U.S. patent application number 12/684792 was filed with the patent office on 2010-12-23 for apparatus for wetting pretreatment for enhanced damascene metal filling.
Invention is credited to Steven T. Mayer, David W. Porter, Mark J. Willey.
Application Number | 20100320081 12/684792 |
Document ID | / |
Family ID | 43353355 |
Filed Date | 2010-12-23 |
United States Patent
Application |
20100320081 |
Kind Code |
A1 |
Mayer; Steven T. ; et
al. |
December 23, 2010 |
APPARATUS FOR WETTING PRETREATMENT FOR ENHANCED DAMASCENE METAL
FILLING
Abstract
Disclosed are pre-wetting apparatus designs and methods. These
apparatus designs and methods are used to pre-wet a wafer prior to
plating a metal on the surface of the wafer. Disclosed compositions
of the pre-wetting fluid prevent corrosion of a seed layer on the
wafer and also improve the filling rates of features on the
wafer.
Inventors: |
Mayer; Steven T.; (Lake
Oswego, OR) ; Porter; David W.; (Sherwood, OR)
; Willey; Mark J.; (Portland, OR) |
Correspondence
Address: |
Weaver Austin Villeneuve & Sampson LLP - NOVL;Attn.: Novellus Systems,
Inc.
P.O. Box 70250
Oakland
CA
94612-0250
US
|
Family ID: |
43353355 |
Appl. No.: |
12/684792 |
Filed: |
January 8, 2010 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
61218024 |
Jun 17, 2009 |
|
|
|
Current U.S.
Class: |
204/242 ;
134/198; 134/56R |
Current CPC
Class: |
C25D 5/003 20130101;
C25D 5/02 20130101; C25D 5/34 20130101; H01L 21/76861 20130101;
H01L 21/76898 20130101; H01L 21/2885 20130101; C25D 17/001
20130101; C25D 3/38 20130101; C25D 21/04 20130101; C25D 7/123
20130101 |
Class at
Publication: |
204/242 ;
134/198; 134/56.R |
International
Class: |
C25B 9/00 20060101
C25B009/00; B08B 3/00 20060101 B08B003/00 |
Claims
1. An apparatus for pre-wetting a wafer substrate prior to
electrolytically processing the wafer substrate, the apparatus
comprising: a degasser configured for removing one or more
dissolved gases from a pre-wetting fluid prior to pre-wetting; a
process chamber having an inlet for admitting the pre-wetting
fluid, wherein the process chamber is configured for pre-wetting
the wafer substrate with a degassed pre-wetting fluid at a
sub-atmospheric pressure; and a wafer holder positioned within the
process chamber and configured to hold the wafer substrate during
the pre-wetting process.
2. The apparatus of claim 1, wherein the degasser is a membrane
contactor degasser.
3. The apparatus of claim 1, wherein the degasser is configured for
cooling the pre-wetting fluid to a temperature of less than about
20.degree. C.
4. The apparatus of claim 1, wherein the process chamber includes
an outlet adapted for connection with a vacuum pump.
5. The apparatus of claim 1, wherein the apparatus is configured
for delivering the pre-wetting fluid onto the wafer substrate in a
liquid form.
6. The apparatus of claim 5, wherein the apparatus is configured
for immersing the wafer substrate into the pre-wetting fluid for a
period of time.
7. The apparatus of claim 5, wherein the apparatus is configured
for spraying or streaming the pre-wetting fluid onto the wafer
substrate for a period of time.
8. The apparatus of claim 1, wherein the wafer substrate has at
least one recessed feature, wherein the apparatus is configured for
delivering the pre-wetting fluid onto the wafer substrate in a
gaseous form, and wherein the pre-wetting fluid has substantially
all incondensable gasses removed and subsequently condenses to form
a liquid film on the wafer substrate that fills the recessed
feature with pre-wetting fluid.
9. The apparatus of claim 1, wherein the apparatus is configured
for maintaining a pressure of less than about 50 Torr in the
process chamber during pre-wetting of the wafer substrate.
10. The apparatus of claim 1, wherein the apparatus is configured
to initiate introduction of the pre-wetting fluid into the process
chamber and onto the wafer substrate after the pressure in the
process chamber has been reduced to less than about 50 Torr.
11. The apparatus of claim 1, wherein the apparatus is configured
to initiate immersion of the wafer substrate into the pre-wetting
fluid after the pressure in the process chamber has been reduced to
less than about 50 Torr.
12. The apparatus of claim 1, wherein the wafer holder is
configured to position the wafer substrate in a substantially
horizontal orientation.
13. The apparatus of claim 1, wherein the wafer holder is
configured to position the wafer substrate in a substantially
vertical orientation.
14. The apparatus of claim 1, wherein the wafer holder is
configured for rotating the wafer substrate.
15. The apparatus of claim 14, wherein the process chamber is
configured to deliver the pre-wetting fluid onto a rotating wafer
substrate.
16. The apparatus to claim 15, wherein the pre-wetting fluid first
contacts the wafer substrate within about 3 cm of the center of the
wafer substrate.
17. The apparatus of claim 1, wherein the process chamber is
further configured for electroplating a layer of metal on the
pre-wetted wafer substrate.
18. The apparatus of claim 1, wherein the apparatus further
comprises a transfer mechanism configured for transferring the
wafer substrate from the process chamber to an electroplating
apparatus.
19. The apparatus or claim 1, wherein the apparatus is a station in
a module, the module further comprising a station configured for an
anodic process selected from the group consisting of electroetching
or electropolishing.
20. The apparatus of claim 1, wherein the apparatus is a station in
a module, the module further comprising an electroplating station
configured to electroplate the wafer substrate with a metal.
21. The apparatus of claim 20, wherein the metal is copper.
22. The apparatus of claim 20, wherein electroplating station is
configured to immerse the wafer substrate in a degassed plating
electrolyte in the electroplating station.
23. The apparatus of claim 22, wherein the electroplating station
is configured to cathodically polarize the wafer substrate before
immersing the wafer substrate in the degassed plating
electrolyte.
24. The apparatus of claim 20, wherein the module comprises a
transfer mechanism configured for transferring the wafer substrate
from the pre-wetting station to the electroplating station in less
than about 1 minute.
25. The apparatus of claim 1, further comprising a controller
comprising program instructions for: reducing pressure in the
process chamber to a subatmospheric pressure; and subsequently
contacting the wafer substrate with the pre-wetting fluid at a
subatmospheric pressure to form a wetting layer on the wafer
substrate.
26. The apparatus of claim 25, wherein the program instructions
further comprise instructions for: rotating the wafer substrate at
a first rotation rate during delivery of a liquid pre-wetting fluid
onto the wafer substrate at a subatmospheric pressure, wherein the
fluid delivery is performed for between about 10 to 120 seconds;
stopping the delivery of the pre-wetting fluid; after stopping the
delivery of the pre-wetting fluid, rotating the wafer substrate at
a second rotation rate to remove excess surface entrained
pre-wetting fluid from the wafer substrate.
27. The apparatus of claim 26, wherein the program instructions
further comprise instructions for: increasing the pressure in the
process chamber to atmospheric pressure after the delivery of the
pre-wetting fluid is stopped and prior to removal of the excess
surface entrained pre-wetting fluid.
28. The apparatus of claim 26, wherein the program instructions
further comprise instructions for: increasing the pressure in the
process chamber to atmospheric pressure after removal of the excess
surface entrained pre-wetting fluid.
29. The apparatus of claim 26, wherein the first rotation rate is
less than about 300 rpm, and wherein the second rotation rate is at
least about 300 rpm.
30. The apparatus of claim 26, wherein the first rotation rate is
about 100 rpm or less, and wherein the second rotation rate is at
least about 500 rpm.
31. The apparatus of claim 25, wherein the apparatus is configured
for removing excess surface entrained pre-wetting fluid from the
wafer substrate by a method selected from the group consisting of
centrifugal spinning, air-knife drying, and wiping.
32. An apparatus for pre-wetting a wafer substrate prior to
electrolytically processing the wafer substrate, the apparatus
comprising: a process chamber having an inlet for admitting a
pre-wetting fluid, wherein the process chamber is configured for
operating at a higher than atmospheric pressure during or after
pre-wetting, to facilitate the removal of bubbles; and a wafer
holder positioned within the process chamber and configured to hold
the wafer substrate during the pre-wetting process.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims benefit under 35 U.S.C. .sctn.119(e)
to U.S. Provisional Application No. 61/218,024, filed Jun. 17,
2009, which is incorporated herein by reference.
FIELD OF THE INVENTION
[0002] The embodiments disclosed herein relate to pre-wetting
apparatus designs and methods. More specifically, embodiments
relate to pre-wetting apparatus designs and methods for pre-wetting
a semiconductor wafer prior to depositing electrically conductive
materials on the wafer for integrated circuit manufacturing.
BACKGROUND
[0003] Wetting is a property of a liquid/solid interface governed
by adhesive forces between the liquid and solid and cohesive forces
in the liquid. Adhesive forces between the liquid and solid cause
the liquid to spread across the solid surface. Cohesive forces in
the liquid cause the liquid to minimize contact with the solid
surface. The wetting of a solid surface by a liquid is important in
many industrial processes where a liquid interacts with a solid
surface. Electroplating (a cathodic process), including
electroplating in integrated circuit manufacturing, is one such
industrial process. Wetting is also important in anodic processes,
including eletroetching and electropolishing.
[0004] For example, in integrated circuit manufacturing, a
conductive material, such as copper, is often deposited by
electroplating onto a seed layer of metal deposited onto the wafer
surface by a physical vapor deposition (PVD) or a chemical vapor
deposition (CVD) method. Electroplating is a method of choice for
depositing metal into the vias and trenches of the wafer during
damascene and dual damascene processing.
[0005] Damascene processing is a method for forming
interconnections on integrated circuits (ICs). It is especially
suitable for manufacturing integrated circuits, which employ copper
as a conductive material. Damascene processing involves formation
of inlaid metal lines in trenches and vias formed in a dielectric
layer (inter-metal dielectric). In a typical damascene process, a
pattern of trenches and vias is etched in the dielectric layer of a
semiconductor wafer substrate. Typically, a thin layer of an
adherent metal diffusion-barrier film such as tantalum, tantalum
nitride, or a TaN/Ta bilayer is then deposited onto the wafer
surface by a PVD method, followed by deposition of an
electroplate-able metal seed layer (e.g., copper, nickel, cobalt,
ruthenium, etc.) on top of the diffusion-barrier layer. The
trenches and vias are then electrofilled with copper, and the
surface of the wafer is planarized.
SUMMARY
[0006] Disclosed are pre-wetting apparatus designs and methods.
[0007] In one embodiment, an apparatus for pre-wetting a wafer
substrate prior to electrolytically processing the wafer substrate
is disclosed. The apparatus includes a degasser configured for
removing one or more dissolved gases from a pre-wetting fluid prior
to pre-wetting and a process chamber having an inlet for admitting
the pre-wetting fluid. The process chamber is configured for
pre-wetting the wafer substrate with a degassed pre-wetting fluid
at a sub-atmospheric pressure. Inside the process chamber is a
wafer holder positioned and configured to hold the wafer substrate
during the pre-wetting process.
[0008] In another embodiment, an apparatus for pre-wetting a wafer
substrate prior to electrolytically processing the wafer substrate
is disclosed. The apparatus includes a process chamber having an
inlet for admitting a pre-wetting fluid. The process chamber is
configured for operating at a higher than atmospheric pressure
during or after pre-wetting, to facilitate the removal of bubbles.
Inside the process chamber is a wafer holder positioned and
configured to hold the wafer substrate during the pre-wetting
process.
BRIEF DESCRIPTION OF THE DRAWINGS
[0009] FIG. 1 depicts a plot of bubble dissolution time versus
feature size.
[0010] FIG. 2 depicts a plot of bubble dissolution time versus the
dissolved gas pressure.
[0011] FIG. 3 depicts a schematic layout of one embodiment of a
pre-wetting apparatus.
[0012] FIG. 4 depicts an embodiment of a pre-wetting chamber.
[0013] FIG. 5 depicts an isometric view of an embodiment of a
pre-wetting chamber.
[0014] FIG. 6 depicts an embodiment of a pre-wetting chamber
configured for a condensation pre-wetting process.
[0015] FIG. 7 depicts an embodiment of a pre-wetting chamber
configured for an immersion pre-wetting process.
[0016] FIG. 8 depicts another embodiment of a pre-wetting chamber
configured for an immersion pre-wetting process.
[0017] FIG. 9 depicts an embodiment of an apparatus in which the
pre-wetting process is performed in a plating cell.
[0018] FIG. 10 depicts an embodiment of an electroplating
system.
[0019] FIGS. 11a and 11b are flow diagrams for embodiments of a
pre-wetting process.
[0020] FIG. 12 is a flow diagram for an embodiment of an
electroplating process for electroplating a layer of metal on a
wafer substrate.
[0021] FIG. 13 depicts a wafer substrate with a feature filled with
pre-wetting fluid.
DETAILED DESCRIPTION
[0022] Reference will now be made to specific embodiments. Examples
of the specific embodiments are illustrated in the accompanying
drawings. While the invention will be described in conjunction with
these specific embodiments, it will be understood that it is not
intended to limit the invention to such specific embodiments. On
the contrary, it is intended to cover alternatives, modifications,
and equivalents as may be included within the spirit and scope of
the invention as defined by the appended claims. In the following
description, numerous specific details are set forth in order to
provide a thorough understanding of the present invention. The
present invention may be practiced without some or all of these
specific details. In other instances, well known process operations
have not been described in detail in order not to unnecessarily
obscure the present invention.
[0023] Disclosed herein are apparatus designs and methods for wafer
pre-wetting, for modifying the conditions of wafer entry and wafer
processing during plating, and pre-wetting fluid compositions. The
pre-wetting process, in accordance with embodiments provided
herein, can be performed in the electroplating chamber or in a
separate pre-wetting station of a module which includes a
pre-wetting station and an electroplating station. In some
embodiments pre-wetting and electroplating are performed in
separate apparatuses.
[0024] The substrate typically is a semiconductor wafer which has a
layer of conductive material residing thereon (e.g., a seed layer
comprising copper or copper alloy). During electroplating,
electrical connections are made to the conductive layer and the
wafer substrate is negatively biased, thereby serving as a cathode.
The wafer is contacted with a plating solution containing a metal
salt (e.g., copper sulfate, a copper alkylsulfonate, or a mixture
of salts), which is reduced at the wafer cathode resulting in metal
deposition on the wafer. In many embodiments, the substrate
contains one or more recessed features (e.g., vias and/or
trenches), which need to be filled by the electroplating process.
The plating solution, in addition to metal salts, may also contain
an acid, and typically contains one or more additives such as
halides (e.g., chloride, bromide, etc.), accelerators, levelers,
and suppressors, used to modulate electrodeposition rates on
various surfaces of the substrate.
[0025] The disclosed processes and associated apparatus designs are
particularly applicable and necessary for electrofilling wider
(e.g., typically greater than 5 .mu.m) and deeper (e.g., typically
greater than 10 .mu.m) damascene structures (vias), such as those
commonly found in emerging copper through silicon via (TSV)
electrofill structures. Through silicon via structures are further
described in U.S. patent application Ser. No. 12/193,644, filed
Aug. 18, 2008, which is herein incorporated by reference. Gas
bubbles, trapped or otherwise residing on the surface or within a
feature, will interfere with the field and feature plating process
either by blocking the feature surface with the non-conducting gas,
or by creating an impediment to the free passage of current. The
disclosed processes and associated apparatus designs enable
void-free copper electrofilling.
[0026] Electroplating and electrofill of TSV interconnections
present a number of challenges. These include long plating times
due to very large and/or deep structures, and formation of side
wall voids due to seed layer corrosion reactions with plating
electrolyte solutions and due to insufficient coverage of lower
sidewalls by PVD-deposited seed layer. Further, it is important to
ensure that the interior of all recessed features are filled with
liquid and that there are no trapped gasses inside the features
that prevents plating therein. It is also advantageous to
simultaneously maintain strong wall and field plating
growth-suppression while removing plating resistance selectively at
the bottom of the feature.
[0027] The pre-wetting apparatus designs and methods described
herein are generally described with respect to electroplating (a
cathodic process) a metal, specifically copper. However, the
pre-wetting apparatus designs and methods described herein are
generally applicable to all electrolytic processes, including
eletroetching and electropolishing, both of which are anodic
processes.
[0028] Methods for forming liquid-filled bubble-free recessed
features that are needed for the plating process are described.
Further, compositions of pre-wetting fluids which minimize seed
layer corrosion and simultaneously increase plating rates are
described.
INTRODUCTION
[0029] The concentration of dissolved gas at the bubble interface
with the liquid is related to internal bubble pressure by Henry's
law, one form of which can be expressed as:
C.sub.i=x.sub.iH.sub.iP.sub.i (1)
where the subscript i depicts "inside" the bubble, C, is the
concentration of a component of dissolved gas molecules in the
liquid phase at the bubble interface (e.g., nitrogen, oxygen, etc.,
each in moles/l), x.sub.i is the mole fraction of that component in
the gas phase inside the bubble itself, H.sub.i is the Henry's law
constant, and P.sub.i is the pressure inside the bubble. This
equation can be written for each molecular component of the gas in
a mixture of gases (e.g., one for oxygen, one for nitrogen, etc.).
