Patent Application
20060133964
The present invention relates generally to chemical analysis, and more particularly to apparatus for the removal of an interfering matrix prior to chemical analysis.
Automated systems for measuring the concentration of analytes in a sample have been developed using a number of analytical techniques such as mass spectrometry. For example, co-assigned U.S. application Ser. No. 10/094,394, entitled “Automated In-Process Ratio Mass Spectrometry For Characterizing Constituents,” filed Mar. 8, 2002, the contents of which are hereby incorporated by reference in their entirety, discloses an automated in-process mass spectrometry (IPMS) apparatus for detecting, identifying, and quantifying chemical constituents and their reaction products in process solutions.
One type of process solution which an IPMS apparatus may analyze is a copper electroplating bath for the deposition of copper structures on semiconductor wafers. The bath comprises a relatively concentrated acidic aqueous copper sulfate solution. Plating topology is controlled by organic plating solution additives within the copper sulfate solution that function to either suppress or accelerate the plating process. These additives experience electrochemical breakdown during the plating process and can be lost by drag out or by becoming trapped within the plated film. However, the achievement of void-free plating in the vias and trenches of sub-micron high-aspect-ratio structures requires very tight control of additive levels. Unlike indirect measurement methods such as cyclic voltametric stripping (CVS) that monitor the effectiveness of the plating solution, the IPMS apparatus discussed above allows a user to directly measure trace component concentration as well as constituent concentration (including breakdown products) in the electroplating bath to ensure a defect-free deposition process.
High sensitivity quantification of the organic additives and their breakdown by-products in the electroplating bath is hampered by the relatively high concentration of a matrix of sulfuric acid and copper sulfate within the bath. These relatively high concentrations of protons, sulfate, and copper ions obscure the detection and quantification of constituents such as organic additives because ionization of the higher concentration ions is statistically more likely in the ionization source of the mass spectrometer. Thus, the matrix of copper sulfate and sulfuric acid should be removed from the sample to more accurately quantify the organic additive concentrations. Similarly, other metrology techniques such as flow injection analysis and chromatography often require the removal of chemical matrices that may hamper the quantification of an analyte of interest. U.S. application Ser. No. 10/641,946 discloses an automated matrix removal module in which a reagent is mixed with a sample having an interfering matrix so that the matrix may be removed prior to quantification of the analyte(s) of interest. The reagent reacts with the matrix to form a precipitate, which is then removed using a filter.
The automated matrix removal module disclosed in U.S. application Ser. No. 10/641,946 advantageously includes a back-flush cycle to cleanse the filter such that the module may remove matrix from samples continually for periods of months. However, addition of an appropriate amount of reagent depends upon the concentration of the interfering matrix. For example, in some instances it may be desirable to under-precipitate the matrix such that some matrix remains in the filtered sample. Alternatively, it may be desirable to “over-precipitate” the matrix such that some un-reacted reagent remains in the filtered sample. Attaining the desired degree of precipitation can be difficult, however, because the concentration of the matrix may change. For example, with regard to a copper plating bath, both copper sulfate and sulfuric acid can be consumed during a plating operation as well as being physically removed from the bath due to wafer removal. In addition, evaporation can occur, thereby changing copper sulfate and sulfuric acid concentrations. Moreover, replenishment of additives in the bath can also cause the matrix concentration to change if corresponding amounts of sulfuric acid and copper sulfate are not added simultaneously.
The potentially-dynamic nature of a matrix concentration thus makes matrix removal problematic. For example, a matrix concentration may be assumed to be static such that a fixed amount of reagent is always added to the sample within the automated matrix removal module. In such a case, the degree of precipitation will increase if the matrix concentration decreases such as described earlier with regard to a copper plating bath. Over time, an undesired over-precipitation may occur. Alternatively, should the matrix concentration gradually increase, the degree of precipitation would gradually decrease until an undesired under-precipitation occurs.
Accordingly, there is a need in the art for automated systems for the removal of interfering matrices prior to a chemical analysis which dynamically respond to changes in concentration of the interfering matrices.
