Patent Application
20080163890
1. Field of the Invention
Embodiments of the present invention generally relate to the field of surface preparation systems and methods. More particularly, embodiments of the present invention relate to systems and methods for cleaning substrates, including silicon substrates used in the manufacture of semiconductors.
2. Description of the Related Art
Cleaning of particles and other contaminants from semiconductor substrate surfaces is one of the critical processes in semiconductor manufacturing. Currently, metal and organic contaminants are removed from substrates using physical energies or forces in combination with various chemistries. Those physical energies include acoustic energy, such as megasonic energy, liquid or aerosol spray, and mechanical brushes.
The requirements for substrate cleaning increase as feature sizes decrease. Of particular note are the sub-angstrom film consumption requirements and the rapidly shrinking killer defect particle size which is well below 50 nm in diameter. At the core of the cleaning dilemma is the reduced efficiency of removing sub-50 nm particles. The challenge in removing these nanoparticles is that the ratio of the particle removal force to the particle adhesion force decreases dramatically with shrinking particle diameter.
In order to account for this reduced megasonics particle removal efficiency, the conventional options include increasing the megasonics power level, and/or increasing the cleaning solution concentration, cleaning time, and or temperature of the cleaning solution. Unfortunately, these options are not suitable for the more demanding sub-65 nm surface preparation requirements. Increasing the megasonic power levels introduces an excessive level of megasonics damage to the smaller geometries. Increasing the cleaning solution concentration, cleaning time, and/or temperature of the solution increases film consumption to intolerable levels.
Batch megasonics can cause extensive damage to sensitive device structures from poor control over megasonic energy distribution. Single substrate tools offer improved control, but can still exhibit isolated damage. In general, there is a linear increase in megasonics damage with increasing particle removal efficiency (“PRE”). Particle removal efficiency is defined as [[(pre-post)/pre]*100] where “pre” is the number of particles measured before cleaning and “post” is the number of particles measured after cleaning at a particle size of 90 nm or below.
While poor cleaning efficiency will have a direct effect on line yields, the damage caused to the device by cleaning-induced film loss is less obvious. This damage can include an increased isolation leakage current, a shorter effective channel length, and increased source/drain resistance.
An additional surface preparation challenge presented at the sub-65 nm node is the introduction of new materials, particularly in the device gate stack. These new materials present very stringent requirements for a native-oxide free surface without particle defects. Unfortunately, this presents an additional dilemma since the HF-last cleans which remove native oxides are notorious for leaving high particle counts, especially in batch tools.
These combined issues have proven especially challenging with the sub-65 nm technology node, and have driven the industry toward single substrate processing tools, such as the Emersion™ system described herein and available from Applied Materials, Inc. of Santa Clara, Calif., which offers the high degree of process control required. Batch tools have proven incapable of achieving high PRE without megasonics damage, watermarks, and film consumption. Furthermore, particle addition with HF-last cleans in batch tools is well above the levels that can be achieved in a single substrate tool. Single substrate cleaning tools offer more precision in terms of megasonic energy distribution, and process uniformity. Thus, there is a need for systems and methods that allow for the precise tailoring of megasonics distribution at a substrate surface to be above the threshold required for PRE, yet below the value which causes structural damage.
Embodiments of the present invention generally relate to systems and methods for cleaning a substrate. More particularly systems and methods that allows for precise tailoring of megasonics distribution at a substrate surface to be above the threshold required for PRE, yet below the value which causes structural damage.
In one embodiment a method for cleaning a substrate is provided. A substrate comprising at least one feature definition is provided. A processing fluid is applied to the substrate. Megasonic energy is directed toward the processing fluid to produce a tunable cavitation zone. The substrate is extracted from the processing fluid through the tunable cavitation zone.
In another embodiment a method for cleaning a substrate is provided. The method comprises creating a tunable cavitation zone in the processing fluid which can be adjusted spatially with angle and power and passing a substrate through the tunable cavitation zone.
In yet another embodiment a method for cleaning a substrate is provided. A substrate comprising at least one feature definition is provided. Megasonic energy is directed toward the substrate. Controlling the megasonic energy to produce a single bubble sonoluminescence region. Extracting the substrate through the single bubble sonoluminescence region.
