Image sensor including optically switchable magnetic tunnel junction array

Abstract

The present invention relates to an image sensor including an optically switchable magnetic tunnel junction array, and discloses an image sensor including an optically switchable magnetic tunnel junction (MTJ) array arranged in columns and rows. The image sensor has a first circuit of transparent conductive material and a second circuit of conductive material. Each first circuit contacts the free layer of the MTJ in the corresponding row. Each second circuit is electrically connected to the fixed layer MTJ in the corresponding column. The first circuits are exposed to radiation simultaneously. The first circuit and the second circuit can be selectively biased. In a global reset operation, the bias condition causes all MTJs to be switched to an antiparallel state. In a global sensing operation, the bias condition causes the MTJs to be switched to a parallel state based on the intensity of radiation received at those portions of the first circuit in contact with the MTJs. In a selective read operation, the bias condition causes the data values ​​stored in the MTJs to be read.

Description

Technical Field

[0001] The present invention relates to sensors and, more particularly, to embodiments of image sensors including optically switchable magnetic tunnel junction arrays (MTJs). Background Art

[0002] Conventional image sensors are based on complementary metal oxide semiconductor (CMOS). Specifically, they include an array of CMOS-based pixels. An exemplary CMOS-based pixel includes a photodiode (e.g., a PIN photodiode) and at least three transistors including a reset transistor (e.g., a P-type field effect transistor (PFET)), an amplifier transistor (e.g., an N-type field effect transistor (NFET)), and an access or select transistor (e.g., another NFET). One disadvantage of such CMOS-based image sensors is that the pixels are not radiation-hardened. Specifically, radiation can cause stored data values ​​to leak. Another disadvantage is that due to the number of devices required for each pixel (e.g., a photodiode and at least three transistors) and the relatively large size of the photodiodes, the pixel array consumes a large amount of chip area. Summary of the Invention

[0003] In general, disclosed herein are embodiments of an image sensor comprising an array of optically switchable magnetic tunnel junctions (MTJs) arranged in columns and rows. The MTJs may include an insulator layer between a free layer and a pinned layer. The image sensor may further include a first circuit. The first circuit may be made, at least in part, of a transparent conductive material. Each first circuit may span a corresponding row of MTJs, thereby extending and contacting the free layer of each MTJ in the corresponding row. The image sensor may also include a second circuit. The second circuit may be made of a conductive material. Each second circuit may be electrically connected to the pinned layer of each MTJ in the corresponding column. The MTJs in such an image sensor may simultaneously store image data in response to specific bias conditions on the first circuit and the second circuit and simultaneous exposure of the first circuit to radiation.

[0004] For example, an image sensor embodiment disclosed herein may include peripheral circuitry connected to first and second lines and a controller in communication with the peripheral circuitry. In response to a control signal from the controller, the peripheral circuitry may apply a first set of bias conditions to the first and second lines to perform a global reset operation with respect to the magnetic tunnel junctions (MTJs), and more specifically, to cause all MJs to simultaneously store a first data value (e.g., a "0" data value). In response to a different control signal from the controller, the peripheral circuitry may apply a second set of bias conditions to the first and second lines to perform a global sensing operation, and more specifically, to cause the MTJs to capture and store image data. Specifically, the second bias conditions may be applied by the peripheral circuitry to the first and second lines such that, in response to simultaneous exposure of all first lines to radiation, switching from the first data value to the second data value occurs only in a given magnetic tunnel junction when the actual intensity level of the radiation received in the immediately adjacent first line portion is greater than a threshold intensity level. Finally, in response to another different control signal from the controller, the peripheral circuitry may apply a third set of bias conditions to select the first and second lines to enable a selective read operation, and more specifically, to read out the data value stored in a given MTJ.

[0005] It should be noted that different image sensor embodiments disclosed herein may include different types of MTJs.

[0006] For example, some image sensor embodiments disclosed herein may include an array of spin Hall effect magnetic tunnel junctions (SHE-MTJs). Each SHE-MTJ may include an insulator layer between a free layer and a pinned layer. These image sensor embodiments may further include a first circuit. The first circuit may include a transparent conductive spin Hall effect material (e.g., a heavy metal or graphene with a thickness less than 3 nm, or some other suitable transparent conductive material exhibiting a spin Hall effect material). Each first circuit may span and contact the free layer of each SHE-MTJ in a corresponding row. These image sensor embodiments may also include a second circuit. The second circuit may include a conductive material. Each second circuit may be electrically connected to the pinned layer of each magnetic tunnel junction in a corresponding column.

[0007] Other image sensor embodiments disclosed herein may include a spin transfer torque magnetic tunnel junction (STT-MTJ) array. Each STT-MTJ may include an insulator layer between a free layer and a pinned layer. These image sensor embodiments may further include a first circuit. The first circuit may include a transparent conductive material (e.g., indium tin oxide or some other suitable transparent conductive material). Each first circuit may span and contact the free layer of each STT-MTJ in a corresponding row. These image sensor embodiments may also include a second circuit. The second circuit may include a conductive material. Each second circuit may be electrically connected to the pinned layer of each magnetic tunnel junction in a corresponding column.

[0008] Depending on the type of MTJ used in the MTJ array of the image sensor (e.g., SHE-MTJ or STT-MTJ), the bias conditions applied to the first and second lines during the above-mentioned global reset, global sensing, and selective read operations will be different (as further discussed in the detailed description section). BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The present invention will be better understood from the following detailed description with reference to the accompanying drawings, which are not necessarily drawn to scale, and in which:

[0010] Figure 1A 、 Figure 1B and Figure 1C Schematic diagrams are provided to illustrate various image sensor embodiments, each of which includes an optically switchable spin transfer torque (STT) magnetic tunnel junction (MTJ);

[0011] Figure 2A 、 Figure 2B and Figure 2C To explain separately Figure 1A 、 Figure 1B and Figure 1C Cross-sectional views of different configurations of the first line employed above the STT-MTJ in any one of the illustrated image sensor embodiments;

[0012] Figure 3 For illustration purposes, Figure 1B and Figure 1C A cross-sectional view of an exemplary on-demand radiation shield in the illustrated image sensor embodiment;

[0013] Figure 4A 、 Figure 4B and Figure 4C The following table shows the reset, sensing and read operations respectively. Figure 1A 、 Figure 1B and Figure 1C An exemplary STT-MTJ in any image sensor embodiment of FIG.

[0014] Figure 5A 、 Figure 5B and Figure 5C Schematic diagrams of image sensor embodiments are respectively illustrated, each embodiment including an optical spin Hall effect (SHE) MTJ.

[0015] Figure 6A 、 Figure 6B and Figure 6C Explain separately in Figure 5A 、 Figure 5B and Figure 5C Cross-sectional views of different configurations of the first line above the SHE-MTJ employed in any of the image sensor embodiments shown; and

[0016] Figure 7A , Figure 7B(1), Figure 7B(2) and Figure 7C The following table shows the reset, sensing and read operations. Figure 5A 、 Figure 5B 、 Figure 5C An exemplary SHE-MTJ in any image sensor embodiment of FIG. DETAILED DESCRIPTION

[0017] As mentioned above, conventional image sensors are based on complementary metal oxide semiconductor (CMOS) and, more specifically, include arrays of CMOS-based pixels.

[0018] An exemplary CMOS-based pixel includes a photodiode (e.g., a PIN photodiode) and at least three transistors including a reset transistor (e.g., a P-type field effect transistor (PFET)), an amplifier transistor (e.g., an N-type field effect transistor (NFET)), and an access or select transistor (e.g., another NFET). In such a pixel, the reset transistor and the photodiode are electrically connected in series between a positive voltage rail and a ground rail. A sensing node is located at the junction between the photodiode and the reset transistor. The amplifier transistor and the access transistor are electrically connected in series between another positive voltage rail and a bit line. The gate of the reset transistor is controlled by a reset signal (RST). The gate of the amplifier transistor is electrically connected to the sensing node, and the gate of the access transistor is electrically connected to the word line. In the array, all pixels in the same column are electrically connected to the same bit line, and all pixels in the same row are electrically connected to the same word line.

