Chemical-Mechanical Impact of Nanoparticles and [617741]

lubricants
Article
Chemical-Mechanical Impact of Nanoparticles and
pH Effect of the Slurry on the CMP of the Selective
Layer Surfaces
Filip Ilie1,* and George Ipate2
1Department of Machine Elements and Tribology, Polytechnic University of Bucharest,
313 Spl. Independentei, 060042 Bucharest, Romania
2Department of Biotechnical Systems, Polytechnic University of Bucharest, 313 Spl. Independentei,
060042 Bucharest, Romania; [anonimizat]
*Correspondence: ilie.filip@yahoo.com; Tel.: +40-21-402-9411
Academic Editor: Costas Charitidis
Received: 29 March 2017; Accepted: 19 May 2017; Published: 23 May 2017
Abstract: This paper provides a tribochemical study of the selective layer surface by chemical
mechanical planarization (CMP). CMP is used to remove excess material obtained in the process
of selective transfer. The paper aims at a better understanding of the planarization (polishing) and
micromachining. The planarization becomes effective if the material removal rate (MRR) is optimal
and the surface defects are minimal. The pH of the slurry plays a very important role in removing
the selective layer by CMP , and hydrogen peroxide (H 2O2) is the most common oxidizer used in
CMP slurry. The purpose of this paper is the analysis of the pH effect on the etching rate (ER) and on
the behavior of selective layer polishing by a constant concentration of H 2O2and the influence of
nanoparticles size and concentration on selective layer surface CMP . The nanoparticle size used is
250 nm. The MRR results through CMP and ER have been shown to be influenced by the presence
of oxides on the selective layer surface and have been found to vary with the slurry pH at constant
H2O2concentrations. The CMP slurry plays an important role in the CMP process performance
and should be monitored for optimum results and minimal surface defects. The paper analyzes
the impact of chemical-mechanical, inter-nanoparticle, and pad-nanoparticle-substrate interactions
on CMP performance, taking into account the state of friction at the interface, by measuring the
friction force. Selective layer CMP optimization studies were required to control the chemical and
mechanical interactions at the interface between the slurry and the selective layer, the slurry chemistry,
the properties, and the stability of the suspended abrasive nanoparticles.
Keywords: nanoparticle interaction; material removal rate; etching rate; friction at interface;
slurry stability; corrosion products
1. Introduction
The selective transfer phenomenon takes place in the friction process of two materials and in the
presence of specific lubricants by transferring the material from one element of a friction couple to
the other. This transfer can be performed safely if there is a favorable energy, a relative movement,
and if, in the friction area, is a copper-based material and a suitable lubricant (glycerin or a special
lubricant) [ 1]. Due to the structure formed on the contact surfaces by selective transfer, the friction force
is very low [ 2,3]. The predominant element in the selective layer structure is copper. By scanning with
an atomic force microscope (AFM) of the selective layer after the chemical mechanical planarization
(CMP) process it is found that the surface of the oxide layer is removed at different rates depending on
the depth of removal and the pH of the solution. CMP is a material removal and surface smoothing
process, which is possible by the combination of chemical and mechanical interactions [4–9].
Lubricants 2017 ,5, 15; doi:10.3390/lubricants5020015 www.mdpi.com/journal/lubricants

Lubricants 2017 ,5, 15 2 of 12
Although the basic principles of CMP are well known, selective layer CMP remains un-optimized
because the mechanism of selective layer removal during planarization (polishing) and the necessary
surface chemical processes are still unknown.
To optimize selective layer CMP it is necessary to control the chemical and mechanical interactions
at the interface between the slurry and selective layer during the planarization process [ 4,5]. It will
also require a deeper understanding of the material removal rate (MRR) during planarization and
a determination of the specific role of each chemical constituent in the slurry [6].
The chemical state of the selective layer surface and the pH of the slurries play a paramount role
in the selective layer CMP removal process. X-ray photoelectron spectroscopy (XPS) is an important
surface analytical tool that can be used to identify compounds formed on the metal surface due to
the action of chemicals present in the CMP slurry. In the literature, there are numerous studies on
similar topics [ 10–26] applied to the copper CMP of integrated circuits. It is worth mentioning that the
present study intends to extend to other fields, namely to surfaces coated through selective transfer
(the surfaces of the selective layer) in the friction process, considering the slurry pH (important for
removal through CMP of the selective layer) [ 7,22–26], H 2O2(the most common oxidant) [ 27–33],
size [ 21,28–31], and concentration [ 12,21,34] of nanoparticles used in the CMP slurry. Thus, this paper
explores the pH effect at a constant H 2O2concentration on the etching and polishing behavior of the
selective layer and the influences on size and concentration of nanoparticles during selective layer
surface CMP .
2. Experimental Details
Selective layer surface investigations related to the CMP process were carried out on a metal
pattern wafer (Workshop of the Department of Machine Elements & Tribology, Bucharest, Romania)
with the selective layer, as there are materials in which under optimum friction conditions, and in
the presence of lubricants, the wear phenomenon itself manifests as a transfer of material from one
element of the friction pair to the other, a specific selective transfer phenomenon. This is a requirement
for any high-efficiency friction couple and a normal process for the self-adjustment phenomenon if
there is a favorable energy in the presence of relative movement.
A selective transfer can be achieved in a friction couple if, in the friction area, there is a material
consisting of copper alloy (bronze) and the adequate lubricant (glycerin). The metal (steel) pattern
wafer with the selective layer was obtained in the friction process between a steel wafer and bronze
in the presence of glycerin, when a material transfer takes place from the bronze surface to the steel
surface, forming a thin selective layer, in which copper is predominant ( 85%), with a different
structure than copper obtained through normal electrolytic procedures.
The planarization was performed on 25 mm 25 mm samples cleaved from a 250 mm diameter
patterned metal wafer with a selective layer (OLC45 steel coated with a selective layer formed in
the friction process with CuSn12T bronze). Sample discs coated with a selective layer containing
85% copper in purity and 25 mm in diameter were polished with 600 grit sandpaper and thoroughly
cleaned with acetone, isopropanol, and distilled water. The specimens were then immersed in pH buffer
solutions with 5.5 vol % H 2O2for a certain period of time. Subsequently, XPS examination (PHI-5702
electron spectrometer, ULVAC-PHI.INC., Chigasaki, Japan) and scanning electron microscopy
(SEM, AA7000 Scanning Electron Microscope, Angstrom Advanced Inc., Stoughton, MA, USA)
was performed.
The polishing experiments were carried out using a CMP tool (UMT TriboLab ™, San Jose, CA,
USA): a rotary type polisher CP-4 with perforated, nonwoven pads. The load is uniformly applied to
the wafer support using air pressure. In this study, the pressure and polishing linear velocity were
maintained constant at 20 kPa and 30 m/min (0.5 m/s), respectively.
The CMP polisher is equipped with an in situ monitoring system of both friction force
(with a piezoelectric force sensor embedded in the polishing head) and temperature. The sensor

