KR20100120787A - Nand flash memory array with extended charge storage node and fabrication method of the same - Google Patents

Abstract

PURPOSE: A NAND flash memory array and a manufacturing method thereof are provided to improve a short channel effect and a retention property by widely forming the width of a word line and/or a bit line to be greater than an interval. CONSTITUTION: A multiple dielectric layer(50) is formed on a substrate(10). The multiple dielectric layer includes a tunneling oxide film(52), a nitride film(54), and a blocking oxide film(56). A gate material(60) is formed on the multiple dielectric layer. A first hard mask material is formed on the gate material. A first hard mask(72) having a minimum feature size is formed by etching the first hard mask material.

Description

NAND flash memory array having an extended charge storage node and a method of manufacturing the same {NAND FLASH MEMORY ARRAY WITH EXTENDED CHARGE STORAGE NODE AND FABRICATION METHOD OF THE SAME}

The present invention relates to a NAND flash memory array and a method of manufacturing the same, and more particularly, a NAND flash memory array and a method of manufacturing the same, which extends the charge storage node area of each cell by reducing the interval between word lines and / or bit lines. It is about.

Recently, as the size of a device is drastically reduced, short channel effects are gradually increasing. This not only degrades the performance of the device itself, but also increases the cell's threshold voltage distribution due to an increase in sub-threshold swing (SS).

In addition, the leakage current (leakage current) is increased due to the deterioration of the on / off characteristics, which serves to lower the potential of the channel during self-boosting, resulting in program disturb.

In particular, since NAND flash memory uses multi-dielectric layers such as ONO as a gate dielectric, unlike an MOSFET, an effective oxide thickness (EOT) increases, thereby reducing the short channel effect. It must be more serious.

In addition, when the gate length gradually decreases due to the reduction in device size, the gate length variation of each cell may increase, which may change the performance of the device.

In addition, if the area of the charge storage node decreases as the gate length decreases, a problem occurs in terms of reliability. First, it may cause a change in the write / erase operation speed. In addition, in terms of retention, a smaller number of charges are stored, which results in a large threshold voltage change even if only a few electrons escape. This may be a bigger problem in the implementation of multi-level cell (MLC) with a small difference in threshold voltages per level.

Therefore, the present invention has been proposed to solve the problems caused by the miniaturization of the device as described above, and the NAND flash memory array and the method of manufacturing the same by reducing the interval between the word line and the bit line to expand the charge storage node area of each cell The purpose is to provide.

In order to achieve the above object, the NAND flash memory array having an extended charge storage node according to the present invention is characterized in that the memory cells having a plurality of charge storage nodes defined between a plurality of word lines perpendicularly cross each bit line. In the NAND flash memory array having a configuration electrically connected in series along each bit line, the line width of each word line defining each charge storage node is larger than the spacing between neighboring word lines.

In addition, the line width of each word line is twice or more than the spacing between adjacent word lines, the spacing between the word line is characterized in that 2 ~ 50 nm.

In addition, the line width of each bit line defining each of the charge storage nodes is larger than an interval between neighboring bit lines, more specifically, it is characterized in that it is more than twice as large.

Meanwhile, a method of manufacturing a NAND flash memory array having an extended charge storage node includes a first step of defining a plurality of bit lines in a predetermined semiconductor substrate; A second step of sequentially stacking a multi-dielectric layer, a gate material, and a first hard mask material on the substrate on which the plurality of bit lines are formed; Etching the first hardmask material to form a first hardmask having a width and a spacing of a minimum feature size; Depositing and etching a material having an etch rate different from that of the gate material on the substrate to form a first sidewall on a side of the first hard mask; And using a first hard mask and the first sidewall as an etch mask and etching the gate material to form a plurality of word lines.

The plurality of bit lines of the first step may include depositing and etching a second hard mask material on the substrate to form a second hard mask having a width and a spacing of a minimum feature size; Depositing and etching a material having an etch rate different from that of the semiconductor material of the substrate to form a second sidewall on a side of the second hard mask; And using the second hard mask and the second sidewall as an etch mask and etching the substrate to form an isolation trench.

According to the present invention, the array structure according to the present invention has the problem of miniaturization of memory cells by making the widths of word lines and / or bit lines wider than their respective intervals, even though the memory density remains the same. ) Is a fundamental solution to the problem.

