Embodiment
The present invention sets forth various in the technology that designs and make the multiple polysilicon layer of use in the various logical elements (for example, NAND door, AND door, NOR door, OR door, XOR gate, sram cell, latch etc.) that is used for semiconductor device.According to particular of the present invention, can reduce gate cell size and memory array cell size by using a plurality of polysilicon layers to make various transistor gates.By this kind mode, can realize reducing of integrated circuit chip area by the standard design rule that reduces corresponding to minimum poly-1 to poly-1 spacing.Therefore, for example, the technology of the present invention can realize the layout and/or the design of the overlapping polysilicon chip of not short circuit each other, and this is because can use a multilayer polysilicon manufacturing technology to form different polysilicon chips.According to a specific embodiment, these superimposed sheets can be made of at least two different polysilicon layers, and those polysilicon layers are separated on vertical by at least one insulating barrier (for example a, oxide layer).By this kind mode, can prevent the electrical short of the polysilicon chip that each is overlapping.And the technology that use multilayer polysilicon of the present invention forms the transistor gates of logic element can provide the extra degree of freedom aspect the accurate adjustment transistor parameter, and these parameters are for example allowed gate voltage etc. for oxide thickness, threshold voltage, maximum.
As mentioned above, traditional knowledge is informed the saving grace of the cost minimization that people reduce or will be relevant with making integrated circuit.Usually, a kind of technology that reduces or make this kind cost minimization is that the quantity that will be used to form the polysilicon layer of integrated circuit (IC) logic element minimizes.Recently, (for example be used to make some type memory, flash memory) manufacturing technology has been used a kind of double level polysilicon layer process, wherein different polysilicon layers is deposited on the silicon wafer to form the control gate and the floating gate of flash memory cell at different time.A design of integrated circuit is intended comprising in flash memory and both application-specific of conventional logic elements therein, and for (for example) forms flash memory cell, the manufacturing of integrated circuit (IC) chip may relate to pair of lamina polysilicon layer technology.Yet, during this kind integrated circuit is made, still keeping the conventional practice of using a single polysilicon layer to make the integrated circuit (IC) logic element.People wish that a reason using the design of single polysilicon layer and make integrated circuit (IC) logic element (even comprising at integrated circuit under the situation of flash memory) is because the design of single polysilicon layer circuit is simpler, therefore be easier to usually make, and foozle still less occurs.In addition, as mentioned above, commercially available the most frequently used circuit simulation software is designed to only compatible standard layout and the manufacturing technology of using the single polysilicon layer at present.The incompatible usually a plurality of polysilicon laminar designs of these circuit simulation softwares.
Yet, opposite with the traditional knowledge and the practice, the present invention informs people especially in the application that will make memory element (for example, flash memory, DRAM) and logic element on same integrated circuit (IC) chip, uses multiple polysilicon layer to make the saving grace of logic element.In those were used, people can utilize multiple polysilicon layer technology (for example, being used to make memory element) by designing the logic element that also uses multiple polysilicon layer.
Fig. 2 A shows the part 200 of the logic element that one specific embodiment is made according to the present invention.More specifically, circuit part 200 shown in Fig. 2 A be how according to the present invention one specific embodiment make the example of cascade transistor circuit (for example circuit shown in Figure 1A).According to specific embodiment, circuit part 200 can be used for making various logical elements, for example, and NAND door, AND door, NOR door, OR door, XOR gate, sram cell, latch etc.
Fig. 2 B-2I in graphic shows that one is used to make the specific embodiment of circuit part 200.Fig. 2 B-2I set forth a kind of according to the present invention one specific embodiment make the technology of a logic element.In the example shown in Fig. 2 B-2I, suppose that a silicon wafer is used for manufacturing one integrated circuit (IC) chip.In the preparation of IC chip manufacturing process, the available p-section bar of some part material of silicon wafer mixes, and forms a P-trap 208 thus.
Shown in Fig. 2 B, one first oxide layer 204a ' is formed on the surface of silicon wafer 210.After forming this first oxide layer 204a ', can on this first oxide layer, deposit one first polysilicon (poly-1) layer 202a '.Then, can remove or the several regions of etching poly-1 layer 202a ', form the first polysilicon layer part 202a by this, shown in Fig. 2 C.After forming the first polysilicon layer part 202a, can remove the several regions of the first oxide layer 204a ', to form one first oxide layer 204a by this, shown in Fig. 2 D.According to a specific embodiment, the formation of polysilicon layer and oxide layer and/or removal can use the known traditional IC chip fabrication techniques of person of ordinary skill in the field to realize.