In accordance with the present invention, an apparatus for the automated removal of a matrix from a solution containing an analyte of interest is provided. The apparatus includes: at least one analytical instrument operable to measure a concentration of the matrix in a sample of the solution; a source of reagent, the reagent being reactive with the matrix to form a precipitate; a reaction vessel; and a filter, wherein the apparatus has a filtering configuration in which a volume of the solution and a volume of the reagent react in the reaction vessel to form a reaction mixture that is filtered through the filter, the volume of the reagent being based upon the matrix concentration measurement, the apparatus having a flushing configuration in which the filter is back flushed with a solvent.
In accordance with another aspect of the invention, an apparatus is provided that includes: a sample extraction module operable to extract sample from a selected one of a plurality of process solution baths; at least one dilution and spiking module operable to spike and dilute extracted sample to form a processed solution; a source of reagent solution reactive with a matrix in the processed solution to form a precipitate; a plurality of matrix removal modules, each matrix removal module operable in a mixing cycle to mix the processed solution with the reagent solution to form a reaction mixture having the precipitate and to filter the precipitate from the reaction mixture through a filter to form a filtered solution, each matrix removal module operable in a flushing cycle to flush the filter; and
In accordance with another aspect of the invention, an apparatus for the automated removal of a matrix from a solution containing an analyte of interest is provided. The apparatus includes: at least one analytical instrument operable to periodically measure a concentration of the matrix; a source of reagent, the reagent being reactive with the matrix to neutralize the matrix; and a reaction vessel, wherein the apparatus has a neutralizing configuration in which a volume of the solution and a volume of the reagent react in the reaction to form a neutralized mixture, the volume of the reagent being based upon a most-recent matrix concentration measurement by the analytical instrument.
Use of the same reference symbols in different figures indicates similar or identical items.
Reference will now be made in detail to one or more embodiments of the invention. While the invention will be described with respect to these embodiments, it should be understood that the invention is not limited to any particular embodiment. On the contrary, the invention includes alternatives, modifications, and equivalents as may come within the spirit and scope of the appended claims. Furthermore, in the following description, numerous specific details are set forth to provide a thorough understanding of the invention. The invention may be practiced without some or all of these specific details. In other instances, well-known structures and principles of operation have not been described in detail to avoid obscuring the invention.
Regardless of the type of chemical metrology instrument that will be used to characterize the analyte of interest in the sample solution, the performance of this instrument may be hampered by the presence of an interfering chemical matrix. As used herein, “matrix” will be understood to denote constituent(s) within the sample solution that hamper analysis of the analyte of interest by the chemical analysis or metrology tool. For example, as discussed earlier, copper sulfate and sulfuric acid act as chemical interferents in the characterization of organic additives within a copper electroplating solution by a mass spectrometer. This interference results from the hydrogen, copper, and sulfate ions being preferentially ionized within the mass spectrometer due to their relatively-high concentration, thereby obscuring the measurement of the organic additive concentrations. The acid and copper sulfate interferents may also be denoted as a “matrix” which must be removed before characterization of the organic additives. In general, a matrix may comprise a plurality of chemical species.
To begin the process of removing the matrix from the sample solution, three-way valve EV1 is actuated so that as the plunger of syringe pump 105 is withdrawn, a portion of the sample carried by conduit 110 is drawn into syringe pump 105. Syringe pump 105 may be controlled by a stepper motor (not illustrated) to precisely control the amount of sample withdrawn into syringe pump 105. A reagent solution may then be added to the contents of syringe pump 105 as follows. The appropriate reagent solution is provided by a reagent source 120 connected to three-way valve EV3. The reagent is selected such that it will react with the matrix such that the matrix is no longer interfering. For example, the reagent-may react with the matrix to produce an insoluble precipitant. In one embodiment, if the matrix is aqueous copper sulfate and sulfuric acid, a suitable reagent solution would be aqueous barium hydroxide. With respect to the removal of copper sulfate and sulfuric acid from an aqueous solution using barium hydroxide, the reactions occur as follows:
H2SO4+CuSO4+2Ba(OH)2→2H2O+2BaSO4+Cu(OH)2
Both BaSO4 and Cu(OH)2 are relatively insoluble in aqueous solutions and would thus precipitate within the syringe pump. As another example, if the matrix comprises Ag+ ions in an aqueous sample, a suitable reagent may comprise a source of Cl− anions such as NaCl.
