Patent Grant
11069745
Embodiments described herein relate generally to a memory device.
In recent years, there has been suggested a memory called a resistive RAM (ReRAM), in which each of memory cells is formed with a non-ohmic element typified by a diode and a variable resistance material. The memory cells of the ReRAM do not use any MOSFET. Therefore, the memory cells are expected in that they enable the achievement of a high integration beyond a conventional trend.
In general, according to one embodiment, a memory device includes: a plurality of first interconnects; a plurality of second interconnects; a plurality of third interconnects; a plurality of memory cells; and a plurality of selectors. The respective first interconnects are provided along a first direction. The respective second interconnects are provided along a second direction different from the first direction. The respective third interconnects are provided along a third direction different from the first and second directions. The memory cells includes variable resistance layers formed on two side surfaces, facing each other in the first direction, of the third interconnects and coupled with the mutually different second interconnects. The selectors couple the third interconnects with the first interconnects. One of the selectors includes a semiconductor layer provided between associated one of the third interconnects and associated one of the first interconnects, and gates formed on two side surfaces of the semiconductor layer facing each other in the first direction with gate insulating films interposed therebetween.
Hereinafter, a memory device according to a first embodiment will be described with respect to an ReRAM as an example.
1. Re: Structure of Memory Cell Array
As shown, global column lines 1, row lines 2 and column lines 3 are provided in the memory cell array. The global column lines 1 are formed in parallel with one another along a first direction, and are arranged in a lowermost layer of the memory cell array. The row lines 2 are formed in parallel with one another along a second direction orthogonal to the first direction, and are provided at positions higher than the global column lines 1. Moreover, layers of the row lines 2 (a first layer, a second layer, a third layer, and the like of
A memory cell MC including a variable resistance element is formed between each of the row lines 2 and each of the column lines 3. In the present example, a variable resistance material 4 is disposed on the whole side surface of the column line 3 (the surface thereof facing the row line 2), and this material functions as the memory cell MC. It is to be noted that the variable resistance material 4 in the present example is provided on two side surfaces (the two side surfaces which face the row lines 2) of the bit line 3 which face each other in the first direction, and are not provided on two side surfaces (the two side surfaces which do not face the row lines 2) which face each other in the second direction in two sets of the facing side surfaces of the bit line 3.
A selection element (sheet selector) SS is provided between the global column line 1 and the corresponding column line. The selection element SS includes a source region 5 formed on the global column line 1, a silicon layer (a channel region) 6 formed on the source region 5, and a drain region 7 formed on the silicon layer 6. Furthermore, a selection gate line 8 along the second direction is formed between the adjacent silicon layers 6. The selection gate line 8 is disposed in parallel with the row lines 2. Furthermore, a gate insulating film 9 is formed between the selection gate line 8 and the silicon layer 6.
It is to be noted that hereinafter, the global column lines 1, the row lines 2 and the column lines 3 are referred to as the global bit lines GBL, the word lines WL, and the bit lines BL, respectively, in the same manner as in a usual MOS type memory device.
The pillar-shaped bit lines BL are formed on the drain region 7 of each of the selection elements SS. Moreover, the variable resistance material A which functions as the memory cell MC is disposed on the side surfaces of the bit lines BL. Furthermore, the word lines WL are formed in a region between the bit lines BL adjacent in the first direction. The variable resistance material 4 is formed by using, for example, HfO as the material so that the material contacts with the bit lines BL and the word lines WL. The variable resistance material 4 typified fay this HfO is a material which transits between at least two resistance values of a low resistance state (LRS) and a high resistance state (HRS). Moreover, it is known that the variable resistance material of the high resistance state transits to the low resistance state when a predetermined or larger voltage is applied and that the variable resistance material of the low resistance state transits to the high resistance state when a predetermined or larger current flows. In particular, an element which performs the transition from the high resistance state to the low resistance state and the transition from the low resistance state to the high resistance state when the voltage is applied with different polarities is called a bipolar operation element. The variable resistance material 4 which performs such an operation can be formed by using a thin film including at least one of materials such as TiO2, ZnMn2O4, NiO, SrZrO3, Pr0.7Ca0.3MnO3 and carbon in addition to HfO.
As described above, the memory cells MC each including the variable resistance material 4 provided between the word line 1 and the bit line 2 are arranged with, for example, a three-dimensional matrix arrangement in the memory cell array. In the present structure, the word lines WL and the bit lines BL simply form a line and space pattern. Moreover, it is required that the word lines WL and the bit lines BL is in an orthogonally positional relation, but any fluctuation in the word line direction and the bit line direction does not have to be taken into consideration. Therefore, about a positioning accuracy in the memory cells during fabrication, it is not necessary to be remarkably nervous, and the fabrication can easily be performed. Moreover, this structure is a highly integrated structure where one bit of information can be stored in a region of 2F2.
As shown, the word lines WL are alternately connected in common. In other words, the memory cell array includes two sets of word lines WL each having a comb-like structure, and a linear region along the second direction of the word lines WL alternately belong to one of the comb-like structures. Furthermore, the present structure can be considered as follows. That is, when the word lines WL are labeled as WL0, WL1, WL2, . . . WL7 in order from the right side of the drawing sheet of
It is to be noted that
2. Re: The Overall Structure of Semiconductor Memory Device
Next, the overall structure of a semiconductor memory device according to the present embodiment will be described with reference to
As shown, a semiconductor memory device 20 includes a memory cell array 21, a WL decoder 22, a GBL decoder 23, a selector decoder 24, a controller 25 and a power supply 26.
The memory cell array 21 his the structure described with reference to
The three-dimensional stacked type memory cell array of
Returning to
The GBL decoder 23 includes a global bit line selection unit and a global bit line driver. Moreover, the global bit line selection unit selects the global bit line GBL on the basis of a column address received from the controller 25. Moreover, the global bit line driver applies, to the selected global bit line and unselected global bit line, the voltage required to read, write and erase the data.
The selector decoder 24 includes a selector selection unit and a selection gate line driver. Moreover, the selector selection unit selects the selection gate line SSG on the basis of a sheet address received from the controller 25. Furthermore, the selection gate line driver applies, to the selected selection gate line and unselected selection gate line, the voltage required to read, write and erase the data.
It is to be noted that “the sheet” indicates a set of the memory cells selected by one of the selection gate lines. That is, the set of the memory cells which are present in the plane formed in the second direction and the third direction in
The controller 25 controls the operation of the whole semiconductor memory device 20. Moreover, the controller transmits the above-mentioned required address to the WL decoder 22, the GBL decoder 23 and the selector decoder 24. Furthermore, when the data is written, the controller 25 instructs the WL decoder 22, the GBL decoder 23 and the selector decoder 24 to apply the required voltage in order to change the resistance state of the variable resistance element of the selected memory cell MC. Moreover, when the data is read, the controller instructs the WL decoder 22, the GBL decoder 23, and the selector decoder 24 to apply the required voltage in order to detect the resistance value of the variable resistance element of the selected memory cell MC as a memory state of the memory cell MC. Furthermore, the controller 25 includes a not-shown sense amplifier, and senses and amplifies the data read through the global bit line GBL by a sense amplifier.
The power supply 26 generates a predetermined voltage set required to read, write and erase the data. The voltage generated by the power supply 26 is given to the word line WL and the bit line BL. For example, when the data is written, a large potential difference is applied between the selected word line and the selected bit line, so that the resistance state of the variable resistance element transits. Moreover, when the data is read, a potential difference is applied between the selected word line and the selected bit line in such a range that any transition of the resistance state does not occur, so that a current flowing through the bit lines or the word lines is detected.
It is to be noted that the CMOS circuit 32 excluding portions for connecting to the memory cell portions 34 can be designed and prepared by, for example, a design rule of 90 nm which is more relaxing than the memory cell portions 34. Moreover, the layer 33 includes portions for electrically connecting to the CMOS circuit 52 around each of the memory cell portions 34, and unit blocks including the memory cell portions 34 and the peripheral connecting portions are arranged in the matrix. Furthermore, in the layer 33, through holes are formed, and in ends of the layer 33, there are formed input/output portions 35 of the present device including terminals electrically coupled with input/output portions of the CMOS circuit 32 via these through holes.
According to such a structure, a function corresponding to a protection film of the CMOS circuit 32 can be performed by insulators formed in the memory cell portions 34. On the other hand, the memory cell portions 34 and the CMOS circuit 32 are coupled each other in a vertical direction to a substrate surface, which enables the shortening of an operation time or the noticeable increase of the number of simultaneously readable/writable cells without involving the increase of chip areas. It is to be noted that the input/output portions 35 of the device are bonded to a lead frame in a packaging step in the same manner as in a usual semiconductor device.
3. Re: Operation
Next, an operation of the semiconductor memory device according to the present embodiment will be described.
In the following description, the selected global bit lines GBL is labeled as GBL_s, and the unselected global line is labeled as GBL_u. Moreover, the selected word lines WL is labeled as WL_s, and the unselected word line is labeled as WL_u. Further among the selection gate lines SSG, two selection elements SS corresponding to the bit line BL connected to the selected memory cell MC are selected, and labeled as SSG_s and SSG_n. The other selection gate lines SSG are unselected, and labeled as SSG_u.
3.1 Writing Operation
First, a writing operation of storing information in the memory cells will be described.
During the writing operation, the GBL decoder 23 applies a writing voltage Vw (>0 V) to the selected global bit line GBL_s, and applies a half (Vw/2) of the writing voltage to the unselected global bit line GBL_u.
Moreover, the WL decoder 22 applies 0 V to the selected word line WL_s, and applies (Vw/2) to the unselected word line WL_u.
Furthermore, the selector decoder 24 applies a writing gate voltage Vg_w (>0 V) to both of two selection gate lines SSG_s and SSG_n, and applies 0 V to the other selection gate lines SSG_u.
