CN114800248A - Monitoring device for dynamic sensing of single-side chemical mechanical planarization processing - Google Patents

Abstract

The invention discloses a monitoring device for dynamic sensing of single-side chemical mechanical planarization processing, and belongs to the technical field of ultra-precision processing monitoring. This monitoring devices installs on single face grinding and polishing machine, can realize the online real-time supervision of plane finishing in-process work piece state, its characterized in that: the sensor unit is included: the method comprises the steps that at least two sensors are used for collecting monitoring signals for joint monitoring; a signal conditioning unit: amplifying, attenuating and/or spectrally filtering the acquired signal amplitude; a data acquisition unit: converting the analog signals processed by the signal conditioning unit into digital signals, and performing data preprocessing of digital filtering, noise reduction and storage; a processor unit: and carrying out mode recognition on the data preprocessed by the data acquisition unit to realize visualization of dynamic indexes of workpiece processing. The on-line real-time monitoring device designed by the invention has the advantages of multiple sensors, interference resistance, high accuracy, wide application range, low cost, convenience in operation, high intelligent level and the like.

Description

A monitoring device for dynamic perception of single-sided chemical mechanical planarization

technical field

The invention relates to the field of ultra-precision machining monitoring, in particular to a monitoring device for dynamic perception of single-sided chemical-mechanical planarization machining.

Background technique

With the increasing progress of optics, optoelectronics, and materials technology, it has also promoted the wide application and rapid development of semiconductor, optical and optoelectronic devices in the fields of electronics, optics, instrumentation, aerospace, and national defense.

Photofinishing of semiconductor, optical and optoelectronic materials is generally achieved by ultra-precision chemical mechanical planarization (ChemicalMechanical Planarization, CMP) methods. Efficient, high-quality ultra-precision CMP processing (nanometer to sub-nanometer precision) depends not only on processing equipment, methods, materials, and precision, but also on detection accuracy, sensitivity, resolution, and depth. The depth is not a scale of physical meaning, but a level based on the mechanism. It is to reveal the physical and chemical mechanism changes during the surface processing of the workpiece from the measurement results of simple physical phenomena, and to be able to process the workpiece based on its processing mechanism and processing characteristics. Identify the surface quality, establish the mapping relationship between changes in material properties, dimensions, and microscopic topography during processing and processing requirements and indicators .

The existing devices for monitoring the ultra-precision chemical mechanical planarization process mainly include: (1) by collecting the reflected light of the polished surface of the workpiece and calculating the predetermined characteristic value to determine the polishing end point (CN200910133188.6, CN200480025849.9 and CN200380101631.2); (2) by measuring the current flowing into the polishing machine and comparing the current values corresponding to different positions with the reference value to determine whether the polishing state is faulty (CN201110059769.7); (3) by setting the The marking piece on the same layer of the polishing layer indicates the consumption of the polishing layer, and the monitoring of the marking piece realizes the monitoring of the consumption of the polishing layer, so as to determine whether the polishing pad needs to be replaced (CN201510898297.2); (4) The measured value is obtained by the acquisition module A time-series data line for characterizing the film layer on the substrate surface with time is formed, the inflection point of the time-series data line is obtained by using the Laida criterion, and the polishing end point is determined according to the time when the inflection point appears (CN201910881051.2); (5) ) By collecting the surface image of the asphalt polishing disc, the gray-scale co-occurrence matrix method is used to monitor the surface texture of the asphalt polishing disc, and the characteristic parameters of the gray-scale co-occurrence matrix are used to quantitatively characterize the passivation state of the asphalt polishing disc surface, so as to realize the surface of the asphalt polishing disc. On-line real-time monitoring of passivation state (CN202010930626.8); (6) by monitoring the strain value in real time during the annular polishing process to solve the friction force between the workpiece and the tool disc, further reflecting the stability of the annular polishing state of the workpiece (CN202011476851. 5).

The above-mentioned various online monitoring devices all have the problems of single test method, great influence by external interference, and low accuracy. For example, diffraction and scattering due to material uniformity problems often affect the accuracy of optical measurement. The monitoring device with high noise ratio, easy acquisition and strong applicability is used to improve the online real-time monitoring system of ultra-precision chemical mechanical planarization.

SUMMARY OF THE INVENTION

The purpose of the present invention is to design a multi-sensing, anti-interference and high-accuracy online real-time monitoring device for the existing ultra-precision chemical-mechanical planarization on-line monitoring devices, such as single testing means, great influence by external interference, and low precision. monitoring device.