To mix the contents of syringe pump 105 with the reagent solution, three-way valves EV1 and EV3 are appropriately actuated to connect syringe pump 105 to reagent source 120. The plunger of syringe pump 105 may then be withdrawn to add the appropriate amount of reagent to the contents of syringe pump 105. The appropriate amount of the reagent depends upon the amount of matrix within the sample solution withdrawn into syringe pump 105. For example, each molecule of sulfuric acid and copper sulfate in a copper plating bath sample requires the addition of a corresponding molecule of barium hydroxide for a perfect removal and neutralization of the matrix. However, as discussed earlier, the analysis following matrix elimination may be better performed if the matrix is “under-precipitated” such that some matrix remains despite the addition of the reagent. Alternatively, the analysis following matrix elimination may be more accurate if the matrix is “over-precipitated” such that the matrix is eliminated yet un-reacted reagent remains.
To provide the appropriate amount of reagent, the concentration of the matrix may be assumed to be static such that a fixed volume of reagent solution may be added based upon the volume of sample solution needing matrix elimination. Alternatively, the matrix concentration may be measured periodically such that the volume of reagent solution being added depends upon the latest measured matrix concentration. For example, it has been shown that the copper sulfate concentration and sulfuric acid concentration in copper electroplating bath solution has appreciable variability in semiconductor manufacturing operations. Thus, an addition of a fixed volume of Ba(OH)2 solution to samples of copper electroplating solutions may result in either under-precipitation or over-precipitation of the copper sulfate/sulfuric acid concentrations. In general, an addition of a variable volume of reagent solution based upon periodic measurements of the concentrations of the matrix avoids such over-precipitation or under-precipitation problems. Such embodiments may be denoted as “closed loop” embodiments in that feedback information (the latest measurement of the matrix concentration to be eliminated) is used to update the amount of reagent being added. In contrast, embodiments in which the amount of reagent being added is static may be denoted as “open loop” embodiments in that no feedback information is utilized.
Regardless of whether a fixed volume or variable volume of reagent solution is used, the flow of reagent solution into syringe pump 105 may be aided by pressuring reagent solution source 120 with an inert gas such as nitrogen. Alternatively, a mechanical pump may be used to assist the flow of reagent solution into three-way valve EV3.
After the addition of reagent solution into syringe pump 105, additional solvent may be optionally added to flush the three-way valves of reagent solution. For example, in the case of removing a copper sulfate and sulfuric acid matrix as discussed above, the additional solvent may be ultra pure water (UPW). In the exemplary embodiment for module 100 shown in
Although no cooling or heating of syringe pump 105 is necessary, the mixture of sample solution and reagent solution within syringe pump 105 benefits from physical mixing to assist the desired reaction between the two solutions. For example, syringe pump 105 may include a side port 149 that is exposed after the plunger is withdrawn. In general, this exposure will occur after the mixture of sample solution and reagent solution has been added to syringe pump 105. Valve EV2 to a drain port 170 is opened and the plunger for syringe 105 is withdrawn until the valves EV1, EV3, and EV5 are cleared of solution. To ensure a homogeneous mixture, valve EV2 to drain port 170 can be left open while the plunger for syringe 105 is withdrawn until side port 149 is connected. Once side port 149 is fluidly connected to drain 170, an inert gas such as N2 may flow through a conduit 151 and a three-way valve 152 and bubble through the contents of syringe pump 105 and out port 170 for a period of time. Advantageously, such bubbling mixes the contents of syringe pump 105 without requiring any physical movement or agitation of syringe pump 105. In this fashion, stress-induced fractures or strain of the various conduits attaching to syringe pump 105 from physical agitation are avoided as well as the cost of an agitating component.
After adequate mixture of the contents of syringe pump 105, the completion of desired precipitation reactions within syringe pump 105 may be assumed. At this point, syringe pump 105 may pump its contents through a filter 140. However, to avoid clogging filter 140 with the full amount of the resulting precipitate, the contents of syringe pump 105 may be allowed to settle. For example, syringe pump 105 may be oriented such that gravity pulls the precipitate towards side port 149. If the plunger of syringe pump 105 is then just partially actuated such that the bulk of the precipitate remains in syringe pump 105, the life of filter 140 is extended. Alternatively, all the contents of syringe pump 105 may be filtered. Before syringe pump 105 may pump its contents through filter 140, three-way valves EV1, EV3, EV5, EV2, and EV4 are appropriately actuated to connect syringe pump 105 to filter 140. The pore size for filter 140 is chosen appropriately as determined by the flow rate, solvent, and expected type of precipitate that will be filtered by filter 140. For example, to filter the Cu(OH)2 and BaSO4 precipitants discussed previously, a suitable pore size for filter 140 is approximately 0.45 um. By proper actuation of three-way valve EV6, the filtered contents pass through filter 140 into conduit 145 for eventual processing by a chemical analysis or metrology instrument (not illustrated).
