Integrated absolute and differential pressure transducer

[0001] 1. Field of the Invention

[0002] This invention is related to pressure sensors and, more particularly, to an integrated absolute and differential pressure sensor that can provide normalized, absolute and differential pressure measurements and outputs over a broad range of sub-atmospheric, atmospheric, and super-atmospheric pressures.

[0003] 2. State of the Prior Art

[0004] In some process, control, or monitoring applications, it would be very beneficial to have the capability of sensing pressure and providing accurate and repeatable pressure measurement or control outputs over a broad pressure range, such as from 10⁻⁸ torr or lower to 10³(1,000) torr or higher. For example, in a physical vapor deposition (PVD) or chemical vapor deposition (CVD) vacuum process chamber for depositing thin films of semiconductor materials on substrates or wafers to fabricate semi-conductor devices, a common deposition process practice may be some variation of the following: (i) Load the substrate or wafer into the vacuum process chamber at atmospheric pressure (e.g., about 600-770 torr); (ii) Close and seal the process chamber and evacuate it to 10⁻⁷ torr or less and hold it there for some period of time to remove all of the air, water vapor, and other potential contaminants, (iii) Back-fill the chamber with inert or over-pressure gas to bring the process chamber back up to about 10⁻³ torr, where it is maintained while process and carrier gasses are fed into the chamber to react or otherwise form a thin film of the desired semiconductor material(s) on the substrate or wafer, while effluents comprising gaseous by-products, unreacted and excess process gasses, and carrier gasses are drawn out of the process chamber; (iv) Stopping the process gasses; and (v) Back-filling the process chamber to increase the pressure in the chamber back to atmospheric pressure so that the chamber can be opened to remove the processed device.

[0005] Another approach is to keep the process chamber at the very low process pressure (vacuum) range used for the deposition processes, while a separate, often smaller, load lock chamber is used to handle the wafers before and after processing, i.e., to cycle between atmospheric pressure and process pressure to move the wafers into and out of the process chamber. The process chamber, when used with such a load lock, is only exposed to atmospheric pressure, therefore, when it is opened for servicing.

[0006] Such vacuum process and load lock systems currently require a plurality of different kinds of individual pressure sensors to measure and/or control pressures over such large ranges. For example, hot cathode pressure sensors are considered to be accurate and dependable for absolute pressure measurements in a range of about 5×10⁻¹⁰ to 5×10⁻² torr, but they are not useful for pressures above 5×10⁻² torr and have to be turned off to avoid burning out the filaments inside the hot cathode gauges. On the other hand, conventional convection pirani pressure sensors have absolute pressure measuring capabilities in a range of about 10⁻³ torr to 1,000 torr, but they are not useful for pressures below 10⁻³ torr, and they have a flat zone in a range of about 10 to 1,000 torr in which accuracy is low. A micropirani pressure sensor, such as the micropirani pressure sensor described in published U.S. patent application Ser. No. 09/907,541, which is incorporated herein by reference, can extend that range down to about 10⁻⁵ torr and alleviate the flat zone, but that range still is not sufficient alone for many processes.

[0007] Further, absolute pressure sensors are problematic in applications such as the vacuum process chamber described above, because, while it may be desirable to open the process chamber door at or very near ambient atmospheric pressure, ambient atmospheric pressure varies, depending on elevation above sea level, weather patterns, and the like, so any particular set point of an absolute pressure sensor is unlikely to match atmospheric pressure consistently. Thus, a differential pressure sensor may be required in addition to the one or two different kinds of absolute pressure sensors described above to provide the required process pressure measurements and controls, which still does not address the flat zone problems, especially where critical process operations are required or desired at pressures that coincide with such flat zones.

[0008] The combination absolute and differential pressure transducer described in the U.S. patent application Ser. No. 09/907,541, published on Jan. 16, 2003, provides a beneficial combination of an absolute pressure sensor with a differential pressure sensor for controlling the opening or closing of interior and exterior doors and other functions of load locks for vacuum processing chambers of transfer chambers. However, the absolute pressure measurements and the differential pressure measurements are separate from each other, and it provides no way to obtain or track absolute pressures above the absolute pressure measuring capability of the absolute pressure sensor and through the differential pressure sensor ranges. Of course, one or more different types of absolute pressure sensors could be added to the combination to provide higher absolute pressure measurements in the higher, differential pressure measurement ranges, but such additional pressure transducers add to the cost of the process equipment and are still not truly integrated in their respective measurements. Many process chamber operators and quality control technicians would like to see an entire process pressure profile on a single absolute pressure scale from atmospheric pressure or higher and down to the lowest vacuum pressure and then back up through those ranges to atmosphere again.

[0009] A general object of this invention, therefore, is to provide a method and apparatus for measuring and/or controlling pressures over wide pressure ranges with absolute and/or differential pressure outputs extending in an integrated manner over such ranges.

[0010] A more specific object of the invention is to provide real time absolute pressure measurements on a single scale extending from above atmospheric pressure to very low vacuum pressures.

[0011] A more specific object of this invention is to provide integrated absolute and/or differential pressure measurement and output capabilities over a pressure range of 10⁻⁸ torr or lower to 10³ torr or higher with as few as one or two absolute pressure sensors and one differential pressure sensor.

[0012] To achieve the foregoing and other objects, a method and apparatus for measuring absolute pressure in a chamber includes determining a correlation factor between absolute and differential pressure measurements taken simultaneously at a pressure where the absolute pressure in the chamber can be measured accurately and reliably, and then adjusting differential pressure measurements with the correlation factor to provide virtual absolute pressure measurements. The absolute pressure measurements are used in chamber pressure ranges where they are more accurate and reliable than the virtual absolute pressure measurements, and the virtual absolute pressure measurements are used in chamber pressure ranges where they are more accurate and reliable than the absolute pressure measurements. The correlation factor is re-determined periodically to adjust for changes in atmospheric pressure. The scope of the invention is defined in the claims below.

