Patent Application
20090016903
The present invention relates generally to apparatus used in metering fluids with high precision, particularly in fields such as semiconductor manufacturing.
Many of the chemicals used in manufacturing integrated circuits, photomasks, and other devices with very small structures are corrosive, toxic and expensive. One example is photoresist, which is used in photolithographic processes. In such applications, both the rate and amount of a chemical in liquid phase—also referred to as process fluid or “chemistry”—that is dispensed onto a substrate must be very accurately controlled to ensure uniform application of the chemical and to avoid waste and unnecessary consumption. Furthermore, purity of the process fluid is often critical. The smallest of foreign particles contaminating a process fluid causes defects in the very small structures formed during such processes. The process fluid must therefore be handled by a dispensing system in a manner that avoids contamination. See, for example, Semiconductor Equipment and Material International, “SEMI E49.2-0298 Guide For High Purity Deionized Water And Chemical Distribution Systems In Semiconductor Manufacturing Equipment” (1998). Improper handling can also result in introduction of gas bubbles and damage the chemistry. For these reasons, specialized systems are required for storing and metering fluids in photolithography and other processes used in fabrication of devices with very small structures.
Chemical distribution systems for these types of applications therefore must employ a mechanism for pumping process fluid in a way that permits finely controlled metering of the fluid and avoids contaminating and reacting with the process fluid. Generally, a pump pressurizes process fluid in a line to a dispense point. The fluid is drawn from a source that stores the fluid, such as a bottle or other bulk container. The dispense point can be a small nozzle or other opening. The line from the pump to a dispense point on a manufacturing line is opened and closed with a valve. The valve can be placed at a dispense point. Opening the valve allows process fluid to flow at the point of dispense. A programmable controller operates the pumps and valves. All surfaces within the pumping mechanism, lines and valves that touch the process fluid must not react with or contaminate the process fluid. The pumps, bulk containers of process fluid, and associated valving are sometimes stored in a cabinet that also houses a controller.
Pumps for these types of systems are typically some form of a positive displacement type of pump, in which the size of a pumping chamber is enlarged to draw fluid into the chamber, and then reduced to push it out. Types of positive displacement pumps that have been used include hydraulically actuated diaphragm pumps, bellows type pumps, piston actuated, rolling diaphragm pumps, and pressurized reservoir type pumping systems. U.S. Pat. No. 4,950,134 is an example of a typical pump. It has an inlet, an outlet, a stepper motor and a fluid displacement diaphragm. When the pump is commanded electrically to dispense, the outlet valve opens and the motor turns to force flow of displacement or actuating fluid into the pumping chamber, resulting in the diaphragm moving to reduce the size of the pumping chamber Movement of the diaphragm forces process fluid out of the pumping chamber and through the outlet valve.
Due to concerns over contamination, current practice in the semiconductor manufacturing industry is to use a pump only for pumping a single type of processing fluid or “chemistry.” In order to change chemistries being pumped, all of the surfaces contacting the processing fluid have to be changed. Depending on the design of the pump, this tends to be cumbersome and expensive, or simply not feasible. It is not uncommon to see processing systems that use up to 50 pumps in today's fabrication facilities.
The invention pertains generally to high precision pumps for use in dispensing process fluids in applications imposing constraints on handling due to corrosiveness of the process fluid, and/or due to sensitivity to contamination (e.g. from other fluids, particulates, etc.), bubbles and/or mechanical stresses. It is particularly useful for pumps in semiconductor processing operations.
In contradistinction to typical deployments of pumps in such applications, particularly those used for high-precision metering, an exemplary pump employing teachings of a preferred embodiment of the invention is capable of pumping more than one type of chemistry or process fluid without requiring cleaning or changing of surfaces contacting the processing fluid. The pump employs multiple pumping heads, each capable of handling a different type of manufacturing fluid. At least two of the pumping heads share a common actuating mechanism. Although a multi-headed pump might be larger when compared to a pump with a single head, utilizing fewer actuating mechanisms than pumping heads saves valuable space in crowded processing facilities, such as those used for fabricating semiconductor components, which use a large number of pumps. Since actuation mechanisms are sometimes the most complex part of a pump, fewer actuating mechanisms in a factory saves money and maintenance time.
