Tribology, the science of
interacting surfaces in relevant motion, encompasses many applications such as
hip implants or ball bearings, where premature failure could end
catastrophically. Further scientific understanding of the behavior of
tribological materials under load is necessary, since lubricants and coatings
are generally applied to these types of material systems to prevent such
unwanted unfortunate events. Tribological coatings typically range from 0.1 –
50 microns thick 1
and tend to have increased resistance to wear and decreased coefficients of
friction compared to the uncoated material 2. Surface properties,
such as friction and anti-wear 2, can be directly
attributed to the hardness of the coating material. Hardness is a material
property that is determined by the structure of the material, which can be
controlled via processing.
Many materials are used as
tribological hard anti-wear coatings, but one of the of the most commonly utilized
materials is diamond. Diamond coatings are often applied to manufactured
cutting tools because of their hardness (Mohs scale 10 out of 10) 3. These coatings
exhibit anti-wear properties and low coefficients of friction (0.05) 1, 4. Cubic boron nitride
(c-BN), a derivative of diamond, is also desired for use as a hard coating
material (Mohs scale 9.9) 4. Many similarities
exist between diamond and c-BN, including structure and bond length, shown in
Figure 1. The close similarity in structure and bond length allows for diamond
and c-BN to have comparable hardness values, enabling them to both be used for
anti-wear coatings. c-BN tends to be more chemically inert than diamond 3, making c-BN a very
attractive material for high temperature ferrous and oxygen-rich environments.
However, the processing of c-BN coatings presents various challenges.
Chemical vapor deposition (CVD) and
physical vapor deposition (PVD) are two common mechanisms used to deposit boron
nitride (BN) coatings. When using CVD and PVD, resulting BN films are generally
contaminated with amorphous/turbostratic or hexagonal boron nitride (h-BN) 5. The differences in
h-BN and c-BN structures results in drastically different material properties. h-BN
is a layered material that possesses similar properties to that of graphite;
the bonds between atoms are short and strong within plane, but are not strong
between multiple planes. The weak van der Waals forces in-between these h-BN
planes cause the structure to shear as a function of wear, making it useful as
a lubricant, but rendering it impractical for use as a hard, anti-wear coating 6. This is due to the
layers flaking off, effectively removing the coating itself until there is
nothing left 6. c-BN possesses three-dimensional bonding, where the
predominant covalent bonding nature of the small atoms creates a very strong,
dense, and close-packed structure. Another roadblock of using CVD and PVD
is that c-BN growth requires ion bombardment from a plasma and/or substrate
bias 7. This processing
step leads to a low-quality c-BN film with a high surface roughness, high
intrinsic stress and contamination which degrade the overall mechanical
Finally, these two techniques have poor conformality and uniformity when
depositing films on non-flat surfaces 8. Therefore, a
different deposition technique is needed.
layer deposition (ALD) is an emerging deposition method that allows for uniform
and conformal coating of non-flat surfaces, as shown in Figure 2 8, 9. This is made
possible because of the gas-phase precursors
that undergo subsequent self-limiting surface reactions as they are introduced
into the system 8, 9. Each precursor is
singularly pulsed into the reaction chamber where it saturates the reactive
sites present on the surface resulting in the formation of a monolayer. The
chemical reaction occurs until all reactive sites are effectively terminated,
allowing the film to conform uniformly to the surface it is bonded to. The
system is then purged with an inert gas to remove all by-products and excess un-reacted
precursor. The thickness of the grown film is therefore solely dependent upon
the number of reaction cycles to which the substrate is subjected 8, 9.
It was suggested by Marlid et al. 10 that temperature and
chemistry both play an important role in the formation of c-BN over h-BN, and
that these parameters need to be investigated via ALD. In response, boron nitride films have been
grown using BCl3/BBr3, and NH3 as precursors 8, 9, 11, 12. Marlid et al. 10 explored the low
pressure, low-temperature growth of boron nitride films on silica substrates using
ALD. In this study, it was found that halogen containing precursors grew
near-stoichiometric smooth films regardless of temperature (400 ºC vs. 750 ºC).
However, films grown at 400 ºC, resulted in amorphous film growth, while those
at 750 ºC resulted in turbostratic film growth 10. Ferguson et al.
explored ALD of boron nitride, also using a halogen containing precursor, to
coat ZrO2 particles 11. The films were
grown at 500K, and transmission electron microscopy (TEM) images suggested that
the boron nitride coatings were amorphous 11. Modeling can offer
insight on how different substrate orientations and precursor
chemistry impacts film growth. c-BN was originally thought not to be the
thermodynamically stable phase of boron nitride at standard conditions, due to
its similarities with carbon and diamond. However, after many calculations,
c-BN was found to the stable phase as shown in Figure 3 3. Despite this, the formation of the commonly
seen metastable h-BN phase during low-pressure experiments is explained by the
Ostwald and Ostwald-Volmer rules 3.
The Ostwald rule states that if energy is removed from a system containing several energy
states, the system will reach the stable ground state only after passing
through all stepwise intermediate metastable states 3.
