is shown in Fig. 4.18. High reﬂectance and strong interference pattern is exhibited
by sample SB1 deposited at 1043 K which shows the highly reﬂective characteristic of diamond
(as can be seen from the bright pink color of the ﬁlm) retained in the ﬁlm. As Td increases, the
reﬂectance from the ﬁlm decreases and the interference pattern grows weaker to a point where it
is almost negligible due to high optical absorption of the graphitized samples. This can be seen
for the highly graphitized sample SB9 grown at 1229 K with the inset image showing the highly
absorptive black color of the ﬁlm..

40

Figure 4.18: Comparison of reﬂectance spectra of UNCD and highly graphitized ﬁlm

4.5 Thickness

The thickness of the ﬁlms was measured using two diﬀerent methods independently, to cross-verify
the calculated values from each method. Thickness calculation from the weight measurement
method was done using the density of pure diamond and that from UV-vis spectroscopy using the
wavelength dependent refractive index of pure diamond. Diﬀerence of up to ∼50 nm was found
between the calculated values from the two methods. Thickness could not be calculated using
UV-vis spectroscopy for the highly graphitized samples due to very weak to negligible interference
patterns obtained. These values of thickness were also compared with the thickness observed
from the SEM images of the cross-section of the ﬁlms (comparison shown in Fig. 4.19). The
calculated values were close to the observed values from SEM images for the low temperature
samples (Samples 1 and 2), thus conﬁrming they have properties close to pure diamond. For
the higher temperature samples, the diﬀerence between the SEM observed and calculated values
increases. For the highest temperature sample, this diﬀerence is maximum (Samples 3 and 4).

41

The SEM observed value of this sample rather matches with the value calculated from the weight
diﬀerence method by substituting the density of diamond (3.51 g/cm3) with the density of graphite
(2.27 g/cm3). This gives a fundamental proof of material transition from UNCD to nanocrystalline
graphite as interpreted by the Raman spectra (shown alongside and discussed before) and SEM
images of the top-view of the ﬁlms.

Figure 4.19: Conﬁrmation of graphitization through thickness measurement

4.6 Deﬁning UNCD based on Raman spectra

Raman spectroscopy reveals and traces the bonding structure of the ﬁlms under study which
undergoes striking changes as Td increases. Namely, the change in the peak positions and widths
of D and G bands indicate a clear material phase transition from diamond grains lined with initially
amorphous carbon GB to nanocrystalline graphite. The ﬁlm remains to be UNCD as long as it has
diamond grains indicated by the presence of 1333 cm-1 peak and sp2 grain boundaries in mixed
amorphous/crystalline state with G band peak located somewhere between 1550 cm-1 and 1590
cm-1. The material transforms completely to nanocrystalline graphite when the D band completely
shifts to nearby wavenumber 1350 cm-1 and therefore no longer exhibits diamond content in the
ﬁlm but rather represents nanographite defect.

42

CONCLUSION

Undoped and nitrogen incorporated ultra-nanocrystalline diamond ﬁlms were successfully grown
in a systematical manner using microwave plasma assisted chemical vapor deposition at various
deposition temperatures for 0, 5, 10 and 20% N2 concentrations in the feed gas of 1% CH4/H2
mixture. Resistivity of the ﬁlms was found to decrease exponentially by over four orders of
magnitude with increasing deposition temperature for all the four nitrogen concentration ﬁlms,
including the nitrogen-free ones. For all the ﬁlms, Raman and SEM analysis revealed a consistent
and reproducible material transformation of the material from UNCD to nanocrystalline graphite
with increasing deposition temperature. We found that
(i) the GB size eﬀects leading to their enhanced connectivity (enhanced number of contacts) within
the GB network promoted by addition of nitrogen; and
(ii) improved crystallinity of individual GBs promoted by the elevated deposition temperature;
are equally important in fundamental understanding of conductivity mechanism in UNCD and have
to be taken into account along with energy space (density of states band diagram) considerations.
Practically speaking, our results outline pathways for engineering resistivity in UNCD ﬁlms toward
targeted electronics applications.

Beyond that, we revisited the basic deﬁnition of UNCD material as it was not fully addressed

in the previous experimental literature.