Shown in Fig. 2 E and 2F, can on the several portions that comprises part 202a and 204a of this silicon wafer, form and/or deposit one second oxide layer 204b ' and one second polysilicon (poly-2) layer 202b ' then.Shown in Fig. 2 G, can remove the several regions of poly-2 layer, to form the second polysilicon layer part 202b.Afterwards, can remove institute's favored area of this second oxide layer, to form the first oxide layer part 204b, shown in Fig. 2 H.Shown in Fig. 2 I, can adopt (for example) traditional ion embedding technology to form doped region 205a and 205b then.According to specific embodiment, those doped regions 205a and 205b can use such as n-section bar material such as arsenic and mix.Perhaps, can mix with a n-section bar material in zone 208, and forming a N-trap, and regional 205a and 205b can be with the doping of p-section bar material.
When being configured to two nmos pass transistors that are connected in series, circuit part 200 will be equivalent to the circuit part 275 of Fig. 2 J, and it can be schematically shown by the Figure 100 of signal shown in Figure 1A.Shown in Fig. 2 J, two gate poles of cascade transistor circuit 275 (for example gate pole B 202b and gate pole A 202a) are used to control from source electrode 205a the electric current to the 205b that drains and flow.
Circuit part shown in Figure 1B 150 (it represents a cascade transistor circuit that adopts conventional art to make) is compared with circuit part 275 shown in Fig. 2 J, can find some differences.For example, compare with Figure 1B, the spacing distance of gate pole A and gate pole B is much smaller among Fig. 2 J.More specifically, shown in Fig. 2 J, the spacing distance of gate pole 202b and gate pole 202a approximates the thickness of the second oxide layer part 204b.In addition, shown in Fig. 2 J, poly-2 layer segment 204b not only be positioned in a continuous manner on the poly-1 layer segment 202a and the first oxide layer part 204a but also with the two adjacency.Poly-2 layer segment 202b also with the region overlapping of poly-1 layer segment 202a.According to different embodiment, overlapping quantity can be in 0% overlapping (for example, with the gate region termination) variation to the scope of about 100% overlapping (for example, overlapping with gate region fully).
In addition, shown in Fig. 2 J, compare, remove a whole doped region from silicon substrate with Figure 1B.For example, shown in Fig. 2 J, circuit part 275 comprises two N
+ Doped region 205a, 205b.On the contrary, circuit part shown in Figure 1B 150 comprises 3 N
+Doped region, i.e. 105a, 105b and 105c.Figure 1B and 2J are compared, can find that promptly the doped region 105b that exists between A of gate pole shown in Figure 1B and the gate pole B is removing in the structure shown in Fig. 2 J.This can reduce the area of logic element on this disk, and the size of circuit small pieces and relevant manufacturing cost are reduced.
Should be appreciated that alternate embodiment of the present invention can comprise the feature that is different from shown in Fig. 2 J circuit part 275.For example, Fig. 2 K shows the alternate embodiment of a circuit part that one specific embodiment is made according to the present invention 280.Shown in embodiment among Fig. 2 K, circuit part 280 comprises two the overlapping polysilicon layer 282a, the 282b that are formed on the substrate 210.In this specific embodiment, substrate 210 is made of N-section bar material, and (p+) doped region 285a, 285b are formed by P-section bar material.Between the circuit part 275 and 280 one significantly difference is: the P-well area 208 of circuit part 275 (Fig. 2 J) is the transistorized local substrate as circuit part 275, and circuit part 280 does not comprise that one is different from the independent well area of substrate 210.But, in circuit part 208 (Fig. 2 H), by the transistorized local substrate of substrate 210 as circuit part 280.
Fig. 3 A-D illustrates the different embodiment of the cascade transistor circuit that can make according to the technology of the present invention.The stereogram of circuit part 200 among Fig. 3 A displayed map 2A.Fig. 3 B shows the stereogram of the circuit part 350 of an alternate embodiment, and circuit part 350 can be used for making up the cascade transistor circuit 100 shown in Figure 1A.