In order to achieve the above object, the present invention proposes a monitoring device for dynamic perception of single-sided chemical mechanical planarization processing, which is installed on a single-sided grinding and polishing machine to realize online real-time monitoring of the state of the workpiece during the planar smoothing process. :include

Sensor unit: contains at least two kinds of sensors, and collects monitoring signals for joint monitoring;

Signal conditioning unit: amplifying, attenuating and/or spectrally filtering the collected signal amplitude;

Data acquisition unit: convert the analog signal processed by the signal conditioning unit into a digital signal, and perform digital filtering, noise reduction, and data preprocessing for storage;

The processor unit: performs pattern recognition on the data preprocessed by the data acquisition unit to realize the visualization of the workpiece machining dynamic indicators, and the dynamic indicators include material removal rate, surface roughness, and damage layer thickness.

Further, a sensor fixture is also included for fixing the workpiece.

Further, the sensor unit and the signal conditioning unit are connected by a precision signal slip ring or a toroidal transformer fixed on the sensor fixture;

The sensor unit is connected to the rotating parts of the precision signal slip ring or toroidal transformer through a line to ensure that the workpiece, the sensor fixture and the sensor unit rotate synchronously; the signal conditioning unit is connected to the precision signal through a line Static parts of signal slip rings or toroidal transformers that do not rotate.

Further, the sensor unit is a combination of two or more of an acoustic sensor, a force sensor, and a temperature sensor.

Further, the sensor layout includes grid, wheel, multi-ring, cross, star or random.

Further, the data sampling unit is composed of one or both of an analog signal amplifier, an attenuator, and a filter.

Further, the data acquisition unit and the processor unit are connected in a wired or wireless manner; the wired manner includes USB and a network port; and the wireless manner includes WIFI and Bluetooth.

Further, the pattern recognition process of the processor unit includes multiple simultaneous compression of data, standardized coding, time-frequency conversion, feature extraction, and aggregation classification, identifying the macro-state features and micro-state features of the workpiece, and realizing the material removal rate, Visualization of surface roughness and/or damage layer thickness.

Further, the macro-state features include slump, fracture, tilt, and removal; the micro-state features include sliding friction, yielding, elastic-plastic deformation, microfracture, and surface/subsurface damage.

The beneficial effects of the present invention are:

The present invention conducts online real-time monitoring of the workpiece state during the single-sided chemical-mechanical planarization process by combining two or more of the acoustic sensor, the force sensor and the temperature sensor, collects various physical quantities of sound, force and heat, and conducts modulus ( A/D) conversion and data processing; through the pattern recognition of the workpiece status characteristic signal, the processing technology conditions (abrasive pad, grinding fluid, load, speed, etc.) and processing status (high efficiency, normal, low efficiency, failure, grinding end point, etc.) are established etc.) mapping relationship, and then realize the prediction of processing indicators (material removal rate, surface roughness, thickness of damaged layer, etc.), and can reverse the process parameters according to processing requirements.

It breaks away from the traditional offline mode of processing-testing-reprocessing-retesting, simplifies the process flow, reduces errors such as clamping and testing, improves processing and manufacturing efficiency, meets the requirements of large-scale uninterrupted processing of enterprises, and conforms to industry Requirements for digitalization, intelligence and sustainable development.

Description of drawings

FIG. 1 is a schematic view of the working structure of the monitoring device for dynamic sensing of single-sided chemical mechanical planarization processing of the present invention.

Among them: 1-sensor unit, 2-signal conditioning unit, 3-data acquisition unit, 4-processor unit, 5-sensor fixture.

Detailed ways

The present invention will be further described below in conjunction with the examples.

The technical scheme of the present invention is: as shown in FIG. 1 , a monitoring device for dynamic perception of single-sided chemical mechanical planarization processing is installed on a single-sided grinding and polishing machine, which can realize real-time online monitoring of the workpiece state during the plane finishing process. monitor. The device consists of a sensor fixture, a sensor unit, a signal conditioning unit, a data acquisition unit, and a processor unit. The back of the workpiece is fixed under the sensor fixture, the sensor unit is fixed by the fixture and completely fits the back of the workpiece, the sensor unit is connected with the signal conditioning unit, the signal conditioning unit is connected with the data acquisition unit, and the data acquisition unit is connected with the control unit.

