A Correlation Study of Thermal Stability on Porous Low k - PowerPoint PPT Presentation

1 / 12
About This Presentation
Title:

A Correlation Study of Thermal Stability on Porous Low k

Description:

... for the 430 C annealed films, and 10% shrinkage for the 450 C annealed films. ... significant change in the field breakdown voltage with either anneal condition. ... – PowerPoint PPT presentation

Number of Views:42
Avg rating:3.0/5.0
Slides: 13
Provided by: tflabS
Category:

less

Transcript and Presenter's Notes

Title: A Correlation Study of Thermal Stability on Porous Low k


1
A Correlation Study of Thermal Stability on
Porous Low k
Y.F. Chow, T.H.Foo, L.Shen1, J.S. Pan1, A.Y. Du,
Z.X. Xing, Y.J.Yuan, C.Y. Li, R. Kumar, P.D.
Foo Singapore Science
Mat. Res. Soc. Symp. Proc. 716 (2002) B12.2
??? 31
2
Introduction
  • To reduce the k value, one either incorporates
    atoms and bonds that have a lower polarizability
    or, lowers the density of the film by making the
    film porous
  • The materials that will be discussed in this
    paper are organic spin on low k from Dow
    Chemical non-porous SiLK and porous SiLK. Porous
    SiLK has been chosen for the study of ultra low k
    since it is formed by lowering the density of the
    film without changing the chemical composition
    from non-porous SiLK.
  • In this paper we investigated the correlation of
    thermal cycles with the film properties in terms
    of film shrinkage, refractive index, uniformity,
    IR spectra, electrical breakdown, hardness and
    pore sizes.

3
Experiment
  • Porous SiLK was coated on the silicon substrate,
    hot-plate baked and cured in the furnace at 430C
    or 450C for 30min in a N2 ambient with
    sacrificial compounds removed and leaving pores
    in the film.
  • Non-porous SiLK films were also prepared with the
    same process condition and used as the baseline
    to compare the change in film properties with
    porous SiLK.
  • Tools for measurements
  • Thermawave Optiprobe (model 5240I) ellipsometer
    thickness, uniformity, refractive
  • index and calculated dielectric constant
  • Bio-rad FTIR, XPS determination of changes in
    chemical bonding
  • SSM mercury probe breakdown voltage
  • AFM surface topography
  • JEOL 6720 SEM cross-section
  • Philip TEM pore size measurement and
    correlation with the thickness obtained by Ellip.
  • MTS Nano Indenter XP hardness

4
Results and Discussion
Shrinkage comparison of non-porous SiLK and
porous SiLK versus temperature and thermal cycles
Non-uniformity comparison of non-porous SiLK and
porous SiLK versus thermal cycles and temperatures
5
Refractive Index comparison with various thermal
cycles for non-porous SiLK and porous SiLK
The refractive index of porous SiLK increases
from 1.53 to 1.56 while that of non-porous SiLK
increases from 1.62 to 1.64 after 9 hours of
curing at 430C. After 9 hours of curing at
450C, the refractive index of non-porous SiLK
and porous SiLK become 1.67 and 1.60
respectively. Since the refractive index of
non-porous and porous organic low k films, is
increased with increasing curing temperature, it
implies that higher curing temperature causes
increased densification and it tends to cause the
pores to collapse. Dielectric constant can be
estimated by using formula k n2 when optical
index of refraction measured at 633nm.
6
Dielectric constant comparison versus temperature
and duration for non-porous SiLK and porous SiLK
Non-porous SiLK and porous SiLK, k value
increases by 7 to 8 with 450C curing , but the
k values increases less than 4 with 430C curing.
7
Surface roughness measurements from AFM
The surface roughness was measured by AFM with a
scanned area of 5 um X 5 um. We have found that
the amount of change in the non-porous SiLK
surface roughness decreases slightly after 9
thermal cycles at 430C while there is about 6.9
decrease at 450C. Higher temperature anneals
produce higher surface roughness but the films
become smoother after several thermal cycles. For
example, non-porous SiLK and porous SiLK film
surface roughness decreases about 0.031nm and
0.164nm respectively in roughness at 450C after
9 thermal cycles. There is no significant change
of surface roughness for both non-porous SiLK and
porous SiLK after 430C anneals. The roughness of
porous SiLK is 1.8 to 2.2nm while that of
non-porous SiLK is 0.4 to 0.46nm. The roughness
of porous SiLK is about 5 times higher than
non-porous SiLK. The results show that there is
minimal impact on the surface roughness with
430C anneals.
8
Cross SEM comparison of non-porous SiLK and
porous SiLK
After 9hours of curing we measured 3 shrinkage
for the 430C annealed films, and 10 shrinkage
for the 450C annealed films.
9
SEM and TEM of porous SiLK before and after
9hours cured at 430 and 450C
The SEM and TEM photos of porous SiLK with 9
hours curing at 430C and 450C. The pores
incorporated in the porous SiLK tend to collapse
at 450C. Porous SiLK with 430C cured has less
deformation of pores compare to 450C cured.
Pores and roughness are observed in SEM but less
obvious under TEM photos. It is probably due to
the sample cleaving and operation voltage. TEM
results show pore size before the 1st cycle of
thermal curing are1025 nm after the 9th thermal
cycle for 430C and 450C. We observed the
boundary that could hardly be justified as pores
or the repeated polymer unit boundary. Both SEM
and TEM are not quantitative techniques to
measure the pore size, and pore size
distribution, but can provide information about
the changes in shapes and sizes of pores.
10
Comparison of Field Breakdown Voltage (FBD) for
non-porous SiLK and porous SiLK versus
temperature and thermal cycles.
Field Breakdown Voltage (FBD) was obtained by
applying a voltage sweep from 0 to 350 V, in
100 steps and 1500ms per step with a compliance
current of 2E-3 amperes. The results show no
significant change in either film type after
several thermal cures at either temperature. The
field breakdown voltage of non-porous SiLK and
porous SiLK are 4 to 4.5 MV/cm and 3 to 3.5 MV/cm
respectively.
11
Comparison of hardness for non-porous SiLK and
porous SiLK versus temperature and thermal cycles.
For hardness measurement, a Berkovich indenter
was applied to the surface with a strain rate of
0.05 to 100 nm depth. Measurements of hardness
for non-porous SiLK and porous SiLK shows that
430C cured films have the similar trends. At
430C, nonporous SiLK has a higher hardness of
0.2 GPa than porous SiLK, while porous SiLK cured
at 450C is harder than non-porous SiLK.
12
There is no significant changed in FTIR spectra
for both 430C and 450C. TGA shows the
decomposition temperature as high as 540C. XPS
analysis found 450C has higher ratio of O-O
bonding to C-O bonding than 430C after the 9th
thermal cycle, but no significant change in ratio
of total carbon to oxygen for both non-porous
SiLK and porous SiLK. It indicated that porous
SiLK is suitable for semiconductor manufacturing
in terms of thermal stability.
Conclusions
Porous SiLK cured at 430C is compatible for the
next generation back end of line processing in
the semiconductor industry. The results show more
film properties change with 450C curing in terms
of film shrinkage, uniformity, refractive index
dielectric constant, pores size, hardness, that
with 430C anneals. There is no significant
change in the field breakdown voltage with either
anneal condition. Porous SiLK users should
minimize thermal cycles above 450C to avoid
collapsed pores and increase in the dielectric
constant.
Write a Comment
User Comments (0)
About PowerShow.com