[0023] The present invention is illustrated for example, but not for limitation, in FIG. 1 by two pressure sensors 20, 30 connected in fluid flow relation to a chamber, which in this example is the interior 61 of a load lock chamber 60, and a microprocessor 80, which receives and processes signals originating from the pressure sensors 20, 30 to provide extended scale absolute pressure measurements of the chamber pressure P_C, as represented by the display 90. The first pressure sensor 20 is an absolute pressure sensor (P_ABS) for sensing the absolute pressure P_C in the chamber 61. The second pressure sensor 30 is a differential pressure sensor (ΔP) for sensing the difference between the atmospheric pressure P_A outside the chamber 61 and the gas pressure P_C inside the chamber 61. In other words, P_ABS=P_P, and ΔP=P_A−P_P.

[0024] To provide some context to aid in understanding the invention, a load lock 60 is often used in semiconductor fabrication to shuttle one or more wafers 73 into and out of a vacuum process chamber 70, where one or more feed gases delivered from feed gas sources 74, 75, 76 are reacted to deposit thin film materials, such as semiconductor material 77, on the wafer 73. A vacuum pump 71 connected to the interior of the vacuum process chamber 70 pumps gases out of the vacuum process chamber 70 to maintain a desired vacuum, i.e., low pressure, usually less than 1 torr and can be down to 10⁻⁸ torr or less, depending on the process requirements. A platform 72 is usually provided in the vacuum process chamber 70 for supporting the wafer 73 during processing.

[0025] The interior chamber 61 of load lock 60 in this example is connected by a passage 69 to the interior of the vacuum process chamber 70. The purpose of the load lock 60 is to facilitate transfer of the wafers 73 from the outside (e.g., ambient) atmosphere into the vacuum process chamber 70 without losing the vacuum in the vacuum process chamber 70 or allowing contaminants into the vacuum process chamber 70. Therefore, an interior door or valve 62 in the passage 69 opens the passage 69 to allow transfer of wafers 73 into and out of the vacuum process chamber 70 and closes the passage way 69 to seal the vacuum process chamber from the load lock chamber 61, when there are no such transfers being conducted. In some process tools, there is an intermediate transfer chamber (not shown) between the load lock and several process chambers to facilitate transferring wafers from one process chamber to another without having to go through the load lock or the atmosphere.

[0026] An exterior door 64 on the load lock 60 opens and closes the load lock chamber 61 to the atmosphere. When a wafer, illustrated in phantom lines 73′ to indicate its transitory position, is being transferred from the outside atmosphere into the load lock chamber 61, the interior door 62 is closed, and the exterior door 64 is open. With the exterior door 64 open, the absolute chamber pressure P_C in the load lock 60 is substantially equal to the absolute atmospheric pressure P_A outside the chamber 61, as indicated at 91 in the display 90, so the differential pressure ΔP is zero. Meanwhile, the vacuum in the vacuum process chamber 70 behind the closed interior door 62 is maintained, i.e., at a much lower pressure P_P than the outside atmospheric pressure P_A. Then, with the wafer 73′ inside the load lock chamber 61, the exterior door 64 is closed, and a vacuum pump 65 connected to the load lock chamber 61 pumps enough of the air and other gases out of the load lock chamber 61 to remove potential contaminants and to reduce the absolute chamber pressure P_C, as indicated by the sloped line 92 in the display 90, to a desired level 93 that substantially matches the low absolute pressure P_P in the vacuum process chamber 70, which can be measured with an absolute pressure sensor 78. At some chamber pressure P_C, such as the pressure point 94, during the pump-down phase 92, where most of the air and potential contaminants have been pumped out of the chamber 61, a common practice is to open a throttle valve 66 to bolster the pump-down speed in the lower absolute chamber pressure P_C portion of the pump-down phase 92. This pressure point 94 is often called a cross-over pressure, because it is where the pump-down of the chamber 91 crosses over from slow to high speed. However, that term for such operation or functionality is not used here to avoid confusion with another cross-over function explained below, which is more central to this invention.

[0027] With the absolute chamber pressure P_C pumped down to substantially the same level 93 as the absolute pressure P_P in the vacuum process chamber 70, the interior door 62 can be opened to allow the wafer 73 to be transferred into the vacuum process chamber 70 and placed on the platform 72. A moveable shuttle (not shown) in the load lock 60 is used to move the wafer 72 into and out of the vacuum process chamber 70, as is well-known to persons skilled in the art.

[0028] With the wafer 73 in place on the platform 72, the interior door can be closed for a time, while the semiconductor material 77 is deposited on the substrate 73. The absolute chamber pressure P_C can be maintained at the low level 93 during the deposition of the semiconductor material 77. When the semiconductor material 77 is deposited on the wafer 73, the interior door 82 is opened again to transfer the wafer 73 back into the load lock chamber 61. Then, the interior door 62 is closed again to isolate the interior of the vacuum process chamber 70 from the load lock chamber 61, so that the absolute chamber pressure P_C can be raised, as indicated at 95, back to atmospheric pressure P_A, as indicated at 96, without affecting the process pressure P_P in the vacuum process chamber 70. A common practice is to use a gas, such as nitrogen, or an inert gas, such as argon, from a source 63 to back-fill the load lock chamber 61 to raise the absolute chamber pressure P_C back up to atmospheric pressure P_A, but air is also sometimes used for this purpose.