[0019] The operation of the image sensor is as follows. During the precharge operation, RST is switched to logic "0" to turn on the reset transistor of each pixel, thereby precharging the sense node. During the sensing operation (also known as the image capture operation), RST is switched to logic "1" to turn off the reset transistor of each pixel, and the photodiode is exposed to light. In each pixel, depending on the intensity level of the light received at the pixel's photodiode, the voltage level on the pixel's sense node will remain at the same precharge voltage level so that the amplifier transistor remains in the on state, or will be discharged to a lower voltage level so that the amplifier transistor switches to the off state. In other words, a sensed data value will be generated on the pixel's sense node. During the read operation, in order to read the stored data value from the selected pixel, the word line of the row containing the selected pixel will be switched to a high voltage level, thereby turning on the access transistor, and any change in the voltage level (or alternatively, current level) on the bit line of the column containing the selected pixel can be sensed (e.g., by a sense amplifier, a transimpedance amplifier (TIA), etc.) and can indicate the stored data value.

[0020] One drawback of such CMOS-based image sensors is that the pixels are not radiation-hardened. Specifically, radiation can cause stored data values ​​to leak. Therefore, the stored data values ​​must be quickly read from the pixels and stored in memory before such leakage occurs to avoid errors. Another drawback is that the pixel array consumes a significant amount of chip area due to the number of devices required per pixel (e.g., a photodiode and at least three transistors) and the relatively large photodiode size.

[0021] Generally speaking, various embodiments of an image sensor are disclosed herein that include an array of optically switchable magnetic tunnel junctions (MTJs) arranged in columns and rows. Each MTJ can be a back-end-of-the-line (BEOL) multilayer structure that includes an insulator layer between a free layer on a first side and a pinned layer on a second side. Each image sensor embodiment can further include a first line. The first line can include a transparent conductive material and can be adjacent to a row of MTJs. Specifically, each first line can span a corresponding row of MTJs so that it extends across and contacts the free layer of each MTJ in the corresponding row. Each image sensor embodiment can further include a second line. The second line can include a conductive material and can be electrically connected to a column of MTJs. Specifically, each second line can be electrically connected to the pinned layer of each MTJ in the corresponding column. In addition, each image sensor embodiment can be configured so that the first line can be exposed to radiation simultaneously, so that both the first line and the second line can be selectively biased to achieve global reset, global sensing, and selective read operations for the MTJs in the array without having to equip each MTJ with a transistor (e.g., an access transistor, etc.).

[0022] For the purposes of this application, a global reset operation refers to an operation in which all MTJs in an array are reset to the same state (e.g., an anti-parallel state) so that they store the same data value (e.g., "0"). A global sensing operation (also referred to herein as an image capture operation) refers to an operation in which all first lines are exposed to radiation, and depending on the intensity level of the radiation, some of the MTJs in the portion of the first line immediately adjacent to the MTJs may switch states (e.g., from an AP state to a parallel (P) state) so that they now store different data values ​​(e.g., "1"). The data values ​​stored in all MTJs after the global sensing operation are referred to herein as image data values. Therefore, the MTJs are considered to be optically switchable. A selective read operation refers to an operation in which the value of a selected pixel is read (i.e., sensed, determined, etc.), the selected pixel being a single MTJ pixel or a multi-MTJ pixel (depending on the embodiment, as described below) and corresponding to a given position of an image captured during the global sensing operation. As discussed below with respect to various embodiments, during different global reset, global sense, and selective read operations, the biasing of the first and second lines will vary depending on the type of MTJ used in the array (e.g., spin transfer torque (STT) MTJ or spin Hall effect (SHE) MTJ).

[0023] More specifically, refer to Figure 1A、 Figure 1B and Figure 1C Some image sensor embodiments 100A, 100B, and 100C disclosed herein include an array 110 of optically switchable magnetic tunnel junctions (MTJs), and in particular, optically switchable spin transfer torque (STT) MTJs 120 arranged in columns and rows. For illustrative purposes, the array 110 is shown as having four columns AD and four rows ad. However, it should be understood that the figures are not intended to be limiting, and alternatively, the array 110 may have any number of two or more columns and two or more rows of STT-MTJs 120.

[0024] Each STT MTJ 120 may be a back-end-of-the-line (BEOL) multilayer structure. That is, each STT MTJ 120 may be a multilayer structure formed in a BEOL metal layer above a front-end-of-the-line (FEOL) device (e.g., a semiconductor device, etc.) on an integrated circuit chip. Each STT-MTJ 120 may include an insulator layer 124 (also referred to as a tunnel barrier layer) between a free ferromagnetic layer 125 (also referred to herein as a free layer or data storage layer) located on a first side and a pinned ferromagnetic layer 123 (also referred to herein as a pinned layer or fixed ferromagnetic layer) located on a second side. In one exemplary embodiment, the pinned ferromagnetic layer 123 may be, for example, a gadolinium (Gd), iron (Fe), and cobalt (Co) alloy layer. The insulator layer 124 (i.e., the tunnel barrier layer) may be, for example, any one of a magnesium oxide (MgO) layer, an aluminum oxide (Al2O3) layer (also referred to as aluminum oxide), and a bismuth ferrite (BiFeO3, also referred to as BFO) layer. The free ferromagnetic layer 125 may be, for example, a cobalt (Co), iron (Fe), or boron (B) alloy layer.

[0025] These image sensor embodiments 100A, 100B, and 100C may further include first lines 126. The first lines 126 may be adjacent to the rows ad of the STT-MTJs 120, respectively. Specifically, each first line 126 may span a corresponding row of STT-MTJs, extending to and contacting the free layer 125 of each STT-MTJ 120 in the corresponding row. For example, the first lines 126 may be relatively thin and flat strips of material (also referred to as nanosheets) that are (a) entirely composed of a transparent conductive material 127 (e.g., Figure 2A (a) or (b) having a first segment and a second segment, the first segment being adjacent to the free layer 125 of the STT MTJ 120 and being composed of a transparent conductive material 127, the second segment being located on opposite sides of the first segment (e.g., such that each first segment is laterally located between two second segments) and being composed of a metal or metal alloy material 128 to increase conductivity (e.g., Figure 2B or Figure 2CIn any case, the transparent conductive material 127 may include, for example, indium tin oxide (ITO), zinc oxide (ZnO), graphene, or any other suitable transparent conductive material. In the case where the first line 126 has a first section of transparent conductive material 127 and a second section of metal or metal alloy material 128, the metal or metal alloy material 128 may be, for example, copper, aluminum, or any other suitable BEOL metal or metal alloy material. In addition, the first section of transparent conductive material 127 and the second section of metal or metal alloy material 128 may have the same thickness (e.g., Figure 2B ) or different thicknesses (eg, the second section of the metal or metal alloy material 128 may be thicker than the first section of the transparent conductive material 127) (eg, Figure 2C shown).

[0026] These image sensor embodiments 100A, 100B, and 100C can further include second lines 121 of conductive material. The second lines 121 can be, for example, conductive wires made of a metal or metal alloy (e.g., copper or any other suitable BEOL metal or metal alloy). The second lines 121 can be electrically connected (e.g., via local interconnects 122) to the MTJs in columns AD of the STT MTJs, respectively. Specifically, each second line 121 can be electrically connected to the pinning layer 123 of each STT-MTJ 120 in the corresponding column AD via the local interconnects 122.

