Layout pattern for SRAM and manufacturing methods thereof

1. Field of the Invention

The invention relates to a static random access memory (SRAM), and more particularly, to a SRAM having the advantage of decreasing the current differences in different regions.

2. Description of the Prior Art

An embedded static random access memory (SRAM) comprises a logic circuit and a static random access memory connected to the logic circuit. SRAM is a kind of volatile memory cell, which means it preserves data only while power is continuously applied. SRAM is built of cross-coupled inverters that store data during the time that power remains applied, unlike dynamic random access memory (DRAM) that needs to be periodically refreshed. Because of its high access speed, SRAM is also used in computer systems as a cache memory.

However, as pitch of the exposure process decreases, it has been difficult for the current SRAM architecture to produce desirable patterns. Hence, how to enhance the current SRAM architecture for improving exposure quality has become an important task in this field.

The present invention provides a layout pattern of a static random access memory (SRAM), comprising a first inverter and a second inverter constituting a latch circuit, wherein the first inverter includes a first pull-up transistor (PL 1 ) and a first pull-down transistor (PD 1 ), and the second inverter includes a second pull-up transistor (PL 2 ) and a second pull-down transistor (PD 2 ), a first inner access transistor (PG 2 A) and a second inner access transistor (PG 2 B) electrically connected to the latch circuit, wherein the first inner access transistor is electrically connected to a first word line and a second bit line, and the second inner access transistor is electrically connected to a second word line and a fourth bit line, and a first outer access transistor (PG 1 A) and a second outer access transistor (PG 1 B) electrically connected to the latch circuit, wherein the first outer access transistor is electrically connected to the first word line and a first bit line, and the second outer access transistor is electrically connected to the second word line and a third bit line, in addition, the first outer access transistor (PG 1 A) has a first gate length, the first inner access transistor (PG 2 A) has a second gate length, and the first gate length is different from the second gate length.

The present invention further provides a method for forming a layout pattern of a static random access memory (SRAM), comprising: firstly, a first inverter and a second inverter are formed, to constitute a latch circuit, wherein the first inverter includes a first pull-up transistor (PL 1 ) and a first pull-down transistor (PD 1 ), and the second inverter includes a second pull-up transistor (PL 2 ) and a second pull-down transistor (PD 2 ), next, a first inner access transistor (PG 2 A) and a second inner access transistor (PG 2 B) which are electrically connected to the latch circuit are formed, wherein the first inner access transistor is electrically connected to a first word line and a second bit line, and the second inner access transistor is electrically connected to a second word line and a fourth bit line, and an ion implantation process is performed on the first inner access transistor or to the second inner access transistor.

The applicant found that due to the reasons such as the layout difference in different regions, the currents passing through different elements are different. In order to solve this issue, the present invention provides a layout pattern of a SRAM and the manufacturing method for forming the layout pattern thereof. The present invention provides two methods: by performing an OPC process and by performing an ion implantation process to solve the issue mentioned above, the current differences in different regions can thereby be decreased and the quality of SRAM can be improved.

These and other objectives of the present invention will no doubt become obvious to those of ordinary skill in the art after reading the following detailed description of the preferred embodiment that is illustrated in the various figures and drawings.

FIG. 1 illustrates a circuit diagram of a SRAM cell according to a first preferred embodiment of the present invention.

FIG. 2 illustrates a layout diagram of a SRAM according to a first preferred embodiment of the present invention.

FIG. 3 illustrates a layout diagram of a SRAM according to a second preferred embodiment of the present invention.

FIG. 4 illustrates a circuit diagram of a SRAM cell according to a third preferred embodiment of the present invention.

FIG. 5 illustrates a layout diagram of a SRAM according to a third preferred embodiment of the present invention.

FIG. 6 illustrates a layout diagram of a SRAM according to a fourth preferred embodiment of the present invention.

To provide a better understanding of the present invention to users skilled in the technology of the present invention, preferred embodiments are detailed as follows. The preferred embodiments of the present invention are illustrated in the accompanying drawings with numbered elements to clarify the contents and the effects to be achieved.