Sharing a single actuating mechanism among multiple heads may seem undesirable, particularly for fluid metering applications. Having a shared actuation mechanism typically means that only one pumping head may be actuated at a time. However, in one exemplary embodiment, the multi-headed pump is capable of fast and frequent switching between pump heads. With actuation between pump heads capable of being switched quickly, there is little delay between demand for dispense and dispense in applications having very short dispense cycles due to relatively small amounts of fluid that are being dispensed.
In the illustrated example, the pumping heads move fluid by drawing it into a pumping chamber and then displacing it. Positive displacement is advantageous for applications requiring precise metering of fluid. The volume of each pumping chamber is increased to suck-in process fluid, and then decreased to push it out. A member that is used to change the volume of a chamber will be called a displacement member. A pumping chamber and displacement member can be implemented a number of different ways. One example includes a piston or piston-like device moving within a cylinder. Other examples include bellows, tubular diaphragms, and rolling diaphragms. The instant example contemplates use of a flexible diaphragm that cooperates with the walls of the pumping chamber to displace fluid. Moving the diaphragm in one direction increases the volume of the pumping chamber, and moving it another direction decreases the volume of the pumping chamber, thus displacing fluid from it. The diaphragms for pump heads 113, 115 and 117 are schematically illustrated in the figure as elements 131, 133 and 135, respectively.
A number of different arrangements can be used to ensure that fluid flows only in one direction through the pump head. In the illustrated example, the pump head includes an inlet (not indicated) for coupling the pump head to a process fluid source, such as sources 101, 103 or 105, and an outlet (not indicated) for coupling the pump head to a dispense point, such as dispense point 107, 109 or 111. The pumping chamber in the pump head has at least one opening, and preferably at least two openings, one being in communication with the inlet and the other in communication with the outlet. Fluid is drawn into the pumping chamber through the inlet opening and is expelled through the outlet opening. This allows for creation of a generally unidirectional flow of process fluid through the pumping chamber, which can assist in reducing pooling of process fluid and accumulation of contaminants in the pump head. The inlet and outlet of each pump head is coupled through valving that ensures, at least normal operation, that fluid flows into the pumping chamber only from the inlet and exits the pumping chamber only through the outlet.
The valving can take different arrangements, depending in part on the number of openings into the pumping chamber and other considerations. In the illustrated example, the valving is comprised of two valves. Check valve 137 ensures one-way flow from the inlet into the pumping chamber, and check valve 139 ensures one-way flow of process fluid exiting the chamber through the outlet. The check valves are self-actuating or lifting, which tends to reduce complexity by avoiding having to implement a mechanism for synchronizing their opening with the pumping action of the pump head. However, it might be advantageous in some circumstances, such as those described below, to incorporate valves whose opening can be independently controlled. Furthermore, use of check valves may not be appropriate for some applications. If the pumping chamber has only one opening, one example of suitable valving includes a three-way valve that selectively couples either the inlet or outlet to the opening, or closes the opening altogether, depending on the stroke of the pump. Other types of valving could be chosen to achieve the same functionality, although possibly at the expense of greater complexity and less reliability.
The plurality of pumping heads share a common actuation mechanism, represented in the figure by drive motor and piston assembly 135. An actuating mechanism includes a force generating component, such as a motor, and a coupling for communicating the force to a fluid displacement member. Sometimes, these components are one and the same. Examples of actuating mechanisms include mechanical, pneumatic and hydraulic mechanisms, and combinations of them. One example of a mechanical actuator is a driver motor coupled to a diaphragm through a purely mechanical coupling, such as a transmission or other mechanical linkage or piston. The linkage or piston converts the output of the motor into movement of the fluid displacement member. A hydraulic coupling can also be used, with the motor moving a piston, which in turn moves hydraulic fluid that pushes against the displacement member. In a purely pneumatic system, for example, gases under high pressure are used to move the displacement member.
In the illustrated example, the force generated by the common actuating mechanism is preferably applied in parallel, rather than serially, to each of the pump heads. Although applying the force in parallel will lead all pump heads to actuate simultaneously, avoiding serial application of the force reduces the complexity by avoiding a mechanism for selectively applying or switching the actuation force between the pump heads. Complexity tends to increase costs and reduce reliability.
