KR102242321B1 - Apparatus and methods for acoustical monitoring and control of through-silicon-via reveal processing - Google Patents

Abstract

Through silicon via (TSV) exposure processes using chemical mechanical polishing (CMP) can be acoustically monitored and controlled to detect and automatically respond to TSV breakage. Acoustic emissions received by one or more acoustic sensors located proximate to the substrate holder and/or polishing pad of the CMP system can be analyzed to detect TSV breakage during the CMP process. In response to detecting TSV breakage, one or more corrective actions may occur automatically. In some embodiments, the polishing pad platen may have one or more acoustic sensors incorporated therein extending to a polishing pad mounted on the polishing pad platen. Methods of monitoring and controlling the TSV exposure process are provided as well as other aspects.

Description

Apparatus and method for acoustic monitoring and control of through silicon via exposure treatment {APPARATUS AND METHODS FOR ACOUSTICAL MONITORING AND CONTROL OF THROUGH-SILICON-VIA REVEAL PROCESSING}

Related applications

This application was filed on May 1, 2013, and the name of the invention is "APPARATUS AND METHODS FOR ACOUSTICAL MONITORING AND CONTROL OF THROUGH-SILICON-VIA REVEAL PROCESSING", US Patent Application No. 13/874,495 (Agent Arrangement No. 20654/ USA), which patent application is incorporated herein by reference in its entirety for purposes of reference.

Technical field

TECHNICAL FIELD The present invention relates generally to semiconductor device fabrication, and more particularly to backside chemical mechanical polishing of through-silicon vias (TSVs).

Chemical mechanical polishing (CMP), also known as chemical mechanical planarization, is a process commonly used in the manufacture of integrated circuits (ICs) on semiconductor substrates. The CMP process can remove topographic features and materials from the partially processed substrate to create a flat surface for subsequent processing. The CMP process may use an abrasive and/or a chemically-active polishing solution on one or more rotating polishing pads that are pressed against the surface of the substrate. The substrate can be held in a substrate holder that rotates the substrate. The substrate holder can also vibrate the substrate back and forth over the surface of the rotating polishing pad(s).

In the manufacture of ICs, 3D packaging can be used to increase circuit functionality and/or performance in a compact footprint. Three-dimensional packaging may involve interconnection of IC chips stacked on top of each other using TSVs (through silicon vias) to electrically connect the stacked IC chips. TSVs are vertical electrical conductors that extend through the substrate. To access TSVs from the back of the substrate (for subsequent electrical connection to another IC below), CMP can be used in the TSV reveal process. The TSV exposure process may include grinding and etching the rear surface of the substrate to expose the TSVs as stubs protruding from the rear surface. Next, a dielectric film can be deposited on the back surface. CMP can be used to complete the TSV exposure process by removing the protruding stubs and polishing the back surface to the desired dielectric film thickness. However, TSV breakage (ie, breakage of one or more stubs) can occur that can destroy the substrate. Therefore, improved TSV exposure processes are needed.

According to one aspect, a platen for a chemical mechanical polishing (CMP) device is provided. The platen comprises: a disk-shaped base configured to receive a polishing pad on a surface, the disk-shaped base having at least one through-hole; And an acoustic sensor received in the at least one through hole and protruding from the surface of the disk-shaped base, the acoustic sensor being configured to be electrically connected to the controller.

According to another aspect, a CMP apparatus configured to perform a chemical mechanical polishing (CMP) process is provided. The CMP apparatus comprises: a platen comprising a polishing pad, a substrate holder configured to hold a substrate to be polished-the platen or substrate holder is configured to bring the substrate and the polishing pad into contact with each other -, positioned close to the polishing pad or substrate during the CMP process An acoustic sensor, and an acoustic processor electrically connected to the acoustic sensor and configured to analyze one or more signals received from the acoustic sensor to detect through silicon via (TSV) breakage.

According to a further aspect, a method of monitoring and controlling a through silicon via (TSV) exposure process is provided. The method includes processing the substrate using a chemical mechanical polishing (CMP) process, sensing acoustic emissions of the CMP process, and analyzing the acoustic emissions to detect TSV breakage.

