Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 102 (2015) 582 – 589

The 7th World Congress on Particle Technology (WCPT7)

Chemical Mechanical Polishing Slurry for Amorphous Ge2Sb2Te5

Zhitang Song1,2*, Weili Liu1,2, Liangyong Wang1,2
1
Shanghai Institute of Microsystem and Information Technology, CAS, 865 Changning Road, Shanghai 200050, PR China
2
Shanghai Xianna Electronic Technology Co.Ltd, No.6 Bldg, 285 Lane,Tiangong Road, Jinshan Development Park, Shanghai, 201506, PR China

Abstract

As flash technologies face scaling issues at 32nm and beyond, phase change memory (PCM) emerges as the best performing
candidate for scaled non-volatile memories of next generation. To fabricate PCM device with high volume and low power
consumption, the structure of PCM has evolved from planar structure to confined cells, which pose challenges for chemical
mechanical polishing (CMP) of the phase change material Ge2Sb2Te5 (GST).In this study, we discuss the properties, polishing
mechanisms and slurry formulations for amorphous GST. Static etching characteristics and zeta potentials of GST powders are
first explored to evaluate the chemical properties of GST. Then the polishing mechanisms of GST at pH 2 and pH 11 are
discussed by means of static etching experiments, polishing performance, GST film hardness measurement, GST solubility and
open circuit potential test. Finally, our efforts for GST slurry development are described. By a combination effect of inhibitors, in
which one can form adsorption layer on GST by carbonyl O and the other one can further protect GST surface by forming
hydrogen bond through multiple OH groups, we have overcome the corrosion issue of GST. Meanwhile, by the adsorption of the
inhibitor layers on GST surface and extremely low removal rate of oxide film (<0.05nm/min), we have achieved a self-stop GST
CMP process with over-polish dishing <8nm and oxide loss <6nm for 60nm GST vias. We have also resolved the issue of
particle residue by using our self-formulated post-cleaning solution.
© 2015
© 2014Published
The Authors. Published
by Elsevier byisElsevier
Ltd. This Ltd. article under the CC BY-NC-ND license
an open access
Selection and peer-review under responsibility of Chinese Society of Particuology, Institute of Process Engineering, Chinese
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Selection
Academyand ofpeer-review under responsibility of Chinese Society of Particuology, Institute of Process Engineering, Chinese Academy
Sciences (CAS).
of Sciences (CAS)
Keywords: chemical mechanical polishing, GST, slurry.

* Corresponding author.
E-mail address: ztsong@mail.sim.ac.cn

1877-7058 © 2015 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Selection and peer-review under responsibility of Chinese Society of Particuology, Institute of Process Engineering, Chinese Academy of Sciences (CAS)
doi:10.1016/j.proeng.2015.01.131
 Zhitang Song et al. / Procedia Engineering 102 (2015) 582 – 589 583

1. Introduction

As flash technologies face major scaling issues at 32nm node [1] and with the release of phase change memory
(PCM) product by Samsung, Numonyx [2], PCM is taken as the best performing candidate for scaled non-volatile
memories of next generation due to its advantages of high write/read speed, bit alterability, erase-free operation,
good data retention, high cycling endurance, and scability [3]. PCM technology stores information in a chalcogenide
material that can be rapidly heated and quenched to either amorphous or crystalline phases that strongly differ in
resistivity [4]. Although many chalcogenide materials have been investigated for the application of PCM, GST is
one of the best materials [5]. To reduce the high reset current and increase the storage density, the structure of PCM
has evolved from a planar structure to a confined cell structure [6], in which GST is first deposited and then the
excess GST is removed by CMP stopping on the SiO2 film [7], leaving only GST confined in the nano-hole. The
confined cell structure makes the CMP of GST film as an essential process for avoiding etching damage and
facilitating easy mass production of PCM devices6. At the same time, polishing of a ternary material GST also poses
new challenges for the CMP process like composition control, thickness control of small device region, and
imperfections (e.g., scratches, defects and corrosion) minimization etc. [8,9]
There has been some work exploring slurry component effects on GST CMP. The components well-documented
to influence GST removal include pH [10], abrasives [6], oxidizers [11,12,13], complexing agents [14], surfactants
[15] using either acidic or alkaline slurry. In this study, we discuss the properties, polishing mechanisms and slurry
formulations for amorphous GST.

