With the growth of modern technology, improved temperature control systems have also been sought for maintaining a thermal load at a precise temperature under energy intensive conditions. Many such control systems also are required to change the temperature of the thermal load in accordance with process conditions, sometimes with great rapidity. As one illustration, semiconductor manufacturing equipment and processes are often dependent upon temperature control of the wafers or other elements on which various surfaces are being deposited or etched, using techniques which are highly energy intensive. It is thus often necessary to maintain a large semiconductor wafer which serves as the base for formation of thousands of minute complex integrated circuits, under precise temperature control, as the wafer is processed, as under plasma bombardment. By such processes, minute patterns may be selectively deposited or etched in the wafer surface.
Semiconductor manufacture is referenced here merely as one example of one process in which there is a need for precise temperature control under dynamic conditions. Other processes in which there are current or prospective demands for such capabilities will present themselves to those skilled in the art.
In the past, temperature stability in the item being processed has often been achieved by using particular fluids and geometries to define effective heat sinks, for withdrawing or supplying thermal energy from the operating zone as needed, to establish a desired effective temperature level in the item. It has been common, heretofore, to employ a thermal transfer medium which remains typically liquid throughout the entire temperature range used in a process. This medium can maintain adequate thermal transfer capability and at the same time avoid the complexity and unpredictability that would be introduced if a change of phase from liquid to vapor were to be introduced, wholly or partially.
Although the state of the art has been constantly evolving, few distinctly different methods were employed until a novel thermal control technique was introduced by Kenneth W. Cowans et al employing energy transfer using different phases of the same medium. Patents entitled “Thermal Control System and Method” (U.S. Pat. Nos. 7,178,353 and 7,425,835) have issued on this concept and are assigned to the assignee of the present application. This concept employs the thermodynamic properties of a refrigerant in both vapor and liquid phases, properly interrelated to exchange thermal energy with a load so as to maintain the temperature at a selected target level within a wide dynamic range. Consequently, the refrigerant can heat or cool a product and process, such as a semiconductor wafer of large size, at a single or a succession of different target temperatures. This concept has been referred to for convenience by the concise expression “Transfer Direct of Saturated Fluid”, abbreviated TDSF. This descriptor recognizes and in a sense summarizes the operative sequence, in which a medium is first compressed to a high temperature gaseous state, then divided, under control, into two interdependent flows. One flow path maintains the fluid in high pressure gaseous phase, but in this flow path the flow rate and mass are varied in accordance with the target temperature to be maintained. Variation of the one flow affects the differential flow in the other path, in which the refrigerant is converted, by cooling, to liquid phase and the flow is then further cooled by expansion. In this path the flow rate is dependent on the heat load presented to the system. Typically, the flow in this liquefied path is regulated by a standard refrigeration thermo-expansion valve (TXV).
As disclosed in the referenced patents, the two flows, of high pressure gas and cooled expanded fluid/vapor, are recombined in a mixer before delivery to the thermal load. The target temperature for the load is established by adjusting the balance between the two flows by admitting a selected amount of hot gas flow, controlled such that needed pressure, temperature and enthalpy are maintained in a continuous loop.
The TDSF concept has numerous advantages. Some can be best expressed in terms of the range of temperatures that can be encompassed from hot (entirely pressurized gas) to maximum cooling (entirely expanded vapor). The concept also enables the load temperature to be maintained with precision. The target temperature can be adjusted bi-directionally and rapidly.