Ultimately, we ﬁnd that sp2 phase plays primary role in setting UNCD conductivity. This
is evidenced by the fact that, unlike in Bhattacharya et al., even undoped samples had high
conductivities and ultimately reached the champion metrics of ∼100 S/cm when fully graphitized.
The most recent results [43, 41] are consistent with our conclusion. In the context of redeﬁned term
UNCD and actual mass density evaluation, the most conductive ﬁlms for all N2 concentrations
studied are no longer UNCD. There is one important diﬀerence between 0% and 20 % series,
though. Thanks to initially enhanced connectivity of GBs, 20 % series retain UNCD bonding
toward the highest deposition temperature: roughly speaking, in the conductivity range 1-10 S/cm,

43

samples 0 and 5 % achieved at Td ≈1150K, are converted to nano-graphite while 20 % series in
the same conductivity range retains UNCD structure. This is important to consider as conductive
UNCD are preferred over nano-graphite due to superior mechanical properties, e.g., for MEMS and
NEMS systems; another example include electron emitters when only diamond grains/graphitic
GB provide for work function engineering through hydrogen surface termination and therefore for
improved electron emissivity.

44

APPENDIX

45

LIST OF TALKS AND PAPERS

(i) "Nanodiamond to Nanographite Conversion During Field Emission", Poster Presentation, An-
nual Graduate Symposium, Michigan Institute for Plasma Science and Engineering (MIPSE),
2018.

(ii) "Nitrogen Doping and Graphitization Eﬀects on Conductivity of Ultra-nanocrystalline Di-
amond", Poster Presentation, Engineering Graduate Research Symposium, Michigan State
University, 2019.

(iii) "Nanodiamond to Nanographite Conversion During Field Emission", Talk, International

Vacuum Nanoelectronics Conference (IVNC), Cincinnati, 2019.

(iv) "Field Electron Emission Induced Glow Discharge in Nanodiamond Vacuum Diode", Publi-

cation, Journal of Physics D: Applied Physics, 2019.

46

BIBLIOGRAPHY

47

BIBLIOGRAPHY

[1]

Jes Asmussen and D K Reinhard. Diamond Films Handbook. 2010. ISBN 0824795776. doi:
10.1201/9780203910603.

[2] Y. Tzeng, M. Yoshikawa, M. Murakawa, and A. Feldman, editors. Applications of diamond
ﬁlms and related materials; Proceedings of the 1st International Conference, Auburn, AL,
Aug. 17-22, 1991, volume 73, 1991.

[3] Orlando Auciello and Anirudha V. Sumant.

science
and application
and technology of ultrananocrystalline diamond (uncd™) ﬁlms
Diamond and Related Materials, 19(7):699 – 718,
to multifunctional devices.
https://doi.org/10.1016/j.diamond.2010.03.015. URL
2010.
http://www.sciencedirect.com/science/article/pii/S092596351000141X. Pro-
ceedings of Diamond 2009, The 20th European Conference on Diamond, Diamond-Like
Materials, Carbon Nanotubes and Nitrides, Part 2.

ISSN 0925-9635.

review of

Status

doi:

the

[4] X. Xiao, J. Birrell, J. E. Gerbi, O. Auciello, and J. A. Carlisle. Low temperature growth
of ultrananocrystalline diamond. Journal of Applied Physics, 96(4):2232–2239, 2004. doi:
10.1063/1.1769609. URL https://doi.org/10.1063/1.1769609.

[5] A.R. Krauss, O. Auciello, D.M. Gruen, A. Jayatissa, A. Sumant, J. Tucek, D.C.
Mancini, N. Moldovan, A. Erdemir, D. Ersoy, M.N. Gardos, H.G. Busmann, E.M.
Meyer, and M.Q. Ding. Ultrananocrystalline diamond thin ﬁlms for mems and mov-
ing mechanical assembly devices. Diamond and Related Materials, 10(11):1952 – 1961,
2001.
doi: https://doi.org/10.1016/S0925-9635(01)00385-5. URL
http://www.sciencedirect.com/science/article/pii/S0925963501003855.

ISSN 0925-9635.

[6] Orlando Auciello, James Birrell, John A Carlisle, Jennifer E Gerbi, Xingcheng Xiao, Bei
Peng, and Horacio D Espinosa. Materials science and fabrication processes for a new MEMS
technology based on ultrananocrystalline diamond thin ﬁlms. Journal of Physics: Con-
densed Matter, 16(16):R539–R552, apr 2004. doi: 10.1088/0953-8984/16/16/r02. URL
https://doi.org/10.1088%2F0953-8984%2F16%2F16%2Fr02.
Diamond power devices.

