Precision screen printing with sub-micron uniformity of metallization material on green sheet ceramic
By combining precision screen printing and symmetrical ink printing (SIP) methods with electric actuators and encoder control, the problems of unevenness and non-repeatability of metallization materials on green ceramic sheets have been solved, enabling high-quality and efficient production of electrostatic chuck (ESC) equipment.
Patent Information
- Application Number
- CN202111293609.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Priority Date
- 2017-08-02
- Filing Date
- 2017-08-04
- Publication Date
- 2026-01-16
- Estimated Expiration
- 2037-08-04
AI Technical Summary
Existing technologies in the fabrication of electrostatic chucks (ESCs) for semiconductor chip manufacturing suffer from inhomogeneity and non-repeatability of the metallization material printed on green ceramic sheets, leading to inconsistent temperatures in circuits and structures, which affects manufacturing quality and yield.
By employing precision screen printing technology, metallization materials are printed on green ceramic sheets. Electric actuators and encoders are used to control the blade gap. Combined with the Symmetric Ink Printing (SIP) method, printing machine hardware deviations are identified and compensated. A knowledge-based model is established to achieve sub-micron level uniformity and repeatability.
It achieves highly uniform and repeatable printing of metallization materials on green ceramic sheets, improves the temperature uniformity and production consistency of ESC equipment, reduces production costs and scheduling complexity, and improves manufacturing quality and output.
Smart Images

Figure CN114005782B_ABST
Abstract
Description
[0001] This application is a divisional application of Chinese Patent Application No. 201780048820.X, filed on August 4, 2017, entitled "Precision Screen Printing with Sub-micron Uniformity of Metallization Material on Green Sheet Ceramics", and claims priority to U.S. Provisional Application Serial No. 62 / 371,636, filed on August 5, 2016, entitled "Precision Screen Printing with Sub-micron Uniformity of Metallization Material on Green Sheet Ceramics", and U.S. Patent Application Serial No. 15 / 667,281, filed on August 2, 2017, entitled "Precision Screen Printing with Sub-micron Uniformity of Metallization Material on Green Sheet Ceramics", the priority of which is hereby claimed.
[0002] CROSS-REFERENCE TO RELATED APPLICATIONS
[0003] This application claims priority to U.S. Provisional Application Serial No. 62 / 371,636, filed on August 5, 2016, entitled "Precision Screen Printing with Sub-micron Uniformity of Metallization Material on Green Sheet Ceramics", and U.S. Patent Application Serial No. 15 / 667,281, filed on August 2, 2017, entitled "Precision Screen Printing with Sub-micron Uniformity of Metallization Material on Green Sheet Ceramics", the priority of which is hereby claimed. TECHNICAL FIELD
[0004] The present invention relates to the field of carriers for workpieces used in the fabrication of microelectronic and micromechanical devices, and in particular to forming such carriers using screen printing on green sheet ceramics. BACKGROUND
[0005] In the fabrication of semiconductor chips, workpieces such as silicon wafers or other substrates are exposed to various different processes in different processing chambers. The chambers can expose the wafers to a number of different chemical and physical processes, thereby creating microscopic integrated circuits and micromechanical structures on the substrate. Layers of material that make up the integrated circuits are created by processes including chemical vapor deposition, physical vapor deposition, epitaxial growth, and the like. Some patterning of the layers of material is performed using photoresist masks and wet or dry etching techniques. The substrate can be silicon, gallium arsenide, indium phosphide, glass, or other suitable materials.
[0006] The processing chambers used in these processes typically include a substrate support, susceptor, or chuck to support the substrate during processing. In some processes, the susceptor can include an embedded heater to control the temperature of the substrate, and in some cases, to provide elevated temperatures that can be used in the process. Electrostatic chucks (ESCs) have one or more embedded conductive electrodes to create an electric field using electrostatic forces to hold the wafer against the chuck.
[0007] The ESC will have a top plate, referred to as a puck, a bottom plate or base, referred to as a susceptor, and an interface or joint that holds the two together. The top surface of the puck has a contact surface that holds the workpiece, which can be made of various materials (e.g., polymer, ceramic, or a combination thereof), and can have a coating over all or over selected locations, etc. Various components are embedded into the puck, including electronic components for holding or clamping the wafer and thermal components for heating the wafer.
[0008] Because the circuits and structures formed on the workpiece are very small, the thermal and electronic environment provided by the workpiece support must be very precise. When the temperature on the workpiece is not uniform or consistent, the circuits and structures will vary. If one support is different from another support, the circuits and structures will vary from different supports. For extreme cases, the processes can need to be adjusted for different supports. This directly impacts the quality and yield of the circuits and structures fabricated on the workpiece. Therefore, the disc with embedded components, thermal components, and electronic components has strict dimensional requirements, both planar and vertical, to ensure consistent performance not only on the surface of a particular ESC, but also from one ESC to another. SUMMARY
[0009] Precision screen printing is described that enables metallization material printed on a green sheet ceramic to have sub-micron uniformity. In some examples, a disc having electrical traces is formed by screen printing a paste containing metal on a ceramic green sheet in a pattern of electrical traces; and processing the printed green sheet to form a disc of a workpiece carrier. In some instances, the printing includes applying a squeegee of a screen printing machine to the printed green sheet in a squeegee direction while the green sheet is on a printer bed of the screen printing machine. The method further includes mapping the printer bed at a plurality of locations along the squeegee direction, identifying non-uniformities in the mapping of the printer bed, and modifying a printer controller of the screen printing machine to compensate for the mapped non-uniformities in the printer bed. BRIEF DESCRIPTION OF DRAWINGS
[0010] Embodiments of the application are illustrated by way of example, and not by way of limitation, in the accompanying drawings wherein:
[0011] Figure 1 is a cross-sectional side view of an electrostatic chuck (ESC) apparatus structure according to an embodiment.
[0012] Figure 2 is a top view of a printer setup for symmetric ink printing (SIP) according to an embodiment.
[0013] Figure 3 is a cross-sectional side view of a conventional squeegee setup for SIP.
[0014] Figure 4 is a cross-sectional side view of an adjustable printer according to an embodiment.
[0015] Figure 5 is a plot of motor encoder transducer feedback versus exposure sequence according to an embodiment.
[0016] Figure 6 is a plot of printed ink thickness versus motor encoder according to an embodiment.
[0017] Figure 7 is a plot of printed ink thickness versus target heater trace resistivity for each design according to embodiments.
[0018] Figure 8 is an isometric view of a printer and squeegee rail to show adjustment points according to embodiments.
[0019] Figure 9 is a top view of a printer setup for symmetric ink printing (SIP) with multiple squeegee positions according to embodiments.
[0020] Figure 10 is a plot of motor encoder values versus each of eight different positions along a squeegee path according to embodiments.
[0021] Figure 11 is a plot of printed ink thickness versus number according to embodiments.
[0022] Figure 12 is a plot of ink thickness versus squeegee gap according to embodiments.
[0023] Figure 13 is a plot of printed ink thickness versus batch size of printed green sheets according to embodiments.
[0024] Figure 14 is a plot of printed ink thickness or printed ink resistivity versus mask thickness design parameter identified as X according to embodiments.
[0025] Figure 15 is a table of data collected when designing a printing screen and squeegee gap configuration according to embodiments.
[0026] Figure 16 is a process flow diagram of a method for highly repeatable squeegee setup according to embodiments.
[0027] Figure 17 is a process flow diagram of a method for conformal precision screen printing according to embodiments.
[0028] Figure 18 is a process flow diagram of a compensation method for ink thickness trends according to embodiments.
[0029] Figure 19 is a process flow diagram of a method for screen mask design according to embodiments.
[0030] Figure 20 is an isometric view of an assembled electrostatic chuck suitable for use according to embodiments.