[0027] In addition, these image sensor embodiments 100A, 100B and 100C can be configured so that the first line and the second line are selectively biased, thereby being able to perform global reset, global sensing and selective read operations of the STT-MTJ without having to configure transistors (e.g., access transistors, etc.) for each STT-MTJ.

[0028] Specifically, image sensor embodiments 100A, 100B, and 100C may further include a controller 180 and peripheral circuits 181 and 183. In response to control signals from the controller 180, the peripheral circuits 181 and 183 may cause specific biasing of the first line 126 and the second line 121 during global reset, global sense, and selective read operations, as described below. Specifically, the peripheral circuit 181 may be connected to the first line 126 of the row and may include, for example, address decoding logic and a first line driver for appropriately biasing the first line, as described below. The peripheral circuit 183 may be connected to the second line 121 of the column and may include column address decoding logic and a second line driver for appropriately biasing the second line, as described below. It is well known in the art that peripheral circuits are configured to selectively bias the lines of an array in response to control signals from a controller. Therefore, to allow the reader to focus on the salient aspects of the disclosed embodiments, the details thereof are omitted from this specification.

[0029] These image sensor embodiments 100A, 100B, and 100C can be further configured such that, at least during the global sensing operation, the first traces 126 are simultaneously exposed to radiation (e.g., ambient radiation of different wavelengths). For example, because the top surface of the first traces 126 is free of any other thin films, or because the top surface of the first traces 126 is covered only by a transparent film (e.g., Figure 1A Alternatively, a single on-demand radiation shield may cover the entire array 110 (not shown), or multiple discrete on-demand radiation shields 130 may be aligned over the first lines 126 (e.g., Figure 1B The image sensor embodiment 100B and Figure 1C 100C). Such an on-demand radiation shield can be configured to switch between an opaque state and a transparent state in response to different bias conditions (e.g., applied by the peripheral circuitry 181 in response to a control signal from the controller 180) such that exposure of the first line occurs only when desired (e.g., during a global sensing operation). That is, in a first state under a first set of bias conditions, the on-demand radiation shield can be substantially opaque to prevent radiation from being transmitted to the first line 526, while in a second state under a second set of bias conditions different from the first set, the on-demand radiation shield can be substantially transparent to allow simultaneous exposure of the first line 126 to radiation.

[0030] Figure 3is a cross-sectional view illustrating an example of an on-demand radiation shield. The on-demand radiation shield includes a plurality of stacked, relatively thin layers of indium tin oxide (ITO) 332(1)-332(5) (e.g., less than 3 nm per layer). For ease of illustration, Figure 3 Five ITO layers are shown; however, a different number of multiple ITO layers (e.g., three or more) may also be used. In any case, the ITO layers 332(1)-332(5) are embedded in the cladding material such that the ITO layers are separated by the cladding material 331. Alternating ITO layers are electrically connected to each other. For example, the odd-numbered ITO layers 332(1), (3), and (5) can be electrically connected to each other via contacts 333a at one end, and the even-numbered ITO layers 332(2) and (4) can be electrically connected to each other via different contacts 333b at the opposite end. Depending on the voltage applied to the odd and even ITO layers via contacts 333a and 333b, respectively, the radiation shield 130 will be transparent or opaque to radiation (e.g., a given wavelength range). For example, if one set of ITO layers (e.g., odd-numbered ITO layers) is negatively biased (e.g., -3 V) and the other set of ITO layers (e.g., even-numbered ITO layers) is grounded, the radiation shield 130 may have a relatively low refractive index (e.g., n-1, e.g., n=1.042) and a high extinction coefficient (e.g., k>0.2, e.g., k=0.273), causing it to behave as a Bragg reflector and result in high transmission loss (e.g., -40 dB or higher, e.g., -47 dB), and thus be substantially opaque. Conversely, if the ITO layers are unbiased, the radiation shield 130 may have a relatively high refractive index (e.g., n-2, e.g., n=1.964) and a low extinction coefficient (e.g., k>0.005, e.g., k=0.002), resulting in low transmission loss (e.g., less than or equal to -4 dB, e.g., -3 dB), and thus be substantially transparent.

[0031] Figure 4A A cross-sectional view illustrating an exemplary STT-MTJ 120 during a global reset operation is shown. During the global reset operation, all first lines 126 may be discharged to ground, and a first positive voltage (e.g., VDD high) may be applied to all second lines 121 so that a sufficiently high current flows from the second lines 121 in a first direction through each of the STT-MTJs 120 (i.e., through the pinned layer 123, the tunnel barrier layer 124, and the free layer 125), thereby simultaneously resetting all STT-MTJs 120 in the array 110 to an antiparallel (AP) state. Therefore, as a result of the global reset operation, the same first data value (e.g., "0") is simultaneously stored in each of the STT-MTJs 120 in the array 110.

[0032] Figure 4Bis a cross-sectional view illustrating an exemplary STT-MTJ 120 during a global sensing operation. During a global sensing operation (also referred to herein as an image capture operation), a second positive voltage (e.g., VDD low) that is less than the first positive voltage (e.g., VDD high) may be applied to all first lines 126, and all second lines 121 may be discharged to ground. Under these bias conditions, a low current flows from the first lines 126 in a second direction through each of the STT-MTJs 120 (i.e., through the free layer 125, the tunnel barrier layer 124, and the pinned layer 123) to the second lines 121. However, switching any given STT-MTJ 120 from the AP state to the parallel (P) state, such that the stored data value switches from a first data value (e.g., “0”) to a second data value (e.g., “1”), will only occur if the actual luminous intensity level of radiation 300 received in the immediately adjacent first line portion is greater than a threshold luminous intensity level (i.e., actual candela (Cd a )>Threshold Candela (Cd t)). That is, during a global sensing operation, all first lines 126 will be exposed to radiation simultaneously because: (a) the top surface of the first lines 126 is continuously exposed regardless of operation (e.g., as in image sensor 100A), or (b) because the state of the on-demand radiation shield 130 has been switched, as described above, to allow transmission of radiation (e.g., in image sensor 100B or 100C). Depending on the specific image captured during the global sensing operation, the actual luminous intensity level of radiation 300 received in any given area of ​​the first lines 126 above any given STT-MTJ may vary. If the actual luminous intensity level of radiation 300 received at a particular first line portion (which is located above and immediately adjacent to a particular STT-MTJ) is below the threshold luminous intensity level, the combined energy from the bias applied to the first lines 126 and the radiation 300 received at the particular first line portion will be insufficient to cause the particular STT-MTJ immediately adjacent thereto to switch from the AP state to the P state. That is, the particular STT-MTJ will continue to store the first data value (e.g., "0"). However, if the actual luminous intensity level of the radiation 300 received at the particular first line portion (which is located above and immediately adjacent to the particular STT-MTJ) is higher than the threshold luminous intensity level, the combined energy from the bias applied to the first line 126 and the radiation 300 received at the particular first line portion will cause the particular STT-MTJ to switch from the AP state to the P state. That is, the data value stored in the particular STT-MTJ will switch from the first data value (e.g., "0") to the second data value (e.g., "1"). Therefore, during this global sensing operation, the STT-MTJs 120 in the array 110 are considered optically switchable.

[0033] Alternatively, the materials of the STT-MTJ and / or the bias conditions used during the global sensing operation can be specifically chosen so that radiation-induced switching occurs only in response to radiation exceeding a threshold intensity value and within a specific wavelength range.

[0034] A selective read operation may then be performed to read pixel values ​​corresponding to different locations of the image captured during the global sensing operation. Figure 4C FIG. 1 is a cross-sectional view illustrating an exemplary STT-MTJ 120 during a selective read operation.