Still other aspects, features, and advantages of the invention may be readily apparent from the following detailed description, in which a number of exemplary embodiments and implementations are described and illustrated, including the best mode contemplated for carrying out the invention. have. The invention may also include other and different embodiments, and some details of it may be modified in various aspects, all of which do not depart from the scope of the invention. Accordingly, the drawings and descriptions should be regarded as illustrative in nature and not limiting. The drawings are not necessarily drawn to scale. The invention covers all modifications, equivalents and alternatives falling within the scope of the invention.

The drawings described below are for illustration only. The drawings are not intended to limit the scope of the invention in any way.
1A to 1C show sequential cross-sectional views of a semiconductor substrate undergoing a TSV exposure process without TSV (through silicon via) breakage according to the prior art.
2 shows a TSV without breakage according to the prior art.
3 shows a cross-sectional view of a semiconductor substrate with TSV breakage according to the prior art.
4 shows a TSV with breakage according to the prior art.
5 shows a schematic partial side view of a chemical mechanical polishing (CMP) system according to embodiments.
6A and 6B show a top view and a side cross-sectional view (cut along line 6B-6B in FIG. 6A), respectively, of a platen and polishing pad of a CMP system according to embodiments.
7 shows a flowchart of a method of monitoring and controlling a TSV exposure process according to embodiments.

Reference will now be made in detail to exemplary embodiments of the present disclosure, which are shown in the accompanying drawings. Wherever possible, the same reference numbers will be used throughout the drawings to refer to the same or similar parts.

In one aspect, the through silicon via (TSV) exposure process using chemical mechanical polishing (CMP) can be acoustically monitored and controlled to detect and automatically respond to TSV breakage. In some IC manufacturing processes, TSV breakage can occur more often during CMP, where the TSV aspect ratio (i.e., exposed TSV stub height versus TSV diameter) can be high (e.g., TSVs with small diameters). High aspect ratio TSVs can allow an IC to have a greater density of chip-to-chip interconnects. However, high aspect ratio TSVs may be less rigid and thus may be more susceptible to breakage during the CMP process removing exposed TSV stubs from the back surface of the substrate.

One or more acoustic sensors may be located in the CMP system to receive acoustic emissions during the CMP process. One or more acoustic sensors may be connected to the substrate holder and/or the polishing pad platen, for example. In some embodiments, the polishing pad platen may have one or more acoustic sensors incorporated therein extending to a polishing pad mounted on the polishing pad platen.

In some embodiments, acoustic emissions received by one or more acoustic sensors may be analyzed by a system controller and/or acoustic processor to detect TSV breakage. The acoustic processor may be part of the CMP system controller, or alternatively may be a separate standalone component connected to the CMP system controller. In response to detecting the TSV breakage, the system controller and/or acoustic processor may automatically initiate one or more remedial actions. For example, in some embodiments, an operator may be notified of a TSV break. Alternatively or additionally, the CMP process pre-determines a down force exerted by one of the substrates or polishing pads on the other, as may be pre-programmed in the system controller and/or acoustic processor. It can be corrected automatically by reducing by an amount, by reducing the rotational speed of the polishing pad and/or the substrate by a predetermined amount, and/or a combination of both. In some embodiments, the CMP process may be automatically stopped and/or control transferred to an endpoint routine of the system controller in response to detecting a TSV break.

In other aspects, methods of monitoring and controlling the TSV exposure process are provided, as described in more detail below in connection with FIGS. 1A-7.

1A-1C illustrate a substrate 100 undergoing a TSV exposure process, which may be referred to as a backside via reveal (BVR) CMP process, according to the prior art. 1A shows a substrate 100 having a back surface 102A partially treated by a TSV exposure process. The substrate 100 includes a silicon base layer 104, a metal (e.g., copper) layer 106, a plurality of TSVs 108 extending from the metal layer 106 and protruding beyond the silicon base layer 104, A barrier layer 110 covering the TSVs 108 and the metal layer 106, and a dielectric layer 112 covering the rear surface 102A. In some fabrication processes, TSVs 108 may have a height H above the silicon base layer 104, which may range from about 2 μm to about 4 μm, and TSV 108 and TSV 108 It can be different between the livers. The substrate 100 with the back surface 102A may be received in a CMP system for further TSV exposure treatment as shown in FIGS. 1B and 1C.