[0031] Figure 21is a schematic diagram of a plasma etching system including a susceptor assembly suitable for use in accordance with embodiments. DETAILED DESCRIPTION
[0032] In the following description, numerous specific details are set forth to provide a thorough understanding of the present application. However, it will be apparent to one skilled in the art that the present application can be practiced without such specific details. In some instances, well-known methods and apparatuses have not been described in detail in order to avoid obscuring the present application. References to "an embodiment" or "one embodiment" in the context of this specification mean that a particular feature, structure, function, or characteristic described in connection with the embodiment is included in at least one embodiment of the present application. The appearance of the phrase "in an embodiment" or "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment. In addition, the particular features, structures, functions, or characteristics can be combined in any suitable manner in one or more embodiments. For example, where a particular feature, structure, function, or characteristic is not explicitly described in connection with both embodiments, it can be combined with either embodiment.
[0033] As described herein, the top plate can be made into a chuck, susceptor, or carrier for supporting a workpiece (e.g., a silicon or other wafer) in a carrier. The top plate can be formed of a ceramic with embedded electronic components, and provide very high precision in the shape and size of the embedded components. This provides better control of process parameters on the workpiece. The components between the top plates are also more consistent. When the top plates wear out and are replaced, this provides more consistent production results. As a result, smaller and more precise features can be formed on the workpiece with higher quality and uniformity, reducing cost, increasing production volume, and reducing downtime to adjust production parameters.
[0034] A method for screen printing with sub-micron uniformity is disclosed herein. The method is applicable to screen printing of certain materials that require precise printing thickness and uniformity throughout the screen mask pattern. The application shown is printing metallization material on green sheet ceramic, which is a core process for making uniform heater trace patterns used in electrostatic chuck (ESC) equipment. ESCs are particularly important for chip device performance and wafer yield for semiconductor processing.
[0035] An ESC with very high temperature uniformity benefits from very highly uniform printing of metallization material on green sheet ceramic. Any material including metallization material printed on a surface will be referred to herein simply as ink. In the example herein, ESC disks are made of multiple green sheets, and some of them are printed with ink in heater trace and electrostatic electrode patterns. The printing uniformity can be on the order of sub-microns.
[0036] Figure 1is a cross-sectional side view of an electrostatic chuck (ESC) apparatus structure. The ESC 2 has a cooling base 4 made of a thermally conductive material such as aluminum or an alloy. There are input and output ports 6 and 8 coupled to external thermal fluid pumps and heat exchangers. Within the cooling base, these ports are coupled to internal cooling channels 20 that circulate thermal fluid through the base to help control the temperature of the ESC. A ceramic puck 12 is attached to the cooling base 4 by a thermally conductive bonding layer 10. The bonding layer is typically an adhesive, but other materials can be used. A workpiece 14 such as a silicon wafer, a glass sheet, a gallium arsenide wafer, or other workpiece is held in place on the puck 12 by electrostatic charge.
[0037] The electrostatic charge is generated and maintained by clamping electrodes 16 within the puck. The electrodes can be charged or discharged using contacts (not shown) on the sides or bottom of the puck. Heater traces 18 can also be formed within the puck to heat the workpiece using, for example, resistive heating from the conductive traces. Multiple traces can be used to apply different amounts of heat through different traces to more precisely control the temperature of the puck, and thus the wafer.
[0038] The ceramic puck can be formed from green sheets that are ink printed, laminated, machined, sintered, polished, and have surface features created in a series of complex processes. Interconnects for the two heater traces and electrodes in the puck are not shown. The quality of the pattern of the heater traces 18 affects the temperature uniformity performance of the ESC.
[0039] The ceramic green sheets can be formed in any of a variety of different ways. In some embodiments, 90-96% of ceramic powder (such as AI2O3 or AIN and glass) is compacted under high pressure with a binder (such as a plasticizer) and then sintered in a short time to form a flexible material that can be handled at room temperature and then hardened later by sintering. The green sheets can have any of a variety of different thicknesses. For example, the green sheets can be from 0.05 mm to 0.5 mm thick and carried in stainless steel frames for handling purposes.
[0040] Alternatively, a green body or any other flexible ceramic material can be used. The green body in some examples is a monolithic compact block of ceramic powder with other fillers and a binder. After forming as described below, it can be sintered in a furnace under heat and pressure. This process can also be applied to the green sheets as described below.
[0041] As described herein, heater trace patterns and other structures can be formed by printing a paste onto a green sheet. The green sheets can be stacked with other green sheets, such that the paste is embedded between the green sheets. Different pastes can be dispensed into different patterns to form different types of electronic components. The pastes are used to form embedded or surface conductive components. The pastes contain a suitable conductive material, such as a metal like tungsten, molybdenum, zinc, silver, gold, or a suitable refractory material, for example, in powder form, and are carried in a suitable suspension and dispersant. The pastes are dried, sintered, or cured with the stacked green sheets to form the finished disc.
[0042] The stacked green sheets are pressed together with sufficient pressure to form a single, stable structure. The paste is dried to be sufficiently hard to be stacked and compacted without too much distortion. The stack of individual ceramic green sheets with the desired thermal and electronic components is then sintered to consolidate the separate sheets into a single solid entity. Thereafter, additional finishing processes, such as machining, grinding, polishing, blasting, washing, etc., can be used to convert it into the final top plate or disc. Because the metal or refractory material can be dispensed onto the intermediate sheet, these features can be embedded into the resulting structure after the sheets are pressed together.
[0043] To have repeatable and consistent quality of the ESCs in a production batch or from various batches, the described method first identifies some hidden printer hardware biases for eliminating or minimizing the root causes of later difficult-to-identify system printing biases.
[0044] Figure 2 is a diagram of a printer setup for symmetrical ink printing (SIP) to show the different possible directions of bias. The printer 22 includes a printing table 24 that carries a green sheet 26 to be printed. The printer brushes the workpiece with a squeegee 28 that passes over the workpiece.
[0045] The main setup biases are first identified and minimized. The center of both the screen mask and the squeegee can be aligned with the center of the heater trace pattern. Specifically, the symmetrical alignment between the heater trace pattern center, the screen mask center, and the squeegee center is managed with tight tolerances. The center of the heater trace pattern is typically located on the center 30 of the screen mask. On the printer, the screen mask is located above the workpiece 26. When the squeegee is pulled across the screen, the ink is pushed by the squeegee through the screen to form the ink pattern on the workpiece. This is shown as starting from a start position and ending at an end position, with the squeegee shown as positioned at the end position in the drawing. The printed pattern that is extruded through the screen mask is the pattern of the heater and other electrical traces on the workpiece.
[0046] The center of the installed squeegee 32 can also be aligned with the center 30 of the heater trace pattern of the screen. Any hidden or overlooked hardware deviations can cause print deviations. Such deviations can invalidate the heater trace design. The print machine hardware calibration principle is then applied to a method referred to herein as Symmetric Ink Printing (SIP) of a metallized material.
[0047] Conventional printing methods use some shimming device to set the squeegee for each production order. Figure 3 is a diagram of a conventional squeegee setup. A shimming device has been added using a setup method. The print machine 302 has a print bed 304 and a sample, coupon, or test workpiece 306 is placed on the bed. A mask 308 is placed over the workpiece. A squeegee 310 is mounted on a carriage or rail 312 and a test is then run. A shimming device 314 is added to the rail to compensate for any parallelism defects. After this process, the squeegee is at a new, more accurate position 316.
[0048] Unfortunately, the actual printed ink thickness and its uniformity are not known at setup. If the test print result is not satisfactory, the operator repeats the complete setup process. The desired squeegee gap should enable adaptive printing, resulting in a highly uniform ink thickness across all heater trace patterns.
[0049] The squeegee gap and the parallelism of the print bed are approximated using shimming devices only, but the printed ink thickness and its uniformity are not known at that time. The print machine setup conditions can also drift after a queue time for measuring the printed ink for a test. Therefore, conventional printing methods cannot produce print uniformity at sub-micron levels.
[0050] Therefore, the repeatability of the squeegee installation on its holder can be important. A special carriage or alignment jig for squeegee installation can be used to ensure high installation repeatability from production batch to batch, as Figure 4 is shown.