2. Experimental Set-up

3wt% colloidal silica with 2wt% H2O2 at pH=2 and pH=11 were chosen as two typical slurries for mechanism
research. 2.5cm×3cm GST and oxide wafers were polished using a CMP Tester (CETR CP-4). The polishing
process parameters were set as follows: pad rotation speed 100rpm, wafer rotation speed 100rpm, down force 1psi,
feed rate of the slurry 100ml/min, polishing time 1min for GST and 3mins for oxide. All experiments were
conducted at room temperature. An electron balance (accuracy 0.01mg) was used to measure the weight of GST and
oxide wafers before and after polishing. Then the weight was converted to removal rate (RR) taking the mass density
of GST and oxide as 6.19g/cm3 and 2.2g/cm3, respectively. Zeta potential measurements were done for 0.1wt% GST
powders in deionized water (DIW) at different pH. Particle size was measured for slurry waste collected every 10s
during polishing. Slurry trial on 12 inch tool was carried out on a Mirra tool using IC1010 pad. SEM and TEM were
used to characterize the polished blanket and patterned wafers.

3. Results and Discussions

Properties of GST
Static Etching. Static etching results are shown in Table 1. The static etching rate (SER) for GST in acidic slurry
(pH=2) is only 4.5nm/min, while it is 25.9nm/min for alkaline slurry (pH=11). The SEM images of GST surface
before and after etching are further shown in Figure 1. Before etching, the as-deposited surface is quite smooth.
After etching ether in pH=2 or pH=11 slurry, the GST surface becomes rough and porous, with gaps lying between
the island-like GST. For the etched GST surface in acidic slurry, silica abrasive residues exist due to the electrostatic
attraction force between GST and silica.

Table 1.Static etch and polishing results of GST and oxide using different silica slurries
GST Oxide
GST/Oxide Polishing
Slurry SER RR RMS RR
MRR/SER Selectivity
(nm/min) (nm/min) (nm) (nm/min)
pH=2 4.5 86.3 0.8 19.2 14.5 5.9
pH=11 25.9 206.7 0.8 8.0 10.3 20.1
584 Zhitang Song et al. / Procedia Engineering 102 (2015) 582 – 589

Figure 1. SEM images of GST film surface. (a).As-deposited; (b) After etching at pH=2; (c) After etching at pH=11.

Zeta Potential. Figure 2 shows the zeta potential values of GST powders as a function of pH. As can be seen in
Figure 2, GST surface is positive charged below pH=2.7 and from pH 3.9 to 5.3, while it’s negative charged for the
other pH values. The isoelectric points (IEP) of GST are listed in Table 2, which are 2.7, 3.9, 5.3 and 10.5. Actually,
for pH of 10.5, it’s negative charged. But since it’s an inflection point and close to zero charged, it’s still taken as a
pseudo IEP. Remarkably, quite different from other materials with only one IEP, there are four IEPs for GST, which
are possibly correlated to the specie equilibriums of HGeO3-/GeO2, Sb2O3/SbO3- and Te/TeO2/TeO32-/HTeO4-/TeO42-.
[13]

Figure 2. Zeta potential values of GST powders as a function of pH

Table 2.IEPs of GST powders

No ķ ĸ Ĺ ĺ
IEP 2.7 3.9 5.3 10.5

GST Polishing Mechanism

Polishing. To further compare the difference of different slurries on GST CMP, polishing was done using the two
chosen slurries. As can be seen in Table 1, the RRs of GST and oxide in acidic slurry are 86.3nm/min and
14.5nm/min, respectively, resulting in a GST over oxide polishing selectivity of 5.9. For alkaline slurry, the RRs of
GST and oxide in acidic slurry are 206.7nm/min and 10.3nm/min, respectively, leading to a GST to oxide polishing
selectivity of 20.1. At the same time, the RMSs of the polished GST surfaces are 0.8nm, which are the same both for
acidic and alkaline slurry. In general, alkaline slurry results in a faster RR of GST and higher polishing selectivity of
GST over oxide, which means alkaline slurry can supply higher processing efficiency and lower oxide loss for GST
CMP. However, taking consideration of the SER of GST at pH=2 and pH=11 (also see Table 1), the ratio of RR to
SER for acidic slurry and alkaline slurry are 19.2 and 8.0, respectively. Since a higher RR/SER ratio generally
results in less dishing and corrosion during CMP, acidic slurry seems to be a better choice.
 Zhitang Song et al. / Procedia Engineering 102 (2015) 582 – 589 585

Phase Change Property. To evaluate different slurries on the phase change properties post-CMP, the resistivity of
GST film was measured before and after polishing. Figure 3 shows the resistivity of GST thin-film as a function of
annealing temperature pre-CMP and post-CMP at pH=2 and pH=11. As can been seem from Figure 2, the three
curves are almost parallel with the same inflection point at around 157ć, which means a transformation of GST
from amorphous state to face-centered cubic crystalline state. The only difference for the three curves is the
resistivity value, which might be due to the surface oxidation. In general, the three almost parallel curves indicate
there is no phase property change either using acidic or alkaline slurry post-CMP.