Consequently, in the selection element SS connected to the selected bit line BL, a channel is formed by the two selection gate lines SSG_s and SSG_n, and the line GBL_s transfers a writing voltage Vw to the selected memory cell MC. On the other hand, 0 V is transferred from WL_s to the selected memory cell MC. In consequence, when a potential difference Vw is applied across the variable resistance element of the memory cell MC, the data is written in the memory cell MC.
3.2 Erasing Operation
Next, an erasing operation of information held in the memory cells will continuously be described with reference to
During the erasing operation, considering that the element performs a bipolar operation, the WL decoder 22 applies a voltage (Ve1), obtained by adding an offset voltage of 1 V to an erasing voltage Ve, to the selected word line WL_s, and applies ((Ve/2)+1) to the unselected word line WL_u.
Moreover, the GBL decoder 23 applies the offset voltage of 1 V to the selected global bit line GBL_s, and applies ((Ve/2)+1) to the unselected global bit line GBL_u.
Furthermore, the selector decoder 24 applies an erasing gate voltage Vg_e to both of the selection gate lines SSG_s and SSG_n, and applies 0 V to the other selection gate lines SSG_u.
In consequence, the voltage Ve is transferred to the selected memory cell MC in the same manner as in the writing. Moreover, when a potential difference Ve is applied across the variable resistance element, the data is erased.
3.3 Reading Operation
Next, an operation of reading the information from the memory cells will continuously be described with reference to
During the reading operation, the GBL decoder 23 applies a voltage (Vr+Vo), obtained by adding an offset voltage Vo to a reading voltage Vr, to the selected global bit line GBL_s and the unselected global bit GBL_u.
Moreover, the WL decoder 22 applies the offset voltage Vo to the selected word line WL_s, and applies (Vr+Vo) to the unselected word line WL_u.
Furthermore, the selector decoder 24 applies a reading gate voltage Vg_r to both of the selection gate lines SSG_s and SSG_n, and applies 0 V to the other selection gate lines SSG_u.
In consequence, the voltage Vr is transferred to the selected memory cell MC in the same manner as in the writing. Moreover, the data read from the selected memory cell MC is transferred to GBL_s via the selection element SS.
4. Effect of the Present Embodiment
According to the structure of the present embodiment, an integration degree of the memory cell array can be enhanced. Hereinafter, the present effect will be described in detail.
In recent years, with the high integration of a semiconductor device, a circuit pattern of an LSI element has increasingly been miniaturized. This miniaturization of the pattern requires not only the reduction of a line width but also the enhancement of a dimensional or positional accuracy of the pattern. This also applies to the memory device. As to the memory device, it has continuously been requested that In memory cells formed by making full use of an accurate processing technology, a predetermined amount of a charge required for storage is held in a smaller region.
Heretofore, various types of memories such as a DRAM, SRAM and a flash memory have been fabricated. All these memories hold the predetermined amount of the charge to store data. Therefore, with the miniaturization of the pattern, there are severe restrictions on fluctuations of the memory cells, and the like. In consequence, a large burden is also imposed on a lithography technology which forms these patterns, which becomes a factor to raise a lithography step cost. Moreover, the lithography step cost occupies a large part of the present mass production cost. Therefore, the burden on the above lithography step becomes the factor for raising a product cost as it is.
On the other hand, in recent years, there has been suggested a memory called an ReRAM in which each memory cell is formed by a non-ohmic element typified by a diode and a variable resistance material, as a technology which conquers such a problem. In this ReRAM, it is not necessary to use an MOSFET in the memory cell, and hence it has been expected that higher integration in excess of a conventional trend can be achieved. Furthermore, the ReRAM has a structure which facilitates three-dimensional stacking. Therefore, it is expected that the integration degree can noticeably be enhanced as compared with a conventional memory in which an only two-dimensional plane is utilized.
However, when a hierarchical bit line structure where bit lines are connected to a global bit line is used, it is actually difficult to form the memory cell with a size of 2F2. This is caused by the selection element SS. The selection element SS switches the connection between the bit line BL and the global bit line GBL, and is usually realized by an MOS transistor.
The selection element SS is positioned immediately under the memory cell MC. Therefore, if the memory cell MC is formed with the size of 2F2, the size of the selection element SS, accordingly, has to be reduced. However, when a usual planar type MOS transistor realizes the selection element SS, it has been difficult to satisfy this requirement.
In the present embodiment, however, the source region 5, the channel region 6 and the drain region 7 are stacked in the same width as that of the bit line BL, and the gates SSG are provided on side surfaces, to realize the selection element SS. In consequence, a size of the selection element SS per bit line BL becomes 4F2. This behavior is shown in
Consequently, the size of the selection element SS is reduced, therefore the size of the memory cell MC can be decreased, so that the integration degree of the memory cell array can be enhanced.
Moreover, according to the present embodiment, any negative voltage is not used, but operation reliability can be enhanced. This is because the offset voltage is used during the erasing operation as described in the above section 3.2. That is, when the potential of GBL_s is set to a value which is about 1 V higher than SSG_u, it is possible to noticeably save a leak current to the unselected cell, because of characteristics of the selection element SS. In this case, when a reference of a bias is set to a positive voltage of 0 V to 1 V, the use of the negative voltage can be avoided. When the negative voltage is used, a circuit for generating the negative voltage is required, and an area of the circuit is comparatively large. In the present embodiment, however, this negative voltage generation circuit is not used, but the leak current is lowered, so that the operation reliability of the semiconductor memory device can be enhanced. It is to be noted that the offset voltage during the erasing is not limited to 1 V, and can appropriately be selected in accordance with a requested performance or the like.
Furthermore, according to the present embodiment, it is possible to realize the lowering of power consumption and/or the raising of an operation speed. This is because the offset voltage is used during the reading operation as described above in section 3.3. Usually, the data is read from the memory cell immediately after performing the writing or the erasing, and it is confirmed whether or not the memory cell obtains a desired resistance value. In consequence, when the resistance value is different from the desired range, the writing or the erasing is additionally performed. Moreover, the voltage required for the writing or the erasing is larger than the voltage required for the reading.
Thus, according to the present embodiment, the reference of the bias during the reading is set at a positive voltage Vo in place of 0 V. That is, a value of the voltage for use in the reading is brought close to that of the voltage for use in the writing and erasing. Therefore, when the device transitions from the writing and erasing operation to the reading operation, or in the opposite case, a large voltage difference is prevented from being made between bias conditions of both the operations. Moreover, a voltage change of a node with a large parasitic capacity can be lowered. Therefore, the increase of an unnecessary power consumption and the delay of an operation time can be suppressed.
It is to be noted that the parasitic capacity of the global bit line GBL is largest in the present embodiment. Therefore, the voltage to be applied to the global bit line GBL is preferably set so that the values thereof become as equal as possible during the writing or erasing and during the reading.
The number of the unselected global bit lines is larger than that of the selected global bit lines. Therefore, the offset voltage Vo may be preferably set so that (Vw/2) becomes substantially equal to (Vr+Vo) during the writing and ((Ve/2)*1) becomes substantially equal to (Vr+Vo) during the erasing, whereby the potential of the unselected global bit line does not noticeably change.
It is to be noted that only one word line WL is usually selected from the memory cell array, but the global bit lines GBL may simultaneously be selected. In consequence, the number of simultaneously writable/eraseable/readable bits increases, and a band width can be increased.
Next, a memory device according to a second embodiment will be described. The present embodiment relates to a fabricating method of an ReRAM described in the above first embodiment.
1. First Fabricating Method
First, a first fabricating method will be described with reference to
First, a usual CMOS circuit 32 which controls an operation of the ReRAM is formed on, for example, a silicon substrate 31. Next, an interlayer insulating film is formed on the silicon substrate 31 to cover the CMOS circuit 32.
Next, as shown in
Next, as shown in
Next, as shown in
Next, as shown in
Next, as shown in
Next, as shown in
Next, as shown in
Next, as shown in
Next, as shown In
Next, as shown in
Next, as shown in
Next, as shown in
Next, as shown in
Afterward, an interlayer insulating film fills in each groove between the adjacent bit lines BL, to complete a memory cell array. Continuously, there are performed a step of connecting the word lines WL and the bit lines BL to the CMOS circuit 32, and the like. Furthermore, a passivation step is performed in the same manner as in a conventional semiconductor device, and an interconnect connecting portion which becomes an input/output portion is further formed. Finally, so-called post steps of inspection, dicing and the like are performed, to complete the ReRAM described in the first embodiment.
2. Re: Second Fabricating Method
Next, a second fabricating method which is different from the first fabricating method will be described with reference to
First, as shown in
Next, as shown in
Next, as shown in
Next, as shown in
Next, a memory device according to a third embodiment will be described. The present embodiment relates to data reading and writing methods of the ReRAM described in the above first embodiment.
1. Data Writing Method
First, the writing method of the data according to the present embodiment will be described.
As described in the first embodiment, when a word line group WLcomb_a is selected, for example, odd word lines WL1, WL3, WL5 to WL(n−1) (n is an even number of 4 or larger) are selected. On the other hand, when a word line group WLcomb_b is selected, for example, even word lines WL0, WL2, WL4 to WL(n−2) are selected. Moreover, two memory cells MC are connected between one global bit line GBL and one word line WL. Therefore, one word line group WLcomb including (n/2) word lines WL is selected, to select n memory cells MC0 to MC(n−1).
The data is written in the n memory cells MC, for example, in order. The reference characters c0 to c(n−1) of
As shown, the word lines WL alternately belong to the word line groups WLcomb_a and WLcomb_a. Therefore, one of the word lines at both ends of the word line group WLcomb is sandwiched between two word lines of the other word line group WLcomb, and the other word line is not sandwiched. More specifically, the word lines at both the ends of the word lines belonging to the word line group WLcomb_a are the word lines WL1 and WL(n−1). One (WL1) of these word lines is sandwiched between two word lines WL0 and WL2 belonging to the other word line group WLcomb_b. However, the other word line (WL(n−1)) is adjacent only to the word line WL(n−2) belonging to WLcomb_b, and is not sandwiched.