[0051] Figure 4 is a cross-sectional side view of an adjustable print machine. The print machine 402 has a print bed 404 that carries a workpiece 406. A print screen 408 is secured to the print bed above the workpiece 406. As Figure 3 is shown, the print screen can or can not be mounted with shims 414 on one or both sides, depending on the implementation. Squeegee carriages 422, 424 are each mounted on each side of the squeegee 410 to hold the squeegee in place as it is brushed across the print screen 408 from its initial starting position (the position of the squeegee 410 in the figure) to its ending position (the position of the squeegee 416 in the figure).
[0052] The right side carriage 422 is aligned with Figure 3The same, but the left side carriage 424 has been adjusted with a shim 426 at the squeegee mount. This shim allows the system to compensate for any misalignment between the two carriages. As shown, there is a larger gap 430 on the left side than on the right. The carriage shim can compensate for this, while the screen shim 414 can compensate for any misalignment with the screen and the workpiece. Optionally, either or both types of shims can be removed when not necessary.
[0053] The encoder from the electric actuator can be used to precisely control the squeegee gap at the resolution of the encoder. The squeegee mounting carriage can be very useful for this approach. The squeegee tilt can be compensated for with the printer's "auto-zero" feature or an alternative.
[0054] Precision screen printing using electric actuators
[0055] Some embodiments use the approach of electric actuators (linear or rotary) with encoders, one on each end of the squeegee mount, which are typically integral with the printer system or can be added to the printer system. The motor encoder can be available from the printer controller. The electric actuators precisely control the squeegee gap and compensate for its misalignment at the resolution (e.g., number of steps) of the single motor encoder. In addition, some printers have a built-in auto-zero feature through some sensor feedback, such as Figure 5 as shown, so that the required encoder compensation can be easily obtained. With the use of a dedicated squeegee mounting carriage, the printer squeegee setup can achieve high repeatability between batches.
[0056] Figure 5 is a plot of the motor encoder transducer feedback on the vertical axis versus the exposure sequence on the horizontal axis. The data has a point for each of the exposures numbered 2 through 10 for this example only, and a straight line is drawn through the data points to show that this is an overall increasing or decreasing trend. In this example data, there is no significant trend in the exposure sequence.
[0057] The squeegee setup can be highly repeatable with the use of a carriage. In this example, the squeegee setup is much more repeatable than the traditional shim method. Misalignment due to any of a variety of reasons can be compensated for by the printer's auto-zero feature, which in this example is estimated to be close to 1 mm.
[0058] With the above described SIP printer setup method using motor encoders for control, the printed heater traces on green sheet ceramic have a high repeatability of the printed ink. After sintering, the metalization process of the printed ink, the resulting heater trace pattern is capable of producing ESC discs with high temperature uniformity repeatability between batches.
[0059] On the other hand, the precise printing of the desired ink thickness can be achieved with a knowledge-based model that is built using a lean DOE (design of experiments) plan to cover the area of interest. For a given screen mask design, a high resolution one-to-one relationship between the heater trace resistivity and the printed ink thickness can thus be derived in the DOE plan.
[0060] Figure 6 is a plot of the printed ink thickness on the vertical axis versus the motor encoder in steps on the horizontal axis. The amount of ink thickness 520 and the motor encoder value or step value 522 intersect at a point 524 on the plot. A correlation curve 526 can then be drawn through the points to determine the correlation based on the DOE. Thus, a high resolution one-to-one relationship between the printed ink thickness and the squeegee motor encoder step can be derived in the same DOE experiment. Once the knowledge-based model is built, the operator can set the printer squeegee by referring to the model and print.
[0061] In a similar manner, a DOE plan can be executed to obtain a highly repeatable correlation that covers the design specification window of the heater trace resistivity. For a target heater trace resistivity, the operator looks up the corresponding motor encoder in the DOE model and then prints the green sheet.
[0062] Figure 7 is a plot of the printed ink thickness on the vertical axis versus the target heater trace resistivity for each design. A particular ink thickness 530 intersects with a particular heater trace resistivity 532 at an intersection point 534 on the plot based on empirical values. A correlation curve 536 can then be constructed using multiple experimental data points. Such a curve, like the thickness / encoder correlation curve, can be used to set the parameters for a particular trace printing job.
[0063] The described method can be implemented in ESC ceramic disc manufacturing in many different ways. The metallization material referred to in this method can be any refractory metal formulated in a liquid form or ink. The green sheet ceramic referred to in this method can be an alumina-based or aluminum nitride-based ceramic regardless of its purity grade or additive formulation material. This method can be used for the printing of non-metallization materials that require precise and highly uniform thickness of the material. The motorized actuator can be linear or rotary. Any encoder information from the motorized actuator can be used to represent the control limits of the squeegee setting.
[0064] Conventional metallization materials printed on green sheet ceramic can have the following disadvantages:
[0065] 1) Inaccurate and non-repeatable print thickness: Current conventional printing press blade settings employ certain types of gap-filling devices. This type of setting process relies on operator experience, and the results are non-repeatable for ESC applications. The thickness of the printed metallized material can only be guessed during trial printing. The settings suffer from low repeatability between batches.
[0066] 2) Time-consuming: Experimental printing of metallized materials is necessary for gap-filling methods. Before thickness measurements can be performed, the printing ink requires a long queuing time to stabilize.
[0067] 3) Squeegee setting drift after test printing: Residual ink material on the screen mask will dry during the measurement queue time. Screen mask cleaning affects subsequent printing. Printing results often drift during production settings.
[0068] 4) Cumbersome production scheduling: If the test run results deviate from the manufacturing specifications, the operator needs to redo the scraper settings. This uncertainty of redoing affects manufacturing scheduling in complex manufacturing environments.
[0069] 5) Guaranteed quality: Manufacturing processes tend to accept below-average quality due to the cumbersome scraper setting process.
[0070] Implementation of SIP printing method
[0071] The SIP printing method for metallizing green ceramic sheets can be directly applied to electrostatic chuck (ESC) equipment. The printing method can be applied to ESCs with single or multiple main heaters in ESC disks used for semiconductor wafer processing applications. It can also be applied to ESCs with symmetrically or asymmetrically arranged micro-heaters in ESC disks. These micro-heaters are used for general thermal control. Furthermore, the printing method can be applied to metallizing green ceramic sheets for non-ESC applications where high printing uniformity on the green ceramic sheet is required.
[0072] Figure 8 This is an isometric diagram showing the adjustment points of the printing press and the squeegee guide rails. The printing press 404 has a right-side squeegee bracket 422 and a left-side squeegee bracket 424 mounted thereon. The squeegee travels across the printing screen from a starting position (where squeegee 410 is located in the diagram) to an ending position (where squeegee 416 is located in the diagram). A lack of flatness in the printing press and a lack of precise parallelism in the squeegee guide rails will affect the consistency of ink printing. The squeegee gaps (h1, h2, h3, h4) are indicated on the left and right sides of the squeegee at each guide rail and at both positions (where squeegees 410 and 416 are located in the diagram). If the printing press is not perfectly flat, the gaps will be inconsistent. Furthermore, if the squeegees are not installed consistently, the gaps will vary with each measurement.
[0073] A method can be applied to assist in accurate screen printing that enables sub-micron uniformity of metallization material on a green sheet ceramic. First, the screen mask and squeegee are set to align with the center of the heater trace. The squeegee gap and parallelism of the printer bed are then optimized using the auto-zero feature of the printer. The results of the auto-zero feature are recorded using a motor encoder. As an example at the center position, the motor encoder can be set to left = 25000, right = 25850.
[0074] Figure 9 is a top view of the same printer bed 404 and squeegee 416. In this example, the heater trace center 30 can be aligned with the center line 432 of the printing screen and the center 4345 of the squeegee, but only at one point. Any misalignment will result in a line that is offset from the center of alignment. Figure 9 Examples of 7 different positions (LI...L7) along the squeegee path from start to finish are also shown.
