Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 1 of 13
`
`Exhibit 28
`
`
`
`
`
`
`
`
`
`
`
`

`

`United States Patent (19)
`Zhou et al.
`
`54 CYCLE PURGING AWACUUM CHAMBER
`DURING BAKEOUT PROCESS
`
`75 Inventors: Jiaxiang Zhou; Stephen D. Dasso,
`both of Austin, TeX.
`
`73 ASSignee: Applied Materials, Inc., Santa Clara,
`
`a.
`
`Appl. No.: 921,811
`21
`22 Filed:
`Sep. 2, 1997
`51
`Int. Cl. ................................ B08B 5/02; B08B 5/04;
`B08B 9/00
`52) U.S. Cl. ............................. 134/19; 134/21; 134/22.1;
`134/22.15; 134/22.18; 134/30; 134/37
`58 Field of Search ............................... 134/19, 21, 22.1,
`134/22.18, 22.15, 30, 37
`References Cited
`
`56)
`
`U.S. PATENT DOCUMENTS
`4,873,833 10/1989 Pfeiffer et al. ........................... 62/55.5
`
`USOO5879467A
`Patent Number:
`11
`(45) Date of Patent:
`
`5,879,467
`Mar. 9, 1999
`
`... 445/40
`5,433,639 7/1995 Zahuta et al. ..
`5,536,330 7/1996 Chen et al. ............................... 134/21
`5,678,759 10/1997 Grenci et al. ........................... 237/1 R
`
`Primary Examiner Zeinab El-Arini
`Attorney, Agent, or Firm-Patterson & Associates
`57
`ABSTRACT
`
`A vacuum system bakeout process is performed by cycling
`the SVStem between two preSSures, pumping the SVStem
`y
`p
`, pumping
`y
`down to a lower preSSure, and holding the System at that
`lower preSSure for a period of time. Agas, Such as argon gas,
`is introduced into the System. This gaS introduction is done
`while cycling between the two preSSures. The pump is used
`to lower the preSSure during cycling, and the gas flow is used
`to raise the pressure. A rough pump performs the cycling
`between pressures, and then a high Vacuum pump evacuates
`the System to the lower pressure.
`
`35 Claims, 6 Drawing Sheets
`
`
`
`
`
`
`
`
`
`24
`ARGON
`SUPPLY
`
`26
`
`FLOW
`CONTROLLER
`
`WACUUM
`
`CHAMB ER
`
`LAMP
`
`32
`
`VACUUM
`PUMP
`
`CRYOPUMP
`ROUGH PUMP
`
`
`
`
`
`PRESSURE
`MEASUREMENT
`28
`
`TEMPERATURE
`MEASUREMENT
`30
`
`Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 2 of 13
`
`

`

`U.S. Patent
`
`
`
`5,879,467
`
`
`
`(LHV XHOIRIA)
`
`Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 3 of 13
`
`90-E00" |
`
`PRESSURE (TORR)
`
`

`

`U.S. Patent
`
`Mar. 9, 1999
`
`Sheet 2 of 6
`
`5,879,467
`
`
`
`N
`N
`
`4
`
`s
`
`
`
`
`
`
`
`N
`
`
`
`
`
`
`
`
`
`
`
`S
`
`Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 4 of 13
`
`|R H.
`
`
`
`II
`
`MC
`
`v
`
`

`

`U.S. Patent
`
`Mar. 9, 1999
`
`Sheet 3 of 6
`
`5,879,467
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`>HETTO? H_LNOO
`
`NAOTH
`
`Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 5 of 13
`
`

`

`Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 6 of 13
`Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 6 of 13
`
`US. Patent
`
`Mar. 9, 1999
`
`Sheet 4 0f6
`
`5,879,467
`
`
`
`TIME(HOURS)
`
`Fig.4
`
`
`
`100E+03
`
`100E+01
`
`1ODE—01
`
`1.00E-O3
`
`1ODE-05
`
`1ODE-07
`
`1DOE-09
`
`400,1
`
`PRESSURE (TORR)
`
`

