Current Monitoring for Plating

The present disclosure relates in general to electroplating and in particular to techniques for monitoring and controlling an electroplating operation.

Electroplating is used in a variety of industries for depositing a layer of metal on a substrate. In various approaches, a substrate and an electrode can be immersed in an electrolyte solution. An electric voltage (electric potential) can be applied between the substrate and the electrode. Metal ions in the electrolyte solution are driven by the voltage toward the substrate, where they are reduced and deposited onto the surface of the substrate.

Electroplating can be used in the fabrication of utensils, jewelry, decorative metal, coinage, machine parts, and other products. Electroplating is also used in various semiconductor manufacturing processes, such as for depositing a layer of metal onto a silicon substrate. The substrate can be a wafer or other component to be used in an integrated circuit or solar cell, for example.

FIG. 1 illustrates one implementation of an electroplating system 100. In this example, three substrates 140a, 140b, 140c (collectively, substrates 140) are held within a plating vessel 120, immersed or partly immersed in an electrolyte solution 122. An electrode 130 is also in contact with solution 122. An electric current is passed through electrode 130, solution 122, and substrates 140. With a suitable choice of electrolyte solution, substrate material, and electrode material, the current causes a metal film to be deposited on substrates 140.

In the example of FIG. 1, an electric voltage is supplied by a power source 180. In this example, power source 180 is a direct current (DC) supply. The positive terminal of power source 180 is coupled via a lead 184 to electrode 130, which serves as an anode in this example. The negative terminal of power source 180 is coupled via a lead 182 to a bus bar 105. Bus bar 105 serves as an electrical power rail and as a mechanical support for substrates 140. In this illustrative example, substrates 140a, 140b, 140c are connected mechanically and electrically to bus bar 105 though holders 145a, 145b, 145c respectively (collectively, holders 145).

In this example, substrates 140 are cathodes in an electrochemical reaction. Current flows from power source 180 to electrode 130, through solution 122, into substrates 140, through holders 145, through bus bar 105, and back to power source 180. In passing from solution 122 into substrates 140, the current causes metal ions to be reduced from aqueous phase into solid-phase atoms. The neutral atoms are deposited onto substrates 140. In various implementations, the source of the metal ions is the material of electrode 130, and the electrode is gradually consumed as the electroplating progresses. In other implementations, the source of the metal ions is the electrolyte solution 122, which may need to be replenished after some time to continue an effective electroplating operation. In various examples, plating vessel 120 is large enough to accommodate large numbers of wafers, e.g., 10, 50, 160, 250, 400 wafers, for simultaneous processing in an electroplating operation.

A variety of different forms of electroplating are used to obtain a desired layer of material on the substrates. Different approaches can be used depending on the substrate material and shape, the desired type of metal layer (e.g. deposited chromium, copper, tin, nickel, other metallic elements, or alloys), the desired purity of the deposited layer, the desired strength of connection between the layer and the substrate, the operating temperature, and other factors. In various situations, multiple plating processes may be conducted in sequence. For example, a silicon wafer may be plated first with one metal (e.g., 30 microns of copper) and then with a second metal (e.g., 6 microns of tin).

In many applications, the final thickness of the deposited layer is a relevant metric in assessing the success of the electroplating process. Similarly, the uniformity of the layer's thickness over the substrate can be a relevant metric in assessing the process. These factors can be of interest in various applications, such as the plating of metal layers onto semiconductor substrates. In various situations, layer thickness (e.g., ones, tens, hundreds, or thousands of microns), layer uniformity (e.g., 1%, 2%, 5%, 8%, 10%, 15%, 20%, 30%, 40%, 60% variations) and other factors can be controlled by monitoring the cleanliness of the substrate, the geometry of the substrate, the purity of the electrolyte solution, the geometry of the current path through the solution between the electrode and the substrates, the total time of the plating process, and other factors.

One factor that may be relevant to the quality of a plating process is the magnitude of the current that is transmitted through a substrate. Increases in the current level can increase the deposition rate, and thus the overall thickness of the deposited metal layer. This current level can be controlled by selecting an appropriate electric voltage to be applied between the electrode and the substrate. The current level can be monitored by a current sensor, such as an in-line ammeter. Alternatively, an electromagnetic sensor can be used to monitor the current without introducing an in-line circuit element.

