Patent Application
20080190558
The present disclosure relates to a method and apparatus for processing of a substrate. More particularly, a method and apparatus for concentrically positioning a substrate relative to an apparatus for processing the edge of the substrate is disclosed. Furthermore, a seal arrangement for the alignment apparatus is also provided. In addition, processes for dry etching of a substrate with a combustion flame are disclosed.
The statements in this section merely provide background information related to the present disclosure and may not constitute prior art.
During the manufacture of integrated circuits, silicon substrate wafers receive extensive processing including deposition and etching of dielectrics, metals, and other materials. At varying stages in the manufacturing process it is beneficial to “clean” the edge area of the wafer to remove unwanted films and contaminants including particles that develop as a result of the wafer processing.
This includes films and contaminants that develop on a near edge top surface (primary processed side), near edge back surface, and edge (including, top bevel, crown and bottom bevel) of the wafer (hereinafter “edge area” refers generally to the near edge top surface, near edge bottom surface, and edge in combination or individually). Removal of films and contaminants is desirable to prevent the potential of particulate migration into the device portion of the wafer. Potential contaminant particles are generated during wafer handling, processing, and as a result of “pop-off” effect due to film stress.
It is a challenge to process and thus remove edge area thin films and contaminants in an efficient and cost effective manner without affecting the remainder of the wafer that contains in-process devices. This challenge is exacerbated by use of chemistries and processes that may adversely impact the in-process device portion of the wafer.
Many of the existing film removal techniques fail to properly remove polymers, edge beads, dielectric or tantalum, particularly from the edge area, as may be desired by the wafer manufacturer. Specifically, it is desirable to maximize the usable surface area of a wafer thus minimizing any unusable edge area with the objective of maximizing die yield. Reduction in functional die produced from the usable surface area is termed yield loss and is generally undesirable and has a negative cost impact. Accordingly, a need in the art exists for improved processing methods and apparatus to remove various front side, back side and edge area films and contaminants in a cost effective and efficient manner.
In accordance with the present teachings, an edge area substrate processing method and apparatus provides advantages over the aforementioned processing methods and systems. An aspect of the present teachings is directed to a method and apparatus for dry chemical processing at atmospheric pressure, the edge area of a substrate in isolation from the remainder of the substrate. The substrate edge area processing apparatus has an isolator for isolating a portion of the substrate edge area to be processed. One or more grooves in the isolator form a plenum for confining flow of a reactive species to the edge area of the substrate. One or more nozzles are disposed in the isolator with at least one of the one or more nozzles at an angle between perpendicular and horizontal to the top surface of the substrate. The one or more nozzles are for emitting a reactive species for reacting with a material on the substrate edge area. Pressure differentials bias the reactive species away from the area of the substrate outside of the isolator.
A substrate edge processing method is disclosed for isolating for isolating and processing a portion of a substrate. The portion to be processed extends from an edge of the substrate radially across the top surface of the substrate to another part of the edge of the substrate, thus isolating an edge area to be processed. A pressure differential barrier is formed between the portion of the substrate being processed and the remainder of the substrate. A reactive species is directed towards the processed portion of the substrate at an angle greater than parallel to the top surface of the substrate and less than vertical to the top surface of the substrate.
In other embodiments, an edge area of the substrate to be processed is isolated from the remainder of the substrate by directing a flow of an inert gas through a plenum near the area to be processed thus forming a barrier while directing a flow of reactive species at an angle relative to the top surface of the substrate towards the substrate edge area thus processing the substrate edge area. A flow of oxygen containing gas into the processing chamber together with a negative exhaust pressure may contribute to the biasing of reactive species and other gases away from the non-processing areas of the substrate.
The described method and apparatus allows for precise processing of portions of the substrate particularly the substrate edge area without allowing for encroachment in the excluded area. Flow control as a part of the apparatus isolator structure in combination with pressure differentials effectively limits movement of reactive species into the area excluded. Using directed flow of the reactive species to the edge area of the substrate allows for a high etch rate and resulting overall significant improvement of throughput of processed substrates. In sum, the system provides for a clean, effective, and efficient method and apparatus for processing the edge area of substrates in a manner that is highly desired for achieving low contamination of the device portion of the substrate.
The present disclosure further provides a method and apparatus for aligning a wafer in a highly concentric and precise fashion. Concentric process application has many benefits over existing technologies. It enables atmospheric pressure, gas phase removal of many undesirable films from the edge area of a semiconductor wafer. The concentric process application measures a radius of a wafer at various locations while the wafer is spinning on a chuck. A determination of a precise center of the wafer is calculated and the wafer is repositioned at the precise center for processing.
Also disclosed is a multi-axis motion seal (i.e. labyrinth) for sealing the processing chamber during processing of the wafer. The seal functions in association with a wafer chuck. The seal and processing chamber define a vacuum chamber connected to a vacuum that is movable in cooperation with the alignment system.