So that the manner in which the above recited features of the present invention can be understood in detail, a more particular description of the invention, briefly summarized above, may be had by reference to embodiments, some of which are illustrated in the appended drawings. It is to be noted, however, that the appended drawings illustrate only typical embodiments of this invention and are therefore not to be considered limiting of its scope, for the invention may admit to other equally effective embodiments.
To facilitate understanding, identical reference numerals have been used, wherever possible, to designate identical elements that are common to the figures. It is contemplated that elements and/or process steps of one embodiment may be beneficially incorporated in other embodiments without additional recitation.
The present invention is described here with respect to a particularly preferred embodiment in which megasonics are used with a processing solution to clean silicon substrates. It will be recognized by those of ordinary skill in the art that these systems and methods can be used to practice a variety of cleaning techniques, on a variety of substrates with a variety of processing solutions. The use of the megasonics/silicon substrate example is intended to be illustrative and not limiting.
The present invention further relates to embodiments of chambers for processing a single substrate and associated processes with embodiments of the chambers. The chambers and methods of the present invention may be configured to perform substrate surface cleaning/surface preparation processes, such as etching, cleaning, rinsing and/or drying a single substrate. Although the illustrative chambers are described for use with one substrate, the embodiments described herein may be used for cleaning a plurality of substrates in a single chamber. Similar processing chambers, methods, and systems may be found in U.S. Pat. No. 6,726,848, which issued on Apr. 27, 2004, U.S. patent application Ser. No. 11/445,707, filed Jun. 2, 2006, U.S. patent application Ser. No. 11/460,172, filed Jul. 26, 2006, and U.S. patent application Ser. No. 11/620,610, filed Jan. 5, 2007, all of which are incorporated herein by reference in their entirety. Embodiments of the invention may be adapted to be disposed on a substrate surface cleaning/surface preparation tool available from Applied Materials, Inc., of Santa Clara, Calif., sold under the trade name “Emersion™.” Embodiments of the invention may also be adapted for use with other substrate surface cleaning/surface preparation tools available from other manufacturers.
Megasonic cleaning is one method of mechanical particle removal used in semiconductor substrate processing. Megasonics is derived from ultrasonic cleaning which has a wider application base and is used in many industries. Both techniques utilize cavitation as a means of particle removal. The cavitation phenomenon can be described as the bubble formation and collapse induced by pressure variations in liquids. Although it is effective for particle removal, collapsing bubbles can also cause material erosion and pattern damage.
Cavitation is generally divided into two classes—transient cavitation (a multiple bubble cavitation mechanism) and stable cavitation (a single bubble cavitation mechanism). Transient cavitation is the process where a void or bubble in a liquid rapidly collapses, producing a shockwave. In transient cavitation, a number of bubbles coalesce leading to asymmetric implosion during a positive pressure cycle. The bubbles involved in transient cavitation are characterized by a large resonance size and a short lifetime measured in nanoseconds. Transient cavitation yields a large number of photon emissions which create microjets and can damage the features on the substrate. Stable cavitation is the repeatable oscillation of bubble diameter without leading to bubble collapse. Stable cavitation is characterized by stable resonance and a long lifetime. The bubble eventually collapses leading to the emission of photons.
Through experimentation using sonoluminescence imaging, a megasonics cleaning method based on single bubble cavitation has been developed. This method utilizes multiple megasonics transducers operated at very low power densities in a single substrate immersion processor. This method is shown to produce high cleaning efficiencies without damage to 45 nm devices. Further, sonoluminescence studies demonstrate that the transducers are operated in the single bubble sonoluminescence (SBSL) regime, well below the cavitation threshold for transient multiple-bubble sonoluminescence (MBSL).