[7] A. Denisenko and E. Kohn.

Diamond

and Related Materials,
doi:

its.
0925-9635.
http://www.sciencedirect.com/science/article/pii/S0925963505000087.
Proceedings of Diamond 2004, the 15th European Conference on Diamond, Diamond-Like
Materials, Carbon Nanotubes, Nitrides and Silicon Carbide.

https://doi.org/10.1016/j.diamond.2004.12.043.

14(3):491

498,

–

concepts
2005.

and lim-
ISSN
URL

[8] M. Kasu, K. Ueda, Y. Yamauchi, A. Tallaire, and T. Makimoto. Diamond-based rf power
transistors: Fundamentals and applications. Diamond and Related Materials, 16(4):1010 –
1015, 2007. ISSN 0925-9635. doi: https://doi.org/10.1016/j.diamond.2006.12.046. URL

48

http://www.sciencedirect.com/science/article/pii/S0925963506005309. Pro-
ceedings of Diamond 2006, the 17th European Conference on Diamond, Diamond-Like
Materials, Carbon Nanotubes, Nitrides and Silicon Carbide.

[9] Toshiharu Makino, Norio Tokuda, Hiromitsu Kato, Masahiko Ogura, Hideyuki Watanabe,
Sung-Gi Ri, Satoshi Yamasaki, and Hideyo Okushi. High-eﬃciency excitonic emission
with deep-ultraviolet light from (001)-oriented diamondp-i-nJunction. Japanese Journal of
Applied Physics, 45(No. 39):L1042–L1044, sep 2006. doi: 10.1143/jjap.45.l1042. URL
https://doi.org/10.1143%2Fjjap.45.l1042.

[10] A. Hokazono, T.

Ishikura, K. Nakamura, S. Yamashita,

of
and Related Materials,

Enhancement/depletion mesfets
amond
9635.
http://www.sciencedirect.com/science/article/pii/S0925963596007261.

https://doi.org/10.1016/S0925-9635(96)00726-1.

diamond
6(2):339

logic
1997.

and
–

their

343,

doi:

circuits.

and H. Kawarada.
Di-
ISSN 0925-
URL

surface

performance

[11] A. Aleksov, A. Denisenko, U. Spitzberg, T. Jenkins, W. Ebert, and E. Kohn.
gate
ISSN
URL

Rf
length.
0925-9635.
http://www.sciencedirect.com/science/article/pii/S0925963501006446.
12th European Conference on Diamond, Diamond- Like Materials, Carbon Nanotubes,
Nitrides and Silicon Carbide.

https://doi.org/10.1016/S0925-9635(01)00644-6.

Diamond and Related Materials,

11(3):382 – 386,

sub-micron

fets with

diamond

channel

2002.

of

doi:

[12] Jan Isberg, Gabriel Ferro, and Paul Siﬀert. Diamond electronic devices. AIP Conference

Proceedings, 1292, 11 2010. doi: 10.1063/1.3518277.

[13] M. I. Landstrass and K. V. Ravi. Resistivity of chemical vapor deposited diamond
ﬁlms. Applied Physics Letters, 55(10):975–977, 1989. doi: 10.1063/1.101694. URL
https://doi.org/10.1063/1.101694.

[14] Johan F. Prins.

n-type semiconducting diamond by means of oxygen-ion implanta-
tion. Phys. Rev. B, 61:7191–7194, Mar 2000. doi: 10.1103/PhysRevB.61.7191. URL
https://link.aps.org/doi/10.1103/PhysRevB.61.7191.

[15] Isao Sakaguchi, Mikka N.-Gamo, Yuko Kikuchi, Eiji Yasu, Hajime Haneda, Toshimitsu
Suzuki, and Toshihiro Ando. Sulfur: A donor dopant for n-type diamond semiconduc-
tors. Phys. Rev. B, 60:R2139–R2141, Jul 1999. doi: 10.1103/PhysRevB.60.R2139. URL
https://link.aps.org/doi/10.1103/PhysRevB.60.R2139.

[16] S. A. Grot, C. W. Hatﬁeld, G. Sh. Gildenblat, A. R. Badzian, and T. Badzian. Electrical
properties of selectively grown homoepitaxial diamond ﬁlms. Applied Physics Letters, 58(14):
1542–1544, 1991. doi: 10.1063/1.105172. URL https://doi.org/10.1063/1.105172.
[17] Oliver A. Williams, Michael D. Whitﬁeld, Richard B. Jackman, John S. Foord, James E. Butler,
and Christoph E. Nebel. Formation of shallow acceptor states in the surface region of thin
ﬁlm diamond. Applied Physics Letters, 78(22):3460–3462, 2001. doi: 10.1063/1.1345806.
URL https://doi.org/10.1063/1.1345806.