[0075] The auto-zero feature of the printer squeegee can be used at each of these multiple positions along the squeegee direction to map out the conformal parallelism. The mapped data can then be programmed as a set of squeegee control software parameters for conformal printing. The auto-zero at multiple positions (LI...L7) can be mapped to the unevenness of the printer bed.
[0076] Figure 10 is a graph showing the measurements that can be used with the motor encoder. Figure 10 is a graph of the motor encoder values or steps on the vertical axis versus each of eight different positions along the squeegee path. The auto-zero feature of the printer squeegee is used at each of the eight multiple positions along the squeegee direction to obtain the conformal parallelism on the heater trace pattern. There are different values on the right rail than on the left rail, and thus two sets of data points for each squeegee position (LI...L8).
[0077] Compensation method for printing thickness trends based on SIP printing method
[0078] Using the principles discussed herein, a method of compensating for printing thickness in SIP printing can be described. This method uses:
[0079] 1) a symmetrical ink printing method that first identifies hidden printer hardware biases for eliminating or minimizing the root cause of those later difficult to identify systematic printing biases, and
[0080] 2) motorized actuator encoders to precisely control the squeegee setup, i.e., the gap and parallelism of the printer bed.
[0081] 3) Using this SIP printing method, a knowledge-based model was then established through a lean DOE plan. This simple and repeatable method can be used by operators for easy setup of the squeegee without trial printing.
[0082] Printer bed flatness and non-parallelism of the squeegee rails along the squeegee direction are another major root cause of the inconsistent temperature profile in the ESC formed using printed green sheets or green bodies. As shown in FIG. 6, the inconsistent temperature profile can affect the chamber matching and CD (critical dimension) tuning. In this figure, seven positions along the squeegee direction are mapped on the printer bed, with five positions covering the heater traces of the screen mask pattern. The set of motor encoder values obtained is equivalent to a fingerprint of the conformal plane relationship during squeegeeing, and this relationship can be programmed into the printer control system for conformal printing. Figure 9
[0083] After the parameters or motor encoder steps are determined for a particular printer bed, screen, and after the ink setup, the ink thickness can change over time and usage. The ink thickness of the print can be uniform across the screen, but the thickness can have an upward or downward trend over large production batches.
[0084] Figure 11 is a plot of the printed ink thickness on the vertical axis versus the batch number on the horizontal axis. For each batch, there is an average ink thickness value, and a line is drawn through these values to show the increasing trend. In this example, the ink thickness increased by 1.5 microns over the printing of 60 ceramic green sheets from the first sheet to the last sheet. This large change is critical for some applications, and thus limits the size of the batch. The root causes of the trend can be complex and chaotic in nature. They can include: changes in ink viscosity during printing, changes in screen tension, screen mask cleaning, screen mask lifting, etc. Regardless of the complex trend mechanisms, the trend can be compensated to flatten the trend.
[0085] For many systems, a linear relationship can be observed between the squeegee gap or motor encoder position and the ink thickness. Figure 12 is a plot of the ink thickness on the vertical axis versus the squeegee gap on the horizontal axis. There are two data points at two different gaps that produce two different thicknesses. If there are additional points, a straight line can be drawn through all of them.
[0086] For any given screen mask design, a flattening method can be used that relies on the linear relationship between the squeegee gap / encoder and the ink thickness. In Figure 11 In the example data shown, the trend was estimated to be 1.5 μm of thickness variation over 60 printed green sheets. To improve print accuracy, this linear relationship can be measured by a simple DOE plan. As Figure 12 The DOE shown can set the blade gap at (x1, x2) μm or equivalent motor encoder steps and then generate the corresponding pair of ink thicknesses (y1, y2) μm. The trend sensitivity can be calculated as the ratio of (y2 - y1 ) = 4 μm of ink thickness variation to (x2 - x1 ) = 28 μm of blade gap variation.
[0087] The print machine blade gap is set in the range of (x1, x2) μm and the corresponding printed ink thickness is in the range of (y1, y2) μm. This relationship is important to calculate the compensation for the ink thickness trend used in printing. The larger the interval between x1 and x2 in the DOE plan, the more accurate and robust the model, but the interval should be narrow enough to stay in the linear region.
[0088] The motor encoder setting can be compensated by combining the two pieces of information above. The compensation can be derived as (28 / 4) x (1.5 / 60) = 0.175 μm / green sheet. The expression equivalent to the encoder is simple. For example, if the encoder unit is equal to 1 μm, then the compensation scheme will be to change the encoder unit of that one encoder for every 6 (i.e., 1 / 0.175) green sheets printed. The printed thickness variation after the compensation has been applied is then reduced to 0.15 μm, or 1 / 10 of the original trend range.
[0089] The encoder compensation can be integrated with the conformal mapping (from the print machine auto-zero feature). Figure 13 The calculation results of the motor encoder compensation are shown. Figure 13 is a plot of the printed ink thickness on the vertical axis versus the batch size of the printed green sheets on the horizontal axis. The upper linear increase line 450 shows the increase in ink thickness if there is no compensation of the ink thickness. The lower line 452 shows the ink thickness and its variation 454 when the trend is compensated by adjusting the motor encoding.
[0090] In this example, the blade gap is set in the range of (x1, x2) μm and the corresponding ink thickness range is (y1, y2) μm. The way to flatten this growth trend is based on the relationship between the ink thickness and the blade gap as described above. As a result, the reduction in screen printing performance is compensated by reducing the blade gap.
[0091] Once the ESC is manufactured with a repeatable temperature profile pattern, the comparison of wafer temperature maps allows information for improving either the heater trace design or the cooling base design or both. The highly repeatable and uniform printing of the metallization material enables the effectiveness and efficiency of the ESC design for advanced electrostatic chucking equipment.
[0092] Implementation of SIP printing method
[0093] The conformal relationship data between the squeegee and the printer bed can be programmed into the printer control software that coordinates the motor encoder compensation for each squeegee end during squeegee. If the temperature deviation is consistent and persistent between ESCs, further optimization can be attempted in motor encoder steps. A DOE can be planned to obtain core data for determining compensation parameters. Figure 11 and 12 Some data that can be used to obtain results as shown in Figure 13 are shown.
[0094] The setup of both the screen mask and the squeegee are aligned with the center of the heater trace. The squeegee gap and its parallelism to the printer bed can be optimized by the auto-zero feature of the printer and recorded with the motor encoder.
[0095] Both the printer bed flatness and the squeegee rail misalignment affect the consistency of the ink printing. Figure 8 The squeegee gap (hi, h2, h3, h4) is not the same for all squeegees because the printer bed is not perfectly flat. If the squeegee is not installed consistently, its measurement is not repeatable.
[0096] The auto-zero feature of the printer squeegee is used at multiple locations along the squeegee direction to map the conformal parallelism. The mapping data can be programmed into the squeegee control software conformal printing.
[0097] Screen mask design method for precision applications based on symmetrical ink printing method
[0098] A highly repeatable squeegee setup procedure for symmetrical ink printing (SIP) method is described, where the metallized material on the green sheet ceramic has sub-micron uniformity. How to use the motor encoder of the linear or rotary squeegee motorized actuator to map the non-flat printer bed for conformal parallelism at the accuracy of the motor encoder is also described. How to flatten the uniform ink printing but with a linear trend of the printing thickness is also discussed. The implementation of all these methods using the printer control software is also discussed.
[0099] The application of the SIP printing method for optimal screen mask design for precision ink printing is described herein. The printing with precision thickness requires some design iterations. The iterations can be effective if the printing deviation is minimized using the SIP printing method. It is useful to shorten the product development cycle that involves a series of lengthy and complex ceramic sintering processes and post-sintering processes.
[0100] Screen mask design involves screen thread selection and screen "height" parameter X. The value of X is related to the ink thickness achieved. Thread selection is straightforward empirically, but design parameter X requires multiple iterations for key applications. A lean process to optimize design parameter X for a given thread diameter is described.
[0101] The new screen mask design starts with an empirical equation developed by the mask designer. The empirical model has only first order accuracy. The screen mask design activity can be divided into two processes: the first process focuses on building a robust model between squeegee gap and printed ink thickness, while the second process focuses on building a correlation between printed ink thickness and heater resistivity. This approach aims to minimize the need for frequent sintering. All trial printed green sheet ceramics are sintered in the same batch.