This is demonstrated in Figure 4. The Ostwald-Volmer rule states that the less
dense phase is formed or nucleated first 3.
Since h-BN is both the metastable and less dense phase of boron nitride,
compared to c-BN, it is the phase that is formed first as demonstrated in previous
CVD and PVD growths 7. To grow c-BN, and not h-BN, the Ostwald and
Ostwald-Volmer rules must be circumvented. This can be achieved by having
adequate surface mobility during deposition, using sp3 hybridized
precursors, and the layer-by-layer growth that is intrinsic to ALD.
A high surface mobility of boron
and nitrogen atoms aids in suppressing h-BN nucleation and growth by allowing
the atoms to re-arrange in an ordered layer. However, mobility increases with
increasing temperature, but as temperature increases the relative thermodynamic
stability of c-BN compared to h-BN decreases 3.
This means that as temperature increases, so does the energy of the system,
thus increasing the likelihood of nucleating h-BN in addition to c-BN. This
makes finding an ideal processing temperature critical in preventing the
nucleation of h-BN while aiding in the mobility of precursor chemistry. In
addition, high mobility comes from small atoms such as hydrogen in comparison
to a halide like those previously mentioned. Hence, hydrogen based precursors, B2H6
and NH3, are chosen in this proposal. These precursors have been
used previously via low-pressure CVD to successfully grow h-BN of controllable
thicknesses 13. Furthermore, these
precursors were chosen due to the sp3 hybridization bonding present
in B2H6, and the formation of sp3 bonds in NH3
that is seen in ALD reactions when growing AlN 14and
Unlike CVD, the separate introduction of precursors that occurs during the ALD process
enables the control of chemical structure. When growing along the (111) face, the
alternating formation of boron and nitrogen planes in addition to sp3
hybridized precursors forces three-dimensional out-of-plane bonding to occur. This
should effectively suppress any h-BN nucleation.
hard, anti-wear coating applied to non-flat surfaces can be achieved by growing
c-BN via ALD. Utilizing the alternating
layer growth method inherent of ALD, c-BN will be synthesized in the (111)
direction using B2H6 and NH3 precursors, which
will favor the formation of sp3 bonds that are present in the cubic
structure but absent in h-BN structure. This will result in a crystalline dense
film with short strong bonds characteristic of other hard anti-wear coating
materials. The c-BN film will be evaluated based on its conformality,
composition, and mechanical properties.
Researchers at Lawrence-Berkeley
national lab created the Materials Project, which utilizes density functional
theory to calculate crystal structures, phase diagrams, chemical reactions, and
material properties 16. The Materials
Project performs high-throughput screening of the structures and properties
such as formation energies, band structures, and surface energies. The
calculation results are stored in an openly accessible database where they can
be used to guide further computation and experimental research. This database
offers insight on how the orientation of the substrate impacts the orientation
of a grown film. For instance, (111) Diamond is among the most favorable of substrates
aiding in the (111) growth of c-BN, having a minimal coincident interface area
(MCIA) criteria of 22.8 18. In addition, the
enthalpy of chemical reactions can be calculated, which will help determine
precursor chemistry aiding in the growth of c-BN. Finally, calculated XRD
patterns can help identify characteristic peaks of structures.
This study will investigate the use of ALD to grow a
conformal hard coating of c-BN through use of B2H6 and NH3
as precursors in addition to understanding the mechanism of increased sp3
bond formation. If these precursors fail, other precursor chemistries
such as N2, BF3, and balancing NH3 with the
addition of H2, can be studied in addition to altering temperatures,
pressures, and dosage times. The chemical reaction of these precursors is shown
in Figure 5. In CVD, boron nitride gas precursors NH3 and BH3
(B2H6 below 600 ºC) are stable 3.
A cold-wall perpendicular flow reactor will be used to deposit the films to
prevent unwanted gas reactions and to ensure uniform
coverage of the substrate. The films will be 100 nm thick because this
thickness should give quantifiable measurements of c-BN while taking a modest
amount of time to grow. The films will be grown on commercially available
diamond (111) 19 substrates where
their surface roughness is known. Diamond is chosen to aid the formation of c-BN
due to a lattice mismatch of 1.4% 3. In addition,
patterned silicon (111) 20 containing trenches will
be used to determine the conformality of the deposited film on non-flat
surfaces and the effects of orientation on the growth of c-BN. The trenches
will be created using standard lithography processes. The patterns will create
trenches of three varying aspect ratios: 1:1, 10:1, 1:10. Argon is chosen as
the purge gas because of its inert nature. Pressure will be held constant at 10
Torr and the flow rates will consist of a few hundred sccm 10.
Knowing that the bond energies for
B-B (310 kJ/mol) and B-N (500 kJ/mol) bonds are larger than that of N-N (150
kJ/mol), B-N bonds are the most favored bonds to form 3. However, B-B bonds
are the longest of the three bonds and unfortunately tend to form in addition
to B-N bonds. Therefore, a higher dosage time is required for NH3
compared to B2H6 to support the formation of B-N bonds
instead of B-B bonds 3.