The PRE results described above suggest the presence of unusual acoustic effects within the chamber. The most likely explanation is cavitation events and possible Lamb wave generation. In order to provide a direct indication of these phenomena, sonoluminescence imaging was performed. Sonoluminescence refers to the photon emission that occurs when a collapsing cavitation bubble heats the gas within the bubble to temperatures high enough to generate incandescent lights. In these tests, an optical imaging system was mounted on top of the Emersion™ system described below to quantitatively measure sonoluminescence within the chamber. Several configurations and power levels are summarized in
Sonoluminescence curve 612 shows the transition from single bubble sonoluminescence (SBSL) to multiple bubble sonoluminescence (MBSL) behavior. The damage curve for 45 nm poly-Si gates 608 and the damage curve for 65 nm poly-Si gates 610 show that 45 nm devices must be operated in the single bubble cavitation regime. The plot 600 also shows that high PRE values are possible in the SB regime.
A megasonics cleaning mechanism based on single bubble cavitation has been demonstrated. This method utilizes multiple megasonics transducers operated at very low power densities in a single substrate immersion processor. This method is shown to produce high cleaning efficiencies without damage to 45 nm devices. Further, sonoluminescence studies demonstrate that the transducers are operated in the single bubble sonoluminescence (SBSL) regime, well below the cavitation threshold for transient multiple-bubble sonoluminescence (MBSL).
The lower portion of the chamber body 701 generally comprises side walls 738 and a bottom wall 703 defining a lower processing volume 739. The lower processing volume 739 may have a rectangular shape configured to retain fluid for immersing a substrate therein. A weir 717 is formed on top of the side walls 738 to allow fluid in the lower processing volume 739 to overflow. The upper portion of the chamber body 701 comprises overflow members 711 and 712 configured to collect fluid flowing over the weir 717 from the lower processing volume 739. The upper portion of the chamber body 701 further comprises a chamber lid 710 having an opening 744 formed therein. The opening 744 is configured to allow the substrate transfer assembly 702 to transfer at least one substrate in and out the chamber body 701.
An inlet manifold 740 configured to fill the lower processing volume 739 with processing fluid is formed on the sidewall 738 near the bottom of the lower portion of the chamber body 701. The inlet manifold 740 has a plurality of apertures 741 opening to the bottom of the lower processing volume 739. An inlet assembly 706 having a plurality of inlet ports 707 is connected to the inlet manifold 740. Each of the plurality of inlet ports 707 may be connected with an independent fluid source, such as chemicals for etching, cleaning, and DI water for rinsing, such that different fluids or combination of fluids may be supplied to the lower processing volume 739 for different processes.
During processing, processing fluid may flow in from one or more of the inlet ports 707 to fill the lower processing volume 739 from bottom via the plurality of apertures 741. In one embodiment, the lower processing volume 739 may be filled in less than about 10 seconds, for example less than about 5 seconds, such as between about 5 seconds and about 1 second.
As the processing fluid fills up the lower processing volume 739 and reaches the weir 717, the processing fluid overflows from the weir 717 to an upper processing volume 713 and is connected by the overflow members 711 and 712. A plurality of outlet ports 714 configured to drain the collected fluid may be formed on the overflow member 711. The plurality of outlet ports 714 may be connected to a pump system. In one embodiment, each of the plurality of outlet ports 714 may form an independent drain path dedicated to a particular processing fluid. In one embodiment, each drain path may be routed to a negatively pressurized container to facilitate removal, draining and/or recycling of the processing fluid. In one embodiment, the overflow member 712 may be positioned higher than the overflow member 711 and fluid collected in the overflow member 712 may flow to the overflow member 711 through a conduit (not shown).
In one embodiment, a draining assembly 708 may be coupled to the sidewall 738 near the bottom of the lower processing volume 739 and in fluid communication with the lower processing volume 739. The draining assembly 708 is configured to drain the lower processing volume 739 rapidly. In one embodiment, the draining assembly 708 has a plurality of draining ports 709, each configured to form an independent draining path dedicated to a particular processing fluid. In one embodiment, each of the independent draining path may be connected to a negatively pressurized sealed container for fast draining of the processing fluid in the lower processing volume 739. Similar fluid supply and draining configuration may be found in FIGS. 9-10 of U.S. patent application Ser. No. 11/445,707, filed Jun. 2, 2006, which is incorporated herein by reference.