[0102] Process 1:
[0103] The purpose of this process is to build a quality correlation with minimal error between squeegee gap and printed ink thickness using the SIP printing method. This is achieved by a wide spacing between x-, xo, and x+. The spacing is as wide as possible but still remains in the linear region.
[0104] Figure 14 is a plot of printed ink thickness or printed ink resistivity on the vertical axis versus mask thickness design parameter identified as X on the horizontal axis. Since ink thickness and resistivity have a linear relationship for any ink type and processing method, they can be plotted on the same vertical axis.
[0105] To use Figure 14 plot, the designer sets X = xo as shown on the horizontal axis and inputs it into the design equation which is only first order accurate. t0 is the output ink thickness using the design equation. After the sintering process is completed, the design can be verified.
[0106] To speed up the mask design, and to obtain a quality model, two additional screen masks with input design parameters x- and x+ respectively are made, also shown on the horizontal axis, so that predicted t- and t+ are 20-25% lower and 20-25% higher than t0 respectively. It is just a DOE plan for a correlation with better signal to noise ratio.
[0107] Following the SIP printing method, which aims to print a highly uniform and repeatable ink thickness on green sheet ceramics, several trial green sheet ceramics with masks A, B, and C are printed. Note that the manufacturing lead time for the three masks is much shorter than the sintering and other processes. With these output data, a quality correlation model between mask design parameter X and printed ink T or resistivity can be built, and the results are plotted in Figure 14In.
[0108] In Figure 14 In this example, there is a gap on the horizontal axis between X0and X0= 50, which is the desired value of the screen mask design parameter X that will allow the printer to print T0= 50 and result in the target heater trace resistivity of 50 / ESC design. In this example, an encoder DOE can be implemented to close this gap using squeegee gap adjustment or another parameter.
[0109] In Figure 15 Three mask designs, referred to as A, B, C, represent three levels of the parameter X at levels X+, X0, and X- respectively. The associated printed ink thicknesses are represented as t+, t0, t- respectively, and the resistivities after sintering are represented as 0+, 0, 0- respectively. In this example, mask B represents the intended design, and masks A and C are used to develop a robust correlation between the mask design parameter and the heater trace resistivity at the printed ink thickness.
[0110] Figure 15 is a table showing an example of the type of data that can be collected when designing a screen for a particular printer configuration. As indicated in the squeegee gap column, a multi-level encoder DOE is planned based on squeegee gap g0. In this example, the encoder DOE for mask design B is used to develop a compensation method for mask design B.
[0111] Five levels of squeegee gap (g0... g4) are set to determine the correlation between the squeegee encoder and the printed ink thickness. At least 2 test prints can be made at each encoder setting to estimate the error term. At least 3 test prints of masks A, B, and C can be made.
[0112] Squeegee setting g0is the previous setting that the mask engineer will use or will need to determine using a DOE. Squeegee gap g0may be based on the mask design engineer's past experience, or the printer manufacturer can recommend a value based on experience. In this example, a five-level encoder DOE as shown in the encoder DOE column can be used if the sinter batch is large enough to accommodate all of the test green sheets in the same sintering process. Otherwise, some levels (such as levels g1and g3) can be skipped. The information for g2and g4may be sufficient to set the construction. In this example, the encoder levels or values are set as follows (however, any other levels can be used):
[0113] g1encoder is set to be 10% more than g0
[0114] g2encoder is set to be 20% more than g0
[0115] g3encoder is set to be 10% less than g0
[0116] The g4 encoder is set to be 20% less than gO
[0117] In this example, as shown in the minimum test print column, there are three test prints for each mask design and two test prints for mask B at each encoding level. Thus, in process A of the five level encoder DOE, a total of 17 green sheets are printed, or 13 green sheets are printed for the 3 level encoder DOE, with limited sintering capacity. Mask A and mask B can repeat the same operation.
[0118] Process 2:
[0119] Review the printing and sintering results.
[0120] If the results produced with mask design B of the encoder DOE meet the ESC design specification, the screen mask design is completed by one sintering process.
[0121] If the results produced with mask design B in process A do not meet the design requirements, or further optimization of the mask design is desired, the quality correlation of the 3 mask designs can be used to indicate the desired mask design parameter resistivity to provide, for example, XΩ = 50, and the output resistance will be Ω = 50 with reasonable error. In the second design iteration, since the encoder step compensation power has been estimated in process A, the encoder DOE can be planned or ignored.
[0122] One of the challenges of the traditional approach to screen mask design for critical applications is the numerous design iterations to achieve design goals. The ink printing quality and consistency are insufficient due to the lack of printing technology optimization. The SIP precision and uniform printing capability breaks the tail chasing pattern and allows the development of an effective design process with sufficient knowledge in a cost-effective manner.
[0123] The above-described techniques can be represented as a series of operations. Four different and related processes have been described and indicated below as methods A, B, C, and D.
[0124] Method A: Figure 16 is a process flow diagram of a method that uses the precision of a motorized encoder for a highly repeatable squeegee setup.
[0125] 1) Starting at 502, a screen mask is manufactured based on a screen mask design for a target printed ink thickness for an ESC application. The alignment gap between the center of the heater trace pattern and the center of the screen mask is measured. The gap is compared to certain tolerances. If the screen mask gap meets the tolerances, go to the next operation.
[0126] 2) At 504, a set of two squeegee mounting jigs are designed to mount the squeegee tightly to the printer squeegee holder. The squeegee mounting jigs are designed such that the center of the mounted squeegee is aligned to the center of the screen mask with certain tolerances. If hardware change is needed, modify the squeegee holder for using the squeegee jigs.
[0127] 3) Check the auto-zero capability of the printer control system. Auto-zero is a feature of the printer control system to equalize some transducer signals (such as pressure or other signals) when both squeegee ends (left end or right end) reach conformal contact of the printer bed. At 506, during the squeegee wiping of the paste on the screen, when the auto-zero function is performed, the printer control system can record the motor encoder values in its memory and display the encoder steps on the control interface at 508. If the auto-zero feature is not available, equip the printer with this capability.
[0128] 4) At 510, verify the repeatability of the squeegee mounted at the center of the heater trace pattern location by the auto-zero feature of the printer. Analyze the encoder readings from the control interface for repeatability. Perform statistical analysis on the encoder data, including mean and variation. These statistical properties of the encoder depend on the resolution or sensitivity of the transducer. Choose the appropriate transducer according to the application requirement. The encoder values can be compared to the resulting roundel. At 512, the repeated printed green sheet can form a carrier roundel, and at 514, the printed electronic traces can be tested for resistivity.
[0129] 5) The difference between the encoder values averaged between the squeegee left end and right end represents the non-parallelism, or tilt bias, of the squeegee to the printer bed. This difference is a characteristic of the squeegee and associated mounting jigs used. At 516, the encoder information and shims can be used to configure the screen printer for the next batch. The encoder information using the auto-zero feature represents the printer hardware capability limit for the highest possible repeatable squeegee setup.
[0130] The conventional squeegee setup method uses shimming devices to set the squeegee gap in a time-consuming manner. This method operates using subjective judgment without knowing the squeegee gap, parallelism level, etc. The repeatability is poor, and it is difficult to apply the adjustment for sub-micron uniformity applications. In addition to the poor repeatability, a green sheet ceramic is used for trial printing for each production batch. This conventional method is inherently inefficient and ineffective.
[0131] The SIP printing method using repeatable squeegee setup and encoder information in the printer control system provides the architecture for conformal and accurate printing of sub-micron uniformity for critical applications. Once the printing recipe is established, any trained operator can repeat the same encoder setup for new production jobs. No shimming method or trial printing is needed.
[0132] Method B: Figure 17 is a process flow diagram for a method of conformal precision screen printing with sub-micron uniformity on a screen mask for printing using Symmetric Ink Printing (SIP).