`

`U.S. Patent
`
`Mar. 9, 1999
`
`Sheet S of 6
`
`5,879,467
`
`OUTGASSING (TORR.LITERISECONDS)
`
`
`
`90-E00" |
`
`Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 7 of 13
`
`90-E00" |
`THROUGHPUT (TORR.LITERISECONDS)
`
`

`

`U.S. Patent
`
`Mar. 9, 1999
`
`Sheet 6 of 6
`
`5,879,467
`
`SanN&N 2
`
`so
`
`Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 8 of 13
`
`3
`
`3.
`
`g
`
`3
`
`g
`
`TEMPERATURE (C)
`
`

`

`5,879,467
`
`1
`CYCLE PURGING AWACUUM CHAMBER
`DURING BAKEOUT PROCESS
`
`BACKGROUND OF THE INVENTION
`
`1. Field of the Invention
`The present invention relates generally to a method for
`achieving the highest vacuum possible in the least amount of
`time for a given vacuum chamber. More specifically, the
`present invention relates to a method of performing a
`bakeout on a vacuum chamber involving cycle purging the
`chamber.
`2. Background of the Related Art
`Vacuum chambers are well known. Vacuum chambers can
`be used to manufacture integrated circuits (ICs) on Silicon
`wafers. Physical vapor deposition (PVD) chambers are one
`type of vacuum chamber for manufacturing ICs. PVD cham
`bers usually must reach about 6.0x10 torr to qualify for
`use in manufacturing ICS.
`The traditional method used for qualifying the vacuum of
`an ultra high vacuum chamber is called a bakeout. The
`bakeout process accelerates the removal of contaminants
`from the chamber, including driving out water vapor and
`other gases from the chamber components. The bakeout
`proceSS is used to determine the highest Vacuum, or lowest
`preSSure, that the vacuum chamber can attain. A vacuum
`chamber's highest vacuum may be limited by a variety of
`reasons, including leaks in the chamber or contaminants in
`the chamber.
`Various gases that have adsorbed onto the walls, instru
`ments or other interior Surfaces of the chamber may com
`prise the contaminants that limit a chamber's highest
`Vacuum potential. When a chamber is Subjected to a
`Vacuum, these gases come out of the interior Surfaces of the
`chamber in a process called degassing, or outgassing or
`desorption. When the pressure is returned to atmospheric
`preSSure, Such as prior to opening the chamber, water vapor
`and other gases in the chamber may be adsorbed into the
`interior Surfaces of the chamber, only to be outgassed later
`when the Vacuum returns.
`The first time that a vacuum chamber is subjected to a
`Vacuum, i.e., during the first bakeout, there may be a
`considerable amount of contaminants, absorbed or adsorbed
`gases, that must be outgassed. Subsequent times that the
`Vacuum chamber is Subjected to a vacuum should not have
`as many contaminants to outgas if the chamber can be kept
`relatively free of Such contaminants in the interim. Thus, the
`first bakeout proceSS is usually the longest period during
`which outgassing must be done.
`The conventional method used to obtain an ultra-high
`Vacuum for a chamber is to pump the chamber using a
`roughing pump to about 100 millitorr, then croSSover to the
`high vacuum pump. After the chamber reaches a threshold
`preSSure, an extensive bakeout, e.g. 36 hours in a PVD
`chamber, is performed. During the bakeout process, the high
`Vacuum pump continuously pumps the chamber.
`Afterwards, the chamber must be cooled to ambient tem
`perature in order to reach the base pressure, which may take
`about ten more hours.
`An example of the conventional bakeout method can be
`Seen in FIG. 1, which is a graph of pressure (in torr) inside
`the vacuum chamber versus time (in hours). Prior to time
`100, the roughing pump brings the vacuum down to about
`0.1 torr. Then the high Vacuum pump, e.g., cryogenic pump,
