Abstract: To provide more test data during the manufacture of non-volatile memories and other integrated circuits, machine learning is used to generate virtual test values. Virtual test results are interpolated for one set of tests for devices on which the test is not performed based on correlations with other sets of tests.

Abstract: To provide more test data during the manufacture of non-volatile memories and other integrated circuits, machine learning is used to generate virtual test values. Virtual test results are interpolated for one set of tests for devices on which the test is not performed based on correlations with other sets of tests. In one example, machine learning determines a correlation study between bad block values determined at die sort and photo-limited yield (PLY) values determined inline during processing. The correlation can be applied to interpolate virtual inline PLY data for all of the memory dies, allowing for more rapid feedback on the processing parameters for manufacturing the memory dies and making the manufacturing process more efficient and accurate. In another set of embodiments, the machine learning is used to extrapolate limited metrology (e.g., critical dimension) test data to all of the memory die through interpolated virtual metrology data values.

Abstract: To provide more test data during the manufacture of non-volatile memories and other integrated circuits, machine learning is used to generate virtual test values. Virtual test results are interpolated for one set of tests for devices on which the test is not performed based on correlations with other sets of tests. In one example, machine learning determines a correlation study between bad block values determined at die sort and photo-limited yield (PLY) values determined inline during processing. The correlation can be applied to interpolate virtual inline PLY data for all of the memory dies, allowing for more rapid feedback on the processing parameters for manufacturing the memory dies and making the manufacturing process more efficient and accurate. In another set of embodiments, the machine learning is used to extrapolate limited metrology (e.g., critical dimension) test data to all of the memory die through interpolated virtual metrology data values.

Abstract: A three-dimensional memory device includes a first-tier alternating stack of first insulating layers and first electrically conductive layers, a second-tier alternating stack of second insulating layers and second electrically conductive layers overlying the first-tier alternating stack, memory openings vertically extending through the second-tier alternating stack and the first-tier alternating stack, memory opening fill structures located in the memory openings, wherein each of the memory opening fill structures includes a respective vertical stack of memory elements and a respective vertical semiconductor channel including a respective portion of a semiconductor material, and hybrid support structures vertically extending at least through a respective subset of layers within the first-tier alternating stack.

Abstract: To provide more test data during the manufacture of non-volatile memories and other integrated circuits, machine learning is used to generate virtual test values. Virtual test results are interpolated for one set of tests for devices on which the test is not performed based on correlations with other sets of tests. In one example, machine learning determines a correlation study between bad block values determined at die sort and photo-limited yield (PLY) values determined inline during processing. The correlation can be applied to interpolate virtual inline PLY data for all of the memory dies, allowing for more rapid feedback on the processing parameters for manufacturing the memory dies and making the manufacturing process more efficient and accurate. In another set of embodiments, the machine learning is used to extrapolate limited metrology (e.g., critical dimension) test data to all of the memory die through interpolated virtual metrology data values.

Abstract: To provide more test data during the manufacture of non-volatile memories and other integrated circuits, machine learning is used to generate virtual test values. Virtual test results are interpolated for one set of tests for devices on which the test is not performed based on correlations with other sets of tests. In one example, machine learning determines a correlation study between bad block values determined at die sort and photo-limited yield (PLY) values determined inline during processing. The correlation can be applied to interpolate virtual inline PLY data for all of the memory dies, allowing for more rapid feedback on the processing parameters for manufacturing the memory dies and making the manufacturing process more efficient and accurate. In another set of embodiments, the machine learning is used to extrapolate limited metrology (e.g., critical dimension) test data to all of the memory die through interpolated virtual metrology data values.

Abstract: To provide more test data during the manufacture of non-volatile memories and other integrated circuits, machine learning is used to generate virtual test values. Virtual test results are interpolated for one set of tests for devices on which the test is not performed based on correlations with other sets of tests. In one example, machine learning determines a correlation study between bad block values determined at die sort and photo-limited yield (PLY) values determined inline during processing. The correlation can be applied to interpolate virtual inline PLY data for all of the memory dies, allowing for more rapid feedback on the processing parameters for manufacturing the memory dies and making the manufacturing process more efficient and accurate. In another set of embodiments, the machine learning is used to extrapolate limited metrology (e.g., critical dimension) test data to all of the memory die through interpolated virtual metrology data values.