[0133] Printing presses for screen printing are not manufactured perfectly. Any printing press has its unique flatness variation measured in Ra, and any one printing press is different from other printing presses. Likewise, the action of any motorized squeegee is unique, and each press is different. The combination of these variations can result in printing deviations that are difficult to identify for screen printing. The screen printing industry currently has no known method or technique to address these variations.
[0134] Conformal screen printing technology is defined as a printing method that applies uniform squeegee pressure along the full squeegee length of the printing press during printing. A conformal screen printing technology for large green sheet ceramic is described below. The SIP method based on Method A above further develops this conformal printing.
[0135] 6) At 520, map the printing press at multiple locations along the squeegee direction. Use equal intervals. The encoder data obtained is given for the printing press system and the characteristics of the squeegee used including the squeegee mounting carriage. The number of mapping locations is chosen to allow proper spline curve fitting.
[0136] 7) At 522, perform statistical spline curve fitting with the mapping data described above, one spline curve for each squeegee end, so two splines are obtained. These spline curves are expressed in encoder and represent the smooth squeegee gap compensation by the motorized actuator that controls the squeegee gap.
[0137] 8) Store the spline data in the control system memory. At 524, the printing press control system uses these profiles in its algorithm to actively drive the squeegee gap in the printing direction with smooth conformal squeegee pressure.
[0138] 9) At 526, review the ESC temperature performance and determine if another design iteration of the heater trace pattern is needed. At this stage of this process, the resulting ESC will exhibit repeatable temperature mapping with little manufacturing variation. There can be temperature non-uniformity, but at this stage, further improvement is available through design compensation for design cycle efficiency. At 516, these settings can be used to configure the printing press for the next set of green sheets.
[0139] Method C: Figure 18 is a process flow diagram for a method of ink thickness trend compensation for large printing batches for Symmetric Ink Printing (SIP).
[0140] Printing with sub-micron uniformity can be achieved in an organized manner by the SIP method. However, printing trends often occur when printing large batches. Printing trends are a combination of many complex and interactive mechanisms between ink viscosity, screen tension, screen mesh cleanliness, etc. This can be addressed as follows.
[0141] 10) At 530, print a large batch (e.g., 60 green sheets) by the SIP method described above.
[0142] 11) At 532, measure the printed ink thickness trend rate. This trend rate can be measured as the ratio of thickness variation to the total number of green sheets. For example, a printed thickness variation of 1.5 μm is found to be associated with a printed batch of 60 ceramic green sheets.
[0143] 12) At 534, plan a two-level DOE with wider spaced blade gaps according to the encoder. This spacing is large enough to unambiguously determine the printed thickness ratio of the two levels. As a guideline, the thickness difference of this DOE is about ±20-30% of the target thickness. This DOE establishes a transition to a useful compensation of the trend by performing printing of multiple green sheets with two or more additional blade gaps.
[0144] 13) At 536, determine the trend rate per sheet in combination of the results of both 11 and 12. The ink thickness trend rate is an indication of the ink thickness at different blade gaps. This relationship can be converted from μm / sheet to equivalent encoder steps or values / sheet.
[0145] 14) The method of trend flattening is based on the equivalent encoder trend rate per successive sheet. At 538, when printing a sheet, adjust the encoder value to compensate for the increase in ink thickness when printing a subsequent sheet. If this trend rate is greater than 1.0 encoder step, the encoder adjustment for compensation is at the resolution of a single encoder step. For example, if this ratio is 2.5 encoder steps per sheet, a first adjustment of 2 encoder steps for one sheet and a next adjustment of 3 encoder steps for the next sheet printing can be done. If the trend rate is a fraction of an encoder step, the inverse of the fraction is the desired adjustment. For example, if the equivalent trend rate is determined to be 0.25 encoder steps / sheet, an adjustment of one encoder step can be made every four (1 / 0.25) successive prints. The trend compensation method can be implemented with software. Manual adjustments can also be made. A software algorithm can be developed to integrate all the features herein.
[0146] Method D: Figure 19 is a process flow diagram of a method of screen mask design suitable for precision applications using SIP printing.
[0147] Screen mask design typically involves two design parameters, one parameter is the line diameter, and the other parameter is a thickness related design parameter, referred to here as X.
[0148] 1) At 540, multiple screen printing masks are produced with different printing thickness parameters. They can then be used to perform tests to optimize the X value and squeegee gap. In one example, three screen masks are made with three different X values. These masks are designated as Mask A, Mask B, and Mask C. Note that the same screen mesh line and the same heater trace pattern are used in these masks, only the X parameter has different values. These values are estimated using a formula derived from theory or experience. The formula provides a first order accuracy. In a specific example, Mask B has X at X0, i.e., X = X0. Mask A has X = X-, and Mask C has X = X+. Take for example, X- / X0 = ~80%, X0 / X = ~80%. Make all masks available in the same printing job described in step 3. Mask B is the candidate mask for the target design.
[0149] 2) At 542, a DOE for squeegee settings is planned. The squeegee settings can be decided based on experience or used as a starting estimate of the parameter g0. A typical range can be from 50 pm to 200 pm for precision screen printing. A five level DOE is created for determining the encoder compensation. This DOE is referred to as the encoder DOE. Figure 15 An example is shown.
[0150] 3) For example, at 544, a total of 17 green sheets are printed using the SIP method described above. In the example described, there are 3 trial prints using Mask A, 3 trial prints using Mask C, and 3 trial prints using B, and all of these trial prints have the same squeegee settings at g0. The encoder DOE uses only Mask B, and 2 trial prints are printed in this DOE with squeegee levels at g1 and g3 as shown. Figure 15 Eight trial prints are printed in the same printing job. Depending on the printer and its parameters, any other number of repeating levels or numbers of prints can be made.
[0151] 4) At 546, the printed ink thickness is measured in a reliable method, either a contact method or a non-contact method.
[0152] 5) At 548, the 17 green sheets are converted to 17 green ESC discs, one disc per printed sheet.
[0153] 6) All 17 discs are sintered in the same sintering job. If the sintering furnace capacity is less than 17 green discs, g1 and g3 in the encoder DOE can be skipped. The same furnace and the same job are used to eliminate sintering variations that can affect the measurements.
[0154] 7) At 550, the resistivity of each disc after sintering is measured and any other manufacturing processes required to complete each disc, such as brazing, are completed.
[0155] 8) At 552, plots are constructed as shown in Figure 14 All correlations between design parameters X, ink thickness T, and resistivity Ω are plotted. These plots are referred to as the knowledge base for screen mask design using the SIP method. The results can be used to determine the parameters for producing a printing screen at 554, and then the appropriate production screen is used at step 516 with the determined printer parameters and configuration.
[0156] 9) If the design error is too large, design iterations can be used. The selection of the X value for the next iteration can be selected using the model or another method. If the design error is small, the built-in encoder DOE can compensate for the design error by the encoder. In this way, screen mask design can be completed in one sintering run or at most two sintering runs. The development time and cost of screen masks for precision applications is greatly reduced.
[0157] Figure 20 is an isometric view of an assembled electrostatic chuck. The support shaft 212 supports the base plate 210 through the isolator 216. The middle isolation plate 208 and the upper cooling plate 206 are carried by the base plate. The top cooling plate 206 carries the dielectric disc 215 on the top surface of the upper cooling plate. The disc has an upper circular platform 205 to support the workpiece 204 and a lower concentric circular base 207 attached to the cooling plate 206. The upper platform has internal electrodes to electrostatically attach the workpiece. The workpiece can alternatively be clamped, vacuumed, or attached in another way. There is a bond joint 218 between the disc 215 and the top cooling plate 206 to hold the ceramic of the top plate to the metal of the cooling plate. As described herein, the heater, the electrodes, or both can be formed in the disc using a printing process on a green sheet. The middle plate can perform cooling, gas flow, and other functions, depending on the particular implementation.