In order to avoid undesirable, simultaneous actuation of all pump heads, yet maintain simplicity, the actuating mechanism in the illustrated example preferably utilizes a fluidic coupling for communicating forces from a motor or other force generating mechanism to the process fluid. The drive assembly for the actuation mechanism in the illustrated example includes a drive motor (not shown) for supplying force for moving the actuating fluid. The drive motor moves a displacement member (e.g. a piston) that, in turn, moves fluid in a manner that causes the pumping head to actuate. Actuating fluid is moved in and out of a chamber on the side of the diaphragm opposite the pumping chamber. Displaced actuating fluid moves into the pumping head, reducing the volume of the pumping chamber and pushing fluid out. Reverse movement of the displacement member causes the actuating fluid to flow from the pumping head, increasing the volume of the pumping chamber and consequently drawing in process fluid. If the fluid is not compressible at least at the pressures at which the pump functions (such fluid being referred to herein as incompressible), and only one pumping chamber is open, the amount of actuating fluid displaced by actuating assembly is proportional to the amount of process fluid displaced from within the pumping chamber.
Blocking flow of process fluid out of the pumping chamber of a pump head in effect blocks the flow of actuating fluid into the pump head, thus causing actuating fluid to be redirected to, and to flow into, another pump head without internal valving to redirect the fluid to different pump heads. Therefore, although internal valving could be used, it is not required in order to ensure only one head is pumping at a time. In this example, a preexisting valve at the outlet—a valve that would otherwise be present for this application—is sufficient, therefore allowing reduction in complexity and the size of the pump without a corresponding increase in the number of external valves that would otherwise be required. Furthermore, existing external valving can be utilized for blocking process fluid flow through the pump heads. In the illustrated example, which uses self-actuating check valves, output valves 119, 121 and 123 are selectively closed to block flow of fluid from the pump heads that are not intended to be pumping during actuation of the pump. The output valves may be located anywhere along the line carrying fluid from the pump head to the dispense point. A controllable valve can be substituted for one or both check valves, or used in addition to them, if an output valve is not available or if there is a preference not to use the output valve. However, this would be at the expense of more cost and complexity. Furthermore, other valving arrangements that are used to ensure one way flow of process fluid through the pump head, such as the three-way valve mentioned above, can also be used for this purpose.
When used for metering fluids, the pump is operated so that only one pump head is active at a time. All actuating fluid is thereby directed only into or out of the active pump. By allowing actuating fluid to flow only out of one pump head at a time, the amount of process fluid being pumped is determined from movement of the displacement member within the actuation mechanism. If more than one pump head is opened for pumping during actuation, a mass flow meter is coupled with the pump head to determine the amount process flowing out of the pump head. However, in applications such as semiconductor manufacturing, dispense cycles are short and demand for dispense from a particular dispense point is not constant and, in some cases, relatively infrequent. Given the absence of internal valving for redirecting the actuating fluid and the simplicity of the mechanism controlling flow of process fluid through a pump head, fast activation of pump heads is possible, thus allowing the actuating fluid to be, in effect, time multiplexed to the pump heads without unduly slowing dispensing.
Referring now to
The body in the illustrated example possesses a square cross-section with four sides. Formed on three of the four sides are faces to which the pumping heads are coupled. The fourth side is used, in this example, to receive a pressure sensor 210. The pressure sensor is used to measure the pressure of hydraulic fluid within the actuation mechanism. Arraying the pumping heads at least partially around channels supplying hydraulic actuation fluid tends to result in more efficient utilization of space as compared to, for example, a configuration in which the heads are arranged in a linear fashion. However, other of the advantages of the exemplary pump illustrated in these figures can be achieved without the pumping heads being arrayed around the central axis. For example, the pumping heads can be arranged in a stacked configuration. More pumping heads can be coupled to the central body by increasing the cross-sectional size, increasing the number of faces disposed around a central axis of the central body, by reducing the size of the pumping heads, and/or by extending the body along its central axis. The size of the pumping head depends in part on the desired volume of the pumping chamber. Preferably, the size of the pumping chamber is such that multiple, incremental dispenses, in which only a portion of the process fluid within the pumping chamber is dispensed during a dispense cycle, are completed before having to draw in more fluid. A face need not be flat, but can be curved if desired. Thus, for example, the central body can have either a polygonal or a generally circular cross section. Although a circular cross-section may take up less space, flat faces have the advantage of a simpler fabrication and connection with the pumping head.