With reference to Fig. 3 A, it should be noted that the design of circuit part 300 is different from traditional circuit design aspect some.For example, as mentioned above, use different polysilicon layers to form transistor gates 202a, 202b.In addition, the position of gate structure and structure are different from traditional circuit design (for example design of traditional circuit shown in Figure 1B).For example, as shown in Figure 3A, gate pole 202b is overlapping with the part of a mode and gate pole 202a, and this mode is sandwiched between the active area of gate pole 202b and substrate 208 part of gate pole 202a.In addition, be different from conventional gate shown in Figure 1B structure (it typically is smooth structure), the structure of gate pole 202b and out-of-flatness, but be one to be included in the stepped configuration of the part of extending on level and the vertical both direction.Therefore, the top of gate pole 202b and lower surface are all not smooth substantially.
The design that also it should be noted that circuit part 300 is different from traditional non-volatile memory architecture aspect some.For example, in traditional non-volatile memory cell structure, can use two-layer polysilicon to form a traditional division gate cell that comprises a control gate and a floating gate.Usually, floating gate is designed to the electric insulation zone, it is a single non-volatile memory cells stored charge as a memory node.Be stored charge correctly, every other structure electric insulation in floating gate and the memory cell is very important.On the contrary, gate structure 202a, 202b shown in Fig. 3 A all are not configured to zone through electric insulation in the mode of the floating gate structure that is similar to nonvolatile memory.But gate structure 202a, 202b all can be electrically coupled to other parts of integrated circuit through one or more joining zone, so that required gate voltage is put on transistor circuit 300.For example, this is shown in Fig. 3 C illustrated embodiment.
Fig. 3 C shows the vertical view of the circuit layout 360 that a specific embodiment of a use the technology of the present invention designs.More specifically, Fig. 3 C illustrated embodiment represents that one is used for the particular of cascade transistor circuit (for example circuit shown in Fig. 3 A) layout.Illustrate traditional schematic diagram of a series connection transistor circuit among Figure 1A.Yet Fig. 3 D shows that one can be used for schematically showing the example of the schematic diagram 370 of circuit 360 shown in Fig. 3 C.
Shown in Fig. 3 C, cascade transistor circuit 360 comprises two the gate pole 382a, the 382b that have lap on active area 365.According to a specific embodiment, circuit 360 can adopt multiple polysilicon layer technology of the present invention manufacturing, and wherein one first gate pole (for example 382a) is formed by the poly-1 layer, and one second gate pole (for example 382b) is formed by the poly-2 layer.A part of overlaid of two gate poles is shown in zone 367.According to a specific embodiment, the width W 1 of gate pole overlapping region 367 is equal to or greater than the width W 2 of active area 365 at least.Shown in Fig. 3 C, each gate pole 382a, 382b include corresponding joining zone 362a, a 362b, think that each gate pole provides electric contact.According to a specific embodiment, poly-1 layer and poly-2 layer can be made of for the known suitable electric conducting material of person of ordinary skill in the field polysilicon or other.
In addition, shown in Fig. 3 C, active area 365 can comprise an one source pole joining zone 364 and a drain contact zone 366.In Fig. 3 C example illustrated, gate pole joining zone 362a, 362b place on the opposite side of active area 365, with guarantee to satisfy with each contact between the relevant design constraint of minimum spacing.Yet, should be appreciated that have many kinds of diverse ways to can be used for circuit 360 shown in the design of graphics 3C.For example, in an alternate implementation (not shown), the contact on each gate pole 382a, 382b can be positioned at the same side of active area 365.Yet should be appreciated that these different constructing plans all have a common ground: some part of gate pole 382a, 382b will overlap each other or termination on active area 365.Should be appreciated that circuit part 300 shows a kind of overlapping gate structure type of multilayer polysilicon that can be used for making the integrated circuit (IC) chip logic element among Fig. 3 A.With a problem of the structurally associated of circuit part 300 be: the variable gate length of gate pole 202b may be with respect to gate pole 202a mismatch.Be used for solving an embodiment who makes the gate pole mismatch problems owing to gate length is variable and be shown in graphic Fig. 3 B.