[0029] While it is important to have accurate and reliable absolute pressure measurements of the chamber pressure P_C going down to the absolute chamber pressure P_C level 93 in order to match the process pressure P_P in the vacuum process chamber 70 before opening the interior door 62, as explained above, it is also important to be able to measure accurately when the chamber pressure P_C matches the atmospheric pressure P_A before the exterior door is opened. Unfortunately, however, there are no absolute pressures sensors that can measure the absolute chamber pressure P_C accurately and reliably over the full pressure range from atmospheric pressure P_A level 91 down to the low pressure level 93 that is needed to match the process pressure P_P in the vacuum process chamber 70. Also, the atmospheric pressure P_A varies significantly with elevation above sea level and with ambient weather conditions, so there is no fixed absolute pressure set point that can be used for the chamber pressure P_C to open the exterior door 64. Consequently, it is necessary to use an absolute pressure sensor 20 that is accurate and reliable at the lower pressure levels to measure absolute chamber pressure P_C for determining when to open the interior door 62, but also to use a differential pressure sensor 30 for determining when the chamber pressure P_C matches the atmospheric pressure P_A in order to open the exterior door 64 at or near atmospheric pressure P_A as indicated at 99 in the display 90. Of course, two higher pressure range absolute pressure sensors (not shown), one for measuring absolute atmospheric pressure P_A and the other for measuring absolute chamber pressure P_C, and an analog or digital comparator circuit (not shown), microprocessor, or other means for comparing such measurements, could be substituted for the differential pressure sensor 30, as is understood by persons skilled in the art. Therefore, the use of the term “differential pressure sensor” herein is meant to include not only conventional, direct read differential pressure sensors or gauges, but also the use of two absolute pressure sensors with circuitry for subtracting measurements of one from measurements of the other to measure differential pressures, as well as any other apparatus or method that is capable of providing differential pressure measurements.

[0030] Use of an absolute pressure sensor 20 for determining when the absolute chamber pressure P_C is low enough to open the interior door 62 in combination with a differential pressure sensor 30 for determining when to open the exterior door 64 of a load lock 60 is well-known, as explained in the published U.S. patent application Ser. No. 09/907,541 and U.S. patent application Ser. No. 09/815,376, both of which are incorporated herein by reference. However, as mentioned above, many vacuum process chamber operators, quality control personnel, and others would like to have a full absolute pressure profile 98 on a single scale that extends over the full chamber pressure P_C range of operation from absolute atmospheric pressure P_A level 91, 96 or above, down to the lowest absolute pressure level 93 or below for diagnostic, quality control, design, maintenance, and other reasons.

[0031] An important feature of this invention, therefore, is to integrate pressure measurements from at least one low or mid-level absolute pressure sensor 20 with differential pressure measurements from the differential pressure sensor 30 to produce accurate and reliable absolute pressure measurements of the chamber pressure P_C on a single scale that extends over a range low enough for opening the interior door 62 in this kind of load lock as well as other process applications and high enough to provide a complete absolute chamber pressure P_C profile 98, including absolute atmospheric pressure P_A. However, this invention is not limited to load lock applications. On the contrary, it is useable for any other application in which such an extended absolute pressure measurement range beyond (above or below) the acceptable accuracy and reliability range of a low or mid-level absolute pressure sensor is needed or desired.

[0032] According to this invention, as illustrated in FIG. 2, the measurements provided by the absolute pressure sensor 20 are used for the absolute chamber pressures P_C below a cross-over pressure P_X level or range on an absolute pressure scale 40 in the display 90, where the absolute pressure sensor 20 has the capability to provide accurate and reliable absolute pressure measurements. However, for absolute chamber pressure P_C measurements above the cross-over pressure level or range P_X, where the absolute pressure sensor 20 does not provide sufficiently accurate and reliable absolute pressure measurements, normalized virtual differential pressure measurements, which are based on differential pressure measurements P₃₀ from the differential pressure sensor 30 are used for the absolute chamber pressure P_C on the absolute pressure scale 40 in the display 90. To normalize such differential pressure measurements P₃₀ from the differential pressure sensor 30 for this purpose, they are correlated with measurements P₂₀ of absolute chamber pressure P_C from the absolute pressure sensor 20 at or below a correlation pressure threshold point P_t, which is preferably at or near a pressure where the practical accuracy of the differential pressure sensor 30 effectively reaches its bottom, i.e., where further decimal places of pressure measurements are not meaningful or significant in a practical application or where physical structure or electric circuit limitations render lower differential pressure ΔP measurements with the differential pressure sensor 30 practically meaningless. This correlation pressure threshold point P_t, or any pressure lower than P_t that is still within an accurate and reliable absolute pressure measuring range of the absolute pressure sensor 20, can be used as a base line for adjusting or normalizing the differential pressure ΔP measurements P₃₀ from the differential pressure sensor 30 to correlate with the absolute chamber pressure P_C measurements P₂₀ from the absolute pressure pressure sensor 20 at a desired level of accuracy, as will be explained in more detail below. Therefore, such adjustment or normalizing correlation factor F can be used to convert or normalize all higher differential pressure ΔP measurements P₃₀ from the differential pressure sensor 30 to the same scale as the absolute chamber pressure P_C measurements P₂₀ from the absolute pressure sensor 20, as is shown in FIG. 3 and will be described in more detail below. Consequently, as the chamber pressure P_C rises above that correlation pressure threshold P_t, the correlation factor F can be added to all of the differential pressure ΔP measurements from the differential pressure sensor 30 to convert them to virtual absolute chamber pressure P_C measurements P_V correlated to the same absolute pressure scale 40 as the absolute pressure measurements P₂₀ produced by the absolute pressure sensor 20.

[0033] Eventually, as the chamber pressure P_C continues to increase, it will reach some chamber pressure P_C level that is still substantially below the atmospheric pressure P_A, but which is above the accurate and reliable absolute pressure measuring capability of the absolute pressure sensor 20. Therefore, the absolute chamber pressure P_C measurements P₂₀ by the absolute pressure sensor 20 above that accuracy and reliability level become unreliable and unusable. However, the differential pressure sensor 30 continues to provide accurate and reliable differential pressure ΔP measurements as the chamber pressure P_C rises all the way up to the atmospheric pressure P_A and beyond. Therefore, in the higher chamber pressure P_C levels, where the absolute pressure measurements P₂₀ by the absolute pressure sensor P₂₀ are unreliable, accurate and reliable virtual absolute chamber pressure P_C measurements P_V can be provided on the same continuous scale 40 all the way up to the atmospheric pressure P_A level and beyond by adding the correlation factor F to the differential pressure ΔP measurements P₃₀ from the differential pressure sensor 30.