1B shows a substrate 100 having a further treated back surface 102B, wherein the dielectric layer 112 and the barrier layer 110 are top surfaces 109 of the TSVs 108 by a CMP process. May have been removed from If TSV breakage does not occur, the CMP process continues to remove materials from the back surface 102B of the substrate 100 and/or polish this back surface until the back surface 102C of FIG. 1C is created. I can. 1C, the TSVs 108 may be flush with the surface 113 of the dielectric layer 112, or in some manufacturing processes when the desired less thick dielectric layer 112 is achieved. Can be slightly lower. As shown, a predetermined copper dishing 111 may occur on the end surfaces of the TSVs 108. In order to control the surface finish and remove small surface defects and flaws, a final soft buff can be provided. When TSV breakage does not occur, the final surface state of the substrate 100 upon completion of the TSV exposure process may appear as shown in FIG. 1C.

2 shows a micrograph of a substrate 200 with a TSV 208 and a peripheral rear substrate surface 202 upon completion of the TSV exposure process without TSV breakage.

3 shows a substrate 300 having a treated back surface 302 with TSV breakage according to the prior art. TSV failure can cause non-reworkable scratches and/or defects across the substrate surface that can adversely affect IC chip yield and reliability. The substrate 300 may have a silicon base layer 304, a metal (eg, copper) layer 306, TSVs 308a and 308b, a barrier layer 310 and a dielectric layer 312. TSV 308b may have been broken during the CMP process. This breakage may have caused oxide gouging 315, which may expose the silicon layer 304 to metal contamination during processing. In some embodiments, TSV 308b may be formed of copper, a relatively soft material, for example. Copper smearing on the silicon layer 304 caused by TSV failure can potentially affect IC quality and/or reliability during post-packaging electrical testing.

4 shows a micrograph of a substrate 400 with a TSV 408 and a peripheral rear substrate surface 402 after TSV failure. As can be seen, significant surface scratches and defects can occur after failure of the TSV 408 during the CMP process. In addition, metal grains that may have been removed by TSV breakage, for example, cause the metal pad 414 (i.e., the top surface of the TSV 408) to meet one or more specifications required for further processing. May not be met, which may further affect IC yield and/or reliability.

5 illustrates a chemical mechanical polishing (CMP) system 500 in accordance with one or more embodiments. The CMP system 500 may be configured to hold the substrate 501 in contact with the polishing pad 516 and may be used to perform the CMP process on the substrate 501 as part of the TSV exposure process. The substrate 501 may be a silicon-containing wafer, such as a patterned wafer comprising partially or fully formed transistors and a plurality of TSVs formed therein. The substrate 501 may be attached (eg, via adhesives) to a second carrier wafer or other suitable backing such that a TSV exposure process can be performed on the substrate 501. The polishing pad 516 may be mounted on a platen 518, which may be disk-shaped and rotated by a suitable motor (not shown) connected to the platen 518 by a shaft 520. The platen 518 may be rotated at about 10-200 rpm. Other rotational speeds may be used.

The substrate 501 may be held in the substrate holder 522. Substrate holders may also be referred to as retainers or carrier heads. In some embodiments, the substrate 501 may be held in the substrate holder 522 via vacuum. Other suitable substrate holding techniques may be used. In some embodiments, the substrate holder 522 may be configured to move the substrate 501 to contact the polishing pad 516 and away from the polishing pad (ie, up and down as shown). The substrate holder 522 may be rotated, and in some embodiments may be vibrated back and forth across the surface of the polishing pad 516 when the polishing pad 516 is rotating in contact with the rear surface of the substrate 501. . In some embodiments, the vibration speed of the substrate holder 522 may be about 0.1 mm/sec to 5 mm/sec. Other vibration speeds may be used. In some embodiments, the substrate holder 522 may be rotated at about 10-200 rpm. Other rotational speeds may be used. Vibration may occur between the center and the radial side of the polishing pad 516. In some embodiments, the substrate holder 522 may be a contour, 5-zone pressure head available from Applied Material of Santa Clara, CA.

In other embodiments, the positions of the polishing pad 516/platen 518 and the substrate 501/substrate holder 522 may be reversed. That is, the polishing pad 516 and the platen 518 move the polishing pad 516 upwardly away from the rear surface of the substrate 501 held in the substrate holder 522 and downwardly in contact with the rear surface. It may be mounted on an overhead assembly or polishing head configured to be used, or may be part of an overhead assembly or polishing head.