According to such a structure, the data c0 is written in the memory cell MC0 connected to the word line WL at the end of the word line group WLcomb and positioned at inner side of the other word line group WLcomb. In other words, the data c0 is written in the memory cell MC0 formed on the side surface of the word line (WL1) at the end of the word line group (e.g., WLcomb_a) and facing the word line (WL0) at the end of the other word line group (WLcomb_b). Further in other words, the data c0 is written in the memory cell MC0 formed on the side surface of the word line (WL1) at the end of the word line group (e.g., WLcomb_a) and positioned between this word line (WL1) and the word line (WL0) at the end of the other word line group (WLcomb_b). Moreover, the data piece c0 among data pieces (c0, c1, . . . c(n−1)) is first written.
The data (c1, c2, . . . c(n−1)) is the data to be written, in order, in the memory cell MC selected in accordance with the word line WL number.
Next, the controller 25 converts the data pieces d0 to d(n−1) (step S11). This conversion rule is as follows.
That is, the data c0 is the data d0 itself transmitted from the host apparatus. On the other hand, the data (c1, c2, . . . c(n−1)) is a value obtained by subtracting the data (c0, c1, . . . c(n−2)) concerning the adjacent word line from the data (d1, d2, . . . d(n−1)) transmitted from the host apparatus. In consequence, the controller 25 obtains the data (c0, c1, . . . c(n−1)). However, when calculation results of the steps S13 and S14 are negative or are not smaller than the maximum value of the data d (“2” in the case of 1-bit data, “4” in the case of 2-bit data, and “8” in the case of 3-bit data), a (mathematically mod (the maximum value of the data values)) procedure to add or subtract the maximum value of the data d is performed.
Then, the controller 25 writes the obtained data (c0, c1, . . . c(n−1)) in the memory cell MC (step S15).
Example 1 in the diagram shows the case of the data value (d0, d1, . . . d7)=(1, 1, 0, 0, 1, 1, 0, 1). In this case, the cell value (c0, c1, . . . c7)=(1, 0, 0, 0, 1, 0, 0, 1).
Example 2 in the diagram snows the case of the data value (d0, d1, . . . d7)=(1, 0, 0, 1, 1, 0, 1, 0). In this case, the cell value (c0, c1, . . . c7)=(1, 1, 1, 0, 1, 1, 0, 0). The value enclosed in a circle mark in the diagram corresponds to a case where “+1” is obtained from an equation shown in the step S14. However, when the maximum value “2” of the 1-bit data is added (mod(2) calculation is performed), “1” is obtained.
Example 3 in the diagram shows the case of the data value (d0, d1, . . . d7)=(1, 3, 1, 2, 3, 0, 2, 0). In this case, the cell value (c0, c1, . . . c7)=(1, 2, 3, 3, 0, 0, 2, 2). The value enclosed in a circle mark in the diagram corresponds to a case where a negative value is obtained from the equation shown in the step S34. However, when the maximum value “4” of the 2-bit data is added (mod(4) calculation is performed), a positive value of 3 or smaller is obtained.
It is to be noted that biases of interconnects when the data is written are shown in
2. Data Reading Method
Next, the data reading method will be described. The data reading method is different from the method described with reference to
3. Re: Specific Examples
Next, specific examples of the above writing operation and reading operation will be described.
3.1 Writing Operation
First, the writing operation will be described. As shown in, for example,
c0=d0=1
c1=d1−c0=1−1=0
c2=d2−c1=0−0=0
c3=d3−c2=0−0=0
c4=d4−c3=1−0=1
Therefore, (c0, c1, c2, c3, c4)=(1, 0, 0, 0, 1) are written in the memory cells MC0 to MC4, respectively.
3.2 Reading Operation
Next, the reading operation will be described. First, there will be described a case where the data is read from the memory cell MC0, with reference to
As shown, when the data is read from a memory cell MC0, a selection gate line SSG0 is selected from two selection gate lines SSG0 and SSG1 of a selection element SS0 corresponding to a selected bit line BL0, and a reading voltage Vg_r is applied to the line. The other selection gate lines SSG1 to SSG(n−1) are unselected, and 0 V is applied to the lines. The other voltages have been described with reference to
As a consequence, a channel is formed in the selection element SS0 by the selection gate line SSG0. Consequently, the bit line BL0 is electrically connected to the global bit line GBL. Any channel is not formed in the other selection elements SS1 to SS(n−1), and hence the other bit lines BL0 to BLn are electrically separated from the global bit line GBL.
Therefore, a current I0 flows from the global bit line GBL to the word line WL1 through the variable resistance material 4 of the selected memory cell MC0. This current I0 has a value corresponding to the data c0 (=d0) written in the memory cell MC0. This current is sensed and amplified, to read the reading data=“1” (=d0).
Next, there will be described a case where the data is read from the memory cell MC1, with reference to
As shown, when the data is read from a memory cell MC1, a selection gate line SSG1 (the selection gate line closer to the memory cell MC0 in which the cell value c0 is written) is selected from two selection gate lines SSG1 and SSG2 of a selection element SS1 corresponding to a selected bit line BL1, and a reading voltage Vg_r is applied. The other selection gate lines SSG0 and SSG2 to SSG(n−1) are unselected, and 0 V is applied. The other voltages have been described with reference to
Then, channels are formed in not only the selection element SS1 but also a selection element SS0 connected to an unselected bit line BL0. Consequently, two bit lines BL0 and BL1 are electrically connected to a global bit line GBL.
Therefore, a current I0 flows from the global bit line GBL to a word line WL1 through a variable resistance material 4 of an unselected memory cell MC0. This current I0 has a value corresponding to the data c0 (=d0) written in the memory cell MC0. Furthermore, a current I1 flows from the global bit line GSL to the word line WL1 through the variable resistance material 4 of the selected memory cell MC1. This current I1 has a value corresponding to data c1 (−d1−c0) written in the memory cell MC1.
Then, the controller 25 senses and amplifies the sum of the above two currents I0 and I1. Then, as shown in
Next, there will be described a case where data is read from the memory cell MC2, with reference to
As shown, when the data is read from the memory cell MC2, a selection gate line SSG2 (the selection gate line closer to a memory cell MC1 in which a cell value c1 is written) is selected from two selection gate lines SSG2 and SSG3 of a selection element SS2 corresponding to a selected bit line BL2, and the reading voltage Vg_r is applied. The other selection gate lines SSG0 and SSG1 and SSG3 to SSG(n−1) are unselected, and 0 V is applied. The other voltages have been described with reference to
Then, channels are formed in not only the selection element SS2 but also a selection element SS1 connected to an unselected bit line BL1, when the selection gate line SSG2 is selected. Consequently, two bit lines BL1 and BL2 are electrically connected to a global bit line GBL.
Therefore, a current I1 flows from the global bit line GBL to a word line WL1 through a variable resistance material 4 of the unselected memory cell MC1. This current I1 has a value corresponding to the data c1 (=d1−c0) written in the memory cell MC1. Furthermore, a current I2 flows from the global bit line GBL to a word line WL2 through the variable resistance material 4 of the selected memory cell MC2. This current I2 has a value corresponding to data c2 (=d2−c1) written in the memory cell MC2.
Then, the controller 25 senses and amplifies the sum of the above two currents I1 and I2. Then, as shown in
The data is read from the other memory cells MC3, MC5 and the like in the same manner as in
In consequence, as shown in
4. Effect by the Present Embodiment
According to the present embodiment, interference with the adjacent cell can be suppressed, and a data reading accuracy can be enhanced. Hereinafter, the present effect will be described in detail.
According to the method of the above-mentioned first embodiment, the voltage Vg_r is applied to two selection gate lines SSG_s and SSG_n. Also by such a method, the data can be read. When such a method is used, the selection element SS has one of the following three states in accordance with voltage states of the two selection gate lines SSG.
In the OFF2-element, a potential which is 1 V or more lower than that of the global bit line is securely supplied to the selection gate lines on both the sides as described above. Therefore, the OFF2-element has a complete OFF-state, and a leak current of the order of pA can only flow.
On the other hand, a channel is formed by the selection gate line set to the high potential, in the ON-element and OFF1-element. Therefore, a certain degree of large current flows.
Moreover, a graph G1 of
During the data writing and erasing, a half voltage is also applied to a half-selected cell, and a half-selected current flows through the cell. Therefore, as shown in
On the other hand, the voltage/current applied to the cell during the reading is set to be sufficiently smaller than that during the writing/erasing. This is because so-called read disturb is suppressed. Additionally, it is preferable to use a region where a gradient of Id-Vd characteristics is as large as possible. A reason is that when a region where the gradient of the Id-Vd characteristics of the selection element is small (so-called saturated region) is an operation point, a difference in detection current, with respect to a difference in cell resistance value, becomes small, and a detection sensitivity deteriorates. Therefore, as shown in
As shown in
In the example of
Then, the current flowing via the OFF2-element is remarkably small, and can be ignored. Therefore, the current flowing through the global bit line GBL is the sum of the current flowing through the selected cell via the ON-element and the current flowing through two half-selected cells (these memory cells are to be unselected cells) via the OFF1-element. Needless to say, a value which can be detected in a circuit is an only value of the current flowing into the global bit line GBL from the outside of the array, and hence the current flowing via the OFF1-element becomes a background noise.
As described above, in bias conditions for the reading, a cell resistance is comparatively large, and the resistance of the selection element S3 is comparatively small. Therefore, the selection element SS does not depend on the ON-element or the OFF1-element, but the current value is determined mainly by the size of the cell resistance sometimes. For example, even when the selected cell has a high resistance, the half-selected cell in a path of the current flowing via the OFF1-element has a low resistance. In this case, the current value detected through the global bit line might be a high current. This inevitably becomes a cause for erroneous reading.