Please note that the figures are only for illustration and the figures may not be to scale. The scale may be further modified according to different design considerations. When referring to the words “up” or “down” that describe the relationship between components in the text, it is well known in the art and should be clearly understood that these words refer to relative positions that can be inverted to obtain a similar structure, and these structures should therefore not be precluded from the scope of the claims in the present invention.

Referring to FIGS. 1-2 , FIG. 1 illustrates a circuit diagram of a SRAM cell according to a first preferred embodiment of the present invention, and FIG. 2 illustrates a layout diagram of a SRAM according to the first preferred embodiment of the present invention.

As shown in FIGS. 1-2 , the SRAM device of the present invention preferably includes at least one SRAM cell, each SRAM cell including an eight-transistor SRAM (8T-SRAM) cell 10 .

In this embodiment, each 8T-SRAM cell 10 is composed of a first pull-up transistor PL 1 , a second pull-up transistor PL 2 , a first pull-down transistor PD 1 , a second pull-down transistor PD 2 , a first outer access transistor PG 1 A, a second outer access transistor PG 1 B, a first inner access transistor PG 2 A and a second inner access transistor PG 2 B. These eight transistors constitute a set of flip-flops. The first and the second pull-up transistors PL 1 and PL 2 , and the first and the second pull-down transistors PD 1 and PD 2 constitute a latch circuit 22 that stores data in the storage nodes 24 and 26 . Since the first and the second pull-up transistors PL 1 and PL 2 act as power load devices, they can be replaced by resistors. Under this circumstance, the static random access memory becomes a four-transistor SRAM (4T-SRAM). In this embodiment, the first and the second pull-up transistors PL 1 and PL 2 preferably share a source/drain region and electrically connect to a voltage source Vcc, and the first and the second pull-down transistors PD 1 and PD 2 share a source/drain region and electrically connect to a voltage source Vss.

Preferably, the first pull-up transistor PL 1 and the first pull-down transistor PD 1 constitute an inverter. Similarly, the second pull-up transistor PL 2 and the second pull-down transistor PD 2 constitute another inverter. In addition, the storage node 24 is connected to the respective gates of the second pull-down transistor PD 2 and the second pull-up transistor PL 2 . The storage node 24 is also connected to the drains of the first pull-down transistor PD 1 , the first pull-up transistor PL 1 , the first inner access transistor PG 2 A, and the second outer access transistor PG 1 B. Similarly, the storage node 26 is connected to the respective gates of the first pull-down transistor PD 1 and first the pull-up transistor PL 1 . The storage node 26 is also connected to the drains of the second pull-down transistor PD 2 , the second pull-up transistor PL 2 , the first outer access transistor PG 1 A and the second inner access transistor PG 2 B. The gates of the first outer access transistor PG 1 A and the first inner access transistor PG 2 A are respectively coupled to a first word line (WL 1 ); the gates of the second outer access transistor PG 1 B and the second inner access transistor PG 2 B are respectively coupled to a second word line (WL 2 ); the source of the first outer access transistor PG 1 A is coupled to a first bit line (BL 1 ), the source of the first inner access transistor PG 2 A is coupled to a second bit line (BL 2 ), the source of the second outer access transistor PG 1 B is coupled to a third bit line (BL 3 ), and the source of the second inner access transistor PG 2 B is coupled to a fourth bit line (BL 4 ).

In this embodiment, an 8 transistors dual port SRAM (8TDP-SRAM) cell 10 is disposed on a substrate 52 , such as a silicon substrate or silicon-on-insulator (SOI) substrate. The substrate may be a planar substrate, or a plurality of fin structures 54 may be formed on the substrate 52 . Next, pluralities of gate structures 56 are formed on the substrate 52 . In this embodiment, take an 8T-SRAM with fin structures 54 as an example, but the present invention is not limited thereto. In another case, the planar SRAM (the SRAM without comprising fin structures) may also be comprised within the scope of the present invention. Besides, in this embodiment, a plurality of contact structures 57 , 58 are disposed on the substrate 52 , and a plurality of contact plugs 59 which are electrically connected to the contact structures 57 , 58 . The other structure features and the operation principle of 8TDP-SRAM are well known to those skilled in the art, and will not be redundantly described here.