A slurry 524 (chemical polishing solution) may be applied to the polishing pad 516 by a dispenser 526 and inserted between the polishing pad 516 and the substrate 501. The distributor 526 can be connected to the slurry supply 528 through one or more suitable conduits. A pump 530, valve 532, or other liquid delivery and delivery mechanism may supply a metered amount of slurry 524 to the surface of the polishing pad 516. In some embodiments, the slurry 524 may be supplied by the distributor 526 onto the surface of the polishing pad 516 prior to the substrate 501, whereby the slurry 524 May be received at, and drawn between the polishing pad 516 and the substrate 501 by rotation of the polishing pad 516.

In some embodiments, one or more portions of the CMP system 500 may be equivalent to or based on those of the Reflexion® GT™ CMP system by Applied Material, Santa Clara, CA, for example.

The CMP system 500 may also include one or more acoustic sensors 534a and/or 534b operable to detect acoustic emissions that occur during a CMP process performed on the substrate 501. In some embodiments, CMP system 500 may include only one of acoustic sensors 534a or 534b. In other embodiments, the CMP system 500 may include both acoustic sensors 534a and 534b. In still other embodiments, the CMP system 500 may include more than two acoustic sensors, and these acoustic sensors may be positioned differently than that shown for the acoustic sensors 534a and 534b.

Acoustic sensors 534a and/or 534b may be positioned proximate to polishing pad 516 and/or substrate 501 during the CMP process. In some embodiments, acoustic sensor 534a may be physically connected to platen 518 (or overhead polishing head) in any suitable manner. For example, the acoustic sensor 534a may be mounted on a bracket that is mechanically fixed to the platen 518. In some embodiments, platen 518 may be an assembly including an upper platen and a lower platen (not shown) attached together. The upper platen may have a polishing pad 516 mounted on the upper platen, wherein the acoustic sensor 534a is integrated or integrated into the outer edge of the lower platen, for example via a bracket or other suitable mechanism. It can be mounted or integrated or mounted under the upper platen. In some embodiments, a bracket or other suitable mechanism may include a spring-loading mechanism to ensure that the acoustic sensor 534a maintains constant contact with the polishing pad. In some embodiments, a bracket or other suitable mechanism may include a cushioning pad to reduce signal attenuation or degradation. In some embodiments, the power supply and signal cable (which may be at least partially represented by signal connection 536a) are routed through platen 518 (or the lower platen of the platen assembly described above) and , Can be connected to the sensor 534a through a high frequency (eg, about 1 MHz) 8-terminal slip ring. In some embodiments, in addition to or as an alternative to the acoustic sensor 534a, the acoustic sensor 534b may be physically connected to the substrate holder 522 in any suitable manner. For example, the acoustic sensor 534b may be mounted on a bracket that is mechanically fixed to the substrate holder 522. Alternatively, acoustic sensors 534a and/or 534b may be located at other suitable locations relative to substrate 501 and polishing pad 516. In some embodiments, acoustic sensors 534a and/or 534b are directly embedded in or integrated with platen 518, substrate holder 522, and/or any other suitable component of CMP system 500. May be (see, for example, platen 618 described below in connection with FIGS. 6A and 6B ).

Acoustic sensors 534a and/or 534b, through wireless or wired signal connections 536a and/or 536b, are configured to detect TSV breakage based on acoustic emissions from the CMP process, and / Or may be configured to be electrically connected to the system controller 540, respectively.

The acoustic processor 538 may be part of the system controller 540 as shown, or may be a separate standalone component that may be electrically connected to the system controller 540. The system controller 540 may include a processor 542, which may control the operation of the CMP system 500, including one or more CMP processes used in the TSV exposure process. In some embodiments, the system controller 540 may not be connected to the acoustic processor 538 and/or may not include an acoustic processor, but instead the processor 542 may be used as the acoustic processor 538 described herein. ) Can be performed additionally.