[0034] It is preferable, however, to not wait until the absolute pressure sensor 20 reaches the end of its accuracy and reliability range before crossing over to the virtual absolute chamber pressure P_C measurements P_V for output by the microprocessor 80 and display by the display device 90. Instead, it is preferable, but not essential, to select a cross-over pressure level P_X, which can be either a distinct cross-over pressure point or a cross-over pressure range with a smoothing function (explained in more detail below), at which the displayed absolute chamber pressure profile 98 is assembled from the virtual absolute chamber pressure P_C measurements P_V rather than from the absolute pressure sensor 20 measurements P₂₀, as will be explained in more detail below.

[0035] To further illustrate one of the principles on which this invention is founded, the absolute pressures P_ABS in a typical vacuum process range are juxtaposed in FIG. 2 to corresponding differential pressures ΔP on a logarithmic scale 40 in units of torr, although any other units of pressure could also be used. In the first column 42 in FIG. 2, a scale of absolute pressures P_ABS is shown extending from 1,000 torr at the upper end of the scale 40 down to 10⁻⁵ torr (0.00001 torr) at the lower end of the scale 40. Some processes are performed in pressures as low as 10⁻⁸ torr or lower, but it is not necessary to extend the scale 40 in this FIG. 2 that low in order to convey the principles of this invention.

[0036] Two arbitrary different example atmospheric pressures P_A—sea level and high plains, e.g., Boulder, Colo. U.S.A.—are used in FIG. 2 to facilitate an explanation of this invention, but any other example atmospheric pressures could also be used. Also, while 760 torr is considered by persons skilled in the art to be a standard atmospheric pressure P_A at sea level, the actual atmospheric pressure P_A will vary above and below that 760 torr level due to various weather patterns. Likewise, while 630 torr is a common atmospheric pressure P_A at high plains and foothills locations, such as Boulder, Colo. U.S.A., the actual atmospheric pressure P_A at any such location will vary above and below that level as weather patterns change. However, such variations are accommodated by this invention in order to continue making very accurate and reliable absolute pressure P_ABS measurements, regardless of atmospheric pressure P_A changes, as will be understood by persons skilled in the art, once they understand this invention. In fact, as mentioned above, one of the purposes of this invention is to provide accurate and reliable chamber pressure P_C measurements up to and beyond whatever local atmospheric pressure P_A might exist at any time, even as such local atmospheric pressure P_A changes from day-to-day, hour-to-hour, or smaller increments of time.

[0037] As illustrated in FIG. 2, when the atmospheric pressure P_A is, for example, 760 torr and the absolute chamber pressure P_C (FIG. 1) is also 760 torr, such as when the exterior door 64 is open, then the differential pressure ΔP between the atmospheric pressure P_A and the chamber pressure P_C in the chamber 61 (FIG. 1) will be, of course, zero, as shown in column 44 in FIG. 2. As the chamber 61 is evacuated by the vacuum pump 65 (FIG. 1), and the absolute chamber pressure P_C is lowered by 660 torr to the absolute chamber pressure P_C of, for example, 100 torr, the differential pressure ΔP will also drop from zero to −660 torr, shown in FIG. 2.

[0038] Likewise, if the process is being performed at a high plains location with an atmospheric pressure P_A of, for example, 630 torr instead of the 760 torr sea level atmospheric pressure, the differential pressure ΔP between the atmospheric pressure P_A and the chamber pressure P_C with the exterior door 64 open, will also be zero, as shown in FIG. 2. However, there is a 130 torr difference between the 760 torr sea level P_A and the 630 torr high plains P_A. Therefore, as the chamber 61 is evacuated down to an absolute chamber pressure P_C of 100 torr, the differential pressure ΔP at the high plains location will be −530 torr instead of the −660 torr at the sea level location.

[0039] The same 130 torr discrepancy between this sea level example and this high plains location example shows in all of the differential pressure ΔP measurements as the chamber pressure P_C is lowered. In other words, lowering the absolute chamber pressure P_C of 100 torr down to 10 torr—a 90 torr drop—correspondingly lowers both the sea level location differential pressure ΔP and the high plains location differential pressure ΔP by 90 torr, i.e., from −660 torr down to −750 torr at the sea level location and −530 torr at the high plains location, as shown in FIG. 2. Likewise, lowering the absolute chamber pressure P_C another 9 torr down to 1 torr would lower the differential pressures ΔP at the sea level location and the high plains locations to −759 torr and −629 torr, respectively. An absolute chamber pressure P_C of 0.1 torr corresponds to differential pressures of ΔP of −759.9 torr at the sea level location and −629.9 torr at the high plains location, and 0.01 torr, 0.001 torr, 0.0001 torr, and 0.00001 torr absolute chamber pressures P_C correspond to −759.99 torr, −759.999 torr, −759.9999 torr, and −759.99999 torr at the sea level location and to −629.99 torr, −629.999 torr, −629.9999 torr, and −629.99999 torr at the high plains location.