Lubricants 2017 ,5, 15 3 of 12
signals were amplified and transferred to a data acquisition system. A thermostat (Carrier®—United
Technologies, Pittsburg, PA, USA) was used for the temperature control.
The colloidal silica-based slurry for polishing, which contains citric acid as a pH buffer and
an oxidizer (H 2O2), was used for the polishing experiments. The colloidal silica concentration of the
slurry was maintained constant at 4.5 wt %, and the pH of the slurry mixture was kept constant at
pH 4. The applied downward pressure was about 0.133 MPa. The polishing time was 5 min and the
polishing rate is an average of three independent polishes.
The removal rate was determined by measuring the weight loss of the platen. To prevent
agglomeration of the nanoparticles the slurry used was continuously agitated in all polishing tests and
directed to the polishing head at a flow rate of 200 mL/min and by 10 mm selective layer wafers.
The chemical elements of the CMP slurry react with the polished wafer surface and form
a chemically-modified top layer with properties different from those of the initial wafer surface.
For example, passivation of the patterned wafer from a metallic selective layer oxide film prevents the
metallic selective layer etching in the low areas, while the high asperities are repeatedly oxidized by
the chemicals and abraded by the slurry nanoparticles until the planarization is reached. Consequently,
slurry oxidizers provide topographic selectivity during metallic selective layer CMP . In selective layer
planarization, the chemical effect is provided by the slurry pH.
The absence of chemically-modified layer is due to the lack of material removed by
polishing [25,34,35]. It was found that the polishing by chemical means leaves an etched surface
with no planarity, whereas polishing only by mechanical means results in a rough surface. Therefore,
the formation of a chemically-altered film is necessary to achieve optimal CMP performance.
3. Results and Discussion
3.1. Technological and Technical Issues
It is essential that, during the CMP process, the superior characteristics of a sample to be polished
at a rate higher than that of the lower characteristics leads to a surface planarization. This requires that
the pad has to be rigid enough so that it does not deform the supplementary recessed sample areas
and the slurry does not etch the lower characteristics during CMP . Thus, the colloidal silica-based
acidic slurry (incorporating silica abrasive particles, H 2O2as an oxidizer, and citric acid as a pH buffer)
used here shows an excellent removal effect and a good behavior of planarization.
The CMP slurry abrasive particles provide mechanical action during the polishing process.
Hence, the chemically-modified wafer surface layer is abraded continuously by the slurry’s abrasive
nanoparticles (with nanometric dimensions), which results in material removal. By optimizing the
chemical modification and mechanical polishing rates optimal polishing performance can be achieved
(minimal deformations and good planarity). The mechanical polishing intensity also varies with the
slurry nanoparticle size and concentration, as factors that determine the load applied per nanoparticle.
In addition, the mechanical polishing depends on the number of abrasive nanoparticles in the slurry
in contact with the wafer surface. These are the very important factors for determining the polishing
performance and for understanding the CMP process. Major changes in the nanoparticles-substrate
interactions can appear as a result of a variation in the nanoparticle size distribution in the slurry,
due to contamination of oversized nanoparticles or loss of the slurry stability [ 4,7,17]. Thus, the MRR
response can vary and can be manifested in the poor control process increasing the number of surface
deformations, giving birth to micro/nano-scale defects. To check the excellent polishing results observed
on selective layer wafers, which are also seen on patterned selective layer samples, the structures were
polished with the same slurry used above for selective layer CMP . Thus, Figure 1 shows MRR by CMP
and the static ER of the selective layer in slurries with 5.5% H 2O2and at different pH values.
A similar trend of MRRs and ERs with the variation of pH was observed. Both the selective layer
MRR and the ER decreased with the increasing pH and reach a minimum at pH 6.5, then the removal
and etching rates start to increase with increasing pH.

Lubricants 2017 ,5, 15 4 of 12

Figure 1

Figure 1. Removal rate and etch rate of selective layer CMP in slurries with 5.5% H 2O2at different pH.
3.2. Effect of the Slurry Nanoparticle Size and Concentration
Since CMP MRR is dependent on the nanoparticle size and concentration in the slurry, the effects
have been investigated by many researchers [ 26–36]. The reported results are often contradictory and
are explained based on the selected ranges of the nanoparticle size and concentration, as well as by
their behavior during the CMP process. Thus, the polishing mechanisms depend on the CMP process,
as well as on the slurry nanoparticle properties, for a given polishing system.
Two material removal mechanisms were considered to elucidate the observed modifications in
polishing rates, which can be explained explicitly on the basis of silicon CMP [ 24,25,37]. Figure 2
illustrates the MRR obtained with the nanoparticle size of 250 nm in silica slurry depending on the
slurry nanoparticle concentration.