49

[18] Satoshi Koizumi, Tokuyuki Teraji, and Hisao Kanda.

vapor deposition of diamond.
2000.
http://www.sciencedirect.com/science/article/pii/S092596350000217X.

Phosphorus-doped chemical
Diamond and Related Materials, 9(3):935 – 940,
doi: https://doi.org/10.1016/S0925-9635(00)00217-X. URL

ISSN 0925-9635.

[19] Hiromitsu Kato, Satoshi Yamasaki, and Hideyo Okushi. n-type doping of (001)-oriented
single-crystalline diamond by phosphorus. Applied Physics Letters, 86(22):222111, 2005.
doi: 10.1063/1.1944228. URL https://doi.org/10.1063/1.1944228.
high

[20] Shinichi
tronics.
0925-9635.
http://www.sciencedirect.com/science/article/pii/S0925963516300759.
Special Issue “26th International Conference on Diamond and Carbon Materials – DCM
2015”.

Single
and Related Materials,

https://doi.org/10.1016/j.diamond.2016.03.013.

diamond wafers
65:168

Diamond
doi:

elec-
ISSN
URL

Shikata.

–

175,

2016.

crystal

for

power

[21] J. W. Liu, M. Y. Liao, M. Imura, E. Watanabe, H. Oosato, and Y. Koide. Diamond
logic inverter with enhancement-mode metal-insulator-semiconductor ﬁeld eﬀect transis-
tor. Applied Physics Letters, 105(8):082110, 2014.
doi: 10.1063/1.4894291. URL
https://doi.org/10.1063/1.4894291.

[22] Orlando Auciello, Pablo Gurman, Maria B. Guglielmotti, Daniel G. Olmedo, Alejandro Berra,
and Mario J. Saravia. Biocompatible ultrananocrystalline diamond coatings for implantable
medical devices. MRS Bulletin, 39(7):621–629, 2014. doi: 10.1557/mrs.2014.134.

[23] Orlando Auciello and Anirudha Sumant. Status review of the science and technology of
ultrananocrystalline diamond (uncd (tm)) ﬁlms and application to multifunctional devices.
Diamond and Related Materials, 19:699–718, 07 2010. doi: 10.1016/j.diamond.2010.03.015.
[24] Piyush Bajaj, Demir Akin, Amit Gupta, Debby Sherman, Bing Shi, Orlando Auciello, and
Rashid Bashir. Ultrananocrystalline diamond ﬁlm as an optimal cell interface for biomedical
applications. Biomedical Microdevices, 9(6):787–794, Dec 2007. ISSN 1572-8781. doi:
10.1007/s10544-007-9090-2. URL https://doi.org/10.1007/s10544-007-9090-2.

[25] Xingcheng Mich. Xiao, Jian Wang, Chao Liu, John Arthur Carlisle, Brian Vernon Mech,
Robert G. Greenberg, Dilek Soydam Guven, Ricardo Freda, Mark S. Humayun, James D.
Weiland, and Orlando H. Auciello. In vitro and in vivo evaluation of ultrananocrystalline
diamond for coating of implantable retinal microchips. Journal of biomedical materials
research. Part B, Applied biomaterials, 77 2:273–81, 2006.

[26] Alan R. Krauss, Dieter M. Gruen, Daniel Zhou, Thomas G. Mccauley, Lu Chang Qin,
Timothy Corrigan, Orlando Auciello, and R. P. H. Chang. Morphology and electron emission
properties of nanocrystalline cvd diamond thin ﬁlms. MRS Proceedings, 495:299, 1997. doi:
10.1557/PROC-495-299.

[27] Dieter M. Gruen.

Nanocrystalline diamond ﬁlms.

als Science, 29(1):211–259, 1999.
https://doi.org/10.1146/annurev.matsci.29.1.211.

doi:

10.1146/annurev.matsci.29.1.211.

Annual Review of Materi-
URL

50

[28] P. Keblinski, D. Wolf, F. Cleri, S.R. Phillpot, and H. Gleiter. On the nature of
grain boundaries in nanocrystalline diamond. MRS Bulletin, 23(9):36–41, 1998. doi:
10.1557/S0883769400029353.