The acoustic processor 538 may be configured to receive one or more signals indicative of acoustic emissions transmitted by the acoustic sensors 534a and/or 534b. The acoustic processor 538 may be configured to detect TSV breakage by analyzing one or more signals received from the acoustic sensors 534a and/or 534b. One or more signals received from acoustic sensors 534a and/or 534b may have amplitudes that may vary over time (eg, representing acoustic emission intensity). The acoustic processor 538 may be configured to receive time-varying signals and may compare the amplitudes of these signals with one or more thresholds and/or threshold bands. Signal amplitudes that exceed those thresholds or are outside of those threshold bands may indicate TSV breakage. In some embodiments, processing of the received signals may include comparing certain aspects or regions of the received signal waveforms with preset thresholds. The acoustic processor 538 may include suitable signal filtering, amplification, transformation (eg, A/D transformation) and processing components, and may include a suitable memory configured to store data and one or more analyses. Data and analysis may be stored in any suitable storage medium (eg, RAM, ROM or other memory) of, for example, acoustic processor 538 and/or system controller 540. One or more stored analyzes and data can be used to monitor and control one or more CMP processes with respect to detection of TSV breakage.

In some embodiments, frequency based analysis may be used to process the acoustic data. High sampling rate acquisition of acoustic signals from acoustic sensors 534a and/or 534b can be achieved by stationary signal analysis such as fast Fourier transformations (FFT), or wavelet packet transformation (WPT). packet transformations). The WPT can decompose the received acoustic signal into two parts: a low-frequency component that can approximate the signal identity, and a high-frequency component that can calculate the details of the signal. The decomposition can be repeated with subsequent approximations decomposed one after the other.

In other embodiments, time-based analysis may be used to process the acoustic data. For example, if the TSV breakage events have signal spikes that are large enough with respect to the signal-to-noise ratio, a simple root mean square (rms) of acoustic signals received from acoustic sensors 534a and/or 534b can be monitored.

In order to correlate the received acoustic signals and TSV break events, the following setup procedure may be used in some embodiments. The first setup substrate without protruding TSV stubs may undergo a CMP process to generate baseline acoustic signal data for normalization. A second setup substrate with very high protruding TSV stubs (eg, 15 μm stub length with 5 μm diameter) can undergo the CMP process. The second set-up substrate can be inspected for TSV breakage after processing, for example using optical inspection or scanning electron microscopy. A comparison of the recorded signal magnitudes from the first and second setup substrates can be made. Any signal spikes visible above the baseline signal in acoustic activity during steady state CMP processing can be classified as break signals, which can then be used to correlate TSV break events.

The acoustic processor 538 can automatically respond to detection of a TSV break by initiating one or more corrective actions. Corrective actions may include, for example, causing an audible alarm or notifying an operator by displaying a warning or other type of message on a display device connected to the system controller 540. Additionally or alternatively, the corrective actions may include automatically stopping the CMP process in response to detecting a TSV break. Additionally or alternatively, the corrective actions may include automatically modifying one or more parameters of the CMP process in response to detecting a TSV break. For example, the acoustic processor 538 may, in response to detecting a TSV breakage, the polishing pad 516 according to one or more programmed routines executed by the acoustic processor 538 or system controller 540. It will be configured to automatically reduce the downward force (or vice versa) exerted by the substrate holder 522 and/or automatically reduce the rotational speed of the substrate holder 522, platen 518, or both. I can. This may allow the CMP system 500 to automatically continue processing of subsequent substrates with modified processing parameters.

6A and 6B illustrate an assembly 600 of a platen 618 and polishing pad 616 that may be used in a CMP apparatus, such as, for example, a CMP system 500, in accordance with one or more embodiments. The platen 618 can include a disk-shaped base 644 configured to receive a polishing pad 616 on a surface 617 of the disk-shaped base 644. The disk-shaped base 644 may have one or more through holes 633a, 633b, and 633c. That is, in some embodiments, the disk-shaped base 644 has only one of the through-holes 633a, 633b and 633c, or only two or three of the through-holes 633a, 633b and 633c. It may have more through-holes than the balls 633a, 633b, and 633c.

The platen 618 may also include one or more acoustic sensors 634a, 634b and/or 634c received in each of the through holes 633a, 633b and 633c. In some embodiments, the acoustic sensors 634a, 634b and/or 634c may be friction fit to each of the through holes 633a, 633b, and 633c. In other embodiments, the acoustic sensors 634a, 634b and/or 634c may be physically connected to or integrally formed with the disk-shaped base 644 in any suitable manner. In some embodiments, the platen 618 may have through-holes 633a, 633b, and 633c in which the acoustic sensor is not accommodated.