Figure 2

Figure 2. Material removal rate (MRR) and silica friction force, based on a 250 nm concentration of
silicon abrasive nanoparticles.

Lubricants 2017 ,5, 15 5 of 12
At a low nanoparticle concentration in the slurry (from 2.0 to 5.0 wt %), MRR increases with
the increasing nanoparticles concentration. The wafer surface micrographs are obtained with help
of an AFM, and at these concentrations show scratches on the polished wafer surface indicating that
the abrasive nanoparticles start rolling and slipping on the wafer surface which results in mechanical
removal. MRR reaches a maximum of 5.5 wt % nanoparticle concentration, after which a significant
decrease is detected. This significant decrease is due to the reduction of the load on the nanoparticles
with the increase in the number of nanoparticles in the wafer contact area, respective of the changes in
the nanoparticle motion.
More nanoparticles tend to start rolling rather than sliding over the wafer surface when the
solids concentration increases from 5.0 to 15.0 wt %. As a result, MRR decreases with the increasing
solids concentration. Furthermore, surface AFM microphotography at 15.5 wt % exhibits pitting
deformations of the wafer surface faster than scratches, proving that the abrasive nanoparticles are in
rolling motion (see Figure 3). Therefore, at these higher solids concentrations, mechanical action is
significantly reduced and MRR takes place mainly due to chemical interactions.
Lubricants 2017 , 5, 15 5 of 12
nanoparticles with the increase in the number of nanoparticles in the wafer contact area, respective
of the changes in the nanoparticle motion.
More nanoparticles tend to start rolling rather than sliding over the wafer surface when the
solids concentration increases from 5.0 to 15.0 wt %. As a result, MRR decreases with the increasing
solids concentration. Furthermore, surface AFM microphotography at 15.5 wt % exhibits pitting
deformations of the wafer surface faster than scratches, provin g that the abrasive nanoparticles are
in rolling motion (see Figure 3). Therefore, at these higher solids concentrations, mechanical action
is significantly reduced and MRR takes place mainly due to chemical interactions.

(a) (b)
Figure 3. AFM mi crographs at 5.0 wt % with scratches ( a), and at 15.5 wt % with pitting
deformations ( b) of the wafer surface.
In support of these mechanisms, in situ friction force measurements conducted on the same
principle also show a similar trend in the total fricti on force depending on the 250 nm nanoparticle
concentration in the slurry (Figure 2). It has also been shown that in situ friction force
measurements in the CMP correlate with the MRR in [38 –43], i.e., the variation is based on the same
considerations (dec rease of the load per particle with the increased concentration and start of the
nanoparticle rolling motion, which leads to a decrease of the frictional force).
At the same time, it is necessary to reduce the number of surface deformations in the CMP
proc ess. For this, it is necessary to act in the contact area by using small -sized nanoparticles at high
nanoparticle concentrations. This allows the determination of polishing performance of the slurry.
Selective layer samples are immersed and polished in the slurry containing H 2O2 and citric acid at
different pHs and then analyzed using XPS. Surface analysis results of the samples immersed in 5.5
vol % H 2O2 are shown in Figure 4.
Figure 4a shows the XPS spectra for selective layer copper at pH 2, 4, 6, and 8 solutions after 30
min of immersion. The binding energy (BE) value of the Cu 2O at 932.2 eV was observed for the
samples of the all four pHs. The relative photoelectron signal intensity to pH 2 and pH 8 samples
compared to pH 4 and pH 6 indicate that signa ls possibly respond not only from Cu 2O on the
surface, but also from the metal substrate of the selective layer.
XPS spectrum analysis shows additional peaks at 933.5 eV and 934.5 eV for samples immersed
in a slurry with pH 2 and pH 8. The selective layer copper in CuO and Cu 2O have BE values of
about 933.4 eV and 934.5 eV, respectively [35,44]. CuO signals can be identified clearly for the
sample of pH 8. These results were further confirmed in the selective layer copper by Auger
electron spectroscopy (AES ), as shown in Figure 4b.
Figure 3. AFM micrographs at 5.0 wt % with scratches ( a), and at 15.5 wt % with pitting deformations
(b) of the wafer surface.
In support of these mechanisms, in situ friction force measurements conducted on the same
principle also show a similar trend in the total friction force depending on the 250 nm nanoparticle
concentration in the slurry (Figure 2). It has also been shown that in situ friction force measurements
in the CMP correlate with the MRR in [ 38–43], i.e., the variation is based on the same considerations
(decrease of the load per particle with the increased concentration and start of the nanoparticle rolling
motion, which leads to a decrease of the frictional force).
At the same time, it is necessary to reduce the number of surface deformations in the CMP process.
For this, it is necessary to act in the contact area by using small-sized nanoparticles at high nanoparticle
concentrations. This allows the determination of polishing performance of the slurry. Selective layer
samples are immersed and polished in the slurry containing H 2O2and citric acid at different pHs
and then analyzed using XPS. Surface analysis results of the samples immersed in 5.5 vol % H 2O2are
shown in Figure 4.
Figure 4a shows the XPS spectra for selective layer copper at pH 2, 4, 6, and 8 solutions after
30 min of immersion. The binding energy (BE) value of the Cu 2O at 932.2 eV was observed for the
samples of the all four pHs. The relative photoelectron signal intensity to pH 2 and pH 8 samples
compared to pH 4 and pH 6 indicate that signals possibly respond not only from Cu 2O on the surface,
but also from the metal substrate of the selective layer.
XPS spectrum analysis shows additional peaks at 933.5 eV and 934.5 eV for samples immersed in
a slurry with pH 2 and pH 8. The selective layer copper in CuO and Cu 2O have BE values of about
933.4 eV and 934.5 eV , respectively [ 35,44]. CuO signals can be identified clearly for the sample of pH 8.
These results were further confirmed in the selective layer copper by Auger electron spectroscopy
(AES), as shown in Figure 4b.
Figure 5 shows the SEM micrographs of the selective layer surface, obtained in solutions with
different pH values after 30 min of immersion.