Figure 3. Resistivity of GST thin-film as a function of annealing temperature. (a).As-deposited; (b) After polishing at
pH=2; (c) After polishing at pH=11

Hardness of GST Film. To understand the polishing mechanism of GST CMP, the hardness of GST film before
and after polishing were measured, which is shown in Table 3. As shown in Table 3, the hardness of GST films pre-
CMP, post-CMP at pH=2 and post-CMP at pH=11 are 2.30GPa, 2.31GPa, 2.41GPa, respectively. At the same time,
the modulus of GST films pre-CMP, post-CMP at pH=2 and post-CMP at pH=11 are 49.85GPa, 47.56GPa,
51.35GPa, respectively. In general, the hardness of GST film post-CMP is almost the same as that pre-CMP, though
there is a slight increase of hardness value for GST film post-CMP at pH=2 and a relative larger increase of value for
that post-CMP at pH=11. The hardness result shows the GST oxides are a little harder than GST itself, not being
softer as expected in many other metal CMPs.

Table 3.Hardness of GST films

GST film Hardness(GPa) Modulus(GPa)
Pre-CMP 2.30 49.85
Post-CMP@pH=2 2.31 47.56
Post-CMP@pH=11 2.41 51.35

Particle Size of Slurry. Figure 4 further shows the particle size of acidic slurry and alkaline slurry as a function of
collection time during polishing of GST. Both for the slurries at pH=2 and pH=11, the mean particle size is around
65nm with a variation within f3nm. No larger variation of particle size during polishing demonstrates the particle
size during polishing is kept almost constant without producing larger or abnormal particles.
586 Zhitang Song et al. / Procedia Engineering 102 (2015) 582 – 589

Figure 4. Particle size of acidic slurry and alkaline slurry as a function of collection time during polishing of GST.

OCP. To further understand the behaviors of GST in acidic and alkaline slurry, OCP of GST was conducted. The
OCP values of GST in the slurries at pH=2 with and without H 2O2 are 0.27V and 0.26V, respectively. While for the
situation at pH=11 with and without H2O2, the OCP values are -0.08V and 0.06V, respectively. According to the
Pourbaix diagrams [13] of Ge, Sb,Te, the stable species for GST in the slurries at pH=2 are GeO 2, Sb2O3 and Te both
for the situations with and without H2O2, while they are HGeO3-, SbO3- and TeO32- for those at pH=11.
Reaction Formula. Since the stable species of GST in acidic and alkaline slurry are GeO 2, Sb2O3, Te and HGeO3-,
SbO3-, TeO32-, the overall reaction formulas during GST polishing process are proposed in Figure 5. For the slurry at
pH=2, Ge, Sb are oxidized to GeO2 and Sb2O3 respectively, while Te still remains as metal state without oxidation.
For the slurry at pH=11, Ge, Sb and Te are oxidized to HGeO 3-, SbO3-, TeO32-, respectively. Faster SER and RR of
GST in alkaline region, quite possibly originates from more water soluble species of HGeO3-, SbO3-, TeO32- at
pH=11, other than water insoluble species of GeO 2, Sb2O3 and Te at pH=2.

pH=2
Ge + 2H2O2 GeO2 + 2H2O

2Sb + 3H2O2 Sb2O3 + 3H2O

Te Te

pH=11
Ge + 2H2O2 + OH- HGeO3- + 2H2O

2Sb + 5H2O2 + 2OH- 2SbO3- + 6H2O

Te +2H2O2 + 2OH- TeO32- + 3H2O

Figure 5. Proposed reaction formulas during GST polishing process

Molecular Removal. Due to the almost constant particle size during polishing process both for the slurries at pH=2
and pH=11, it can also hypothesize that GST is removed as molecular instead of as a lump. Otherwise,
agglomeration or abnormal particles can occur in the polishing waste during the polishing process.
Advances for Slurry Formulation
Corrosion. Due to the thermal budget in GST film deposition and the differences of standard reduction potentials
of Ge, Sb, Te, slurry chemicals and underlying metal contacts, corrosion is a big challenge for 12inch GST CMP. By
a combination effect of inhibitors, in which one can form adsorption layer on GST by carbonyl O and the other one
can further protect GST surface by forming hydrogen bond through multiple OH groups, we have overcome this
issue. Figure 6 shows the OM images of GST surface post-CMP processed by conventional slurry and our slurry.
 Zhitang Song et al. / Procedia Engineering 102 (2015) 582 – 589 587