In this respect, according to the present embodiment, the data d0 received from the host apparatus is written, as it is, in the memory cell MC (the memory cell MC0 of
Moreover, the current flowing through the global bit line GBL is measured, during the reading, in a state where one of the two selection gate lines SSG of the selection element SS directly connected to the bit line BL belonging to the selected cell is provided with a high potential, as described with reference to
In this case, the selection gate line SSG provided with the high potential satisfies the following conditions. That is, the data written in the memory cell MC which becomes a reading object is a cell value c(k+1). Moreover, this memory cell is labeled as MC(k+1). Then, the selection gate line SSG which satisfies both of the following two conditions is selected.
That is, when the bit line BL connected to the selected memory cell MC(k+1) is BL(k′+1), the selection gate line SSG positioned between the bit line BL(k′+1) and the bit line BLk′ is provided with the high potential.
Therefore, the current flowing through the global bit line GBL is a value corresponding to the data of the memory cells MC(k+1) and MCk. That is, the value corresponds to (c(k+1)+ck). Then, a relation of d(k+1)=c(k+1)+ck, and hence the value of this current flowing through the global bit line GBL corresponds to the data value d(k+1) as it is. That is, when the current is detected once, it is possible to read each data value while suppressing the influence of the adjacent unselected memory cell (the half-selected cell). Therefore, a high-accuracy and high-speed reading operation is enabled.
It is to be noted that the selection gate line SSG to which a high potential (Vg_w, Vg_e) is applied during the data writing and erasing may be two selection gate lines described in the first embodiment, or one selection gate line described at the reading time of the present embodiment.
Next, a memory device according to a fourth embodiment will be described. The present embodiment relates to a data writing method of the ReRAM described in the above first to third embodiments, and relates to the order of memory cells MC in which data is to be written.
1. Re: Writing Method
Furthermore, the controller 25 sets a word line address to a WL decoder 22. In consequence, the WL decoder 22 selects a word line WI. (step S22). Then, the WL decoder 22 applies 0 V to the selected word line WL_s (step S23). Furthermore, the decoder applies Vw/2 to an unselected word line WL_u.
Moreover, the controller 25 converts writing-data as described in the third embodiment (step S24). That is, the controller 25 converts a data value d received from a host apparatus into a cell value c.
Then, the selector decoder 24 sets a selection gate line address (the sheet selector address) to a selector decoder 24. In consequence, the selector decoder 24 selects a selection gate line SSG (a selection element SS (step S25). This selection/non-selection is performed in accordance with the cell value obtained In the step S24. Then, the selector decoder 24 applies Vg_w to the selected selection gate line SSG for a predetermined pulse time (step S26). A potential of the unselected selection gate line SSG is set to 0 V. In this case, when Vg_w is applied to the selection gate line SSG, the corresponding selection element SS is turned on. Therefore, a potential difference Vw is applied to a variable resistance material 4 of the selected memory cell, and the data is written in the cell. It is to be noted that a state where a high potential is applied to two selection gate lines SSG_s and SSG_n is preferably utilized to impart a sufficient current supply ability to the selection element SS and to prevent wrong writing in the unselected cell.
The controller 25 successively performs the data writing of the above steps S25 and S26 in the memory cells MC connected to the selected word lines WL (a selected word line group WLcomb) (step S27). Moreover, when the writing in all the memory cells MC connected to the selected word line group WLcomb is completed (the step S27, YES), the controller 25 selects another word line group WLcomb (step S28), to write the data in the same manner (the steps S22 to S27). In this case, not only the word line group WLcomb which is present in the same layer but also the word line group WLcomb which is present in a different layer may be selected. However, even when the selected word line group WLcomb changes, the selected global bit line GBL does not change.
When the writing through the predetermined word line group WLcomb in accordance with the selected global bit line GBL is completed (the step S26, YES), the controller 25 verifies that the data has correctly been written (step S29), to end the writing operation.
It is to be noted that another global bit line GBL nay be selected to perform the writing after the step S29 or before the step S29 (after the step S28). That is, the controller nay return to the step S20 after the step S28 or S29.
2. Re: Specific Example
A specific example of the above writing method described with reference to
As shown, a global bit line GBL0 is first selected. In this state, word line groups WLcomb_a1, WLcomb_b1, WLcomb_a2 and WLcomb_b2 are successively selected. In the diagram, (1) to (4) indicate that the word line groups WLcomb are not simultaneously selected, but are successively selected. Every time each word line group WLcomb is selected, the selection gate line SSG is selected in accordance with a cell value, thereby writing the cell value c in the memory cell MC.
Behaviors for writing the data in the memory cell MC are shown in
Afterward, the word line groups WLcomb of a second layer and upper layers are similarly selected, and the data is written in the memory cells MC connected to the groups. Also in this case, the global bit line GBL0 remains to be selected.
When the writing of the data in the predetermined memory cell MC in the group G0 is completed, the data is next written in the predetermined memory cell MC of the group G1. This behavior is shown in
As shown, a global bit line GBL1 is first selected. In this state, the word line groups WLcomb_a1, WLcomb_b1, WLcomb_a2 and WLcomb_b2 are successively selected. Then, every time the word line group WLcomb is selected, the selection gate line SSG is selected in accordance with the cell value, thereby writing the cell value c in the memory cell MC, as described with reference to
In consequence, the global bit line GBL, the word line group WLcomb and the selection gate line SSG need to be selected, to select the memory cell MC. In this case, the global bit line GBL is finally selected in the present embodiment.
3. Effect of the Present Embodiment
According to the present embodiment, a writing operation can be speeded up. Hereinafter, the present effect will be described.
The global bit line is disposed to extend through the overall area of the chip (or a half area, ¼ area, or block area) over a large number of memory cell arrays. Additionally, the global bit line has a substantially minimum dimension, to enhance a chip integration degree. Therefore, the global bit line becomes an interconnect having a length of several millimeters or larger, and a line width and an adjacent space of several tens of nanometers or smaller. Therefore, a product of a parasitic capacity and an interconnect resistance becomes approximately several hundreds of nanoseconds (e.g., when an interconnect length is 3.2 mm, an interconnect width and an adjacent space are 24 nm and a sheet resistance is 1.5Ω/□, an interconnect capacity becomes up to 1.6 pF and an interconnect resistance becomes up to 200 kΩ. Therefore, a CR product becomes up to 320 nsec).
On the other hand, local or sub-local (within several hundreds of micrometers at the highest) interconnects are used in the word line group and the selection gate line. Therefore, the product of the interconnect resistance and the parasitic capacity becomes about hundred nsec or smaller, and becomes a fraction of the global bit line or smaller.
Therefore, a time required to stabilize the potential when the potential of the global bit line is changed is considerably longer than a time required to stabilize the potential when the potential of the word line or the selection gate line is changed.
When a certain global bit line GBL is selected by the method of the present embodiment under such a situation, the word line group WLcomb and the selection gate line SSG are selected to successively select all the required memory cells while maintaining the state where the global bit line GBL is selected. Then, the next global bit line GBL is selected after selecting all the memory cells MC which are connected to the global bit line GBL and in which it is necessary to write the data. That is, the number of times to change the potential of the global bit line GBL set to be smaller than the number of times to change the potential of the word line group WLcomb and the selection gate line SSG as much as possible. Therefore, a waiting time to stabilize the potential of the interconnect can be saved, and the writing operation can be speeded up.
Hereinafter, a comparative example will specifically be described.
To simplify the description of such a comparative example, there is considered a case where details of a data structure are Ignored and the data is written in all the memory cells of one memory cell array. Then, when the method of comparative example is used, a time required to write the data in all the memory cells is substantially represented by the following equation.
NW(NS(NG(tG+tP)+tS)+tW),
where NG is the number of the global bit lines in the memory cell array, NS is the number of the selection gate lines, NW is the number of word line groups (=the number of layers of word lines×2), tG, tS and tW are times required to stabilize the potentials of the global bit line, the selection gate line and the word line group, respectively, and tP is a pulse applying time.
As described above, tG is longest among tG, tS and tK. Moreover, in the method of the comparative example, at least a time which is (Nw×NS×NG) times as long as tG is required, and a very long time is required to write the data.
On the other hand, in the method of the present embodiment, a time required to write the data in all the memory cells is substantially represented by the following equation.
NG(NW(NS(tS+tP)+tW)+tG).
That is, a coefficient of tG with the longest time to stabilize the potential is NG only. Therefore, it is possible to noticeably shorten the time as compared with the comparative example.
More specifically, when a required time is estimated with NG=72, NS=16, NW=32, tG=300 nsec, tS=100 nsec, tW=100 nsec and tP=50 nsec, about 13.0 msec is required in the comparative example, whereas in the present embodiment, about 5.8 msec is required and the time becomes half or shorter.
4. Modification Example of the Present Embodiment
Hereinafter, a modification example of the present embodiment will be described.
However, a flowchart shown in
Moreover, the loop processing concerning the global bit lines GBL is not necessarily positioned on the outermost side.
Moreover, in the above embodiment, there has been described an example where the data writing and reading methods described in the third embodiment are applied. That is, data conversion is performed in the step S24. However, the method described in the first embodiment may be applied in place of the method described in the third embodiment. In this case, the step S24 is not required.
Next, a memory device according to a fifth embodiment will be described. The present embodiment relates to data mapping for performing the writing and reading of data by a page unit in the ReRAM described in the above third embodiment.
1. Re: Mapping of Data
There will be described an example where the data of one page described with reference to
First, the blocks which are present in the same row of the memory cell array are selected (called a selected blocks row). Specifically, the blocks are selected by using a global word line. The global word line is an interconnect connected in common to gates of MOSFETs which become drivers of word line groups WLcomb_a and WLcomb_b of the blocks present in the same row. The selected blocks row includes 4608 blocks as described above. The 4608 blocks are divided into 16 groups each including 238 blocks. The number of the blocks included in each group is the number of the simultaneously accessible blocks (288 blocks in the present example) during the reading. Furthermore, 288 blocks included in each group are divided into 18 columns each including 16 blocks. The number of the blocks included in each column is the number of the simultaneously accessible blocks (16 blocks in the present example) during the writing.