[29] F Cleri, P Keblinski, L Colombo, D Wolf, and S. R Phillpot. On the electrical activity ofsp2-
bonded grain boundaries in nanocrystalline diamond. Europhysics Letters (EPL), 46(5):
671–677, jun 1999. doi: 10.1209/epl/i1999-00318-5. URL https://doi.org/10.1209.

[30] Peter Zapol, Michael Sternberg, Larry A. Curtiss, Thomas Frauenheim, and Dieter M. Gruen.
Tight-binding molecular-dynamics simulation of impurities in ultrananocrystalline diamond
grain boundaries. Phys. Rev. B, 65:045403, Dec 2001. doi: 10.1103/PhysRevB.65.045403.
URL https://link.aps.org/doi/10.1103/PhysRevB.65.045403.

[31] James Birrell, J. A. Carlisle, O. Auciello, D. M. Gruen, and J. M. Gibson. Morphology and
electronic structure in nitrogen-doped ultrananocrystalline diamond. Applied Physics Letters,
81(12):2235–2237, sep 2002. ISSN 00036951. doi: 10.1063/1.1503153.

[32] S. Bhattacharyya, O. Auciello, J. Birrell, J. A. Carlisle, L. A. Curtiss, A. N. Goyette, D. M.
Gruen, A. R. Krauss, J. Schlueter, A. Sumant, and P. Zapol. Synthesis and characterization
of highly-conducting nitrogen-doped ultrananocrystalline diamond ﬁlms. Applied Physics
Letters, 79(10):1441–1443, sep 2001. ISSN 00036951. doi: 10.1063/1.1400761.

[33] Dieter M. Gruen, Xianzheng Pan, Alan R. Krauss, Shengzhong Liu, Jianshu Luo, and Christo-
pher M. Foster. Deposition and characterization of nanocrystalline diamond ﬁlms. Journal
of Vacuum Science & Technology A, 12(4):1491–1495, 1994. doi: 10.1116/1.579343. URL
https://doi.org/10.1116/1.579343.

[34] Hak-Joo Lee, Hyeongtag Jeon, and Wook-Seong Lee. Ultrananocrystalline diamond ﬁlm
deposition by direct-current plasma assisted chemical vapor deposition using hydrogen-rich
precursor gas in the absence of the positive column. Journal of Applied Physics, 109(2):
023303, 2011. doi: 10.1063/1.3533764. URL https://doi.org/10.1063/1.3533764.
[35] Sobia Allah Rakha, Guojun Yu, Jianqing Cao, Suixia He, and Xingtai Zhou. Inﬂuence of
ch4 on the morphology of nanocrystalline diamond ﬁlms deposited by ar rich microwave
plasma. Journal of Applied Physics, 107(11):114324, 2010. doi: 10.1063/1.3410804. URL
https://doi.org/10.1063/1.3410804.

[36] Dieter M. Gruen. Ultrananocrystalline diamond in the laboratory and the cosmos. MRS

Bulletin, 26(10):771–776, 2001. doi: 10.1557/mrs2001.204.

[37] P. Achatz, O. A. Williams, P. Bruno, D. M. Gruen, J. A. Garrido, and M. Stutzmann. Eﬀect
of nitrogen on the electronic properties of ultrananocrystalline diamond thin ﬁlms grown on
quartz and diamond substrates. Physical Review B - Condensed Matter and Materials Physics,
74(15), 2006. ISSN 10980121. doi: 10.1103/PhysRevB.74.155429.

[38] Oliver A. Williams, Stephane Curat,

Richard B. Jackman.
plied Physics Letters, 85(10):1680–1682, 2004.
https://doi.org/10.1063/1.1785288.

and
n-type conductivity in ultrananocrystalline diamond ﬁlms. Ap-
URL

Jennifer E. Gerbi, Dieter M. Gruen,

doi:

10.1063/1.1785288.

51

[39] James Birrell, J. E. Gerbi, O. Auciello, J. M. Gibson, D. M. Gruen, and J. A. Carlisle. Bonding
structure in nitrogen doped ultrananocrystalline diamond. Journal of Applied Physics, 93(9):
5606–5612, 2003. ISSN 0021-8979. doi: 10.1063/1.1564880.