[0035] In some image sensor embodiments (see, for example, Figure 1A The image sensor embodiment 100A and Figure 1BIn an image sensor embodiment 100B, the pixels may be single MTJ pixels such that, in array 110, the number of columns of pixels and the number of rows of pixels are the same as the number of columns of STT MTJs and the number of rows of STT MTJs. In this case, the image sensor may include sensing circuitry 190, and this sensing circuitry 190 may include amplifiers 191 (e.g., sense amplifiers, operational amplifiers (OPAMPs), or any other suitable type of amplifier) ​​electrically connected to the second lines 121. It should be noted that the ground connection to the second lines 121 may also be implemented using a virtual ground connection using an offset-compensated OPAMP. During a selective read operation of a selected single MTJ pixel (in this case, a read operation of the single STT-MTJ therein), a read voltage (Vread) may be applied to a specific first line 126 in contact with the STT-MTJ of the selected single MJT pixel (i.e., the specific first line 126 of the row containing the STT-MTJ of the selected single MJT pixel). In the presence of radiation (as in image sensor 100A) or in the absence of radiation when the on-demand radiation shield 130 is opaque (as in image sensor 100B), Vread can be less than the voltage level required to switch the STT-MTJ of the selected single MTJ pixel from the AP state to the P state. All other first lines can be discharged to ground. As a result of these bias conditions, the output of the particular amplifier 191 for the particular column containing the STT-MTJ of the selected single MTJ pixel (i.e., the output of the particular amplifier 191 on the particular second line 121 electrically connected to the STT-MTJ of the selected single MTJ pixel) will represent the pixel value.

[0036] It should be noted that, alternatively, the STT-MTJ can be designed as a high resistance device (e.g., by increasing the thickness of the insulator layer 124) to suppress the effects of sneak current from unselected rows during selective read operations. It should also be noted that, given the above-described bias conditions, a concurrent read operation of pixel values ​​can be performed, which is read as the stored data values ​​of all single STT-MTJs of all single MJT pixels in a given pixel row.

[0037] Alternatively, in other image sensor embodiments (see, for example, Figure 1CIn the image sensor embodiment 100C), the pixels may be multi-MTJ pixels. That is, each pixel corresponding to a given position of an image captured during a global sensing operation may be associated with multiple STT-MTJs 120. For example, each multi-MTJ pixel 129 corresponding to a given position of an image captured during a global sensing operation may be associated with two or more adjacent STT MTJs 120 from two or more adjacent columns and / or two or more adjacent STT MTJs 120 from two or more adjacent rows. For ease of illustration, in Figure 1C In the embodiment of the present invention, each multi-MTJ pixel 129 is associated with four STT MTJs, including two from adjacent columns of STT MTJs and two from adjacent rows of STT MTJs. Therefore, the exemplary array 110 having four columns AD of STT MTJs and four rows AD of STT MTJs will have only two columns of pixels and two rows of pixels. In this case, the sensing circuit 190' may include switches 192 (e.g., transmission gates) between adjacent second lines in the second line group associated with each of the pixel columns. That is, each pair of adjacent second lines in a given second line group associated with a given pixel column will be electrically connected (as needed) using the switches 192 (e.g., transmission gates). The sensing circuit 190' may further include amplifiers 191 (e.g., sense amplifiers or any other suitable type of amplifier) ​​electrically connected to the second line groups, respectively. That is, each amplifier 191 may be electrically connected to a given second line group associated with a given pixel column. During a selective read operation of a selected multi-MTJ pixel 129 (in this case, a read operation of the combined data values ​​stored in four adjacent STT MTJs), the switch 192 will be opened in response to one or more control signals from the controller 180 (e.g., in response to an enable signal and a reverse enable signal), thereby electrically connecting adjacent second lines within each given second line group associated with each given pixel column. A read voltage (Vread) can then be applied only to those first lines 126 that are in contact with the STT MTJs of the selected multi-MTJ pixel 129. In the presence of radiation (e.g., when no radiation shield is present (not shown)), or alternatively, in the absence of radiation when the on-demand radiation shield 130 is opaque (as in image sensor 100C), Vread can be less than the voltage level required to switch any STT-MTJ in the selected pixel 129 from the AP state to the P state. All other first lines can be discharged to ground. As a result of opening the switch 192 , and further as a result of the bias conditions, the output of the particular amplifier 191 for the particular pixel column containing the selected multi-MTJ pixel 129 will indicate the combined value of the stored data values ​​of all the STT-MTJs in the selected multi-MTJ pixel 129 .

[0038] Image capture accuracy can be statistically improved by considering the results of the switching behavior of all STT-MTJs in a multi-MTJ pixel in response to radiation exposure during a global sensing operation, rather than the switching behavior of a single STT-MTJ in a single-MTJ pixel. That is, rather than associating a given location in a captured image with high-intensity light when the STT-MTJ in the single-MTJ pixel switches its stored data value in response to radiation exposure during a global sensing operation, accuracy is improved by associating a given location with high-intensity light only when at least half of a group of adjacent STT-MTJs in a multi-MJT pixel switches their stored data values. In the former case, if the STT-MJT in a single-MJT pixel erroneously switches its stored data value, the value associated with that pixel will be erroneous. In the latter case, if only one of the STT-MJTs in a multi-MJT pixel erroneously switches its stored data value, the value associated with that pixel will still be correct. It should be noted that, given the biasing conditions described above, parallel readout of pixel values ​​can be performed from the multiple-MJT pixels in a given pixel row. In any case, during a selective read operation, the value read out from the single-MTJ pixel or multi-MTJ pixel will be binary, indicating whether high-intensity light (optionally of a particular color) is present.

[0039] Alternatively, for multi-MJT pixels, different STT-MTJs within a given group (i.e., within a multi-MTJ pixel) can be configured to have different energy barriers. For example, the STT-MTJs within each multi-MTJ pixel can each have a different cross-sectional area. In this case, the analog intensity of incident radiation can be sensed by detecting the total number of STT-MTJs within the group that switch during a global sensing operation.

[0040] refer to Figure 5A 、 Figure 5B and Figure 5C Other image sensor embodiments 500A, 500B, and 500C disclosed herein include an array 510 of optically switchable magnetic tunnel junctions (MTJs), particularly optically switchable spin Hall effect (SHE) MTJs 520, arranged in columns and rows. For illustrative purposes, the array 510 is shown as having four columns AD and four rows ad. However, it should be understood that these numbers are not intended to be limiting, and that the array 510 may have any number of two or more columns and two or more rows of SHE-MTJs 520. As discussed in more detail below, the SHE-MTJs 520 may be conventional SHE-MTJs or SHE-MTJs assisted by voltage-controlled magnetic anisotropy (VCMA).

[0041] In any case, each SHE-MTJ 520 can be a back-end-of-the-line (BEOL) multilayer structure. That is, each SHE-MTJ 520 can be a multilayer structure formed in a BEOL metal layer above a front-end-of-the-line (FEOL) device (e.g., a semiconductor device) on an integrated circuit chip. Each SHE-MTJ 520 can include an insulator layer 524 (also referred to as a tunnel barrier layer) between a free ferromagnetic layer 525 (also referred to herein as a free layer or data storage layer) on a first side and a pinned ferromagnetic layer 523 (also referred to herein as a pinned layer or fixed ferromagnetic layer) on a second side. In one exemplary embodiment, the pinned ferromagnetic layer 523 can be, for example, a gadolinium (Gd), iron (Fe), and cobalt (Co) alloy layer. The insulator layer 524 (i.e., tunnel barrier layer) can be, for example, any of a magnesium oxide (MgO) layer, an aluminum oxide (Al2O3) layer (also referred to as aluminum oxide), and a bismuth ferrite (BiFeO3, also referred to as BFO) layer. The free ferromagnetic layer 525 may be, for example, a cobalt (Co), iron (Fe), boron (B) alloy layer.