In some embodiments, acoustic sensors 634a, 634b and/or 634c may protrude from surface 617 of disk-shaped base 644 by a distance D1. Distance D1 may be selected to reduce acoustic signal attenuation, which may occur, for example, in the polyurethane soft SUBA™ portion of some embodiments of polishing pad 616. In some embodiments, the distance D1 may be about 50 mils (about 1.27 mm). This may ensure that one or more of the acoustic sensors 634a, 634b and/or 634c may be very close to the polishing surface 621 of the polishing pad 616 but are less likely to be damaged during polishing.

In some embodiments, acoustic sensor 634a may be located approximately at the center of platen 618. This central position can ensure that the distance of the acoustic sensor 634a with respect to the substrate being processed is kept constant. The acoustic sensor 634b may be located radially outward by a distance D2 from the center of the platen 618, and the acoustic sensor 634c may be located radially outward by a distance D3 from the center of the platen 618. have. In one or more embodiments, the distance D2 may be about 5 inches (about 12.7 cm) radially outward from the center of the platen 618 and the distance D3 is about 10 inches radially outward from the center of the platen 618 ( About 25.4cm). In some embodiments, a D3 distance of about 10 inches (about 25.4 cm) can ensure that the acoustic sensor 634c is positioned closest to the substrate in every rotation pass. In some embodiments, as the substrate moves away from the acoustic sensor 634c during CMP processing, the received acoustic data may be filtered. The distances D2 and/or D3 may alternatively have other suitable dimensions.

Each of the acoustic sensors 634a, 634b and/or 634c may be configured to be electrically connected to a controller or an acoustic processor through a wired or wireless connection. In some embodiments, acoustic sensors 634a, 634b and/or 634c have electrical connectors 646a, 646b and/or 646c accessible below platen 618 (i.e., opposite surface 617). ) Can be included.

6B, the polishing pad 616 may be mounted on the disk-shaped base 644, and one or more non-through holes (non-through holes) on the base-side surface 619 of the polishing pad 616 ( 635a, 635b and 635c). The non-penetrating holes 635a, 635b and 635c may have a depth of about a distance D1 and be configured to receive a protruding portion of one or more of the acoustic sensors 634a, 634b and/or 634c therein. I can. The number and position of the non-penetrating holes 635a, 635b, and 635c may correspond to the number and positions of the through-holes 633a, 633b, and 633c of the platen 618, respectively. The polishing pad 616 may be the same as or similar to an IC1000™ polishing pad with SUBA™ IV subpads having one or more non-through holes 635a, 635b and/or 635c formed therein, for example.

In some embodiments, any one or more of the acoustic sensors 534a, 534b, 634a, 634b and/or 634c may be piezoelectric, transducer and/or accelerometer type sensors, each of which may have a high signal-to-noise ratio. have. In some embodiments, acoustic sensors 534a, 534b, 634a, 634b, and/or 634c may include a flat frequency response over a range of about 100-500 kHz. In some embodiments, any one or more of the acoustic sensors 534a, 534b, 634a, 634b, and/or 634c may amplify the acoustic signal with a gain of about 40-60 dB. In some embodiments, acoustic sensors 534a, 534b, 634a, 634b, and/or 634c may include a high pass filter having a range of about 50 Hz-100 Hz. Any suitable acoustic sensor may be used for the sensors 534a, 534b, 634a, 634b and/or 634c.

7 illustrates a method 700 of monitoring and controlling a TSV exposure process in accordance with one or more embodiments. At process block 702, method 700 may include processing the substrate using a CMP process. The CMP process can be part of the TSV exposure process. For example, referring to FIGS. 1A-1C and 5, a substrate 100 having a rear surface 102A may be received in a CMP system 500. The substrate 100 may be mounted or attached to the substrate holder 522, and a polishing pad 516 for CMP processing as shown and described with respect to the rear surfaces 102B and 102C, or possibly 302. Pressed against.