In FIG. 1, a hand-held clamp meter 190 is used to measure the current that flows though substrate 140c. Clamp meter 190 has two prongs that can be moved to engage or disengage with each other around a conductor. When engaged around a conductor, the prongs form a sensor loop around the conductor. Depending on the type of clamp meter, the sensor loop employs a transformer winding, magnetic vanes, Hall sensors, or other techniques to measure the enclosed current. In the example of FIG. 1, clamp meter 190 is engaged around a narrow section of holder 145c. As a result, a display on clamp meter 190 shows the instantaneous current flowing though substrate 140c.

A system operator may move clamp meter 190 to engage around the holders 145a, 145b, in order to see the instantaneous current flowing though the corresponding substrates 140a, 140b. Similarly, the operator may move clamp meter 190 to engage around the holders lead 182 or 184 in order to see the instantaneous total current flowing though all of the substrates 140.

The use of clamp meter 190 involves some manual intervention during the electroplating process. An operator needs to manually engage the clamp meter, hold, observe, record, and disengage the clamp meter for each measurement. The manual effort can be labor-intensive for a long process (e.g., multiple hours, multiple days). Also, the manual operation introduces some risk of mechanical trauma—the operator may inadvertently knock or even dislodge a substrate from bus bar 105. Similarly, clamp meter may inadvertently be dropped into solution 122.

Other sensors can be used instead of a hand-held device. One approach is to affix a current sensor to one or more support structures, such as holders 145, that are used to conduct current into the substrates.

FIGS. 2 and 3 show one implementation of a system 200 with a current sensor 220a, 220b mounted on a conducting element 210. FIG. 2 is a perspective view and FIG. 3 is a side view of the arrangement. The sensor has two components, 220a and 220b (collectively, sensor 220). Each component is mounted on an opposing side of conducting element 210. In this implementation, the components 220a, 220b are each a Hall sensor, and are mounted in a differential measurement configuration. In some other implementations, a single Hall sensor is used for measuring the current.

In this example, sensor 220 is mounted onto conducting element 210. A connector 233 connects components 220a and 220b, enabling comparison between components 220a and 220b. Sensor 220 may be mounted using screw holes 231 that enable the components to be attached to conducting element 210. An electrical connector 240 supplies power to sensor 220, and communicates data measured by sensor 220 to a computer or other monitoring device. In various implementations, electrical connector 240 is a Universal Serial Bus (USB) or Ethernet connector. In other implementations, sensor 220 draws power from an internal battery and electrical connector 240 is replaced by a wireless data link.

FIG. 4 is a perspective view of one implementation of a system 400 with a current sensor 420 mounted on a plating hangar 410. In this example, a semiconductor wafer 440 is attached to a jig 430, which in turn is attached to plating hangar 410.

In this example, hangar 410 is configured to be electrically and mechanically connected to a bus bar (not shown). A top portion 412 of hangar 410 has a ridge 418 that can stably rest on the bus bar. A bottom portion 416 of hangar 410 has features to which various jigs can be connected. Jig 430 is configured to connect with the hangar and with the wafer: a top section 432 of jig 430 has features that attach to hangar 410, and a bottom section 434 of jig 430 has features that attach to wafer 440. In various implementations, the holders 145 from FIG. 1 each include a hangar 410 and a jig 430.

In hangar 410, top portion 412 and bottom portion 416 are connected to each other by a neck portion 414. Current flowing from the bus bar to the wafer passes through neck portion 414 of hangar 410. In various implementations, neck portion 414 is relatively narrow so that the current is throttled, concentrated, and/or collected in neck portion 414. For example, the neck portion may be constructed of steel, aluminum, copper, or other conductive material with a cross section of 0.6 in.×2.6 in. In various other implementations, a neck portion can have other dimensions, such as thicknesses of 0.125 in., 0.1875 in., 0.250 in., 0.3750 in. Sensor 420 is mounted onto one face of neck portion 414. In various implementations, a corresponding sensor (not shown) is mounted in a differential configuration on an opposing face of neck portion 414. In various implementations, a channel or groove can be formed or cut into neck portion 414 to accommodate sensor 420. In various implementations, sensor 420 is mounted in a location where it can measure all of the current passing into a substrate, or in a location where it can measure a substantial fraction of the current passing into the substrate. Additionally, the location of the sensor can be chosen so that it is not subject to jostling or other undesired mechanical contact during regular use and storage of a hangar.