Although filter 140 will have thus removed any solid precipitate in the filtered solution provided to the chemical metrology instrument, unfiltered solution and precipitate will now be contaminating module 100. Thus, module 100 may be flushed as follows before another cycle of receiving a sample, mixing the sample with a reagent, and filtering the mixed solution may begin. To begin the flush cycle, the plunger of syringe pump 105 is withdrawn a sufficient amount to expose back flush port 149. With back flush port 149 exposed, three-way valves EV5, EV3, EV1, and 152 are appropriately actuated so that solvent source 130 is connected through syringe pump 105 to a drain 155. Because solvent source 130 may be pressurized, solvent will then flush from solvent source 130 through back flush port 149 into drain 155. Alternatively, a mechanical pump may be used to force solvent into three-way valve EV5 and eventually to drain 155.
While syringe pump 105 is flushed, filter 140 may also be back-flushed. To back-flush filter 140, three-way valves EV6, EV4, and a three-way valve 162 are appropriately actuated to connect a solvent source 160 or a solvent source 161 to a drain 165. An inert gas such as nitrogen may be used to pressurize solvent source 160 such that solvent will then flow through three-way valve EV6, and through filter 140 and three-way valve EV4 into drain 165, thereby back-flushing filter 140 of the precipitate from the previous filtering cycle. Solvent sources 161 and 162 may comprise different solvents such as ultra pure water or dilute nitric acid. The life of filter 140 may be substantially extended in this fashion.
An additional back flush of filter 140 may be performed by appropriately actuating three-way valves EV1, EV3, EV5, EV2, EV4, EV4, EV6, and 162 to connect the remaining one of solvent source 160 or 161 to syringe pump 105. Syringe pump 105 is withdrawn to the side port 149 whereby three-way valve 152 is appropriately actuated to connect the backflush port 149 to drain 155. Solvents 160 or 161 are then allowed to flow to drain 155 for a predetermined amount of time. After the predetermined flushing time, the three-way valves are actuated to their default, normally open state.
To complete the flushing cycle, three-way valves EV1, EV3, EV5, and EV2 are appropriately actuated to connect syringe pump 105 to a drain 170. The plunger of syringe pump 105 is then depressed. Because syringe pump 105 will have been filled with clean solvent after sufficient flushing through back flush port 149, clean solvent will then flush through three-way valve EV2 into drain 170, thereby flushing three-way valve EV2. It will be appreciated that the presence of drain 170 is a result of the use of three-way valves—three-way valve EV4 cannot be actuated so as to connect drain 165 to syringe pump 105. Thus, after flushing syringe pump 105, its contents must be emptied into another drain such as drain 170. In alternative embodiments that do not use three-way valves, this extra drain would be unnecessary. For example, one or more two-position multi-way valves such as rotary valves could be employed to alternatively connect syringe pump 105 to the sample source, to the reagent source 120, to filter 140, and finally to a drain.
Consider the advantages provided by module 100. Because the various components may all be actuated according to commands from a microprocessor or state machine, the operation is entirely automated and requires no human intervention. Moreover, because of the flush cycle, filter 140 may be reused for many cycles, thereby keeping operation costs low. However, the time required to (if desired) mix the contents of syringe pump 105 to ensure completion of desired precipitation reactions, allow the contents to settle, and flush syringe pump 105 may be problematic.