Please refer to FIG. 2 again, in the layout pattern of each 8TDP-SRAM cell 10 , the region that is relatively closer to the latch circuit (the composition of the first pull-up transistor PL 1 , the first pull-down transistor PD 1 , the second pull-up transistor PL 2 and the second pull-down transistor PD 2 in FIG. 1 ) is defined as an inner region 2 . On the other hand, the region that is relatively far apart from the latch circuit is defined as an outer region 1 . Therefore, the “inner” and the “outer” in the names of the first outer access transistor PG 1 A, the second outer access transistor PG 1 B, the first inner access transistor PG 2 A and second inner access transistor PG 2 B also show these transistors disposition within the outer region 1 or disposition within the inner region 2 . For example, both the first outer access transistor PG 1 A and the second outer access transistor PG 1 B are disposed within the outer region 1 ; the first inner access transistor PG 2 A and the second inner access transistor PG 2 B are disposed within the inner region 2 .

The applicant found that during the operation of SRAM, due to the reasons such as the layout difference in different regions, or the energy consumption, etc. . . . one issue may occur: the current passing through the elements within the inner region is different from the current passing through the elements within the outer region. For example, in 28 nanometers (nm) 8TDP-SRAM, according to the experiment of the applicant, the current ID 1 that passes through the first outer access transistor PG 1 A (the current that passes from the source of the PG 1 A to the drain of the PG 1 A) or passes through the second outer access transistor PG 1 B is about 120-130 micro amperes (μA), however, the current ID 2 that passes through the first inner access transistor PG 2 A (the current that passes from the source of the PG 2 A to the drain of the PG 2 A) or passes through the second inner access transistor PG 2 B is lower than ID 1 , it is about 95-100 μA. In other words, in 28 nm 8TDP-SRAM, compared with the current ID 1 passing through the outer region 1 , the current ID 2 passing through the inner region 2 is decreased about 20-30%.

The issue mentioned above may influence the operation of SRAM, or further decrease the yield for manufacturing the SRAM. One purpose of the present invention is to solve the issue mentioned above, in other words, to decrease the current differences in different regions.

Please refer to FIG. 2 again, which illustrates a layout diagram of a SRAM according to a first preferred embodiment of the present invention. Take the first outer access transistor PG 1 A and the first inner access transistor PG 2 A as an example, it is noteworthy that the gate of the first outer access transistor PG 1 A directly contacts the gate of the first inner access transistor PG 2 A. More precisely, in this embodiment, the first outer access transistor PG 1 A and the first inner access transistor PG 2 A comprise a same gate structure (labeled as 56 A in FIG. 2 ). In other words, a gate of the first outer access transistor PG 1 A and a gate of the first inner access transistor PG 2 A are arranged along a same symmetry axis, similarly, a gate of the second outer access transistor PG 1 B and a gate of the second inner access transistor PG 2 B are arranged along a same symmetry axis too. However, the current ID 2 passing through the first inner access transistor PG 2 A is smaller than the current ID 1 passing through the first outer access transistor PG 1 A about 20%. Please also refer to the equation 1 as below:

I DS =μ n C ox W/L [( V GS −V th ) V DS −V 2 DS /2]  Equation 1

wherein:

I DS is the current passing through the transistor; W is the gate width (the channel width) of the transistor; L is the gate length (the channel length) of the transistor; μ n is the carrier mobility; C ox is the unit capacitance of the gate oxide layer; V GS is the voltage from the gate to the source; V DS is the voltage from the drain to the source; and V th is the threshold voltage.