Abstract: To provide more test data during the manufacture of non-volatile memories and other integrated circuits, machine learning is used to generate virtual test values. Virtual test results are interpolated for one set of tests for devices on which the test is not performed based on correlations with other sets of tests. In one example, machine learning determines a correlation study between bad block values determined at die sort and photo-limited yield (PLY) values determined inline during processing. The correlation can be applied to interpolate virtual inline PLY data for all of the memory dies, allowing for more rapid feedback on the processing parameters for manufacturing the memory dies and making the manufacturing process more efficient and accurate. In another set of embodiments, the machine learning is used to extrapolate limited metrology (e.g., critical dimension) test data to all of the memory die through interpolated virtual metrology data values.

Abstract: To provide more test data during the manufacture of non-volatile memories and other integrated circuits, machine learning is used to generate virtual test values. Virtual test results are interpolated for one set of tests for devices on which the test is not performed based on correlations with other sets of tests. In one example, machine learning determines a correlation study between bad block values determined at die sort and photo-limited yield (PLY) values determined inline during processing. The correlation can be applied to interpolate virtual inline PLY data for all of the memory dies, allowing for more rapid feedback on the processing parameters for manufacturing the memory dies and making the manufacturing process more efficient and accurate. In another set of embodiments, the machine learning is used to extrapolate limited metrology (e.g., critical dimension) test data to all of the memory die through interpolated virtual metrology data values.

Abstract: A multi-tier three-dimensional memory array includes multiple alternating stacks of insulating layers and electrically conductive layers that are vertically stacked. Memory stack structures including memory films and semiconductor channels extend through the alternating stacks. The alternating stacks are formed as alternating stacks of insulating layers and sacrificial material layers, and are subsequently modified by replacing the sacrificial material layers with electrically conductive layers. Structural support during replacement of the sacrificial material layers with the electrically conductive layers is provided by the memory stack structures and dielectric support pillar structures. The dielectric support pillar structures may be formed only for a first-tier structure including a first-tier alternating stack of first insulating layers and first spacer material layers, or may vertically extend over multiple tiers.

Abstract: A multi-tier three-dimensional memory array includes multiple alternating stacks of insulating layers and electrically conductive layers that are vertically stacked. Memory stack structures including memory films and semiconductor channels extend through the alternating stacks. The alternating stacks are formed as alternating stacks of insulating layers and sacrificial material layers, and are subsequently modified by replacing the sacrificial material layers with electrically conductive layers. Structural support during replacement of the sacrificial material layers with the electrically conductive layers is provided by the memory stack structures and dielectric support pillar structures. The dielectric support pillar structures may be formed only for a first-tier structure including a first-tier alternating stack of first insulating layers and first spacer material layers, or may vertically extend over multiple tiers.

Abstract: Data stored in a plurality of charge storage elements in a three-dimensional memory device can be read with high speed by measuring a majority charge carrier current passing through a vertical semiconductor channel. A memory film is provided in a memory opening extending through an alternating stack of insulating layers and electrically conductive layers. A set of doped semiconductor material regions having a doping of a first conductivity type can collectively extend continuously from underneath a top surface of a substrate through the memory film to a level of a topmost layer of the alternating stack. A well contact via structure can contact a doped contact region, which is an element of the set of doped semiconductor material regions. A p-n junction is provided within each memory opening between the doped vertical semiconductor channel and an upper doped semiconductor region having a doping of a second conductivity type.

Abstract: A memory opening can be formed through an alternating stack of insulating layers and sacrificial material layers provided over a substrate. Annular etch stop material portions are provided at each level of the sacrificial material layers around the memory opening. The annular etch stop material portions can be formed by conversion of surface portions of the sacrificial material layers into dielectric material portion, or by recessing the sacrificial material layers around the memory opening and filling indentations around the memory opening. After formation of a memory stack structure, the sacrificial material layers are removed from the backside. The annular etch stop material portions are at least partially converted to form charge trapping material portions. Vertical isolation of the charge trapping material portions among one another around the memory stack structure minimizes leakage between the charge trapping material portions located at different word line levels.