Lubricants 2017 ,5, 15 6 of 12
Lubricants 2017 , 5, 15 6 of 12

Figure 4. Selective layer copper XPS ( a) and AES spectra ( b) of Cu 2O films formed in solutions with
5.5% H 2O2 at different pHs after 30 min of immersion.
Figure 5 shows the SEM micrographs of the selective layer surface, obt ained in solutions with
different pH values after 30 min of immersion.

(a) (b)
Figure 4. Selective layer copper XPS ( a) and AES spectra ( b) of Cu 2O films formed in solutions with
5.5% H 2O2at different pHs after 30 min of immersion.
Lubricants 2017 , 5, 15 6 of 12

Figure 4. Selective layer copper XPS ( a) and AES spectra ( b) of Cu 2O films formed in solutions with
5.5% H 2O2 at different pHs after 30 min of immersion.
Figure 5 shows the SEM micrographs of the selective layer surface, obt ained in solutions with
different pH values after 30 min of immersion.

(a) (b)
Figure 5. Cont .

Lubricants 2017 ,5, 15 7 of 12
Lubricants 2017 , 5, 15 7 of 12

(c) (d)
Figure 5. SEM micrographs for the selective layer surface formed in solutions with 5.5% H 2O2 at
various pH after 30 min of immersion: ( a) pH 2, ( b) pH 4, 9( c) pH 6, and ( d) pH 8.
Figure 5a shows the selective layer surface morphology, obtained in pH 2 solution. The surface
morphology indicates a strong etching reaction and the surface was coated with a discontinuous
layer of amorphous corrosion products, whic h are Cu 2O and CuO, according to XPS analysis. In
this situation, Cu dominates the selective layer so that a higher MRR through CMP was obtained.
Figure 5b shows the selective layer surface morphology obtained in solution at pH 4. Polishing
scratches on th e substrate surface could be seen in the micrograph even after a 30 min immersion,
indicating that the layer formation rate is quite low. This is also confirmed by XPS results. Figure 5c
shows the surface morphology of the selective layer obtained in solut ion with pH 6. The whole
surface was covered with CuO and Cu 2O crystals. This layer was porous, but compact, and thus
protective. Since the CuO oxide film is much tougher than the selective layer [1], a relatively lower
MRR through CMP was observed. Figure 5d shows the surface morphology of the selective layer
obtained in a solution at pH 8. A much thicker film formed on the substrate surface can be
observed, but it is slightly anchored and oxide film cracks are seen observed everywhere. It can,
therefore, be easier to remove it by CMP.
The study was conducted at a constant H 2O2 concentration. In order to observe the slurry pH
effects, the MRR and ER of the selective layer were measured as a function of pH at 5.5 vol % H 2O2
concentration. Figure 6 shows that the selective layer MRR and ER at 5.5 vol % H 2O2 concentration
in the slurry decrease with increasing pH.

Figure 6. Material removal rate (MRR) and etch rate (ER) vs. pH at 5.5 vol % H 2O2 slurry.
Figure 5. SEM micrographs for the selective layer surface formed in solutions with 5.5% H 2O2at
various pH after 30 min of immersion: ( a) pH 2, ( b) pH 4, 9( c) pH 6, and ( d) pH 8.
Figure 5a shows the selective layer surface morphology, obtained in pH 2 solution. The surface
morphology indicates a strong etching reaction and the surface was coated with a discontinuous
layer of amorphous corrosion products, which are Cu 2O and CuO, according to XPS analysis. In this
situation, Cu dominates the selective layer so that a higher MRR through CMP was obtained. Figure 5b
shows the selective layer surface morphology obtained in solution at pH 4. Polishing scratches on the
substrate surface could be seen in the micrograph even after a 30 min immersion, indicating that the
layer formation rate is quite low. This is also confirmed by XPS results. Figure 5c shows the surface
morphology of the selective layer obtained in solution with pH 6. The whole surface was covered with
CuO and Cu 2O crystals. This layer was porous, but compact, and thus protective. Since the CuO oxide
film is much tougher than the selective layer [ 1], a relatively lower MRR through CMP was observed.
Figure 5d shows the surface morphology of the selective layer obtained in a solution at pH 8. A much
thicker film formed on the substrate surface can be observed, but it is slightly anchored and oxide film
cracks are seen observed everywhere. It can, therefore, be easier to remove it by CMP .
The study was conducted at a constant H 2O2concentration. In order to observe the slurry pH
effects, the MRR and ER of the selective layer were measured as a function of pH at 5.5 vol % H 2O2
concentration. Figure 6 shows that the selective layer MRR and ER at 5.5 vol % H 2O2concentration in
the slurry decrease with increasing pH.

Figure 2

Figure 6. Material removal rate (MRR) and etch rate (ER) vs. pH at 5.5 vol % H 2O2slurry.