In various situations, wafer 440 is fully immersed into an electrolyte solution during a plating process. (For example, the illustration in FIG. 1 showed wafers 140 being fully immersed.) In these situations, some of the bottom section 434 of jig 430 may also be in contact with the electrolyte solution. Depending on the electroplating chemistry, the exposed section of a jig may be plated or degraded, limiting the reusable lifetime of the jig. In other situations, jig 430 is held above the surface of the electrolyte solution. In these other situations, a top portion 442 of wafer 440 may not be exposed to the solution, while a remainder portion 444 is immersed in the solution.

FIG. 5 is a flowchart showing one implementation of a method for controlling current delivered to an element during an electroplating operation 500. In various implementations, some or all of the acts in this method may employ various components discussed with regard to the systems depicted in FIGS. 1-4 or 6-7. For example, in the illustrated implementation, the method commences in act 510 when an element, such as a semiconductor wafer, is prepared for electroplating. The wafer is then mounted to a holding jig in act 520. The jig can be, in various implementations, a single-use or few-use component for mating a specialized plating element to a generalized electroplating hangar. The jig is affixed to a hangar in act 530, and the hangar is mounted on a bus bar in act 540. The bus bar provides mechanical support and electrical current to the wafer via the hangar and the jig. The wafer is then immersed or partly immersed in an electrolyte solution that is a suitable electroplating bath.

In act 550, an electric voltage is applied between the bus bar and an electrode in the electrolyte solution. The electric voltage causes an electric current to follow a path from the electrolyte solution through the wafer, the jig, and the hangar. In act 560, the current is measured. The measurement can be made, for example, at a constricted portion of the hangar that carries all or most of the current which passes through the wafer. The measurement is transmitted in act 570 to a control unit. The control unit analyzes the measurement in act 580 and determines, at decision block 585, what adjustment (if any) needs to be made to raise or lower the current. If no adjustment is needed, the present voltage level (from act 550) is maintained in act 592 and the procedure returns to act 550. Otherwise, the voltage level is adjusted in act 590 and the procedure returns to act 550.

Acts 550-560-570-580-585-590/592 form a feedback loop that holds the current level at (or near) a desired value. The desired value can be target constant value or can vary according to a target profile (e.g., based on a plating rate, a desired plating thickness, considerations of the geometry of the substrate, etc.). The loop can terminate at any point (act 550, 560, 570, 580, 585, 590, or 592), at a time when the electroplating procedure is complete.

FIG. 6 illustrates a second implementation of an electroplating system 600. In this example, three substrates 640a, 640b, 640c (collectively, substrates 640) are held within a plating vessel 620, immersed or partly immersed in an electrolyte solution 622. An electrode 630 is also in contact with solution 622. An electric current is passed through electrode 630, solution 622, and substrates 640. With a suitable choice of electrolyte solution, substrate material, and electrode material, the current causes a metal film to be deposited on substrates 640.

In this example, power source 680 is a signal-controllable DC supply. The positive terminal of power source 680 is coupled via a lead 684 to electrode 630. The negative terminal of power source 680 is coupled via a lead 682 to a bus bar 605. Bus bar 605 serves as an electrical power rail and as a mechanical support for substrates 640. In this example, substrates 640a, 640b, 640c are connected mechanically and electrically to bus bar 605 though holders 645a, 645b, 645c respectively (collectively, holders 645). Current flows from power source 680 to electrode 630, through solution 622, into substrates 640, through holders 645, through bus bar 605, and back to power source 680. In passing from solution 622 into substrates 640, the current causes metal to be plated onto substrates 640.