(a) (b)
Figure.6 OM images of GST surface post-CMP by:(a).Conventional slurry; (b) Our slurry

Self-stop polishing. Besides corrosion, dishing is also a big issue for GST slurry. As we know, GST is much
softer than oxide. However, to ensure a large process window post-CMP, GST slurry is required to have low oxide
RR, which inevitably can lead to the occurrence of GST dishing when polishing patterned wafers. To resolve this
issue, a surfactant inhibitor was adopted to be absorbed on GST surface. Once the end-point was achieved, GST in
the via would be protected from polishing by the absorbed surfactant layer, which ensured the polishing process
could be self-stopped.
End-point detection (EPD) trace. Endpoint is also an important factor for process control. To clearly catch the
endpoint, the signals should be different from each other for polishing GST and oxide. Figure 7 shows a typical
endpoint curve for 12 inch GST patterned wafers when using our formulated slurry, which suggests that the endpoint
is caught very well.

Figure.7 Endpoint curve for 10% over-polishing a 12 inch GST patterned wafer

Particle residue. Another possible issue is particle residue post-CMP. It’s originated from the attraction between
the opposite charged GST and particles. By formulating a solution, in which GST and particles are strongly repelled,
particle residues can be thoroughly removed (see Figure 8).
588 Zhitang Song et al. / Procedia Engineering 102 (2015) 582 – 589

(a) (b)
Figure.8 SEM images of GST surface. (a).Post-CMP; (b) Post-CMP and Post-cleaning.

Table 4 further summarizes the performance of our XAN GST 11015 slurry, which shows high GST RR, high
GST/oxide selectivity, low oxide loss, low dishing and long POU life-time.

Table 4.XAN GST11015 slurry performance on 12 inch tool

Oxide Dishing
GST RR GST RR NU GST/Oxide POU life-time
Slurry loss (nm, OP 10% for
(nm/min) (%) Selectivity (day)
(nm) 60nm CD@2psi)
XAN
194.9 1.9 1742:1-4000:1 <6 <8 >7
GST11015

4. Summary

Static etching experiments of GST film showed GST were etched faster in alkaline region than acidic region, both
resulting a porous island-like surface. Zeta potential measurement demonstrated GST has four IEPs. A further
comparison of polishing performance in the two slurries showed alkaline slurry leads to faster removal rate of GST
and higher polishing selectivity of GST over oxide, while acidic slurry results in a higher RR/SER ratio, both not
changing the phase change property of GST post-CMP. A slight harder surface of GST in the slurry than as-
deposited, almost constant particle size during polishing and stable species of GST in acidic and alkaline slurry
being GeO2, Sb2O3, Te and HGeO3-, SbO3-, TeO32-, suggested the polishing of GST was accompanied with oxidation
reactions, removal of GST as molecular, possible softer hydration forming in alkaline region and electrochemical
process between Ge and Te/Sb. Our slurry formulation efforts further show that by a combination effect of inhibitors,
in which one can form adsorption layer on GST by carbonyl O and the other one can further protect GST surface by
forming hydrogen bond through multiple OH groups, we have overcome the corrosion issue of GST. Meanwhile, by
the adsorption of the inhibitor layers on GST surface and extremely low removal rate of oxide film (<0.05nm/min),
we have achieved a self-stop GST CMP process with over-polish dishing <8nm and oxide loss <6nm for 60nm GST
vias. We have also resolved the issue of particle residue by using our self-formulated post-cleaning solution.