[0040] While there are absolute pressure sensors available that can measure pressures down to 0.00001 torr (i.e., 10⁻⁵ torr) and below quite accurately and reliably, the available differential pressure sensors effectively bottom out in accuracy at about one decimal place, i.e., when the absolute chamber pressure P_C is about 0.1 torr. At about that level, differential pressure ΔP measurements down to any further decimal places, such as −759.99 torr or −759.999 torr for the example sea level location and −629.99 torr or −629.999 torr for the example high plains location, are practically meaningless, just as the difference between differential pressure ΔP measurement P₃₀ with the differential pressure sensor 30 of −660.99 torr and −660.999 torr at the sea level location would be practically meaningless, because the available differential pressure sensors are just not accurate to more than one or two decimal places in torr units. However, in the upper pressure ranges, for example, above 100 torr, measurement accuracies to one or even zero decimal places is usually sufficient for most process operations. Therefore, correlating absolute chamber pressure P_C measurements P₂₀ from the absolute pressure sensor 20 to the effectively bottomed-out differential pressure ΔP measurements P₃₀ from the differential pressure sensor 30 when the absolute chamber pressure P_C is below a threshold pressure P_t of 1 torr or perhaps 0.1 torr or lower, depending on the decimal places of accuracy needed at higher differential pressure ΔP measurements, provides an effective, accurate, and repeatable baseline for normalizing the differential pressure sensor 30 measurements to the absolute pressure sensor 20 measurements. With such baseline normalization, the differential pressure sensor 30 measurements P₃₀ can, according to this invention, be normalized or converted with a normalizing correlation factor to virtual absolute chamber pressure P_C measurements P_V, including in the higher chamber pressure P_C ranges, where the differential pressure sensor 30 is most accurate and dependable and the absolute pressure sensor 20 losses its accuracy and dependability. For example, if the differential pressure ΔP measurements P₃₀ of differential pressure pressure sensor 30 are normalized to the accurate and reliable absolute chamber pressure sensor 20 measurements P₂₀, when the chamber pressure P_C is at a baseline threshold pressure P_t of 1 torr or below, then, as the absolute chamber pressure P_C rises above some cross-over pressure level P_X, such as 100 torr, where the differential pressure sensor 30 is more accurate and reliable than the absolute pressure sensor 20, the differential pressure ΔP measurements P₃₀ of the differential pressure sensor 30 can be converted to accurate and reliable virtual absolute chamber pressure P_C measurements P_V all the way up to atmospheric pressure P_A level and above. In fact, a piezo differential pressure sensor is accurate from about 100 torr below atmospheric pressure P_A to about 1,500 torr above atmospheric pressure P_A, which is a range of about 1,600 torr. Therefore, normalizing such differential pressure measurements P₃₀, as described above, can provide accurate and dependable virtual absolute pressure measurements P_V over a 1,600 torr range. The exact lower and upper ends 43, 45 of such virtual absolute pressure measurement P_V range on the absolute pressure scale 40 will depend on what the atmospheric pressure pressure P_A happens to be at any particular time, as illustrated in FIG. 4, which is described in more detail below.

[0041] An example normalizing method for implementing this invention to produce such virtual absolute chamber pressure P_C measurements P_V is illustrated in FIGS. 3 and 4, although other methods or variations can be devised by persons skilled in the art, once they understand the principles of this invention. For purposes of explanation of this method in FIG. 3, reference is also made to a chart of effective measuring ranges of example absolute and differential pressure sensor 20, 30 ranges in FIG. 4 and an example load lock chamber 61 pressure profile in FIG. 5, as well as with continuing reference to the FIGS. 1 and 2 discussed above.

[0042] In this example, the absolute pressure sensor 20 is shown in FIG. 4 to have an effective measuring range extending from about 100 torr down to about 10⁻⁵ torr, and the absolute pressure measurements by the absolute pressure sensor 20 are called P₂₀ for convenience. The micropirani absolute pressure sensor described in U.S. patent application No. 09/907,541, which is incorporated herein by reference, has this kind of range, while conventional convection pirani absolute pressure sensors have a somewhat narrower range of about 100 torr down to about 10⁻³ torr. Other thermal conductivity type absolute pressure sensors as well as diaphragm based absolute pressure sensors, such as low range capacitive manometers, low range piezo, strain gauge, and others also are effective in the upper portions of this example absolute pressure sensor range.

[0043] The differential pressure sensor 30 for this example description is shown in FIG. 4 to have an effective measuring range from above atmospheric pressure (e.g., about 1,500 torr above P_A) down to about minus 99.9 percent of atmospheric pressure (−99.9% Atm), i.e., down to about 10⁻¹ torr, although it is more accurate and reliable above about 1 torr. The differential pressure measurements by the differential pressure sensor 30 are called P₃₀ for convenience, and, as mentioned above, absolute pressure measurements by the absolute pressure sensor 20 are called P₂₀.

[0044] The bar chart in FIG. 4 is divided into a differential pressure domain 42 on top and an absolute pressure domain 44 on bottom to aid in illustrating the concept that the differential pressure sensor 30 output measurements P₃₀ shift laterally in relation to the absolute pressure scale 40 and to the absolute measurements P₂₀, as indicated by arrow 41 and phantom lines 43, 45 in FIG. 4, with the amount of such shift depending on the changes in atmospheric pressure. If the atmospheric pressure P_A is higher, the differential pressure measurements P₃₀ shift to the right in relation to the absolute pressure seal 40 in FIG. 4, and, if the atmospheric pressure P_A moves lower, the differential pressure measurements shift to the left in relation to the absolute pressure scale 40.

[0045] The absolute pressure profile 98 in FIG. 5 is an enlargement of the absolute pressure profile 98 for a typical load lock cycle in the display 90 of FIG. 1. The scale 40 at the left is absolute pressure in torr, the scale 140 at the right is differential pressure in torr, and the scale 142 at the bottom is time in minutes, although any suitable pressure and time units can be used. To re-cap, the cycle starts at atmospheric pressure P_A indicated at 91, and it descends downwardly 92 during evacuation of the load lock chamber 61 to a base level pressure 93, where it is held for a time during transfer of a wafer 73 into (and, if desired, back out of) the reaction chamber 70 (FIG. 1). Then, as the load lock chamber 61 (FIG. 1) is back-filled, the absolute chamber pressure P_C rises back, as shown at 95, to atmospheric pressure P_A as indicated at 96.

[0046] The absolute chamber pressures P_C for the absolute pressure profile 98, as well as for certain switching functions for the load lock operation (e.g., opening the throttle valve 66 at a pressure point 94, and opening the interior door 62 at a pressure point 97) can be provided, for example, by the method 10 shown in FIG. 3. At the start of the method 10, it is desirable, although not necessary, to choose an initial correlation factor F₀ 46 for use in converting differential pressure measurements P₃₀ to normalized virtual absolute pressure measurements P_V. For example, the initial correlation factor F₀ could be at 760 torr for a sea level location or at 630 torr at a high plains location, which would provide an approximation good enough for the first part of the absolute pressure profile 98 down to the cross-over pressure P_X for the first load lock cycle. Therefore, the correlation factor F is set at 47 in FIG. 3 equal to the initial correlation factor F₀. The absolute pressure measurement P₂₀ and the differential pressure measurement P₃₀ are read at 48 from the absolute pressure sensor 20 and from the differential pressure sensor 30.