Power source 680 is controlled by a control unit 690. Control unit 690 includes an input 692, an output 698, a processor 694, and a memory 696. Input 692 receives one or more signals that provide information about current flowing into one or more of the substrates 640. In the depicted example, input 692 is a wireless receiver and is shown as receiving a wireless signal 691 that represents a current measurement through holder 645a. This current measurement is generated by a sensor 612 mounted on holder 645a. Input 692 is shown as also receiving a wireless signal 697 that represents a second current measurement, through holder 645c. This second current measurement is generated by a second sensor 614 mounted on holder 645c. The measurement and transmission operations (whether wireless or wired, analog or digital) can be adapted in various implementations to tolerate electrically noisy environments, such as those that may be present in the vicinity of a high-current electroplating procedure.

In various implementations, signals 691 and 697 carry information based on Hall measurements from sensors 612 and 614. In various implementations, these measurements are indicated as total currents (detected amperes); in other implementations, these measurements are indicated in other physical units (e.g., detected amperes per cm̂2); in yet other implementations, these measurements are indicated in other scalar units (e.g., a unitless 8-bit or 20-bit number that is a linear or nonlinear representation of detected current). These indications are computed, in some implementations, based on device calibrations stored internally in sensors 612 and 614. For example, sensor 612 may include an analog detector that uses a generates an output signal in the range of −4000 mV to +4000 mV, with the output being proportional to a detected current and a conversion constant of 4 to 4.5 mV output per Ampere of detected current.

Processor 694 uses the received current measurements to determine whether to change the voltage provided between bus bar 605 and electrode 630. Based on this determination, output 698 provides an output signal that increases, decreases, or maintains the electric voltage generated by power source 680. In various implementations, processor 694 may be configured to execute instructions stored in memory 696 to regulate power source 680. By executing the instructions, processor 694 may operate, for example, to maintain a predetermined DC current level supplied to wafer 640a. This operation would be based on wireless signal 691 received by input 692. In other examples, processor 694 and the software it executes may be configured to regulate power source 680 based on multiple current measurements (e.g. as received in wireless signals 691 and 697), environmental factors, or other factors, or combinations thereof. Some additional examples of considerations used in the regulation of a power source are discussed below.

In the example of FIG. 6, the currents through substrates 640a and 640c are measured, while the current through substrate 640b is not. Depending on the overall operation, it may not be necessary to monitor each of the substrates in an electroplating procedure. For example, various electroplating procedures may be performed with multiple substrates 640 that have similar or identical geometries, and with holders 645 that also have similar or identical geometries. This uniformity can be helpful in inferring that the current that flows through each of the substrates are the same or similar. With this inference, it may not be necessary to monitor each substrate in a batch of substrates. In some situations it may be sufficient to monitor 90%, 75%, 50%, 20%, 10%, 5%, or 1% of the substrates in a batch. Similarly, it may be sufficient to monitor a selected number such as 20, 10, 8, 5, 4, 3, 2, or just 1 of the substrates in a batch.

The determinations made by processor 694 can be based on various factors in addition to the current measurement(s). These additional factors can be encoded as comparison parameters or other information stored, for example, in memory 696. For example, processor 694 may be configured (using hardware, software, firmware, or combinations thereof) to hold a current level constant, within a fixed target range of values. For example, historical data from past experience may indicate that a particular process yields suitable platings if the current is held to 50 amperes through each wafer throughout a four-hour plating period. The historical data may further show that variations of up to +/−10 amperes also yield suitable platings. Based on these observations, processor 694 and/or the instructions that it executes may be configured to adjust power source 680 on an ongoing basis to ensure that the current measured at one of the sensors (e.g. sensor 614) stays in the range of 40 to 60 amperes during the plating period. This operation forms a feedback control loop: the adjustments made by processor 694 lead to changes in the voltage provided by power source 680, which affects the total current driven into bus bar 605, which affects the current that enters into wafer 640c, which affects the measurement made by sensor 614, which can lead to further changes by processor 694. Processor 694 and/or the instructions that it executes can be configured to use hysteresis, filtering, and/or other techniques to avoid positive-feedback or other control issues, and to keep the feedback loop stable.