`reduces the vacuum as low as about 2x10 torr. At time 100,
`the bakeout lamps are turned on and the chamber will heat
`
`15
`
`25
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 9 of 13
`
`2
`up, causing extra outgassing in the chamber. This extra
`outgassing causes a rise in the pressure immediately after
`time 100 for 3-4 hours, after which, the pressure slowly
`drops for the next 10-15 hours as fewer and fewer contami
`nants remain to be outgassed from the chamber. At time 102,
`if the chamber has a wafer bakeoutheater, the heater may be
`turned on, thus causing another rise in the pressure after time
`102 when more outgassing from the heater occurs. At time
`104, the bakeout test is done, the lamps and heater are turned
`off, and the chamber is permitted to cool until the final
`qualifying preSSure may be determined. The time to perform
`this bakeout test is about 36 hours for a PVD chamber.
`The problem that slows down the conventional method of
`performing the bakeout is the rise in preSSure just after the
`bakeout lamps are turned on. This preSSure rise is due to the
`Speed at which the gases desorb from the chamber and are
`removed. For example, the outgassing rate can be calculated
`from the outgassing rate per area (typically 5.0x107
`torrliter/S'cm) and the total interior Surface area (typically
`10x10" cm for a PVD chamber). The product of these two
`parameters gives a typical outgassing rate of 5.0x10
`torrliter/s. By comparison, the cryopump's throughput can
`be calculated from its effective pumping speed (typically
`300 to 500 liter/s with the cryopump restrictor) and the
`pressure in the chamber (about 5.0x10 torr to 2.0x10
`torr). The product of these two parameters is the typical
`throughput for a cryopump: between about 1.5x10
`torr liter/s and about 1.0x10 torr liter/s. Thus, contami
`nants are outgassing almost five times as fast as the cry
`opump can remove them, So the pressure goes back up
`almost an order of magnitude after initially being reduced at
`time 100.
`The difference between the outgassing rate and the pump
`throughput indicates a very inefficient procedure that is
`permitting part of the desorbed contaminants to re-adsorb
`onto the interior Surface of the chamber. The typical gas
`composition for a typical high vacuum chamber with an
`O-ring Seal may contain as much as 50% water vapor, So
`water vapor and other contaminants are continually
`re-adsorbing onto the chamber interior, causing a long Slow
`bakeout process.
`For the process described by FIG. 1, the increase in
`pressure when the lamps are turned on at time 100 is known
`to have two parts to it. The pressure is related to temperature
`according to the general gas law:
`
`P=nKT,
`Equation 1
`where P is preSSure, n is the number density of the gas
`molecules, K is Boltzman's constant, and T is temperature.
`The change in pressure may be due to a change in either the
`temperature T or the density n. Therefore, for a change in
`preSSure, the general gas law may be expressed as:
`
`AP-nkATKTAn.
`Equation 2
`If Equation 2 is divided by Equation 1, the result is:
`APP=ATIT-Anfn.
`Equation 3
`
`The values for AP/P and AT/T can be calculated from the
`readings from the pressure measurement 28 and the tem
`perature measurement 30 (see FIG. 3). A typical value for
`AP/P during bakeout is in the range of 5-7, while a typical
`value for AT/T is less than 1. From Equation 3, a typical
`value for An/n is in the range of 4-6, which is greater than
`the range of values for AT/T. Therefore, a substantial amount
`of the cause for the rise in pressure after time 100 in the
`
`