In addition, the method of manufacturing an array according to the present invention can reduce the spacing of word lines and / or bit lines to less than the minimum feature size that is allowed in the photo etch through the sidewall process while using the existing process as it is, and the reduced By increasing the word line and / or bit line width by the length, there is an effect that a NAND flash memory array having a charge storage node can be economically easily expanded.

Hereinafter, preferred embodiments of the present invention will be described with reference to the accompanying drawings.

1 to 9 are process perspective views showing a first embodiment of a method for manufacturing a NAND flash memory array having an extended charge storage node according to the present invention, and FIG. 10 is a first embodiment of the manufacturing method according to the present invention. 11 is a perspective view showing an array structure manufactured according to an example, and FIG. 11 is a cross-sectional view comparing the cross section (b) of FIG. 10 with the conventional one (a), and FIGS. 12 and 13 are related to the manufacturing method according to the present invention. 14 is a perspective view showing a second embodiment of the present invention, FIG. 14 is a perspective view showing an array structure manufactured according to a second embodiment of the manufacturing method according to the present invention, and FIGS. 15 to 21 are structural and electrical characteristics by simulation. The figure showing.

First, a NAND flash memory array having an extended charge storage node according to the present invention includes a plurality of word lines WL1 perpendicularly intersecting each bit line BL1, BL2, and BL3, as shown in FIGS. 10 and 11. A NAND flash memory array having a configuration in which memory cells having a plurality of charge storage nodes defined between WL2 and WL3 are electrically connected in series along each bit line, wherein each word defining each charge storage node is defined. The line width Lg 'of the line is larger than the gap Wgap' between neighboring word lines.

That is, as shown in FIG. 11, one memory cell occupies one pitch and remains the same as before, but the line width of the word line is larger than the conventional one (Lg '> Lg), and the distance between neighboring word lines is conventional. It is characterized in that the charge storage node of each cell is extended in the bit line direction by making it smaller (Wgap '<Wgap).

By doing so, the contact margin of each word line can be increased and the contact resistance can be reduced.

In addition, when the line width of the word line is expanded as described above, the gate physical length of each cell is increased, so that the short channel effect can be significantly suppressed, and the sub-threshold slope can be significantly increased.

In order to confirm the electrical characteristics as described above, simulations with the structure as shown in FIG.

Through this, when the half pitch (minimum feature size) is extended to the line width (Lg ') of the word line to 50 nm at 30 nm, it can be seen that the sub-threshold slope is significantly increased than before expansion.

In the above embodiment, the line width Lg 'of each word line is preferably more than twice the spacing between adjacent word lines Wgap', which is a source / drain formation method of each cell and an interference problem with neighbor cells. It depends on the gap between the word lines (Wgap ') determined in consideration of.

That is, as shown in FIG. 9, when the source / drain 86 of each cell is formed by impurity ion implantation, the gap (Wgap ') between the word lines may be 50 nm or more. In the case where the source / drain of each cell is electrically formed by the fringing field 66 due to the pass voltage of Vpass, the gap between the word lines is considered in consideration of the interference problem with the neighboring cells. It is preferable that silver is 2-50 nm.

The lower limit of the gap between the word lines Wgap 'may be 2 nm depending on the development of the dielectric material.

As shown in FIG. 13, when the source / drain is formed by the inversion by the fringing field 66, a simulation for determining the peak inversion charge according to the spacing between word lines is performed. It carried out by and obtained the same result as FIG.

As can be seen in FIG. 18, when the same pass voltage Vpass is applied to neighboring word lines at the same pitch, the peak inversion charge increases when the word line width is increased and the gap width between word lines is decreased.

When the source / drain is formed in inversion, the short channel effect is generally significantly improved by increasing the effective channel length. When the word line width is increased, the physical gate length is further increased. Likewise, the sub-threshold slope can be increased further.

Increasing the word line width eventually increases the charge storage node area of each cell, which results in more charge being stored (as shown in the simulation of FIG. 20, which is increased by about 53%). Even if the resultant stored charge is partially lost, as shown in the simulation result of FIG. 21, the change of the threshold voltage does not occur significantly, thereby improving the retention characteristic. This is particularly effective in suppressing read errors in multi-level cell (MLC) implementations with low threshold voltage margins per level.