Sensor fixture for holding workpiece, sensor unit and signal conditioning unit. The sensor unit is connected to the precision signal slip ring or the rotating parts of the toroidal transformer through the line to ensure that the workpiece, the sensor fixture and the sensor unit rotate synchronously; the signal conditioning unit is connected to the precision signal slip ring or the stationary part of the toroidal transformer through the line, and does not rotate.

The sensor unit consists of two or more combinations of acoustic sensor, force sensor, and temperature sensor; the sensor layout includes grid, wheel, multi-ring, cross, star, and random. The sensor unit is used to collect the physical quantities during the ultra-precision chemical-mechanical planarization process and convert them into analog signals.

The signal conditioner unit is composed of one or both of an analog signal amplifier or an attenuator and a filter, which can amplify, attenuate and/or filter the spectrum of the initially collected signal.

The data acquisition unit is used to convert the analog signal processed by the signal conditioning unit into a digital signal, and perform digital filtering, noise reduction, and preprocessing of the stored data.

The processor unit is used to dump the data preprocessed by the data acquisition unit and perform data signal processing of standardization and sparseness, time-frequency conversion, feature extraction, and modal recognition, so that the macroscopic state of the workpiece such as slump and fracture can be changed. Visualization of micro-states such as damage and micro-cracks, prediction of material removal rate and surface roughness, and then processing states such as high efficiency, normal, low efficiency, failure, and end point are obtained.

The data acquisition unit and the processor unit are connected by wire or wireless. The wired method includes USB and network port, and the wireless method includes WIFI and Bluetooth.

The working principle of the present invention is:

During the ultra-precision chemical-mechanical flattening process, the physical quantity emitted by the workpiece is collected and converted into an electrical signal by the sensor in direct contact with it. ) conversion and digital filtering, noise reduction, preprocessing of the stored data, and the preprocessed data is transferred to the computer for standardization and sparseness, time-frequency conversion, feature extraction, and modal recognition. Visualize macro-states such as fractures and micro-states such as scratches and micro-cracks, predict material removal rate and surface roughness, and then obtain processing states such as high efficiency, normal, low efficiency, failure, and end point.

Example 1:

For example, in the process of grinding sapphire with free abrasives, the acoustic emission sensor and the three-dimensional force sensor arranged in a cross shape on the sensor fixture respectively collect the physical quantities of stress waves generated by abrasive rolling and scribing the workpiece surface and convert them into analog signals of sound and force. , through the signal conditioner to amplify the perceived analog signal and filter the clutter, the 10M sampling rate, 16-bit channel data collector collects the adjusted analog signal and converts it into a digital signal, and then through digital filtering, Noise reduction, preprocessing of stored data, transfer of preprocessed data to computer and data signal processing for standardization and sparseness, time-frequency conversion, feature extraction, modal identification, and display of acoustic emission time-frequency-intensity curve and time - Friction coefficient curve, the time-material removal rate curve is drawn by the weighing method in different periods, and the mathematical relationship between the acoustic emission signal intensity and friction coefficient and the abrasive material removal rate is established. Perceive, estimate the sapphire material removal rate during free abrasive grinding.

Example 2:

For example, in the process of grinding silicon carbide with bonded abrasives, the acoustic emission sensor and temperature sensor arranged in a multi-ring layout on the sensor fixture respectively collect the stress waves and frictional thermal physical quantities generated on the surface of the workpiece due to the frictional mechanical action of the abrasive pads. The hot analog signal is amplified, amplified and filtered by the signal conditioner. The 15M sampling rate, 19-bit channel data collector collects the adjusted analog signal and converts it into a digital signal. After digital filtering, noise reduction, and preprocessing of stored data, the preprocessed data is transferred to the computer and subjected to data signal processing of standardization and sparseness, time-frequency conversion, feature extraction, and modal identification, showing that the acoustic emission time-frequency- Intensity curve and time-temperature curve, the time-surface roughness curve is drawn by the time-based white light interferometry surface topography method, and the mathematical relationship between the acoustic emission signal intensity and temperature and the surface roughness of the workpiece is established. The perception of signal intensity and temperature can estimate the surface roughness of silicon carbide during the grinding process of fixed abrasives. In addition, when there are signal characteristics corresponding to defects such as workpiece slumps, scratches, and micro-cracks, the operator can appropriately carry out the grinding process. operation to improve the surface quality of the workpiece.