Processor 694 and/or the instructions that it executes can use other approaches for determining how to control power source 680. For example, instead of maintaining a current level to be close to a fixed value (e.g., 20, 35, 50, 65, 70 amperes) throughout a plating procedure, a control system may be configured with a changing temporal profile. For example, the target value of a measured current may be 25+/−5 amperes during a first 1-hour phase of a plating process, 65+/−15 amperes during the next 1.5-hour phase of the plating process, and 65+/−5 amperes during a final 2-hour phase of the plating process. Other temporal profiles may also be used. For example, various implementations of processor 694 and associated software may be configured to accept linear or nonlinear algebraic expressions for the target value of a current being controlled (e.g., I_target=0.5 amps×time_hours; I_target=40 amps+3 amps×time_hours; I_target=40 amps+3 amps×sin(time_hours/1_hour)).

Moreover, the target profile may be adjusted based on the history of a plating process. For example, various implementations of processor 694 and associated software may be configured so that if a plating current has consistently been measured at the upper range of acceptable values through the first half of a plating process, then the current is subsequently maintained in the lower range of acceptable values for the second half of the plating process. Similarly, the duration of a plating process can be shortened (or extended) if the plating current has largely been measured on the high side (or on the low side) of acceptable values.

Also, various implementations of processor 694 and the instructions it executes (or both) may be configured or programmed so that the target current is based on other measured factors (e.g., ambient temperature) or input parameters (the target thickness of the plating layer, the desired tolerance of that target thickness, the number of substrates in a plating bath, the locations of the substrates in a plating bath, the arrangement of hangars on a bus bar, the volume or other geometry of an electrolyte bath, the cleanliness of the plating solution, whether the plating solution is fresh or has been used for a previous plating, or other factors).

As another example, various implementations of processor 694 and associated software can be configured to monitor the currents supplied to multiple wafers (e.g. wafers 640a and 640c, as depicted in FIG. 6), and to attempt to keep all of the measured currents within a target range. For example, consider a situation where memory 696 indicates that a target range is 50 amperes with a tolerance of +/−5 amperes, and processor 694 determines that wafer 640a is consistently receiving a little more current than wafer 640c. Such a situation may arise, for example, due to a faulty mounting, an adverse mounting position in the electrolyte bath for one of the wafers, or other reasons. In that situation, processor 694 may adjust power source 680 so that wafer 640a receives 52 amperes and wafer 640c receives 48 amperes of current. In effect, this mode of operation involves some degree of compromise among the measured values, with both of the measured values being kept within the tolerances. Similar compromises may be configured for three or more wafers.

In some situations, processor 694 may determine that a compromise is not possible. For example, consider a situation where a target current is 60 amperes, with a tolerance of +/−3 amperes. If processor 694 detects that five wafers in a batch are each receiving 62 amperes of current (just a bit more than the target value) while one anomalous wafer is receiving 47 amperes (substantially less than the target value), the processor may be unable to compensate in a way that keeps all the wafers in range. In such a situation the processor may adjust the power supply in a manner that deliberately sacrifices the anomalous wafer. For example, processor 694 may adjust the power source so that the five wafers receive the optimal 60 amperes, while the one anomalous wafer receives only 44 amperes. Alternatively, or in addition, the processor may trigger an alarm signal, alerting an operator that the anomalous wafer may need attention or adjustment.

In various implementations, processor 694 converts the information received in signals 691 and 697 based on system calibrations that are stored in memory 696. For example, after sensor 612 is affixed to holder 645a, an operator may perform an initial calibration procedure. In the calibration procedure, the operator can compare the measurements that are received in signal 691 against clamp-meter measurements for a variety of operating conditions. These comparisons can be used to generate a lookup table for use during subsequent electroplating operations.

In the illustrated example, control unit 690 is depicted with a single processor and a single memory. In other implementations, a control unit includes a data collection unit (e.g., having an input, a processor, and a memory) and also includes a feedback control unit (e.g., having an output, a processor, and a memory) that responds to information received by the data collection unit.

In various implementations, electroplating system 600 operates so that control unit 690 monitors a current and provides feedback control during an entire time duration of a plating process. In other implementations, a plating process may be performed with the monitoring during only a portion of the plating process. For example, the monitoring may operate substantially continuously during at least 25%, 40%, 50%, 75%, 90%, 98%, or 100% of a full duration of an electroplating activity. Similarly, automated feedback control may be used for all or part of a time duration of a plating process, such as substantially continuously during at least 15%, 25%, 35%, 40%, 50%, 75%, 80%, 90%, 95%, 98%, or 100% of a full duration of an electroplating activity.