The data of one page is mapped in a bit group specified by a selected column [0:17], a selected block [0:15] and a selected sheet selector (the selection gate line) [0:15], in a region specified by the selected blocks row [0:4095], a selected group [0:15], a selected global bit line [0:63] and a selected word line [0:31] (because in each block, the number of the word line groups WLcomb is two, and the layer number of the word lines WL is 16 layers) in the memory cell array. In this case, a selected column address corresponds to the page data index i, a selected block address corresponds to the page data index j, and a selected gate line address corresponds to the page data index k.
The above groups and columns will specifically be described with reference to
As shown, a group number is 16 groups. Then, blocks BLK(16h*g)(h=0 to 287) are assigned to a group GRg (g=0 to 15). Therefore, 288 blocks BLK0, BLK16, BLK32, . . . BLK4592 belong to the group GR0. 288 blocks BLK1, BLK17, BLK33, . . . BLK4593 belong to the group GR1. 288 blocks BLK2, BLK18, BLK34, . . . BLK4594 belong to the group GR2. Afterward, 288 blocks BLK15, BLK31, BLK47, . . . BLK4607 similarly belong to the final group GR15.
Then, the blocks are assigned to a column Cq (q=0 to 17) by a unit of 16 blocks in each group. Blocks BLK(283r+16q+g)(r=0 to 15) are assigned to the column Cq. Therefore, 16 blocks BLK0, BLK238, BLK576, . . . BLK4320 are assigned to the column C0 of the group GR0. 16 blocks BLK16, BLK304, BLK592, . . . BLK4336 are assigned to the column C1 of the group GR0. 16 blocks BLK32, BLK320, BLK608, . . . BLK4342 are assigned to the column C2 of the group GR0. Afterward, 16 blocks BLK272, BLK560, BLK848, . . . BLK4592 are similarly assigned to the final column C17 of the group GR0.
Moreover, 16 blocks BLK1, BLK289, BLK577, . . . BLK4321 are assigned to the column C0 of the group GR1. 16 blocks BLK17, BLK305, BLK593, . . . BLK4337 are assigned to the column C1 of the group GR1. 16 blocks BLK33, BLK321, BLK609, . . . BLK4343 are assigned to the column C2 of the group GR1. Afterward, 16 blocks BLK273, BLK561, BLK849, . . . BLK4593 are similarly assigned to the final column C17 of the group GR1.
The same applies to the other groups GR3 to GR15.
2. Writing Operation of Data
A data writing operation by using the above mapping will be described.
During the writing, the data of 16 bits belonging to a selection gate line in the selected column is simultaneously written. In this case, the data of one page is stored once in the page register (included in, e.g., the controller 25) in the chip, and the corresponding index is attached thereto. Afterward, the data is taken in the corresponding data order. i.e., each data for 16 bits of j[0:15] belonging to the same (i, k) is taken. Afterward, after the conversion of the data value and the cell value is performed, the data is written in the cells of 16 bits belonging to the same selection gate line. After the loop of the selected gate line, the loop processing of the selected column is performed, to end the writing of the data of one page.
The behavior of the data writing is shown in
When the data of pages is written, a series of processing may be expanded, in order, to different selected word line loop, selected global bit line loop, selected group loop and selected blocks row loop.
3. Data Reading Operation
Next, the reading operation will be described. During the reading, the data of 288 bits belonging to the selection gate line SSG is simultaneously read, and the page data is successively read In a selected gate line address order. As described above, the page data corresponds to the selection gate line address, and configured in order of the index k. Therefore, when the loop of the selection gate line is repeated, the data of one page can be reconstructed in the original order.
When the pages of the data are read, the series of processing nay be expanded to different selected word line loop, selected global bit line loop, selected group loop and selected blocks row loop, in order, in the same manner as in the writing.
4. Effect of the Present Embodiment
When the data in one page is processed with different simultaneously parallel blocks during the writing and the reading and the cells successively selected by the innermost loop are arranged in a direction (along the GBL) different from a parallel block direction during the writing and the reading, the mapping as in the present embodiment is preferably performed. That is, the data can be taken in order from the top bit of the page data during the reading, and hence a reading latency can be shortened.
It is to be noted that as in the step S29 of
In this case, the above-mentioned data of one page is used as a unit. After writing each page of data, the same page is read. Then, a read result is stored in another page register, and values of both the page registers may be compared. Moreover, when a non-matched portion is present, the data is written again. After repeating this rewriting processing, the processing may advance to the processing of the next page (the selected word line loop, the selected global bit line loop, the selected group loop and the selected blocks row loop), if necessary.
Next, a memory device according to a sixth embodiment will be described. Unlike the above third embodiment, the present embodiment increases the number of word line groups WLcomb from two to three or more to suppress the occurrence of erroneous reading.
1. Re: Structure of Word Line WL
As shown, according to the structure described in the first embodiment, the word lines in a block belong to one of two word line groups WLcomb_a and WLcomb_b as shown in, for example,
On the other hand, according to the structure of the present embodiment, word lines in a block belongs to one of four word line groups WLcomb_a, WLcomb_b, WLcomb_c and WLcomb_d. Moreover, two word lines WL belonging to the same word line group WLcomb_a are disposed adjacent, to sandwich two word lines WL belonging to the other word line groups WLcomb_b and WLcomb_c, respectively. The same also applies to the word lines WL belonging to the word line group WLcomb_b, WLcomb_c and WLcomb_d. In other words, both sides of the word line belonging to the word line group WLcomb are sandwiched between the other word lines, two word lines which sandwich the word line belong to the word line groups WLcomb which are different from the word line group WLcomb to which the corresponding word line belongs and which are different from each other.
In an example of, for example,
2. Re: Data Reading Method
Next, a data reading method will be described. Biases of interconnects during the data reading are similar to those of the third embodiment. Moreover, as to selected selection gate lines SSG, the selection gate lines SSG facing the selected memory cell MC in the first direction via the selected bit line BL are selected.
In other words, in two selection gate lines of a selection element SS connected to the selected bit line BL, the selection gate line SSG is unselected which forms a current path between a global bit line GBL and the memory cell MC disposed between the selected word line WL and the unselected bit line BL, and the selection gate line SSG is selected which forms a current path between the global bit line GBL and the memory cell MC disposed between the unselected word line WL and the unselected bit line BL.
In the present example, data is read from a memory cell MC connected to the word line WL1 and a bit line BL1. Selection gate lines which form a channel in a selection element SS connected to the bit line BL1 are two selection gate lines SSG1 and SSG2. In these selection gate lines, the selection gate line SSG facing the selected memory cell in the first direction via the bit line BL1 is the selection gate line SSG2. That is, the current path is formed between the global bit line GBL and the memory cell disposed between the selected word line WL1 and an unselected bit line BL0, by the selection gate line SSG1, but the current path is not formed by the selection gate line SSG2. Therefore, a selector decoder 24 selects the selection gate line SSG2, applies, for example, a voltage Vg_r thereto, and applies 0 V to the other selection gate lines.
Consequently, a cell current flows through the selected word line WL1 from the global bit line GBL via a channel formed by the selection gate line SSG2, the bit line BL1 and the selected memory cell MC.
3. Re: Data Writing Method and Erasing Method
A data writing method is similar to the third embodiment except that data conversion is not required. That is, only one selection gate line SSG described in the above 2 may be selected. The same applies to an erasing time. Needless to say, two selection gate lines SSG may be selected in the same manner as in the first embodiment.
4. Effect of the Present Effect
Also according to the present embodiment, a data reading accuracy can be enhanced while suppressing interference with adjacent cells. Hereinafter, the present effect will be described in detail.
In the present example, three adjacent word lines belong to mutually different word line groups WLcomb. In other words, different word line drivers apply a voltage to these word lines.
When one selection gate line SSG is selected during the data reading, a current path is formed between two bit lines BL and the global bit line GBL. Moreover, these two bit lines BL are associated with three word lines WL. However, these three word lines WL belong to the mutually different word line groups WLcomb, and two of the lines are unselected. Therefore, any potential difference is not made in the current path formed for the unselected memory cell MC, and any current substantially does not flow through this current path. Moreover, the current substantially flows only through the current path formed for the selected memory cell MC.
This respect will be described again with reference to
Consequently, a potential between both ends of each of the unselected memory cell MC connected to the unselected word line WL2 and the bit line BL1, the unselected memory cell MC connected to the unselected word line WL2 and the bit line BL2 and the unselected memory cell MC connected to the unselected word line WL3 and the bit line BL2 becomes the same potential, and any current flows through the current path for these memory cells. That is, the current path through which the current flows when the selection gate line SSG2 is selected is the only current path for the selected memory cell MC (the path reaching the selected word line WL1 from the selected global bit line GBL through the bit line BL1 and the memory cell MC).
Thus, according to the present example, it is not necessary to convert the data as in the third embodiment, and the reading current cars be detected without being influenced by the unselected memory cell MC adjacent to the selected memory cell MC.
5. Modification Example of the Present Embodiment
Hereinafter, a modification example of the present embodiment will be described. In the example of
A relation between the word line WL and the word line group WLcomb can appropriately be selected, and is not limited to the above examples, as long as three adjacent word lines belong to different word line groups WLcomb.
Next, a memory device according to a seventh embodiment will be described. The present embodiment is different from the above third and sixth embodiments in chat a position where a selection element SS is disposed is shifted to suppress the occurrence of erroneous reading.
1. Re: Structure of Selection Element SS
As shown, in a structure of the present embodiment, the position of the selection element SS is shifted in a first direction from the structure of
2. Re: Data Reading Method
Next, the data reading method will be described. Biases of interconnects during the data reading are similar to those of the third embodiment. Moreover, the selection gate line SSG immediately under the selected bit line BL is selected. In other words, the selection gate line SSG connected to a gate shared by transistors TR1 and TR2 connected to the selected bit line BL is selected.