There is a similar expression for the concentration of dissolved
gas in the bulk solution, where the subscript b is used to indicate
the solution "bulk", for example, with P equal to P.sub.b,
signifying the gas phase pressure which would be in equilibrium
with a concentration of the species in the bulk, C.sub.b. Ignoring
2D and 3D dispersion effects, and assuming that the diffusion of
gas molecules from within the bubble gas phase to the bubble/liquid
surface is not rate limiting (so that an equilibrium condition
between the dissolved gas at the bubble interface and the
concentration of gas inside the bubble is maintained), a useful
approximation for the rate of gas dissolution from the bubble
trapped inside a feature can be obtained, expressed as:
R=dV/dt=DH(x.sub.iP.sub.i-x.sub.bP.sub.b)/h (2)
where V is the bubble gas volume, t is time, D is the diffusion
coefficient of the gas in the solution, h is the distance from the
top of the trapped bubble to the edge of the boundary layer
thickness, which resides at a distance .delta. above the upper
wafer plane, and the subscript b corresponds to conditions in the
bulk of the solution at the diffusion boundary layer interface. For
a given chemical system at a fixed temperature (constant Henry's
law constant and diffusion coefficient), two factors can lead to
relatively rapid bubble dissolution: 1) a large concentration
difference/driving force (x.sub.iP.sub.i-x.sub.bP.sub.b); and 2) a
short diffusion distance h.
[0030] If the value of the driving force term
H(x.sub.iP.sub.1-x.sub.bP.sub.b) is zero, the rate of dissolution
is zero. Generally, this term is very small. Since the gas within
the bubble typically comes from air inside a via in a wafer prior
to the pre-wetting process, and the liquid is typically saturated
with the same air prior to the pre-wetting process, the mole
fraction at the bubble interface and in the bulk solution will
initially be the same as that of air (e.g., x=0.21 for oxygen, both
in the bubble and in the bulk solution). Therefore, for this
situation and in general (i.e., unless other mechanisms are
employed to enhance bubble dissolution), it is primarily the
natural capillary difference in pressure within the bubble versus
outside the bubble that results in bubble dissolution.
[0031] Trapped gas held within a small damascene feature (e.g., a
via) can exhibit a very large internal pressure, due to strong
internal capillary forces. The total internal capillary pressure is
proportional to the contact angle and the surface tension, and
inversely proportional to radius of curvature of the bubble,
P.sub.i=P.sub.ext+.delta. cos .theta./r (3)
where P.sub.i is the total internal pressure within the bubble,
P.sub.ext is the external pressure of the fluid (typically about 1
atmosphere), a is the liquid/gas surface tension, .theta. is the
solid/liquid/gas contact angle, and r is the radius of curvature.
Note that the radius of curvature r cannot significantly differ
from the feature width, so one can often substitute the radius of
the via as an approximation for a bubble's radius of curvature.
With a small via, the total internal pressure (and hence the
partial pressure of each component) can become very large,
exceeding several atmospheres or more. These large internal
pressures then drive a non-equilibrium condition with respect to
the bulk of the solution, and the bubble interface becomes
significantly supersaturated with respect the amount of dissolved
gas in the bulk of the solution at the same pressure (i.e., the
amount of dissolved gas at the bubble's interface exceeds the gas
solubility in the liquid). This satisfies one of the conditions for
rapid bubble dissolution. For small vias, the small diffusion
distance, "h", also aids in a rapid dissolution rate.
[0032] In contrast, large vias with larger radius bubbles have both
small excess internal pressure and much larger diffusion distances.
Calculation/modeling of the time for complete bubble dissolution
for various conditions (i.e., partial pressure of dissolved gas,
rotation rate of wafer) as a function of the via depth for vias of
a 3:1 aspect ratio (depth to width) with the via initially 50%
filled of gas at atmospheric pressure, are shown in FIG. 1. For all
of processes shown in FIG. 1, .sigma.=60 dyne/cm (e.g., a value for
water), D=1.9E-5 cm.sup.2/sec (e.g., a value for air in water),
T=20.degree. C., and Vi=50% of a via.
[0033] Vi is the initial volume of the bubble, under 1 atmosphere
of pressure (i.e., only 50% of each via is filled with a bubble for
generating these plots). For the P.sub.ext=0.2 cases, the pressure
on the fluid is still one atmosphere, but the partial pressure of
dissolved gas in the bulk of the liquid is only equivalent to that
in equilibrium with 0.2 atmospheres of gas pressure. This condition
could be achieved, for example, by having a trapped bubble form by
flooding the surface with a degassed fluid with a pressure of 0.2
atmospheres while the pressure of gas over the fluid is 1
atmosphere. For the P.sub.ext=3 cases, the amount of dissolved gas
in the liquid is equal to that in equilibrium with 1 atmosphere of
pressure, but the pressure on the liquid and bubble is equal to 3
atmospheres of pressure. This condition could be achieved, for
example, by having a trapped bubble form by flooding the surface
with atmospheric-saturated liquid, and then applying an external
pressure over that via/liquid/wafer of 3 atmospheres. In this case,
the bubble immediately shrinks in size to 1/3 its original
volume.
[0034] Comparing curves A and F (non-degassed pre-wetting fluid,
with the amount of gas equal to that in equilibrium with 1
atmosphere of air) to curves B and C (a pre-wetting fluid degassed
to a partial pressure equal to 0.2 of an atmosphere), the degassed
solution cases have a lower bubble dissolution time. Curves F and C
are similar by comparison, but the boundary layer thickness and
dissolution time are greater because the wafer was rotated at a
slower rate (12 rpm versus 90 rpm in A and B).
[0035] Curves A and F of FIG. 1 show that the time of dissolution
of a bubble inside a via where the solution is saturated with air
changes more than 5 orders of magnitude between a via 0.2 .mu.m in
size versus 50 .mu.m in size. In small, submicron features, bubbles
are unstable and dissolve quickly, but in larger features, bubbles
will persist for very long times. For example, calculations
indicate that a relatively large front end of the line structure 1
.sub.[tm in diameter and 4 .mu.m in depth completely filled with
gas will have that gas completely dissolved in less than a 4
seconds. In contrast, a 0.25 .mu.m feature, 1 .mu.m deep, is so
unstable that it would dissolve in less than 0.4 seconds, and
smaller structures essentially dissolve instantaneously. However,
both of the favorable factors (i.e., high internal pressure and
short diffusion distances) are missing in large TSV scale
structures. In contrast, calculations show that it could take over
2 hours for a 25 .mu.m wide, 100 .mu.m deep feature to dissolve.
Even if this feature were only filled 10% with gas at its bottom,
it would still take 20 minutes or more for that gas to be
removed.
[0036] Removing gas from the pre-wetting fluid reduces the time to
dissolve a trapped bubble. In this case, the right term of the
driving force (x.sub.bP.sub.b in equation 2) is diminished by
stripping the gas from the solution, for example by reducing the
partial pressure of the gas exposed to the pre-wetting fluid in a
degassing unit under partial vacuum (i.e., by driving down the
magnitude of this product in the gas side of the degassing unit,
gas is driven out of the fluid). The gas in the trapped bubble is
at approximately 1 atmosphere of pressure or more (when there is
significant capillary pressure). At the bubble interface, the
concentration of gas will be at near equilibrium with that of the
same 1 atmosphere or more or pressure, but in the solution, the
concentration as a whole, due to the degassing operation, is at a
much lower concentration. This creates a significant concentration
driving force and degree of sub-saturation of the gas in the
solution (chemical "capacity") to enable rapid dissolving the
bubble.
[0037] This procedure at first may appear to be appealing, but may
suffer from two limitations. First, for large deep vias, the
diffusion distance for the gas may still be a significant limiting
factor. Second, since the amount of gas in the solution can never
be less than zero, the magnitude of the driving force for
dissolution is limited to be no more than approximately Hx.sub.iP
(P=1 atmosphere). Comparing curves B and C to A and F of FIG. 1,
the rate of dissolution of large features (e.g., 50 .mu.m)
decreases by a more than one order of magnitude over the
non-degassed gas, but the dissolution time is generally still
unacceptably long (e.g., at least 5-10 minutes). Note that the rate
of dissolution of smaller features are not significantly affected
by the use of a degassed solution, because the process is dominated
by the large excess internal bubble pressure compared to the
increase in 1 atmosphere of dissolve gas driving force.
[0038] FIG. 2 shows the bubble dissolution time for various feature
dimensions (at 90 rpm rotation, 60 dyne/cm), where the amount of
dissolved gas is the independent parameter. In each case, the
bubble is initially 50% of the via size, and there is an external
pressure of 1 atmosphere on the fluid and bubble, though the
dissolved partial pressure changes as a function of the x axis. To
clarify, in FIG. 2, the concentration of dissolved gas corresponds
to the dissolved gas pressure on the x-axis, related by Henry's
law. These partial pressure would be obtained, for example, by
degassing the contacting fluid to the extent of the x-axis
parameter. Bubbles in smaller, less deep features dissolve more
quickly, the rate aided by the large internal capillary pressure.
Again reducing the partial pressure for smaller features has less
of a relative effect on reducing the dissolution time. For larger
features (e.g., 50 .mu.m.times.150 .mu.m), there is diminishing
benefit of reducing the partial pressure of dissolved gas below 30
to 40% of the saturated condition. In all but the smallest and most
shallow features, the time of dissolution exceeds 100 seconds.
Feature depth is a significant limiting factor in all cases, with
deep features having long dissolution times.
Apparatus
[0039] Generally, the apparatus designs and methods described
herein avoid the formation of a bubble within a recessed feature
(e.g., a via) on a wafer substrate by first removing gas, primarily
all non-condensable gasses (e.g., nitrogen and oxygen), from within
the feature before pre-wetting the surface and feature with a
fluid. To accomplish this, the wafer with the recessed features is
placed in a vessel that is suitable for both holding the wafer and
removing the gas from the wafer surfaces (e.g., a vacuum vessel).
In addition to the vessel itself, a mechanism for removing the gas
(e.g., a line connected to a vacuum source such as a pump) and a
mechanism for depositing liquid onto the surface while the vacuum
conditions are maintained, are needed.
[0040] Described herein are various apparatus designs for
pre-wetting a wafer prior to, or within a short time after, the
initiation of a plating process, where bubbles and gas that might
otherwise be trapped within features recessed in a surface are
avoided. Embodiments of a pre-wetting apparatus include various
elements. Typically, a pre-wetting apparatus includes a pre-wetting
fluid storage and return tank, including liquid mixing devices and
liquid level controllers and sensors. In some embodiments, the
apparatus includes a pre-wetting fluid degassing flow loop. Such a
degassing flow loop includes a circulating pump, routing/diverting
valves, a liquid degassing element, and a connection between the
liquid degassing element and the system vacuum pump (used to pump
down and apply vacuum to various liquid degassing elements on the
tool and the pre-wetting chamber), in some embodiments. A
pre-wetting apparatus also includes a pre-wetting chamber. A
pre-wetting chamber, in some embodiments, includes a two position
(open/closed) vacuum wafer access door or lid for access to the
chamber and a combined door or lid and splash shield that prevents
liquid from hitting and subsequently dropping from the upper walls
or door onto the wafer surface. In some embodiments, inside the
chamber is a wafer holder for supporting and rotating the wafer
within the chamber. In some embodiments, the chamber includes an
air-dome chamber-heater, used to prevent liquid condensation on the
walls of the chamber that would otherwise reside above the wafer
and the vacuum wafer access door and potentially drip onto the
wafer. Pre-wetting chambers typically include an inlet port for
pre-wetting fluid to enter the chamber and to direct pre-wetting
fluid to land on the upper surface of the rotating wafer and an
inlet line and chamber port for drawing and releasing vacuum on the
chamber, the inlet line containing a particle filtration device and
the inlet port containing a flow diffuser configured to disperse
incoming gas flow and minimize chamber flow turbulence. In some
embodiments, the chamber includes liquid level sensors for
monitoring an empty/ready and overflow/over-full condition.
Pre-wetting chambers also typically include a drain for removing
liquid from the chamber and directing the drained fluid back to the
storage tank.
[0041] Embodiments described herein overcome the deleterious
effects of trapped bubbles, particular those bubbles which can be
formed in larger vias or trenches in a wafer, by: (1) avoiding
trapping gas in the via during pre-wetting altogether by removing
substantially all of the atmospheric non-condensable gasses above
the wafer and from within the via, and then pre-wetting the wafer
with pre-wetting fluid; and/or (2) significantly increasing the
rate a bubble will dissolve by applying a large external pressure
on the fluid, thereby driving the bubble to dissolve in the fluid
by creating a large supersaturated condition at the bubble
interface. In addition to these pre-treating and pre-plating
measures, in some embodiments plating is performed in a plating
solution that is maintained in a degassed state, and in other
embodiments, the plating solution is degassed in the line just
prior to being exposed to the wafer surface.
[0042] In some embodiments it is possible to perform pre-wetting
within an electroplating cell, where the pre-wetting fluid has the
same composition as the plating solution. However, for a variety of
reasons, including the hardware complexity of combining plating
processes with vacuum processes, pre-wetting (including vacuum
feature-backfilled pre-wetting) is often performed in a different
cell, sub-cell, or module than the plating cell. When pre-wetting
under vacuum is performed in a distinctly different area of the
plating cell, or in a distinctly separate module from the plating
cell, rather than in the plating solution, the composition of the
pre-wetting fluid can be selected. The pre-wetting fluid may have
the same, or very similar, composition as that subsequently used
for plating the wafer. The pre-wetting fluid may include all the
elements of the plating bath (e.g., the same solvent(s) and same
dissolved metal ions, acids, cation, additives and halides, at the
same or very similar concentrations as in the plating solution).
Such a pre-wetting fluid may work in some embodiments.
Alternatively, in other embodiments, a pre-wetting fluid that is
very different from the plating solution may be used. For example,
in some embodiments, a pre-wetting fluid of 1) water, 2) a fluid
with a substantially higher metal ion concentration than that of
the plating solution, 3) a fluid having either a lower, different
combination of, or no dissolved halides, 4) a fluid substantially
free of one, a few, or all of the plating additives, or 5)
water-miscible solvents may be used as pre-wetting fluids. Such
pre-wetting fluids are further described herein.
[0043] A number of factors should be considered when selecting a
pre-wetting fluid composition, including the possibilities of: a)
corroding the metal layers on a wafer substrate before initiation
of plating; b) inhibiting the plating process (i.e., slowing down
or inhibiting altogether the feature metal-filling process); c) the
loss of pre-wetting fluid to subsequent pre-wetting fluid reuse;
and d) altering (by adding, diluting, or concentrating) various
critical species concentrations within the plating bath over time.
The latter process may alter the metal ion concentrations, halide
concentrations, organic additives, etc., in the plating bath. These
effects can be quite substantial. Furthermore, when using a
pre-wetting fluid of a different composition than the plating bath,
performing the pre-wetting process in the same module without
enabling suitable mechanisms of removing and recovering excess
entrained pre-wetting fluid that would be added to the plating
solution would generally require mechanisms for mitigating,
monitoring and/or otherwise correcting for plating solution
modification over time. On the other hand, the use of hardware and
a process wherein the pre-wetting operation is performed in a
separate treatment station, module, vessel, or sub-vessel of the
plating cell that allows for separation and recovery of this fluid
may be advantageous because it can avoid such issues. With this
background, and in order to simplify the description of core
concepts of embodiments, many embodiments are described hereafter
in the context of a separate pre-wetting "station" and a separate
"plating station", with the wafer being transferred from the former
to the latter. However, while perhaps favorable in some
circumstances (e.g., for avoiding mixing of unlike liquids or for
other reasons), the aspect of embodiments related to the particular
choice of pre-wetting materials, the general fluid, and plating
processing sequences are not intended to be so limited.
[0044] FIG. 3 depicts a schematic layout of one embodiment of a
pre-wetting apparatus (i.e., chamber 301 and associated hardware).
The chamber 301 is connected to vacuum pump 303 though an outlet in
the chamber and through a three-way valve connection 305. On the
other side of the three-way valve is degassing loop 306 that
includes a pre-wetting fluid tank 307, a degassing device 309, and
a pump 311 for circulating the pre-wetting fluid around the
degassing loop. In another embodiment, the pre-wetting fluid feed
line and the vacuum line are not connected except at the chamber,
and each has its own valve (i.e., there is no three-way valve). In
an alternative embodiment, the chamber has an inlet for admitting
pre-wetting fluid and an outlet adapted for connection with a
vacuum pump. The position of the pump 311 can after the degassing
element, if it is desired to drive the fluid into the chamber by
the pump rather than by being sucked into the chamber by a pressure
differential between the pre-wetting fluid tank 307 and the chamber
301.
[0045] In some embodiments, the area in the pre-wetting fluid
holding tank 307 gasses by applying a vacuum to the holding tank
using a vacuum pump (not shown) so that a minimum amount of
dissolved gas is achieved. The rate or removal of the gas from the
pre-wetting fluid can also be increased by increasing the exposed
surface of the fluid to the vacuum, for example, by having the
fluid re-enter the chamber from the circulation loop in a spray or
through a spray column. In the embodiment of the system shown in
FIG. 3, pre-wetting fluid is circulated though the degassing device
309 (e.g., in some embodiments, a membrane contact degasser) for
removing one or more dissolved gases (e.g., both O.sub.2 and
N.sub.2) from the pre-wetting fluid prior to pre-wetting. Examples
of commercially available degassing devices include the
Liquid-Cel.TM. from Membrana of Charlotte, N.C. and the pHasor.TM.
from Entegris of Chaska, Minn. The amount of dissolved gas can be
monitored with an appropriate meter (e.g., a commercial dissolved
oxygen meter (not shown)). The removal of the dissolved gas prior
to the pre-wetting fluid entering chamber 301 can improve the
pre-wetting process, as is described herein. After degassing the
pre-wetting fluid, optionally, the valve 315 between the vacuum
side of the degassing chamber 309 and the vacuum pump 303 is closed
(this prevents gas initially in the chamber from becoming dissolved
in the degassed pre-wetting fluid; in some embodiments, separate
pumps can be used for these two functions).
[0046] Unlike conditions that exist when using an apparatus
similarly configured to that of FIG. 3, if the pre-wetting fluid is
not degassed prior to exposing it to a wafer under vacuum,
dissolved gas from the fluid can be released from the fluid as it
enters the chamber. This results in bubbles forming inside the
vias. While not wanting to be limited by a particular model or
theory, a via bottom is a location of negative curvature, and it is
believed that this location is a particularly susceptible to
nucleating a bubble and releasing gas from the pre-wetting fluid.