[0047] The differential pressure ΔP is set at 49 to the value of the measurement P₃₀ and is output at 50 for any desired display or control functions, such as to open the exterior door 64 of the load lock 60 (FIG. 1) when ΔP=0, as indicated at 99 in the example process pressure profile in FIG. 5. Then, the virtual absolute chamber pressure measurement P_V is calculated at 51 by adding the correlation factor F to the differential pressure measurement P₃₀. For example, if the correlation factor F is 760 torr, then the virtual absolute chamber pressure measurement P_V is the differential pressure measurement P₃₀ plus 760 torr.

[0048] Next, as shown in FIG. 3, if the absolute chamber pressure P_C measurement P₂₀ is not greater than the cross-over pressure level P_X at 52, then the absolute chamber pressure P_C is set equal to the absolute pressure measurement P₂₀ at 53. If P₂₀ is greater than the cross-over pressure level P_X at 52, then the absolute chamber pressure P_C is set equal to the virtual absolute chamber pressure measurement P_V at 54. Then, the absolute chamber pressure P_C, whether set equal to P₂₀ at 53 or equal to P_V at 54, is output at 55 for any desired control and display functions. Alternatively, the test at 52 could be whether P_V>P_X instead of whether P₂₀>P_X, with an equivalent effect.

[0049] Finally, if the absolute chamber pressure measurement P₂₀ is not less than the correlation threshold pressure P_t at 56, then the method 10 loops back via 57 to obtain another reading of the absolute chamber pressure measurement P₂₀ and of the differential pressure measurement P₃₀ for another iteration through the logic to obtain new ΔP and P_C values as the chamber pressure P_C decreases. However, if the absolute chamber pressure measurement P₂₀ is less than the correlation threshold pressure P_t at 56, then the correlation factor F is recalculated at 58 before the method 10 loops back via 59 for another iteration. As discussed above the correlation threshold pressure P_t is preferably, but not necessarily, low enough to be at or near the bottom of the differential pressure measuring capability of the differential pressure sensor 30 at whatever minimum decimal place is needed for the precision desired in the virtual pressure measurements P_V. However, P_t should not be so low that the actual absolute chamber pressure P_C never or rarely gets that low, because, according to the method 10, the correlation factor F only gets updated to compensate for any atmospheric pressure P_A changes due to weather changes or other causes, when the absolute chamber pressure P_C falls below P_t. Therefore, if the chamber pressure P_C is cycled to drop below P_t every hour, for example, the correlation factor F will be updated every hour to compensate for any atmospheric pressure P_A changes due to weather or otherwise. Such updates keep the P_C outputs at 55 accurate, even above P_X, regardless of atmospheric pressure P_A changes.

[0050] While setting the correlation pressure threshold P_t at or very near a level where the differential pressure sensor 30 measuring capabilities essentially zero out is very convenient and effective, especially for processes that cycle down below that level quite often, any pressure level where the relationship between an accurate, reliable, absolute pressure measurement and a differential pressure is known can be used to determine the correlation factor. For example, if the absolute atmospheric pressure P_A is known from some other source at a particular instant in time when the differential pressure sensor measures zero, such absolute atmospheric pressure P_A value can be used to set the correlation factor F. Such other source could be, for example, another absolute pressure sensor that is accurate and reliable at that level.

[0051] As mentioned above, and as shown in FIG. 4, the absolute chamber pressure P_C output at 55 in the method 10 in FIG. 3 comprises a continuous composite of absolute pressure measurements P₂₀ and virtual absolute pressure measurements P_V, which effectively extends the range of accurate and reliable absolute chamber pressure P_C measuring capability higher than the capability of the absolute pressure sensor 20 alone—all the way up to atmospheric pressure P_A and beyond. The cross-over pressure P_X is preferably selected to be at the level at which virtual absolute pressure measurements P_V are more accurate and reliable than absolute pressure measurements P₂₀ from absolute pressure sensor 20 and vice-versa. This simple cross-over at P_X, where P_X is a single pressure point, is acceptable for many applications. However, any common smoothing function, which is known to persons skilled in the art, can be used, if desired, to blur or spread the cross-over level over a range 103 to ensure that the pressure profile 98 in FIG. 5 does not have a sharp bend or kink where the cross-over occurs. Basically, a smoothing function starts by weighting the P_C value more with P₂₀ at one end of the range 103 and changing gradually to weighting the P_C value more heavily with P_V at the other end of the range 103 so that the P_C output in the range 103 would be a blended value of P₂₀ and P_V.

[0052] With the correlation factor F updated at 58 in FIG. 3 after the absolute chamber pressure measurement P₂₀ falls below the correlation threshold pressure P_t, as explained above, the output P_C at 55 is an accurate measurement of the absolute chamber pressure P_C all the way up the rising portion 95 of the chamber pressure profile 98 to atmospheric pressure 96 and above. However, the differential pressure ΔP output at 50 in FIG. 3 can still be used to open the exterior door 64 (FIG. 1), if desired. That same updated correlation factor F continues to be used for the correlation of the differential pressure measurements P₃₀ to virtual absolute pressure pressure measurements P_V at 49 in FIG. 3 for the descending portion 92 (FIG. 5) of the next load lock evacuation cycle, since it is the correlation factor F that is related most recently to the actual atmospheric pressure P_A and is, therefore, usually more accurate than the initial correlation factor F₀.

