CARBON 6 8 ( 2 0 1 4 ) 4 7 3 –4 7 9

Available at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier.com/locate/carbon

Chemical mechanical polishing of thin film

diamond
a,*
Evan L.H. Thomas , Geoffrey W. Nelson b, Soumen Mandal a, John S. Foord c
a,*
Oliver A. Williams
a
School of Physics and Astronomy, Cardiff University, Cardiff CF24 3AA, UK
b
Department of Materials, Imperial College, London SW2 2BP, UK
c
Department of Chemistry, University of Oxford, Oxford OX1 3TA, UK

A R T I C L E I N F O A B S T R A C T

Article history: The demonstration that Nanocrystalline Diamond (NCD) can retain the superior Young’s
Received 6 August 2013 modulus (1100 GPa) of single crystal diamond twinned with its ability to be grown at low
Accepted 8 November 2013 temperatures (<450 C) has driven a revival into the growth and applications of NCD thin
Available online 16 November 2013 films. However, owing to the competitive growth of crystals the resulting film has a rough-
ness that evolves with film thickness, preventing NCD films from reaching their full poten-
tial in devices where a smooth film is required. To reduce this roughness, films have been
polished using Chemical Mechanical Polishing (CMP). A Logitech Tribo CMP tool equipped
with a polyurethane/polyester polishing cloth and an alkaline colloidal silica polishing fluid
has been used to polish NCD films. The resulting films have been characterised with
Atomic Force Microscopy, Scanning Electron Microscopy and X-ray Photoelectron Spectros-
copy. Root mean square roughness values have been reduced from 18.3 nm to 1.7 nm over
25 lm2, with roughness values as low as 0.42 nm over 0.25 lm2. A polishing mechanism of
wet oxidation of the surface, attachment of silica particles and subsequent shearing away
of carbon has also been proposed.
 2013 Elsevier Ltd. All rights reserved.

1. Introduction However, diamond does not grow epitaxially on silicon,

requiring wafers to be seeded with nanodiamond particles
The demonstration that Nanocrystalline Diamond (NCD) prior to growth [4]. The subsequent competitive growth of
retains many of the superlative properties of single crystal these nanocrystals into coalesced NCD films results in a sur-
diamond in a low cost, large area wafer scale package, as well face roughness that evolves with film thickness and exceeds
as the possibility of CMOS integration due to the possibility of that of cleaved single crystal diamond. The increased rough-
growth at low temperatures (<450 C) has driven a resurgence ness of NCD films can be detrimental for many of its key
in research into the use of thin diamond films [1,2]. With a applications, such as the integration of AlN as a piezoelectric
high Young’s modulus of 1100 GPa, the highest phase velocity in MEMS and SAW applications, decreased Q [5–7] – spoiling of
of all materials of 12,000 m/s and thermal conductivity up to MEMS devices, and enhanced friction in tribological coatings.
2000 W/mK [3], applications include Micro-Electro-Mechani- To work around this roughness, previous reports have
cal Systems (MEMS), Surface Acoustic Wave (SAW) devices, used the nucleation side of freestanding NCD films, either
thermal management and tribological coatings. locally removing the silicon substrate or bonding/glueing

* Corresponding authors: Fax: +44 29208 74056.

E-mail addresses: [email protected], [email protected] (E.L.H. Thomas), [email protected] (O.A. Williams).

0008-6223/$ - see front matter  2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.carbon.2013.11.023
474 CARBON 6 8 ( 2 0 1 4 ) 4 7 3 –4 7 9