If this occurs, bubbles will be formed from the pre-wetting fluid
containing dissolved gas because it is supersaturated with gas
under the pre-wetting conditions (e.g., vacuum in the chamber). The
bubbles so formed can remain there after the pre-wetting process,
which in turn can inhibit plating there and lead to associated
defects. Therefore, in some embodiments (including the embodiment
shown in FIG. 3), the pre-wetting fluid used in the pre-wetting
process is a degassed pre-wetting fluid. In some embodiments, the
degassed pre-wetting fluid may be a plating solution, and the
pre-wetting methods described herein may be preformed in the same
chamber as the plating chamber itself. If a separate pre-wetting
chamber and apparatus are employed, but the pre-wetting fluid is
not degassed, then intermittent and unreliable filling results may
be observed. For example, when vias on a wafer are filled with
pre-wetting fluid (with the wafer under vacuum) without first
degassing the pre-wetting fluid, it has been found that
approximately 15% of the vias still have air bubbles in them (as
indicated by the same percentage having post-plating voids,
indicative of trapped gas bubble therein). Thus, in some
embodiments, it is important to perform pre-wetting under vacuum
(i.e., at a subatmospheric pressure) and with a degassed fluid.
[0047] In contrast, the use of a degassed pre-wetting fluid in
combination with a pre-wetting operation under vacuum (i.e., at
subatmospheric pressure) leads, in some embodiments, to
significantly fewer feature voids than when pre-wetting under
vacuum alone is employed. In specific embodiments that give good
protection against forming voids, a combination of a degassed
pre-wetting fluid with pre-wetting under vacuum is further combined
with plating in a plating solution that is degassed. The plating
solution may be degassed only in the initial stages of plating
(e.g., for only about the first 10 minutes of the plating process),
or remain degassed for the entire plating process (e.g., if the
plating time is greater). Experiments performed under these
conditions produced vias that were void free.
[0048] Returning to FIG. 3, after the pressure in chamber 301 has
reached a low value (i.e., a subatmospheric pressure), the
three-way valve 305 to the vacuum pump location is switched to
connect to the line from the degassing loop 306, and the three-way
valve 313 of the degasser loop is set to allow fluid to be directed
into the vacuum chamber 301. In some embodiments, the
subatmospheric pressure is about equal to that of the boiling
pressure of the pre-wetting fluid at the temperature of operations,
which for water at ambient temperature is about 20 torr. In other
embodiments, the subatmospheric pressure is about 50 torr. In
further embodiments, the pressure of 50 torr is maintained while
pre-wetting the wafer substrate. In alternative embodiments, the
pre-wetting system is configured to initiate introduction of the
pre-wetting fluid into the chamber and onto the wafer substrate
after the pressure in the chamber has been reduced to less than
about 50 Torr. In embodiments where the pre-wetting fluid tank 307
is at atmospheric pressure, liquid is drawn into the chamber 301 by
the pressure differential between the vacuum chamber and the
pre-wetting fluid tank.
[0049] The pre-wetting fluid wets the device side of the wafer
surface of a wafer in the chamber 301. Needle valve 317 can be used
to meter the flow of the pre-wetting fluid into chamber 301.
Embodiments of chamber 301 are described herein. Chamber 301, in
some embodiments, is a pressure chamber configured to apply an
external pressure to increase the rate of bubble dissolution, as
described herein. In further embodiments of a pre-wetting
apparatus, the pre-wetting apparatus includes a transfer mechanism
configured for transferring the wafer substrate from the
pre-wetting chamber to an electroplating apparatus.
[0050] In some embodiments, the pre-wetting fluid is cooled prior
to injection into the pre-wetting chamber (e.g., 0.degree. C. for
water, or -10.degree. C. for suitable electrolytes). In other
embodiments, the degasser is configured for cooling the pre-wetting
fluid to a temperature of less than about 20.degree. C. Other
examples of methods for cooling the pre-wetting fluid include
passing the fluid over a heat exchanger in the pre-wetting fluid
holding tank or though a in-line cooler (both not shown in FIG. 3).
Cooling the pre-wetting fluid reduces the partial vapor pressure of
the solvent of the pre-wetting fluid, which allows for greater
applied vacuum, for example, to the degassing device. Lowering the
temperature of the pre-wetting fluid can also be effective in
increasing both the surface tension and viscosity of the
pre-wetting fluid, which tends to make the phenomena of degassing
device "blow through" or "weeping" less prevalent. Weeping can be a
particularly difficult problem when dealing with salt containing
pre-wetting fluids, because weeping salt laden fluids tend to dry
and destroy the pores of the degassing device. Using lower
temperature fluids reduces the tendency of salt laden electrolyte
to evaporate and flow, avoiding this known source of degassing
device failure. For example, the vapor pressure of water (with a
small amount of salt) is about 2.7 to at -10.degree. C. versus 17.5
torr at 20.degree. C. and 32 torr at 30.degree. C. With a 20 torr
vacuum (yielding about 0.5 ppm dissolved atmospheric gas) applied
to a degassing device, a 30.degree. C. pre-wetting fluid will
literally boil and leave salts around the pores of the degassing
device, and a 20.degree. C. pre-wetting fluid will evaporate
rapidly. Very little degassing device salting occurs when using a
-10.degree. C. pre-wetting fluid, however. Thus, in general, more
dissolved gas can be removed more efficiently from a degassing
device with a lower temperature fluid. In some embodiments, the
pre-wetting fluid is cooled to a temperature of less than
20.degree. C., for example 0.degree. C. or less, while it is
degassed and before it enters the processing chamber. Also,
reducing the temperature of the pre-wetting fluid reduces the rate
of metal corrosion in the pre-wetting system.
[0051] In some embodiments of a pre-wetting apparatus, the surface
of a wafer is wetted with a pre-wetting fluid followed by the
application of an external pressure to the fluid. The wafer surface
is first contacted with the fluid using an appropriate mechanism,
usually immersing the wafer in a pre-wetting fluid (described
herein). In these embodiments, the pre-wetting chamber includes an
inlet for admitting pre-wetting fluid and the chamber is configured
for operating at a higher than atmospheric pressure during or after
pre-wetting. The application of an external pressure to the fluid
facilitates the removal of bubbles. In some embodiments, the
pre-wetting fluid is preconditioned so that it is substantially
free not just of oxygen (e.g., to minimize corrosion of the metal
on the wafer), but of all dissolved non-condensable gasses, such as
nitrogen and carbon dioxide, prior to the pre-wetting of the
surface, to accelerate the dissolution rate of any trapped bubble
in a recessed feature. Exposure of a wafer to deoxygenated
processing fluid for use in the treatment of a semiconductor wafer
is described in U.S. Pat. Nos. 6,021, 791 and 6,146,468, which are
incorporated herein by reference.
[0052] After immersion of the wafer into a pre-wetting fluid or
covering of the wafer with a pre-wetting fluid, the region around
the wafer (e.g., a pressure chamber) is closed and sealed, and an
external pressure is applied to the chamber and fluid. Pressure may
be applied pneumatically (e.g., introducing high pressure gas into
the chamber in the area over the fluid), or hydraulically (e.g.,
with the chamber substantially free of non-dissolved gas and using
a hydraulic piston or other suitable device to apply external
pressure to the fluid). As the pressure in the chamber increases,
the bubble will decrease from its original size. When using
pneumatic (gas) pressure to compress the trapped bubble, it may be
important to avoid dissolving substantial amounts of gas into the
pre-wetting fluid, particularly in the vicinity of the bubble. In
some embodiments, a stagnant, relatively thick layer of fluid, for
example, greater than 1 cm in thickness, is used. In other
embodiments, the pneumatic pressure is applied to the chamber
though a long tube with a substantial resistance for dissolution of
gas from reaching the interface so that the gas that contacts the
liquid does so over a relatively small surface area and has a
relatively long diffusion path, limiting the amount of gas that can
dissolve in the fluid over a period of time. However pressure is
applied, the driving force for dissolution of the trapped bubble
will increase with applied pressure. For a large bubble without
significant capillary pressure effects, the driving force for
dissolution will be approximately equal to the product of the
initial mole fraction of the particular gas component in the bubble
and the difference in applied pressure to the chamber and the
initial partial pressure of dissolved gas in the fluid. This later
quantity will vary depending on the extent of degassing performed
on the pre-wetting fluid.
[0053] While pressure can be applied either pneumatically or
hydraulically, in pre-wetting embodiments that are not immersion
embodiments, but rather coverage of a wafer with a thin layer of
pre-wetting fluid, the pneumatically applied external pressure will
potentially allow gas to redissolve quickly into a (e.g., degassed)
thin layer of pre-wetting fluid. There is a competition between gas
uptake from the external pressurized gas source versus gas
dissolution into the liquid from the bubble. Therefore, a
relatively thick layer of pre-wetting fluid should be used for
non-immersion pre-wetting operations. Also, there are a limited
number of practical mechanisms for applying hydrostatic pressure to
the thin layer of pre-wetting fluid on a wafer. One possible
mechanism for doing so is to create a face-up wafer and a
pre-wetting liquid fluid containing cup. In contrast, there is a
much wider tolerance with thick layers of pre-wetting fluid and the
immersion pre-wetting method. This is because pressure can be
transmitted to the bubble by a purely hydrostatic mechanism, and
alternatively, application of pneumatic pressure will not quickly
re-saturate the pre-wetting fluid around a bubble in a via with gas
because of the relatively long diffusion distances involved.
[0054] When pressure is applied, with the gas partial pressure in
the bubble exceeding that in the pre-wetting fluid, the bubble will
begin to dissolve. Eventually the bubble will completely dissolve,
the total time for which depends on parameters such as its initial
size, the applied pressure, and the original depth of the bubble
inside the feature. After the bubble has completely dissolved, some
time should generally be allowed to pass before the pressure is
released, so that any excess dissolved gas (beyond that which will
be soluble at 1 atmosphere) can equilibrate into the pre-wetting
fluid as a whole. This avoids the possibility of re-nucleating a
bubble inside the feature. When this is procedure is followed, the
bubble will be removed from the feature and not reform upon release
of the excess external pressure.
[0055] Referring to FIG. 1, curves D and E (90 versus 12 rpm
rotation in a plating bath respectively) are calculated for the
rates of bubble dissolution as discussed above, but in this case a)
with the amount of initial gas dissolved in the solution equal to
that in equilibrium with 1 atmosphere air (same as condition A,
i.e., no degassing of the contacting fluid) and b) with an external
applied pressure of 3 atmospheres. For this case, the total
pressure of dissolved gas in the bulk fluid is equal to air at 1
atmosphere, and at the interface of the bubble, in equilibrium with
3 atmospheres of pressure. Comparing cases A and F (no degassing
and no pressurization) with cases B and C (degassing but no
pressurization) and cases D and E (no degassing but with
pressurization) in FIG. 1, the pressurization of the fluid appears
to be a good method in terms of achieving the shortest time of
dissolution. Using a previously degassed pre-wetting fluid (0.2
atmospheres) in combination with a 3 atmosphere external fluid
pressurization (a case not shown in FIG. 1) will generally lead to
a further 50% reduction in degassing time for large features (3-1=2
atmospheres driving force versus 3-0=3 atmospheres driving force),
according to calculations.
[0056] Note, however, that there is a potentially significant added
advantage of using degassed fluid in this operation beyond just the
reduction in gas removal time (which could be achieved, for
example, by simply increasing the pressure to, say 4 atmospheres of
pressure in this case). After the release of the externally applied
pressure to the chamber, gas from the bubble and some of the gas
from the external source (if pneumatically driven) will have
dissolved into the pre-wetting fluid. As indicated above, unless
one waits for equilibration (which can be a relatively slow
process, taking several minutes or more), there is a tendency to
re-nucleate and re-form a bubble inside the via, since after
releasing the pressure, the fluid (particularly inside the feature)
still can contain gas at a concentration in excess to that soluble
at ambient conditions/pressures (i.e., in excess of that which
would be in equilibrium with a pressure one atmosphere). In
contrast, if the fluid is degassed before application of the
externally applied pressure, this equilibration time can be largely
eliminated, since there is a substantial excess capacity to absorb
the gas from the bubble and thereby avoid re-nucleation and
precipitation of the bubble.
[0057] Finally, depending on the wafer's orientation and the
surface tension between the bubble and the inner via surface, it is
possible that the shrinking of the trapped bubble by external
compressive pressures down to a size significantly less than the
diameter of the via could allow that bubble to detach itself from
the wall and subsequently rise out of the via mouth due to its own
buoyancy. Once the bubble has exited the via, the pressure can be
removed without the possibility of the bubble being trapped inside.
The terminal rise velocity of a bubble less than about 0.5 mm
rising in an infinite media (no wall effects) depends on its
diameter a, the kinematic viscosity .nu., and the Reynolds number
Re, and can shown to be approximately given by:
V = 2 9 ga 2 v for Re < 1.0 ( 4 ) V = 1 9 ga 2 v for 20 < Re
< 100 ( 5 ) ##EQU00001##
where g is the acceleration of gravity and .nu. is the pre-wetting
fluids kinematic viscosity (fluid viscosity divided by fluid
density).
[0058] The difference in the behavior of these cases (i.e., (4) and
(5)) is that for low Re, convection is negligible and no wake is
developed behind the rising bubble, versus the irrotational case
(i.e., when the Reynolds number is high), where wake drag is
considered and results in twice the drag. The time it takes for a
bubble to rise the depth of the via can be calculated as t
(sec)=h/V, which, for example, for a 10 .mu.m diameter bubble in a
100 .mu.m deep via (0.01 cm) would be about just under one second.
Typically a 100 .mu.m deep via might have a 25 .mu.m diameter
opening, so the assumption of a bubble rising in an infinite media
is not correct, as wall flow-slip effects will increase the time.
It is recognized that one could further speed up the process if an
external body force were applied to the system in excess to or
instead of gravity. For example, a centripetal force could be
applied by spinning the wafer with the wafer opening pointed
towards the center of rotation, helping to drive the bubble
inwards.
[0059] Equations 4 and 5 underestimate the actual bubble rise time
when the bubble diameter is close to the size of the via. This
underestimation becomes a factor when the assumption of the bubble
rising in an infinite media becomes fundamentally incorrect (i.e.,
for bubble diameters greater than about 1/4 the feature diameter
size). The shear flow stresses between the movement of the rising
bubble and the via wall begin to dominate under such circumstances.
Still, the conditions that satisfy the assumptions can be achieved
simply by applying more pressure to the system (shrinking the
bubble further), or by simply accounting for the expected longer
bubble rise/clearing time when the bubble diameter is close to the
via diameter.
[0060] Different designs of pre-wetting chambers are described
herein. One embodiment of a pre-wetting chamber is shown in FIG. 4.
In this embodiment, the pre-wetting chamber is configured for
delivering the pre-wetting fluid onto the wafer substrate in a
liquid form. The pre-wetting chamber may also be configured for
spraying or streaming the pre-wetting fluid onto the wafer
substrate for a period of time. In FIG. 4, a wafer 401 is held
face-up in pre-wetting chamber 403 with wafer holder 402. In some
embodiments, the wafer holder is configured to hold the wafer
substrate in substantially a horizontal orientation (i.e., parallel
to the Earth's horizon), during the pre-wetting process. In other
embodiments, the wafer holder is configured to hold the wafer
substrate in substantially a vertical orientation during the
pre-wetting process.
[0061] In a typical operation, vacuum is first pulled on chamber
403 though vacuum port 409, which is connected to a vacuum system
(not shown). This reduces the pressure in the chamber to a
subatmospheric pressure. After much of the gas in the chamber is
removed by the vacuum, pre-wetting fluid is delivered onto the
wafer surface from the nozzle 405 or other mechanism. In some
embodiments, the pre-wetting fluid is degassed prior to contacting
the wafer surface, again, to avoid gas being released as the
pre-wetting fluid enters the vacuum environment. The wafer may be
rotated with motor 407 during the pre-wetting fluid delivery
process to insure complete wetting and exposure of the wafer. In
some embodiments, the pre-wetting chamber is configured to deliver
the pre-wetting fluid onto the wafer substrate. In some embodiment,
the pre-wetting fluid is liquid. In some embodiment, the
pre-wetting fluid (a liquid) first contacts the rotating wafer
substrate within about 3 cm of the center of the wafer substrate.
After pre-wetting, the wafer is spun at a low rpm with motor 407 to
remove entrained pre-wetting fluid, but leaving a thin layer of
fluid on the wafer surface. Excess pre-wetting fluid is drained and
exits the vacuum chamber through port 411. The wafer is then
transferred to a standard plating cell such as a Novellus clamshell
cell for plating with a thin layer of pre-wetting fluid retained by
surface tension on its surface and within its features.
[0062] FIG. 5 depicts an isometric view of an embodiment of a
pre-wetting chamber suitable for performing pre-wetting process
described herein. FIG. 5 is a detailed schematic of a pre-wetting
chamber similar to the embodiment shown in FIG. 4. Pre-wetting
chamber 501 includes a motor 503 for rotating the wafer during
processing is anchored below the chamber via the chuck to chamber
base 504 by a motor-and-bearing supporting member 505, both of
which also create a fluid seal between the bearing 507 and the
underside of the chamber and the bearing. The bearing is a
commercially available vacuum-pass-through central shaft rotary
bearing. The motor is attached via a coupling 509 to a drive shaft
511 that passes though a vacuum isolating bearing to the chuck base
513. The chuck has three arms (515 is one arm) to support the wafer
(wafer not shown), confinement pins, and other alignment apparatus
517 as appropriate.