Lubricants 2017 ,5, 15 8 of 12
The growth of the Cu 2O AES peak (see Figure 4b, where BE~337 eV) with increasing pH is
demonstrated by the fact that there is a larger amount of Cu 2O on the surface at higher pH values.
The higher amount of Cu 2O protects (passivates) the metal surface with the selective layer and reduces
the oxidation rate of the metallic selective layer at higher pH values. The presence of these cupric
oxides (Cu 2O, CuO) is indicated by the Auger lines at BE 934.5 eV , as seen in the spectrum in Figure 4a.
The MRR behavior in Figure 6 can be explained by the pH-induced effects. At 5.5 vol % H 2O2
concentrations, a CuO film is present on the metallic selective layer surface during polishing. In this
case, we note that the MRR decreases with increasing pH due to decreased CuO solubility, possibly
limited by the solubility of CuO film that forms on the selective layer metallic surface, during CMP .
3.3. Slurry Nanoparticle Size Distribution
To obtain an optimum polishing performance with minimal surface deformations, it is necessary
to use mono-modified nanoparticles in the CMP slurries [ 35]. However, in practice, there may
be nanoparticles of larger size in the slurries due to insufficient filtration (hard nanoparticles), or
primary slurry nanoparticles due to poor stability (soft nanoparticles). The presence of different
size nanoparticles in the CMP slurries unevenly distributes the load applied by the polisher head
on the abrasive particles, resulting in surface deformations [ 32,37]. Although the presence of hard
nanoparticles was suspected of causing major surface deformations [ 32,35,37], the polishing tests
made it possible to establish degradation detection limits to check planarization performances.
Surface analysis of selective layer wafers polished with hard nanoparticles in the slurry showed
an increased surface roughness and several relative surface deformations, compared to baseline
polishing, as shown in the AFM images from Figure 7a,b. In addition, significant modifications were
detected in the presence of these nanoparticles and in the MRR response, indicating that measures
must be taken to protect the surface quality and also to achieve a suitable MRR, by removing them
from the slurry.
Lubricants 2017 , 5, 15 8 of 12
The growth of the Cu 2O AES peak (see Figure 4b, where BE~ 337 eV) with increasing pH is
demonstrated by the fact that there is a larger amount of Cu 2O on the surface at higher pH values.
The higher amount of Cu 2O protects (passivates) the metal surface with the selective layer and
reduces the oxidation rate of th e metallic selective layer at higher pH values. The presence of these
cupric oxides (Cu 2O, CuO) is indicated by the Auger lines at BE 934.5 eV, as seen in the spectrum in
Figure 4a.
The MRR behavior in Figure 6 can be explained by the pH -induced effects. A t 5.5 vol % H 2O2
concentrations, a CuO film is present on the metallic selective layer surface during polishing. In this
case, we note that the MRR decreases with increasing pH due to decreased CuO solubility, possibly
limited by the solubility of CuO film that forms on the selective layer metallic surface, during CMP.
3.3. Slurry Nanoparticle Size Distribution
To obtain an optimum polishing performance with minimal surface deformations, it is
necessary to use mono -modified nanoparticles in the CMP slurries [35]. However, in practice, there
may be nanoparticles of larger size in the slurries due to insufficient filtration (hard nanoparticles),
or primary slurry nanoparticles due to poor stability (soft nanoparticles). The presence of different
size nanoparti cles in the CMP slurries unevenly distributes the load applied by the polisher head on
the abrasive particles, resulting in surface deformations [32,37]. Although the presence of hard
nanoparticles was suspected of causing major surface deformations [32,35 ,37], the polishing tests
made it possible to establish degradation detection limits to check planarization performances.
Surface analysis of selective layer wafers polished with hard nanoparticles in the slurry showed an
increased surface roughness and se veral relative surface deformations, compared to baseline
polishing, as shown in the AFM images from Figure 7a,b. In addition, significant modifications
were detected in the presence of these nanoparticles and in the MRR response, indicating that
measures must be taken to protect the surface quality and also to achieve a suitable MRR, by
removing them from the slurry.

(a) (b)
Figure 7. Surface response quality of the selective layer wafers polished with: ( a) baseline slurry; ( b)
hard nanoparticles in slurry and 1 wt % 250 nm nanoparticles.
To remove larger nanoparticles, a filtration of CMP slurries is usually conducted. Nevertheless,
even after filtering of the slurries, defects are often observed on the polished surfaces and may be
higher than expect ed [44]. As a result, there is a suspicion that some of the defects may be created
by the agglomeration of nanoparticles formed during the CMP process. These observations indicate
that CMP slurries must remain stable during polishing to achieve optimal pla narization
performances.
3.4. Stabilization of CMP Slurries
In the planarization process, the CMP slurry has to be the stabilized in the presence of reactive
additives and under extreme conditions of pH, pressure, temperature, and ionic strength. For the
stabilization of the nanoparticles a surfactant is used at the solid –liquid interface [16,44 –50] because
the presence of surfactants allows lubrication between the abrasive and the surface to be polished
[50]. Thus, friction forces can decrease, resulting i n reduced MRR in the presence of the surfactants.
Figure 7. Surface response quality of the selective layer wafers polished with: ( a) baseline slurry;
(b) hard nanoparticles in slurry and 1 wt % 250 nm nanoparticles.
To remove larger nanoparticles, a filtration of CMP slurries is usually conducted. Nevertheless,
even after filtering of the slurries, defects are often observed on the polished surfaces and may be
higher than expected [ 44]. As a result, there is a suspicion that some of the defects may be created by
the agglomeration of nanoparticles formed during the CMP process. These observations indicate that
CMP slurries must remain stable during polishing to achieve optimal planarization performances.
3.4. Stabilization of CMP Slurries
In the planarization process, the CMP slurry has to be the stabilized in the presence of reactive
additives and under extreme conditions of pH, pressure, temperature, and ionic strength. For the
stabilization of the nanoparticles a surfactant is used at the solid–liquid interface [ 16,44–50] because
the presence of surfactants allows lubrication between the abrasive and the surface to be polished [ 50].