According to equation 1, it shows that the current passing through the transistor is proportional to the gate width W (the channel width) of the transistor, but it is inversely proportional to the gate length L (the channel length) of the transistor. In other words, if other conditions are fixed, the smaller the gate length L is, the larger the current passing through the transistor is. Therefore, as the issue that the current difference in different regions occurs, by adjusting the gate length L in different regions, the current can be changed again, and thereby eliminate the current differences.

More precisely, as shown in FIG. 2 , the gate length of the first outer access transistor PG 1 A is labeled as L 1 , and gate length of the first inner access transistor PG 2 A is labeled as L 2 , by performing an optical proximity correction (OPC) process, to adjust the gate length L 1 of the first outer access transistor PG 1 A, or to adjust the gate length L 2 of the first inner access transistor PG 2 A. In this embodiment, if the ratio between the gate length L 1 and the gate length L 2 is adjusted, the ration of the current ID 1 passing through the first outer access transistor PG 1 A and the current ID 2 passing through the first inner access transistor PG 2 A can be changed. For example, in one embodiment, the gate length L 1 is fixed at 24 nm, and the gate length L 2 is changed by the OPC process, to decrease the gate length L 2 to about 20 nm, in this way, since the current passing through the transistor is inversely proportional to the gate length L of the transistor, after the OPC process, the current ID 1 is not changed, but the current ID 2 is increased, so the current differences between the current ID 1 and the current ID 2 can be decreased, to solve the issue mentioned above.

It can be understood that the OPC process of the present invention is not only limited to adjust the gate length L 2 , it can also be used for adjusting the gate length L 1 . In this embodiment, since the current ID 1 is larger than the current ID 2 , so the gate length L 2 should be adjusted to be smaller than the gate length L 1 . Preferably, the ration between the gate length L 2 and the gate length L 1 is between 0.58-0.99. In addition, in 28 nm SRAM, the gate length L 2 is preferably between 14-24 nm, but is not limited thereto. Those parameters mentioned above can be adjusted according to actual requirements.

Furthermore, in another case, the gate lengths of the other two transistors of the 8TDP-SRAM cell 10 , the second outer access transistor PG 1 B and the second inner access transistor PG 2 B (labeled as the gate length L 3 and the gate length L 4 in FIG. 2 ), can also be simultaneously adjusted by the OPC process mentioned above or by another OPC process in different steps. After the OPC process, the gate length L 3 is adjusted to enlarge the gate length L 4 , therefore the current differences between the current passing through the second outer access transistor PG 1 B and the current passing through the second inner access transistor PG 2 B can be decreased, and it should also be within the scope of the present invention.

In the first preferred embodiment above, the ratio between different gate lengths is changed by further performing an OPC process, thereby decreasing the current differences in different regions and improving the quality of SRAM.

The following description will detail the different embodiments of the SRAM of the present invention. To simplify the description, the following description will detail the dissimilarities among the different embodiments and the identical features will not be redundantly described. In order to compare the differences between the embodiments easily, the identical components in each of the following embodiments are marked with identical symbols.

Please refer to FIG. 3 , which illustrates a layout diagram of a SRAM according to a second preferred embodiment of the present invention. Similar to the first preferred embodiment, this embodiment also comprises at least one eight-transistor dual port SRAM (8TDP-SRAM) cell 10 , and has the same circuit diagram as the first preferred embodiment has (please refer to FIG. 1 ). As shown in FIG. 1 and FIG. 3 , the 8TDP-SRAM cell 10 comprises a latch circuit 22 , and the latch circuit 22 includes a first pull-up transistor PL 1 , a second pull-up transistor PL 2 , a first pull-down transistor PD 1 and a second pull-down transistor PD 2 . In addition, the 8TDP-SRAM cell 10 also comprises a first outer access transistor PG 1 A, a second outer access transistor PG 1 B, a first inner access transistor PG 2 A and a second inner access transistor PG 2 B. Similarly, an outer region 1 and an inner region 2 are defined on the substrate 52 .