The central body preferably also houses, as in this example, at least one hydraulic actuation mechanism. The mechanism includes a fluid reservoir as well as a displacement element. In the illustrated embodiment, the actuation fluid reservoir is comprised of a cavity 207 (see
In the illustrated embodiment, pumping head structures 202, 204 and 206 are coupled respectively with a face portion 211 formed on each of three side walls of body 208.
In each of the pumping head structures, diaphragm 212 extends across the face portions and cooperates with a pumping head to define a pumping chamber 214 on one side of the diaphragm, and with a depression 216 formed in the body, at the face portion, to define an actuating fluid chamber 218 on the opposite side of the diaphragm. In this preferred embodiment of the exemplary pump the diaphragm can be easily removed and replaced by removing the pumping head assembly. The diaphragm is sealed against the cooperating face of body by O-ring seal 220. Plate 222 attaches the diaphragm to the face of the body. Among other advantages, attaching the diaphragm with the plate allows the pump to be built and charged with actuation fluid—preferably a substantially incompressible fluid (at least at the pressures typically encountered in the application), such as glycol—prior to the pump heads being assembled with the body. The diaphragms are preferably made from a translucent material in order to permit visual identification of any air or gas bubbles within the actuation fluid prior to attaching the pump heads. Although one diaphragm per pump head is being used in the illustrated embodiment, two or more adjacent pump heads could instead use a different area of one, larger diaphragm, isolated by a seal or other structure so that process fluid does not leak between the pump heads. Vent line 223 permits air to be purged from the actuation fluid chamber 218 in each pumping head. Vent lines 223 are sealed with plugs that are not shown in the figures. Air entrapped in the actuation fluid and/or process fluid, pumping chamber, actuation fluid chamber, cavity 207, or any of the channels within the pump carrying the fluids, can also be detected by charging the pumping chambers with process fluid, closing each of them so that process fluid cannot flow out, pumping the actuation fluid and monitoring the pressure of the actuation fluid using pressure sensor 210. Because air bubbles are compressible, the measured pressure will be less than expected if a substantial amount of air is entrapped in the system.
Each pump head structure 202, 204 and 206 is an assembly that includes a pumping chamber cover 224 with a cavity or depression 226. The cover cooperates with the diaphragm 212 to form pumping cavity 214. O-ring 225 forms a seal between the cover 224 and the diaphragm mounting plate 216. Inlet orifice 228 and outlet orifice 230 extend through cover 224 for permitting flow of process fluid into and out of, respectively, the pumping chamber. The inlet orifice is located near the bottom of the pumping chamber so that fluid flows upward, against gravity, when the pump is in a normal operating position, toward the outlet orifice. This arrangement and the elongated form of the pumping chamber tends to reduce pooling of process fluid within the pumping chamber and encourages migration of bubbles toward the outlet to assist with purging. The generally curved shape of the depression 226 and obtuse angles at the junctions of straight surfaces within the pumping cavity avoid sharp corners in which process fluid and micro-bubbles might collect and be difficult to purge, thus further reducing the risk of entrainment of bubbles during normal operation.
Each pump head structure includes connectors for connecting lines carrying process fluid into and out of the pump head. In order to save space, the connectors are preferably oriented in a direction that is generally parallel to the elongated axis of the pumping chambers and the body 208. If oriented with their axes perpendicular to the axis of the body 208, the pump 200 would occupy more space in lateral directions, and additional space would be required to accommodate the process fluid lines that will be connected to the inlet and outlet connectors. Inlet fitting 232 and outlet fitting 234 are threaded into a connector block 236. The illustrated inlet and outlet fittings are examples of flare type fittings typical in semiconductor manufacturing. They are intended to be representative generally of fittings for connecting lines to the pump. Other types of fittings can be used, depending on the application. Other examples of high purity fittings used in the semiconductor industry include Super Type Pillar Fitting® and Super 300 Type Pillar Fitting® of Nippon Packing Co., Ltd., Flowell® flare fittings, Flaretek® fittings from Entegris, “Parflare” tube fittings from Parker, LQ, LQ1, LQ2 and LQ3 fittings from SMC Corporation, Furon® Flare Grip® fittings and Furon® Fuse-Bond Pipe from Saint-Gobain Performance Plastics Corporation. The connector block 236 and the cover 224 are, in this example, fabricated separately and assembled into a head assembly. However, the assembly could be fabricated using fewer or more components.