[0042] Those skilled in the art will recognize that similar or identical materials can be used to form STT-MTJs, conventional SHE-MTJs, and VCMA-assisted SHE-MTJs. However, the insulator layer 524 employed in conventional SHE-MTJs can use a different thickness than the insulator layer 524 used in VCMA-assisted SHE-MTJs to optimize performance.

[0043] These image sensor embodiments 500A, 500B, and 500C may further include a first line 526. The first line 526 may be adjacent to rows ad of the SHE-MTJs 520, respectively. Specifically, each first line 526 may span a corresponding row of SHE-MTJs such that it extends and contacts the free layer 525 of each SHE-MTJ 520 in the corresponding row. For example, the first line 526 may be a relatively thin and flat strip of material (also referred to as a nanosheet) that is: (a) entirely made of a transparent conductive SHE material 527 (e.g., Figure 6A (b) having a first segment and a second segment, the first segment being adjacent to the free layer 525 of the SHE-MTJ 520 and being made of a transparent conductive SHE material 527, the second segment being located on the opposite side of the first segment (e.g., such that each first segment is laterally located between two second segments) and being made of a metal or metal alloy material 528 to improve conductivity (e.g., Figure 6B or Figure 6C). In any case, the transparent conductive SHE material 527 may comprise a relatively thin (e.g., less than 3 nm) and may comprise, for example, a heavy metal (e.g., gadolinium, tantalum, platinum, tungsten, or any other suitable heavy metal having a density of at least 5 g / cm3), graphene, or any other suitable transparent conductive material having a spin Hall effect. It should be noted that indium tin oxide (ITO) or zinc oxide (ZnO) may also be used, but those skilled in the art will recognize that these materials have relatively low spin Hall angles. Where the first line 526 has a first segment of transparent conductive SHE material 527 and a second segment of metal or metal alloy material 528, the metal or metal alloy material 528 may be, for example, copper, aluminum, or any other suitable BEOL metal or metal alloy material. Additionally, the first segment of transparent conductive SHE material 527 and the second segment of metal or metal alloy material 528 may have the same thickness (e.g., Figure 6B As shown) or different thicknesses (eg, the thickness of the second section of the metal or metal alloy material 528 may be greater than the thickness of the transparent conductive SHE material 527) (eg Figure 6C shown).

[0044] These image sensor embodiments 500A, 500B, and 500C may further include second lines 521 of conductive material. The second lines 521 may be, for example, conductive wires made of a metal or metal alloy (e.g., copper or any other suitable BEOL metal or metal alloy). The second lines 521 may be electrically connected (e.g., via local interconnects 522) to the MTJs in columns AD of the SHE-MTJs, respectively. Specifically, each second line 521 may be electrically connected to the pinning layer 523 of each SHE-MTJ 520 in the corresponding column AD via the local interconnects 522.

[0045] Furthermore, these image sensor embodiments 500A, 500B, and 500C may be configured such that both the first line and the second line are selectively biased, thereby enabling global reset, global sensing, and selective read operations with respect to the SHE-MTJs to be performed without requiring configuration transistors (e.g., access transistors, etc.) for each SHE-MTJ.

[0046] Specifically, image sensor embodiments 500A, 500B, and 500C may further include a controller 180 and peripheral circuits 581-583. In response to control signals from the controller 180, the peripheral circuits 581-583 may cause biasing of the first line 126 and the second line 121 during global reset, global sense, and selective read operations, as described below. Specifically, the peripheral circuits 581 and 582 may be connected to opposite ends of the first line 526 of the row and may include, for example, address decoding logic and a first line driver for appropriately biasing the opposite ends of the first line, as described below. The peripheral circuit 583 may be connected to the second line 521 of the column and may include column address decoding logic and a second line driver for appropriately biasing the second line, as described below. It is well known in the art that the peripheral circuits are configured to selectively bias the lines of the array in response to control signals from the controller. Therefore, in order to allow the reader to focus on the salient aspects of the disclosed embodiments, the details thereof are omitted from this specification.

[0047] These image sensor embodiments 500A, 500B, and 500C can be further configured such that, at least during the global sensing operation, the first traces 526 are simultaneously exposed to radiation (e.g., ambient radiation of different wavelengths). For example, because the top surface of the first traces 526 is free of any other film, or because the top surface of the first traces 126 is covered only by a transparent film (e.g., Figure 5A In an embodiment of an image sensor 500A (shown in FIG. 5 ), the first lines 526 can be continuously exposed to ambient light. Alternatively, a single on-demand radiation shield (not shown) covering the entire array 510 or a plurality of discrete on-demand radiation shields 530 can be aligned over the first lines 526. Such an on-demand radiation shield can be configured to switch between opaque and transparent states in response to different bias conditions (e.g., applied by peripheral circuitry 581 and / or 582 in response to control signals from controller 580) such that exposure of the first lines occurs only when desired (e.g., during global sensing). That is, in a first state under a first set of bias conditions, the on-demand radiation shield can be substantially opaque to prevent radiation from being transmitted to the first lines 526, while in a second state under a second set of bias conditions different from the first set, the on-demand radiation shield can be substantially transparent to allow simultaneous exposure of the first lines 526 to radiation. Figure 3 is a cross-sectional view of an exemplary on-demand radiation shield (see above Figure 3 for detailed discussion).

[0048] Figure 7Ais a cross-sectional view illustrating an exemplary SHE-MTJ 520 during a global reset operation. For example, during a global reset operation, a first positive voltage (e.g., VDD high) may be applied to the first ends of all first lines 526 (i.e., the transparent conductive SHE material), all second ends of all first lines 526 opposite the first ends may be grounded, and all second lines 521 may also be grounded, so that a sufficiently high current flows through the first lines 526 in a first direction (e.g., in the X direction) to simultaneously reset the SHE-MTJs to an antiparallel (AP) state. That is, in each SHE-MTJ, the polarity of the magnetism in the free layer 525 will remain or switch to a second direction that is perpendicular to the first direction of the current flowing through the first lines 526 and opposite to the polarity of the magnetism in the pinned layer 523. For example, if the magnetic polarity in the pinned layer 523 of each SHE-MTJ is in the Y+ direction, a current flowing in the X direction through the first line 526 (containing a transparent conductive SHE material) can cause the magnetic polarity in the free layer 525 to remain in the Y direction or switch to the Y direction. In this way, the same first data value (e.g., “0”) is simultaneously stored in each SHE-MTJ 520 in the array 510.

[0049] FIG7B(1) and FIG7B(2) are different cross-sectional views illustrating an exemplary conventional SHE-MTJ or an exemplary VCMA-assisted SHE-MTJ, respectively, during a global sensing operation. During a global sensing operation (also referred to herein as an image capture operation), all first lines may be simultaneously exposed to radiation 300, however, the bias conditions will vary depending on whether the SHE-MTJ is a conventional SHE-MTJ or a VCMA-assisted SHE-MTJ.