In addition, processes for combustion flame based processing of the wafer are disclosed. The disclosed chemistries react in a combustion flame to produce a reactive species for processing the wafer in a precise and efficient manner.
In another embodiment, a system is provided for dielectric film removal from near edge regions. These films are etched using H2:NF3 dominant chemistries. Certain metal films can also be removed. Examples include tungsten and tantalum. Many metal oxide or nitride films can also be etched.
Further areas of applicability of the present invention will become apparent from the detailed description provided hereinafter. It should be understood that the detailed description and specific examples, while indicating the preferred embodiment of the invention, are intended for purposes of illustration only and are not intended to limit the scope of the invention.
The drawings described herein are for illustration purposes only and are not intended to limit the scope of the present disclosure in any way.
a and 22b represent nozzle subplates as shown in
The following description is merely exemplary in nature and is not intended to limit the present disclosure, application, or uses.
Central to the present disclosure's near edge film removal technology is the ability to apply reactive gas to a wafer in a highly concentric and precise fashion. Process application is typically sensitive to wafer or substrate eccentricity variation in the range of 50 to 100 μm. Multiple subsystems are required to achieve this type of process application.
The equipment front end module 17 contains a robot and the pre-aligner station 19. Wafers are processed from a front opening unified pod. The utility cabinet 20 contains control electronics, computer(s), endpoint equipment, gas delivery equipment and other facilities interconnects. Process gases 21 are connected to the module and flow regulated by appropriate mass flow controllers (MFC's) 52. Other facilities connections such as exhaust 56 and cooling water 58 are also connected.
Referring generally to
Disposed within the upper section 38 of the isolator 25 are a first nozzle 45 and a second nozzle 49. Both nozzles are configured to emit a directed flow of reactive species towards the edge area 33 of the wafer 26. First nozzle 45 is offset from an axis perpendicular to a plane that is common with the top surface 30 of the wafer 26 (the “wafer plane”). First nozzle 45 is pointed towards the top surface 30 at an angle of 80°±/−5°relative to the wafer plane. Second nozzle 49 is offset by an angle of 45°±/−5° to the wafer plane. Second nozzle 49 is also offset by ˜15° from a plane perpendicular to the wafer plane that runs through the center of the isolator 25 and center of the wafer 26.
First nozzle 45 is connected to a first channel 48 disposed in the upper section 38. First channel 48 leads to a gas line 47. Second nozzle 49 is connected to a second channel 53 disposed in the upper section 38. Second channel 53 leads to the gas line 47. First nozzle 45 and second nozzles 49 are connected via the gas line 47 to a reactive gas species source. Optionally, the first and second channels 48 and 53 can be coupled to sources having differing chemistry.
First nozzle 45 is positioned for bevel and crown processing at a distance of 0.1 to 0.5 mm from the edge of the wafer 26 and 1.3 to 1.8 mm distance from the top surface 30 of the wafer 26. Second nozzle 49 is positioned 0.5 to 3.0 mm in from the edge of the wafer 26 and 0.6 to 1.1 mm distance from the top surface 30 of the wafer 26. Radial position of the nozzles and distance from the wafer surface is dependent upon desired edge exclusion area and is also process and film dependent.
Reactive gas species source either provides a reactive gas species or component reactants for forming the reactive gas species. Reactive gas species can be generated via near atmospheric pressure techniques. This includes near atmospheric capacitively coupled plasma source (i.e., APJET), as described in U.S. Pat. No. 5,961,772, incorporated herein by reference or inductively coupled plasma discharge (i.e., ICP torch), as described in U.S. Pat. No. 6,660,177, incorporated herein by reference or combustion flame.
Spontaneous etchants, for example F2, O3, or HF can also be used. Advantageously, none of these reactive species techniques produce ion bombardment characteristic of an ionic plasma thus minimizing surface and device damage potential. Further, although envisioned, none of these techniques requires a vacuum chamber together with associated equipment.
An upper purge plenum 88 disposed in the upper section 38 extends at or near the edge of the top surface of the wafer 26, above and across an area of the wafer to be processed to at or near another edge of the top surface 30 of the wafer 26. The upper purge plenum 88 is ˜3.0 mm wide and extends for a total path length of ˜37.5 mm. The upper purge plenum 88 is part of a tuned flow system which prevents reactive gas migration out of the processing area.
The upper purge plenum 88 is connected to a first purge channel 92 that is connected to a purge gas source 96 via a purge gas line 94. The purge gas source 96 supplies an inert gas, for example, argon that is fed via the first purge channel 92 into the upper purge plenum 88. Alternatively, the upper purge plenum 88 can provide CDA or oxygen containing gas, which augments the reaction of the reactive gas.