Fig. 3 B shows the stereogram of an alternate embodiment logic element circuit part 350 that one specific embodiment structure forms according to the present invention.Shown in Fig. 3 B, circuit part 350 comprises one first hydride layer 304a, a poly-1 gate pole 302a, one second oxide layer 304b and a poly-2 door-plate 302b.Shown in Fig. 3 B, the both sides of poly-2 gate pole 302b and poly-1 gate pole 302a are all adjacent.In addition, poly-2 gate pole 302b is overlapping with poly-1 gate pole 302a above substrate active area 308.Therefore, shown in Fig. 3 B, at least a portion of gate pole 302a is sandwiched between the active area 308 of gate pole 302b and silicon substrate.In addition, shown in Fig. 3 B, gate pole 302a and the gate pole 302b spacing that approximates the second oxide layer 304b thickness of being separated by.
Gate structure configuration advantage is to alleviate or to eliminate the described variable gate length issue above with reference to Fig. 3 A shown in Fig. 3 B.For example, according to a specific embodiments, because the length of (for example) gate pole 302b is to be determined by the mask edge that can correctly aim at, so the overall width W of gate pole 302b can keep constant.Therefore, the gate configuration of Fig. 3 B can be used for relaxing the problem of mismatch between poly-1 gate pole 302a and the poly-2 gate pole 302b.
It should be noted that circuit part 300 and 350 only is intended to explain the structure on the active area of each circuit (for example, 208,308), and may not reflect all features of each circuit.Therefore, should be appreciated that circuit part 300 and 350 can comprise other features that do not show among Fig. 3 A and the 3B.For example, this kind feature relates to the contact that is used to contact gate structure 202a, 202b, 302a, 302b.Another feature relates to the structure of gate structure 202a, 202b, 3028,302b.For example, in a constructing plan, each gate structure all can be configured to a polysilicon lines that can extend along either direction in the X-Z plane.Another feature relates to interpolation, and other can be used for making up the transistor of required logic element.
Should be appreciated that the technology that use multilayer polysilicon of the present invention forms logic element can provide the extra degree of freedom aspect the various transistor parameters of accurate adjustment, these parameters are for example allowed gate voltage etc. for oxide thickness, threshold voltage, maximum.For example, according to different embodiment, be the various transistor parameters of accurate adjustment, poly-1 and poly-2 gate can be made for different thickness respectively.According to a specific implementation, two same sizes (for example, width and length can be benefited because of having 2 different threshold voltages, because its gate separately can be made of 2 different oxide layers by) logic transistor.In addition, should be appreciated that in traditional MOS transistor, drain electrode and all horizontal proliferation below gate region of source junction can reduce effective gate length thus and aggravate short-channel effect.Yet, use series transistor circuit configurations of the present invention, get final product (for example) and eliminate a knot (and corresponding horizontal proliferation) in to the transistor that is connected in series at each, improve the short-channel effect of this (a bit) cascade transistor circuit thus.
Employed another common circuit is shown among Fig. 6 A in the conventional logic elements design.Fig. 6 A shows that one comprises the circuit part 600 (being referred to herein as parallel transistor circuit 600) of 2 parallel connected transistors.A kind of conventional design layout that is used for making parallel transistor circuit 600 is shown in Fig. 6 C.Shown in Fig. 6 C, traditional parallel transistor circuit layout 670 is included in and forms 2 poly-1 gate pole 652a, 652b on the active area 681 of silicon substrate.Gate pole 652a, 652b are to use the single polysilicon layer to form.According to traditional design rule, those gate poles minimum spacing 679 that need be separated from each other.In the embodiment shown in Fig. 6 C, source area 672a, the 672b of parallel transistor circuit via an electric connection line 677 electrical couplings together.
Fig. 6 B illustrates a profile that uses the parallel transistor circuit part 650 that traditional IC manufacturing technology makes.Circuit part 650 shown in Fig. 6 B is to use the single polysilicon layer, adopts the technology that is similar to cascade transistor circuit part 150 among above-mentioned manufacturing Figure 1B to make.Shown in Fig. 6 B, circuit part 650 comprises the first oxide layer part 604a, 604b (its two all formed by the first identical oxide layer), poly-1 gate pole part 602a, 602b (it is all formed by a single polysilicon layer) and 3 different doped region 605a, 605b, 605c.