[0050] Specifically, for a conventional SHE-MTJ (as shown in FIG. 7B(1)), a second positive voltage (e.g., VDD low) less than the first positive voltage (e.g., VDD high) may be applied to all second ends of the first lines 526, all first ends of the first lines 526 may be grounded, and all second lines 521 may be discharged to ground. Under these bias conditions, a low current flows through the first lines 526 in the X+ direction. However, by switching the magnetic polarity in the free layer 525 to the same direction as the magnetic polarity in the pinned layer 523 (e.g., to the Y+ direction), any given conventional SHE-MTJ is switched from the AP state to the parallel (P) state, so that the actual luminous intensity level of the radiation 300 received by the immediately adjacent first line portion is greater than the threshold luminous intensity level (i.e., actual candela (Cd a )>Threshold Candela (Cd t)), the switching of the stored data value from the first data value (e.g., “0”) to the second data value (e.g., “1”) will only occur when the first lines 526 are simultaneously exposed to radiation 300 during the global sensing operation, either because: (a) the top surface of the first lines 526 is continuously exposed regardless of the operation (e.g., as in image sensor 500A), or (b) because the state of the on-demand radiation shield 530 has been switched (as needed, as described above) to allow transmission of radiation 300 (e.g., in image sensors 500B or 500C). Depending on the particular image captured during this global sensing operation, the actual luminous intensity level of radiation 300 received in any given area of ​​the first lines 526 above any given conventional SHE-MTJ can vary. If the actual luminous intensity level of the radiation 300 received at a particular first line portion (which is located above and immediately adjacent to the particular SHE-MTJ) is less than the threshold luminous intensity level, the combined energy of the low current in the X+ direction passing through the first line 526 and the radiation 300 received at the particular first line portion will be insufficient to cause the particular conventional SHE-MTJ adjacent thereto to switch from the AP state to the P state. That is, the particular conventional SHE-MTJ will continue to store the first data value (e.g., "0"). However, if the actual luminous intensity level of the radiation 300 received at the particular first line portion (which is located above and immediately adjacent to the particular SHE-MTJ) is greater than the threshold luminous intensity level, the combined energy of the low current in the X+ direction passing through the first line 526 and the radiation 300 received at the particular first line portion will cause the particular SHE-MTJ to switch from the AP state to the P state (i.e., will cause the polarity of the magnetism in the free layer 525 to switch so that it is the same polarity as the magnetism in the pinned layer 523, e.g., in the Y+ direction). That is, the data value stored in a particular conventional SHE-MTJ will switch from a first data value (eg, “0”) to a second data value (eg, “1”).

[0051] For a VCMA-assisted SHE-MTJ (as shown in FIG. 7B (2)), the first and second ends of all first lines 526 can be electrically grounded, and a voltage-controlled magnetic anisotropy (VCMA) voltage can be applied to all second lines. The VCMA voltage can be a high positive voltage, in particular, at a level that is the same as or higher than the first positive voltage (e.g., VDD high) described above. Under these bias conditions, the VCMA-assisted SHE-MTJ is switched from the AP state to the parallel (P) state by switching the polarity of the magnetism in the free layer 525 to the same direction as the polarity of the magnetism in the pinned layer 523 (e.g., in the Y+ direction), so that the actual luminous intensity level of the radiation 300 received by the immediately adjacent first line portion is greater than the threshold luminous intensity level (i.e., actual candela (Cd a )>Threshold Candela (Cd t)), the switching of the stored data value from the first data value (e.g., “0”) to the second data value (e.g., “1”) will only occur when the first lines 526 are simultaneously exposed to radiation 300 during the global sensing operation, either because: (a) the top surface of the first lines 526 is continuously exposed regardless of the operation (e.g., as in image sensor 500A), or (b) because the state of the on-demand radiation shield 530 has been switched (as needed, as described above) to allow transmission of radiation 300 (e.g., in image sensors 500B or 500C). Depending on the particular image captured during this global sensing operation, the actual luminous intensity level of radiation 300 received in any given area of ​​the first lines 526 above any given VCMA-assisted SHE-MTJ can vary. If the actual luminous intensity level of the radiation 300 received at a particular first line portion (which is located above and immediately adjacent to a particular VCMA-auxiliary SHE-MTJ) is less than the threshold luminous intensity level, the combined energy from the VCMA voltage applied to the second line and the radiation 300 received at the particular first line portion will be insufficient to cause the particular VCMA-auxiliary SHE-MTJ adjacent thereto to switch from the AP state to the P state. In other words, the particular VCMA-auxiliary SHE-MTJ will continue to store the first data value (e.g., “0”). However, if the actual luminous intensity level of the radiation 300 received at a particular first line portion (which is located above and immediately adjacent to the particular VCMA-assisted SHE-MTJ) is higher than the threshold luminous intensity level, the combined energy from the VCMA voltage applied to the second line and the radiation 300 received at the particular first line portion will cause the particular VCMA-assisted SHE-MTJ to switch from the AP state to the P state (i.e., it will cause the polarity of the magnetism in the free layer 525 to switch so that it is the same polarity as the magnetism in the pinned layer 523, for example, in the Y+ direction). In other words, the data value stored in the particular VCMA-assisted SHE-MTJ will switch from a first data value (e.g., “0”) to a second data value (e.g., “1”).

[0052] Therefore, the SHE-MTJs 520 (conventional or Vvcma assisted) in the array 510 are considered optically switchable during this global sensing operation.

[0053] Alternatively, the materials of the SHE-MTJ 520 and / or the bias conditions used during global sensing operations may be specifically chosen so that radiation-induced switching occurs only in response to radiation exceeding a threshold intensity value and within a specific wavelength range.

[0054] A selective read operation may then be performed to read pixel values ​​corresponding to different locations of the image captured during the global sensing operation. Figure 7C A cross-sectional view of an exemplary SHE-MTJ 520 during a selective read operation is shown.

[0055] In some image sensor embodiments (see, for example, Figure 5A The image sensor embodiment 500A and Figure 5B In an image sensor embodiment 500B, the pixels may be single MTJ pixels such that, within the array 510, the number of pixel columns and the number of pixel rows are the same as the number of SHE-MTJ columns and the number of SHE-MTJ rows. In this case, the image sensor may include a sensing circuit 590, and the sensing circuit 590 may include amplifiers 591 (e.g., sense amplifiers or any other suitable type of amplifier) ​​electrically connected to the second lines 521. During a selective read operation of a selected single MTJ pixel (in this case, a read operation of a single SHE-MTJ therein), a read voltage (Vread) may be applied to opposite ends (i.e., first and second ends) of a particular first line 526 in contact with the SHE-MTJ of the selected single MJT pixel (i.e., to the particular first line 526 of the row containing the SHE-MTJ of the selected single MJT pixel). In the presence of radiation (as in image sensor 500A) or in the absence of radiation when the on-demand radiation shield 530 is opaque (as in image sensor 500B), Vread can be less than the voltage level required to switch the SHE-MTJ of the selected single MTJ pixel from the AP state to the P state. All other first lines can be grounded. As a result of these bias conditions, the output of the particular amplifier 591 for the particular column containing the SHE-MTJ of the selected single MTJ pixel (i.e., the output of the particular amplifier 591 on the particular second line 521 electrically connected to the SHE-MTJ of the selected single MTJ pixel) will indicate the pixel value.

[0056] It should be noted that under these bias conditions, a concurrent read operation can be performed that reads the pixel values ​​as stored data values ​​for all single SHE-MTJs of all single MJT pixels in a given pixel row.