The dosage time for the B2H6, using argon as a carrier
gas, will be 2 seconds followed by a minute purge of pure argon gas to ensure
all excess precursor and by-products are removed from the system. Following the
purge, a dosage of NH3, using hydrogen as a carrier gas, will occur
lasting 10 seconds. This will then be
followed by another argon gas purge to remove by-products and excess precursor.
Characterization of the films will guide adjustments to the processing of these
films to further understand the effects of dosage time on film structure.
Calculated Gibbs free energy values
of c-BN and h-BN show that c-BN is the thermodynamically stable phase at
standard conditions up until about 850 ºC 3.
Therefore, at temperatures up to 850 ºC, c-BN should be the favored phase. To
identify the effects of temperature on the crystal structure of boron nitride,
three temperatures will be tested; 300 ºC, 600 ºC, and 900 ºC. The lower two temperatures
are chosen due to the effects of B2H6 decomposition. At 300
ºC, the B2H6 precursor is stable but decomposes at
temperatures near 600 ºC 3. Furthermore, h-BN is
expected to grow at 900 ºC.
X-ray diffraction (XRD) will be
used to determine the presence of the sp3 cubic and/or the sp2
hexagonal phases of boron nitride in the deposited coating. This analysis will
allow for the development of a structural trend as a function temperature.
There are four characteristic 2-theta peaks located at 43°, 50°, 75°, 91° that
correspond to the crystalline planes of c-BN (111), (200), (220), and (311),
If these are the only peaks present, then the film will consist of
predominantly the cubic phase.
spectroscopy (FTIR) will be used to compliment XRD and aid in distinguishing
between h-BN and c-BN. The characteristic c-BN peaks consist of the B-N bending
vibration and the B-N-B bending vibration that occur at the wavenumbers ~1400
cm-1 and 800 cm-1, respectively 7, 21. Bulk c-BN crystals
exhibit a peak at about 1065 cm-1 7.
X-ray photoelectron spectroscopy
(XPS) will be used to determine whether the deposited films have a
stoichiometric B:N ratio of 1:1. If there are excess of either element, this
means that B-B bonds, N-N bond, or possibly hydrogen contamination exist as defects
within the films leading to a degradation of mechanical properties of the film.
The hydrogen contamination is not measurable via XPS, but if only boron and
nitrogen are present within the XPS data and their ratio is not 1:1, hydrogen
will be assumed to be the cause. B-B and N-N bonds in c-BN will result in
crystal defects, whereas B-B bonds will specifically result in twinning on the
Cubic Boron Nitride thickness
The conformality and thickness of
the films will be measured using transmission electron microscopy (TEM) providing
insight on the growth pattern of the boron nitride coating on non-flat surfaces
of the silicon substrates. Focused ion beam (FIB) will be used to prepare these
samples. Under TEM, a cross-section of the sample will be examined allowing for
the average thickness, conformality, and crystal structure of the boron nitride
coating to be measured.
Hardness, anti-wear, and friction
The hardness of the grown c-BN coatings
will be measured following the procedure by Yamada et al 22, utilizing a hysitron
nanoindentor equipped atomic force microscope (AFM) using a diamond tip with a
10 mN load. The hardness should be
around 50-60 GPa 23. A ball-on-plate
tribometer will be used to determine the wear and friction behavior of the
deposited films. Si3N4 will be used as a counter material
for 10,000 cycles 6.
c-BN is harder than the Si3N4, therefore it is expected
that the c-BN should only undergo polishing and exhibit a similar
root-mean-square (rms) roughness value (<0.5nm 10) before and after the wear test. This surface roughness comparison will determine how much wear the material has undergone after the wear test. If there is minimal change in surface roughness, which will be measured using AFM, then the material must have undergone minimal wear and thus had low friction and high hardness. The AFM scanning size will be 5x5 microns and will be measured using tapping mode. Summary: The proposed work will focus on the processing of a hard, anti-wear coating (c-BN) on flat and non-flat surfaces employing the technique of ALD. The Materials Project will be used to help determine the effects of substrate orientation on the preferred growth of c-BN in addition to exploring precursor chemistry favoring the formation of c-BN. The effects of temperature and dosage time will be explored to determine how processing conditions change material structure and properties. It is expected that the higher the temperature, longer dosage time, and (111) direction will favor the growth of c-BN. The formation of conformal c-BN, achieved by the alternating layer growth method of ALD and reactants B2H6 and NH3, will be determined using XRD and FTIR to confirm that c-BN has formed as opposed to h-BN. Nano-hardness tests will be used to determine whether the hardness of the close-packed cubic boron nitride film lies within the characteristic range of 50-60 GPa. Tribological wear tests will be conducted using the ball-on-plate tribometer, where a material of known hardness and coefficient of friction will repeatedly be scratched under a known load until the coating has worn off. Finally, the surface roughness of the deposited c-BN film will be measured before and after tribological testing using AFM, where a low surface roughness will result in low friction and a longer life-time of the coating and coated material.