the wafer to another support and completely removing the converted to bound silanol groups (Si(OH)4) by the liquid
silicon [8]. However, this process is complicated, time con- polishing fluid, silica particles in the slurry then bond to the
suming and incompatible with some applications of NCD surface of passivation layer [15]. The moving polishing pad,
such as tribology and most MEMS structures. Therefore, there if sufficiently rough, will then create a force on the silica par-
is a real need for a polishing step to produce low roughness ticle. As long as the energy from this shearing action is larger
NCD films. than the binding energy, the polishing pad then removes the
Diamond polishing can be broadly split into two catego- particle and attached silanol molecule from the surface [16].
ries: contact and non-contact polishing [9,10]. Traditional In this paper, CMP of NCD films is reported with the use of
mechanical polishing of diamond involves pressing a sample silica based polishing fluid and a polyester/polyurethane pol-
against a fast rotating iron scaife, >2500 rpm, in the presence ishing pad at room temperature. It is important to note that
of a diamond grit and binder. With applied carrier down no diamond-based products were used in either the pad or
forces greater than 10 N micro-cracking of the diamond oc- slurry, unlike previous studies [17,18] making post CMP re-
curs, with roughness values dependent on the grade of grit moval of polishing materials simpler. Films have been studied
used [10,11]. However with this technique removal rates are with Scanning Electron Microscopy (SEM) to deduce morphol-
generally low, 10 nm/h, and the high forces on the sample ogy, Atomic Force Microscopy (AFM) to deduce roughness,
can cause deep fissures and create surface pits [11]. To en- and X-ray Photoelectron Spectroscopy (XPS) in an effort to
hance the polishing rate and reduce surface damage the hy- explain the polishing mechanism.
brid technique of Chemically Assisted Mechanical Polishing
and Planarisation (CAMPP) was developed. With this tech-
nique an oxidiser, typically potassium nitrate or potassium 2. Experimental procedure
hydroxide, heated to around 360 C is added to the mechani-
cal process [10,12]. After cracking by the scaife, the molten Silicon (1 0 0) p-type 2-inch wafers of 500 lm thickness were
oxidisers enter and convert diamond to carbon dioxide and used as substrates throughout. Before deposition all wafers
carbon monoxide weakening the surface and allowing further were cleaned using the standard SC-1 process of 30% H2O2:
micro-cracking to occur. Through this technique faster re- NH4OH:DI H2O (1:1:5) at 75 C for 10 min. The substrates were
moval rates and arithmetic roughness (Ra) values of 2.8 nm then rinsed in DI H2O in an ultrasonic bath for 10 min and
are achievable when used in conjunction with an initial spun dry. To seed, the wafers were placed in a mono-dis-
mechanical polish [12]. Other chemical based techniques persed nanodiamond/H2O colloid and agitated in an ultra-
have replaced the metal scaife with a roughened glass plate sonic bath for 10 min. This process is known to produce
and the diamond particles with slightly softer boron carbide nucleation densities exceeding 1011 cm 2 [4]. Once seeded
grit [13]. However, while these techniques makes it possible the wafers were rinsed, spun dry at 3000 rpm, and then
to polish films of several tens of microns thickness, for films immediately placed inside the CVD chamber.
with thickness in the hundreds of nanometres the wafer Chemical Vapour Deposition (CVD) of Nanocrystalline Dia-
bow can be significantly greater than the thickness of the film mond was carried out in a Seki 6500 series Microwave Plasma
(typically> 10 lm over a 2-inch Si wafer), as shown schemati- Reactor under 3% CH4/H2 conditions at 40 Torr and 3.5 kW
cally in Fig. 1. This will prevent uniform polishing across the microwave power. Upon termination of growth all films were
entire film and possibly cause shattering due to the rigidity cooled down in hydrogen plasma to ensure hydrogen termina-
of the scaife. Therefore a more flexible polishing pad is re- tion and prevent deposition of non-sp3 material. Substrate
quired in order to conform to the bowed sample. One such temperatures were approximately 840 C as determined by
method that is commonly used in the IC fabrication industry dual wavelength pyrometry, with heating solely from the
for the polishing of dielectric and metal interconnects is microwave induced plasma. Films were grown to 360 nm deter-
Chemical Mechanical Polishing (CMP). With this technique a mined in situ through the use of pyrometric interferometry,
softer polyester based polishing pad is used rather than a and ex-situ with a Filmetrics F-20 Spectral Reflectance system.
hard metal scaife with the aid of a colloidal silica (50– Chemical Mechanical Polishing was performed with a Log-
200 nm) [14] based polishing slurry at room temperatures. In itech Tribo polishing system in conjunction with a SUBA-X
conventional dielectric polishing Silicon Dioxide (SiO2) is polishing pad and Logitech supplied Syton SF-1 alkaline col-
loidal silica polishing slurry (15–50% SiO2, 9.2–10.1 pH, 4–5%
ethylene glycol). Before use, the pad was conditioned for
30 min to ensure a high surface roughness to maximise pol-
ishing action and slurry distribution [19]. During polishing
both pad and carrier were kept at 60 rpm rotating in opposite
directions, while the carrier swept across the pad as shown
schematically in Fig. 2. Down pressure was kept at 4 psi, while
a backing pressure of 20 psi was used in an attempt to present
a flat NCD film surface to the polishing pad. After initial wet-
Fig. 1 – Schematic of exaggerated wafer bow seen with NCD ting of the plate, the feed slurry rate was kept at 40 ml/min.
films. Due to the differing coefficients of thermal expansion Three films were polished for durations of 1, 2, and 4 h. After
of diamond and silicon, upon cooling from growth polishing the films were cleaned in an attempt to remove any
temperatures significant bowing will occur. (A colour remaining polishing slurry with a standard SC-1 clean as
version of this figure can be viewed online.) detailed previously.
 CARBON 6 8 (2 0 1 4) 4 7 3–47 9 475