Lubricants 2017 ,5, 15 9 of 12
Thus, friction forces can decrease, resulting in reduced MRR in the presence of the surfactants.
The result was that the surface quality was optimal, with the roughness and deformation of the selective
layer surface being minimal. Therefore, this emphasizes the importance of nanoparticle–substrate
interactions along with the nanoparticle-nanoparticle interactions to optimize the CMP process.
To do this, it was necessary to calculate the applied force on a single abrasive nanoparticle during
polishing. This value was estimated to be 100–1000 nN for a 200 nm nanoparticle following the method
used in [ 51,52] and was in line with the modeling values from the literature [ 5,10–12,14,51,52]. The use
of surfactants showed a nonlinear behavior of the lateral forces, observing a normal force, for 250 nm
nanoparticles, as shown in Figure 8.
Figure 6

Figure 8
Figure 8. AFM friction force measurements on a selective layer wafer with 250 nm nanoparticles
attached to the tip. Note: Slurries contain surfactants at 35, 70, and 140 mM concentrations at pH 8.
Although the MRR values were low compared to the baseline, this approach allowed a new
method to determine the friction forces occurring in the CMP process. Several in situ studies of the
friction force response and nanoparticle-substrate friction simulations with AFM for slurries with
added surfactant have proved that the adhesion strength of the surfactant and slurry chemistry
(composition, pH, ionic strength, and surfactant) can be adapted to provide a sufficient frictional force
obtaining an optimal planarization (polishing) performance [ 51,52]. Hence, the effect of surfactant
concentration on stabilization ability of the CMP slurry revealed an optimal concentration range
for a certain surfactant. However, by use of a surfactant in the CMP slurry, the surface quality was
improved. Therefore, surfactants can be used to stabilize nanoparticles in a CMP slurry, to modify
the interactions between nanoparticles, and between nanoparticles and the substrate, during CMP to
optimize the process.
4. Conclusions
The MRR of the selective layer during CMP or etching is influenced by the presence of selective
layer oxides on the surface.
Surface analysis revealed the presence of Cu 2O on the selective layer surface after immersing
selective layer samples in a slurry with different pHs containing H 2O2.
The presence of the Cu 2O film reduced the ER of the selective layer. Surface analysis also revealed
the presence of CuO on the selective layer metallic surface. Corrosion products formed on the selective
layer surface were found to be responsible for MRR through CMP .
MRR through CMP and ER of the selective layer in slurries with 5.5 vol % H 2O2have been shown
to vary with pH. When the pH of the slurry was varied, the polishing rate decreased with increasing
pH at constant H 2O2concentrations.

Lubricants 2017 ,5, 15 10 of 12
The MRR of the selective layer during CMP for the slurry used in this paper appears to be
determined by the oxidation rate of the metallic selective layer and by the solubility of a Cu 2O or
CuO film.
In the CMP process, the slurry plays an important role in the process performance and should be
closely monitored for optimal results. Slurry chemistry should be tailored to contain no hard particles
or soft particles in order not to deteriorate the surface quality and lead to inconsistent MRRs.
For an effective planarization, MRR must be optimal and surface defects must be minimal. In order
to optimize selective layer CMP , it is necessary to control the chemical and mechanical interactions
at the interface between the slurry and selective layer surface, the slurry chemistry, the properties,
and the stability of the abrasive nanoparticles, respectively, leading to a deeper understanding of
MRR during the planarization process and the specific role of each chemical constituent in the slurry.
Therefore, the chemical state of the selective layer surface and the pH of the slurries play a very
important role in the selective layer CMP removal process, and H 2O2is the most common oxidizer
used in the CMP slurry.
By using surfactants at the solid-liquid interface to stabilize the slurry in extreme conditions,
the CMP process is improved with optimal effects on the surface quality (minimal roughness
and deformations).
For the CMP process to be effective, the friction state at the interface must also be taken into
account by measuring the friction force due to the impact of chemical-mechanical, inter-particle,
and pad-nanoparticle-substrate interactions.
Acknowledgments: We would like to express the sincere gratitude of the technicians of the Department of
Machine Elements and Tribology (DME&T) for the support in the realization of the metal pattern wafer with
the selective layer in the workshop and the laboratories of the department, from which samples cleaved
for experiments.
Author Contributions: Filip Ilie and George Ipate conceived and designed the experiments; Filip Ilie and
George Ipate performed the experiments; Filip Ilie analyzed the data; George Ipate contributed with materials,
and prepared solutions and the analysis tools; and Filip Ilie wrote the paper.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Ilie, F. Studytribological of Thin Superficial Layers Formed in the Friction Couples Through Selective Transfer ;
Technical Publishing House: Bucharest, Romania, 2002.
2. Ilie, F.; Tita, C. Comparative analysis of tribological performances of materials that can achieve thin layers
from selective transfer. J. BalkanicalTribol. Assoc. 2006 ,12, 404–411.
3. Garkunov, D.N. Tribotechnology (Wear and Non-Wear) ; Publishing House MSKHA: Moscow, Russia, 2001.
(In Russian)
4. Ilie, F. Tribochemical interaction between nanoparticles and surfaces of selective layer during chemical
mechanical polishing. J. Nanopart. Res. 2013 ,15, 1997. [CrossRef]
5. Ilie, F. Models of nanoparticles movement, collision, and friction in chemical mechanical polishing (CMP). J.
Nanopart. Res. 2012 ,14, 752–759. [CrossRef]
6. Armini, S.; Whelan, C.M.; Moinpour, M.; Maex, K. Copper CMP with Composite Polymer Core Silica Shell
Abrasives: A Defectivity Study. J. Electrochem. Soc. 2009 ,156, H18–H26. [CrossRef]
7. Nolan, L.; Cadien, K. Chemical Mechanical Polish for Nanotechnology. In Nanolubrification ; Stepanova, M.,
Dew, S., Eds.; Springer: Vienna, NY, USA, 2012.
8. Guo, D.; Xie, G.; Luo, J. Mechanical properties of nanoparticles: Basics and applications. J. Phys. D 2014 ,47,
1–25. [CrossRef]
9. Dubin, V .M. Copper Electroplating for On-Chip Metalization. In Advanced Interconnects for ULSI Technology ;
Baklanov, M.R., Ho, S., Zschech, E., Eds.; Jhon Wiley and Sons Ltd.: West Sussex, UK, 2012.
10. Luo, J.; Dornfeld, D.A. Material removal mechanism in chemical mechanical polishing: Theory and modeling.
IEEE Trans. Semicond. Manuf. 2001 ,14, 112–133.