The purpose of this embodiment is same as the purpose of the first preferred embodiment: to decrease the current differences between the current passing through the transistor within the outer region and the current passing through the transistor within the inner region. In this embodiment, for example, a current measurement process can be performed on the first outer access transistor PG 1 A and on the first inner access transistor PG 2 A, so as to obtain the current ID 1 passing through the first outer access transistor PG 1 A and the current ID 2 passing through the first inner access transistor PG 2 A. According to the applicant's experiment, due to the layout difference in different regions or other reasons, the current ID 1 should larger than the current ID 2 . Usually, the ratio of the current ID 1 to the current ID 2 is larger than 1.20, but is not limited thereto. Therefore, one purpose of this embodiment is to decrease the current differences in different regions.

The difference between this embodiment and the first preferred embodiment is in this embodiment, an ion implantation process is performed on the substrate 52 within the outer region 1 or within the inner region 2 , to dope or implant the suitable ions into the substrate 52 within the outer region 1 or within the inner region 2 , and to form the doping region 60 (within the outer region 1 ) or to form the doping region 62 (within the inner region 2 ). In this way, the threshold voltage of each transistor can be changed, so as to decrease the current ID 1 within the outer region 1 or to increase the current ID 2 within the inner region 2 . More precisely, take a n-type transistor (NMOS) as an example, by doping the phosphorus (P) ions or the arsenic (As) ions into the substrate 52 within the inner region 2 , to help increase the current passing through the NMOS (the current from the source to the drain of the NMOS), in this way, the current ID 2 can be increased. Or in another case, by doping the boron (B) ions or the boron fluoride (BF) ions into the substrate 52 within the outer region 1 , this helps to decrease the current ID 1 . According to the applicant's experiment, the current ID 2 will be decreased about 9% when doping per 10 13 (1/cm 2 ) concentration BF ions into the substrate within the outer region 1 . Therefore, by adjusting the concentration of the doping ions, the current differences between the current ID 1 and the current ID 2 can be effectively decreased.

It is noteworthy that the first preferred embodiment and the second preferred embodiment can be combined together. In other words, in one case, an OPC process and an ion implantation process can be performed on a same 8TDP-SRAM, so as to adjust the gate length and also to change the threshold voltage, and to change the current passing through the transistor. It should also be within the scope of the present invention.

The current difference issue mentioned above not only occurs in 8TDP-SRAM, it may also occur in other SRAM structures, such as 8-transistor register file (8TRF-SRAM) or the SRAM including more transistors (such as 10T-dual port SRAM).

In another embodiment of the present invention, please refer to FIGS. 4-5 , FIG. 4 illustrates a circuit diagram of a SRAM cell according to a third preferred embodiment of the present invention, and FIG. 5 illustrates a layout diagram of a SRAM according to a third preferred embodiment of the present invention.

In this embodiment, an 8TRD-SRAM cell 12 includes a first pull-up transistor PL 1 ′, a second pull-up transistor PL 2 ′, a first pull-down transistor PD 1 ′ and a second pull-down transistor PD 2 ′, a first access transistor PG 1 and a second access transistor PG 2 . In addition, the 8TRD-SRAM cell 12 further comprises two read transistors: RPG and RPD which are series connected to each other. The first pull-up transistor PL 1 ′, the second pull-up transistor PL 2 ′, the first pull-down transistor PD 1 and the second pull-down transistor PD 2 constitute a latch circuit 22 ′. Besides, in this embodiment, the first pull-up transistor PL 1 ′ and the second pull-up transistor PL 2 ′ share a source/drain region and electrically connect to a voltage source Vcc, and the first and the second pull-down transistors PD 1 ′ and PD 2 ′ share a source/drain region and electrically connect to a voltage source Vss.

Furthermore, the gates of the first access transistor PG 1 and the second access transistor PG 2 are respectively coupled to a word line (WL); the source of the first access transistor PG 1 is coupled to a first bit line (BL 1 ), the source of the second access transistor PG 2 is coupled to a second bit line (BL 2 ); the source of the read transistor RPG is coupled to a read bit line (RBL), the gate of the read transistor RPD is coupled to the latch circuit 22 ′, and the drain of the read transistor RPD is coupled to the voltage source Vss.