Such delay is avoided in an alternative automated matrix removal module 200 as illustrated in
It will be appreciated that numerous pumping configurations may be used to effect the desired flow rates into mixing tee 205. For example, a first syringe pump 220 may withdraw a desired volume of the sample solution from a sample source 225 through appropriate actuation of a three-way valve 230. After three-way valve 230 is appropriately actuated so that syringe pump 220 is connected to mixing tee 205, syringe pump 220 may force its contents into mixing tee 205. Similarly, a syringe pump 240 may withdraw a desired volume of reagent solution from a reagent solution source 245 through appropriate actuation of a three-way valve 250. After three-way valve 250 is appropriately actuated so that syringe pump 240 is connected to mixing tee 205, syringe pump 240 may pump its contents into mixing tee 205 simultaneously with the flow of sample solution pumped from syringe pump 220. The resulting mixture from mixing tee 205 flows through conduit 210 into filter 140 as forced by the actuation of syringe pumps 220 and 240. Filter 140 may then be back-flushed with solvent from sources 160 and 161 through appropriate actuation of three-way valves EV4, EV6, and 162 as discussed with regard to
Multiple automated matrix elimination modules such as module 100 or 200 may be employed in a chemical analysis or metrology system. For example, automated matrix elimination modules may be advantageously incorporated into an IPMS system used to analyze the concentrations of constituents in copper electroplating baths. Turning now to
Regardless of how many spikes are added, the multi-way valve is then actuated such that the loops are connected with a diluent source (such as a syringe pump filled with diluent) which may then mix the contents of the connected loops with the diluent. Should additional dilution be required, the diluted and spiked sample may then be processed in additional dual-loop multi-way valves. In these additional valves, one of the loops is filled with diluent rather than spike, since spiking has already taken place. The remaining loop is filled with diluted and spiked sample from the previous multi-way valve. It will be appreciated that a single spike and dilution module may be used rather than multiple modules 360 and 365. However, by using two or more of such modules, a diluted and spiked sample may be delivered by one of the modules while another is still performing its dilution and spiking operations, thereby enhancing throughput.
Through appropriate actuation of either selection valve 330 or selection valve 370, processed sample from modules 360 and 365 may have its matrix eliminated in a selected one of a plurality of matrix elimination modules 340. Each of the matrix elimination modules 340 may be implemented such as described with regard to modules 100 and 200. It will be appreciated that just a single matrix elimination module 340 could be used in IPMS system 300. However, by providing a plurality of modules 340, each module may be specialized to the analysis of a given analyte. For example, a first module 340a may be dedicated to the analysis of SPS. Such analyses are best performed such that the matrix is under-precipitated, leaving the treated solution slightly acidic. In contrast, the analysis of polyethylene glycol (PEG) is best performed such that the matrix is over-precipitated, leaving the treated solution slightly basic. Thus, a second module 340b may be dedicated to the analysis of polyethylene glycol (PEG), and so on. In addition, the provision of a plurality of modules 340 allows for pipelining such that as one module is performing a back flush or otherwise getting prepared for another cycle of matrix elimination, another module is actively eliminating its matrix, and so on.
The following discussion will assume that IPMS system is analyzing baths 310 of a copper plating solution as an exemplary embodiment. As discussed earlier, an open loop addition of reagent solution such as a Ba(OH)2 solution by modules 340 may be problematic in that matrix concentration (such as copper sulfate and sulfuric acid in the copper electroplating solutions) is often variable and thus dependent upon sampling time. To allow for closed loop control of the addition of reagent solution, IPMS system 300 includes one or more analytical instruments. For example, IPMS system may monitor the copper sulfate and sulfuric acid concentrations in baths 310 using an optical spectroscopy module 350 and a pH ion selective electrode 355. In an alternative embodiment, the mass spectrometer in IPMS system 300 may itself serve as the analytical instrument that monitors the matrix concentration. The location of the analytical instruments may thus be “in-line” in that they are located in a sample path connecting to the mass spectrometer. Alternatively, they may be located “off line” in a sample path that does not connect to the mass spectrometer. In the embodiment illustrated in
As known in the optical absorption arts, optical spectroscopy module 350 may include a liquid flow cell coupled to an optical spectrometer. A light source shines light through the liquid flow cell into the optical spectrometer so that a measurement of the optical transmittance through the diluted sample may be conducted. Periodically, optical spectroscopy module 350 may be recalibrated by determining the absorption of a blank solution (such as UPW) and a control solution of copper sulfate having a known concentration.