[0053] The logic in the process discussed above and illustrated in FIGS. 3-5 can be implemented in any manner, such as a microprocessor 80, as shown in FIG. 1 or any other convenient manner known to persons skilled in the art. Signals from the absolute and differential pressure sensors 20, 30 can be communicated to the microprocessor 80 by any convenient communication links 21, 31, and signals from the microprocessor to operate the doors 62, 64, the throttle valve 66, and the display 90 can be via suitable communication links 84, 83, 68, 86, respectively. Such communications links can be hard wired, radio frequency, infrared, sound, or any other signal communication technique. Also, the signals and processed information can be handled or stored in any buffers, filters, amplifiers, analog-to-digital converters, memory devices, and other conventional signal processing components (not shown), as is known to persons skilled in the art. The display 90 is used generically here and can be visual, printed, projected, or any component for receiving, using, storing, or displaying the pressure outputs ΔP and/or P_C from the process described in relation to FIG. 3 or otherwise in accordance with this invention. The pressure sensor 78 can be, but does not have to be, connected to the microprocessor 80 via communication link 85 for monitoring, comparison, display, or the like.

[0054] As explained above and illustrated in FIG. 4, the absolute pressure sensor 20, with current technologies, is unlikely to be able to measure very low absolute pressures, e.g., below 10⁻⁴ or 10⁻⁵, and still extend high enough, e.g., 1 to 100 torr, to get within a useable correlation range, e.g., 10⁻² to 1 torr. Absolute pressure sensors 20 in these ranges are considered to be mid-range absolute pressure sensors. If it is desired to evacuate the chamber pressure P_C down to even lower pressure levels, such as the 10⁻⁷ torr process base pressure level 113 in the example process pressure profile 98 in FIG. 5, a second, low absolute pressure sensor 25, as shown in FIGS. 8 and 9 can be added to this invention. The low absolute pressure sensor 25 for example, an ion gauge, a hot cathode pressure sensor, or a cold cathode pressure sensor, can extend the range of accurate and reliable absolute chamber pressure P_C outputs down to 10⁻⁸ torr or below, as illustrated in FIG. 7. An ion gauge, a hot cathode gauge, and a cold cathode gauge are examples of absolute pressure sensors that are capable of providing accurate and reliable absolute pressure measurements P₂₅ at these low absolute pressure levels.

[0055] To illustrate the operation of this second, lower range, absolute pressure sensor 25 in combination with the mid-range absolute pressure sensor 20 and the differential pressure sensor 30 in this invention, reference is made now primarily to FIGS. 6-9. In this example, the process chamber 70, shown in FIG. 9, is not equipped with a load lock, and the exterior door 164 is positioned to open and close the passage 69 into the process chamber 70. The two absolute pressure sensors 20, 25 and the differential pressure sensor 70 are all connected directly in fluid flow relation to the interior 161 of the chamber 70. The chamber pressure P_C in this example, therefore, is the pressure in the interior 161 of the process chamber 70. The pressure scales 40, 140 and time scale 142 in FIG. 8 are similar to those in FIG. 5.

[0056] The mid-range absolute pressure sensor 20 and the differential pressure sensor 30 are correlated together in substantially the same manner as described above for the FIGS. 1-5 example. The initialization of the correlation factor F with an initial F₀ can be the same, as shown in FIG. 6. All three sensors 20, 25, 30 are read at 48 in FIG. 6 for the two absolute pressure measurements P₂₀, P₂₅ and for the differential pressure measurement P₃₀. When a process, such as that illustrated by the pressure profile 110 in FIG. 8, starts with the chamber pressure P_C equal to atmospheric pressure P_A, when the door 164 is closed, the differential pressure sensor 30 is the only one that outputs accurate and reliable measurements at that pressure level 112 and in the initial portion 114 of the chamber pressure P_C descent. Therefore, the differential pressure ΔP is the same as the differential pressure measurement P₃₀, as shown at 49, and it is output at 50 for whatever functionalities are desired, such as displays of differential pressure ΔP on a differential pressure scale 140 and opening the door at 132 after the process is completed. However, as explained above, adjusting the differential pressure measurement P₃₀ with the correlation factor F at 51 in FIG. 6 also produces a virtual absolute chamber pressure measurement P_V, which is as accurate as the correlation factor F, and which is output as the absolute chamber pressure P_C at 55 for that initial portion 114 down to the cross-over pressure P_X. Then, when the chamber pressure P_C is lowered to a range in which mid-range absolute pressure sensor 20 provides more accurate and reliable pressure measurements P₂₀, for example, in the range 116 below the first cross-over pressure P_X in FIG. 8, the chamber pressure P_C output at 55 in FIG. 6 is equal to the absolute pressure measurements P₂₀. A pressure point 94 in this mid-range 116 can be used to open the throttle valve 66 in FIG. 9 to bolster pumpdown speed, as explained for the FIGS. 1-5 example above. Also, as explained above, when the chamber pressure P_C drops below the correlation pressure threshold P_t, the correlation factor F is updated at 58 in FIG. 6 to compensate for any change in the atmospheric pressure P_A that may have occurred since the previous update of the correlation factor F. The connections, signal communications links, microprocessor 80, and other signal processing and handling components in the example of FIG. 9 can be similar to those described for FIG. 1, as would be obvious to persons skilled in the art, once they understand the principles of this invention, thus need not be described further herein.

[0057] Unlike the example process pressure profile 98 in FIG. 5, however, the example process pressure profile 110 illustrated in FIG. 8 continues below the pressure measuring capability of the mid-range pressure sensor 20. Therefore, from a second cross-over pressure level P_XX or range 118, preferably where the pressure measurements P₂₀, P₂₅ are both still accurate and reliable, and down through a lower portion 120 of the process pressure profile 110 under P_XX to the base pressure 122, the absolute chamber pressure P_C output at 55 in FIG. 6 is equal to the absolute pressure measurement P₂₅ from the low range pressure sensor 25. This second cross-over at P_XX is implemented at 124 and 126 in FIG. 6.

[0058] The base pressure level 122 is commonly used to draw as many impurities out of the process chamber 161 as practically possible before back-filling the process chamber 161 with an inert or an overpressure gas 63 to raise the process chamber pressure P_C to a more mid-level pressure level 126, where the process feed gases 74, 75, 76 are flowed into the process chamber 161 to react and deposit the semiconductor material 77 on the substrate 73. The process pressure level 126 can be above, below, or equal to the second cross-over pressure P_XX, as desired by an operator.