Acknowledgements

Acknowledgements and Reference heading should be left justified, bold, with the first letter capitalized but have
no numbers. Text below continues as normal.
 Zhitang Song et al. / Procedia Engineering 102 (2015) 582 – 589 589

References

[1] D.T. Castro, L. Goux, G.A.M. Hurkx, K. Attenborough, R. Delhougne, J. Lisoni, F.J. Jedema, M.A.A. Zandt, R.A.M.Wolters, D.J.
Gravesteijn, M.A. Verheijen, M. Kaiser, R.G. R. Weemaes, D.J. Wouters:”Evidence of Thermo-Electric Thomson Effect and Influence on
the Program Conditions and Cell Optimization in Phase-Change Memory Cells”, IEDM 2007, 1-4244-0439-X, pp 315-318.
[2] http://en.wikipedia.org/wiki/Phase-change_memory
[3] R. Bez:”Chalcogenide PCM: a memory technology for next decade”, IEEE IEDM Tech Dig, 2009, pp 5.1.1-4.
[4] D.L. Kencke, I.V. Karpov, B.G. Johnson, S.J. Lee, D.C. Kau, S.J. Hudgens, J.P. Reifenberg, S.D. Sarvansky, J.Y. Zhang, M.D. Giles, G.
Spadini:”The Role of Interfaces in Damascene Phase-Change Memory”, IEDM 2007, 1-4244-0439-X, pp 323-326.
[5] Z.M. Sun, J. Zhou, R. Ahuja:”Structure of Phase Change Materials for Data Storage”, Phys.Rev.Lett, 96 (2006), pp 055507 1-3.
[6] J.H. Park, H.Cui, S.H. Yi, J.G. Park:”Effect of abrasive material properties on polishing rate selectivity of nitrogen-doped Ge2Sb2Te5 to SiO2
film in chemical mechanical polishing”, J.Mater.Res, 23, 12 (2008), pp 3323-3329.
[7] J.Y. Cho, H. Cui, J.H. Park, S.H.Yi, J.G. Park:”Role of Hydrogen Peroxide in Alkaline Slurry on the Polishing Rate of Polycrystalline
Ge2Sb2Te5 Film in Chemical Mechanical Polishing”, Electrochem.Solid-State Lett, 13, 5 (2010), pp H155-H158.
[8] M. Krishnan, J.W. Nalaskowski, L.M. Cook:”Chemical Mechanical Planarization: Slurry Chemistry, Materials, and Mechanisms”,
Chem.Rev., 110, 1 (2010), pp 178-204.
[9] T.Y. Lee, I.K. Lee, B.H. Choi, Y.S. Park:”Chemical Mechanical Polishing Slurry Composition for Polishing Phase-Change Memory Device
and Method for Polishing Phase-Change Memory Device Using the Same”, US Pat. App. 11967446 (2007).
[10] Z.F. Zhang, W.L. Liu, Z.T. Song:”Influence of pH and abrasive concentration on polishing rate of amorphous Ge2Sb2Te5 film in chemical
mechanical polishng”, J.Vac.Sci.Technol.B., 29, 1 (2011), pp 011020 1-5.
[11] H. Cui, J.Y. Cho, H.S. Hwang, J.H. Lim, J.H. Park, H.S. Park, K. Hong, J.G. Park:” Potassium Permanganate as Oxidizer in Alkaline Slurry
for Chemical Mechanical Planarization of Nitrogen-doped Polycrystalline Ge2Sb2Te5 film”, J.Electrochem.Soc.,157, 11 (2010), pp H1036-H
1041.
[12] M. Zhong, Z.T. Song, B. Liu, S.L. Feng, B. Chen:”Oxidation Addition Effect on Ge2Sb2Te5 Phase Change Film Chemical Mechanical
Polishing”, J.Electrochem.Soc., 155, 11 (2008), pp H929-H931.
[13] H. Cui, J.Y. Cho, J.H. Park, H.S. Park, J.G. Park:”Chemical Mechanical Planarization Mechanism for Nitrogen-Doped Polycrystalline
Ge2Sb2Te5 Film Using Nitric Acidic Slurry Added with Hydrogen Peroxide”, J.Electrochem.Soc., 158, 6 (2011), pp H666-H670.
[14] J.H. Park, J.Y. Cho, H.S. Hwang, U. Paik, J.G. Park:”Effect of Alkaline Agent on Polishing Rate of Nitrogen-Doped Ge2Sb2Te5 Film in
Chemical Mechanical Polishing”,Electrochem.Solid-State Lett., 11, 10 (2008), pp H288-H291.
[15] L.Y. Wang, B. Liu, Z.T. Song, S.L. Feng, Y.H. Xiang, F.X. Zhang:”Acid and Surfactant Effect on Chemican Mechanical Polishing of
Ge2Sb2Te5”, J.Electrochem.Soc.,156, 9 (2009), pp H699-H702.