In one embodiment, a megasonic transducer 704 is disposed behind a window 705 in the bottom wall 703. The megasonic transducer 704 is configured to provide megasonic energy to the lower processing volume 739. The megasonic transducer 704 may comprise a single transducer or an array of multiple transducers, oriented to direct megasonic energy into the lower processing volume 739 via the window 705. When the megasonic transducer 704 directs megasonic energy into processing fluid in the lower processing volume 739, acoustic streaming, i.e. streams of micro bubbles, within the processing fluid may be induced. The acoustic streaming aids the removal of contaminants from the substrate being processed and keeps the removed particles in motion within the processing fluid hence avoiding reattachment of the removed particles to the substrate surface.
In one embodiment, a pair of megasonic transducers 715a, 715b, each of which may comprise a single transducer or an array of multiple transducers, are positioned behind windows 716 at an elevation below that of the weir 717, and are oriented to direct megasonic energy into an upper portion of lower processing region 739. The transducers 715a and 715b are configured to direct megasonic energy towards a front surface and a back surface of a substrate respectively.
The transducers 715a and 716b are preferably positioned such that the energy beam interacts with the substrate surface at or just below a gas/liquid interface (will be described below), e.g., at a level within the top 0-20% of the liquid in the lower processing volume 739. The transducers may be configured to direct megasonic energy in a direction normal to the substrate surface or at an angle from normal. Preferably, energy is directed at an angle of approximately 0-30 degrees from normal, and most preferably approximately 5-30 degrees from normal. Directing the megasonic energy from the transducers 715a and 715b at an angle from normal to the substrate surface can have several advantages. For example, directing the energy towards the substrate at an angle minimizes interference between the emitted energy and return waves of energy reflected off the substrate surface, thus allowing power transfer to the solution to be maximized. It also allows greater control over the power delivered to the solution. It has been found that when the transducers are parallel to the substrate surface, the power delivered to the solution is highly sensitive to variations in the distance between the substrate surface and the transducer. Angling the transducers 715a and 715b reduces this sensitivity and thus allows the power level to be tuned more accurately. The angled transducers are further beneficial in that their energy tends to break up the meniscus of fluid extending between the substrate and the bulk fluid (particularly when the substrate is drawn upwardly through the band of energy emitted by the transducers) thus preventing particle movement towards the substrate surface.
Additionally, directing megasonic energy at an angle to the substrate surface creates a velocity vector towards the weir 717, which helps to move particles away from the substrate and into the weir 717. For substrates having fine features, however, the angle at which the energy propagates towards the substrate front surface must be selected so as to minimize the chance that side forces imparted by the megasonic energy will damage fine structures.
It may be desirable to configure the transducers 715a and 715b to be independently adjustable in terms of angle relative to normal and/or power. For example, if angled megasonic energy is directed by the transducer 715a towards the substrate front surface, it may be desirable to have the energy from the transducer 715b propagate towards the back surface at a direction normal to the substrate surface. Doing so can prevent breakage of features on the front surface by countering the forces imparted against the front surface by the angled energy. Moreover, while a relatively lower power or no power may be desirable against the substrate front surface so as to avoid damage to fine features, a higher power may be transmitted against the back surface (at an angle or in a direction normal to the substrate). The higher power can resonate through the substrate and enhance microcavitation in the trenches on the substrate front, thereby helping to flush impurities from the trench cavities.
Additionally, providing the transducers 715a, 715b to have an adjustable angle permits the angle to be changed depending on the nature of the substrate (e.g. fine features) and also depending on the process step being carried out. For example, it may be desirable to have one or both of the transducers 715a, 715b propagate energy at an angle to the substrate during the cleaning step and then normal to the substrate surface during the drying step (see below). In some instances it may also be desirable to have a single transducer, or more than two transducers, rather than the pair of transducers 715a, 715b.
In one embodiment, the chamber lid 710 may have integrated vapor nozzles (not shown) and exhaust ports (not shown) for supplying and exhausting one or more vapor into the upper processing volume 713. During processing, the lower processing volume 739 may be filled with a processing liquid coming in from the inlet manifold 740 and the upper processing volume 713 may be filled with a vapor coming in from the vapor nozzles on the chamber lid 710. A liquid vapor interface 743 may be created in the chamber body 701. In one embodiment, the processing liquid fills up the lower processing volume 739 and overflows from the weir 717 and the liquid vapor interface 743 is located at the same level as the wire 717.