[0059] At the completion of the example semiconductor 77 deposition process in FIGS. 6-9, the feed gasses 74, 75, 76 are turned off, and the back-fill gas 63 or another back-fill gas (not shown) is used to raise the process chamber pressure P_C back up to the atmospheric pressure P_A. Through the mid-range 128 between the second cross-over pressure P_XX and the first cross-over pressure P_X, the absolute pressure of the profile 110 is provided by the absolute pressure measurements P₂₀. Finally, in the pressure range 130 above the first cross-over pressure P_X, the absolute pressure P_C measurements for the process pressure profile 110 are again provided by the virtual absolute pressure measurements P_V, which are calculated by adding the updated correlation factor F to the differential pressure measurement P₃₀ as explained above and shown at 52, 54, 55 of FIG. 6. Finally, when the chamber pressure P_C reaches atmospheric pressure P_A, the differential pressure sensor 30 senses zero differential pressure ΔP, and the ΔP output at 53 can be used to open the door 164 at 132, as explained above.

[0060] As explained above, any pressure level at which an accurate relationship between absolute pressure and differential pressure is known or can be measured or otherwise determined can be used to determine a correlation factor and, with the differential pressure measurements, to extend absolute pressure measurements beyond the accurate and reliable absolute pressure measuring capabilities of an absolute pressure sensor. The examples described above show this invention extending the range of absolute pressure measurements above the accurate and reliable absolute pressure measuring capabilities of an absolute pressure sensor by adding the correlation factor F to the differential pressure measurements P₃₀. However, the principles of this invention also work for determining and using a correlation factor along with differential pressure measurements to extend absolute pressure measurements below the accurate and dependable pressure measuring capabilities of a high range absolute pressure sensor. For example, if an absolute pressure sensor (not shown) is capable of measuring high absolute pressures, such as from 500 to 3,000 torr accurately and reliably, but could not measure absolute pressures below 1,000 torr, a differential pressure sensor that is accurate and reliable from +200% atmosphere, down to −99.9% atmosphere (approximately 1,200 to 1,500 torr down to about −760 to −600 torr, depending on the specific atmospheric pressure at the time) along with an appropriate correlation factor could be used to extend the absolute pressure measuring range below the 1,000 torr low range limit of the absolute pressure sensor, e.g., down to 0.1 torr. The correlation factor could be determined, for example, at atmospheric pressure, where the differential pressure is zero. If desired, the absolute pressure measurements could then be extended down to even lower pressures, e.g., below 1 torr down to 10⁻⁸ torr, by combining a mid-range absolute pressure sensor and a low-range absolute pressure sensor, as explained above.

[0061] Consequently, a differential pressure sensor can, according to this invention, be used to provide virtual absolute pressure measurements P_V within its accurate and reliable differential pressure measuring range, on a common absolute pressure scale with absolute pressure measurements of one or more absolute pressure sensors above and/or below the differential pressure sensor range. This capability is advantageous, even if absolute pressure sensors with accurate and reliable pressure measuring capabilities are available for the same range as the differential pressure sensor in some circumstances. For example, in the FIG. 4 example described above, the absolute pressure sensor 20 could be a micropirani sensor, which, like a number of other absolute pressure sensors, a thermal conductivity type pressure sensor. Pressure readings from thermal conductivity pressure sensors change with different kinds of gasses, i.e., with different molecular contents, at higher pressures, such as over about 1 torr. In other words, for example, a thermal conductivity type absolute pressure sensor will output different pressure readings P₂₀ when the gas in the chamber is changed or mixed with another gas, even if the actual absolute pressure P_C in the chamber does not change. Direct reading differential pressure sensors, such as piezo and capacitance diaphragm gauges, are not gas-type dependent, thus provide the same pressure readings regardless of the kinds of gasses that are introduced into the chamber. Therefore, for gas-independent absolute pressure measurements, use of the differential pressure measurements P₃₀ with a correlation factor F according to this invention has advantages over a thermal conductivity absolute pressure sensor for the same range. Therefore, the absolute pressure sensor not only performs two functions simultaneously according to this invention, i.e., measuring and monitoring both differential and absolute pressures in a range of about 10 torr to 1,500 torr or higher, it can also provide the absolute pressure readings in that range better than at least some absolute pressure sensors that are gas-type dependent.

[0062] The foregoing description is considered as illustrative of the principles of the invention. Furthermore, since numerous modifications and changes will readily occur to those skilled in the art, it is not desired to limit the invention to the exact construction and process shown and described above. Accordingly, resort may be made to all suitable modifications and equivalents that fall within the scope of the invention. For example, while the comparisons in decision boxes 50, 56 in FIGS. 3 and 6 are expressed in terms of “greater than” and “less than”, they could be “greater than or equal to” or “less than or equal to”, respectively, because these pressure levels PX and Pt do not have to be exact for this invention to function properly. Therefore, the terms “greater than” (>) is considered to include “greater than or equal to” (≧), and the term “less than” is considered to include “less than or equal to” (≦), for the explanations or recitations relating to the methods in FIGS. 3 and 6. Also, in both FIGS. 3 and 6, the comparison at 52 can be “Is P_V>P_X?” instead of “Is P₂₀>P_X?”, as indicated above. As mentioned above, example current technology absolute pressure sensors suitable for use in this invention include the thermal conductivity-type sensors, micropirani, conventional convention pirani, hot cathode, cold cathode, ion gauge, etc,. as well as low-range diaphragm sensors such as capacitive, piezo, strain gauge, and the like. Any diaphragm differential pressure sensor and combinations of absolute pressure sensors, as mentioned above, are suitable for use in this invention. Of course, the invention would also work with any future technology absolute or differential pressure sensors. The words “comprise,” “comprises,” “comprising,” “include,” “including,” and “includes” when used in this specification are intended to specify the presence of stated features, integers, components, or steps, but they do not preclude the presence or addition of one more other features, integers, components, steps, or groups thereof.