Abstract: A memory opening can be formed through an alternating stack of insulating layers and sacrificial material layers provided over a substrate. Annular etch stop material portions are provided at each level of the sacrificial material layers around the memory opening. The annular etch stop material portions can be formed by conversion of surface portions of the sacrificial material layers into dielectric material portion, or by recessing the sacrificial material layers around the memory opening and filling indentations around the memory opening. After formation of a memory stack structure, the sacrificial material layers are removed from the backside. The annular etch stop material portions are at least partially converted to form charge trapping material portions. Vertical isolation of the charge trapping material portions among one another around the memory stack structure minimizes leakage between the charge trapping material portions located at different word line levels.

Abstract: Disclosed is a method for producing a hollow fiber carbon membrane using a hollow fiber carbon membrane-forming material by means of a dry method or dry-wet method comprising a spinning step, a drying step, an infusibilization step, and a carbonization step as basic steps; wherein in the infusibilization step, heat treatment is performed at least two times at different temperatures, with the second temperature being higher than the first temperature. It is preferable that the second heat treatment is performed directly after the first treatment is performed, without once cooling to room temperature. The obtained hollow fiber carbon membrane has improved permeability, without reducing its separation coefficient.

Abstract: Threshold voltage shift due to programming of a neighboring memory element can be reduced or suppressed by forming a compositionally modulated charge storage layer in a three-dimensional memory device. The compositionally modulated charge storage layer can be formed by providing an oxygen-containing dielectric silicon compound layer outside a tunneling dielectric layer, and subsequently nitriding portions of the oxygen-containing dielectric silicon compound layer only at levels of the control gate electrodes. An alternating stack of sacrificial material layers and insulating layers can be employed to form a memory stack structure therethrough.

Abstract: Corner rounding of electrically conductive layers in a replacement electrode integration scheme can be alleviated by employing compositionally modulated sacrificial material layers. An alternating stack of insulating layers and compositionally modulated sacrificial material layers can be formed over a substrate. Each of the compositionally modulated sacrificial material layers has a vertical modulation of material composition such that each compositionally modulated sacrificial material layer provides greater resistance to conversion into a silicon-oxide-containing material at upper and lower portions thereof than at a middle portion thereof during a subsequent oxidation process. Bird's beak features can be formed with lesser dimensions, and electrically conductive layers formed by replacement of remaining portions of the sacrificial material layers with a conductive material can have less corner rounding.

Abstract: Data stored in a plurality of charge storage elements in a three-dimensional memory device can be read with high speed by measuring a majority charge carrier current passing through a vertical semiconductor channel. A memory film is provided in a memory opening extending through an alternating stack of insulating layers and electrically conductive layers. A set of doped semiconductor material regions having a doping of a first conductivity type can collectively extend continuously from underneath a top surface of a substrate through the memory film to a level of a topmost layer of the alternating stack. A well contact via structure can contact a doped contact region, which is an element of the set of doped semiconductor material regions. A p-n junction is provided within each memory opening between the doped vertical semiconductor channel and an upper doped semiconductor region having a doping of a second conductivity type.

Abstract: A hollow fiber membrane module comprising a bundling tube including a body part and head parts placed at both terminals of the body part and O-rings with which the bundling tube is fitted into a housing being attached to outer circumferential surfaces of the head parts, wherein at least one of the head parts are divided into two in a place closer to a terminal of the bundling tube than places where the O-rings are attached while maintaining a cylindrical shape of the head part, and the divided parts have a slidable structure while maintaining a coupled condition therebetween that forms the cylindrical shape. The hollow fiber membrane module is capable of preventing damage effectively on a hollow fiber membrane or the bundling tube due to expansion or contraction of the hollow fiber membrane to be caused by change of an external environment such as temperature.

Abstract: Disclosed is a method for producing a hollow fiber carbon membrane using a hollow fiber carbon membrane-forming material by means of a dry method or dry-wet method comprising a spinning step, a drying step, an infusibilization step, and a carbonization step as basic steps; wherein in the infusibilization step, heat treatment is performed at least two times at different temperatures, with the second temperature being higher than the first temperature. It is preferable that the second heat treatment is performed directly after the first treatment is performed, without once cooling to room temperature. The obtained hollow fiber carbon membrane has improved permeability, without reducing its separation coefficient.