[0063] At the lower section of the chamber is a drain 519 for
removing excess pre-wetting fluid that may accumulate there after
it is applied to the rotating wafer. The fluid is flung out toward
the chamber walls and drops to the chamber base. In some
embodiments, a wafer peripheral "fluid defector shield" (not shown)
is positioned approximately in the plane of the wafer to deflect
fluid emanating from the wafer edge downward before hitting the
chamber wall. The deflector shield may be moveable, or the wafer
and wafer chuck plane may be adjusted by appropriate vertical
moving mechanisms and seals. Also at the base of the chamber is a
vacuum inlet and vacuum release line 521, housed in a fluid
protecting shield 523 is some embodiments. This shield helps
prevent surges of gasses from unnecessarily disturbing fluids
within the chamber as well as minimizing the amount of liquid drawn
into the vacuum line by isolating the two. While the vacuum line
(and shield) can be located in the upper section of the chamber, it
is advantageous to draw vacuum from below the wafer so as to
minimize the propensity of any particles falling onto the wafer and
forming defects. This can occur if particles or other materials
enter the chamber during back filling the chamber with a gas or
from the ambient environment while the chamber door is open. To
minimize particles and other materials from entering the chamber,
the chamber is typically backfilled with a particle-filtered inert
gas such as nitrogen, carbon dioxide, or argon, and a slight
positive pressure of clean particle free gas is feed to the chamber
while the door is open. The backfill gas is typically filtered and
the entering fluid enters a flow diffuser mounted on the wall of
the chamber, so as to avoid a gas flow jet that might dry the wafer
or disturb any chamber contents unnecessarily.
[0064] In some embodiments, a pre-wetting fluid nozzle 525 is
located above and to the side but not over the centrally located
wafer and wafer chuck, oriented and configured to spray or stream
fluid to reach the wafer central regions. In other embodiments, the
pre-wetting fluid nozzle is attached to a movable arm which can be
positioned over the wafer. In the embodiment shown in FIG. 5, the
chamber vacuum door 527 is located along the walls of the chamber,
and configured to seal to the chamber itself. It can be moved away
from the chamber as well as downwards (or upwards) so a wafer can
enter the chamber freely, and then be repositioned to the sealing
position after a wafer is placed onto the wafer holding chuck. The
doors and other elements that potentially could hold entrained
fluid should be designed such that the fluid may not drip onto the
wafer. For example, the door's retracted position and associated
hardware may be positioned below the plane that the wafer creates
on insertion into the chamber, so as to avoid dripping fluid of
otherwise contaminating the wafer during transit in or out of the
chamber.
[0065] In some embodiments, the upper section of the chamber,
particularly the areas above the plane where the wafer sits in the
chuck and is extracted through the door, are heated above the
temperature of the wafer that is to be pre-wetted. This includes
both the areas that reside above the wafer (top surface or vacuum
dome, not shown in FIG. 5) as well as the peripheral areas around
the wafer. This heating is useful in avoiding liquid from dropping
from the ceiling of the chamber onto the wafer before vacuum
conditions are established, potentially trapping an air bubble
inside a via where the drop fell, circumventing the desired process
of putting a pre-wetting fluid onto the wafer only when air is
first removed from the vias. Similarly, during the placement of a
wafer into the chamber, liquid falling from the walls onto the
wafer surface would have a similar effect. By heating the chamber
walls, condensation on the walls and ceiling is avoided, as well as
enabling the rapid evaporation of any stray droplets that might
otherwise reach those positions, thereby keeping these areas
dry.
[0066] While not shown in FIG. 5, in some embodiments a vertically
moveable and automatable splash shield is positioned peripheral to
the wafer and chuck and inside the chamber. The splash shield can
be moved upwards during the application of fluid or at other times
as suited to minimize and avoid, among other things, liquid from
contacting the chamber door or upper walls. Alternatively, the
wafer chuck can be moved downwards deeper into the chamber and
below the plane of the vacuum door after wafer insertion,
accomplishing the same purpose.
[0067] In other embodiments, rather than delivering a pre-wetting
fluid to the wafer surface, the wafer is immersed in or otherwise
covered with a pre-wetting fluid (e.g., by condensation) while
vacuum conditions are maintained above the fluid and wafer. Since
the creation of a vacuum within the chamber creates conditions
where there is substantially no non-condensable gas in the chamber,
the pre-wetting fluid is not impeded from entering a via. Put
another way, the liquid does not need to displace any gas located
within a via during pre-wetting, since the gas has been removed in
a separate operation (pulling vacuum) prior to the pre-wetting
operation.
[0068] For example, in one embodiment, after vacuum is applied to
the pre-wetting chamber, a condensable fluid vapor is created
within the chamber or introduced to the chamber (e.g., vapors of
water (e.g., low pressure steam), methyl alcohol,
dimethylcarbonate, diethylcarbonate, isopropyl alcohol, dimethyl
sulfoxide, and dimethyl formamide, or other liquid used as the
subsequent plating electrolyte, easily dissolvable in a subsequent
rinse, or soluble in the subsequent plating electrolyte). In
embodiments where the wafer substrate has at least one recessed
feature and the pre-wetting chamber is configured to deliver
pre-wetting fluid onto the wafer substrate in a gaseous from, the
pre-wetting fluid condenses to form a liquid film on the wafer
surface that fills the recessed feature with the pre-wetting fluid.
FIG. 6 depicts an embodiment of a pre-wetting chamber that is
configured for such a condensation pre-wetting process. FIG. 6
shows chamber 601 that has a moveable vacuum lid (alternatively an
access door) 609 allowing access to the chamber, a line to a vacuum
source 611, a vacuum release line 613, and a condensable fluid
inlet 615. Vacuum seal 617 seals the lower vacuum containment
vessel 619 from the rest of the chamber. The wafer 603 sits on a
wafer cooling element (chiller) 605 that is part of the wafer
holding fixture (chuck) 607. The wafer cooling element 605 reduces
the wafer substrate surface temperature to a temperature below the
condensation temperature of the pre-wetting fluid that flows into
the chamber thorough inlet 615 as a vapor. In another embodiment,
after creating a vacuum and removing the condensable gasses (e.g.,
air) from the chamber 601 with vacuum, water is simply heated and
allowed to vaporize (i.e., boil) in the chamber, and allowed to
condense on the surfaces, including and preferentially on the
cooler wafer 603, inside the chamber. For example, in a chamber
without vacuum seals 617, a small amount of water in the lower
section 619 of the chamber could be heated and allowed to flash
simultaneously while vacuum is pulled inside the chamber. The
connection to the vacuum can be removed (closed) at some point
during the process.
[0069] In another embodiment, the wafer substrate is immersed into
a bath of pre-wetting fluid for a period of time. FIG. 7 depicts an
embodiment of a pre-wetting chamber configured for such an
immersion pre-wetting process. In FIG. 7, the wafer 701 is held in
wafer holder 702 in a chamber 703. Chamber 703 has an inlet 711 for
admitting pre-wetting fluid. As shown, the wafer is held in the
wafer holder face up, and held by an appropriate mechanism that
still allows fluid to reach wafer from the peripheral edge. Vacuum
is pulled on chamber 703 though vacuum port 707, which is connected
to a vacuum system (not shown). Then, the wafer is wetted with a
pre-wetting fluid by, for example, either 1) the wafer and wafer
holder moving down into the pre-wetting fluid 713 or 2) the
pre-wetting fluid level rising by fluid entering through inlet 711.
During the pre-wetting process, the wafer may be slowly rotated
with motor 705. After the pre-wetting process, the liquid level is
lowered, or the wafer raised, and the wafer is spun at low rpm with
motor 705 to remove excess entrained fluid, leaving a thin
pre-wetting fluid layer. A flow of nitrogen gas through port 709
may also be used to dry the backside of the wafer while the
frontside of the wafer remains wetted. The wafer is then
transferred to a standard clamshell for plating.
[0070] In other embodiments of the pre-wetting chamber shown in
FIG. 7, the wafer can be held in a face down position. In some
embodiments of a pre-wetting apparatus with a pre-wetting chamber
as shown the FIG. 7, the pre-wetting apparatus is configured to
initiate immersion of the wafer into the pre-wetting fluid after
the pressure in the chamber has been reduced to less than about 50
Torr. The pre-wetting chamber 703 shown in FIG. 7 can be used in
embodiments in which an external pressure is applied to dissolve
bubbles, as described herein. The chamber and other components
would need to withstand internal pressures instead of or in
addition to vacuum.
[0071] FIG. 8 depicts another embodiment of a pre-wetting chamber
configured for an immersion pre-wetting process. FIG. 8 shows
pre-wetting chamber 801, wafer 809, and fluid 813 or wafer holder
803 moving relative to each other. In this embodiment, the chamber
and wafer holder 803 can be tilted for precise control of the
pre-wetting front and complete liquid removal from the chamber.
Also, the gap between the wafer 809 and the bottom of the chamber
is small. As in FIG. 7, pre-wetting fluid in FIG. 8 may enter/exit
though port 811 and a vacuum may be pulled on chamber 801 though
vacuum port 807, which is connected to a vacuum system (not shown).
Excess entrained fluid may be removed from the wafer surface by
spinning it at a low rpm with motor 805. The embodiment shown in
FIG. 8 is particularly useful when pre-wetting the wafer substrate
surface with a high-cost pre-wetting fluid, or when it is otherwise
desirable to use a minimal amount of pre-wetting fluid (e.g., so
the level of dissolved gas can be maintained at low levels). After
pre-wetting, the wafer is transferred to a standard clamshell for
plating. A similar design of a narrow-gap, tilted surface
pre-wetting apparatus, but without a mechanism for applying a
vacuum during the pre-wetting operations, is described in U.S.
patent application Ser. No. 11/200,338, filed Aug. 9, 2005, herein
incorporated by reference.
[0072] The chamber shown in FIG. 8 can also be used in the
embodiment in which an external pressure is applied, as described
above. In this embodiment, the chamber and other equipment is
designed or modified to be able to withstand and maintain internal
positive pressures.
[0073] An embodiment of an apparatus in which the pre-wetting
process is performed in a plating cell is shown in FIG. 9.
Alternatively, this can also be stated that the pre-wetting chamber
is configured to both pre-wet a wafer substrate and to electroplate
a layer of metal on the pre-wetted wafer substrate. In FIG. 9,
chamber 901 is a plating cell, with a vacuum sealing surface being
a section of the cell wall 903. Wafer holding fixture 905 holds
wafer 915. In the embodiment depicted in this figure, the plating
cell contains an ionically resistive ionically permeable high
resistance virtual anode (HRVA) 907 and a separated anode chamber
(SAC) region 909. One example of an HRVA containing apparatus is
described in U.S. patent application Ser. No. 12/291,356, filed
Nov. 7, 2008, which is incorporated herein by reference in relevant
part. See also U.S. patent application Ser. No. 11/506,054, filed
Aug. 16, 2006, which is incorporated herein by reference in
relevant part.
[0074] Initially wafer 915 is held above the plating solution 913
and the vacuum is drawn on the chamber through vacuum port 911.
When vacuum is drawn on the chamber, vacuum typically should be
drawn on the backside of the wafer through the wafer holding
fixture so that the wafer does not fracture. Afterwards, the fluid
level 913 is raised, wetting the wafer surface. In some
embodiments, this fluid is pre-wetting fluid, and in other
embodiments, this fluid is a plating solution. In some embodiments,
the fluid is degassed prior to contacting the wafer surface. Since
there is no gas in the chamber, the fact that the wafer is face
down does not lead to any trapped gas-containing bubbles below the
surface or inside the vias. After the pre-wetting is complete, the
vacuum can be released. Electroplating a metal (in some
embodiments, copper) on wafer 915 can then begin. It is generally
simpler (mechanically and processing conditions) to perform plating
at ambient pressures, with or without wafer rotation.
Alternatively, the vacuum can be held throughout the electroplating
process. Again, it is advantageous in this and other embodiments to
have the fluid degassed prior to performing the pre-wetting
operations. Otherwise the fluid may release dissolved gas, forming
bubble inside the features or on the surface as gas is driven out
of the liquid by the lower pressure.
[0075] A general description of a clamshell-type plating apparatus
having aspects suitable for use with embodiments described herein
is described in detail in U.S. Pat. Nos. 6,156,167 and 6,800,187,
which are incorporated herein by reference for all purposes.
[0076] FIG. 10 depicts an embodiment of an electroplating
system/module 1001 for processing wafers. The particular tool
layout shown contains two separate wafer handling robots, 1003
which moves a "dry" wafer from a cassette stationed in a front FOUP
(front opening unified pods) loader 1005 to an aligner
module/transfer station (not shown) and a transfer chamber robot
1004. The aligner module ensures that the wafer is properly aligned
on a transfer chamber robot 1004 arm for precise delivery to other
chambers/modules of the system. In some embodiments, the aligner
module both aligns the wafer azimuthally (so called "wafer notch
aligning") as well as in the vertical and horizontal planes to a
particular location (i.e., fixes the wafer's x, y, and z position
registry).
[0077] The same or a different transfer chamber robot is used to
feed the wafer back from the back end "wet processing area" of the
tool to the FOUP after processing and drying is completed. A back
end robot (not shown) may contain two or more arms, each with a
single or multiple "end-effectors" to grip the wafer. Some
"end-effectors" grip the wafer at the bottom of the wafer with a
vacuum "wand", and others may hold the wafer only at is peripheral
edge. In some embodiments, one robot wafer handling arm
end-effectors is used only for handling a wafer that has a wet
surface, and the others reserved for handling only fully dry
wafers, thereby minimizing contamination.
[0078] After a wafer enters the transfer station (containing
transfer chamber robot 1004), the wafer typically is fed to a
pre-wetting chamber 1013 (i.e., a pre-wetting apparatus is a
station in a module, the module further including an electroplating
station configured to electroplate the wafer with a metal, the
metal being copper in some embodiments), the various embodiments of
which are described herein. In other embodiments, system 1001 is
configured for an anodic process. In such embodiment, the module
further includes a station configured for an anodic process, such
as electroetching or electropolishing.
[0079] The pre-wetting chamber 1013 is either configured to pre-wet
a wafer under vacuum or to apply pressure to a wetted wafer, and in
some embodiments, both. Using a pre-wetting chamber configured to
pre-wet a wafer under vacuum as an example, ambient air is removed
from the chamber while the wafer is spun. Once vacuum is achieved,
the device side of the wafer is exposed to degassed pre-wetting
fluid (degassed in module 1015 with a degassing flow loop). After
wetting is complete, excess fluid is removed, gas is reintroduced
to the chamber to atmospheric pressure, and the chamber is opened
to allow the wafer to be extracted by the robot or other transfer
mechanism. In some embodiments, the transfer mechanism is
configured to transfer the pre-wetted wafer substrate from the
pre-wetting station to an electroplating station in less than about
one minute.
[0080] In some embodiments, the wafer is then placed into an
aligner (not shown), such as a notch aligner. By passing though a
high-accuracy notch aligner, accurate placement into a edge sealing
plating cell, which excludes plating solution from the back and the
very small device side edge exclusion region (e.g., about 1 mm from
the edge), is possible. The plating cell may be specially designed
to have a seal that traverses the notch area. Plating and feature
filling (i.e., a layer of metal is electroplated on the wafer
substrate) occurs in plating cell 1021, 1023, or 1025, (i.e.,
electroplating stations) and in some embodiments, the plating
solution is a degassed solution. In some embodiments, the metal is
copper. The electroplating stations are configured to immerse the
wafer in a degassed plating electrolyte in the electroplating
station. In some embodiments, an electroplating station is
configured to cathodically polarize the wafer substrate before
immersing the wafer substrate in a degassed plating electrolyte.
The plating solution can be recycled though a separate degassing
loop different from a flow loop between a main plating bath and the
plating sell, or by passing though the degassing element in the
same loop as the bath/plating cell loop, being degassed just prior
to entering the plating cell.
[0081] After plating is completed the wafer is rinsed with water
above the plating cell and spun to remove excess entrained fluid,
the wafer holding clamshell apparatus is opened releasing the edge
seal and allowing wafer extraction. In some embodiments, the wafer
is then picked-up from the plating cell and transported in to a
metal removing isotropic etching module (ITE module) 1031. The ITE
module is a wet etching module used to remove metal from the top of
the wafer primarily in the field region over the feature of the
plated wafer, while leaving at least some metal inside the recessed
features. Various designs of suitable equipment, etching processes,
and etching chemical formulations are described in U.S. Pat. Nos.
5,486,235, 7,189,649, 7,338,908, 7,315,463 and U.S. patent
application Ser. No. 11/602,128, filed Nov. 20, 2006, U.S. Ser. No.
11/888,312, filed Jul. 30, 2007, and 11/890,790, filed Aug. 6,
2007, each incorporated herein by reference.
[0082] In addition, metal at the edge of the wafer is removed in
the ITE module 1031. Because a wafer is often held in an edge
excluding clamshell apparatus, only a thin layer of seed metal
exists at the outermost periphery (the original seed layer) prior
to the top side global etching performed here. Therefore, after
processing here, it is common for the extreme edge of the wafer to
be completely bare of metal, while the more central, non-plating
protected and edge excluded region may have some metal remaining
(however, in other cases, the metal is removed from that region as
well). This module can therefore perform both the global etch
removal of metal from the wafer as a whole, as well as remove metal
from the outer periphery edge and outer peripheral bottom of the
wafer, often eliminating the need to perform a more complex edge
specific etch process, edge bevel removal (EBR), as described in,
for example, U.S. Pat. No. 6,309,981, incorporated herein by
reference.
[0083] In some embodiments, the progress of the etching process and
the thickness distribution of the film is monitored in the etching
module, for example, by measuring the cross wafer sheet resistance
with an eddy current meter, or the reflection of an acoustic
signal. Alternatively, the thickness after etching can be measured
in the transfer station in the dry state later in the process, and
the process results monitored or modified as appropriate to
minimize any wafer-to-wafer performance drift. After etching, the
wafer can be either rinsed and dried in the etching module, or
moved to a separate module, a wafer rinsing, cleaning and drying
station 1041. There, any oxide film that may have formed in the
process sequence is removed or reduced (e.g., by applying a dilute
acid solution to the surface), any residual chemical not removed by
a more cursory rinse in the etching station is removed (both front
and back of the wafer), and an edge bevel removal operation is
performed as desired (see, for example, U.S. Pat. No. 6,309,981).