[0158] The ESC can control the temperature of the workpiece using the resistive heater in the disc, the coolant fluid in the cooling plate, or both. Power, coolant, gas, etc. are supplied to the cooling plate 206 and the disc 215 through the support shaft. The ESC can also be manipulated and held in position using the support shaft.
[0159] Figure 21is a partial cross-sectional view of a plasma system 100 having a pedestal 128 according to embodiments described herein. Although a pedestal is shown here, the principles described herein can be used on any of a variety of different workpiece carriers, including different types of chucks, carriers, and pedestals. Although a chamber pedestal is shown, the principles described can also be applied to workpiece carriers used outside of the processing chamber. The pedestal 128 has an active cooling system that allows for active control of the temperature of a substrate located on the pedestal over a wide temperature range as the substrate is subjected to a number of processes and chamber conditions. The plasma system 100 includes a processing chamber body 102 having a sidewall 112 and a bottom wall 116 that define a processing region 120.
[0160] The pedestal, carrier, chuck, or ESC 128 is disposed in the processing region 120 through a passage 122 formed in the bottom wall 116 in the system 100. The pedestal 128 is adapted to support a substrate (not shown) on its upper surface. The substrate can be any of a variety of different workpieces for processing applied by the chamber 100 made of any of a variety of different materials. The pedestal 128 can optionally include a heating assembly (not shown) (e.g., a resistive assembly) to heat and control the substrate temperature at a desired process temperature. Alternatively, the pedestal 128 can be heated by a remote heating assembly such as a lamp assembly.
[0161] The pedestal 128 is coupled to a power socket or power box 103 through a shaft 126, which can include a drive system to control the elevation and movement of the pedestal 128 within the processing region 120. The shaft 126 also contains a power interface to provide power to the pedestal 128. The power box 103 also includes an interface for power and temperature indicators (such as a thermocouple interface). The shaft 126 also includes a base assembly 129 adapted to be detachably coupled to the power box 103. A circumferential ring 135 is shown above the power box 103. In one embodiment, the circumferential ring 135 is a shoulder adapted as a mechanical stop or landing configured to provide a mechanical interface between the base assembly 129 and an upper surface of the power box 103.
[0162] A lever 130 is disposed and used to actuate a substrate lift pin 161 disposed through the pedestal 128 through a passage 124 formed in the bottom wall 116. The substrate lift pin 161 lifts the workpiece off the pedestal top surface to allow the workpiece to be removed in and out of the chamber, typically using a robot (not shown) through a substrate transfer port 160.
[0163] A chamber lid 104 is coupled to the top of the chamber body 102. The lid 104 houses one or more gas distribution systems 108 coupled thereto. The gas distribution system 108 includes a gas inlet passage 140 that delivers reactant and cleaning gases into the processing region 120 through a showerhead assembly 142. The showerhead assembly 142 includes an annular bottom plate 148 having a blocker plate 144 disposed intermediate the faceplate 146.
[0164] A radio frequency (RF) source 165 is coupled to the showerhead assembly 142. The RF source 165 powers the showerhead assembly 142 to facilitate the creation of a plasma between the faceplate 146 of the showerhead assembly 142 and the heated susceptor 128. In one embodiment, the RF source 165 can be a high frequency radio frequency (HFRF) power source such as a 13.56 MHz RF generator. In another embodiment, the RF source 165 can include a HFRF power source and a low frequency radio frequency (LFRF) power source such as a 300 kHz RF generator. Alternatively, the RF source can be coupled to other portions of the processing chamber body 102 such as the susceptor 128 to facilitate plasma generation. A dielectric isolator 158 is disposed between the lid 104 and the showerhead assembly 142 to prevent the conduction of RF power to the lid 104. A shadow ring 106 can be disposed on the periphery of the susceptor 128 to engage a substrate at a desired height of the susceptor 128.
[0165] Optionally, cooling passages 147 are formed in the annular bottom plate 148 of the gas distribution system 108 to cool the annular bottom plate 148 during operation. A heat transfer fluid such as water, ethylene glycol, gas, etc. can be circulated through the cooling passages 147 such that the bottom plate 148 is maintained at a predetermined temperature.
[0166] A chamber liner assembly 127 is disposed within the processing region 120 in close proximity to the sidewall 112 of the chamber body 102 to prevent the sidewall 112 from being exposed to the processing environment within the processing region 120. The liner assembly 127 includes a circumferential pumping cavity 125 coupled to a pumping system 164 configured to exhaust gases and byproducts from the processing region 120 and control the pressure within the processing region 120. A plurality of exhaust ports 131 can be formed on the chamber liner assembly 127. The exhaust ports 131 are configured to allow the flow of gases from the processing region 120 to the circumferential pumping cavity 125 in a manner that facilitates processing within the system 100.
[0167] The system controller 170 is coupled to various subsystems to control the manufacturing process in the chamber. The controller 170 can include a temperature controller 175 for executing temperature control algorithms (e.g., temperature feedback control) and can be software or hardware, or a combination of both software and hardware. The system controller 170 also includes a central processing unit 172, memory 173, and input / output interface 174. The temperature controller receives temperature readings 143 from sensors (not shown) on the susceptor. The temperature sensors can be near the coolant channels, near the wafer, or placed in the dielectric material of the susceptor. The temperature controller 175 uses the sensed temperature or temperatures to output control signals that affect the rate of heat transfer between the susceptor assembly and heat sources and / or heat sinks (such as heat exchanger 177) outside the plasma chamber 105.
[0168] The system can also include a controlled heat transfer fluid loop 141 with flow controlled based on a temperature feedback loop. In an example embodiment, the temperature controller 175 is coupled to a heat exchanger (HTX) / chiller 177. The heat transfer fluid flows through valves (not shown) at a rate controlled by valves through the heat transfer fluid loop 141. The valves can be incorporated into the heat exchanger or a pump inside or outside the heat exchanger to control the flow rate of the heat fluid. The heat transfer fluid flows through conduits in the susceptor assembly and then back to the HTX 177. The temperature of the heat transfer fluid is increased or decreased by the HTX and then the fluid is returned through the loop to the susceptor assembly.
[0169] The HTX includes a heater 186 to heat the heat transfer fluid and thereby heat the substrate. The heater can use resistive coils around pipes inside the heat exchanger or the heat exchanger is formed with the heated fluid conducting heat through the exchanger to the conduits containing the hot fluid. The HTX also includes a chiller 188 to extract heat from the heat fluid. This can use a radiator to put heat into ambient air or a coolant fluid or in any of a variety of other ways. The heater and chiller can be combined so that the temperature controlled fluid is first heated or cooled and then exchanges heat from the control fluid with the heat fluid in the heat transfer fluid loop.
[0170] Valves (or other flow control devices) between the HTX 177 and the fluid conduits in the susceptor assembly can be controlled by the temperature controller 175 to control the flow rate of the heat transfer fluid to the fluid loop. The temperature controller 175, temperature sensors, and valves can be combined to simplify the structure and operation. In an embodiment, the heat exchanger senses the temperature of the heat transfer fluid after it returns from the fluid conduits and heats or cools the heat transfer fluid based on the temperature of the fluid and the desired temperature of the chamber 102 operating state.
[0171] An electrical heater (not shown) can also be used in the ESC to apply heat to the workpiece assembly. The electrical heater, typically in the form of a resistive assembly, is coupled to a power supply 179 controlled by the temperature control system 175 to energize the heater assembly to achieve a desired temperature.
[0172] The heat transfer fluid can be a liquid such as, but not limited to, deionized water / glycol, Fluorinert® from 3M (such as Fluorinert® FC-72®), or Galden® from Solvay Solexis, Inc., or any other suitable dielectric fluid such as those containing perfluorinated inert polyethers. Although the present specification describes a susceptor for a PECVD processing chamber in context, the susceptor described herein can be used in a variety of different chambers and for a variety of different processes. The heat transfer fluid can be a liquid such as, but not limited to, deionized water / glycol, Fluorinert® from 3M (such as Fluorinert® FC-72®), or Galden® from Solvay Solexis, Inc., or any other suitable dielectric fluid such as those containing perfluorinated inert polyethers. Although the present specification describes a susceptor for a PECVD processing chamber in context, the susceptor described herein can be used in a variety of different chambers and for a variety of different processes.