Lubricants 2017 ,5, 15 11 of 12
11. Gopal, T.J.; Talbot, J.B. Use of Slurry Colloidal Behavior in Modeling of Material Removal Rates for copper
MP . J. Electrochem. Soc. 2007 ,154, H507–H512. [CrossRef]
12. Ihnfeldt, R.; Talbot, J.B. Modeling of Copper CMP Using the Colloidal Behavior of an Alumina Slurry with
Copper Nanoparticles. J. Electrochem. Soc. 2007 ,154, H1018–H1026. [CrossRef]
13. Pirayesh, H.; Cadien, K. The Effect of Slurry Properties on the CMP Removal Rate of Boron Doped Polysilicon.
ECS J. Solid State Sci. Technol. 2016 ,5, P233–P238. [CrossRef]
14. Ihnfeldt, R.; Talbot, J.B. The Effects of Copper CMP Slurry Chemistry on the Colloidal Behavior of Alumina
Abrasives. J. Electrochem. Soc. 2006 ,153, G948–G955. [CrossRef]
15. Ihnfeldt, R.; Talbot, J.B. Modeling Material Removal Rates for Copper CMP Using Copper Nanohardness
and Etch Rates. J. Electrochem. Soc. 2008 ,155, H582–H588. [CrossRef]
16. Luo, J.; Dornfeld, D.A. Effects of Abrasive Size Distribution in Chemical Mechanical Planarization: Modeling
and Verification. IEEE Trans. Semicond. Manuf. 2003 ,16, 469.
17. Chen, H.; Guo, D.; Xie, G.; Pan, G. Mechanical model of nanoparticles for material removal in chemical
mechanical polishing process. Friction 2016 ,4, 153–164. [CrossRef]
18. Banerjee, G. Chemical Mechanical Planarization for Cu-Low-k Integration. In Advanced Interconnects for ULSI
Technology ; Baklanov, M.R., Ho, S., Zschech, E., Eds.; Jhon Wiley and Sons Ltd.: West Sussex, UK, 2012.
19. Choi, S.; Doyle, F.M.; Dornfeld, D.A. Material Removal Mechanism during Copper Chemical Mechanical
Planarization Based on Nano-Scale Material Behavior. ECS J. Solid State Sci. Technol. 2017 ,6, P235–P242.
[CrossRef]
20. Basim, G.B.; Moudgil, B.M. Effect of Soft Agglomerates on CMP Slurry Performance. J. Colloid Interface Sci.
2002 ,256, 137–142. [CrossRef]
21. Brahma, N.A. Effects of Slurry Chemistry on the Rate of Agglomeration of Alumina Nanoparticles for
Chemical Mechanical Planarization. Ph.D. Thesis, University of California, San Diego, CA, USA, 2013.
22. Jindal, A.; Babu, S.V . Effect of pH on CMP of Copper and Tantalum. J. Electrochem. Soc. 2004 ,151, G709–G716.
[CrossRef]
23. Du, T.; Desai, V . The pH Effect on Chemical Mechanical Planarization of Copper. Proc. MRS 2003 , 767.
[CrossRef]
24. Wang, Y.G.; Zhang, L.C.; Biddut, A. Chemical effect on the material removal rate in the CMP of silicon
wafers. Wear 2011 ,270, 312–316. [CrossRef]
25. Bielman, M. 1998. Chemical Mechanical Polishing of Tungsten. Master’s Thesis, University of Florida,
Gainesville, FL, USA, 1998.
26. Jiang, L.; He, Y.; Yang, Y.; Luo, J. Chemical Mechanical Polishing of Stainless Steel as Solar Cell Substrate.
ECS J. Solid State Sci. Technol. 2015 ,4, P162–P170. [CrossRef]
27. Kao, M.J.; Hsu, F.C.; Peng, D.X. Synthesis and Characterization of SiO 2Nanoparticles and Their Efficacy in
Chemical Mechanical Polishing Steel Substrate. Adv. Mater. Sci. Eng. 2014 ,2014 , 691967.
28. Deshpande, S.; Kuiry, S.C.; Klimov, M.; Obeng, Y.; Seal, S. Chemical Mechanical Planarization of Copper:
Role of Oxidants and Inhibitors. J. Electrochem. Soc. 2004 ,151, G788–G794. [CrossRef]
29. Eom, D.H.; Kim, I.K.; Han, J.H.; Park, J.G. The Effect of Hydrogen Peroxide in a Citric Acid Based Copper
Slurry on Cu Polishing. J. Electrochem. Soc. 2007 ,154, D38–D44. [CrossRef]
30. Kim, I.K.; Kang, Y.J.; Kim, T.G.; Park, J.G. Effect of Corrosion Inhibitor, Benzotriazole, in Cu Slurry on Cu
Polishing. Jpn. J. Appl. Phys. 2008 ,47, 108–112. [CrossRef]
31. Aksu, S.; Wang, L.; Doyle, F. Effect of Hydrogen Peroxide on Oxidation of Copper in CMP Slurries Containing
Glycine. J. Electrochem. Soc. 2003 ,150, G718. [CrossRef]
32. Du, T.; Vijayakumar, A.; Desai, V . Effect of hydrogen peroxide on oxidation of copper in CMP slurries
containing glycine and Cu ions. Electrochim. Acta 2004 ,49, 4505–4512. [CrossRef]
33. Jiang, L.; He, Y.; Liang, H.; Li, Y.; Luo, J. Effect of Potassium Ions on Tantalum Chemical Mechanical Polishing
in H 2O2-Based Alkaline Slurries. ECS J. Solid State Sci. Technol. 2016 ,5, P100–P111. [CrossRef]
34. Xie, Y.; Bhushan, B. Effects of particle size, polishing pad and contact pressure in free abrasive polishing.
Wear 1996 ,200, 281. [CrossRef]
35. Cook, L.M. Chemical Mechanical Planarization of Microelectronic Materials. J. Non-Cryst. Solids. 1990 ,120,
152. [CrossRef]
36. Brown, N.J.; Baker, P .C.; Maney, R.T. A new model for surface roughness evolution in the Chemical
Mechanical Polishing Process. Proc. SPIE 1981 ,306, 42.