After rinsing the wafer with water, it is spun dry and then removed
to a transfer station, where the front end robot re-deposits it
into the wafer holder cassette.
[0084] One concern with the pre-wetting process is that in the time
between pre-wetting and plating (i.e., after exposing the wafer to
pre-wetting fluid while under vacuum in a pre-wetting chamber but
before plating commences), it is possible to have the wafer surface
"de-wet". Dewetting may be described as a physical draining and
coagulation of the pre-wetting fluid from the surface (i.e., rather
than a drying of the surface), leaving one section of the surface
with a thicker film of pre-wetting fluid, and another section with
no pre-wetting fluid thereupon. This characteristic behavior is
generally associated with a highly hydrophobic surface with respect
to the pre-wetting fluid. If the wetting layer pulls back or
coagulates from a previously wetted surface, then the attributes of
the pre-wetting process are lost. To avoid this phenomenon, wetting
agents can be added to the pre-wetting fluid to avoid the pooling
of the fluid into puddles.
[0085] Surface oxides, surface contaminants, and other deposited
materials on the wafer surface that are exposed to air and humidity
may be very hydrophobic. For example, thin copper metal seed
layered wafers exposed to air and water vapor will form a thin
cuprous oxide layer, which is hydrophobic with respect to water. To
avoid this potential problem, the oxide film can be removed in
certain embodiments by, for example, by adding a small amount of
acid (e.g., H.sub.2SO.sub.4, H.sub.3PO.sub.4) at a pH where the
oxide is no longer stable, to the pre-wetting fluid used in the
pre-wetting process. The acid will react with the oxide to form
water and the metal salt. The pre-wetting fluid can also contain
small amounts of surface tension and contact angle lowering wetting
agents (e.g., surfactants, alcohols) which also avoid the
phenomena. Pre-wetting fluid chemistry is discussed further
herein.
[0086] In some embodiments where the pre-wetting operation is
performed in a separate chamber prior to plating, the pre-wetting
fluid may contain a small amount of metal ions, for example, to aid
in avoiding the formation of bacteria in the system or modifying
the oxide removal properties. Alternatively, a metal oxide suitable
reducing agent may be added to the wetting solution, such as
formaldehyde, glyoxylic acid or dimethyl-amine borane, or with a
metal ion complexing additive (for copper, examples might include
ammonia, glycine, ethylene diamine). Further, the surface oxide or
other contaminants can be removed by treating the wafer in a
reducing atmosphere (e.g., forming gas or hydrogen in argon), with
or without heating, prior to the pre-wetting operation. The
temperature of the pre-wetting fluid and the wafer surface can also
be increased or decreased from ambient conditions to optimize the
retention of fluid on the surface.
[0087] In some embodiments, operations in a pre-wetting chamber or
a pre-wetting chamber that is part of an electroplating system are
controlled by a computer system. The computer includes a controller
including program instructions. The program instructions may
include instructions to perform all of the operations needed to
pre-wet a wafer substrate. In one embodiment, the instructions are
for reducing pressure in the process chamber to a subatmospheric
pressure and subsequently contacting the wafer substrate with the
pre-wetting fluid at a subatmospheric pressure to form a wetting
layer on the substrate surface. The wafer substrate may be rotated
at a first rotation rate during delivery of a liquid pre-wetting
fluid onto the wafer substrate at a subatmospheric pressure, the
fluid delivery being performed for between about 10 to 120 seconds.
Then, delivery of the pre-wetting fluid is stopped. After stopping
the delivery of the pre-wetting fluid, the wafer substrate is
rotated at a second rotation rate to remove excess surface
entrained pre-wetting fluid from the wafer substrate. In some
embodiments, the vacuum in the process chamber is released after
the delivery of the pre-wetting fluid is stopped and prior to
removal of the excess entrained pre-wetting fluid. In alternative
embodiments, the vacuum is released after removal of the excess
entrained pre-wetting fluid. The wafer may be rotated at different
rates in different embodiments. In some embodiments, the first
rotation rate during delivery of a liquid pre-wetting fluid onto
the wafer substrate is less than about 300 rpm and the second
rotation rate to remove excess entrained pre-wetting fluid from the
wafer substrate is at least about 300 rpm. In other embodiments,
the first rotation rate is about 100 rpm or less, and the second
rotation rate is at least about 500 rpm. In yet further
embodiments, the pre-wetting apparatus is configured for removing
excess entrained pre-wetting fluid from the wafer substrate by a
method selected from the group consisting of centrifugal spinning,
air-knife drying, and wiping and the controller includes program
instructions for performing these operations.
Process/Method
[0088] In a general pre-wetting method for some embodiments
disclosed herein, a vacuum is first created in the environment
around the wafer. Then, the wafer surface is sprayed with, streamed
with, covered with, or immersed in sufficient (in some embodiments,
degassed) pre-wetting fluid, eventually exposing the entire wafer
to a sufficiently thick liquid layer. The layer may not cover the
entire surface at all times, until later in the process. The wafer
surface is then left immersed or otherwise exposed to a pre-wetting
fluid layer for a period of time (e.g., by continuing to spray,
stream, cover, or immerse the surface with additional fluid) until
adsorption (or reaction) of any pre-wetting fluid constituents at
the wafer surface have substantially reached completion/equilibrium
and a favorable/uniform wetting character (hydrophilicity, low
contact angle) is achieved. After pre-wetting, the spraying,
streaming, or covering of the wafer with pre-wetting fluid is
stopped. In some embodiments, vacuum is released and then excess
entrained fluid is removed from the (now) completely hydrophilic
surface (e.g., by centrifugal spinning, air-knife drying, squeegee
wiping, etc.), leaving a thin uniform adherent layer of pre-wetting
fluid on the surface. In other embodiments, excess entrained fluid
is removed before releasing vacuum. Finally, the wafer is
transferred to plating cell to plate the wafer.
[0089] Because there may be anywhere from a few seconds to over a
minute between the time that entrained pre-wetting fluid is removed
from the wafer surface to the initiation of metal deposition, it is
important that the wafer is globally hydrophilic and remain
completely coated with fluid over the entire surface. In the
ensuing time, a hydrophobic surface/fluid combination can lead to
the fluid receding from, and uncovering a portion of, the wafer
surface, for example, starting from the wafer edges. This
de-wetting may cause the fluid to be drawn out from within any
recessed features within the wafer substrate, possibly leading to
gas being trapped within the feature on immersion into the plating
bath. Hydrophobic surfaces, particularly those that have completely
de-wetted in some regions, have non-uniform fluid pre-wetting layer
thickness over the wafer substrate. In the case that the
pre-wetting fluid in use has a different composition than the
plating bath, the subsequent immersion of the pre-wetted wafer into
the plating solution will not allow for a uniformly wetted surface
if the pre-wetting fluid has not wetted the wafer properly. The
non-uniformly wetted wafer will cause the diffusion times and
concentrations various components to be different across the
wafer's surface due to the thickness of the wetted layer. This can
lead to variation in feature filling behavior or the creation of
various wafer surface defects, such as lines of entrapped bubbles,
metal pits, metal thickness variations, or growth protrusions.
Therefore, after the pre-wetting process, the pre-wetting fluid
should create a uniform, small contact angle with respect to the
entire wafer surface, for example, a contact angle of about 45
degrees or less, if possible. When a lower contact angle is
possible, a very thin and adherent pre-wetting fluid layer can be
created.
[0090] It is often observed that the contact angles of a surface
can change with time, and that hydrophobic surfaces may become more
hydrophilic over time when exposed to certain liquids. Certain
wafer surfaces, such as those coated with copper films by, for
example, plasma vapor deposition, can exhibit a significant
decrease in the liquid/surface contact angle with time upon
continuously exposing the surface to the pre-wetting fluid. In
particular, the continual exposure of such a surface, while under
vacuum conditions, can lead to rapid and complete transformation of
the surface from a generally de-wetted, hydrophobic state, to a
wetted, hydrophilic state.
[0091] Furthermore, this transformation, specifically when
occurring under vacuum and with a degassed pre-wetting fluid, leads
to particularly favorable low defectivity when combined with the
subsequent plating operation. While not wanting to be bound by any
particular wetting model or theory, if sufficient time to pass is
allowed to pass (e.g., 5 seconds to 1 minute) while the surface is
immersed in, sprayed with, streamed with, covered with, or
otherwise treated with a surface-tension-lowering pre-wetting
fluid, the surface can undergo a conversion from a hydrophobic to a
hydrophilic state. For example, by allowing time for low
concentration constituents (e.g., wetting agents) time to adsorb to
the wafer interface, or alternatively, time for spurious adsorbed
species that reside on the surface (e.g., from atmospheric
exposure) to be desorbed from the surface, suitable stable wetting
behavior can be obtained. Alternatively, agents in the pre-wetting
fluid may react to slightly roughen the surface and/or remove thin
surface layers such as surface oxides, nitrides, or carbonates.
[0092] As a specific example, the transformation of a cuprous or
cupric oxide laden surface, which tends to be inherently quite
hydrophobic to water, to a hydrophilic metallic surface, is needed.
By simple exposure to either de-ionized (DI) water (which does not
react with the oxide), the surface may remain largely hydrophobic.
Alternatively, by exposing the surface to a slightly acidic oxide
removing solution, such as DI water containing a small amount of
either dissolved acid (e.g., sulfuric, methane sulfonic, or acetic
acid, resulting in a pH of between about 2 to 4), with or without
dissolved metal ions and salts, a small amount of metal (e.g.,
copper) complexing agent (e.g., citrate, pH of between about 3 to
6, glycine or ethylene diamine, pH of between about 6 to 12) or to
a solution containing an appropriate metal oxide reducing
agent/compound (e.g., formaldehyde, glycolic acid, dimethylamine
borane), is effective in removing the surface oxide and
transforming a hydrophobic interface to a hydrophilic interface.
Two examples of copper surface oxide removing reactions in a weak
acid are:
CuO+2H.sup.+.fwdarw.Cu.sup.+2+H.sub.2O, and (6)
Cu.sub.2O+2H.sup.+.fwdarw.2Cu.sup.-+H.sub.2O.fwdarw.Cu.sup.+2+Cu+H.sub.2-
O (7)
[0093] A thin oxide surface layer of cuprous and cupric oxide on
copper is formed almost immediately and continues to grow in
thickness over time, simply as a result of exposure of the
sputtered copper surface to atmospheric air, particularly moist air
(i.e., air with humidity). The oxide can be converted/removed by
exposure to an appropriate removal agent (such as those listed
herein), but it is important to consider complete oxidation of the
copper layer (also, e.g., within the feature). Subsequent removal
of the metal oxide layer by use of an oxide removal process (as
opposed to an oxide reducing process) may inhibit subsequent film
growth for a fully oxidized copper layer. Also, the wetting
conversion process (such as those listed herein) are chemical
reactions having finite reaction rates. For example, exposing the
wafer to the oxide removing pre-wetting fluid or plating bath will
begin to form a layer of hydrophilic surface at the point of fluid
contact. Areas with longer exposure to the pre-wetting (e.g., oxide
removing) fluid may prevent other areas of the wafer from becoming
wetted in the process. The hydrophilic areas that may be created
can tend to channel fluid flow thereupon, preventing the wetting of
other areas. One objective, therefore, is a modification of the
contact angle, wetting properties, and general wetting process to
enable the entire surface to become eventually uniformly covered
with liquid, both macroscopically and microscopically.
[0094] By applying a degassed pre-wetting fluid to the surface
while simultaneously maintaining a low pressure/vacuum atmosphere,
the impediment of simultaneously expanding, flushing, or otherwise
removing trapped gas from the surface is substantially eliminated,
and so the impediment of exposing areas of the wafer that are still
hydrophobic due to no or limited previous exposure to the
pre-wetting fluid may be reduced. Considering the process without
employing the vacuum and wetting combination, the various regions
of the wafer surface will fall into 5 wetted categories: 1)
Hydrophobic Wetted: Covered with and wetted with pre-wetting fluid
but for an insufficient time, so it is still hydrophobic; 2)
Hydrophilic Wetted: Covered with and wetted with pre-wetting fluid
for a sufficient time, so it is has become hydrophilic; 3)
Un-wetted: Hydrophobic, exposed to air, and never exposed to
pre-wetting fluid; 4) De-wetted: Previously wetted, but having
become de-wetted, and again exposed to air; 5) Trapped Bubble:
Containing a bubble containing trapped air at the surface and under
a layer of pre-wetting fluid.
[0095] It is important to note that an area in state 3, 4, or 5
will not undergo any adsorption or chemical reaction, leading to
the absence of any hydrophobic-to-hydrophilic surface
transformation unless and until that region later becomes wetted.
Furthermore, areas around state 3 that are in state 1 or 2 are
wetted and are or will become hydrophilic, allowing fluid to flow
freely and continuously over this surface and making the removal of
the bubble or wetting of adjacent surfaces considerably more
difficult. Also, a currently hydrophobic surface region, having
previously been exposed to pre-wetting fluid, may repeatedly go
between the states of liquid-coverage-free and covered but
hydrophobic. The process continues converting between these states
as fluid wicks away to adjacent hydrophilic area, oscillating back
and forth from state 1 to state 3 multiple times, until eventually
it either it i) changes to state 2 and becomes hydrophilic and
wetted, thereafter stays in state 2, or ii) become surrounded by
areas that are more wetting, encapsulates a bubble, and transforms
to state 4.
[0096] The above processes, performed under atmospheric conditions
(i.e., in air), should be contrasted with processes performed under
vacuum (and with degassed pre-wetting fluid). In these processes,
there are only three wetted categories that exist: 1) Wetted:
Covered and wetted with pre-wetting fluid; 2) Un-wetted: Exposed to
vacuum and never exposed to pre-wetting fluid; 3) De-wetted:
Previously wetted, but having become de-wetted and re-exposed to
vacuum.
[0097] A pre-wetting process performed under vacuum ensures that,
as long as a particular part of the wafer has been exposed to
pre-wetting fluid (state 1) for sufficient time, the particular
part of the wafer will eventually become hydrophilic. Unlike a
pre-wetting process performed in atmosphere, a high fluid velocity
pre-wetting fluid stream is not required to "flush away" trapped
bubbles. Furthermore, bubble flushing is not 100% effective, and
will often lead to bubble fragmentation, leaving a large number of
smaller, hard to remove bubble behind. Hence, pre-wetting under
vacuum is a much more reliable low defect process over simply
spraying, covering, or immersing the wafer into a pre-wetting fluid
under atmosphere. Other factors that favor pre-wetting under vacuum
are that a) surface energies of the vacuum/liquid/metal interface
are different and the contact angle is often lower than the
air/liquid/metal interface, b) metal oxide/nitride/carbonate
reformation is avoided, and c) using degassed fluid prevents the
possibility of gas precipitating out of the fluid, for example, as
a result of a spurious temperature or pressure change at some
points at the liquid-wafer interface.
[0098] FIG. 11a is a flow diagram for a general embodiment of a
pre-wetting process (1100). A wafer substrate having an exposed
metal layer on at least a portion of its surface is provided to a
pre-wetting process chamber (1105). The pressure in the process
chamber is then reduced to a subatmospheric pressure (1110). The
wafer substrate is then contacted with a pre-wetting fluid at a
subatmospheric pressure to form a wetting layer on the wafer
substrate surface (1115). Such a pre-wetting process can be
preformed in the pre-wetting apparatus designs described
herein.
[0099] The wafer substrate has different features in different
embodiments. The wafer substrate may have at least one recessed
feature. The recessed feature may be a damascene feature, which are
formed by damascene patterning processes. A damascene plating
process is a process in which a recess in a dielectric layer of a
semiconductor wafer formed by a damascene patterning process is
filled with a metal film. A recessed feature may also be a
though-mask feature.
[0100] In some embodiments, the pre-wetting fluid is substantially
free of dissolved gases. In some embodiments, one or more dissolved
gases are removed from the pre-wetting fluid prior to contacting
the wafer with the pre-wetting fluid. To aid in removal of
dissolved gases, in some embodiments, the pre-wetting fluid is
cooled to less than about 20.degree. C. during removal of the
gases. To remove gases from a pre-wetting fluid to obtain, in some
instances, a pre-wetting fluid that is substantially free of
dissolved gases, a pre-wetting fluid treatment tank has the
pre-wetting fluid circulating for a specific time period (typically
1/2 an hour, depending on the capabilities and capacity of the
degasser) though a degassing loop before contacting the wafer
substrate with pre-wetting fluid. This is discussed herein with
respect to FIG. 3. Typically this implies that fluid is flowing
through the loop while the vacuum pump is on and at vacuum, and the
valve connecting the degasser and to the pre-wetting tank to the
pump is open. This ensures that the pre-wetting fluid that is
subsequently applied to the wafer surface is substantially free of
dissolved gases. Measurements of a system so designed shows
residual levels of dissolved oxygen reaching as little as about
1-2% or less of that saturated with oxygen from air.
[0101] Furthermore, dome and wall heaters on the process chamber
may be turned on, set to a temperature of about 10.degree. C., and
in some instances about 20.degree. C. or greater, than that of the
pre-wetting fluid temperature. For example, if the fluid
temperature is about 20.degree. C., a wall temperature of about 40
to 50.degree. C. is appropriate. Dome and wall heaters avoid
condensation on the surfaces and the potential for liquid droplets
falling onto the exposed surface prior to pre-wetting under vacuum.
A purge of the chamber surfaces can be accomplished by bringing the
chamber to vacuum with the door closed and the walls at the target
heated temperature. For example, without a wafer present in the
chamber and the walls heated, the chamber is brought to vacuum and
remains at vacuum for about 10 minutes or more, so as to remove any
liquid which might have accumulated on the chamber ceiling and
upper walls. The vacuum can be removed by backfilling with, for
example, clean dry nitrogen. This procedure removes any possible
condensate from the chamber walls and minimizes the formation of
gas born particles. After confirming that a) all chamber fluid
level sensors are at appropriate values (e.g., tank full, chamber
empty), b) the heater is on, and c) the vacuum is ready for
processing, the pre-wetting chamber process door can be opened and
the door shield (if so equipped) dropped. Next, a wafer is placed
into the chuck and the robot arm is retracted, the vacuum door is
closed, and the liquid splash shield is raised or the wafer lowered
below the shield (if so equipped).