`

`3
`conventional bakeout proceSS is due to the change in density
`of the gas, which confirms that the cryopump is not able to
`pump out the desorbed gases fast enough to keep up with the
`outgassing rate.
`Another method of performing a bakeout is shown in U.S.
`Pat. No. 5,536,330, issued Jul. 16, 1996, assigned to Applied
`Materials, Inc. of Santa Clara, Calif., and incorporated
`herein by reference as if fully Set forth below, and assigned
`in common with the present application. This patent
`describes a bakeout process that Starts with preheated gas
`Sweeping the Vacuum chamber. While maintaining a low
`vacuum, about 50 to 750 torr, hot gas is flowed through the
`chamber to Sweep out as much of the degassing molecules
`as possible. Then the vacuum is pumped down to an ultra
`high vacuum to continue the bakeout process. The gas is
`heated before flowing into the chamber. This Step requires a
`heater, which adds to the complexity of the test Setup.
`Additionally, the high pressure requires a very high purity of
`gas to water vapor ratio, because the partial preSSure due to
`the water vapor in the chamber will not go below that of the
`gas flowing into the chamber. Furthermore, the outgassing
`rate at the low vacuum is not very great.
`Attempts have been made to enhance this procedure by
`raising and lowering the pressure while Sweeping with
`preheated argon gas. However, all Such attempts have been
`proposed at the lower vacuums that require a greater gas
`purity. These attempts may have had a high throughput up to
`about 6.0x10 torr liter/s, but a low outgassing rate.
`It is, therefore, desirable to have a method for performing
`a bakeout process that can more quickly remove the des
`orbed gases from a vacuum chamber, and thereby reduce the
`total time for the bakeout test.
`
`SUMMARY OF THE INVENTION
`A bakeout process of an ultra-high vacuum System is
`performed by cycling the chamber between two pressures at
`medium vacuum, then pumping the chamber to a high
`Vacuum, and continuing to pump the System at the high
`Vacuum until the end of the process.
`A gas, Such as argon gas, may be introduced into the
`System while cycling between the two pressures. The argon
`gas may have a purity level of 99.99%, with less than 1 ppm
`of water vapor. The cycling part of the process performs an
`initial rapid Sweep of other gases or contaminants out of the
`System.
`To perform the pressure cycling, the pump is used to
`lower the preSSure, and the argon gas flow is used to raise the
`preSSure. The introduction of argon gas is done periodically
`to create the cycles, or it may be done continuously while the
`pumping action is varied to create the pressure cycles.
`Aroughing pump performs the preSSure cycles, and a high
`Vacuum pump pumps down to and holds the vacuum at the
`lower preSSure. Thus, the cycle preSSures are kept within the
`range of the roughing pump, about 10 torr and 0.05 torr. At
`this preSSure range, the pump throughput is significantly
`greater than the outgassing rate of the System, So the initial
`removal of most of the gas in the chamber occurs much more
`rapidly than in the prior procedure, and the gases cannot
`re-adsorb onto the chamber Surfaces.
`In one embodiment of the invention, a bakeout lamp is
`used during the cycling. The heat from the lamp enhances
`the outgassing rate from the chamber interior Surfaces.
`BRIEF DESCRIPTION OF THE DRAWINGS
`So that the manner in which the above recited features,
`advantages and objects of the present invention are attained
`
`Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 10 of 13
`
`5,879,467
`
`4
`and can be understood in detail, a more particular descrip
`tion of the invention, briefly summarized above, may be had
`by reference to the embodiments thereof which are illus
`trated in the appended drawings.
`It is to be noted, however, that the appended drawings
`illustrate only typical embodiments of this invention and are
`therefore not to be considered limiting of its Scope, for the
`invention may admit to other equally effective embodi
`mentS.
`FIG. 1 is a graph of a prior art bakeout method.
`FIG.2 shows a vacuum system that may be used with the
`present invention.
`FIG. 3 is a Schematic outline of the equipment used to
`carry out the process of the present invention.
`FIG. 4 is a graph of a bakeout method incorporating the
`present invention.
`FIG. 5 is a graph of a bakeout method showing a
`comparison of pump throughput with chamber outgassing.
`FIG. 6 is a graph of gas temperature during a bakeout
`method incorporating the present invention.
`DETAILED DESCRIPTION OF THE
`PREFERRED EMBODIMENT
`FIG. 2 generally shows a vacuum system 10, which may
`incorporate the present invention, having a pump 12
`mounted on a chamber 14. A gate valve 16 generally
`separates the pump 12 and the chamber 14. When gate valve
`16 is open, the pump 12 and the chamber 14 are in
`communication with each other; and when gate valve 16 is
`closed, the pump 12 and the chamber 14 are isolated from
`each other.
`The vacuum system 10 may be any kind of system that
`has a pump that reduces the pressure of the chamber. The
`described embodiment relates to any vacuum System used in
`manufacturing ICS on Silicon wafers, including a physical
`vapor deposition (PVD) vacuum system, a chemical vapor
`deposition (CVD) vacuum system, an etch vacuum system,
`etc. The described embodiment specifically shows a PVD
`Vacuum System, but it is to be understood that the present
`invention is not restricted to this one embodiment.
`Pump 12 includes a pump for high Vacuum pumping.
`Additionally, a roughing pump is connected through a
`foreline to chamber 14 in a known manner. Except where
`Specifically noted, references to a pump in this description
`are to the high vacuum pump Since the vacuum test
`described herein operates in the high vacuum region. High
`vacuum pumps include a cryogenic pump (Sometimes called
`a cryopump), a turbomolecular pump, and a getter pump. In
`the following description, the invention is described as
`including a cryopump; however, it should be understood that
`any of the other types of high Vacuum pumps could be used
`in accordance with the present invention.
`When gate valve 16 is open, the pump can pump down
`chamber 14 and maintain both Sections in a high vacuum.
`When gate valve 16 is closed, the pump will usually
`continue to try to maintain the high vacuum, but Since it is
`cut off from its pump, the vacuum in chamber 14 has nothing
`to maintain it, So the pressure in the chamber can rise if there
`is a Source for gases to enter it. In fact, it is expected that
`Small amounts of gases may continue to desorb from the
`interior Surfaces of the chamber causing a normal rise in the
`chamber pressure when the gate valve is closed. A failure in
`chamber 14, however, may cause its pressure to rise unac
`ceptably.
`Chamber 14 includes an ion gauge, not shown, for mea
`Suring the low preSSures in the chamber 14 at which most
`
`15
`
`25
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`