The use of a refrigerant having a temperature/pressure transition that is somewhere in mid-range relative to the operating temperature band, however, creates possibilities for undesired changes in refrigerant state under certain operating conditions. Situations have been encountered in which performance limitations have been imposed on TDSF systems because of installations which introduce substantial pressure drops or long transport lines for the refrigerant. These conditions can arise because, in a two-phase medium, pressure drops are also accompanied by temperature variations. For example, long line lengths from compressor and condenser units to a semiconductor processing site may be required for operative or geometrical considerations. Heretofore, installations which have inherently required the use of long transport distances for refrigerant media have sometimes imposed restraints on the use of the TDSF concept or the use of special expedients which add undesirable complexity and cost. It is also true that long lines can introduce another complication, that of ‘puddling’: If this occurs, the liquid phase can separate from the two-phase mixture creating variations in mass flow at the line's end. This can adversely alter control characteristics due to surging conditions as pure liquid and pure gaseous phases alternate with mixed two phase flow.
The present invention discloses a novel implementation of the TDSF concept of separating and later recombining a high pressure gas phase of a two-phase refrigerant medium with a cooled, liquefied and then expanded differential flow of the same medium, and application of the medium to the thermal load. In accordance with the invention the principal phase of the refrigerant that is propagated through the thermal load while the load is being heated is the cooled expanded differential flow. The combination of cooled expanded flow through the thermal load with the modulated high pressure gas flow occurs after as well as before the thermal load, so that this approach has been termed “Post Load Mixing” (PLM). The media fed into the thermal load heat exchanger is stabilized in temperature throughout its flow through that exchanger because it is responsive both to the enthalpy of the expanded component and the pressure modulated by the hot gas in the mixing process.
The PLM approach uses the two different phase states of the refrigerant in a uniquely integrated manner. The pressure of the suction line to the compressor is influenced by the mass of refrigerant received, since the compressor is a device that processes a fixed volume per unit of time. In the PLM system the flow through the thermal load has a smaller differential in temperature than would exist with unidirectional transport of fully mixed dual flows, and the thermal load temperature can be thus more tightly controlled. Essentially, the flow through the thermal load is so controlled as to be mainly or completely the cooled expanded component, and in consequence the pressure drop undergone by the refrigerant in passing through the load is lessened. Furthermore, by post load mixing after the refrigerant has passed through the load, the refrigerant passing through the thermal load has a greater percentage of liquid than if all the hot gas had been mixed before the load and thus has a higher heat transfer coefficient, so that thermal exchange is more efficient, particularly at and near the last portions of the heat exchanger passage.
The PLM concept employs some mixing of the two flows both before and after the thermal load, but in a selectable proportionality. This is done in a preferred embodiment by including two impedances in the paths supplying the high pressure hot gas to the mixing tees. Said impedances are settable as to magnitude. A flow of high pressure gas is branched off and combined with the cooled expanded flow at an input mixer coupled to the input to the thermal load. The flow bypassing the thermal load is also directed through a series-coupled solenoid valve which can be controlled so as to enable rapid changes of operating mode between post load mixing and fast heating of the thermal load. Said solenoid valve is closed when rapid heating of the thermal load is desired. This is usually employed when switching the load from one temperature to a hotter temperature, as when a chuck that is normally cold during processing is removed from the system to allow repair to be accomplished. Rapid heating will thus minimize the time needed for such repair and changeover.
The post load mixing approach may be used in certain geometries or applications requiring that the refrigerant be transported over a relatively large distance between the energizing (compressing and condensing) sites and the sites at which thermal exchange occurs. In accordance with the invention, substantial advantages are achieved in these situations by deploying the principal flow adjusting, combining and mixing circuits in a geometrically compact and thermodynamically adapted post load mixing unit, denoted the PLM line box (LB).
The PLM LB is for disposition in proximity to the thermal load and incorporates conduits for high pressure gas flow, liquefied refrigerant low, and return flow, as well as a thermo-expansion valve (TXV), an equalizer for the TXV, and check valves and mixing tees. The configuration, which forestalls mixing before the transport lines, is realized within a volume that is about one cubic foot or less. This unit may be described as comprising a remote control box.