Fig. 7 A shows the stereogram of the parallel transistor circuit part 700 that one specific embodiment is made according to the present invention.Shown in Fig. 7 A, circuit part 700 comprises one first oxide layer part 704a, a poly-1 gate pole 702a, one second oxide layer part 704b, a poly-1 gate pole 702b and 2 doped region 705a, 705b, and these 2 doped region 705a, 705b can be used as the source electrode and the drain region of parallel transistor circuit.According to a specific implementation, the technology type that is used for parallel transistor circuit part 700 shown in the shop drawings 7A is similar to the described technology above with reference to graphic Fig. 2 B-2I.Therefore, for example, poly-1 gate pole 702a can be formed by one first polysilicon layer, and poly-2 gate pole 702b can be formed by second polysilicon layer that is different from first polysilicon layer.In addition, oxide layer part 704a can be formed by one first oxide layer, and oxide layer part 704b then can be formed by second oxide layer that is different from first oxide layer.
Fig. 7 C shows that one can be used for making the vertical view of the layout 760 of parallel transistor circuit shown in Fig. 7 A.Fig. 7 D shows a schematic diagram 770, and it is used to schematically show the parallel transistor circuit design 760 shown in Fig. 7 C.Shown in Fig. 7 C, parallel transistor circuit design 760 comprises a poly-1 gate pole 782a and a poly-2 gate pole 782b.Each gate pole 782a, 782b all to small part source electrode and the drain electrode between active area 765 on.Each gate pole includes corresponding contact areas 762a, 762b.The part of two gate poles is overlapping at regional 767 places as shown in the figure.According to a specific embodiment, the width W 2 of gate pole overlapping region 767 is less than the width W 1 of active area 760.
A difference shown in parallel transistor circuit design shown in Fig. 7 C and Fig. 6 C between traditional parallel circuits design is that circuit shown in Fig. 6 C comprises two via an electric connection line 677 electrical couplings source area 672a, 672b together.On the contrary, shown in Fig. 7 C, 760 in parallel transistor circuit comprises a single source area 764 and a single drain region 766.
Fig. 7 B shows the stereogram of the alternate embodiment of a parallel transistor circuit part 750 of making according to the technology of the present invention.Shown in Fig. 7 B, circuit part 750 comprises one first oxide layer part 754a, a poly-1 gate pole 752a, one second oxide layer part 754b, a poly-2 gate pole 752b and 2 doped region 755a, 755b, and these 2 doped region 755a, 755b can be used as the source electrode and the drain region of parallel transistor circuit.According to a specific embodiments, the technology type that is used for parallel transistor circuit part 750 shown in the shop drawings 7B is similar to the described technology above with reference to graphic Fig. 2 B-2I.Therefore, for example, poly-1 gate pole 752a can be formed by one first polysilicon layer, and poly-2 gate pole 752b then can be formed by second polysilicon layer that is different from first polysilicon layer.In addition, oxide layer part 754a can be formed by one first oxide layer, and oxide layer part 754b can be formed by second oxide layer that is different from first oxide layer.
It should be noted that one of them structural distinctions between the parallel transistor design is shown in the design of serial transistor shown in Fig. 3 C and Fig. 7 C: shown in Fig. 3 C, each gate pole 382a, 382b all can cut off the electric current that flow to drain electrode 366 from source electrode 364.Yet shown in Fig. 7 C, gate pole 782a, 782b but all can not control the electric current that flow to drain electrode 766 from source electrode 764 fully.But each gate pole only control flows is crossed the part of the electric current of active area.Yet,, for example, preferablely apply suitable control voltage to two gate pole 782a, 782b and stop electric current to flow to drain electrode from source electrode according to embodiment shown in Fig. 7 C.
Should be appreciated that each circuit shown in Fig. 3 A-3D and the 7A-7D can be used for making the different logic element of various formation integrated circuit (IC) chip parts.Those logic elements comprise NAND door, AND door, NOR door, OR door, XOR gate, latch etc.In addition, those logic elements can comprise static storage cell, for example SRAM.Fig. 4 shows that one is used to make the vertical view of the conventional design layout of SRAM memory cell 400.Usually, the SRAM memory cell is to adopt the standard design that only comprises a single polysilicon layer to make.This design makes manufacturing process more uncomplicated relatively and more cheap.Therefore, as shown in Figure 4, traditional sram cell design comprises three poly-1 parts 402,404a, 404b, and wherein each part forms by same polysilicon layer.