[0057] Alternatively, in other image sensor embodiments (see, for example, Figure 5CIn the image sensor embodiment 500C), the pixels may be multi-MTJ pixels. That is, each pixel corresponding to a given position of an image captured during a global sensing operation may be associated with multiple SHE-MTJs 520. For example, each multi-MTJ pixel 529 corresponding to a given position of an image captured during a global sensing operation may be associated with two or more adjacent SHE-MTJs 520 from two or more adjacent columns and / or two or more adjacent SHE-MTJs 520 from two or more adjacent rows. For ease of illustration, in Figure 5C In the example, each multi-MTJ pixel 529 is associated with four SHE-MTJs, including two from adjacent columns of SHE-MTJs and two from adjacent rows of SHE-MTJs. Thus, an exemplary array 510 having four columns AD and four rows ad of SHE-MTJs would have only two columns of pixels and two rows of pixels. In this case, the sensing circuit 590′ may include switches 592 (e.g., transmission gates) between adjacent second lines in the second line group associated with each of the pixel columns. That is, each pair of adjacent second lines in a given second line group associated with a given pixel column can be electrically connected (as needed) through the use of switches 592 (e.g., transmission gates). The sensing circuit 590′ may also include amplifiers 591 (e.g., sense amplifiers or any other suitable type of amplifier) ​​electrically connected to each of the second line groups. That is, each amplifier 591 can be electrically connected to a given second line group associated with a given pixel column. During a selective read operation of a selected multi-MTJ pixel 529 (in this case, a read operation of the combined data values ​​stored in four adjacent SHE-MTJs), the switch 592 will respond to one or more control signals from the controller 580 (e.g., an enable signal and an inverse enable signal) to electrically connect adjacent second lines within each given second line group associated with each given pixel column. A read voltage (Vread) can then be applied only to the opposite end of the first line 526 that contacts the SHE-MTJ of the selected multi-MTJ pixel 529. In the presence of radiation (e.g., when there is no radiation shield (not shown)), or alternatively, in the absence of radiation when the on-demand radiation shield 530 is opaque (as in image sensor 100C), Vread can be less than the voltage level required to switch any SHE-MTJ in the selected multi-MTJ pixel 529 from the AP state to the P state. All other first lines can be discharged to ground. As a result of opening the switch 592 , and further as a result of the bias conditions, the output of the particular amplifier 591 for the particular pixel column containing the selected multi-MTJ pixel 529 will indicate the combined value of the stored data values ​​of all the SHE-MTJs in the selected multi-MTJ pixel 529 .

[0058] By considering the switching behavior of all SHE-MTJs in a multi-MTJ pixel in response to radiation exposure during a global sensing operation, rather than the switching behavior of only a single SHE-MTJ in a single-MTJ pixel, image capture accuracy can be statistically improved. That is, rather than associating a given location in a captured image with high-intensity light when the SHE-MTJ in the single-MTJ pixel switches its stored data value in response to radiation exposure during a global sensing operation, accuracy is improved by associating a given location with high-intensity light only when at least half of a group of adjacent SHE-MJTs in a multi-MJT pixel switches their stored data values. In the former case, if the SHE-MJT in a single-MJT pixel erroneously switches its stored data value, the value associated with that pixel will be erroneous. In the latter case, if only one of the SHE-MJTs in a multi-MJT pixel erroneously switches its stored data value, the value associated with that pixel will still be correct. It should be noted that, given the biasing conditions described above, concurrent read operations can be performed for the pixel values ​​of all multi-MJT pixels in a given pixel row. In any case, during a selective read operation, the value read from the single-MTJ pixel or multi-MTJ pixel will be binary, indicating whether high-intensity light (optionally of a particular color) is present.

[0059] Alternatively, for multi-MJT pixels, different SHE-MTJs within a given cluster (i.e., within a multi-MTJ pixel) can be configured to have different energy barriers. For example, the SHE-MTJs within each multi-MTJ pixel can each have a different cross-sectional area. In this case, the analog intensity of incident radiation can be sensed by sensing the total number of SHE-MTJs within the cluster that switch during a global sensing operation.

[0060] In each of the above-described image sensor embodiments (e.g., Figure 1A-1C The image sensor embodiments 100A-100C and Figures 5A-5CIn the image sensor embodiments 500A-500C), during a global sensing operation, all MTJs 120, 520 in the array 110, 510 are exposed to radiation simultaneously, and when specific bias conditions applied to the first and second lines are applied, only the MTJs that receive radiation above a threshold luminous intensity level switch their states from AP to P simultaneously. The state of the array 110, 510 in which the MTJs are simultaneously exposed to radiation and can be simultaneously switched according to the intensity of the radiation is referred to herein as a global shutter effect. This global shutter effect is achieved due to the specific bias conditions applied to the first and second lines during the global sensing operation (as described above). Other bias conditions, particularly the specific bias conditions applied to the first and second lines during reset and read operations (as described above), are not sufficient to achieve this optical switching under the MTJ structure. In any case, the global shutter effect can be enhanced, particularly by better defining the global shutter effect through the use of an on-demand radiation shield 130, 530, which specifically limits the time of radiation exposure.

[0061] The above-mentioned image sensor embodiments (eg, Figure 1A-1C The image sensor embodiments 100A-100C and Figures 5A-5C Advantages of the image sensor embodiments 500A-500C) include the ability of the MTJ (e.g., STT-MTJ or SHE-MTJ) to store sensed image data, which is captured during a global sensing operation, until a global reset operation is performed and without leakage issues. That is, the array has inherent memory storage and is radiation-hardened, so there is no need to read the captured data immediately or avoid data loss. In addition, the MTJ is a relatively small back-end-of-the-line (BEOL) data storage device, and unlike conventional image sensors, which require multiple front-end-of-the-line (FEOL) devices for each pixel in the array (e.g., relatively large photodiodes and at least three transistors per pixel), the disclosed image sensor embodiments do not require FEOL devices within the array 110, 510. Therefore, the disclosed image sensor embodiments consume less chip area (e.g., approximately 2-4 μm) than conventional image sensors. 2 ), consuming significantly less chip area (e.g., less than 0.022 μm 2 ).

[0062] It should be understood that the terms used herein are for describing disclosed structures and methods, rather than for limiting. For example, as used herein, the singular forms "one", "an" and "the" are intended to include plural forms, unless the context clearly indicates otherwise. In addition, the terms "include", "comprise", "include", "include" and / or "include" as used herein specify the presence of the features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or their groups. In addition, the terms as used herein, such as "right", "left", "vertical", "horizontal", "top", "bottom", "lower", "below", "below", "bottom", "up", "above", "parallel", "vertical", etc., are intended to describe the relative position (unless otherwise stated) of orientation and description in the drawings, and terms such as "contact", "direct contact", "adjacent", "directly adjacent", "closely adjacent" are intended to represent that at least one element is in physical contact with another element (other elements without separating the elements). The term "laterally" is used herein to describe the relative position of elements, and more specifically, to indicate that one element is positioned to the side of another element, rather than above or below the other element, as the elements are oriented and illustrated in the drawings. For example, an element with a side adjacent to another element would be located beside the other element, an element with a side immediately adjacent to another element would be located directly beside the other element, and an element that laterally surrounds another element would be adjacent to and bordered by an outer sidewall of the other element. The corresponding structures, materials, acts, and equivalents of all means or step plus function elements in the following claims are intended to include any structure, material, or act to perform the function in combination with other elements specifically claimed.

[0063] The description of various embodiments of the present invention is presented for illustrative purposes and is not intended to be exhaustive or limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terminology used herein is selected to best explain the principles of the embodiments, practical applications, or technical improvements over existing technologies on the market, or to enable others skilled in the art to understand the embodiments disclosed herein.

Claims

1. A sensor comprising: an array of magnetic tunnel junctions arranged in columns and rows, wherein the magnetic tunnel junctions include spin Hall effect magnetic tunnel junctions including an insulator layer between a free layer and a pinned layer; first lines comprising a transparent conductive spin Hall effect material, wherein each first line spans the free layer contacting each magnetic tunnel junction in a corresponding row and is exposed to radiation; and second lines, wherein each second line is electrically connected to the pinned layer of each magnetic tunnel junction in a corresponding column, and wherein the magnetic tunnel junction is configured to store a data value that depends on an intensity level of the radiation received at a portion proximate to the first line during a sensing operation.

2. The sensor according to claim 1, wherein The transparent conductive spin Hall effect material includes heavy metal or graphene with a thickness of less than 3 nm.

3. The sensor according to claim 1, wherein Each first line includes a first segment comprising the transparent conductive spin Hall effect material in contact with the free layer of each magnetic tunnel junction in the corresponding row and a second segment located on an opposite side of the first segment and comprising a metal or metal alloy material.