are removed, while there is also a lack of voids opening up to

the substrate. Very little contamination from silica can also be
seen on the SEM images, initially suggesting an SC-1 clean is
enough to clear the surface of any loose polishing introduced
contamination.
Fig. 4 shows the AFM images of the as grown and 1–4 h
CMP films, while Table 1 shows the average roughness over
the 5 scans of 25 lm2 for each film. As can be seen the micro-
graphs reiterate this steady polishing, showing a decrease in
roughness from the as grown 18.3 nm rms to 1.7 nm rms over
the 25 lm2 scans. Also shown in blue on Fig. 4D is a smaller
area of 0.25 lm2 showing that a local roughness of
Fig. 2 – Schematic diagram of CMP tool. Sample held in
0.42 nm rms is achievable with CMP and the parameters used.
rotating carrier and swept across counter rotating polishing
The removal rate is approximately 16 nm/h for the three
cloth. Slurry distributed continually onto top of polishing
polished films.
cloth. (A colour version of this figure can be viewed online.)
As attempts to use Raman and the surface enhanced tech-
nique of Shell Isolated Nanoparticle Enhanced Raman
SEM images were taken with the SE2 detector of a Raith Spectroscopy (SHINERS) [23] were deemed inconclusive due
e-line SEM, operated at 10 kV, 10 mm working distance and to the swamping of the surface signal by the signal from the
20 k magnification. AFM was performed with a Veeco Dimen- bulk, XPS has been used to deduce polishing mechanism.
sion 3100 AFM operated in tapping mode and equipped with a
TESPA tip of 320 kHz resonant frequency, 8 nm radius, and 3.2. X-ray photoelectron spectroscopy
42 N/m spring constant. 5 areas of 25 lm2 were taken around
the centre of each sample, with post AFM analysis being car- XPS has been widely used to study CVD diamond films and
ried out with Gwyddion SPM analysis software. Removal rates previous studies are the basis for the present analysis
were calculated by comparing the average thickness of 13 [24,25]. Survey XPS spectra are shown in Fig. 5A, with major
points on each film before and after polishing with the Filmet- photoelectron and Auger peaks labelled. Unpolished and pol-
rics F-20 system. ished diamond films have significant C1s (285.0 eV) and O1s
XPS experiments were conducted using a VG ESCA Lab XPS (531.0 eV) character. Photoelectron signal originating from F,
spectrometer at 1 · 10 9 Torr, using an Al Ka radiation source S, Cl, and Si core levels are seen, particularly on polished
(1486.3 eV) at 10 kV anode with 10 mA emission current. The diamond films. Clearly the chemical polishing process is
Fixed Analyser Transmission (FAT) mode was used to obtain introducing non-diamond contamination to the surface.
spectra, using a pass energy of 50 eV or 25 eV for survey and However, this level of contamination is not sufficient enough
‘narrow’ XPS scans, respectively. All peak fitting was done to be detrimental to the use of NCD in MEMS devices and the
using XPS Peak Fit (v. 4.1) software. The reported binding ener- applications mentioned earlier. It is also highly probable a
gies have an error of ±0.25 eV, based on the calibration to the hydrofluoric acid clean or that similar post CMP cleaning pro-
C1s peak. Peak areas were normalised to the XPS cross-sec- cesses to the CMOS industry, such as hydrogen or oxygen
tion of the F1s photoelectron signal by use of the atomic sen- plasma exposure, can be developed to remove this surface
sitivity factors [20]. Elemental ratios were calculated from the contamination.
normalised peak areas and have an error of about 15–20% [21]. The C1s region is typically used to characterise changes to
the surface chemistry of diamond thin films. Representative
3. Results and discussion C1s spectra are shown in Fig. 5B. The C1s spectra were decon-
voluted into four chemical environments, as done for previ-
3.1. Morphology ous studies on CVD diamond thin films [24,25]: diamond
(CAC, 285.0 eV), hydrocarbon (CAH, 285.5 eV), ether (CAO,
SEM images of the as grown and 1–4 h polished films are 286.5 eV), and carbonyl (C@O, 287.5 eV). No evidence for the
shown in Fig. 3. The as grown film of Fig. 3A shows clear fac- presence of a significant amount of a fifth form, carboxyl
eting with crystal sizes of approximately 100–250 nm, as is (C(@O)OH, 288.5 eV) was found. By comparing Fig. 5B(i) and
typical for films grown under 3% methane admixture condi- 5B(ii), it can be seen that CMP does not significantly change
tions [22]. When comparing the as grown film to the 1 h the chemical termination of the CVD diamond surface,
CMP film of Fig. 3B a clear polishing action can be seen. Peaks although it does lead to subtle changes in the concentrations
of the crystals that come into contact with the polishing pad of the differing carbon species. However the most important
are removed first, followed by a progression down to the point conclusion from the chemical analysis is that significant
at which a neighbouring crystal is met for the 2 h and 4 h amounts of graphite or graphite related defects do not devel-
films of Fig. 3C and D. The resulting crystal plateaus appear op on the surface: it is well-known that treatments such as Ar
very smooth with little evidence of cracking, suggesting a ion bombardment [26] and electrochemical anodisation [27]
significant chemical action to the polishing. After 4 h of produce an sp2 type defect structure on the diamond inter-
polishing it can be seen that the film appears close to face, which is visible in XPS as a peak shifted by about 1 eV
optimum. With reference to Fig. 2 it can also be seen that a to a lower binding energy than that of the main diamond
point has been reached at which the majority of crystal peaks peak. It can be seen from Fig. 5B(ii) that such defects are not
476 CARBON 6 8 ( 2 0 1 4 ) 4 7 3 –4 7 9