The use of oxygen containing gas allows the reaction of un-reacted H2. This also compensates for extreme length limitations and allows for a higher volume fraction of NF3. The increased NF3 volume fraction leads to enhanced etched rates as well as an enhancement of throughput. Although one purge channel is seen disposed in the upper section 38 of the isolator 25, more than one channel may be present for directing a flow of purge gas into the upper purge plenum 88. Purge channels have an inside diameter of 2.00 mm. The flow of purge gas into the upper purge plenum 88 creates a pressure differential in the area of the top surface 30 surrounded by the upper purge plenum 88 resulting in a barrier between the top surface 30 and the edge area 33 of the wafer 26 being processed.
The upper purge plenum 88 is separated from the top surface 30 of the wafer 26 by an inside baffle 100. Inside baffle 100 follows along the inside perimeter of the upper purge plenum 88 and is separated from the wafer 26 by a gap of 0.30 to 0.80 mm. An outside baffle 104 follows along the outside perimeter of the upper purge plenum 88 and is separated from the wafer 26 by a gap of 0.50 to 1.10 mm. As seen, outside baffle 104 is wider and closer to the top surface 30 of the wafer 26 than the inside baffle 100. This facilitates forming a pressure induced barrier around the in-process portion of the wafer 26 by creating a pressure differential biasing a flow of a purge gas in a direction across inside baffle 100 into the processing area of the isolator 25.
A second purge channel 108 is disposed in the lower section 39 of the isolator 25. This is connected by the purge gas line 94 to the purge gas source 96. Second purge channel 108 is for feeding purge gas to a lower purge plenum 114. Similarly to the upper purge plenum 88, the lower purge plenum 114 extends from at or near the edge area 33 of the wafer 26 below and across the bottom surface 32 to at or near another location of the edge of the wafer 26. Similarly to the upper purge plenum 88, the lower purge plenum 114 is disposed between a lower inside baffle 112 and a lower outside baffle 118. The lower purge plenum 114 together with the lower inside baffle 112 and lower outside baffle 118 bias a flow of purge gas in a direction across the lower inside baffle 112 and across the bottom surface 32.
Wafer chuck 28 is movable in r-θ-z or xyz-θ directions, using module 27, for positioning the wafer 26 and rotating it within a slot of the isolator 25 defined between the upper section 38 and lower section 39. Alternatively, the isolator 25 structure can also be moved in r with the chuck moving in θ and z. Once in position the distance between each side of the wafer 26 and the upper section 38 or lower section 39 is 0.30 to 0.80 mm. The slot open area without a wafer 26 is 124.20 to 216.20 mm2. The slot open area with a wafer 26 present is 55.20 to 147.20 mm2. The exhaust slot width is 93.0 mm.
A gas diffuser 24 extends into the processing chamber 22 providing a flow of inert or oxygen containing gas to the processing chamber 22. The gas diffuser 24 is typically of the shower head type design and is connected via a diffuser 24 gas line 148 to the purge gas source 96.
The exhaust plenum 41 together with the exhaust system 56 are an additional part of the tuned flow system which prevent reactive gas migration out of the processing area. Exhaust system 56 creates a negative pressure in the exhaust plenum 41 that draws active species gases together with the inert gas, processed byproducts, and condensation away from the processing area and prevents migration of these gases into the device area of the wafer 26.
A heater element 122 is connected by a heater line to a heater power supply 126. The heater element 122 heats the isolator 25 and to a lesser extent, the wafer 26. Heating the isolator 25 is desirable to prevent condensation of gases that can be corrosive to the isolator 25 and potentially introduce contamination into the processing area.
The nozzles of the edge area processing system 20, including the first nozzle 45 and second nozzle 49 are made of sapphire. Sapphire is advantageously non-reactive to the chemistries used in substrate processing. This is desirable since the processing of semiconductor substrates requires trace material contamination analysis at the parts per million level with acceptable addition to the substrate being less than approximately 1010 atoms/cm2. Further, particle additions to the substrate should be zero for sizes greater than approximately 0.1 micron.
It is also, in many situations, desirable to achieve a laminar gas flow from the nozzles. This requires setting the aspect ratio of the nozzle at greater than or equal to 10× length to diameter. With some reactive gases, aspect ratios of greater than 40:1 or preferably 80:1 are desirable. Nozzle inside diameters are around 0.254 to 0.279 mm which requires a uniform smooth nozzle bore length of approximately 2.50 mm.
The isolator 25 nozzles, including the first nozzle 45 and second nozzle 49, while described as angled relative to the wafer plane at ˜80 degrees and ˜45 degrees, respectively, can advantageously be angled in a different direction relative to the wafer plane in order to facilitate processing including etching or deposition of a thin film.
In operation, a wafer 26 is centered on the wafer chuck 28 and then the wafer chuck 28 positions the wafer 26 in the slot of the isolator 25 between the upper section 38 and the lower section 39 for processing. The movement system 27 rotates wafer chuck 28, and thus the wafer 26.