[0173] A backside gas source 178, such as a pressurized gas supply or a pump, and a gas reservoir are coupled to the susceptor assembly through a mass flow meter 185 or other type of valve. The backside gas can be helium, argon, or any gas that provides thermal convection between the wafer and the puck without affecting the process of the chamber. The gas source pumps gas through gas outlets of the susceptor assembly, described in more detail below, to the backside of the wafer under control of a system controller 170 connected to the system.
[0174] The processing system 100 can also include other systems not specifically shown in Figure 21 The illustrated chamber is provided as an example, and depending on the nature of the workpiece and the process desired, any of a variety of other chambers can be used with the present application. The susceptor and heat fluid control system described can be adapted for use with different physical chambers and processes.
[0175] As used in the embodiments of the application and the appended claims, the singular "a," "an," and "the" include plural references unless the context clearly dictates otherwise. It will also be understood that the term "and / or" as used herein refers to and encompasses any and all possible combinations of one or more of the associated listed items.
[0176] The terms "coupled" and "connected," along with their derivatives, can be used herein to describe functional or structural relationships between components. It should be understood that these terms are not intended as synonyms for each other. Rather, in particular embodiments, "connected" can be used to indicate that two or more elements are in direct physical, optical, or electrical contact with each other. "Coupled" can be used to indicate that two or more elements are in either direct or indirect (with other intervening elements between them, e.g., in an effect relationship) physical, optical, or electrical contact with each other, and / or that the two or more elements co-operate or interact with each other (as in a cause-and-effect relationship).
[0177] The terms "over," "under," "between," and "on" as used herein refer to a relative position of one component or material layer with respect to other parts or layers noted in the physical relationship. For example, in the context of material layers, a layer disposed over or under another layer can be directly in contact with the other layer or can have one or more intervening layers. Moreover, a layer disposed between two layers can be directly in contact with the two layers, or can have one or more intervening layers. In the context of a component assembly, similar distinctions will be made. The terms "first," "second," "third," "fourth," and the like in the context of an element or article of manufacture do not necessarily indicate that the element or article of manufacture must be in a particular order, but rather, these terms can be used solely to distinguish between two or more elements or articles of manufacture. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the present disclosure. It is further noted that the use of particular terms to describe embodiments is by way of illustration only and not by way of limitation. The scope of the present disclosure is defined by the appended claims and equivalents thereto, and not by the terminology used herein.
[0178] It is to be understood that the above-referenced elements of the present application are merely preferred embodiments of the present application, and that numerous changes in the elements can be made without departing from the spirit and scope of the application. Although specific terms are employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation.
Claims
1. A top plate of a workpiece carrier, comprising: a plurality of ceramic green sheets; and a printed ceramic green sheet having a conductive pattern printed thereon, the printed sheet material being embedded in the plurality of ceramic green sheets, wherein the plurality of ceramic green sheets are sintered and hardened, wherein a metallization material is printed on the ceramic green sheet by: experimenting printing a plurality of ceramic green sheets with different screen masks having different squeegee settings of a screen printer; converting the experimentally printed ceramic green sheets into a top plate; measuring the resistivity of the top plate; correlating the resistivity with the different screen masks; generating a screen mask for use with the screen printer based on the correlation between the resistivity and the different screen masks; and screen printing a metallization material on the ceramic green sheet using the screen mask.
2. The top plate of claim 1, wherein the conductive pattern forms an electrically resistive heater.
3. The top plate of claim 1, wherein the printed sheet material further comprises an electronic component attached to the sheet material and coupled to the pattern.
4. The top plate of claim 1, further comprising a frame for carrying the hardened plurality of green sheets.
5. The top plate of claim 1, wherein the ceramic green sheets are formed from ceramic powder and glass compacted with a binder.
6. The top plate of claim 1, wherein the printed conductive pattern comprises a conductive metal in the form of a paste, the conductive metal including a suspension and a dispersant dispersed in the pattern.
7. The top plate of claim 1, wherein the conductive pattern includes a plurality of heater traces, each heater trace having a resistivity of at least 50 Ω.
8. The top plate of claim 1, wherein the printed ceramic green sheet includes AI2O3.
9. The top plate of claim 1, wherein the printed ceramic green sheet includes AIN.
10. The top plate of claim 1, wherein the conductive pattern includes a conductive material selected from the group consisting of tungsten, molybdenum, zinc, silver, and gold.
11. The top plate of claim 1, further comprising: an upper circular platform above a lower concentric circular base, the lower concentric circular base having a diameter greater than a diameter of the upper circular platform.
12. The top plate of claim 11, wherein the upper circular platform has an internal electrode to electrostatically attach a workpiece to the top plate.
13. The top plate of claim 1, further comprising: an internal electrode to electrostatically attach a workpiece to the top plate.
14. An electrostatic chuck, comprising: a top plate, the top plate comprising: a plurality of ceramic green sheets; and a printed ceramic green sheet having a conductive pattern printed thereon, the printed sheet material being embedded in the plurality of ceramic green sheets, wherein the plurality of ceramic green sheets are sintered and hardened; a cooling plate below the top plate; and an adhesive for attaching the top plate to the cooling plate, wherein a metallization material is printed on the ceramic green sheet by: experimenting printing a plurality of ceramic green sheets with different screen masks having different squeegee settings of a screen printer; converting the experimentally printed ceramic green sheets into a top plate; measuring the resistivity of the top plate; correlating the resistivity with the different screen masks; generating a screen mask for use with the screen printer based on the correlation between the resistivity and the different screen masks; and screen printing a metallization material on the ceramic green sheet using the screen mask. experimentally printing a plurality of ceramic green sheets with different screen masks having different squeegee settings of a screen printer; converting the experimentally printed ceramic green sheets into a top plate; measuring the resistivity of the top plate; correlating the resistivity with the different screen masks; generating a screen mask for use with the screen printer based on the correlation between the resistivity and the different screen masks; and screen printing a metallization material on the ceramic green sheets using the screen mask.
15. The electrostatic chuck of claim 14, further comprising: a bottom plate under the cooling plate.
16. The electrostatic chuck of claim 15, further comprising: an insulating plate between the cooling plate and the bottom plate.
17. The electrostatic chuck of claim 14, wherein the top plate further comprises: an upper circular platform above a lower concentric circular base, the lower concentric circular base having a diameter greater than a diameter of the upper circular platform.
18. The electrostatic chuck of claim 14, further comprising: an internal electrode to electrostatically attach a workpiece to the top plate.
19. A plasma system, comprising: a process chamber body having sidewalls and a bottom wall defining a process region; an electrostatic chuck in the process region, the electrostatic chuck comprising: a top plate including: a plurality of ceramic green sheets; and a printed ceramic green sheet having a conductive pattern printed thereon, the printed sheet embedded in the plurality of ceramic green sheets, wherein the plurality of ceramic green sheets are sintered and hardened; and a cooling plate under the top plate; and an adhesive to attach the top plate to the cooling plate; a lid coupled to a top of the chamber body; and a showerhead between the lid and the process region, wherein a metallization material is printed on the ceramic green sheets by: experimentally printing a plurality of ceramic green sheets with different screen masks having different squeegee settings of a screen printer; converting the experimentally printed ceramic green sheets into a top plate; measuring the resistivity of the top plate; correlating the resistivity with the different screen masks; generating a screen mask for use with the screen printer based on the correlation between the resistivity and the different screen masks; and screen printing a metallization material on the ceramic green sheets using the screen mask.
Citation Information
Patent Citations
Ceramic substrate for semiconductor fabricating device
US20020158328A1
Electrostatic chuck heater
US20080047952A1
Plasma reactor electrostatic chuck with cooled process ring and heated workpiece support surface
US20130277339A1
Electrostatic chuck and method for manufacturing electrostatic chuck
US20130321974A1
Semiconductor manufacturing equipment component and method of making the same
US20160099163A1