At process block 704, detection of acoustic emissions of the CMP process may occur. 5, 6A and 6B, the detection of acoustic emissions may be performed by any one or more of the acoustic sensors 534a, 534b, 634a, 634b and/or 634c. Acoustic emissions are, for example, a substrate 100 having a rear surface 102A of FIG. 1A or a substrate such as the substrate 501 of FIG. 5, the polishing pad 516 of FIG. 5 or the polishing of FIGS. 6A and 6B. It may be from a CMP process being processed with a polishing pad such as pad 616. Acoustic sensors 534a, 534b, 634a, 634b and/or 634c can detect acoustic emissions due to processing of the substrate 100 or 501, and provide electrical signals indicative of such acoustic emissions, for example a system controller. 540 and/or a controller such as the sound processor 538 and/or the sound processor.

At process block 706, method 700 may include analyzing acoustic emissions to detect TSV breakage. Analysis of acoustic emissions may include comparing one or more parameters (eg, amplitude) of one or more received signals with one or more thresholds and/or threshold ranges. One or more received signals may indicate acoustic emissions from the CMP process, and one or more thresholds and/or threshold ranges may indicate whether TSV breakage occurred during the CMP process. During one or more baseline CMP processes performed on the first and second setup substrates where TSV failure occurred and not, respectively, one or more thresholds and/or threshold ranges may have been predetermined.

If TSV breakage is detected at decision block 708, the method 700 may proceed to process block 710. For example, in some embodiments, a high spike in the received acoustic signal may be, for example, part of a program running on the acoustic processor 538 or part of the endpoint software running on the system controller 540, for example. The method 700 may be triggered to proceed to process block 710 according to a pre-defined algorithm that may be used. If TSV breakage is not detected, method 700 may proceed to decision block 712.

At process block 710, method 700 may include automatically responding to detection of a TSV break. In some embodiments, this step may include automatically notifying an operator who may be reworkable of the substrate currently being processed. In some embodiments, method 700 may additionally or alternatively break TSV by automatically modifying the CMP process, e.g., by reducing the downward force, reducing the rotational speed, or both. Can respond to the detection of. The method 700 may additionally or alternatively respond to detection of a TSV break by automatically stopping the CMP process. This may occur either by going directly to the end block 714 (the path shown by the broken line), or in some embodiments, by going to decision block 712, in which a "yes" response may be automatically triggered. And, thereby effectively stopping the CMP process. Method 700 may proceed to decision block 712 in another manner.

At decision block 712, method 700 may include determining whether an endpoint of the CMP process has been detected. Endpoint detection may be performed by a system controller of a CMP system, such as, for example, system controller 540 of CMP system 500. In some embodiments, endpoint detection for the TSV exposure process may include detecting the point at which TSVs are planarized coplanar with the dielectric oxide surface, as shown in FIGS. 1C and 2. This endpoint detection can be determined, for example, by acoustic analysis and/or motor torque feedback of the motor drive, for example rotation of the polishing pad. Both acoustic analysis and motor torque feedback can be based on friction changes that may occur when the material(s) being processed is changed. For example, when the CMP process changes from predominantly removing/polishing metal material on the substrate surface to predominantly removing/polishing oxide material on the substrate surface, one or more received acoustic signals, between the substrate surface and the polishing pad. Friction changes may occur that may be exhibited at and/or in the received motor torque feedback. Additionally or alternatively, endpoint detection for the TSV exposure process can be determined based on a white light spectrograph indicating a specific oxide thickness. If an endpoint is detected at decision block 712, the method 700 may proceed to an end block 714. If the endpoint has not been detected, the method 700 may return to process block 704.

At the end block 714, the method 700 and processing of the substrate using the CMP process may end.

The process and decision blocks of method 700 may be executed or performed in an order or sequence that is not limited to the order and sequence shown and described. For example, in some embodiments, process block 704 may be performed concurrently with process blocks 706 and/or 710 and/or decision blocks 708 and/or 712.

Those of ordinary skill in the art will readily appreciate that the invention described herein permits a wide range of applications and applications. Many embodiments and adaptations of the invention other than those described herein, as well as many modifications, modifications, and equivalent configurations, other than those described herein, without departing from the nature or scope of the invention, are intended for use in the invention and the foregoing description. From or will be appropriately suggested by him. Accordingly, although the present invention has been described in detail with reference to specific embodiments herein, the present disclosure is only illustrative, and provides examples of the present invention, and provides a complete and usable disclosure of the present invention. It should be understood that it is made only for the purpose of doing so. This disclosure is not intended to limit the invention to the specific apparatus, device, assembly, system or method disclosed, and, on the contrary, is intended to cover all modifications, equivalents, and alternatives included within the scope of the invention.