Inert gas or CDA is allowed to flow into the upper purge plenum 88 and lower purge plenum 114 from the purge gas source 96. The inert gas or CDA flows into the upper purge plenum 88 and lower purge plenum 114 at a rate of 100 sccm to 8,000 sccm. Inert gas or CDA is also allowed to flow into the processing chamber 22 through the gas diffuser 24. This gas flows into the processing chamber 22 at a rate of 500 sccm to 10,000 sccm.
The exhaust system 56 is activated to draw gases and process byproducts including condensation through the exhaust plenum 41. Next, reactive species 130 emit from first nozzle 45 and second nozzle 49. The igniter power supply 126 energizes the clean igniter system 78 and the first gas line 93 and second gas line 98 are opened to allow a flow of hydrogen and nitrogen trifluoride gases into the nozzle assembly 84 and through the four nozzles 84. The gas mixture is frequently different during the ignition stage. The igniter nozzle uses H2 and O2 only at higher total flow rates than the processing nozzles 45, 49. Typically, the initiator nozzle uses approximately 800 sccm H2 and 200 sccm. The process nozzles typically ignite with a Lo NF3 fraction. Typically about 20 sccm max. Reactive species (or gases in the case of a combustion flame) flow through the nozzles at a rate of between 200 and 800 sccm and preferably between 375 sccm to 475 sccm. The reactive species 130 impinge upon the edge area 33 of the wafer 26 as the wafer 26 rotates. The reactive species 130 react with a thin film or contaminant in the edge area 33 of the wafer 26 resulting in a reactant byproduct 66. Alternate nozzle configurations are envisioned. For example, referring briefly to
Heater 122 is energized to heat the wafer top surface 30. This optional step is intended to prevent vapor produced as a byproduct of the chemical reaction, for example water vapor, from condensing on the wafer top surface 30. Condensation can be prevented by heating the wafer top surface 30 to a temperature at or above the boiling point for the reactant byproducts, for example heating the wafer top surface 30 above 100° C. to prevent the condensation of water. Alternatively, wafer 26 surface heating can be supplied via a heated substrate holder 82 or via infrared energy directed at the wafer perimeter, or via other heat sources such as a flame.
The reactive species 130 are prevented from passing out of the isolator 25 by the flow of inert gas working in concert with a pressure differential drawing gases into the exhaust plenum 41 and into the exhaust system 56. This inert gas forms a pressurized barrier in the upper purge plenum 88 and lower purge plenum 114 surrounding the in-process edge area of the wafer. The inside baffle member 61 in cooperation with the outside baffle member 63 biases the flow of insert gas towards the in-process area of the wafer 26. Reactant byproducts formed as a result of the reactive species 130 reacting with a thin film on the wafer 26 surface are drawn away from the in-process area of the wafer 26 into the exhaust plenum 41. Thus, advantageously, reactive species 130 and reactive byproducts 142 are confined to the edge area of the wafer 26 and prevented from migration into other areas of the wafer 26 that may damage wafer component devices. In addition, the pressure differential induced by the exhaust plenum 41 further biases gas flow away from the central portion of the wafer 26.
As the wafer 26 rotates either the wafer chuck 28 translates with respect to the nozzle assembly 84 and the combustion flame across the wafer top surface 30. As a result a desired section of the wafer top surface 30 is processed. Processing includes the removal of a thin film, for example, silicon dioxide or tantalum as described above in relation to the substrate processing method.
After the wafer is processed, the first gas controller 102 and second gas controller 106 are closed. Simultaneously, the fourth gas controller 49 is opened to allow a flow of argon gas or CDA into the edge-type nozzle assembly 84 and through the first and second nozzles 45, 49 to “blow out” the combustion flame. The controller 140 additionally allows blow off of the nozzles if EMO or a power failure occurs. Additionally, the controller 52 can extinguish the flames upon low gas delivery pressure, if the enclosure is opened, or if there is a loss of control air. Also coupled to the controllers are a plurality of H2 sensors which will shut off the system or signal an alarm should the H2 level in the chamber 22 be above a predetermined level. The wafer 26 may be removed after the chamber 22 is evacuated of process gases and byproducts.
Processing of the edge area 33 of the entire wafer may be accomplished with a single rotation of the wafer 26. Alternatively, more than one rotation may occur and more than one process may be performed including deposition and etching. After the flow of reactive species is stopped a flow of the inert gas continues until the processing chamber 22 is sufficiently evacuated of other gases and condensations. Then, the heater element 122 is turned off and the flow of inert or CDA gas from the purge gas source 96 is stopped and the wafer 26 is removed and replaced with another wafer for processing.
The described system 20 and associated method for using the system is suitable for etching of target thin films. This includes, but is not necessarily limited to, tantalum and tantalum nitride; inter-layer dielectrics; backside polymers; and photoresist edge bead.