During processing, a substrate being processed in the substrate processing chamber 700 is first immersed in the processing liquid in the lower processing volume 739, and then pulled out of the processing liquid. It is desirable that the substrate is free of the processing liquid after being pulled out of the lower processing volume 739. In one embodiment, the Marangoni effect, i.e. the presence of a gradient in surface tension will naturally cause the liquid to flow away from regions of low surface tension is used to remove the processing liquid from the substrate. The gradient in surface tension is created at the liquid vapor interface 743. In one embodiment, an isopropyl alcohol (IPA) vapor is used to create the liquid vapor interface 743. When the substrate is being pulled out from the processing liquid in the lower processing volume 739, the IPA vapor condenses on the liquid meniscus extending between the substrate and the processing liquid. This results in a concentration gradient of IPA in the meniscus, and results in so-called Marangoni flow of liquid from the substrate surface.
As shown in
Using the aforementioned data and system, a method that allows for precise tailoring of megasonics distribution at the substrate surface by achieving the threshold required for PRE, while remaining below the value that causes structural damage is provided. This method 800 may be performed in a process chamber similar to that described above with reference to
At step 806, megasonic energy is directed into the process chamber to produce a tunable cavitation zone. When the upper transducers 715a, 715b are powered on, the upper transducers form a zone Z of optimum performance. This zone Z is a band of megasonic energy extending across the chamber, preferably slightly below the gas/liquid interface. In one embodiment, the lower megasonic transducer 704 is also powered on thus contributing megasonic energy to zone Z forming a three phase interface of megasonic energy. The area of the zone Z is preferably selected such that when the substrate passes through the zone Z, up to 30 percent of the surface area of a face of the substrate is positioned within the zone. Most preferably, as the center of the substrate passes through the zone, approximately 3-30 percent of the surface area of a face of the substrate is positioned within the band.
The upper transducers 715a, 715b may be configured to direct megasonic energy in a direction normal to the substrate surface or at an angle from normal. Preferably, energy is directed at an angle of approximately 0 degrees to 30 degrees from normal, and most preferably approximately 5 degrees to 30 degrees from normal.
The tunable cavitation zone can be adjusted by modifying the power level and the angle of each transducer. As discussed above, it is preferable that the power levels be adjusted so the cleaning process can take place in the single bubble sonoluminescence region.
At step 808, the substrate is extracted through the tunable cavitation zone. A “sweep” is performed when the wafer is extracted from the chamber and inserted into the chamber through the zone Z of optimum performance. When the wafer is extracted from the chamber, the wafer is swept through this zone of optimum performance. The substrate may be translated through the zone to achieve a rate of approximately 25-300 mm/sec, such as between about 100 mm/sec and about 200 mm/sec, for example about 150 mm/sec. In one embodiment, upon initiation of the “sweep” the upper transducers and the lower transducer are powered on. The three transducers are powered to between about 0.04 W/cm2 to about 0.2 W/cm2 each, such as between about 0.10 W/cm2 to about 0.15 W/cm2, for example about 0.12 W/cm2 each. In one embodiment, the three transducers are powered off after the extraction step. In one embodiment, the upper transducers 715a, 715b have different power levels. In another embodiment, the upper transducers 715a, 715b have the same power level. In another embodiment, all three transducers have different power levels. In another embodiment, all three transducers have different power levels.
Multiple embodiments utilizing principles of the present invention have been described. These embodiments are given only by way of example and are not intended to limit the scope of the claims—as the apparatus and method of the present invention may be configured and performed in many ways besides those specifically described herein. Moreover, numerous features have been described in connection with each of the described embodiments. It should be appreciated that the described features may be combined in various ways, and that features described with respect to one of the disclosed embodiments may likewise be included with the other embodiments without departing from the present invention.
While the foregoing is directed to embodiments of the present invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow.
This application claims benefit of U.S. provisional patent application Ser. No. 60/884,362, filed Jan. 10, 2007, which is herein incorporated by reference.
| Number | Date | Country | |
|---|---|---|---|
| 60884362 | Jan 2007 | US |