In this embodiment, as shown in FIG. 5 , the 8TRF-SRAM cell 12 is disposed on a substrate 52 , such as a silicon substrate or silicon-on-insulator (SOI) substrate. The substrate may be a planar substrate, or a plurality of fin structures 54 may be formed on the substrate 52 . Next, pluralities of gate structures 56 are formed on the substrate 52 . In this embodiment, take an 8T-SRAM with fin structures 54 as an example, but the present invention is not limited thereto. In another case, the planar SRAM (the SRAM without comprising fin structures) may also be comprised within the scope of the present invention. Besides, in this embodiment, a plurality of contact structures 57 , 58 are disposed on the substrate 52 , and a plurality of contact plugs 59 which is electrically connected to the contact structures 57 , 58 . The other structure features and the operation principle of 8TDP-SRAM are well known to those skilled in the art, and will not be redundantly described here.

Similar to the preferred embodiment mentioned above, due to the reasons such as the layout difference in different regions, or the energy consumption, etc. . . . one issue may occur: the current passing through the first access transistor PG 1 (the current from the source to the drain, labeled as ID 1 ′) is different from the current passing through the second access transistor PG 2 (the current from the source to the drain, labeled as ID 2 ′). For example, in 28 nanometers (nm) 8TRF-SRAM, according to the experiment of the applicant, the current ID 1 ′ that passes through the first access transistor PG 1 is smaller than the current ID 2 ′ that passes through the second access transistor PG 2 about 7-13%.

Hence, performing the OPC process mentioned above can adjust the current differences between the current ID 1 ′ and the current ID 2 ′. More precisely, both the gate length L 1 ′ of the first access transistor PG 1 and the gate length L 2 ′ of the second access transistor PG 2 can be changed through an OPC process, for example, decreasing the gate length L 1 ′ or increasing the gate length L 2 ′ to decrease the current differences between the current ID 1 ′ and the current ID 2 ′. In other words, the OPC process mentioned in the first preferred embodiment is applied to the 8TRF-SRAM of this embodiment. Except for the features mentioned above, the other components, material properties, and manufacturing method of this embodiment are similar to the first preferred embodiment detailed above and will not be redundantly described.

Please refer to FIG. 6 , which illustrates a layout diagram of a SRAM according to a fourth preferred embodiment of the present invention. Similar to the third preferred embodiment, this embodiment also comprises at least one 8TRF-SRAM cell 12 , and having the same circuit diagram as the third preferred embodiment has (please refer to FIG. 4 ).

The difference between this embodiment and the third embodiment is in this embodiment, an ion implantation process is performed on the substrate surrounding the first access transistor PG 1 or the second access transistor PG 2 , to form the doping region 60 , 62 in the substrate. In this way, the threshold voltage of each transistor can be changed, so as to increase the current ID 1 ′ or to decrease the current ID 2 ′. Except for the features mentioned above, the other components, material properties, and manufacturing method of this embodiment are similar to the second preferred embodiment detailed above and will not be redundantly described.

It is noteworthy that the third preferred embodiment and the fourth preferred embodiment can be combined together. In other words, in one case, an OPC process and an ion implantation process can be performed on a same 8TRF-SRAM, so as to adjust the gate length and also to change the threshold voltage, and to change the current passing through the transistor. It should also be within the scope of the present invention.

In summary, due to the reasons such as the layout difference in different regions, the currents passing through different elements are different. In order to solve this issue, the present invention provides a layout pattern of a SRAM and the manufacturing method for forming the layout pattern thereof. The present invention provides two methods: by performing an OPC process and by performing an ion implantation process to solve the issue mentioned above, the current differences in different regions can thereby be decreased and the quality of SRAM can be improved.

Those skilled in the art will readily observe that numerous modifications and alterations of the device and method may be made while retaining the teachings of the invention. Accordingly, the above disclosure should be construed as limited only by the metes and bounds of the appended claims.