The translational ‘R-axis’ gap and the ‘Z-θ axis’ gap are shown in
A minimum pressure differential between the seal exhaust ports, and the process chamber 22 was found to be −2 water column inches. Larger differential pressure values can be used and a practical upper limit is not known. Pressure differential between the process chamber and atmosphere should be at least −0.4 water column inches. This results in a seal exhaust to atmosphere pressure differential of at least −2.4 water column inches.
Referring again to
Gaseous reactants react in combustion flame to form gaseous hydrogen fluoride (HF) (a reactive species) and gaseous nitrogen (N2) effluents. The following chemical equation describes the production of gaseous hydrogen fluoride and gaseous nitrogen from gaseous reactants based on a stoichiometric mixture (a 3:2 molar ratio):
3H2(gas)+2NF3(gas)→6HF(gas)+N2(gas)
Advantageously, this reaction is performed substantially at atmospheric pressure. This allows for use of viscous (rather than molecular) flow properties to precisely treat portions of the substrate surface 18 and minimize exposure of other substrate areas to the reactive process. Although a 3:2 molar ratio is described higher or lower ratios may be used depending on the desired result.
Further, this reaction is not induced by an ion producing field consistent with a plasma. It is believed that a plasma is a collection of charged particles where the long-range electromagnetic fields set up collectively by the charged particles have an important effect on the particles' behavior. It is also believed that the combustion flame 12 has substantially no ionic species present. As a result, there is no risk of ionic damage to the substrate.
Substantial heat is generated from the exothermic chemical reaction of H2 and NF3. This effect allows a small volume of highly reactive species in the form of HF to be generated due to the amount of energy represented by the resultant temperature. Elevated temperature in turn substantially increases reaction rates which results in higher etch rates. The result is higher process throughput.
A silicon dioxide thin film can be etched by the gaseous hydrogen fluoride according to the following overall reaction:
4HF(gas)+SiO2(solid)→SiF4(gas)+2H2O(gas)
Gaseous silicon tetrafluoride and water vapor leave the surface of the silicon dioxide thin film. Advantageously, this reaction provides for a change of silicon dioxide thin film from a solid to a gas byproduct that can be easily evacuated.
Gaseous hydrogen fluoride will also etch a substrate surface of silicon. Silicon etching follows the following overall reaction:
4HF(gas)+Si(solid)→SiF4(gas)+2H2(gas)
In this reaction, gaseous silicon tetrafluoride and gaseous hydrogen leave the silicon substrate surface. This reaction provides for a change of silicon on the substrate surface from a solid to a gas byproduct that can be evacuated.
Similarly, etching of a tantalum thin film follows the following overall reaction:
10HF(gas)+2Ta(solid)→2TaF5(gas)+5H2(gas)
In this reaction, gaseous tantalum pentafluoride and gaseous hydrogen leave the tantalum substrate surface. This reaction provides for a change of the tantalum on the substrate surface from a solid to a gas byproduct that can be evacuated. For this reaction, preheating of the wafer using an O2+H2 flame is desirable to prevent the condensation of reaction products on the wafer.
Organic and polymer films can also be removed using the above described chemistry however selectivity issues to Si and SiO2 may in some instances make this less desirable. The above chemistry for example can be used to etch SiO2 over Si where etching of oxide is desirable but Si is not. Passivation of exposed Si to the etch chemistry can be promoted by first exposing an etch field to a hydrogen rich flame with oxygen. The etch field is then exposed to the combustion flame of H2 and NF3 where the oxide is etched.
Other desirable non-oxygen oxidizers for reaction with hydrogen in a combustion flame for substrate etching include fluoride (F2), chlorine (Cl2), and chlorine trifluoride (ClF3). Hydrogen and fluoride react in a combustion flame as follows:
H2(gas)+F2(gas)→2HF(gas)
Similarly to the combustion flame of H2 and NF3 the resulting HF reactive species is a desirable etchant as described above.
Hydrogen and chlorine react in a combustion flame as follows:
H2(gas)+Cl2(gas)→2HCl(gas)
Hydrogen and chlorine trifluoride react in a combustion flame as follows:
4H2(gas)+2ClF3(gas)→6HF(gas)+2HCl(gas)
In both the proceeding combustion flame reactions, the resultant hydrogen chloride reactive species can be advantageously used for etching when materials not readily etched by fluorine are present in the film stack. This includes a film stack comprising aluminum. Hydrogen chloride as a reactive species etches aluminum as follows:
2Al(solid)+6HCl(gas)→2AlCl3(gas)+3H2(gas)
Hydrogen chloride etches silicon as follows:
Si(solid)+4HCl(gas)→SiCl4(gas)+2H2(gas)
Hydrogen chloride etches silicon oxide as follows:
SiO2(solid)+4HCl(gas)→SiCl4(gas)+2H2O(vapor)
Chlorine trifluoride represents a hybrid etch chemistry where both fluorine and chlorine based etchant reactive species are produced. Often this compound is combined with another fluorine containing gas (such as NF3 or CF4) or with Cl2 is used in varying ratios when multiple materials are present in the film stack, requiring both fluorine and chlorine based chemistry for removal.