`

`S
`other gauges cannot operate. The ion gauge also shows the
`change in pressure over time. When gate valve 16 is open,
`the ion gauge shows the pressure in both chamber 14 and
`pump 12. When gate valve 16 is closed, the ion gauge shows
`the pressure only in chamber 14.
`Chamber 14 also includes one or more bakeout lamps, as
`shown in FIG. 3. A typical bakeout lamp contains about 99%
`Ar and about 1% N. The lamps may be used during bakeout
`and during actual manufacturing of ICS on the wafers to
`heat the chamber and the wafer.
`FIG.3 shows a schematic of the parts of a vacuum system.
`Chamber 14 has a lamp 32 as described above. The pressure
`measurement 28 for chamber 14 may be an ion gauge as
`described above. The temperature measurement 30 for
`chamber 14 may be any Suitable device, Such as a thermo
`couple. Vacuum pump 18 may be a System of one or more
`pumps. Vacuum pump 18 is part of vacuum Section 12 and
`is shown with a cryopump 20 and a rough pump 22 as
`described above. Alternatively, the rough pump 22 connects
`directly to the chamber 14 rather than going through the
`Vacuum Section 12. An argon Supply 24 provides argon for
`the bakeout test. A flow controller 26 regulates the flow of
`argon from argon Supply 24 into chamber 14.
`FIG. 4 shows a graph of a bakeout procedure for a
`chamber 14 that takes only about nine hours to perform.
`Prior to time 400, the gas in chamber 14 is cold purged by
`cycling between just below Standard pressure and about 0.1
`torr while flowing argon through chamber 14 in order to
`remove Some contaminants quickly. Then between times
`400 and 402, the cryopump 20 pumps down chamber 14 to
`less than 3x10' torr to further remove contaminants. At
`time 402, the bakeout process actually begins. The pressure
`in chamber 14 is brought up and cycled between about 0.1
`torr and about 10 torr, as described below, until time 404.
`The cycling time is called the ramp purge, chamber tem
`perature ramp or cycle purge. Each cycle may have a period
`about five to ten minutes long. The ramping down to the
`lower pressure and then back up to the higher pressure may
`be done as quickly as possible, and then the preSSure may be
`held at the higher preSSure for the remainder of the cycle
`period. The cycling continues for about twenty cycles until
`time 404. At time 404, chamber 14 is brought to a high
`vacuum and held there for the remainder of the bakeout test
`until time 406. The high vacuum is maintained after time
`406 while chamber 14 cools down. At time 406, the pressure
`in the chamber 14 is about 5x10' torr, and the temperature
`of the interior of the chamber 14 is about 250 C. As the
`chamber temperature cools to about room temperature over
`the next few hours, the pressure will decrease to about
`6x10 torr.
`As shown in FIGS. 4, 5 and 6, the inventive method uses
`the high flow characteristics of non-cryogenic pumps, in
`combination with pulsing of the outgassing caused by low
`preSSure and higher temperatures, in pressure regions where
`the higher capacity pumps may be used to advantage, to
`decrease the bakeout time for a high vacuum chamber. In the
`prior art, the paradigm for achieving high Vacuum with good
`integrity, i.e. with a slow rate of rise due to outgassing, was
`Simply to pump continuously Smaller and Smaller Volumes
`of outgassed material from the chamber using a cryopump.
`Thus, in the prior art, any increase in preSSure during the
`cryopumped portion of bakeout was unacceptable, because
`the cryopump had limited pumping capacity. The present
`invention counter-intuitively uses higher pressure cycling at
`the beginning of the bakeout period to rapidly purge or
`bakeout the chamber Surfaces, which allows the use of
`higher flow rate pumps which cannot effectively operate at
`
`15
`
`25