In the present example, data is read from a memory cell MC2 connected to the word line WL2 and a bit line BL2. Therefore, a selector decoder 24 selects a selection gate line SSG2 positioned immediately under the bit line BL2, applies, for example, a voltage Vg_r, and applies 0 V to the other selection gate lines.
In consequence, a cell current flows through the selected word line WL2 from a global bit line GBL via a channel formed by the selection gate line SSG2, a bit line BL1 and the selected memory cell MC.
3. Re: Data Writing Method and Erasing Method
A data writing method is similar to the third embodiment except that data conversion is not required. That is, only one selection gate line SSG described in the above 2 may be selected. The same also applies to an erasing time.
4. Effect of the Present Embodiment
Also according to the present embodiment, interference with adjacent cells can be suppressed, and a data reading accuracy can be enhanced. Hereinafter, the present effect will be described in detail.
In the present example, one selection gate line SSG corresponds to one bit line BL (one column of bit line BL disposed in a second direction). Therefore, a current path is not formed between the global bit line GBL and another bit line BL (another bit line row disposed in the second direction) corresponding to the unselected selection gate line SSG.
In, for example, the example of
Therefore, a relation of con1>con2>con3 is obtained. For example, con1 is a current flowing through the channel formed by the selected selection gate line SSG2. Moreover, con2 is a leak current which is influenced by the potential of the selection gate line SSG2 to flow into the global bit line GBL from a bit line BL3 through the silicon layer 6 between the selection gate line SSG2 and a selection gate line SSG3. Furthermore, con3 is a leak current flowing through the silicon layer 6 between the unselected selection gate lines.
Thus, the current con1 flowing through the selected bit line BL2 becomes largest, and hence the influence of the adjacent cell can be suppressed. Moreover, the leak current con2 can be as low as the leak current con3, depending on an element size.
It is to be noted that when the present embodiment is considered to have a structure where the position of the selection element SS of the first embodiment is shifted along the first direction to connect the drain region 7 to two bit lines BL, the one selection element SS is formed by two transistors TR1 and TR2 connected to two selection gate lines having different gates, as shown in
5. Modification Example of the Present Embodiment
Moreover, in the structure according to the present embodiment, the selection gate line SSG is formed immediately under the bit line BL, and two portions of the drain region 7 come in contact with the silicon layer 6. Therefore, a line width of the bit line BL (or the drain region 7) along the first direction is larger than a line width of the selection gate line SSG along the first direction. In this case, the line width of the bit line BL (or the drain region 7) along the first direction may be larger than a minimum processing dimension F (a half pitch). In this case, a space between the adjacent bit lines BL (i.e., the line width of the word line WL along the first direction) is formed to be smaller than the minimum processing dimension F (the half pitch) by use of a side wall processing technology or the like. Moreover, a width (the period or the pitch) of a set of one word line WL and one bit line BL is preferably 2F. In consequence, a size of one memory cell MC can be 2F2.
Next, a memory device according to an eighth embodiment will be described. In the present embodiment, a selection element SS is shifted to a position different from that of the seventh embodiment, to suppress the occurrence of erroneous reading.
1. Re: Structure of Selection Element SS
As shown, the selection element SS according to the present embodiment is different from the seventh embodiment in shift amount of a silicon layer 6, a source region 5 and a selection gate line SSG along a first direction. More specifically, the silicon layer 6 of the one selection element SS is connected to only one bit line BL. Both sides of the silicon layer 6 are in contact with the selection gate lines SSG via gate insulating films 9. In this case, a region where a channel is formed by the one selection gate line SSG contacts with the drain region 7, and functions as a current path between the corresponding bit line BL and a global bit line GBL. A region where a channel is formed by the other selection gate line SSG does not contact with the drain region 7. Therefore, the region does not function as the current path of a cell current. Moreover, the one selection gate line SSG is positioned immediately under the corresponding bit line BL, and the other selection gate line SSG is positioned immediately under another bit line BL.
Therefore, as shown in
2. Re: Data Reading Method
Next, a data reading method will be described. Biases of interconnects during the reading are similar to those of the third embodiment. That is, in
In the present example, data is read from a memory cell MC2 connected to the word line WL2 and a bit line BL2. Therefore, a selector decoder 24 selects a selection gate line SSG2 positioned immediately under the bit line BL2 to apply, for example, a voltage Vg_r thereto, and applies 0 V to the other selection gate lines.
In consequence, a cell current flows into the selected word line WL2 from the global bit line GBL through a channel formed by the selection gate line SSG2, the bit line BL2 and the selected memory cell MC. The present embodiment is different from the seventh embodiment only in that only one current path is present in the selection element SS.
3. Re: Data Writing Method and Erasing Method
A data writing method is similar to the third embodiment except that data conversion is not required. That is, only one selection gate line SSG nay be selected as described in the above 2. The same applies to an erasing time.
4. Effect of the Present Embodiment
Also according to the present embodiment, interference with the adjacent cell can be suppressed, and a data reading accuracy can be enhanced. Hereinafter, the present effect will be described in detail.
In the present example, one selection gate line SSG is associated with one bit line BL (one column of bit lines BL arranged in a second direction) in the same manner as in the seventh embodiment. Therefore, a current path substantially is not formed between the global bit line GBL and another bit line BL corresponding to the unselected selection gate line SSG (another bit line column disposed in the second direction) (i.e. any effective current path is not formed).
For example, in the example of
In consequence, the only selected bit line BL2 is substantially connected to the global bit line GBL. Therefore, an influence of a leak current flowing through another unselected bit line can remarkably be lowered, and the data reading accuracy can further be enhanced.
5. Modification Example of the Present Example
However, the drain region 7 is diffused to such an extent that when one of two selection gates is selected in the selection element SS corresponding to an unselected bit line BL, a channel formed by this selected selection gate line SSG does not contact with the drain region 7. This is because if this channel contacts with the drain region 7, a current path is unfortunately formed between the unselected bit line and the global bit line GBL.
It is to be noted that in the present example, a space between the adjacent bit lines BL may be formed with a minimum processing dimension F, or a smaller dimension.
Next, a memory device according to a ninth embodiment will be described. In the present embodiment, unlike the above third, sixth, seventh and eighth embodiments, the selection gate line SSG is not shared with two selection elements SS, but the line is separated for each selection element SS, whereby wrong reading is suppressed. 1. Re: Structure of Selection Element SS
As shown, in the present embodiment, one selection gate line SSG shared by two selection elements SS disposed adjacent in a first direction in
Furthermore, the selection gate lines SSG are drawn to a hook-up region outside the memory cell array. The hook-up region is a region for connecting interconnects such as the selection gate lines SSG and word lines WL to a peripheral circuit such as a decoder. Also in the hook-up region, the selection gate lines SSG has the ring-like shape.
2. Re: Data Writing, Erasing and Reading Methods
A data writing method is similar to the third embodiment except that data conversion is not required. That is, as to the selection gate line SSG, only one selection gate line SSG of the selection element SS connected to the selected bit line BL may be selected.
Biases during data erasing and reading are similar to those of the first embodiment. However, as to the selection gate line SSG, only one selection gate line SSG of the selection element SS connected to the selected bit line BL is selected.
3. Re: Fabricating Method of Selection Gate Line
Next, a fabricating method of the selection gate line SSG according to the present embodiment will be described with reference to
First, the structure of
Next, as shown in
Next, as shown in
Afterward, the selection gate lines 48 are etched back to complete the selection gate lines SSG shown in
4. Effect of the Present Embodiment
Also in the present embodiment, interference with adjacent cells can be suppressed, and a data reading accuracy can be enhanced. Hereinafter, the present effect will be described in detail.
In the present example, one selection gate line SSG shared by two selection elements SS in the first embodiment is divided into two selection gate lines SSG. Moreover, one selection gate line SSG obtained by the dividing is used as a gate electrode of one selection element SS (a column of selection elements SS arranged along the second direction), and the other selection elements SS (another column of selection elements SS arranged along the second direction) are not used as the gate electrodes.
Therefore, when a certain selection gate line SSG is selected during the data reading, a current path is formed only between the selected bit line BL corresponding to the selection element SS and the global bit line GBL, and any current path is not formed between the unselected bit line BL and the global bit line GBL. Therefore, the influence of a leak current flowing through the other unselected bit line BL can remarkably be lowered, and the data reading accuracy can further be enhanced.
Next, a memory device according to a tenth embodiment will be described. In the present embodiment, unlike the above third, sixth, seventh, eighth and ninth embodiments, a plurality of reading operations is performed to verify reading data, thereby suppressing erroneous reading.
1. Structure of Sense Amplifier
A memory cell array according to the present embodiment is shown in
Gates of the transistors 71 and 72 are connected in common to form a current mirror circuit CM1. Sources of the transistors 71 and 72 are connected to a power source potential Vdd. A drain of the transistor 71 is connected to the gate thereof and a corresponding global bit line GBL. A drain of the transistor 72 is connected to one of nodes P1 to P3 by the switch element SW1. An operation of the switch element SW1 is controlled by, for example, a controller 25.
Gates of the transistors 79 and 80 are connected in common to form a current mirror circuit CM2-1. Sources of the transistors 79 and 80 are grounded. A drain of the transistor 79 is connected to the gate thereof and the node PI. Gates of the transistors 81 and 82 are connected in common to form a current mirror circuit CM2-2. Sources of the transistors 81 and 82 are grounded. A drain of the transistor 81 is connected to the gate thereof and the node P2. One electrode of the capacitor element C2 is connected to gates of the transistors 81 and 82, and the other electrode is grounded. Gates of the transistors 83 and 84 are connected in common to form a current mirror circuit CM2-3. Sources of the transistors 83 and 84 are grounded. A drain of the transistor 83 is connected to the gate thereof and the node P3. One electrode of the capacitor element C3 is connected to gates of the transistors 83 and 84, and the other electrode is grounded.