The chemical equations shown above are a simplified view of the real reactions taking place within the combustion flame and on the substrate surface. The reaction chemistries occurring are quite complex resulting in intermediate and final reaction products.
A nozzle assembly 84 is held by a support member 46 over a wafer 26 retained on the substrate holder 82. Four nozzles 45 are disposed in the nozzle assembly 84. The nozzle assembly 84 is maintained at a distance of ˜1.5 mm from the wafer top surface 30 during processing.
A hydrogen gas source and nitrogen trifluoride gas source 55 are connected by a first gas line 48 and second gas line 53 through a first gas controller 102 and second gas controller 106 to a common mixing gas line 110 connected to the nozzle assembly 84 for combining and mixing H2 and NF3. An exhaust scoop 116 is adjacent to the substrate holder 82 for exhausting gases and reactant byproducts. The exhaust scoop is connected by a plenum 67 to a blower device 124. The exhaust scoop 116 draws gases and reactant byproducts out of the processing chamber 22 through the blower device 124.
In one embodiment, an argon gas source 96 is connected by a third gas line 132 through a third gas controller 49 to the processing chamber 22. In another embodiment, a CDA (clean dry air) or oxygen containing gas 72′ is connected by the third gas line 132 through a third gas controller 49 to the process wafer. The argon or CDA gas source 131 is also connected by a fourth gas line 134 through a fourth gas controller 49 to the common mixing gas line 110. An igniter assembly 78 positioned close to the nozzle assembly 84 is connected by wires 83 to an igniter power supply 126.
In operation, the robot unloads wafer from front opening unified pod (FOUP) and places the wafer on a pre-aligner 19. Once the pre-alignment routine is completed, the robot retrieves wafer from pre-aligner and places it into the chamber 22 on lift pins 16. Wafer chuck 28 moves up in z and lifts wafer 26 from lift pins 16 and rotates and positions the wafer edge to allow measurement using laser micrometer 15. Wafer center offset direction and magnitude is computed as described above. Wafer 26 is then rotated to align offset direction with the ‘r’ axis. The chuck 28 then descends in ‘z’ axis to return wafer to lift pins 16. The wafer movement system 27 moves chuck assembly increments in ‘r’ by the offset magnitude to center the chuck 28 with respect to the wafer 26. The movement system 27 then elevates in ‘z’ axis to lift wafer from lift pins 16. The chuck rotates and the edge position is re-measured to validate centering. The wafer is then ready for concentric process application as described above.
A heater 122 is positioned proximately to the area of the wafer 26 to be processed. The heater 122 (shown in
Referring to
Referring to
Referring to
Now referring to
Although NF3 is used in the above embodiments as the non-oxygen oxidizer other non-oxygen oxidizers as previously discussed are suitable for use in the preferred embodiments. Further, additional embodiments for isolating and processing a wafer according to the above-described method are disclosed in U.S. patent application Ser. No. 11/230,263, filed on Sep. 19, 2005 and titled “Method and Apparatus for Isolative Substrate Edge Area Processing.” The disclosure of this application is incorporated herein by reference.
Removal of dielectric thin films such as silicon oxide from substrates using H2 and NF3 gas mixtures is performed with a hydrogen fraction in the range of 0.5 to 0.7. For example, if the total flow is 800 sccm, H2 flow will be in the range of 400 sccm to 560 sccm with NF3 flow in the range of 400 sccm to 240 sccm. IR preheat is used in cases where ambient oxygen is present to discourage combustion products from condensing on the substrate.
Removal of tantalum from the near-edge region of the substrate is carried out using an etch nozzle configuration similar to that detailed for dielectric removal. Total gas flow per nozzle is approximately 400 sccm with an H2 fraction in the range of 0.6 to 0.7. The primary tantalum etch product is TaF5 which has a boiling point of ˜230° C. Substrate surface temperatures in the etch region must be kept about this temperature to prevent condensation of the etch product. This is readily achieved using an additional combustion flame nozzle (not shown) positioned to impinge a flame on the substrate immediately prior to the impingement of the etch flame. This pre-heat nozzle discharges a flame of H2 and O2 preferably in the range of 0.5 to 0.8, H2 fraction at a total flow of ˜400 sccm for a single nozzle.