[0102] A target level of vacuum for the pre-wetting process in some
embodiments is between about 10 and 100 torr, for example about 40
torr. In some embodiments, the vacuum (i.e., subatmospheric
pressure) is about 50 torr. In some embodiments, after pump down is
complete, the vacuum line can be closed, while in other
embodiments, the pump continues to pull a vacuum while pre-wetting
fluid is injected into the chamber and onto the wafer.
[0103] In some embodiments, a liquid pre-wetting fluid is delivered
onto the wafer substrate surface. This may entail immersing the
wafer substrate in the pre-wetting fluid. Alternatively, this may
entail spraying or covering the wafer substrate with the
pre-wetting fluid. In other embodiments, contacting the wafer
substrate with a pre-wetting fluid is performed by delivering a
gaseous pre-wetting fluid onto the wafer substrate. The gaseous
fluid is allowed to condense and form the wetting layer on the
wafer substrate. In these embodiments, the temperature of the wafer
substrate may be reduced below the condensation temperature of the
pre-wetting fluid before exposing the wafer substrate to the
pre-wetting fluid.
[0104] In some embodiments, the wafer may be rotated while a liquid
pre-wetting fluid is delivered onto the wafer substrate surface. In
some embodiments, the wafer substrate is rotated at a rate of
between about 10 rpm to 300 rpm. In further embodiments, the wafer
substrate is rotated at a rate of between about 10 rpm to 100 rpm.
In other embodiments, the wafer substrate is spun at speed of from
about 100 to 400 rpm, for example at about 300 rpm. In some cases a
higher rotation rate (e.g., about 400 to 800 rpm), or a cycling of
rotation rate, may be used for a short time (about 2 to 10 seconds)
where overcoming fluid wetting resistance of a highly hydrophobic
wafers is an issue. Chamber pump down may be initiated before of
after wafer rotation is started.
[0105] In embodiments where a liquid pre-wetting fluid in used,
flow of the pre-wetting fluid is initiated into the chamber and
onto wafer surface. A typical flow rate of between about 0.5 and 2
pm, for example, about 0.8 lpm, is used, for between about 3
seconds and 1 minute or more, for example, for about 20 seconds,
depending on the necessary time to achieve full wetting of a
particular surface, rotation rate of the wafer, and the wetting
properties of the fluid. In some embodiments, the pre-wetting fluid
is contacted with the wafer substrate from about 10 second to 120
seconds. After the wetting process is complete, the pre-wetting
fluid flow is stopped, for example, by closing a pre-wetting fluid
flow valve.
[0106] Next, the chamber is brought to an atmospheric pressure. In
some embodiments, the chamber is brought to an atmospheric pressure
with an oxygen-free gas, e.g., dry nitrogen.
[0107] In some embodiments, excess pre-wetting fluid is removed
from the substrate surface. This can be done before or after
bringing the chamber to an atmospheric pressure. In some
embodiments, excess pre-wetting fluid is removed from the wafer
substrate surface by rotating the wafer substrate. The wafer
substrate rotation rate is increased to a value where excess
entrained fluid can be removed from the wafer substrate surface,
but a thin layer of liquid remains. The wafer substrate may be
rotated from about 300 rpm to 1000 rpm during removal of the excess
pre-wetting fluid. The wafer substrate may be rotated less than
about 20 seconds during removal of the excess pre-wetting fluid. In
other embodiment, the wafer substrate rotation rate is increased to
between about 250 and 800 rpm for between about 5 and 60 seconds,
while avoiding the complete drying of the pre-wetting fluid. While
the rotation process generally can be initiated prior to the
release of vacuum, by performing this step after the release of
vacuum, it is believed that the potential for the wafer drying is
reduced, because the evaporative drying from a thin layer and the
possibility of creating a dry surface at some point on the wafer
may be less.
[0108] After removing the excess entrained fluid from the wafer
substrate surface, the wafer substrate rotation is stopped, the
splash shield lowered and/or the wafer substrate raised (if so
equipped), the vacuum door opened, and the wafer removed from the
chamber and placed in an electroplating chamber. In some
embodiments, the pre-wetted wafer substrate is exposed to an
environment outside of the chamber and the electroplating chamber
for less than about one minute. In other embodiments, the
pre-wetted wafer substrate has a wetting layer having a thickness
of between about 50 to 500 .mu.m immediately prior to
electroplating when it is transferred to the electroplating
chamber. After the wafer substrate is in the electroplating
chamber, the wafer substrate is electroplated using a degassed
plating solution, in some embodiments. In some embodiments, the
pre-wetted wafer substrate is cathodically polarized with respect
to a plating solution before contacting the wafer substrate with
the plating solution. The pre-wetting process chamber and the
electroplating chamber may be distinct stations of one apparatus
module. In other embodiments, the wafer substrate is electroplated
in the same chamber than was used for pre-wetting. In these
embodiments, the electroplating may be performed using a degassed
plating solution.
[0109] In alternative embodiments, after removing the pre-wetted
wafer substrate from the pre-wetting process chamber, the
pre-wetted wafer substrate is transferred to a chamber configured
to perform an anodic process such as electroetching and
electropolishing.
[0110] FIG. 11b is a flow diagram for another embodiment of a
pre-wetting process (1150). A wafer substrate having an exposed
metal layer on at least a portion of its surface is provided to a
pre-wetting process chamber (1155). The pressure in the process
chamber is then reduced to a subatmospheric pressure (1160). The
wafer substrate is then contacted with a pre-wetting fluid at a
subatmospheric pressure (1165). The pressure in the process chamber
is then increased to facilitate the removal of bubbles (1170). Such
a pre-wetting process can be preformed in pre-wetting apparatus
designs described herein.
[0111] The apparatus designs and methods described herein may be
used to pre-wet a partially fabricated semiconductor device
structure. In some embodiments, a pre-wet partially fabricated
semiconductor device structure includes at least one recessed
feature. The recessed feature has a layer of metal lining the
feature. The recessed feature also includes a substantially
gas-free pre-wetting fluid filling the feature, the pre-wetting
fluid comprising an aqueous metal salt solution substantially free
from plating accelerators and levelers.
[0112] Different combinations of pre-wetting fluid compositions and
plating solution compositions can be used in a pre-wetting process
combined with an electroplating process, as described herein. FIG.
12 is a flow diagram for an embodiment of an electroplating process
1200 for electroplating a layer of copper on a wafer substrate. A
wafer substrate having an exposed metal layer on at least a portion
of its surface is provided to a pre-wetting process chamber (1205).
The wafer is then contacted with a pre-wetting fluid to form a
layer of pre-wetting fluid on the wafer substrate (1210). The
pre-wetted wafer is then contacted with a plating solution that
includes metal ions to electroplate a layer of metal on the wafer
substrate (1215).
[0113] The apparatus designs and methods described herein are
useful in various other liquid semiconductor processes and
circumstances, beyond electroplating/feature filling, where bubbles
or trapped gasses within a high aspect ratio feature may pose a
problem.
[0114] All operations described herein, including the various
wetting, pre-wetting, degassing, alignment, transfer, and plating
operations, may be configured or programmed in one or more
controllers provided on or otherwise in communication with the
described modules and systems. Any combination or sequence of such
operations, as described herein, may be programmed or configured as
such using such controller(s). Firmware, software macros,
application specific integrated circuits, shareware, and the like
may be used to implement the controller instructions.
Chemistry of the Pre-Wetting Fluid
[0115] By properly controlling the chemistry of the pre-wetting
fluid, further benefits of the pre-wetting process described herein
may be realized, including a 50% or greater reduction in the
filling time of a feature with a metal. Furthermore, the feature
filling process may start significantly more rapidly, reflected by
the fact that under similar conditions (i.e., the same conditions,
with the exception of the composition of the pre-wetting fluid),
the amount of metal deposited selectively at the bottom of the
feature in the same time is much greater. With particular organic
and inorganic additive combinations to the pre-wetting fluid, the
pre-wetting process allows for an excellent side wall and field
(field refers to wafer substrate regions that are flat and outside
of features) metal growth selectivity versus bottom of a feature
metal growth selectivity, allowing for high rate selective
deposition with greater than an order of magnitude relative plating
rate/growth at the feature bottom versus the upper side walls and
field. The selectivity achieved by controlling the chemistry of the
pre-wetting fluid allows for bottom-up, often plug-fill, growth and
the ability to rapidly fill, without voiding, high aspect ratio
features.
[0116] Historically, a number of different plating bath solutions
used to deposit copper have been used to meet various needs/goals.
Copper sulfate and copper methane sulphonate are the most commonly
used metal salts for electroplating copper, particularly in the
integrated circuit industry. The acid copper fluoroborate bath
(mixture of copper and fluoroboric acid with boric acid), with its
high solubility of copper and potential for high deposition rates,
is also used, but has largely fallen out of favor and replaced by
the methane sulphonate system (which also has high copper
solubility), at least in part because of the tendency for the
BF4.sup.- anion to decompose and form hazardous HF. Alkaline copper
cyanide and copper pyrophosphate baths have also been widely used,
with cyanide baths having generally good plating performance, but
have fallen out of favor for toxicity and stability reasons.
[0117] While the scope of this description is not limited to
electroplating of s specific metal, or to specific plating solution
and pre-wetting fluid combinations described in the examples,
electroplating of copper in plating solution baths containing
copper sulfate and/or copper methanesulphonate will be used as an
illustration of specific embodiments. It is understood that
embodiments disclosed herein can be employed for deposition of
metals other than copper, such as nickel, iron, gold, silver, tin,
lead, zinc, as well as copper and other metal co-deposited alloys
(e.g., various solders such as lead-tin and silver-tin, or magnetic
alloy materials containing iron, cobalt, and nickel). It is also
understood that in copper electroplating, a variety of other salts
beyond copper sulfate and copper methanesulphonate may be used.
[0118] Copper sulfate and methanesulphonic acid plating bath
solutions typically contain three or more materials (so called
plating "additives") in small concentrations (10 ppb to
approximately 1000 ppm) that affect the surface electrodeposition
reactions. Typically, additives include accelerators (mercapto
containing species, for example; also referred to as brighteners),
suppressors (typically polymers such as polyethylene glycol, for
example; also referred to as carriers), levelers, and halides
(e.g., chloride ion and bromide ion), each having a unique and
beneficial role in creating a copper film with desired micro- and
macro-characteristics.
[0119] The pre-wetting fluid and plating solution compositions
described herein may be used with any of the apparatus designs or
methods. For example, the pre-wetting fluid and plating solution
compositions are able to be used with the methods described in
FIGS. 11a, 11b, and 12.
[0120] There are several different categories of process
interactions which should be considered in selecting the optimal
pre-wetting fluid for a wafer substrate. These various issues are
discussed herein, along with the presumed or measured examples of
their effect on feature filling.
[0121] One consideration is that the surface tension of the
pre-wetting fluid should be sufficiently compatible with the wafer
substrate surface (e.g., hydrophilic) so that the entire surface
remains covered with pre-wetting fluid from the period after the
pre-wetting fluid is applied to the surface under vacuum and the
time the wafer is moved to and is immersed into the plating bath.
In some embodiments, the pre-wetting layer, just before immersion
into the plating solution, is thin (e.g., about 50 to 500 .mu.m
thick) and uniform. By the film being thin, the amount of
concentration increase or dilution/modification of the plating bath
concentrations is kept small, and the film has a minimal delay in
adsorption of plating additives to the general plating surface
(i.e., field region). By the film thickness being uniform, the
uniform transition from the state of being covered with a solution
of pre-wetting fluid composition to plating bath solution
composition is possible and much more easily controllable.
[0122] Another consideration is that when transferring the wafer
from a pre-wetting station to a plating station, the features are
filled and the general surface is coated with the pre-wetting
fluid. During the ensuing time between the initial exposure of the
surface to the plating solution and the initiation of plating,
various unfavorable reactions with the constituents of the
pre-wetting fluid, either alone, or in combination with gasses,
coming from the atmosphere, may occur. By degassing the pre-wetting
fluids (e.g., by using a degasser, as described herein), those
reactions that involve dissolved gasses may be reduced or
eliminated. Still, if and when the liquid surface layer of the
pre-wetted wafer is exposed to air, gas re-adsorption into the
degassed pre-wetting fluid will occur (e.g., after 15 seconds or
more), and may lead to deleterious corrosion or other effects.
Alternatively, with an appropriate choice of components included in
the composition of the pre-wetting fluid and/or timely/rapid wafer
transfer to the plating cell, such reaction and effects may be
reduced or altogether avoided.
[0123] In general, a reaction between the pre-wetting fluid and the
seed layer on a wafer results from the existence of a chemical
driving force (i.e., a negative free energy for reaction) with
moderate activation energies. Eliminating the driving force, or
inhibiting the kinetics, can forestall a deleterious reaction. The
reactions involve a combination of one or more solvents (e.g.,
water, alcohols, carbonates, or ketones), pre-wetting fluid solutes
(e.g., acids, inorganic salts, organic electrolytic or neutral
plating additive species), and dissolved gasses. An example of a
particularly deleterious reaction is the corrosion reaction of the
metal seed layer. The seed corrosion rate will depend, for example,
on parameters such as the pre-wetted wafer transfer time, the
temperatures of the pre-wetting fluid and plating solution bath,
the choice of pre-wetting solvent(s), the pH of the pre-wetting
fluid, the particular dissolved constituents in the pre-wetting
fluid, and any spatial and time-varying distribution or
redistribution (i.e., concentration difference due to the diffusion
into or out of the features) during the initial immersion of the
wafer into the plating solution bath. These different reactions are
described herein.
[0124] Any electrolytic reaction for the corrosion of a metal can
be represented as two half-reactions, coupled by the transfer of an
electron in the metal. For example, the reduction of oxygen or
other oxidizing agent in the solvent (the element that is reduced)
couples with the oxidation of copper metal. The reaction of copper
metal with oxygen occurs in two steps, to cuprous ion, and
depending on the solvent environment, presence of complexing
agents, and pH, to cupric ion.
Cu.fwdarw.Cu.sup.++e-.fwdarw.Cu.sup.+2+e- (8)
[0125] The reduction reactions of oxygen, written for either acid
or alkaline conditions, are
O.sub.2+4H.sup.++4e.sup.-.fwdarw.H.sub.2O (9a)
O.sub.2+2H.sub.2O+4e.sup.-.fwdarw.4OH.sup.- (9b)
[0126] Using an oxygen free pre-wetting fluid prevents reaction 9a
or 9b from occurring altogether, and hence the corrosion of copper
is prohibited from this source. Therefore, removal of oxygen from
the pre-wetting fluid is desirable in some embodiments. However, if
oxygen is reintroduced into the electrolyte from the environment
(e.g., during a transfer from the pre-wetting location to the
plating solution), reaction 9a or 9b is again allowed to occur.
Similarly, if the supply of protons is small (e.g., a pH of greater
than about 3), reaction 9a will be reduced.
[0127] Referring to an example situation depicted in FIG. 13, a
structure 1301 in a wafer substrate 1302 consists of a cavity
filled with pre-wetting fluid 1303. The feature surface 1305, walls
1306 and bottom 1307 of the structure are typically coated with a
barrier layer (not shown) below an electroplateable "seed layer"
(e.g., copper, 1304). The thickness of the metal along the walls,
particularly on the lower wall 1308, is typically much thinner than
that on the surface 1305 (and often also thinner than that at the
feature bottom 1307) due to the nature of the seed deposition
process (e.g., PVD). Initially, a degassed pre-wetting fluid 1303
is introduced to the surface under vacuum that contains no bubbles
and little or no dissolved gas (e.g., oxygen). However, some gas
may subsequently be introduced into the liquid from the atmosphere
during the transfer of the wafer, with a near saturated condition
created at the exposed liquid layer surface 1308. With a much
shorter diffusion distance and resistance to arriving at the
surface 1305, oxygen reduction reaction 9a will preferentially
start to occur there first. While reaction 8 may occur anywhere
along the surface, the reaction may preferentially occur at point
on the surface where it is roughest and its effect is most
detrimental where the film is thinnest (with a potential for loss
of all seeded metal). Also, the metal corrosion half reaction may
occur preferentially at locations where the oxygen reduction
reaction is not co-occurring, such as at deep inside the feature
(e.g., at locations 1307 and 1308). The overall reaction is
completed by having the electrons created by reaction 8 inside the
feature traveling through the metal along the wall to the feature
top and field, where they combine with the oxygen via reaction 9a
or 9b. The walls may be rough on a microscopic scale due to the
feature creation method (e.g., from a repetitions application of a
SF.sub.6 isotropic RIE etch/C.sub.4F.sub.8 passivation sequence
known as an Advance Silicon Etch or the "Bosche" process) and/or
the deposition process. Rough metal surfaces tend to have a higher
local electrochemical activity, so corrosion will be greater at
rough metal surfaces than for a uniform smooth surface. These
phenomena will increase the driving for metal loss from these rough
metal surfaces. See, for example, a discussion of these phenomena
in U.S. Pat. No. 6,946,065.
[0128] In some embodiments, a substantially non-conductive (i.e.,
non-ionic and electrolyte free) solvent can be effectively used for
pre-wetting fluid for a pre-wetting process performed under vacuum.