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 11 of 13
`
`5,879,467
`
`6
`very low preSSures, to rapidly remove the outgassed material
`from the chamber. Thus, when the bakeout period reaches
`the low pressure, low pumping speed level (the low pumping
`Speed caused by the inherent pumping Speed limitation of
`the cryopump), the pumping load caused by outgassing is
`Significantly reduced, thereby allowing the cryopump to
`reach its base pressure far more rapidly.
`With the procedure shown in FIGS. 4, 5 and 6, the
`bakeout test may take only about nine hours to perform. A
`Shorter bakeout time has also been done Successfully.
`Bakeout lamp 32 is turned on at time 402, the beginning
`of the bakeout process, and held on until time 406, the end
`of the bakeout test. The temperature in the chamber will
`slowly increase from room temperature to more than 250
`C., but will vary rapidly during the cycling time, as will be
`described below with reference to FIG. 6.
`A flow of non-preheated argon gas through chamber 14 is
`maintained throughout the cycle purge time by argon Supply
`24 and flow controller 26. Vacuum pump 18 takes the argon
`back out of chamber 14. The argon may be heated before it
`enters chamber 14, Since a hotter purge gas enhances the
`desorption of gases from the chamber interior Surfaces, but
`it has been found not to be necessary to preheat the argon in
`this procedure. Instead, the bakeout lamp 32 provides Suf
`ficient heat to heat chamber 14 to enhance the desorption and
`removal of the contaminants, as described below with ref
`erence to FIG. 6. This setup is simpler than one which
`requires a heater to preheat the gas. At time 404, the argon
`gas flow is turned off.
`The downward pressure ramp in a cycle is performed by
`Vacuum pump 18. The upward pressure ramp is performed
`by the argon gas flowing into chamber 14. The argon Supply
`24 may be periodically opened and closed to cause these
`cycles. This cycling procedure Strikes a balance between a
`high outgassing rate and a high throughput, because at the
`low-pressure end of each cycle the outgassing is greater, but
`the throughput is lower, than at the high-pressure end of each
`cycle, where the opposite is true. Therefore, by cycling the
`preSSure, the procedure alternates between releasing more
`gas into the chamber and then flowing it out of the chamber
`faster. This procedure removes the gas, and hence, the
`contaminants, faster than a steady flow of gas does. Thus, by
`the time the procedure reaches the end of the preSSure
`cycling Steps, considerably more of the adsorbed gases have
`been removed than in the same time period for the conven
`tional bakeout process. As a result, the rise in preSSure Seen
`in FIG. 1 following time 100 does not occur in the present
`procedure, So the present procedure can proceed to the end
`of the process much more quickly than can the conventional
`proceSS.
`The throughput of the pump compared with the outgas
`sing of the chamber is shown in FIG. 5 for the cycle purging
`method 450 as well as for the conventional bakeout proce
`dure 452. As can be seen beginning at time 454, the
`beginning of the bakeout procedure, the throughput for the
`cycle purge method 450 is considerably greater than the
`outgassing rate 456. Even during the first two hours, when
`the outgassing rate 456 rises considerably, the throughput of
`the cycle purge method 450 is always greater than the
`outgassing rate 456. Thus, the desorbed gases cannot be
`re-adsorbed onto the interior Surface of the chamber 14. The
`throughput for the conventional bakeout method 452,
`however, is seen shortly after time 454 to be below the rise
`in the outgassing rate 456 and remains below the outgassing
`rate 456 for almost two hours before the outgassing rate
`finally drops below the throughput of the conventional
`
`