A rate of etching of the edge portion of the wafer 26 can be calculated based on consideration of exposure width, wafer circumference and rotational speed. For example, consider a 200 mm circumferential wafer with 2,000 Å of SiO2 that is rotated at 2 rpm and the SiO2 thin film on the edge area is completely removed in one rotation. Assuming a conservative exposure width of mm of the combustion flame effluent on the wafer edge (using a 0.256 mm nozzle bore) an exposure fraction can be calculated as 5 mm/(628 mm×2 rev/min)=0.004 min/rev. The etch rate can then be approximated by dividing the 2,000 Å/rev removal by the exposure fraction. That is 2,000 Å/rev/0.004 min/rev=500,000 Å/min SiO2 removal. If a smaller 2 mm exposure width is assumed then the removal rate becomes 1,256,000 Å/min. Based on these considerations and assumptions a poly-silicon thin film would be etched at an approximate rate of 3×106 Å/min; a photoresist thin film would be etched at an approximate rate of 4.6×106 Å/min; and a tantalum thin film would be etched at an approximate rate of 1×106 Å/min. This is a significantly high rate of etching resulting in a high rate of processing throughput of wafers.
One configuration is optimized for EBR from spin-on films on the top surface and edge region of wafers. This configuration uses reactive gas generated by a combustion flame of H2 and O2 to remove the resist. The present disclosure defines an optimized process using a minor fraction of the non-oxygen oxidizer NF3 in the gas mixture for photoresist EBR. This addition increases the combustion flame temperature and chemical reactivity. These modifications to the combustion flame mixture substantially enhance sharpness of the etch interface and increase slope of the transition to full film thickness, both highly desirable enhancements.
For spin on films with low or minimal etch rate in the H2:O2 dominant chemistry such as organosilicates, inorganic polymers, and spin on glass materials, increasing amounts of fluorine containing gases such as NF3 can be added to further increase etch rate. In this embodiment reactive gas application to the near edge area of the wafer is achieved using the invention disclosed in “Method and Apparatus for Isolative Substrate Edge Area Processing,” previously incorporated by reference.
Undesirable dielectric films can be removed from the front surface of in process semiconductor wafers. These films can also flake and result in defects which cause yield loss. Concentric process application is critical in these processes where reactive gas application must be targeted to the edge region while not affecting the device area of the wafer.
Tantalum removal is similar in configuration to the front side dielectric removal module. Differences exist in the use of a preheat nozzle to reach a higher surface temperature (>230° C. target) to prevent TaF5 condensation in the etch region. Surface temperature pre-heat target for typical film removal is ˜120° C. and is primarily to prevent condensation of water vapor byproduct from the combustion reaction.
The in-situ wafer centering sequence typically takes 8 to 15 seconds. This overhead can be overlapped with gas flow stabilization time or ignition sequence. Wafer ‘z’ plane displacement is measured during rotation and can be used to map out ‘z’ displacement due to wafer bow or warp.
Process operation and details for Ta and dielectrics is discussed at length in the “Substrate Processing Method and Apparatus Using a Combustion Flame” patent application, previously incorporated by reference. This process operation can be applied to backside polymer and edge bead removal.
Backside polymer removal according to the principles of the present disclosure is accomplished by using four nozzles located in the isolator structure. As shown in
Typically, the thickest polymer is located on the bevel region of the wafer. Consequently the NF3 fraction in the 105° jets is higher than the 45° jets aimed at the thinner polymer on the back surface. Currently the method process uses 210 sccm H2, 80 sccm O2, and 100 sccm NF3 in each 105° (high fraction) nozzle. Flows of 240 sccm H2, 120 sccm O2, and 20 sccm NF3 are used in each 450 (low fraction) nozzle. The nozzles are constructed from sapphire with an ID of approximately 254 μm and an aspect ratio of greater than or equal to 10:1. Rotational speeds using during process are typically in the 1 to 6 RPM range. Surface heating for condensation prevention (>100° C. target) is done using a fiber coupled laser diode array.
Chemistry used for EBR depends on the film being removed. For photoresist removal 240 sccm H2, 120 sccm O2, and 20 sccm NF3 performs well. Rotation rate to remove 15,000 Angstroms of resist is typically 1 to 3 RPM. Two nozzles are used for the photoresist EBR process, one at 450 and one at 650. In cases where minimum edge exclusion is desired (˜0.5 mm) only the 650 jet is used. Films with low removal rate, typically silicon containing films, require higher NF3 fraction. The high fraction process used for backside polymer is an example (25% NF3) although higher fractions can be used, frequently without oxygen addition, to ˜50%.
Nozzle aiming for backside polymer removal is shown in
The centering routine measures and records θ, Ti, (1 . . . n) and the laser micrometer 15 reading, Li, (1 . . . n) which represents the edge position. Typically n=50 in this application. The true radius of the wafer is assumed (100 mm or 150 mm). Theta is referenced using the wafer notch position. The following values are computed for each data point:
X
i=(R+Li)·cos(Ti) 1a
Y
i=(R+Li)·sin(Ti). 1b
The objective is to minimize the sum of squares of the deviations given by
D
i=(Xi−Xc)2+(Yi−Yc)2−Rc2 2
where Xc is the x-axis center point, Yc is the y-axis center point and Rc is the assumed radius. The Gauss-Newton method is used to solve the set of non-linear equations. An example of this method is given in “Least-Squares Fitting of Circles and Ellipses” by Gander, et. al. published in BIT, vol. 34, 1994, pp. 558-578.