Fig. 3 – SEM micrographs of as grown and polished films. CMP was used on three different films for the indicated amount of
time under identical conditions. (A) As grown, (B) 1 h CMP film, (C) 2 h, and (D) 4 h.

Fig. 4 – Corresponding AFM micrographs for as grown and polished films shown in Fig. 3. (A) As grown, (B) 1 h CMP film, (C)
2 h, and (D) 4 h. (A colour version of this figure can be viewed online.)
 CARBON 6 8 (2 0 1 4) 4 7 3–47 9 477

Table 1 – Roughness values over 25 lm2 for the as grown Table 2 – Elemental ratios of O1s, F1s, and Si2s with respect
and 1–4 h polished films. to C1s as a function of polishing duration.

Polishing duration (hrs.) Roughness (nm rms) Polishing duration (hrs.) O1s/C1s F1s/C1s Si2s/C1s

0 18.3 0 0.022 0.017 0.004

1 11.0 1 0.142 0.076 0.013
2 4.5 2 0.120 0.032 0.009
4 1.7 4 0.100 0.010 0.005

produced by the polishing procedure, again emphasising the cleaning of the samples before XPS analysis) do form on the
gentle nature of CMP. surface being polished.
The polishing process permanently increases the oxygen The F1s signal after 1 h of polishing was unexpected. After
character of the diamond surface. This is clearly demon- 4 h, this signal is reduced to near-negligible levels (see Table 1).
strated by the dramatic increase in the O1s/C1s ratio after The origin of this signal is likely the polymer-based pads used
1 h of polishing (see Table 2). Given that the XPS sampling to polish the diamond substrates, surfactants in the chemical
depth in diamond is probably up to 10 C layers, based on typ- solution, or solvent residue. All of these sources could be
ical electron elastic mean free paths, O:C ratios of 0.1 signify additional sources of Cl, S, and O photoelectron signal.
at least monolayer coverage. Curiously however the level of
C oxidation does not seem to be changed by this increase in 3.3. Discussion
oxygen. However noting the presence of the additional
elements F, Cl, Si, and S, it is clear the associated molecular The SEM and AFM images of Fig. 4 and Fig. 5 show a steady
species could contain O, and of course adsorbed water could polishing action with time. The polishing begins with the
also be present. The trend in the O1s/Si2s ratio in Fig. 5C removal of peaks due to the contact with polishing pad,
shows this increases with polishing time. This emphasises followed by a progression down to the intersection with
the point that the source of O in the XPS spectra is not limited neighbouring crystals. This initial polishing of high points
to silica, but has a main component from other species. Over- with the smoothness and apparent crack free nature of crys-
all the conclusion is that a range of molecular species which tal tops suggests a true chemical and mechanical synergy to
are fairly strongly bound (and therefore surviving the the polishing. This steady polishing is reiterated by the AFM

A B

Fig. 5 – (A) Survey XPS spectra of diamond substrates with main photoelectron and Auger signals indicated. From top to
bottom, the spectra represent as grown, 1 h CMP, 2 h, and 4 h respectively. The insert contains data from the main body of the
figure, showing XPS features at low binding energy.; (B) Representative C1s spectra before (i) and after 4 h (ii) of polishing. The
experimental data (black) is shown above the curve-fits (red).; (C) O1s/Si2s ratio plotted as a function of the polishing
duration. (A colour version of this figure can be viewed online.)
478 CARBON 6 8 ( 2 0 1 4 ) 4 7 3 –4 7 9

with an as grown rms roughness of 18.3 nm being reduced to

1.7 nm after 4 h. It can also be seen that for a smaller area,
closer to the size of an individual crystal, the roughness can
be as low as 0.42 nm. The removal rate seen of approximately
16 nm/h exceeds that typically possible with traditional
mechanical polishing [11], however this is heavily dependant
on the age and condition of the pad. While this is less than
the lm/h polishing rates of CAMPP, films in the hundreds of
nanometres with lower initial roughness can be polished
without possible cracking.
With regards to mechanism, contact polishing can be
broadly divided into three mechanisms: micro-chipping, con-
version to graphite, and chemical reaction [10]. Due to the
comparatively low hardness and flexibility of the polyester/
polyurethane polishing cloth as well as the lack of any dia- Fig. 6 – Proposed polishing mechanism. Wet oxidation of
mond based products in both the slurry and cloth, it is unli- hydrogen terminated diamond by polishing fluid increases