The connector block 236 includes a passageway that carries fluid from the inlet into the connector block toward the inlet orifice 228 of the pumping chamber. In this example, the passageway is formed by a channel 238 formed on the surface of a block and a cooperating gasket 240. The gasket also seals the pumping chamber cover 224 with the connector block 236. A hole 242 allows fluid to flow into channel 244 (see
In the illustrated example, a one-way check valve 246 is integrated into the connector block that allows fluid to flow only from the inlet fitting 232 to the pumping chamber. The check valve is inserted into the same bore as the inlet fitting 232. It is comprised of an orifice plate 248 and an umbrella-shaped valve 250 that cooperates with the orifice plate 248. The valve's stem attaches the valve to the orifice plate. Fluid flowing under pressure through the holes in the orifice plate, toward the valve, tends to cause the edges of the valve to curl up or lift, while the center of the valve remains stationary. The valve has an inverted shape. When it is assembled, the stem pulls the edges of the valve against the orifice plate, thereby creating a seating force that presses the perimeter of the valve against the plate. This forms a good seal. More details about this particular type of check valve can be found in commonly assigned U.S. patent application Ser. No. 11/612,408, filed on Dec. 18, 2006, which is incorporated herein by reference.
The connector block also includes a passageway that carries fluid exiting pumping chamber 214 to the outlet connector 234. It also incorporates a one-way check valve 252 that allows fluid flow in the direction of the outlet connector. Check valve 252 is substantially similar to check valve 246. It includes an orifice plate 254 that sits in a recess 255 formed on the back of pumping chamber cover 224. Umbrella-shaped valve 256 is attached to the orifice plate 254. Fluid flowing out of the pumping chamber 214, through the outlet orifice 230, flows through the check valve 252 and into a passageway that connects with outlet fitting 234. That passageway is formed in part by channel 258, formed in one surface of connector block 236, and cooperating gasket 240. Segment 260 (see
Incompressible actuating fluid is stored in the central chamber or cavity 207 of the actuating mechanism. When piston 209 translates within the cavity 207, passageways 264 communicate fluid between the cavity and an actuating fluid chamber 218, associated with each of the pumping heads 202, 204 and 206. Fluid is capable of moving in parallel between the cavity 207 and each actuating fluid chamber 218. Therefore, actuating fluid will, unless otherwise stopped, flow into each actuating chamber 218 when the piston displaces actuating fluid from the cavity 207. Similarly, actuating fluid will, unless otherwise stopped, flow out of the actuating fluid chamber 218 associated with each pump head when the piston is retracted, causing the actuating fluid to be drawn into the cavity 207.
Assuming that the pumping chamber 214 and the corresponding actuating fluid chamber 218 contain no gas, air or other compressible substance, flow of fluid through a given passageway is controlled in the illustrated embodiment by whether the diaphragm is permitted to move in the corresponding pump head. If it cannot move, actuating fluid will tend not to flow in either direction through the passageway between the cavity 207 and the actuating fluid chamber 218 that is associated with that diaphragm. Whether a diaphragm moves depends on whether process fluid can be drawn into the pumping chamber 214 during flow of actuating fluid out of the actuating fluid chamber 218, and whether it can flow out of the pumping chamber during flow of the actuating fluid from the cavity 207 and into the actuating fluid chamber 218. Given that process fluid can only flow in one direction through the pumping chamber of the illustrated embodiment, opening and closing a valve (not shown in these figures) located in the outlet flow path for process fluid from the pumping chamber 214 will thus determine whether the diaphragm can be moved to displace the process fluid in the pumping chamber, which in turn determines whether actuating fluid flows into the actuating fluid chamber for the given pump head. By opening the outlet of only one pumping head, all the actuating fluid caused by displacement of piston 209 will be forced to flow into only the actuating fluid chamber of the pump head with the open outlet. The volume of actuating fluid displaced by movement of piston 209 will equal the volume of process fluid displaced by the diaphragm of the pump head with the open outlet. In other words, there is a linear relationship between the movement of the piston and the volume of process fluid pumped.