In this combination, the thermo-expansion valve is proximately coupled to a temperature sensing bulb responsive to the temperature in the return line from the load after the mixing tee located downstream from the thermal load. Said thermo-expansion valve is also coupled with a pressure sensing line to the return line in a position proximate said temperature sensing bulb, which coupling serves to establish the external equalizer function. In those installations displaying a minimal pressure drop through the thermal load said thermo-expansion valve can be of the internally equalized type. When such non-equalizing valves are employed said coupling to the return line is not used. The two mixing tees are disposed separately, one before and one after the thermal load. The system may include a check valve before the first mixing tee, and, for flow regulation, a flow orifice is disposed before each mixing tee. A solenoid valve is located in series with the second mixing tee. Consequently, despite the fact that long transport lines may be needed between the phase conversion, energy demanding portions of the system and the thermal load at the process site, needed phase conversions and flow modulations are effected reliably without the danger of accumulation of internal liquids.
In accordance with other features of the invention, where transport lines and conditions present only marginal probability of liquefaction, the transport lines from the proportional valve and the thermo-expansion valve can be disposed to parallel but insulated externally from each other before being coupled to a mixer in the PLM configuration.
A better understanding of the invention may be had by reference to the following description taken in conjunction with the accompanying drawings, in which:
A generalized system utilizing post load mixing (PLM) is shown in
A separate branch from the compressor 12 output 13 is taken from a junction before the condenser 14 to direct pressurized hot gas from the compressor 12 into a second flow path 22. This second flow path 22 includes a proportional valve 24 that is operated by a controller 18 so as to adjust the proportion (in mass flow rate) or hot gas that is to be used out of the compressor 12 output. This adjustment modulates the two flows and ultimately determines the proportion of hot gas to be employed in the consequent mixture of the two flows, as described below. The adjustment consequently sets the target temperature for the thermal load 30.
In the first branch 20 the output from the condenser 14 is applied to a thermo-expansion valve TXV 26, this output being dependent on and determined by the differential temperature between the superheated gas as sensed a proximate by bulb 35 and the temperature of output fluid from the second mixer 32 a point in line 51 adjacent where the bulb 35 is located. The thermo-expansion valve 26 thus senses the pressure difference between liquid contained within bulb 35 and the pressure sensed by a line 48 connected to externally equalized TXV 26. The output flow from the TXV 26 is here coupled to the thermal load 30, which is depicted only generally. Said output flow from the TXV 26 travels through a delta P valve 49 which valve performs the same function as disclosed in U.S. Pat. No. 7,178,353. After passing through valve 49 the expanded cooled output from the TXV 26 mixes with some of the hot gas in the first mixing tee 50. The output 31 from the load 30 is, in accordance with the PLM approach, returned to the input of the compressor 12 via one input of a second mixing tee 32, which also receives, at a separate input, some of the output from the proportional valve 24. The output line from the second mixing tee 32 returns to the compressor 12, but the input pressure of this return flow is sensed on route to the compressor 12 input by the external equalization bulb 35 which is coupled into the TXV 26 via the line 36. This connection also provides the known external equalization feature disclosed in the patents referred to above and in other patents and applications on the TDSF system, so that it need not be described in further detail. In addition, the controller 18 for the proportional valve 24 receives a temperature input from a sensor 38 that is responsive to the temperature level at the thermal load 30. Alternatively, said temperature sensor 38 may be mounted so as to sense any other location that is desired to regulate.
The PLM dual flow, dual mixing system, has other features and advantages. A solenoid valve, labeled SXV 54 is in the path from the proportional valve 24 to the second mixer 32. The SXV 54 is controlled by the controller 18, so it can be shut off whenever the system is programmed to make a change in the target temperature from one level to a higher level. Shutting off this path at the SXV 54 assures that all hot gases flow to the input of the first mixer 50, and more rapidly increase the temperature of the flow into the thermal load 30. In the input to the SXV 54, a settable impedance, shown symbolically, constituting a controllable orifice 78 is included, in parallel to a comparable settable impedance or controllable orifice 79 in the direct path to the first mixer 50. By the use of these control orifices 78 and 79, the two separate flows of pressurized gas fed into the first mixer 50 and second mixer 32 can be proportioned and balanced as desired. The system also includes, as shown, a heater 117 in the input to the compressor 12, which heater 117 may be activated by the controller 18 to convert a liquid containing mixture returning from the second mixer 32 to the wholly gaseous phase for proper operation of the compressor 12.