kely that micro-chipping is the cause of polishing. Coupled the carbonyl (C@O), and hydroxide (OH) content on the
with the low temperatures and pressures used, and the fact surface. As with the CMP of SiO2, hydroxide ions facilitate
that Ultrananocrystalline Diamond (UNCD), a material with attachment of silica particles to surface. Shearing forces
a higher sp2 content [1], burns at 600 C [28] it is also unlikely generated on silica particle by asperities of the rough
that conversion to CO and CO2 can be responsible with the polishing pad then removes carbon atom from surface,
polishing rates seen. providing polishing.
The lack of significant change in the graphitic content of
the films as indicated by XPS indicates that conversion to
graphite is not responsible for polishing. Typical techniques expensive diamond grit, or cast iron scaifes and instead uses
that rely on conversion to graphite utilise catalytic materials polyester/polyurethane polishing pads commonly found in
such as iron cobalt or nickel to lower the activation energy the IC fabrication industry. As shown, considerable action
and operate at temperatures of approximately 750 C [10], sig- can be seen without the need for raised temperatures or high
nificantly higher than the 30–50 C temperature of the waste pressures, simplifying the equipment required. Therefore
slurry. CMP is a promising method of achieving low roughness
Therefore it is proposed that the polishing mechanism fol- diamond surfaces at low cost.
lows that of the CMP of Silicon Dioxide. In traditional CMP,
hydroxide ions within the polishing fluid react with the sur- 4. Conclusion
face siloxane (SiAOASi) bonds, creating a silanol based pas-
sivation layer (Si(OH)4) [15,16]. Silica particles within the NCD films have been polished by CMP with the use of a poly-
polishing fluid will then attach themselves to the hydrated urethane/polyester felt and an alkaline colloidal silica polish-
groups of the passivation layer. Should the polishing pad then ing fluid (Syton SF-1). No diamond based products were used
be sufficiently rough, a shearing force will be created on the in either the slurry or polishing cloth. A final rms roughness
silica particles. If the energy from this shearing action is lar- value of 1.7 nm was achieved over 25 lm2, with values as
ger than the binding energy, the molecule will be removed low as 0.42 nm over 0.25 lm2. The polishing mechanism
resulting in polishing of the surface. proposed consists of the wet oxidation of the surfaces with
With diamond, XPS has shown that CMP leads to general the polishing fluid facilitating the attachment of silica parti-
oxidation of the interfacial region; increasing the carbonyl cles to the diamond film, followed by the shearing away of
and hydroxyl content of the surface. Drawing parallels to the particle due to forces from the polishing pad. Thus with
the hydroxyl bonding seen in the polishing of SiO2, we believe its low temperature, simple operation, ability to polish wafers
the OH termination facilitates the bonding of silica particles with significant bow, and already common CMOS industry
to the surface, as shown schematically in Fig. 6. As with supplies, CMP is an attractive method for the polishing of thin
SiO2 CMP the rough pad surface will then create a shear force film diamond.
on the silica particle. Due to the bond strengths of SiAO, OAC
and CAC being 800 kJ/mol, 1077 kJ/mol and 610 kJ/mol respec- Acknowledgments
tively [29], it is believed that when this force is applied the
CAC bond will break, polishing the film surface. Alternatively The authors wish to acknowledge the financial support of the
an oxidised silica particle can directly attach itself without EPSRC under the grant ‘Nanocrystalline diamond for Micro-
the need for intermediate wet chemical oxidation. As this is Electro-Mechanical-Systems’ reference number EP/J009814/1.
only a proposed model based on the mechanism on SiO2, fur- OAW wishes to acknowledge the support of Marie Curie
ther work is needed for validation and optimisation of the Actions for his intra-EU fellowship. The authors also wish to
CMP of diamond films. thank Sam Ladak at the Cardiff School of Physics and Astron-
Through the use of CMP it has been shown that bowed omy for assistance with AFM, and Mark Kennedy and John
thin film diamond can be polished without fear of cracking McCrossan at Logitech for their help with the early CMP
of films. The technique removes the need for the use of process.
 CARBON 6 8 (2 0 1 4) 4 7 3–47 9 479