As process fluid is always permitted to flow into each of the pumping chambers in the illustrated embodiment, actuating fluid will always flow from each actuating fluid chamber 218 during retraction of piston 209, at least until the diaphragm reaches the surface of the wall forming depression 216 for that particular process fluid chamber. The wall forming depression 216 preferably includes a channel 217 to ensure that the diaphragm is pulled evenly against the wall. Thus, the illustrated embodiment of pump 200 will simultaneously recharge, or will recharge in parallel, each pumping chamber in the pump, regardless of the number of pumping heads.
Piston 209 include a sliding seal 262. Displacement of the piston within cavity 207 is preferably controlled by a stepper motor 264, which turns a drive screw 266. Clamp 268 attaches the drive screw to output shaft 270 of the motor. Thrust bearing 272 prevents the drive shaft from axially loading the output shaft of the motor. The threads on the drive screw 266 couple with threads on the inside of the piston 209. The angular position of the piston is fixed by a guide 274, which is clamped to the piston and cooperates with slot 276 to prevent rotation of the piston. Turning the drive screw moves the piston. Other types of mechanisms for translating the piston could, however, be substituted. An optical sensor 278 detects when guide 274, and thus piston 209, is at a predetermined limit during upstroke. This is used to calibrate the pump. Cover 280 seals an opening that allows access to the cavity 207 for assembly and cleaning.
For semiconductor and other high purity applications, it is preferred that all surfaces of the pump that contact the process fluid are made of non-contaminating or non-reacting material. One example of such a material is polytetrafluoroethylene, which is sold by DuPont under the trademark Teflon®.
An exemplary application of multiple head dispense pump 200 is illustrated by
The outlet fitting 234 (see
Operation of the pump 200 and dispense valves 312 is controlled by a controller 314. Preferably, the controller is programmable, microprocessor-based, but could be implemented using any type of analog or digital logic circuitry. The same controller can be used to control more than one multi-head pump. The controller typically receives a demand for dispense signal from a manufacturing line, where the wafer is being processed. However, the control processes can be implemented in the line controller or other processing entity associated with the fabrication facility.
When the controller receives a request for dispense of process fluid, as indicated by blocks 402, 404, and 406, the controller signals the other interfaces that the pump is busy and sets a flag indicating that dispense is active for that interface. Thus, if the request is received on interface 1, the controller communicates to interfaces 2 and 3 at step 408 that the pump is busy, so that production tracks or lines that communicate with it know that dispense is not available. It also sets at step 410 a stored flag, dispense 1, active. Similarly, if a dispense request is received on interface 2, a pump busy signal or state is communicated to interfaces 1 and 3 at step 412 and a dispense 2 flag is set active at step 414. Finally, if the request for dispense is received on interface 3, the pump busy signal or state is communicated to interfaces 1 and 2 at step 416, and the dispense 3 flag is set active at step 418.
As indicated by decision step 420, the controller determines whether there is an optional dispense delay set up or programmed for that interface. In a dispense delay, as indicated by steps 422, 424 and 426, the dispense valve corresponding to the active dispense flag is opened for a predetermined period of time prior to the pump being actuated. This might be used in applications in which, for example, it is desirable for the rate of dispense to start slow and then increase. If there is no dispense delay, the pump is started at step 428. The controller can be set up or programmed to open the dispense valve corresponding to the active dispense flag either immediately or after a predetermined or programmed delay, as indicated by steps 430, 432 and 434.
Once the dispense valve is opened and the pump is started, the controller actuates the pump so that a preset or otherwise determinable amount of process fluid is dispensed at a predefined rate or rates (the rate can be varied by, or be a function of, time and/or other parameters, if desired), as indicated by step 436. In the embodiment illustrated in
Once suck-back is completed, an end of dispense state or signal is communicated to the interface with the active dispense flag, as indicated by steps 472, 474, 476, 478, 480, and 482. The controller then waits for the interface to release the dispense, as indicated by steps 484, 486, and 488. The release occurs when the track or line controller signals acknowledge the end of dispense.