The optical transmittance of the diluted sample is related to measurements by the optical spectrometer. For example, a UV/VIS spectrometer may read units of “counts,” where each count is equivalent to a number of photons (for example, 100 photons) impacting a CCD detector within the spectrometer. Thus, a blank count would correspond to counts measured by the UV/VIS spectrometer when the blank solution (such as UPW) is flowing through the illuminated liquid flow cell. In contrast, a dark count would correspond to counts measured when no light is transmitted into the liquid flow cell. A sample count would correspond to the counts measured by the UV/VIS spectrometer when sample is flowing through the illuminated liquid flow cell. Given these definitions, the transmittance of a sample corresponds to the ratio: (sample count-(blank count-dark count)/(blank count-dark count).
Given the transmittance, the concentration of copper (and thus copper sulfate) may be determined through application of Beer's law:
A=εbc
where A is absorbance, E is emissivity coefficient, b is the optical path length in the liquid flow cell, and c is concentration. In turn, the absorbance is related to the transmittance as:
A=−log(transmittance)
Standard techniques to achieve optical instrument stability can be used. For example light from a source can be split with one beam passing through the flow cell to a detector to provide a signal. A second beam is directed at a detector to provide a reference signal thereby allowing corrections to be made for varying light source intensity. In one example, a preferred range of copper concentration for optimal performance of optical spectroscopy module 350 is between 0.5 g/L and 5 g/L of copper. If the concentration of copper in the copper electroplating bath solutions is approximately 40 g/L, sample dilution module 320 should dilute the samples for optical analysis by a factor of approximately 20:1.
Ion selective electrode 355 may be located inline or offline as discussed with regard to optical spectroscopy module 350. To receive a sample at ion selective electrode 355, selection valve 330 and a three-way valve EV16 are actuated accordingly. It will be appreciated that the mass spectrometer itself may comprise the analytical instrument that measures the matrix concentration. In general, however, it will not be the mass spectrometer (or whatever type of chemical metrology instrument being implemented) if the matrix is problematic to that instrument.
As discussed analogously in U.S. application Ser. No. 10/094,394, IPMS system 300 includes one or more processors (not illustrated) that control the automated analysis of analytes of interest in the sampled solutions by the remaining components in IPMS system 300. Thus, a processor may be configured to determine the amount of Ba(OH)2 added within modules 340 from measurements by optical spectroscopy module 350 and the analyte pH using ion selective electrode 355 as follows. From the copper concentration of a sampled bath solution (in grams/L) determined by the optical absorption measurements in optical spectroscopy module 350 (with consideration of the dilution as well), the moles/liter of Cu in the undiluted sample is given by (concentration in grams/L)/(molecular weight of Cu). Similarly, the measured pH from ion selective electrode 355 may be used to determine the concentration of sulfuric acid in grams/L in the undiluted sample. In turn, the concentration of sulfuric acid may then be divided by the molecular weight of sulfuric acid to obtain the concentration of sulfuric acid in moles/L. The concentration of sulfate in the sampled bath (in moles/L) is thus given by the sum of the ((moles of Cu)/L) and the ((moles of sulfuric acid)/L) concentrations.
Given the concentration of sulfate in the sampled bath, the concentration of sulfate in the diluted bath sample from sample dilution module 320 will depend upon the dilution ratio being implemented. For each mole in the diluted sample, a corresponding mole of Ba(OH)2 may be added in matrix elimination modules 340. However, a factor may be either subtracted from the calculated moles of Ba(OH)2 or added to precisely obtain the desired pH level. For example, referring again to
The above-described embodiments of the present invention are merely meant to be illustrative and not limiting. For example, there may be embodiments in which a matrix need not be precipitated but merely neutralized. For example, if the matrix comprises sulfuric acid but the metrology instrument is only affected by the acidity, the sulfuric acid may be neutralized by an appropriate addition of a reagent such as NaOH. In such an embodiment, the filter would be unnecessary. Moreover, dilution module 100 of
This application is a continuation-in-part of U.S. application Ser. No. 10/641,946, filed Aug. 15, 2003, the contents of which are hereby incorporated by reference in their entirety.
| Number | Date | Country | |
|---|---|---|---|
| Parent | 10641946 | Aug 2003 | US |
| Child | 11254030 | Oct 2005 | US |