In this embodiment, the line width of each of the bit lines BL1, BL2, and BL3 defining each of the charge storage nodes is also larger than a distance between neighboring bit lines, and thus the area of each charge storage node is in the word line direction. It is preferable to extend the width of each word line for the same reason of extending the area of each charge storage node in the bit line direction. Of course, this is also true under the condition that the minimum feature size (or half pitch) is the same as before.

In this case, the line width of each bit line is more preferably two times larger than the distance between neighboring bit lines, which is dependent on the interval between bit lines determined in consideration of the electrical insulation between neighboring bit lines when defining each bit line. do.

Next, a method of manufacturing a NAND flash memory array having an extended charge storage node according to the present invention is as follows.

First, a plurality of bit lines are defined in a predetermined semiconductor substrate (first step), which may be defined to have minimum feature sizes having the same bit line width and spacing as usual, but with FIGS. 1 to 4. Similarly, it is desirable to define the width of the bit line larger than the interval of the bit line.

Subsequently, as shown in FIG. 5, the multiple dielectric layer 50, the gate material 60, and the first hard mask material 70 are sequentially stacked on the substrate 10 on which the plurality of bit lines are formed (second step).

In this case, the multiple dielectric layer 50 is formed of a tunneling oxide film 52 / nitride film 54 / blocking oxide film 56, the gate material 60 is preferably a doped silicon-based material (polysilicon, etc.), The first hard mask material 70 is preferably an oxide film.

Next, the first hard mask material 70 is etched to form a first hard mask 72 having a width and an interval of a minimum feature size (third step).

Subsequently, a material having an etch rate different from that of the gate material 60 is deposited and etched on the substrate to form a first sidewall 82 on the side of the first hard mask 72 as shown in FIG. step).

In this case, a material having an etch rate different from that of the gate material 60, that is, the material of the first sidewall 82 may be an oxide film or a nitride film.

Subsequently, the gate material 60 is etched using the first hard mask 72 and the first sidewall 82 as an etch mask, and a plurality of word lines 64 are formed as shown in FIG. 13 ( 5th step).

Through the above steps, as shown in FIG. 13 or 14, a NAND flash memory array having an extended charge storage node forming an inversion source / drain into the fringing field 66 may be manufactured.

However, using the first hard mask 72 as an etch mask between the third and fourth steps, as shown in FIG. 6, etching the gate material 60 to a predetermined depth to form a trench. Further, after performing the fourth step sidewall process, as shown in FIG. 7, the first sidewall 80 is formed by the sidewalls of the first hard mask 72 and the gate material 62 exposed by the etching. Is formed on the trench side of the.

Subsequently, the first hard mask 72 and the first sidewall 80 are used as an etching mask, and as shown in FIG. 8, the remaining of the gate material 62 and the multi-dielectric layer 50 are further etched to form trenches. After forming the 84, the first hard mask 72 and the first sidewall 80 are removed, and an ion implantation process is performed on the substrate. As shown in FIG. 9, the source / drain 86 is formed. The method may further include forming a film, and depositing an insulating film on the entire surface of the substrate and performing a planarization process, thereby manufacturing an array as shown in FIG. 10.

As shown in FIG. 6, the gate material 60 may be etched to a predetermined depth to form a trench, and then the sidewall process may be performed to adjust the deposition height of the gate material 60 and to give a tilt during ion implantation. When the gate (word line) is too high and narrow, there is an advantage that can solve the problem that the ion implantation is not good.

On the other hand, when defining a plurality of bit lines in the first step, it is preferable that the width of each bit line is also defined to be larger than the interval of the bit line to make the charge storage node area of each memory cell larger. same.

First, as shown in FIG. 1, the second hard mask material 20 is deposited on the semiconductor substrate 10, the photoresist film PR, and the like are applied, and then the width and interval of the minimum feature size possible in the photolithography process. The second hard mask 30 is formed by etching the second hard mask material 20 using the pattern 30 as an etch mask.

Subsequently, a second sidewall is formed directly on the side surface of the second hard mask by depositing and etching a material (eg, an oxide film or a nitride film) having a different etching rate from that of the semiconductor material of the substrate.

Thereafter, the second hard mask and the second sidewall are used as an etching mask, and the substrate is etched to form an isolation trench 12.