4. The sensor according to claim 1, further comprising: Controller; and a peripheral circuit in communication with the controller, the peripheral circuit being configured to selectively bias the first line and the second line in response to a control signal from the controller.

5. The sensor according to claim 4, in, During a global reset operation, the peripheral circuit applies a first positive voltage to first ends of all the first lines, grounds second ends of all the first lines opposite to the first ends, and grounds all the second lines, so that all the magnetic tunnel junctions in the array store a first data value simultaneously. During a global sensing operation, the peripheral circuit applies a second positive voltage less than the first positive voltage to the second ends of all the first lines, grounds the first ends of all the first lines, and grounds all the second lines, so that in response to simultaneous exposure of all the first lines to radiation, switching from a first data value to a second data value occurs only in a given magnetic tunnel junction when an actual intensity level of the radiation received at a portion proximate to the first line is greater than a threshold intensity level.

6. The sensor according to claim 5, in, The pixel consists of a single magnetic tunnel junction, Wherein, the sensor further includes amplifiers electrically connected to the second circuits respectively; and In which, during a selective read operation, the peripheral circuit applies a read voltage to a first line contacting the single magnetic tunnel junction of a selected pixel and grounds all other first lines, so that the output of the amplifier on the second line electrically connected to the single magnetic tunnel junction indicates the data value stored in the single magnetic tunnel junction.

7. The sensor according to claim 5, in, The pixel comprises a group of at least two magnetic tunnel junctions such that the array of magnetic tunnel junctions comprises an array of pixels arranged in columns and rows; The sensor further includes: switches respectively located between adjacent second lines in the group of second lines associated with the column of pixels; and amplifiers, electrically connected to the group of second circuits respectively; and The switch is turned on in response to an enable signal from the controller and electrically connects the second lines within each group, and during a selective read operation, the peripheral circuit applies a read voltage to all first lines in contact with any magnetic tunnel junction of a selected pixel and further grounds all other first lines, so that the output of the amplifier of the group connected to the second line electrically connected to the magnetic tunnel junction of the selected pixel indicates the combined data value of all the magnetic tunnel junctions in the selected pixel.

8. The sensor of claim 1 , further comprising at least one on-demand radiation shield selectively operable in a first state and a second state, wherein In the first state, the at least one radiation on demand shield is substantially opaque to prevent the radiation from being transmitted to the first line, and wherein in the second state, the at least one radiation on demand shield is substantially transparent to allow the first line to be simultaneously exposed to radiation.

9. A sensor comprising: an array of magnetic tunnel junctions arranged in columns and rows, wherein the magnetic tunnel junctions include an insulator layer between a free layer and a pinned layer; first lines comprising a transparent conductive material, wherein each first line spans the free layer contacting each magnetic tunnel junction in a corresponding row and is exposed to radiation; and second lines, wherein each second line is electrically connected to the pinned layer of each magnetic tunnel junction in a corresponding column, wherein the magnetic tunnel junction is configured to store a data value that is dependent on an intensity level of the radiation received at a portion proximate to the first line during a sensing operation, and wherein the magnetic tunnel junction simultaneously stores image data in response to specific bias conditions on the first and second lines and simultaneous exposure of the first line to radiation.

10. The sensor according to claim 9, wherein Each first line includes a first segment comprising the transparent conductive material in contact with the free layer of each magnetic tunnel junction in the corresponding row and a second segment located on an opposite side of the first segment and comprising a metal or metal alloy material.

11. The sensor according to claim 9, wherein The magnetic tunnel junction includes a spin transfer torque type magnetic tunnel junction.

12. The sensor according to claim 11, wherein The transparent conductive material includes any one of indium tin oxide (ITO), zinc oxide (ZnO) and graphene.

13. The sensor according to claim 11, further comprising: Controller; and a peripheral circuit in communication with the controller, the peripheral circuit being configured to selectively bias the first line and the second line in response to a control signal from the controller.

14. The sensor according to claim 13, in, During a global reset operation, the peripheral circuit grounds all the first lines and applies a first positive voltage to all the second lines, so that all the magnetic tunnel junctions in the array store a first data value simultaneously; as well as During a global sensing operation, the peripheral circuit applies a second positive voltage less than the first positive voltage to all of the first lines and grounds all of the second lines, such that in response to simultaneous exposure of all of the first lines to radiation, switching from a first data value to a second data value occurs only in a given magnetic tunnel junction when an actual intensity level of the radiation received proximate a portion of the first lines is greater than a threshold intensity level.

15. The sensor according to claim 13, in, The pixel comprises a single magnetic tunnel junction; Wherein, the sensor further includes amplifiers electrically connected to the second circuits respectively; and In which, during a selective read operation, the peripheral circuit applies a read voltage to a first line contacting the single magnetic tunnel junction of a selected pixel and grounds all other first lines, so that the output of the amplifier on the second line electrically connected to the single magnetic tunnel junction indicates the data value stored in the single magnetic tunnel junction.

16. The sensor according to claim 13, in, The pixel comprises a group of at least two magnetic tunnel junctions such that the array of magnetic tunnel junctions comprises an array of pixels arranged in columns and rows; The sensor further includes: switches, respectively located between adjacent second lines in the group of second lines associated with the column of pixels; amplifiers, electrically connected to the group of second circuits, respectively; and The switch is turned on in response to an enable signal from the controller and electrically connects the second lines within each group, and during a selective read operation, the peripheral circuit applies a read voltage to all first lines in contact with any magnetic tunnel junction of a selected pixel and grounds all other first lines, so that the output of the amplifier of the group connected to the second line electrically connected to the magnetic tunnel junction of the selected pixel indicates a combined data value of all the magnetic tunnel junctions in the selected pixel.

17. The sensor of claim 13, further comprising at least one on-demand radiation shield operable in a first state and a second state, wherein In the first state, the at least one radiation on demand shield is substantially opaque to prevent the radiation from being transmitted to the first line, and wherein in the second state, the at least one radiation on demand shield is substantially transparent to allow the first line to be simultaneously exposed to radiation.

18. A sensor comprising: an array of magnetic tunnel junctions arranged in columns and rows, wherein the magnetic tunnel junctions include an insulator layer between a free layer and a pinned layer; first lines comprising a transparent conductive material, wherein each first line spans the free layer contacting each magnetic tunnel junction in a corresponding row and is exposed to radiation; second lines, wherein each second line is electrically connected to the pinned layer of each magnetic tunnel junction in a corresponding column, and wherein the magnetic tunnel junction is configured to store a data value that depends on an intensity level of the radiation received at a portion proximate to the first line during a sensing operation; and a peripheral circuit connected to the first circuit and the second circuit; wherein the peripheral circuit applies a first set of bias conditions to the first circuit and the second circuit so as to simultaneously store a first data value in the magnetic tunnel junction; and The peripheral circuit applies a second set of bias conditions to the first and second lines such that, in response to all of the first lines being simultaneously exposed to radiation, switching from a first data value to a second data value occurs only in a given magnetic tunnel junction when an actual intensity level of the radiation received proximate a portion of the first lines is greater than a threshold intensity level.

19. The sensor according to claim 18, wherein The magnetic tunnel junction includes a spin transfer torque type magnetic tunnel junction, and wherein the transparent conductive material includes any one of indium tin oxide (ITO), zinc oxide (ZnO), and graphene.

20. The sensor according to claim 18, wherein The magnetic tunnel junction comprises a spin Hall effect type magnetic tunnel junction, and wherein the transparent conductive material comprises a transparent conductive spin Hall effect material, and the transparent conductive spin Hall effect material comprises a heavy metal or graphene with a thickness less than 3 nm.

Images