Gates of the transistors 77 and 78 are connected in common to form a current mirror circuit CM3-1. Sources of the transistors 77 and 78 are connected to the power source potential Vdd. A drain of the transistor 77 is connected to the gate thereof and a drain of the transistor 80. Gates of the transistors 75 and 76 are connected in common to form a current mirror circuit CM3-2. Sources of the transistors 75 and 76 are connected to the power source potential Vdd. A drain of the transistor 75 is connected to the gate thereof and a drain of the transistor 82. Gates of the transistors 73 and 74 are connected in common to form a current mirror circuit CM3-3. Sources of the transistors 73 and 74 are connected to the power source potential Vdd. A drain of the transistor 73 is connected to the gate thereof and a drain of the transistor 84.
A non-inverting input terminal of the operation amplifier OP1 is connected to drains of the transistors 76 and 74. A inverting input terminal of the operation amplifier OP2 is connected to a drain of the transistor 78, and a non-inverting input terminal thereof is connected to an output terminal of the operation amplifier OP1.
2. Re: Reading Operation
Next, a data reading method according to the present embodiment will be described with reference to
As shown, for example, the controller 25 first connects the selected global bit line GBL_s to the sense amplifier 70 (the drain of the MOS transistor 71), and the GBL decoder 23 applies a predetermined reading bias (e.g., Vr+Vo) to GBLs. At this time, a WL decoder 22 applies a predetermined reading bias (Vr+Vo) to both of the selected word line WL_s and the unselected word line WL_u of a measurement object cell arrangement. Continuously, the WL decoder 22 lowers the voltage of the selected word line (WL_s: only one WLcomb is selected from one cell arrangement) to a predetermined reading selection voltage (e.g., Vo). In this state, data is roughly read by the following three steps.
(First Step)
First, the controller 25 connects the switch element SW1 to the node P3. Moreover, a selector decoder 24 applies the selection gate voltage (e.g., Vg_r) to the selection gate line SSG_n. SSG_s and SSG_u are set to 0 V. In consequence, a current I3 flowing through the global bit line GBL is transferred to the current mirror circuit CM2-3 of a second stage via the current mirror circuit CM1 of a first stage. Then, a gate voltage corresponding to the current I3 is held in the capacitor element C3.
(Second Step)
Next, the controller 25 connects the switch element SW1 to the node P2. Then, the selector decoder 24 applies 0 V to the selection gate line SSG_n to which the selection gate voltage has been applied in the first step, and applies the selection gate voltage to SSG_s. Heedless to say, SSG_u is 0 V. In consequence, the current I2 flowing through the global bit line GBL is transferred to the current mirror circuit CM2-2 of the second stage via the current mirror circuit CM1 of the first stage. Then, a gate voltage corresponding to a current I2 flowing through the global bit line GBL is held in the capacitor element C2.
(Third Step)
Finally, the controller 25 connects the switch element SKI to the node P1. Then, the selector decoder 24 applies the selection gate voltage to two selection gate lines SSG_s and SSG_n. In consequence, a current I1 flowing through the global bit line GBL is transferred to the current mirror circuit CM2-1 of the second stage via the current mirror circuit CM1 of the first stage.
At this time, the current mirror circuits CM2-3 and CM2-2 of the second stage output a current equivalent to the currents I3 and I2 by a charge accumulated in the capacitor elements C3 and C2. Therefore, the currents I3 and I2 are transferred to the current mirror circuits CM3-3 and CM3-2 of a third stage. In consequence, the output of the operation amplifier OP1 constituted as an addition circuit is equivalent to the sum of the currents I3 and I2.
Then, the output of the current mirror circuit CM2-1 of the second stage is transferred to an input of the current mirror circuit CM3-1 of the third stage, and hence the output of the current mirror circuit CM3-1 becomes the current I1. In consequence, the output of the operation amplifier OP2 constituted as a differential amplification circuit is equivalent to a value of the current (I2+I3−I1).
It is to be noted that the above operation is repeatedly performed for different selected word lines if necessary, whereby it is possible to successively read stored values of the cells on the same global bit line. Moreover, it is also possible to read, in parallel, the stored values of the cells on the plural, global bit lines by use of plural reading circuits.
3. Effect of the Present Embodiment
Also according to the present embodiment, interference with adjacent cells can be suppressed, and a data reading accuracy can be enhanced. Hereinafter, the present effect will be described in detail.
As described in the third embodiment, the current flowing through the OFF2-element is remarkably small, and can be ignored. Therefore, the current I1 flowing through the global bit line GBL becomes the sum of a current Ia flowing through the selected memory cell MC via the ON-element and currents Ib and Ic flowing through the unselected memory cell via the OFF1-element. Then, the current Ib and Ic flowing through the OFF1-element become background noise as described above.
In this respect, not only the reading of
In consequence, when (I2+I3−I1)=((Ia+Ib)+(Ia+Ic)−(Ia+Ib+Ic)) is detected, the current other than the current flowing through the selected memory cell can be cancelled, and the only current Ia flowing through the selected memory cell can be detected. That is, the background noise can be cancelled, and the data reading accuracy can be enhanced.
It is to be noted that a method of using an analog circuit for detecting (I2+I3−I1) was utilized in the present embodiment. However, it is possible to use a method of performing digitization (4-level digitalization) in a detection stage of I1, I2 and I3, and finally digitally calculating (I2+I3−I1).
[Modification Example, etc.]
As described above, the memory device according to the above first to tenth embodiments includes first interconnects(GBL in
The embodiments are not limited to the above-mentioned configurations, and can variously be modified. For example, the respective embodiments can appropriately be combined, or can be performed alone. For example, the operations of the above third to tenth embodiments have been described on the assumption that the offset voltages Vo and 1 V are used in the same manner as in the first embodiment, but the offset voltage may not be used. Moreover, the writing method described in the fourth embodiment may be applied to the other embodiments. Furthermore, the pattern of the word lines described in the sixth embodiment may be applied to the seventh to tenth embodiments.
Moreover, in the third embodiment, the data does not necessarily have to be written in order of the memory cells MC0 to MC(n−1). There is not any special restriction on the writing order, as long as the cell values c0 to c(n−1) converted on the basis of the above-mentioned rule are written in the memory cells MC0 to MC(n−1). Therefore, there may be a case where, for example, the cell value c(n−1) is first written in the memory cell MC(n−1), and then the cell value c0 is written in the memory cell MC0. The same applies to the reading, and it is not necessary to read the data In order of the memory cells MC0 to MC(n−1). Even when the data is read in any order, the influence of the adjacent cells is suppressed.
Further in the fabricating methods described in the second and tenth embodiments, the order can be changed as much as possible, or the steps may simultaneously be performed. Furthermore, the above-mentioned materials or specific values of the film thicknesses and the like can appropriately be changed.
Moreover, the bit lines, the word lines and the global bit lines do not necessarily have to orthogonally cross one another, as long and the directions thereof are different from one another. Moreover, a sectional shape and/or a planar shape of a stacked structure of the memory cell is not limited to the square shape, and another polygonal or round shape may be used.
Furthermore, the circuit diagrams and the like described in the above embodiments are merely examples, and there is not any special restriction, as long as the function of each embodiment can be realized. For example, the circuit of
It is to be noted that the above embodiments can include the following configurations.
[1] A memory device comprising:
a plurality of first interconnects(GBL in
a plurality of second interconnects(WL in
a plurality of third interconnects(BL in
a plurality of memory cells(MC in
a plurality of selectors (SS in
wherein one of the selectors(SS in
a plurality of gates(SSG in
[2] The device according to [1], wherein first one of the selectors shares one of the gates with second one of the selectors adjacent to the first one of the selectors.
[3] The device according to [2], wherein in one of the selectors associated with one of the third interconnects coupled with a memory cell which is selected to be read data, a selection voltage is applied to one of the gates provided on the two side surfaces, and a non-selection voltage is applied to another of the gates.
[4] The device according to [3], wherein the second interconnects alternately belong to the sane interconnect group (WLcomb_a or WLcomb_b in
the same voltage is applied to two or more of the second interconnects belonging to the same interconnect group in an operation.
[5] The device according to [4], further comprising:
a controller which controls a writing operation of data into the memory cell,
wherein the controller writes cell values (c0, c1, . . . c(n−1)) obtained according to the following equations in the memory cells (MC0, MC1, . . . MC(n−1)), when receiving, from a host apparatus, data values (d0, d1, . . . d(n−1)) as writing-data into the memory cells (MC0, MC1, . . . MC(n−1)) (
c0=d0, and
c(k+1)=d(k+1)+ck,
where (MC0, MC1, . . . MC(n−1)) indicate the memory cells coupled with the second interconnects belonging to the same interconnect group and arranged in order from the memory cell MC0 positioned at the endmost portion, and
n is a natural number of 2 or larger, and (k+1) is a natural number of 1 to (n−1).
[6] The device according to [5], wherein one of the second interconnects(WL) which is coupled with the memory cell MC0 is located between other two of the second interconnects belonging to the different interconnect group(
[7] The device according to [6], wherein when data is read from a memory cell MC(k+1) in data reading, the selection voltage is applied to one of the gates which forms a channel between a memory cell MCk and one of the first interconnects in the two gates of one of the selectors associated with one of the third interconnects coupled with the memory cell MC(k+1) (
[8] The device according to [5], wherein the controller writes data in the memory cells (MC0, MC1, . . . MC(n−1)) of one of the interconnect group (WLcomb), and then repeats the writing data to the different interconnect group(WLcomb) (steps S41 to S46 in
[9] The device according to [8], wherein the controller repeats the writing data to the interconnect group(WLcomb) in the different first interconnect(GBL) (steps S40 and S47&48 in
[10] The device according to [5], wherein the controller writes the data in the memory cells(MC0, MC1, . . . MC(n−1)) of one of the interconnect groups(WLcomb), and then repeats the writing data to the different first interconnect(GBL)(
[11] The device according to [10], wherein the controller repeats the writing data to the first interconnect(GBL) in the different interconnect group(WLcomb)(
[12] The device according to [2], wherein each of the second interconnects belongs to one of three or more interconnect groups (WLcomb_a to WLcomb_d in
the same voltage is applied to one or more of the second interconnects belonging to the same interconnect group during the operation, and
the three adjacent second interconnects belong to the mutually different interconnect groups.