Example 3:

For example, in the chemical mechanical polishing of silicon wafer from copper layer to tantalum layer, the three-dimensional force sensor and temperature sensor on the sensor fixture in a star-shaped layout respectively collect the stress wave and frictional heat generated on the surface of the workpiece and convert them into sound and heat. The analog signal is amplified, amplified and filtered by the signal conditioner. The 20M sampling rate, 15-bit channel data collector collects the adjusted analog signal and converts it into a digital signal. Filtering, noise reduction, stored data preprocessing, the preprocessed data is transferred to the computer and subjected to data signal processing for standardization and sparseness, time-frequency conversion, feature extraction, and modal identification, showing the time-friction coefficient curve and time -Temperature curve, qualitatively analyze the type of wafer based on the friction coefficient and temperature signal, monitor the friction coefficient and temperature of the different layers of the wafer and its softened layer surface, determine the polishing end point time, and prevent "under-polishing" or "over-polishing".

The preferred embodiment of the present invention has been described, and various changes or modifications can be made by those skilled in the art without departing from the scope of the present invention. The parts not involved in the present invention are the same as or can be implemented by using the prior art.

Claims (9)

1. The utility model provides a monitoring devices for single face chemical mechanical planarization processing developments perception, installs on single face grinding and polishing machine, realizes the online real-time supervision of plane finishing in-process work piece state which characterized in that: comprises that

A sensor unit: the system comprises at least two sensors, a monitoring signal acquisition unit and a monitoring signal acquisition unit, wherein the at least two sensors are used for acquiring monitoring signals for joint monitoring;

a signal conditioning unit: amplifying, attenuating and/or spectrally filtering the acquired signal amplitude;

a data acquisition unit: converting the analog signals processed by the signal conditioning unit into digital signals, and performing data preprocessing of digital filtering, noise reduction and storage;

a processor unit: and carrying out pattern recognition on the data preprocessed by the data acquisition unit to realize the visualization of dynamic indexes of workpiece processing, wherein the dynamic indexes comprise material removal rate, surface roughness and damage layer thickness.

2. The apparatus for monitoring dynamic sensing of single-sided chemical mechanical planarization processing of claim 1, wherein: the sensor fixture is used for fixing the workpiece.

3. The apparatus for monitoring dynamic sensing of single-sided chemical mechanical planarization processing of claim 2, wherein: the sensor unit and the signal adjusting unit are connected by a precise signal slip ring or a ring transformer fixed on the sensor clamp; the sensor unit is connected with a rotating part of the precision signal slip ring or the annular transformer through a circuit, so that the workpiece, the sensor clamp and the sensor unit can synchronously rotate; and the signal regulating unit is connected with the static part of the precision signal slip ring or the annular transformer through a line and does not rotate.

4. The apparatus for monitoring dynamic sensing of single-sided chemical mechanical planarization processing of claim 1, wherein: the sensor unit is a combination of two or more of an acoustic sensor, a force sensor and a temperature sensor.

5. The apparatus for monitoring dynamic sensing of single-sided chemical mechanical planarization processing of claim 1, wherein: the sensor layout mode comprises grid, wheel, multiple rings, cross, star or random.

6. The apparatus for monitoring dynamic sensing of single-sided chemical mechanical planarization processing of claim 1, wherein: the data sampling unit is composed of one or two of an analog signal amplifier, an attenuator and a filter.

7. The apparatus for monitoring dynamic sensing of single-sided chemical mechanical planarization processing of claim 1, wherein: the data acquisition unit is connected with the processor unit in a wired or wireless mode; wired modes include USB and network port; the wireless mode comprises WIFI and Bluetooth.

8. The apparatus for monitoring dynamic perception of single-sided chemical mechanical planarization processing of claim 1, wherein: the mode identification process of the processor unit comprises the steps of carrying out data multiple synchronous compression, standardized coding, time-frequency conversion, feature extraction and aggregation classification, identifying the macro-state features and the micro-state features of the workpiece, and realizing visualization of the material removal rate, the surface roughness and/or the damage layer thickness.

9. The apparatus for monitoring dynamic sensing of single-sided chemical mechanical planarization processing of claim 8, wherein: the macroscopic state features comprise edge collapse, fracture, inclination and removal amount; the micro-regime features include sliding friction, yield, elasto-plastic deformation, micro-fracture, surface/sub-surface damage.

Images