Sensors 612, 614 may each incorporate a Hall sensor or a differentially-aligned pair of Hall sensors. The signal from the sensor can be conditioned by an amplifier, converted from analog to digital form, and possibly pre-processed by a microcontroller for transmission or local storage. A variety of technologies are contemplated for transmitting wireless links 691 and 697. In various implementations, wired links can be used instead of wireless links.

In various implementations of an electroplating system, the data can be logged in a unit on the holder instead of (or in addition to) being transmitted to a control unit. In such implementations, a holder or hangar can be equipped with a memory that stores current measurements for future reading. In various situations, an operator may manually remove the memory during a plating process for analysis and possible adjustment or fine-tuning of the plating process. In yet other implementations, current readings are stored in a memory buffer on a hangar, and are intermittently transmitted to a control unit.

FIG. 7 is a block diagram of one embodiment of a computer system 700 for controlling an electroplating activity. For example, computer system 700 may be an embodiment of control unit 690. Computer system 700 may include a processor 710 and a memory 720 coupled together by a communications bus 705. Processor 710 may be a single processor or a number of individual processors working together. Memory 720 is typically random access memory (RAM), or some other dynamic storage device, and is capable of storing data 726 and instructions to be executed by the processor, e.g., operating system 722 and applications 724. The applications 724 and operating system 722 may include software, firmware, and/or ROM instructions. Memory 720 may also be used for storing temporary variables or other intermediate information during the execution of instructions by the processor 710. In various situations, computer instructions, such as those used by processor 694 or other instructions used in a plating operation, may be stored on a non-transitory computer-readable storage medium.

Computer system 700 may also include input devices such as a keyboard, mouse, or touch screen 750, a USB interface 752, communications input and output components 754, output devices such as graphics & display 756, a magnetic memory storage such as hard disk 758, an optical memory storage such as CD-ROM 760, and a semiconductor memory storage such as removable flash memory card 770, all of which are coupled to processor 710, e.g., by communications bus 705. It will be apparent to those having ordinary skill in the art that computer system 700 may also include numerous elements not shown in the figure, such as additional storage devices, communications devices, input devices, and output devices.

The flow chart of FIG. 5 illustrates some of the many operational examples of the techniques disclosed in the present application. Those having ordinary skill in the art will readily recognize that certain steps or operations illustrated in FIG. 5 may be eliminated or taken in an alternate order, or with alternate configurations such as various electroplating arrangements or alternating or reversed current flows. Moreover, various aspects of these steps or operations can be implemented as one or more software programs for a computer system (e.g., computer system 700) and are encoded in a computer readable medium (e.g., memory 720) as instructions executable on one or more processors (e.g., processor 710). A tangible computer readable medium may include, for example, an electronic storage medium, a magnetic storage medium, or an optical storage medium, or combinations thereof. The software programs may also be carried in a communications medium conveying signals encoding the instructions. Separate instances of these programs may be executed on separate computer systems, e.g., in a multi-processor architecture. Thus, although certain steps have been described as being performed by certain devices or software programs, this need not be the case and a variety of alternative implementations will be understood by those having ordinary skill in the art.

Those having ordinary skill in the art will readily recognize that the techniques and methods discussed below may be implemented in software using a variety of computer languages, including, for example, traditional computer languages such as assembly language, Pascal, and C; object oriented languages such as C++, C#, and Java; and scripting languages such as Perl and Python. Additionally, software 724 may be provided to the computer system via a variety of computer readable media including electronic media (e.g., flash memory), magnetic storage media (e.g., hard disk 758, a floppy disk, etc.), optical storage media (e.g., CD-ROM 760 or DVD-ROM), other tangible storage media, and communications media conveying signals encoding the instructions (e.g., via a wired or wireless network coupled to communications input and output components 754).

Although the present disclosure has been described in connection with several embodiments, the disclosure is not intended to be limited to the specific forms set forth herein. On the contrary, it is intended to cover such alternatives, modifications, and equivalents as can be reasonably included within the scope of the disclosure as defined by the appended claims.