Lubricants 2017 ,5, 15 12 of 12
37. Trogolo, J.A.; Rajan, K. Near surface modification of the silica structure induced by chemical/mechanical
polishing. J. Mater. Sci. 1994 ,29, 4554–4558. [CrossRef]
38. Basim, G.B.; Adler, J.J.; Mahajan, U.; Singh, R.K. Effect of particle size of chemical mechanical polishing
slurries for enhanced polishing with minimal defects. J. Electrochem. Soc. 2000 ,147, 3523–3528. [CrossRef]
39. Sivaram, S.; Bath, M.H.M.; Lee, E.; Leggett, R.; Tolles, R. Chemical Mechanical Polishing of Inter-level
Dielectrics: Models for Removal Rate and Planarity. In MRS Proceedings ; Cambridge University Press:
Cambridge, UK, 1992.
40. Basim, G.B.; Vakarelski, I.U.; Brown, S.; Moudgil, B.M. Strategies for the Optimization of Chemical
Mechanical Polishing (CMP) Slurries. J. Dispers. Sci. Technol. 2003 ,24, 499–515. [CrossRef]
41. Xu, X.; Luo, J.; Lu, X.; Zhang, C. Effect of Nanoparticle Impact on Material Removal. J. Tribol. Trans. 2008 ,51,
718–722. [CrossRef]
42. Scarfo, A.M.; Manno, V .P .; Rogers, C.B.; Anjur, S.; Moinpour, M. In-Situ Measurement of Pressure and
Friction During CMP of Contoured Wafers. J. Electrochem. Soc. 2005 ,152, G477–G481. [CrossRef]
43. Khanna, A.J. 2010. Quantification of Particle Agglomeration during Chemical Mechanical Polishing of
Metals and Dielectrics. Ph.D. Thesis, University of Florida, Gainesville, FL, USA, 2010.
44. Mcintyre, N.S.; Sunder, S.; Shoesmith, D.W.; Stanchell, F.W. Chemical information from XPS—Applications
to the analysis of electrode surfaces. J. Vac. Sci. Technol. 1981 ,18, 714. [CrossRef]
45. Colic, M.; Fuerstenau, D.W. Influence of the dielectric constant of the media on oxide stability in surfactant
solutions. Langmuir 1997 ,13, 6644. [CrossRef]
46. Solomon, M.J.; Saeki, T.; Wan, M.; Scales, P .J.; Boger, D.V .; Usui, H. Effect of adsorbed surfactants on the
rheology of colloidal zirconia suspensions. Langmuir 1999 ,15, 20. [CrossRef]
47. Koopal, L.K.; Goloub, T.; deKaiser, A.; Sidorova, M.P . The effect of cationic surfactants on wetting, colloid
stability and flotation of silica. Colloids Surf. 1999 ,151, 15. [CrossRef]
48. Bremmel, K.E.; Jameson, G.J.; Biggs, S. Adsorption of ionic surfactants in particulate system: Flotation,
stability, and interaction forces. Colloids Surf. 1999 ,146, 755. [CrossRef]
49. Evanko, C.R.; Dzombak, D.A.; Novak, J.W. Influence of surfactant addition on the stability of concentrated
alumina dispersions in water. Colloids Surf. 1996 ,110, 219. [CrossRef]
50. Adler, J.J.; Singh, P .K.; Patist, A.; Rabinovich, Y.I.; Shah, D.O.; Moudgil, B.M. Correlation of particulate
dispersion stability with the strength of self-assembled surfactant films. Langmuir 2000 ,16, 7255–7262.
[CrossRef]
51. Basim, G.B.; Vakarelski, I.U.; Moudgil, B.M. Role of interaction forces in controlling the stability and polishing
performance of CMP slurries. J. Colloid Interface Sci. 2003 ,263, 506–515. [CrossRef]
52. Bozkaya, D.; Müftü, S. Effects of Surface Forces on Material Removal Rate in Chemical Mechanical
Planarization. J. Electrochem. Soc. 2010 ,157, H287–H296. [CrossRef]
©2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).