R E F E R E N C E S [16] Hocheng H, Tsai HY, Su YT. Modeling and experimental

analysis of the material removal rate in the chemical
mechanical planarization of dielectric films and bare silicon
wafers. J Electrochem Soc 2001;148(10):G581–6.
[1] Williams OA. Nanocrystalline diamond. Diam Relat Mater
[17] Hocheng H, Chen CC. Chemical-assisted mechanical
2011;20(5–6):621–40.
polishing of diamond film on wafer. In: Jywe W, Chen CL, Fan
[2] Butler JE, Sumant AV. The CVD of nanodiamond materials.
KC, Fung RF, Hanson SG, Hsieh WH, et al., editors. Progress
Chem Vap Deposition 2008;14(7–8):145–60.
on advanced manufacture for micro/nano technology 2005,
[3] Philip J, Hess P, Feygelson T, Butler JE, Chattopadhyay S, Chen
Pt 1 and 2. Trans Tech Publications; 2006. p. 1225–30.
KH, et al. Elastic, mechanical, and thermal properties of
[18] Haisma J, Vanderkruis F, Spierings B, Oomen JM, Fey F.
nanocrystalline diamond films. J Appl Phys
Damage-free tribochemical polishing of diamond at room-
2003;93(4):2164–71.
temperature – a finishing technology. Precision Eng–J Am Soc
[4] Williams OA, Douheret O, Daenen M, Haenen K, Osawa E,
Precision Eng 1992;14(1):20–7.
Takahashi M. Enhanced diamond nucleation on
[19] McGrath J, Davis C. Polishing pad surface characterisation in
monodispersed nanocrystalline diamond. Chem Phys Lett
chemical mechanical planarisation. J Mater Process Technol
2007;445(4–6):255–8.
2004;153:666–73.
[5] Palasantzas G. Random surface roughness influence on gas
[20] Wagner CD, Davis LE, Zeller MV, Taylor JA, Raymond RH, Gale
damped nanoresonators. Appl Phys Lett 2007;90(4).
LH. Empirical atomic sensitivity factors for quantitative-
[6] Ergincan O, Palasantzas G, Kooi BJ. Influence of surface
analysis by electron-spectroscopy for chemical-analysis. Surf
modification on the quality factor of microresonators. Phys
Interface Anal 1981;3(5):211–25.
Rev B 2012;85(20).
[21] Seah MP. Quantification of AES and XPS. In: Briggs Das MP,
[7] Imboden M, Mohanty P. Dissipation in
editor. Practical surface analysis by auger and X-ray
nanoelectromechanical systems. Phys Rep 2013.
photoelectron spectroscopy. Chichester, UK: John Wiley &
[8] Rodriguez-Madrid JG, Iriarte GF, Williams OA, Calle F. High
Sons; 1983. p. 197–210.
precision pressure sensors based on SAW devices in the GHz
[22] Williams OA, Kriele A, Hees J, Wolfer M, Mueller-Sebert W,
range. Sens Actuators A–Phys 2013;189:364–9.
Nebel CE. High Young’s modulus in ultra thin nanocrystalline