When the interface releases the dispense, the controller clears all dispense flags at step 490, communicates to all dispense interfaces that the pump is busy at step 492, and recharges the pump at step 494. To recharge the pump, the stepper motor is stepped in a direction opposite of the direction it is stepped for dispense, until the pumping chambers in each pump are fully charged. In the embodiment illustrated in
Referring now to
In each of the examples of a two-stage pumping system, a pumping chamber 506 is used as a first stage, and pumping chamber 508 is used as a second stage. The volume of each pumping chamber is changed to draw in and expel process fluid using a diaphragm, bellows, rolling diaphragm, tubular diaphragm or other arrangement. In examples 500, 502, and 504, pumping chambers 506 and 508 can be two different heads of a multi-headed pump, such as the one described in
The first stage of the pump is used to pull fluid from a source 509 and push it to a filtering system, generally designated by filter 510. The second stage is used for pulling the fluid from the filtering system and dispensing it, in a metered fashion, onto, for example, a wafer 512. Fill valve 513 is opened to allow fluid to be drawn from the source 509 and into the first stage, and then closed when the first stage pumps. The fill valve can be alternatively implemented as a check-valve. The filtering system typically includes a vent, controlled in these examples by valve 514, and a drain, controlled in these examples by valve 516. Each of the examples also includes a dispense valve 518, for controlling dispensing, and an optional suck-back valve 520. Each of the two-stage pumping systems in the examples includes a valve 522 for preventing reverse flow of processing fluid from the pumping chamber 508. A check valve is preferred. Two-way and other types of valves can be substituted for the check valve, but they will need to be opened and closed synchronously with the operation of the pumping system, thereby complicating the control processes. Each two-stage pumping system includes a recirculation loop 521 that is opened and closed by recirculation valve 523. The two, two-stage pumping systems 505 shown in
The two-stage pumping systems 500 and 505 shown in
In all examples 500, 502, 504, and 505, multiple pumping chambers are driven by a single drive mechanism, which in these examples is comprised of stepper motor 526, turning a screw 528, which in turn causes translation of a piston within cylinder 530. In the two-stage pumping systems 500, 502, and 504, each drive mechanism is coupled in parallel to pumping chambers 506 and 508. In the two-stage pumping systems 505, shown in
For semiconductor and other high purity applications, it is preferred that all surfaces of the pump that contact the process fluid be made of non-contaminating or non-reacting material. One example of such a material is polytetrafluoroethylene, which is sold by DuPont under the trademark Teflon®. Other examples include high density polyethylene and polypropylene, and PFA (perfluoroalkoxy copolymer resin).
The drive mechanism operates substantially similarly to the actuation mechanism described in connection with
In two-stage pumping systems 500, 502 and 505, shown in
The operation of the two-stage pumping systems, which is described below, is controlled by one or more controllers, executing predetermined control routines to open and close the various valves and to cause turning of the motor of the drive mechanism.
Referring now only to
The two-stage pumping system of
Each of the two, two-stage pumping systems 505 in
Valves 532 and 534 are optional for each of the drive mechanisms, although they can provide greater control and accuracy. Furthermore, no valve 536 on the outlet of the first stage pump is required when valves 532 and 534 are omitted, since the first stage of each of the two pumping systems is being operated independently of the second stage of each of the two pumping systems. However, if the reservoirs or filters of the respective of the two-stage pumping systems 505 need to be filled independently, then an output valve, like valve 536, would be desirable to have.
The foregoing description is of an exemplary and preferred embodiment of a multiple dispense head pump employing at least in part certain teachings of the invention. The invention, as defined by the appended claims, is not limited to the described embodiments. Alterations and modifications to the disclosed embodiments may be made without departing from the invention. The terms used in this specification are, unless expressly stated otherwise, intended to have ordinary and customary meaning and are not intended to be limited to the details of the illustrated structures or the disclosed embodiments. None of the descriptions in the present application should be read as implying that any particular element, step, or function is an essential element which must be included in the claim scope. The scope of patented subject matter is defined only by the allowed claims. Moreover, none of these claims are intended to invoke paragraph six of 35 USC §112 unless the exact words “means for” or “steps for” are followed by a participle.