`

`7
`bakeout method 452. This difference in throughput and
`outgassing rate is responsible for the rise in pressure begin
`ning at time 100 shown in FIG. 1. The throughput level of
`the cycle purge method 450, by comparison, does not drop
`down to or below the throughput level of the conventional
`bakeout method 452 until after the outgassing rate 456 has
`decreased, So the throughput is always greater than the
`outgassing rate.
`The cycling procedure also Strikes a balance between high
`temperature and high throughput of gas. FIG. 6 shows the
`temperature of chamber 14 during an exemplary cycling
`procedure that is similar to the one shown in FIG. 4. The
`bakeout test begins at time 500, where the pressure cycling
`begins. When the preSSure rises, the argon gas is flowing
`through chamber 14 at a rate of about 100 to 200 sccm. Since
`the argon is not preheated, it lowers the temperature as it
`flows into the chamber during the cycle. When the argon is
`turned off to let the preSSure drop during a cycle, the lamp
`32 heats chamber 14 to a higher temperature. At the start of
`the first cycle, time 500, the temperature at the high-pressure
`end of the pressure cycle is about 40 C. to 50 C., and the
`temperature at the low-pressure spike is about 70° C. to 80
`C. These high/low temperature variations slowly rise until
`the last cycle when the temperature at the high pressure is
`about 120° C. to 130° C., and the temperature at the low
`pressure is about 150° C. to 160° C. Afterwards, since the
`argon flow is turned off, the temperature rises dramatically
`to above 250 C., where the temperature rise begins to slow
`down, as the pressure drop approaches its limit. Therefore,
`Since higher temperatures cause greater outgassing rates, by
`cycling the temperature in this manner, the procedure once
`again alternates between releasing more gas from the cham
`ber interior Surfaces and then flowing it out of the chamber
`faster.
`The purity of the argon gas is preferably at least 99.99%
`argon and less than one part per million water vapor. The
`purge gas may also be nitrogen or other inert gas.
`If chamber 14 has a water bakeout heater, then it may be
`turned on during the cycle purge time. The heater may be
`slowly ramped up from 100° C. to 400° C.
`During the cycle purge period, the preSSure range is well
`within the range of rough pump 22, and during most of the
`time the chamber pressure is above the high pressure limit
`of the cryopump. Therefore, the cryopump 20 cannot be
`used for the cycling. Instead, the gate valve 16 has to be
`closed until the chamber 14 is crossed-over to the high
`vacuum pump at time 404.
`In one example of the procedure in FIG. 4, rough pump
`22 may be a Scroll pump or dry pump with an effective Speed
`of 10 liter/second. This effective speed is considerably lower
`than the 500 liter/second for the cryopump described above
`in the background of the invention. However, at a purge
`preSSure of 10 torr, the throughput for the Scroll pump may
`be 100 torr liter/s, which is considerably higher than the
`1.0x10 torrliter/s for the cryopump. Thus, the scroll pump
`may pump 100,000 times as much gas as the cryopump. In
`addition, the 100 torr-liter/s is considerably faster than the
`5.0x10 torrliter/s outgassing rate, So the Scroll pump is
`more than capable of removing gas fast enough to keep up
`with the outgassing of gases from the interior of chamber 14.
`AS a result, the flow of argon gas and the pumping action in
`this bakeout procedure Sweep away contaminants consider
`ably faster than Seen in the conventional bakeout process.
`By the time the procedure reaches the end of twenty
`cycles, the bakeout lamp 32 has been on for about 2-3 hours,
`and the chamber temperature has already increased to about
`
`15
`
`25
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`Case 6:20-cv-00636-ADA Document 92-7 Filed 03/31/21 Page 12 of 13