Claims (15)

As a platen for a chemical mechanical polishing (CMP) device,
A disk-shaped base, wherein the disk-shaped base is configured to receive a polishing pad on a surface of the disk-shaped base, the disk-shaped base having at least one through hole;
An acoustic sensor received in the at least one through hole and protruding from a surface of the disk-shaped base, the acoustic sensor being configured to be electrically connected to a controller; And
A bracket connected to the disk-shaped base and holding the acoustic sensor-the bracket is a cushioning pad configured to reduce signal attenuation and the acoustic sensor and the polishing pad when the polishing pad is mounted on the disk-shaped base Includes a bias mechanism configured to maintain constant contact between polishing pads-
Platen comprising a. The method of claim 1,
The polishing pad, the platen having a non-through hole configured to receive the acoustic sensor therein on the base side surface of the polishing pad. The method of claim 1,
The at least one through hole is located at the center of the disk-shaped base, or 5 inches radially outward from the center, or 10 inches radially outward from the center. The method of claim 1,
And the acoustic sensor protrudes by 50 mils from the surface of the disk-shaped base. A CMP device configured to perform a chemical mechanical polishing (CMP) process, comprising:
A platen configured to receive a polishing pad;
A substrate holder configured to hold a substrate to be polished, the platen or the substrate holder configured to cause the substrate and the mounted polishing pad to contact each other;
An acoustic sensor positioned close to the polishing pad or the substrate during the CMP process; And
An acoustic processor electrically connected to the acoustic sensor and configured to analyze the one or more signals to detect TSV failure by correlating one or more signals received from the acoustic sensor with recorded TSV failure events
Including, the recorded TSV damage event,
Recording a baseline acoustic signal using a first setup substrate without protruding TSV stubs, recording a TSV failure acoustic signal using a second setup substrate with protruding TSV stubs, and identifying signal spikes indicative of TSV failure events Generated by comparing the baseline acoustic signal with the TSV damaged acoustic signal. The method of claim 5,
The acoustic processor is further configured to automatically notify an operator in response to detecting a TSV break. The method of claim 5,
The acoustic processor automatically triggers the CMP process in response to detecting a TSV break, by reducing a down force in response to detecting a TSV break, reducing a rotational speed, or both. Also configured to stop or modify as, a CMP device. The method of claim 5,
The acoustic sensor comprises a high pass filter having a flat frequency response over a range of 100-500 kHz and a range of 50-100 Hz. The method of claim 5,
The acoustic sensor amplifies an acoustic signal with a gain of 40-60dB, CMP device. The method of claim 5,
The acoustic sensor is incorporated into the platen such that the acoustic sensor protrudes from the surface of the platen to the polishing pad. The method of claim 5,
The acoustic sensor is incorporated in the platen, at the center of the platen, or at 5 inches radially outward from the center, or 10 inches radially outward from the center. A method of monitoring and controlling a through silicon via (TSV) exposure process, comprising:
Recording a baseline acoustic signal using a first setup substrate that does not have protruding TSV stubs;
Providing a second setup substrate having protruding TSV stubs;
Recording a TSV breakage sound signal using the second setup substrate having protruding TSV stubs;
Comparing the baseline acoustic signal with the TSV vandal acoustic signal to identify signal spikes indicative of TSV vandal events;
Processing the substrate using a chemical mechanical polishing (CMP) process;
Sensing acoustic emissions of the CMP process; And
Analyzing the acoustic emissions, based on the correlation between the acoustic emissions and recorded TSV break events, detecting TSV break
How to include. The method of claim 12,
In response to detecting the TSV breakage, the method further comprises automatically notifying an operator. The method of claim 12,
Responsive to detecting TSV breakage, automatically stopping or modifying the CMP process. The method of claim 14,
Wherein the automatically correcting comprises automatically correcting the CMP process by reducing a downward force, reducing a rotational speed, or both, in response to detecting a TSV breakage. Way.

Images