Mixing the hot gas from the proportional valve 24 with the cooled expanded flow from the TXV 26 after the thermal load 30 retains the essential benefits of the TDSF system, but offers particular added benefits. These are particularly applicable where substantial pressure drops or differentials in heat transfer coefficients may be encountered or exist within thermal load 30. The mass flow from the proportional valve 24, when combined with the system flow at the second mixing tee 32 and also with the TXV 26 output to the first mixing tee 50, modulates the pressure within the load 30. This variation affects the temperature within the circuit and thereby controls the temperature of the load. With PLM, the temperature level across a thermal load, such as a semiconductor chuck can be contained within tolerances that are more precise than previously expected. Tests of a practical system show a reduction in temperature differential to 3° C. from a prior 10° C. differential.
The media fed into the thermal load 30 is stabilized in temperature throughout its flow path in the heat exchanger therein because of the total pressure of the refrigerant fluid, which pressure is controlled by the proportion of hot gas propagated into the circuit. The pressure of the refrigerant in the suction line to the compressor 12 is influenced by the mass passed into the compressor, which compressor 12 processes a fixed volume per unit of time. Because of these interrelated factors, the thermal load 30 is more tightly temperature controlled than in non-PLM based systems. In the system shown, the flow through the thermal load 30 is generally restricted so as to be completely or almost completely that refrigerant that flows through the thermo-expansion valve 26. By so limiting the flow, the pressure drop undergone by the refrigerant passing through the load is lessened. Also, since the hot gas is mixed at the second mixer 32 with the two-phase output of the TXV 26 after the output has passed through the load 30 there is a greater percentage of liquid in the mix at this point. Thus the heat transfer coefficient is maintained high throughout the thermal load 30. Therefore, adjustments in the two flows can also be made after sensing the thermal load temperature, in order to anticipate temperature differentials.
Reference should now be made to the Mollier diagram of
This conclusion is exemplified by factual results achieved in the use of the PLM concept in controlling the temperature of an electrostatic chuck used in semiconductor processing. In prior systems, temperature control units have used a liquid mix of thermal exchange fluid, and provided temperature differentials of the fluid through the chuck typically averaging 10° C. (±5° C.). Using post load mixing, however, the temperature differential through the entire area of the chuck was reduced to no more than about ±3° C.
In order efficiently to utilize the thermal and fluid pressure energy in the lines 63 and 64 in propagating fluids to and from the physically well separated TDSF system 10, the operative elements for mixing and control are principally located relatively remotely in what is here called a “PLM Line Box” 70, as shown in both
Also consistent with the arrangement of
The arrangement of elements inside the PLM Remote Box 70 is shown three dimensionally in
Incorporating the operative control elements for unification and mixing of the two flows of refrigerant in the very small volume illustrated in
As a qualitatively limited alternative, when substantial line lengths might introduce problems with liquid puddling within transport lines, unstable temperature changes due to puddling can be limited or avoided using the insulation technique depicted in
Although there have been described above and illustrated in the drawings various forms and expedients for post load mixing, the invention is not limited thereto but incorporates all features and alternatives within the coverage of the appended claims.
This application relies for priority on similarly titled applications of the same inventors, namely a first provisional application filed Sep. 23, 2008, Ser. No. 61/192,881 and a second provisional application filed May 20, 2009, Ser. No. 61/179,745.
Number | Date | Country | |
---|---|---|---|
61192881 | Sep 2008 | US | |
61179745 | May 2009 | US |