This is despite factors that otherwise would lead one to avoid the
use such a fluid. One such conceptually negative factor is that the
conductivity of such a pre-wetting fluid is quite small. At the
time immediately after immersion of a wafer into a plating bath
solution, deposition at the bottom of the feature filled with a
non-conductive or low conductivity solvent is expected to be
hindered by the inability to support plating because of its
inability to support ionic current flow. Another potentially
detrimental factor is the formation, after entry of a wafer into an
electroplating bath, of a potential and the establishment of an
internal corrosion cell due to the different activities of
dissolved metals at the wafer surface and within the feature. The
electrochemical potential difference in the solution, between the
bottom of the feature and the top of the feature can be expressed
by a form of the Nernst equation:
.DELTA. V = RT nF ln C ( feature ) C ( surface ) ( 10 )
##EQU00002##
In equation 10, R is the universal gas constant, T is the absolute
temperature, n is the number of electrons for the corrosion
reaction, F is Faraday's constant, and C(feature) and C(surface)
are the concentrations of metal ions at the two locations. A
concentration cell is created, with the corrosion driving potential
created by a difference in concentrations as given by equation 10.
When using a pre-wetting fluid free of dissolved metal ions, the
bottom of a feature will encounter a C(feature) concentration that
is smaller than the C(surface) for some period of time after
immersion into the plating bath containing metal ions. Therefore, a
corrosive potential difference will exist between location the
bottom of a feature and the surface, with the corrosion potential
causing the metal on the walls and bottom of the feature to
preferentially oxidize, release electrons, and complete the cycle
by combining with the metal ions from the solution at the
surface.
[0129] Specifically, the reaction
Cu.sup.++=.fwdarw.2e-+Cu (11)
will occur at the surface region, and will be coupled with the
oxidation reaction
Cu.fwdarw.2e-+Cu.sup.++ (12)
occurring on the walls and lower surface at the bottom of the
feature. To avoid this undesirable process when using this type of
pre-wetting fluid, it is important to establish a cathodic
(plating) polarization of the wafer surface versus the plating
solution prior to, or very shortly after, immersion of the wafer
surface into the plating solution. (See U.S. Pat. Nos. 7,211,175,
6,562,204 and 6,551,483, which deal with entry cathodic protection
and potentiostatic entry, herein incorporated by reference in
relevant part.) This is accomplished by applying a cathodic
potential difference or a small cathodic current between the wafer
and the solution before wafer bath entry. Alternatively, or
additionally, in some embodiments, a rapid rinse of the wafer
surface with a solution having a relatively low metal ion
concentration can be used (e.g., DI water), followed by a high
speed spin or other method to remove the surface solution. This
process reduces the concentration of the metal at the surface
relative to that in the feature, but also removes electrolyte from
the wafer edge, reducing the propensity for that electrolyte to be
plated on the edge of the wafer and plating apparatus contacts
(e.g., when plating in a closed or seal contact "plating cup"). As
another alternative, the metal ion concentration in the pre-wetting
solution can be at least equal to or greater than that of the
subsequent plating bath.
[0130] Examples of embodiments of pre-wetting fluids of the
substantially non-conductive class are an electrolyte-free
isopropyl alcohol or other water soluble non-aqueous solvents
(i.e., water-miscible solvents). Other embodiments include
alcohols, a dialkylcarbonate, dimethylformamide, and dimethyl
sulfoxide. Another embodiment is a solution of water containing a
small concentration of the non-metal-complexing tetramethylammonium
sulfate and/or tetramethylammonium hydroxide, in the pH range
between about 3.5 and 11.5. Still another embodiment is a solution
of water containing a surfactant such as the anionic surfactant
laurilsulfate (with an alkali metal cation or tetramethylammonium
cation). Pre-wetting fluids that have a reduced surface tension
compared to water, relatively small conductivities (e.g., compared
to acids or strong bases), and are non-copper complexing are used
in some embodiments.
[0131] In some embodiments of an electroplating process 1200 for
electroplating a layer of metal on a wafer substrate shown in FIG.
12, a wafer substrate having an exposed metal layer on at least a
portion of its surface is provided to a pre-wetting process chamber
(1205). The wafer is then contacted with a pre-wetting fluid to
form a layer of pre-wetting fluid on the wafer substrate (1210).
The pre-wetting fluid includes a water-miscible solvent. The water
miscible solvent may be an alcohol, ketone, dimethylcarbonate,
diethylcarbonate, dimethyl sulfoxide, or dimethyl formamide. The
pre-wetted wafer is then contacted with a plating solution that
includes metal ions to electroplate a layer of metal on the wafer
substrate (1215). In some embodiments, the plating solution
includes copper ions to electroplate a layer of copper on the wafer
substrate.
[0132] To elaborate, because the half oxidation reactions (e.g.,
reactions 8 and 9) must complete an electrical circuit and
therefore pass ionic electrical current between the two locations
where oxidation and reduction reactions are occurring, in some
embodiments it is advantageous to use a pre-wetting fluid that has
a small ionic conductivity, such as a low conductivity solvent that
itself will not react directly with copper, or a solvent
substantially free of ionically disassociated and conductive ions
(i.e., dissolved acid, bases, and salts). Many water soluble solute
free solvents, such as DI water, isopropyl alcohol, ethylene
glycol, propylene glycol, propylene carbonate, etc., have high
electronic resistances when free of solutes, and the solubility of
the cupric or cuprous ion in a neutral pH solvent in the same are
also generally very small. Because of these factors, corrosion of
the metal in these solvents can only occur by the direct oxidation
with dissolved oxygen, generally a very slow process at ambient
temperature and oxygen concentrations:
Cu+1/2O.sub.2.fwdarw.Cu.sub.2O (13)
[0133] Therefore, performing pre-wetting using ionic-solute-free
solvents (i.e., a non-electrolytic solution) such as water or DI
water is one embodiment, and using deoxygenated deionized solvents
such as deoxygenated DI water is another embodiment (to avoid
reaction 13). In some embodiments, the plating solution is also
deoxygenated/degassed prior to and during contact/exposure to the
wafer surface, and a potential or current is applied to the wafer
prior to entry, establishing a voltage greater than that given by
equation 10, thereby preventing reaction 12 within the feature from
occurring. With respect to preventing side wall corrosion,
pre-wetting fluids that contain non-ionic dissolved species (e.g.,
non-ionic surfactants or switter-ionic surfactants, added to lower
surface tension, or organic materials added to react with oxygen)
are useful in some embodiments over highly conductive ionic solutes
such as acids and bases. This is because of the generally lower
solution specific conductance and ion current coupling of the
oxidation and corrosion half reactions. An exception to this is the
addition of surface adsorbing electrochemically active non-ionic
materials (e.g., a non-ionic leveler compounds). A further example
of an unfavorable pre-wetting fluid combination for copper plating,
in some embodiments, is polyethylene glycol or
polyethylene/polypropylene oxide copolymer (which are known to act
as plating "suppressors") at all but very low concentrations
combined with small amount of dissolved halide (e.g. chloride)
ions. A suppressor without the adsorption and electrochemical
activity enhancing halides appears to be unfavorable, in some
embodiments, unless at very low concentrations.
[0134] In one experiment performed according to an embodiment
described herein, a 60 .mu.m deep/10 .mu.m wide TSV via structure
having a 8000 .ANG. copper seed layer was electroplated with
copper. The feature was pre-wetted with deoxygenated, deionized
water. After the wafer was exposed to the atmosphere for 5 minutes,
it was then transferred to a plating cell, followed by immersion
into the plating solution. The plating solution was a deoxygenated
plating bath sold under the trademark DVF 200.TM. by Enthone Inc.
with added components (DVF 200.TM. is a copper methane
sulfonate/methane sulfonic acid plating solution to which
accelerators, suppressors, and leveler additives, and 50 ppm
chloride ions, were added). The filling characteristics of multiple
individual features across multiple runs using this method showed
complete void-free feature filling in almost all cases. In some
experiments, the wafers were cathodically polarized prior to entry
into the plating solution. This result shows the process robustness
for creating void-free bottom-up filling using a combination of a
degassed DI water pre-wetting process performed under vacuum,
followed by potentiostatic entry into a copper plating
solution.
[0135] In other embodiments, substantially non-conductive
pre-wetting fluids that contain some dissolved compounds other than
metals (e.g., either electrolytic or non-ionic, organic or
inorganic, added, for example, in relatively small quantities to
reduce surface tensions and aid in wetting), but are substantially
free of materials that are electrochemically active and/or are
considered plating bath additives, are used. For example, in some
embodiments, a pre-wetting fluid that is substantially free from
any accelerator/brighteners or levelers (that might typically be
found in a subsequently used plating bath) is used rather than
those that do contain such electrochemically active agents.
[0136] In one experiment performed according to an embodiment
described herein, a wafer with a 60 .mu.m deep/10 .mu.m wide TSV
via structure having a 8000 .ANG. copper seed layer was
electroplated with copper. The feature was pre-wetted with a
pre-wetting fluid containing copper methane sulfonic acid (copper
salt, 80 g/L copper ion), 20 g/L methane sulfonic acid, 50 ppm
chloride ion, and 3 or 12 ppm of the copper plating accelerator
dimercapto-propane sulfonic acid (SPS). After pre-wetting, the
wafers were exposed to the atmosphere for about 1 minute, and then
transferred to a plating cell, followed by immersion into the
plating solution. The plating solution was a deoxygenated plating
bath sold under the trademark DVF .sup.200TM by Enthone Inc. with
additive components. Copper was then plated onto the wafer. In both
cases (i.e., pre-wetted with a solution containing 3 ppm of
dimercapto-propane sulfonic acid and 12 ppm of dimercapto-propane
sulfonic acid), sidewall voids were formed.
[0137] In some embodiments, the pre-wetting fluid includes water
and a copper salt. This helps to avoid corrosion of a seed layer
due to setting up the electrochemical difference discussed in
relation to equation 10. In certain embodiments, the copper salt is
at a concentration of at least about 50% of a saturation limit. In
certain embodiments, the copper salt is copper sulfate, a copper
alkylsupphonate, and mixtures thereof. In certain embodiments, the
copper salt is at a concentration greater than about 20 g/L of
copper. In some embodiments, after pre-wetting the wafer substrate
with a pre-wetting fluid that includes water and a copper salt, the
pre-wetted wafer substrate is electroplated with copper with a
copper-containing plating solution; the pre-wetting fluid contains
a copper salt at a copper concentration that is the same or higher
than the copper concentration in the plating solution. In some
embodiments, the copper concentration in the pre-wetting fluid is
at least about 25% greater than the copper concentration in the
plating solution. In other embodiments, the pre-wetting fluid
consists essentially of water and a copper salt.
[0138] In some embodiments, a pre-wetting fluid having the same or
a very similar composition to that of the plating solution (i.e., a
solution having the same metal salts and/or the same metal ion
concentrations, the same acids and/or the same acid
concentration(s), the same halides and/or the same concentrations
of halides, and the same additives and/or the same concentrations
of additives) is used. In embodiments when the pre-wetting fluid
and the plating solution have the same composition, a layer of
metal may be plated on the wafer substrate in the same chamber as
used for pre-wetting. However, when a seed layer is marginal (e.g.,
rough and thin within the feature), such a pre-wetting fluid (i.e.,
the same or very similar to the plating solution) may be
susceptible to feature fill voiding due to seed corrosion from the
pre-wetting fluid. The feature rate of filling may also be improved
by using a different solution for pre-wetting than that of the
plating bath, as described herein.
[0139] In one experiment performed according to an embodiment
described herein, a 60 .mu.m deep/10 .mu.m wide TSV via structure
having a 8000 .ANG. copper seed layer was electroplated with
copper. The feature was initially pre-wetted with a plating
solution (i.e., the pre-wetting fluid had the same composition as
the plating solution). The feature/wafer was pre-wetted via a
pre-wetting process performed under vacuum described herein with
commercially available deoxygenated plating bath sold under the
trademark DVF 200.TM. by Enthone Inc., with plating additive
components (i.e., the plating additive components used with DVF
200.TM. in the other experiments described herein). The surface was
exposed to the deoxygenated plating bath and was then exposed to
the atmosphere for either 1 or 3 minutes between release of vacuum
and the transfer/immersion into the plating bath and initiation of
metal deposition. The wafers were cathodically polarized
immediately upon entry into the plating solution. In one case where
the surface was exposed to the atmosphere for 1 minute, the feature
was filled with metal and void free, with no evidence of side wall
corrosion. However, features from the same wafer show that some
features are not filled, typically with one side of the feature
having an irregularly shaped void. This is generally believed to be
associated with the loss of seed metal at the side of the feature.
For a wafer prepared and processed in exactly the same fashion,
except that there was 3 minutes exposure to atmospheric between
release of vacuum and the initiation of plating, the feature
filling was grossly incomplete. In many cases, the entire bottom of
the feature was unplated. A similar trend (i.e., transition from
void free to significant side wall voiding) also occurs for a fixed
atmospheric exposure time, but with decreasing seed layer
thickness. Therefore, the use of the plating solution as a
pre-treatment solution is less than optimal in some embodiments
because of its significant sensitivity to incomplete feature
filling due to side wall corrosion. Particularly in situations
where the seed layer thickness is quite thin, the number of side
wall void-type defects increased markedly in both situations,
indicating a narrow tolerance of seed layers to this pre-wetting
fluid.
[0140] Referring back to reactions 8 and 9, the metal ion created
by the coupling of reaction 8 with reaction 9a or 9b must be able
to pass electrical (ionic) current back to the surface though the
fluid, so having a solution of substantial conductivity is an
unfavorable pre-wetting fluid property, in some embodiments. This
is in contrast the substantial conductivity that is generally
desired in a plating solution, where conductivity is tailored so as
to minimize voltage drops in the solution and within the feature to
facilitate the deposition process. Of particular interest is the
high ionic mobility of protons, which is the highest of any cation.
This property tends to impart very high conductivities to acid
solutions of given molarities. Therefore, as a general rule,
pre-wetting fluids having highly dissociated acids of high
concentrations (e.g., those that generate a pH of less than about 2
or create greater than about 0.01 molar of free protons) are not
favored in some embodiments because they facilitate corrosion
reactions due to their high conductivities. Under such acidic
conditions, the metal at the bottom of the feature's wall 1308
(FIG. 13) is subject to unfavorable conditions and could
potentially result in the area corroding and producing a side wall
without an electroplateable seed layer.
[0141] As a primary consideration, it is desirable to avoid this
inside-the-feature corrosion and feature filling defects such as
voiding. The combination of the high conductivity, acidity, and
potentially unfavorable adsorption and reaction of additives and
halide(s) with the metal can lead to, for example, feature sidewall
corrosion and filling defects, as well as inhibiting or delaying
the establishment of a favorable distribution of the additives on
the various surfaces required for optimal feature filling rates or
void free filling. Because the metal on the side wall can be thin
and oxidized prior to the pre-wetting fluid exposure, a corrosion
reaction involving acid or other components can lead to the loss of
all the platable metal, leaving a non-plateable metal such as the
copper diffusion barrier tantalum or tantalum nitride with an
exposed oxide layer thereupon. Long unpolarized (cathodically
protected) exposure of the surfaces to inappropriate component
mixtures of the pre-wetting fluid may therefore lead to poor
feature filling. Contrary to using a strongly acidic electrolyte
(pH of less than about 2), using more neutral or near neutral
pre-wetting fluid can limit the supply of protons for reaction 9a,
reducing the corrosion rate, reducing the defectivity, and
generally improving the reliability and success of the overall
pre-wetting. Note that pre-wetting fluids of this description would
generally not be optimal or acceptable for copper metal deposition,
but they are favored for pre-wetting, in some embodiments.
Solutions in the pH ranges of about 2 to 12, free of dissolved
metal ion complexing anions, do not allow reactions such as 8 and 9
to occur at appreciable rates.
[0142] In some embodiments, the pre-wetting fluid includes
deionized water, an acid, and a copper salt, with the pre-wetting
fluid pH not lower than about 2. In further embodiments, the pH of
such a pre-wetting fluid is between about 2 and 4. The acid in such
embodiments may be sulfuric acid, an alkylsuphonic acid, and
mixtures of these acids. The pre-wetting fluid may also include
less than about 2 g/L or sulfuric acid or methane sulfonic acid in
some such embodiments. In other embodiments, the pre-wetting fluid
consists essentially of water, an acid, and a copper salt, with the
pre-wetting fluid pH greater than about 2. In yet other
embodiments, the pre-wetting fluid includes water and an acid, and
the pre-wetting fluid has a pH of greater than about 2.
[0143] According to the various published pH/potential stability
diagrams and calculations (also known as Pourbaix diagrams),
exposure of copper metal to a non-complexing electrolytic solution
of about pH 3 or greater and further having an oxidizing source
(such as dissolved oxygen) is expected to form a metal surface
oxide. The formation of the oxide instead of a dissolved metal salt
of copper may inhibit further oxidation. It is thermodynamically
favored for cuprous ions formed at the interface to react with
water or hydroxide directly to form cuprous oxide or hydroxide
(rather than form dissolved cuprous or cupric salts).
2Cu+4OH.sup.-.fwdarw.Cu.sub.2O+H.sub.2O+4e.sup.- (15a)
Cu+2OH.sup.-.fwdarw.Cu(OH).sub.2+2e.sup.-.fwdarw.CuO+H.sub.2O+2e.sup.-
(15b)
[0144] At a very high pH, the hydroxide of copper is slightly
soluble, so this condition is slightly unfavorable from this
prospective. The coupling of the copper oxidation half reaction
with the oxygen reduction reaction may be reduced in a neutral
solution, and so pre-wetting fluids free of copper complexing
agents in a pH range of from about 2 to 12, more particularly
between about 3.5 to 10.5, are useful classes of pre-wetting fluids
for use under vacuum. Pre-wetting fluids of this class are
solutions that may contain some dissolved compounds (both
electrolytic and non-ionic, organic or inorganic, for example, to
reduce surface tensions and aid in wetting), but are substantially
free of materials that electrochemically alter the plating of the
metal and active and/or are considered plating bath additives. The
presence of a copper complexing agent also changes conditions,
allowing the formation of a complex instead of the passivating
oxide/hydroxides; if oxygen is present, unfavorable high rate
corrosion is expected in metal complexing agent containing
solutions having a dissolved oxidizer. Some materials that are
typically bath additives can form metal complexes, such as
mercapto-group containing brighteners/accelerators (e.g.,
mercapto-propanesulfonic acid, di-mercaptopropane sulfonic acid,
etc.) and various nitrogen group containing levelers (e.g., diazine
black and Janus Green B). For example, a pre-wetting fluid that
does not contain any brighteners or levelers that might typically
be found in a subsequently used plating bath may avoid related
pre-wetting seed metal corrosion. Suppressors such as polyethers
(e.g., polyethylene glycol, polypropylene oxide, etc.) or metal ion
complexing agents by themselves are not particularly corrosive, and
since they tend to reduce the surface tension as wetting agents,
can be added when high rate fill is not a primary concern. However,
the addition of suppressor in combination with chloride ions,
generally considered a necessary co-constituent to achieve
suppressor electrochemical activity, is not favored in some
embodiments.