Conventional design constraints requires each poly-1 part (for example, 402,404a, 404b) minimum spacing that separates each other (for example, spacing A) to prevent short circuit with (for example).In addition, as shown in Figure 4, traditional sram cell design comprises a P
+ Doped region 406 and N
+Doped region 408.The part of sram cell 400 is formed on the P-trap 420.In this figure, each metal interconnecting wires between not shown regional 406,408 and 404.
Relevant with traditional sram cell manufacturing technology various design constraints require various structural designs with sram cell be at least one minimum prescribed amount of separation of other structures in the sram cell (or with SRAM in other structures overlapping).For this reason, the size of traditional sram cell can not reduce to less than minimum designated size.For example, if minimum feature size is 100nm, then the sram cell size has 1m at least usually
2Area.Yet, use manufacturing technology of the present invention, can reduce the memory array cell size by using multiple polysilicon layer to make each transistor.In this way, can realize reducing of memory array cell dimensioned area corresponding to the IC design rule of minimum poly-1 to poly-1 spacing by reducing.
Fig. 5 shows that one can utilize the example of the SRAM memory cell design layout that the technology of the present invention makes.As shown in Figure 5, sram cell 500 comprises at least one poly-1 layer 502 and multiple poly-2 layer 504a, 504b, and wherein poly-2 layer 504a, 504b are formed by a polysilicon layer that is different from poly-1 layer 502.Each polysilicon layer 502,504a, 504b all comprise a corresponding gate pole district 530 and a corresponding interconnection district 532.According to a specific embodiment, interconnection district can be equivalent to be formed at some part of the polysilicon layer in passive (or) district of sram cell 500.In the embodiment shown in fig. 5, the part of sram cell 500 is to be formed on the P-trap 520.In addition, in this figure, each metal interconnecting wires between not shown regional 506,508 and 504.
According to a specific embodiment, can adopt one with make the various transistors that comprised in the sram cell 500 shown in Figure 5 with reference to the described transistor manufacturing technology of Fig. 2 B-2I similar methods in graphic.In one of many possible embodiment, can make a multiple polysilicon layer sram cell, wherein the transmission gate gated transistors comprises the gate pole that is formed by the poly-1 layer, and on draw and/or pull-down transistor comprises the gate pole that is formed by the poly-2 layer.
As shown in Figure 5, the size of sram cell 500 can (for example) some part by making poly-1 layer 502 and some overlap (shown in 515) of poly-2 layer 504a, 504b reduced.This designing technique also helps to reduce poly-1 layer 502 and N
+Spacing (shown in B ') between the zone 508.In addition, shown in embodiment among Fig. 5, overlapping interconnect area 532 places (for example, on the passive region of sram cell 500) that appear at polysilicon layer of 515 place's polysilicon regions.
Although show in Fig. 5, sram cell design 500 also can comprise at least two different oxide layers, insulated from each other and and the surrounding structure insulation to help each polysilicon layer.For example, first oxide layer can be positioned under the poly-1 part 502, and second oxide layer can be positioned under poly-2 layer 504a, the 504b, makes poly-2 layer and poly-1 layer electric insulation thus.
As in circuit shown in Fig. 3 A-D and the 7A-D, use multiple polysilicon layer providing the extra degree of freedom aspect the accurate adjustment transistor parameter in sram cell design shown in Figure 5, these parameters are for example allowed gate voltage or the like for oxidated layer thickness, threshold voltage, maximum.
Although this paper has elaborated some preferred embodiments of the present invention with reference to accompanying drawing, but should be appreciated that, the present invention is not limited only to those definite embodiment, and the person of ordinary skill in the field can carry out various changes and modification to it, and this does not deviate from the present invention's spirit category that is defined in the claims of enclosing.For example, according to some specific embodiment, the transistor gates material that is used to form logic element of the present invention can be made of the combination (for example, titanium silicide) of electric conducting material (for example, titanium), semi-conducting material (for example, polysilicon) or two kinds of materials.In addition, the insulating barrier described in each embodiment of the application's case (for example, oxide layer) can be made of the insulation or the dielectric material of silicon dioxide and/or other types.