2D Optical Detector With Shift Register Arrays
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional optical metrology systems face challenges in capturing a wide range of physical defects on semiconductor wafers with high throughput due to slow data acquisition and resolution limitations, especially when dealing with multiple sets of system parameters like polar angles of incidence, wavelengths, and polarization states.
Innovation Solution
A two-dimensional detector with multiple independent linear arrays of shift registers subdivides the photosensitive area into stripes, allowing high-resolution detection of one optical property along each stripe and low-resolution detection of another property across several stripes, while minimizing signal contamination through opaque materials or optical shutters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If a conventional two-dimensional detector reads one row at a time through a single shift register, then the detector can maintain simple structure, but the readout speed becomes prohibitively slow (100-200 milliseconds for 1024×38 pixels)
Solution Approach 1:
The detector is divided into multiple independent segments, each with its own shift register. The active area is subdivided into multiple photosensitive stripes, with each stripe having an associated linear array of shift register elements. This segmentation allows parallel readout of multiple rows simultaneously, dramatically increasing readout speed while keeping each segment's structure simple and manageable.
Solution Approach 2:
The patent transitions from a single-dimensional sequential readout approach to a two-dimensional parallel readout architecture. By introducing multiple shift registers arranged in arrays corresponding to multiple photosensitive stripes, the system exploits the second dimension (across multiple stripes) to enable simultaneous readout operations, thereby achieving high-speed readout without excessive complexity.
2Quantity of substance
If the detector is exposed to light during readout to maintain continuous charge accumulation, then photon flux is maximized, but signal contamination occurs due to mixing of information across rows
Solution Approach 1:
The detector active area is segmented into multiple independent photosensitive stripes, each with its own shift register array. This segmentation isolates the charge accumulation and readout processes for each stripe, preventing signal mixing between different rows or stripes. Each stripe can be read out independently while maintaining continuous exposure, thus preserving signal integrity without sacrificing photon flux.
Solution Approach 2:
The patent introduces opaque materials or optical shutters as intermediaries to control light exposure during readout. These intermediaries can selectively block light to specific regions (such as shift register areas) while allowing continuous exposure to the photosensitive stripes, thereby preventing signal contamination in non-photosensitive areas without reducing the photon flux to the active detection regions.
3Loss of information
If conventional shutters are used to block light during readout, then signal contamination is prevented, but the shutter inertia is too high to follow millisecond integration times
Solution Approach 1:
The patent extracts the light-blocking function from a mechanical shutter system and relocates it to the detector structure itself through opaque materials integrated into the shift register regions. This extraction eliminates the need for fast mechanical shutters by building the light-blocking capability directly into the detector architecture, achieving instantaneous response without mechanical inertia limitations.
Solution Approach 2:
The patent uses opaque materials as static intermediaries that permanently block light to non-photosensitive regions. These materials serve as fixed mediators between the light source and the shift register areas, preventing signal contamination without requiring any moving parts or fast response mechanisms. The light-blocking function is achieved through the inherent optical properties of the opaque materials rather than dynamic shutter control.
4Speed
If vertical binning mode is used to transfer all rows to a single shift register, then readout speed improves (M+N clock cycles), but vertical resolution is completely lost
Solution Approach 1:
The detector is segmented into multiple photosensitive stripes, each maintaining its own linear array of shift register elements. This segmentation preserves the vertical dimension information for each stripe while enabling parallel readout. The vertical resolution within each stripe is maintained because each stripe's pixels are independently addressed and transferred to their dedicated shift register, avoiding the resolution loss inherent in vertical binning.
Solution Approach 2:
The patent resolves the contradiction by utilizing the horizontal dimension (across multiple stripes) to achieve parallelism and speed improvement, while preserving the vertical dimension within each stripe for resolution maintenance. This two-dimensional parallel architecture allows simultaneous readout of multiple stripes (improving speed) while keeping the vertical pixel structure intact within each stripe (preserving resolution).
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This configuration enhances the throughput of metrology systems by enabling simultaneous detection of multiple optical properties with different resolution requirements, maintaining high sensitivity and reducing signal contamination, thus improving the efficiency of defect detection on semiconductor wafers.
Implementation Method 1
A two dimensional detector (e.g., a charge coupled device camera) is employed to resolve two beam properties
Implementation Method 2
State of the art two dimensional detectors (e.g., back-thinned CCD detectors) allow for a two dimensional readout of the active area by shifting pixels vertically to a shift register at the bottom of the active area
Data Source
AI summary
Methods and systems for enhancing the throughput of a metrology system generating measurement signals based on at least two different optical properties of the illumination light are presented. A detector having a two dimensional photosensitive area is subdivided into multiple photosensitive stripes by multiple, independent linear arrays of shift register elements located within the photosensitive area. Charge transfer from pixels within each stripe is directed to a distinct linear array of shift register elements. Each photosensitive stripe is able to resolve an optical property dispersed across the length of each stripe with relatively high resolution. In addition, the detector is able to resolve another optical property dispersed across several photosensitive stripes in a direction orthogonal to each linear array of shift registers at a relatively low resolution.