As best in
As shown in
As can be seen in
As best seen in
Adjacent to the bore 172, the sealing plate 168 can define second groove 180. A second vacuum chamber 182 can be defined between the second groove 180 and the spindle 60. This second vacuum chamber 182 can be independently coupled to the vacuum source 176. As best seen in
With reference to
As shown in
Disposed immediately before the lead-in portion 224 of the nozzle 45 is a blowback flash suppressor device 226. This device 226 is a chamber 228 having a volume significantly larger than the volume of the lead-in portion 224. Disposed within the volume is a porous stainless steel member 228 which functions as an energy sink to prevent the flame front from traveling up through the nozzle 45,49 and into the gas supply in the event of a system failure.
As shown in
Disposed within the mass flow controller 52 is a normally open valve (not shown) which functions to dump CDA into the fuel supply source should the power be interrupted. Additionally, should the system 20 desire to shut off the processing nozzles 45,49 the normally opened valve is actuated and allows CDA at a pressure higher than the pressure of the fuel source to flow into the processing nozzles 45, effectively extinguishing the flames without the risk of a system explosion.
In operation, fuel is provided to the nozzles 45, through the flash suppressor device 226 from the mass flow controller 52. The vacuum source draws a vacuum in the vacuum chamber 236 preventing corrosive reaction gases from leaking past the nozzle assembly 84.
As seen in
Disposed at a distal end of the elongated cavity 244 is the heating element 246. This heating element 246 can be electrically coupled to a power source which functions to provide electric current to heat the heating element. Alternatively, this element can be inductively heated.
As shown in
In operation, the system for initiating a clean flame, needed in the processing of the wafer 26, includes disposing the heating element 246 within an igniter assembly 78 and energizing the heating element 246 so as to bring the assembly 78 to a predetermined ignition temperature. Gas is then passed through an ignition nozzle assembly 248 at a first gas rate pass the igniter assembly 78 to ignite an initiation flame. The initiation flame is then passed by a plurality of nozzles of a nozzle assembly 84 to ignite a plurality of flames from the nozzles. After the plurality of nozzles of the nozzle assembly 84 have been lit, an air dam is passed in front of the initiation flame by actuating the air knife 250. A non-flammable gas is then passed through the initiator nozzle 248 at a second predetermined rate. In this regard, a second predetermined rate can be greater than the rate of fuel passing through the nozzle. This prevents blow back into the ignition system to the equipment. The use of the air knife 250 allows for the extinguishment of the initiation flame without disruption of the processing flames.
With reference to
As shown in
The spectral region of interest used for flame sensing with H2 and O2 dominated gas mixtures is between about 300 and 325 nm. Emissions around 309 nm is from an intermediate O—H species generated in the flame.
It is envisioned that the mass flow controller 52 of the present system can be coupled to the spectral analyzer 260. In this regard, it is envisioned that should the system determine that one or more nozzles has not be properly emitted, the system will signally fault and can shut the system down. As shown in
The foregoing discussion discloses and describes exemplary embodiments of the present invention. One skilled in the art will readily recognize from such a discussion, and from the accompanying drawings and claims that various changes, modifications, and variations can be made therein without departing from the spirit and scope of the invention.
This application claims the benefit of U.S. Provisional Application No. 60/819,521, filed on Jul. 7, 2006. This application is a continuation-in-part of U.S. patent application Ser. No. 11/131,611, filed on May 18, 2005, which is a divisional application of 10/401,074, filed on May 27, 2003, now U.S. Pat. No. 6,936,546, issued Aug. 30, 2005, which claims priority U.S. Provisional Application 60/376,154, filed Apr. 26, 2002. This application is also a continuation-in-part of U.S. patent application Ser. No. 11/230,261, filed Sep. 19, 2005. This application is also a continuation-in-part of U.S. patent application Ser. No. 11/230,263, filed Sep. 19, 2005. This application is also a continuation-in-part of U.S. patent application Ser. No. 11/417,297, filed May 2, 2006. The disclosure of the above applications are incorporated herein by reference.
| Number | Date | Country | |
|---|---|---|---|
| 60819521 | Jul 2006 | US | |
| 60376154 | Apr 2002 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 10401074 | Mar 2003 | US |
| Child | 11131611 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 11131611 | May 2005 | US |
| Child | 11825659 | US | |
| Parent | 11230261 | Sep 2005 | US |
| Child | 10401074 | US | |
| Parent | 11230263 | Sep 2005 | US |
| Child | 11230261 | US | |
| Parent | 11417297 | May 2006 | US |
| Child | 11230263 | US |