[13] The device according to [12], wherein a selection voltage is applied to one of the gates(SSG2 in
[14] The device according to [2], wherein one of the selectors is shared by two adjacent ones of the third interconnects(
[15] The device according to [14], wherein the one of the selectors further includes a source region coupled with the associated one of the first interconnects, the semiconductor layer stacked on the source region, and a first drain region and a second drain region coupled with two adjacent ones or the third interconnects, and
the gates are provided immediately under the third interconnects(
[16] The device according to [15], wherein a selection voltage is applied to one of the gates(SSG2 in
[17] The device according to [16], wherein the one of the gates to which the selection voltage is applied allows two of the selectors (SS1,SS2 in
[18] The device according to [2], wherein the one of the selectors further includes a source region coupled with the associated one of the first interconnects, the semiconductor layer provided on the source region, and a drain region provided on the semiconductor layer and coupled with the associated one of the third interconnects, and
the drain region overlaps, in the first direction, with the semiconductor layer in a partial region(
[19] The device according to [18], wherein the semiconductor layer includes a first region in which a first channel is formed by one of the gates formed on the two side surfaces of the semiconductor layer and a current path is formed between the drain region and the source region by the first channel, and
a second region in which a second channel is formed by the another of the gates formed on the two side surfaces and the second channel is isolated from the drain region (
[20] The device according to [19], wherein the selection voltage is applied to the one of the gates(SSG2 in
[21] The device according to [20], wherein the second channel does not function as an effective current path between the one of the third interconnects(BL2 in
[22] The device according to [1], wherein in each of the selectors, gates(SSG in
the gates are separated between the selectors.
[23] The device according to [22], wherein a size of the gate along the first direction is smaller than ½ of a size of the second interconnect along the first direction (
[24] The device according to [23], wherein a size of the gate along the first direction is smaller than a minimum processing dimension in a photolithography technology (
[25] The device according to [2], further comprising:
a controller which controls a reading operation of data from the memory cells,
wherein the controller detects a first current(I1 in
a second current(I2 in
a third current(I3 in
determines reading data on the basis of the first to third currents.
[26] The device according to [25], wherein the controller subtracts the first current from the sum of the second current and the third current, to determine the reading data(
[27] The device according to [1], wherein each of the second interconnects belongs to one of two or more interconnect groups(WLcomb_a,WLcomb_b in
the same voltage is applied to one or more of the second interconnects belonging to the same one of the interconnect groups in an operation,
wherein the controller writes data in the memory cells of one of the interconnect groups (WLcomb_a1 in
[28] The device according to [27], wherein the controller repeats the writing data of the interconnect groups(WLcomb_a,WLcomb_b in
[29] The device according to [1], wherein each of the second interconnects belongs to one of two or more interconnect groups (WLcomb_a, WLcomb_b in
the same voltage is applied to one or more of the second interconnects belonging to the same interconnect group in an operation,
wherein the controller writes data in the memory cells of one of the interconnect groups (WLcomb_a, WLcomb_b), and then repeats the writing data to different one of the first interconnects (GBL).
[30] The device according to [29], wherein the controller repeats the writing data of the first interconnects(GBL) to the different one of the interconnect groups(WLcomb_a,WLcomb_b).
[31] A fabricating method of a memory device comprising:
forming a plurality of global bit lines along a first direction(
forming a semiconductor layer on the global bit lines(
patterning the semiconductor layer along a second direction orthogonal to the first direction(
forming a plurality of gate insulating films on side surfaces of the patterned semiconductor layer(
forming gate electrodes on the side surfaces of the patterned semiconductor layer(
forming a plurality of layers of word lines on the patterned semiconductor layer(
forming a variable resistance material on side surfaces of the word lines; and
forming a plurality of bit lines in contact with the variable resistance material and the upper surface of the semiconductor layer (
[32] A fabricating method of a memory device comprising:
successively forming global bit line film and a semiconductor layer (
patterning the global bit line film and the semiconductor layer along a first direction (
filling in grooves generated by the patterning with an insulating film (
patterning the semiconductor layer and the insulating film along a second direction orthogonal to the first direction (
forming a plurality of gate insulating films on side surfaces of the patterned semiconductor layer (
forming a plurality of layers of word lines on the patterned semiconductor layer (
forming a variable resistance material on side surfaces of the word lines (
forming a plurality of bit lines in contact with the variable resistance material and the upper surface of the semiconductor layer (
[33] The method according to [32], wherein the gate electrode fills in a groove between adjacent patterned semiconductor layers (
[34] The method according to [32], wherein the gate electrode does not fill in a groove between adjacent patterned semiconductor layers (
While certain embodiments have been described, these embodiments have been presented by way of example only, and are not intended to limit the scope of the inventions. Indeed, the novel embodiments described herein may be embodied in a variety of other forms; furthermore, various omissions, substitutions and changes in the form of the embodiments described herein may be made without departing from the spirit of the inventions. The accompanying claims and their equivalents are intended to cover such forms or modifications as would fall within the scope and spirit of the inventions.
This application is a continuation of and claims the benefit of priority under 35 U.S.C. § 120 from prior U.S. patent application Ser. No. 16/172,088, filed Oct. 26, 2018, which is a continuation of and claims the benefit of priority under 35 U.S.C. § 120 from prior U.S. patent application Ser. No. 15/650,526, filed Jul. 14, 2017, which is a continuation application and claims the benefit of priority under 35 U.S.C. § 120 from prior U.S. patent application Ser. No. 15/152,342, filed May 11, 2016, which is a continuation of and claims the benefit of priority under 35 U.S.C. § 120 from prior U.S. patent application Ser. No. 14/472,094, filed Aug. 28, 2014, which is a continuation of and claims the benefit of priority under 35 U.S.C. § 120 from prior U.S. application Ser. No. 13/313,186, filed on Dec. 7, 2011. The entire contents of each of which are incorporated herein by reference.
| Number | Name | Date | Kind |
|---|---|---|---|
| 6649953 | Cha | Nov 2003 | B2 |
| 6759707 | Prall | Jul 2004 | B2 |
| 7952163 | Baek et al. | May 2011 | B2 |
| 8084830 | Kanno et al. | Dec 2011 | B2 |
| 8254160 | Murooka et al. | Aug 2012 | B2 |
| 8295077 | Murooka | Oct 2012 | B2 |
| 8406034 | Murooka | Mar 2013 | B2 |
| 8891277 | Murooka | Nov 2014 | B2 |
| 8963115 | Murooka | Feb 2015 | B2 |
| 8971093 | Murooka | Mar 2015 | B2 |
| 9224459 | Murooka | Dec 2015 | B1 |
| 9368160 | Murooka | Jun 2016 | B2 |
| 9741766 | Murooka | Aug 2017 | B2 |
| 10141374 | Murooka | Nov 2018 | B2 |
| 10600845 | Murooka | Mar 2020 | B2 |
| 20040090815 | Tajiri | May 2004 | A1 |
| 20080265235 | Kamigaichi | Oct 2008 | A1 |
| 20090141547 | Jin et al. | Jun 2009 | A1 |
| 20090273054 | Kim et al. | Nov 2009 | A1 |
| 20100027316 | Yoon et al. | Feb 2010 | A1 |
| 20100259960 | Samachisa | Oct 2010 | A1 |
| 20100259961 | Fasoli | Oct 2010 | A1 |
| 20100259962 | Yan et al. | Oct 2010 | A1 |
| 20110261606 | Sandhu et al. | Oct 2011 | A1 |
| 20110261607 | Tang et al. | Oct 2011 | A1 |
| 20120147644 | Scheuerlein | Jun 2012 | A1 |
| 20120147645 | Scheuerlein | Jun 2012 | A1 |
| 20120147646 | Scheuerlein | Jun 2012 | A1 |
| 20120147647 | Scheuerlein | Jun 2012 | A1 |
| 20120147648 | Scheuerlein | Jun 2012 | A1 |
| 20120147649 | Samachisa et al. | Jun 2012 | A1 |
| 20120147650 | Samachisa et al. | Jun 2012 | A1 |
| 20120147651 | Scheuerlein et al. | Jun 2012 | A1 |
| 20120147652 | Scheuerlein | Jun 2012 | A1 |
| 20120147689 | Scheuerlein | Jun 2012 | A1 |
| 20130043455 | Bateman | Feb 2013 | A1 |
| 20150103581 | Murooka | Apr 2015 | A1 |
| Number | Date | Country |
|---|---|---|
| 2007-536680 | Oct 2007 | JP |
| 2008-181978 | Aug 2008 | JP |
| 4299882 | Jul 2009 | JP |
| 2010-287872 | Dec 2010 | JP |
| 2011-129639 | Jun 2011 | JP |
| 4719233 | Jul 2011 | JP |
| Entry |
|---|
| Office Action dated Jun. 16, 2015 in Japanese Patent Application No. 2012-197238 (with English translation). |
| Number | Date | Country | |
|---|---|---|---|
| 20200176514 A1 | Jun 2020 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 16172088 | Oct 2018 | US |
| Child | 16787225 | US | |
| Parent | 15650526 | Jul 2017 | US |
| Child | 16172088 | US | |
| Parent | 15152342 | May 2016 | US |
| Child | 15650526 | US | |
| Parent | 14472094 | Aug 2014 | US |
| Child | 15152342 | US | |
| Parent | 13313186 | Dec 2011 | US |
| Child | 14472094 | US |