`
`5,879,467
`
`8
`150 C. or more, so the Subsequent temperature rise after the
`argon flow is turned off is not as Significant as it is in the
`conventional proceSS when the bakeout lamp is turned on.
`Additionally, by this time, most of the desorbed gases from
`the chamber 14 have been efficiently removed, so the
`chamber outgassing rate has decreased. As a result, the
`pressure rise seen at time 100 in FIG. 1 does not occur in
`FIG. 4.
`The higher pressure in the cycle purge time in FIG. 4 is
`shown to be about 10 torr, but higher cycle pressures are
`possible, even above 100 torr. In another example of the
`procedure in FIG. 4, rough pump 22 may be a dry pump with
`an effective Speed of 5 liter/s. If the higher pressure in a
`cycle can be 100 torr, then the throughput for the dry pump
`may be 500 torr liter/s, again considerably faster than the
`outgassing rate. The higher cycle pressure, however,
`requires a higher purity of argon gas, i.e., leSS Water Vapor,
`because of the partial pressure of the water. Such higher
`purity argon gas is difficult to obtain on a commercial basis.
`In contrast, the relatively lower purity argon gas is readily
`available commercially. It may be possible to take commer
`cially available ultra-pure liquid argon and run it through an
`argon purifier to reduce the water vapor partial pressure to
`a level of less than 10 parts per billion prior to flowing it
`through chamber 14.
`In yet another example, rough pump 22 may be a dry
`pump with an effective Speed of 5 liter/s. At a purge preSSure
`of 300 mtorr (common when there is a low flow rate from
`the flow controller), the throughput would be 1.5 torr liter/s.
`This is still 1000 times higher than the throughput of a
`cryopump.
`Experiments have shown that an optimum pressure for the
`high end of a cycle is in the range from about 0.3 torr to
`about 10 torr, with a preferred range being from about 0.5
`torr to about 10 torr, while an optimum pressure for the low
`end of a cycle is in the range from about 0.05 torr to about
`0.1 torr. The pressure at the low end of the cycle is actually
`only limited by the range of the roughing pump, Since it is
`preferred not to have to use the cryopump.
`While the foregoing is directed to the preferred embodi
`ment of the present invention, other and further embodi
`ments of the invention may be devised without departing
`from the basic Scope thereof, and the Scope thereof is
`determined by the claims which follow.
`We claim:
`1. A method of qualifying a vacuum System, comprising:
`(a) cycling the vacuum system between a first pressure in
`a range of about 0.3 torr to about 100 torr and a second
`preSSure in a range lower than the first preSSure, and
`(b) pumping the vacuum system down to a third pressure
`lower than the Second pressure.
`2. The method of claim 1, wherein cycling the vacuum
`System comprises introducing a gas into the vacuum System.
`3. The method of claim 2, wherein the gas comprises at
`least about 99.99% by volume argon.
`4. The method of claim 2, wherein the gas comprises less
`than about 1 part per million water vapor.
`5. The method of claim 2, wherein the gas is not preheated
`before it is introduced into the vacuum System.
`6. The method of claim 2, wherein cycling the vacuum
`System comprises:
`(d) lowering the pressure in the vacuum system by
`pumping the vacuum System down to the Second pres
`Sure, and
`(e) raising the pressure in the vacuum System by injecting
`the gas into the vacuum System until the pressure in the
`Vacuum System approximates the first pressure.
`
`

`

`5,879,467
`
`9
`7. The method of claim 6, further comprising:
`(f) repeating (d)-(e).
`8. The method of claim 2, wherein introducing a gas is
`performed periodically during cycling.
`9. The method of claim 1, further comprising:
`(c) heating the vacuum system during cycling.
`10. The method of claim 9, wherein heati