[0145] In some embodiments, the pre-wetting fluid may aid in
removing the oxide surface. In some embodiments of an
electroplating process 1200 for electroplating a layer of metal on
a wafer substrate shown in FIG. 12, a wafer substrate having an
exposed metal layer on at least a portion of its surface is
provided to a pre-wetting process chamber (1205). The pressure in
the pre-wetting process chamber is then reduced to a subatmospheric
pressure (not shown). The wafer is then contacted with a
pre-wetting fluid to form a layer of pre-wetting fluid on the wafer
substrate (1210). In one embodiment, the pre-wetting fluid includes
an acid to at least partially remove surface oxide from the seed
layer and the pre-wetting fluid has a pH of between about 2 to 6.
The pre-wetted wafer is then contacted with a plating solution that
includes metal ions to electroplate a layer of metal on the wafer
substrate (1215). The plating solution has a pH of between about 2
to 6 and the plating solution and the pre-wetting fluid have
different compositions.
[0146] In other embodiments, the pre-wetting fluid may aid in
transforming a metal oxide laden surface to a metallic surface
(e.g., cuprous or cupric oxide--see reactions 6 and 7 and the
associated discussion) or in removing the oxide surface. In some
embodiments of an electroplating process 1200 for electroplating a
layer of metal on a wafer substrate shown in FIG. 12, a wafer
substrate having an exposed metal layer on at least a portion of
its surface is provided to a pre-wetting process chamber (1205).
The wafer is then contacted with a pre-wetting fluid to form a
layer of pre-wetting fluid on the wafer substrate (1210). In one
embodiment, the pre-wetting fluid includes a small amount of a
reducing agent to at least partially reduce surface oxide on the
seed layer. In another embodiment, the pre-wetting fluid includes a
metal complexing agent to at least partially remove surface oxide
on the exposed metal layer and the pre-wetting fluid has a pH of
between about 4 to12. The pre-wetted wafer is then contacted with a
plating solution that includes metal ions to electroplate a layer
of metal on the wafer substrate (1215).
[0147] In some embodiments, the plating solution includes copper
ions to electroplate a layer of copper on the wafer substrate. In
such embodiments, the exposed metal layer on the wafer substrate is
generally copper or a copper alloy. Examples of reducing agents for
copper include formaldehyde, glycolic acid (and salts thereof), and
dimethylamine borane. When the exposed metal layer is copper, the
pre-wetting fluid includes, in some embodiments, a copper
complexing agent to at least partially remove surface oxide on the
exposed copper layer and the pre-wetting fluid has a pH of between
about 4 to12.
[0148] In general, small concentrations (e.g., parts per million,
typically 10 to 100 ppm) of halide ions such as chloride or bromide
are present and often critical in many plating bath solutions.
Halides are also well known corrosion agents. It is generally known
that a solution containing halides will corrode a surface faster
than an identical solution (i.e., consistent pH and ionic strength)
without halides. Because they are critical in successful plating
and their concentrations are low, one might assume that not having
them present in the pre-wetting fluid would inhibit their uniform
exposure to the features internal surfaces and thereby have a
detrimental effect to the feature filling process. In some
embodiments, however, it is useful to not include or add even these
very low levels of halides to the pre-wetting fluids. In some
embodiments, the pre-wetting fluid is substantially free from
halides. Even with low, parts per million levels of halides, alone
or in combination with other plating bath additives, dramatic
increases in the corrosion rate of the metal on side wall features
have been observed. While not wanting to be held to any
particularly theory, corrosion of the metal, as a whole, is perhaps
catalyzed or stabilized by the formation of cuprous halide
reactants.
[0149] In one TSV feature filling experiment, similar to the other
experiments described herein, performed according to an embodiment
described herein, the pre-wetting fluid contained 100 g/L copper
methane sulfonic acid, 16 g/L methane sulfonic acid, and either no
chloride or 50 ppm of chloride ions. The copper plating was then
performed using the same solution and processes as described for
the other experiments described herein. Side wall seed layer
corrosion was greatly exacerbated by even small quantities of
halide ions in the pre-wetting fluid
[0150] Plating bath suppressors include polyethylene glycol (PEG),
polypropylene glycol (PPG), polyethylene oxide (PEO), polypropylene
oxide (PPO), and various copolymers of these monomers. A suppressor
is used to suppress copper plating outside the features on a wafer,
allowing for copper deposition inside such features. They are also
good surface tension reducing agents (surfactants), and hence,
might be considered to be useful component in the pre-wetting
fluid. As noted herein, the suppressive plating characteristics of
these compounds are generally derived in combination with halides,
and the presence of halides can cause feature side wall corrosion.
In some embodiments, the pre-wetting fluid is substantially free
from halides, plating accelerators, and plating levelers, and
includes a plating suppressor at a small concentration (e.g.,
typically less than about 15 ppm).
[0151] An experiment was performed using pre-wetting fluid
containing 100 g/L copper methane sulphonate, 16 g/L methane
sulfonic acid (sometimes referred to as virgin makeup solution or
"VMS"), no chloride ions, and various amount of an 8000 molecular
weight polyethylene glycol to determine the effects of suppressors
on a pre-wetting fluid. Side wall corrosion was generally not
observed in these chloride-free pre-wetting fluids. Feature
filling, however, was significantly impacted as the concentration
of the suppressor reached somewhere between about 5 parts and 25
parts per million. The fill characteristics transitioned from a
bottom-up filling at from about 0 to 5 ppm of suppressor to a
bottom void at about 25 ppm suppressor. At about 50 ppm of
suppressor and above, the plating was largely conformal. Thus, the
use of above about 15 ppm of a suppressor is undesirable from a
feature filling prospective in some embodiments.
[0152] In some embodiments of an electroplating process 1200 for
electroplating a layer of copper on a wafer substrate shown in FIG.
12, a wafer substrate having an exposed metal layer on at least a
portion of its surface is provided to a pre-wetting process chamber
(1205). The wafer is then contacted with a pre-wetting fluid to
form a layer of pre-wetting fluid on the wafer substrate (1210).
The pre-wetting fluid includes water and copper ions, and is
substantially free of plating additives. In this embodiment, the
plating solution includes plating additives. The concentration of
copper ions in the pre-wetting fluid is greater than the
concentration of copper ions in the plating solution. In some
embodiments, the pre-wetting fluid is substantially free of
additives, including halides, accelerators, and levelers, and
combinations thereof. In some embodiments, the pre-wetting fluid
includes polyethylene oxide at a concentration of less than about
15 ppm. In some embodiments, plating solution additives include
halides, accelerators, suppressors, and combinations thereof. The
pre-wetted wafer is then contacted with a plating solution that
includes copper ions to electroplate a layer of copper on the wafer
substrate (1215).
[0153] In addition to the possibility of a particular pre-wetting
fluid either aiding or hindering the feature filling process (e.g.,
the avoidance of side wall corrosion or creating a conformal type
filling behavior), there is also a relationship between the
composition of the pre-wetting fluid and the rate of feature
filling. In experiments performed to compare the rates of feature
filling, the plating bath solution composition and plating currents
versus time were fixed and the amount of feature filling at the end
of the process was monitored. The experiments showed that the
choice of pre-wetting fluid can have a dramatic effect on the
feature filling rate and time, sometimes increasing the filling
rate and reducing the filling time by a factor of two or more.
[0154] While not wanting to be held to any particular explanation
or model for this effect, it is believed that having a conductive
electrolyte (e.g., as opposed to, for example, DI water) that
principally contains significant amount of the metal salt is
necessary to rapidly initiate and sustain plating at the bottom of
a feature. In some embodiments, the pre-wetting fluid should not
include some or all of the plating bath additives which are
necessary in the field to suppress plating thereupon (i.e.,
levelers and suppressor). In some embodiments, the pre-wetting
fluid is substantially free from plating levelers. The addition of
accelerator to the pre-wetting fluid is useful in some embodiments,
but accelerators (such as dimercaptopropane sulfonic acid, SPS)
tend to bifurcate and form a very strongly adsorbing accelerator
monomer (e.g., mercaptopropane sulfonic acid, MPS) simply by
exposure of the surface to the pre-wetting fluid. This bifurcation
of accelerator may be sufficiently fast that, prior to initiating
plating, the entire surface is saturated with the accelerator
adsorbate. Therefore, upon starting the subsequent plating, the
field and upper side walls may require the accelerator to be
removed or deactivated in order to cause the current to be driven
into the intricacies of the features. In contrast, a pre-wetting
fluid that contains metal ions, but no accelerator or other
additives, will plate at a high rate upon entry in the bath until a
suppressing additive become surface-active. In some embodiments,
the pre-wetting fluid is substantially free from one or more of
halides, plating accelerators, and plating levelers.
[0155] The smaller accelerator can diffuse quickly from the plating
bath into the mouth and lower regions of a feature, while the
suppressor and leveler molecules will diffuse slower and initially
act primarily on the upper side walls of the feature, hence
creating polarization contact and driving the current into the
feature. The accelerator molecule primarily serves the purpose of
deactivating, removing, or preventing the polarization developed by
the suppressor/halide combination. An accelerator in itself is only
a weak polarizing agent with respect to an additive free solution
that is free of the polarizing suppressor/halide combination. While
some halides, such as chloride, may be relatively small and diffuse
with similar alacrity to the accelerator molecule, without either
suppressor and leveler present at the feature bottom, the surface
kinetics will still be fast and the plating resistance very low
(halides alone are generally non-polarizing, and in fact some
literature says they are depolarizing, by themselves). Furthermore,
the ability to have the electrochemical conversion of SPS to the
strongly adsorbing MPS accelerator molecule occur for some duration
of time in an local environment free of suppressor (i.e., before
any suppressor can arrive sets up an "inertial" plating condition
that tend to reduce the propensity for the suppressor later absorb)
avoids polarization and increases the relative plating to the
bottom of the feature. In contrast, at the upper walls and field of
the feature, very shortly after exposure to the plating solution,
suppressor is adsorbed, and leveler competes with the adsorption of
accelerator to remove the polarization there, so polarization
develops there very rapidly. As time progresses after immersion
into the plating bath, the small molecules such as the accelerator
and halide will diffuse rapidly into the feature, but the larger
levelers and much larger suppressor will diffuse much more slowly,
delaying their inhibitive effect inside and allowing for a fast
fill
[0156] Depending on the feature size, quality of seeding inside the
feature, various processing costs, and other goals, one or more
pre-wetting fluids may be favored over another, in certain
embodiments. Tables 1 to 4 are based a large number of feature
filling experiments and observations, similar to those described
herein, that qualitatively categorized and compared the propensity
of feature corrosion and enhancement/retardation of feature filling
(filling rate) for a number of different pre-wetting fluid
combinations. The table term "EXCELLENT" indicates a generally
highly favorable result (e.g., little evidence of seed corrosion or
enhanced or otherwise high feature filling rate). The table term
"GOOD" indicates a potentially acceptable result, though one that
may not be optimal in all cases (e.g., depending on seed quality,
plating baths, etc.). The table term "FAIR" includes performances
that are typically quite marginal or may be unreliable, and may
often lead to negative or poor results. Finally, the table term
"POOR" indicates almost invariably unacceptable seed corrosion or a
significantly altered (e.g., conformal filling) or diminished
filling rate behavior.
[0157] Results for different concentrations of acids are given.
Results on metals of either sulfate or methane sulfonic acids are
also given, although the difference between the two was generally
found to be minimal. In tables where the labels "sulfuric or
methane sulfonic" acid are given, the additional component creates
a mixture of the two types of acids (not just more acid of the same
chemical). For example, in Table 2, all of the pre-wetting fluids
included greater than 2 g/L or either sulfuric acid or methane
sulfonic acid. The pre-wetting fluid in row 1 [i.e., >2 g/L
Methane Sulfonic (or Sulfuric) Acid] contained BOTH 2 g/L of
sulfuric acid AND 2 g/L of methane sulfonic acid. The pre-wetting
fluid in row 2 [i.e., <2 g/L Methane Sulfonic (or Sulfuric)
Acid] contained EITHER 2 g/L of sulfuric acid OR 2 g/L of methane
sulfonic acid. Sulfuric acid has about the same molecular weight as
methane sulfonic acid, so the concentrations are about the same in
the two cases, but it is recognized that sulfuric acid is both
diprotic and has a different dissociation constant
(H.sub.2SO.sub.4: MW=98, pKa.sub.1=3.0, pKa.sub.2=2;
CH.sub.3SO.sub.3H: MW=96, pKa=1.9), so the pH of a solution
containing the same amount of sulfuric acid will be lower. Finally,
the concentrations of the copper solutions from different salts are
in grams per liter of cupric ions (Cu.sup.++), not of the anhydrous
salt or hydrated salt.
[0158] From these tables a number general trends of good
pre-wetting fluids can be identified, specifically those containing
little or no acid (pH of 2 or greater), moderate to high metal ion
concentrations, little or no halides (<10 ppm), less than about
15 ppm PEG like suppressors, and neither leveler or accelerator
plating additive. A solution containing from 20 to 100 g/L metal
ion with no other components other than solvent (water) and a small
concentration of a surfactant (or no surfactant) is an example of a
good pre-wetting fluid composition.
TABLE-US-00001 TABLE 1 Pre-wetting fluid of DI water and one other
component. Feature Feature Sidewall Bottom Up Material Corrosion
Filling DI water and One Other component Susceptabilty Rate DI
Water Only Excellent Fair-Good >2 g/L Sulfuric Acid Poor Good
>2 g/L Methane Sulfonic Acid Poor Good <2 g/L Sulfuric Acid
Fair-Good Fair-Good <2 g/L Methane Sulfonic Acid Fair-Good
Fair-Good <20 g/L Copper Sulfate Excellent Good <20 g/L
Copper Methane Sulfonate Excellent Good >20 g/L Copper Sulfate
Excellent Excellent >20 g/L Copper Methane Sulfonate Excellent
Excellent >10 ppm Chloride/Bromide Poor Poor >15 ppm
Suppressor Good Fair Accelerator Fair Poor Leveler Fair Poor
TABLE-US-00002 TABLE 2 Pre-wetting fluid of DI water, greater than
2 g/L sulfuric acid or methane sulfonic acid, and one other
component. Material Feature Feature DI water, >2 g/L Sulfuric
Sidewall Bottom Up Acid or Methane Sulfonic Acid, Corrosion Filling
and One Other component Susceptabilty Rate >2 g/L Methane
Sulfonic (or Sulfuric) Acid Poor Good <2 g/L Methane Sulfonic
(or Sulfuric) Acid Poor Good <20 g/L Copper Sulfate Fair Good
<20 g/L Copper Methane Sulfonate Fair Good >20 g/L Copper
Sulfate Fair Excellent >20 g/L Copper Methane Sulfonate Fair
Excellent >10 ppm Chloride/Bromide Poor Poor >15 ppm
Suppressor Good Fair Accelerator Poor Poor Leveler Poor Poor
TABLE-US-00003 TABLE 3 Pre-wetting fluid of DI water, less than 2
g/L sulfuric acid or methane sulfonic acid, and one other
component. Material Feature Feature DI water, <2 g/L Sulfuric
Acid Sidewall Bottom Up or Methane Sulfonic Acid Corrosion Filling
and One Other Component Susceptabilty Rate >2 g/L Methane
Sulfonic (or Sulfuric) Acid Poor Good <2 g/L Methane Sulfonic
(or Sulfuric) Acid Fair-Good Good <20 g/L Copper Methane
Sulfonate Good Good <20 g/L Copper Sulfate Good Good >20 g/L
Copper Sulfate Excellent Excellent >20 g/L Copper Methane
Sulfonate Excellent Excellent >10 ppm Chloride/Bromide Poor Poor
>15 ppm Suppressor Good Fair Accelerator Fair Poor Leveler Fair
Poor
TABLE-US-00004 TABLE 4 Pre-wetting fluid of DI water, less than 2
g/L sulfuric acid or methane sulfonic acid, greater than 20 g/L
copper sulfate, and one other component. Material DI water, <2
g/L Sulfuric Acid Feature Feature (or Methane Sulfonic Acid),
>20 g/L Sidewall Bottom Up Copper Sulfate, Corrosion Filling and
One Other Component Susceptabilty Rate >2 g/L Sulfuric (or
Methane Sulfonic) Acid Fair Excellent <2 g/L Sulfuric Acid (or
Methane Sulfonic) Good Excellent Acid <20 g/L Copper Methane
Sulfonate Excellent Excellent >20 g/L Copper Methane Sulfonate
Excellent Excellent >10 ppm Chloride Poor Poor >15 ppm
Suppressor Excellent Fair Accelerator Fair Poor Leveler Fair
Poor
CONCLUSION
[0159] Although the foregoing apparatus designs and methods have
been described in some detail for purposes of clarity of
understanding, it will be apparent that certain changes and
modifications may be practiced within the scope of the appended
claims. It should be noted that there are many alternative ways of
implementing both the process and compositions described herein.
Accordingly, the present embodiments are to be considered as
illustrative and not restrictive, and embodiments are not to be
limited to the details given herein.
* * * * *