When the isolation trench 12 is formed, the second hardmask material is etched only a predetermined depth, that is, as shown in FIG. 2, the trench 24 is formed in the second hardmask material, and as shown in FIG. 3, The second sidewall 40 is formed on the side of the trench 24, and the second hardmask material and the second sidewall 40 remaining using the second hardmask 22 and the second sidewall 40 as an etch mask and the second sidewall 40 are formed as an etch mask. It is more preferable to form the isolation trench by sequentially etching the substrate 10.

The isolation trench 12 may be left void without filling anything, but preferably filled with a high dielectric material.

As described above, the preferred embodiment of the present invention has been described, but those skilled in the art can be replaced with equivalents or limited to the configuration based on the embodiment, various descriptions, even if the description thereof is omitted, the appended claims It belongs to the scope.

1 to 9 are process perspective views for illustrating a first embodiment of a method for manufacturing a NAND flash memory array having an extended charge storage node according to the present invention.

10 is a perspective view showing an array structure manufactured according to the first embodiment of the manufacturing method according to the present invention;

FIG. 11 is a cross-sectional view comparing the cross section (b) of FIG. 10 with the conventional one (a),

12 and 13 are process perspective views for showing a second embodiment of a manufacturing method according to the present invention,

14 is a perspective view showing an array structure manufactured according to the second embodiment of the manufacturing method according to the present invention;

15 to 21 are diagrams showing the structure or electrical characteristics by simulation.

Claims (10)

A NAND flash memory array having a configuration in which memory cells having a plurality of charge storage nodes defined between a plurality of word lines perpendicularly intersecting each bit line are electrically connected in series along the respective bit lines. And a line width of each word line defining each charge storage node is greater than a distance between neighboring word lines. The method of claim 1, And a line width of each word line is more than twice as large as a distance between neighboring word lines. In the second, NAND flash memory array having an extended charge storage node, characterized in that the spacing between the word line is 2 ~ 50 nm. 4. The method according to any one of claims 1 to 3, And a line width of each bit line defining the respective charge storage nodes is greater than a distance between neighboring bit lines. The method of claim 4, wherein And a line width of each bit line is at least two times larger than a distance between neighboring bit lines. A first step of defining a plurality of bit lines in a predetermined semiconductor substrate; A second step of sequentially stacking a multi-dielectric layer, a gate material, and a first hard mask material on the substrate on which the plurality of bit lines are formed; Etching the first hardmask material to form a first hardmask having a width and a spacing of a minimum feature size; Depositing and etching a material having an etch rate different from that of the gate material on the substrate to form a first sidewall on a side of the first hard mask; And a fifth step of forming the plurality of word lines by using the first hard mask and the first sidewall as an etch mask and etching the gate material to form a plurality of word lines. Method of making an array. The method of claim 6, Between the third step and the fourth step further comprises the step of forming a trench by using the first hard mask as an etching mask and etching the gate material to a predetermined depth, The first sidewall of the fourth step is formed on the side of the first hard mask and the trench side of the gate material exposed by the etching, After the fifth step, after removing the first hard mask and the first sidewall, the multi-dielectric layer is further etched using the plurality of word lines as an etch mask, and an ion implantation process is performed on the substrate to perform source / application. A method of manufacturing a NAND flash memory array having an extended charge storage node, further comprising forming a drain. 8. The method according to claim 6 or 7, The plurality of bit lines of the first step, Depositing and etching a second hardmask material on the substrate to form a second hardmask having a width and a spacing of a minimum feature size; Depositing and etching a material having an etch rate different from that of the semiconductor material of the substrate to form a second sidewall on a side of the second hard mask; And forming an isolation trench by etching the substrate using the second hard mask and the second sidewall as an etch mask, and forming an isolation trench. The method of claim 8, The second hard mask material may be etched only a predetermined depth to have a width and an interval of a minimum feature size to form the second hard mask, and then the second hard mask and the second sidewall are etched masks. And forming a separate trench by sequentially etching the second hard mask material and the substrate to form a plurality of bit lines, wherein the plurality of bit lines are defined. . The method of claim 9, The multiple dielectric layer is an oxide film / nitride film / oxide film, The first hard mask material is an oxide film, The second hard mask material is a nitride film, And said first sidewall material and said second sidewall material are oxide or nitride films.

Images