[40] Tomohiro Ikeda, Kungen Teii, C. Casiraghi, J. Robertson, and A. C. Ferrari. Eﬀect of the s p2
carbon phase on n -type conduction in nanodiamond ﬁlms. Journal of Applied Physics, 104
(7), 2008. ISSN 00218979. doi: 10.1063/1.2990061.

[41] J.J. Alcantar-Peña, J. Montes, M.J. Arellano-Jimenez, J.E. Ortega Aguilar, D. Berman-
Mendoza, R. García, M.J. Yacaman, and O. Auciello. Low temperature hot ﬁlament
chemical vapor deposition of ultrananocrystalline diamond ﬁlms with tunable sheet re-
sistance for electronic power devices. Diamond and Related Materials, 69:207 – 213,
2016.
https://doi.org/10.1016/j.diamond.2016.09.007. URL
http://www.sciencedirect.com/science/article/pii/S0925963516303284.

ISSN 0925-9635.

doi:

[42] James E. Butler and Anirudha V. Sumant. The cvd of nanodiamond materials. Chem-
doi: 10.1002/cvde.200700037. URL

ical Vapor Deposition, 14(7-8):145–160, 2008.
https://onlinelibrary.wiley.com/doi/abs/10.1002/cvde.200700037.

[43] Neda Wiora, Michael Mertens, Kai Brühne, H.-J Fecht, Ich C. Tran, Trevor Willey, Anthony
van Buuren, Juergen Biener, and Jun-Sik Lee. Grain boundary dominated electrical conduc-
tivity in ultrananocrystalline diamond. Journal of Applied Physics, 122:145102, 10 2017.
doi: 10.1063/1.4993442.

[44] James Birrell,

J.E. Gerbi, O. Auciello,
raman

the

Interpretation

Carlisle.
mond.
0925-9635.
http://www.sciencedirect.com/science/article/pii/S0925963504002596.

https://doi.org/10.1016/j.diamond.2004.07.012.

and Related Materials,

Diamond

doi:

of

J.M. Gibson,
of
spectra
14(1):86

Johnson,

J.
ultrananocrystalline
–

2005.

92,

and J.A.
dia-
ISSN
URL

[45] James E. Butler and Richard L. Woodin. Thin ﬁlm diamond growth mechanisms, pages 15–30.

Springer Netherlands, Dordrecht, 1994. ISBN 978-94-011-0725-9.

[46] H. R. Phillip and E. A. Taft. Kramers-kronig analysis of reﬂectance data for diamond.
doi: 10.1103/PhysRev.136.A1445. URL

Phys. Rev., 136:A1445–A1448, Nov 1964.
https://link.aps.org/doi/10.1103/PhysRev.136.A1445.

[47] M. A. Pimenta, G. Dresselhaus, M. S. Dresselhaus, L. G. Cançado, A. Jorio, and R. Saito.
Studying disorder in graphite-based systems by Raman spectroscopy. Physical Chemistry
Chemical Physics, 9(11):1276–1291, 2007. ISSN 14639076. doi: 10.1039/b613962k.

[48] A. C. Ferrari and J. Robertson. Interpretation of raman spectra of disordered and amorphous
carbon. Phys. Rev. B, 61:14095–14107, May 2000. doi: 10.1103/PhysRevB.61.14095. URL
https://link.aps.org/doi/10.1103/PhysRevB.61.14095.
[49] O. Gröning, O. M. Küttel, P. Gröning, and L. Schlapbach.

ties of nanocrystalline chemically vapor deposited-diamond ﬁlms.

Field emission proper-
Journal of Vacuum

52

Science & Technology B: Microelectronics and Nanometer Structures Processing, Mea-
surement, and Phenomena, 17(5):1970–1986, 1999.
10.1116/1.590857. URL
https://avs.scitation.org/doi/abs/10.1116/1.590857.

doi:

[50] R. Pfeiﬀer, H. Kuzmany, N. Salk, and B. Günther. Evidence for trans-polyacetylene
Ap-
URL

in nanocrystalline diamond ﬁlms from h–d isotropic substitution experiments.
plied Physics Letters, 82(23):4149–4150, 2003.
https://doi.org/10.1063/1.1582352.

doi:

10.1063/1.1582352.

[51] Stanislav Baturin, Tanvi Nikhar, and Sergey Baryshev. Field electron emission induced glow
discharge in nanodiamond vacuum diode. Journal of Physics D: Applied Physics, 2019. URL
http://iopscience.iop.org/10.1088/1361-6463/ab2183.

53