[9] Widiez J, Rabarot M, Saada S, Mazellier JP, Dechamp J, Delaye
diamond. Chem Phys Lett 2010;495(1–3):84–9.
V, et al. Fabrication of silicon on diamond (SOD) substrates by
[23] Li JF, Huang YF, Ding Y, Yang ZL, Li SB, Zhou XS, et al. Shell-
either the bonded and etched-back SOI (BESOI) or the smart-
isolated nanoparticle-enhanced Raman spectroscopy. Nature
cut (TM) technology. Solid-State Electron 2010;54(2):158–63.
2010;464(7287):392–5.
[10] Malshe AP, Park BS, Brown WD, Naseem HA. A review of
[24] Ferro S, Dal Colle M, De Battisti A. Chemical surface
techniques for polishing and planarizing chemically vapor-
characterization of electrochemically and thermally oxidized
deposited (CVD) diamond films and substrates. Diam Relat
boron-doped diamond film electrodes. Carbon
Mater 1999;8(7):1198–213.
2005;43(6):1191–203.
[11] Schuelke T, Grotjohn TA. Diamond polishing. Diam Relat
[25] Wilson JIB, Walton JS, Beamson G. Analysis of chemical
Mater 2013;32:17–26.
vapour deposited diamond films by X-ray photoelectron
[12] Ollison CD, Brown WD, Malshe AP, Naseem HA, Ang SS. A
spectroscopy. J Electron Spectrosc Relat Phenom 2001;121(1–
comparison of mechanical lapping versus chemical-assisted
3):183–201.
mechanical polishing and planarization of chemical vapor
[26] Lau WM, Huang LJ, Bello I, Yiu YM, Lee ST. Modification of
deposited (CVD) diamond. Diam Relat Mater
surface band bending of diamond by low-energy argon and
1999;8(6):1083–90.
carbon ion-bombardment. J Appl Phys 1994;75(7):3385–91.
[13] Yuan Z, Jin Z, Zhang Y, Wen Q. Chemical mechanical
[27] Foord J, Hu JP. Electrochemical oxidation and reduction
polishing slurries for chemically vapor-deposited diamond
processes at diamond electrodes of varying phase purity.
films. J Manuf Sci Eng 2013;135(4):041006.
Phys Status Solidi A–Appl Mater Sci 2006;203(12):3121–7.
[14] Zantye PB, Kumar A, Sikder AK. Chemical mechanical
[28] Kim HJ, Moldovan N, Felts JR, Somnath S, Dai Z, Jacobs TDB,
planarization for microelectronics applications. Mater Sci
et al. Ultrananocrystalline diamond tip integrated onto a
Eng R Rep 2004;45(3–6):89–220.
heated atomic force microscope cantilever. Nanotechnology
[15] Krishnan M, Nalaskowski JW, Cook LM. Chemical mechanical
2012;23(49):495302.
planarization: slurry chemistry, materials, and mechanisms.
[29] Lide DR. CRC handbook of chemistry and physics. Boca
Chem Rev 2010;110(1):178–204.
Raton, FL: CRC Press; 2005.