(→Specifications) |
|||
(41 intermediate revisions by 28 users not shown) | |||
Line 1: | Line 1: | ||
{{lithography processes}} | {{lithography processes}} | ||
− | The '''3 | + | The '''3 nanometer ([[3 nm]] or 30 Å) lithography process''' is a [[technology node]] semiconductor manufacturing |
+ | :process following the [[5 nm lithography process|5 nm process]] node. | ||
+ | Commercial [[integrated circuit]] manufacturing using 3 nm process is set to begin some time around [[2023]]. | ||
+ | |||
+ | The term "[[3 nm]]" is simply a commercial name for a generation of a certain size and its technology, and | ||
+ | :'''does not''' represent any geometry of the transistor. | ||
== Industry == | == Industry == | ||
− | {{ | + | ==== [[Intel]] ==== |
+ | P1278 Intel's 5-nanometer (renamed as Intel 20A) process node is expected to ramp around the 2H [[2024]]/[[2025]] timeframe. | ||
+ | |||
+ | === [[TSMC]] === | ||
+ | N3 technology will offer up to 70% logic density gain, up to 15% speed improvement at the same power and up to | ||
+ | :30% power reduction at the same speed as compared with N5 technology (According to TSMCs website). | ||
+ | If this holds true we could see 300+ MT/mm2. | ||
+ | |||
+ | === [[Samsung]] === | ||
+ | On May 24, [[2017]] Samsung announced they will be switching to a transistor they call ''Multi-Bridge-Channel FET'' | ||
+ | :(''MBCFET''), an extension of a [[Gate-all-around]] (GAA) [[FET]]. | ||
+ | This is planned for somewhere after the [[5 nm]] node but the exact timeline or specification is currently unknown. | ||
+ | |||
+ | == Specifications == | ||
+ | |||
+ | {{finfet nodes comp | ||
+ | <!-- Intel --> | ||
+ | | process 1 fab = [[Intel]] | ||
+ | | process 1 name = P1276 (CPU/Chipset) | ||
+ | | process 1 date = 2024 | ||
+ | | process 1 lith = EUV | ||
+ | | process 1 immersion = Yes | ||
+ | | process 1 exposure = SE EUV+DUV SAPQ | ||
+ | | process 1 wafer type = Bulk | ||
+ | | process 1 wafer size = 300 mm | ||
+ | | process 1 transistor = FinFET | ||
+ | | process 1 volt = | ||
+ | | process 1 delta from = [[4 nm]] Δ | ||
+ | | process 1 fin pitch = 30 nm | ||
+ | | process 1 fin pitch Δ = 1.0x | ||
+ | | process 1 fin width = | ||
+ | | process 1 fin width Δ = narrower | ||
+ | | process 1 fin height = | ||
+ | | process 1 fin height Δ = taller | ||
+ | | process 1 gate len = | ||
+ | | process 1 gate len Δ = 1.0x | ||
+ | | process 1 cpp = 50 nm | ||
+ | | process 1 cpp Δ = 1.0x | ||
+ | | process 1 mmp = 30 nm | ||
+ | | process 1 mmp Δ = 1.0x | ||
+ | | process 1 sram hp = 0.0300 μm<sup>2</sup> | ||
+ | | process 1 sram hp Δ = 1.0x | ||
+ | | process 1 sram hd = 0.0240 μm<sup>2</sup> | ||
+ | | process 1 sram hd Δ = 1.0x | ||
+ | | process 1 sram lv = | ||
+ | | process 1 sram lv Δ = | ||
+ | | process 1 dram = | ||
+ | | process 1 dram Δ = | ||
+ | <!-- TSMC --> | ||
+ | | process 2 fab = [[TSMC]] | ||
+ | | process 2 name = N3(B) | ||
+ | | process 2 date = 2023 | ||
+ | | process 2 lith = EUV | ||
+ | | process 2 immersion = Yes | ||
+ | | process 2 exposure = SALELE | ||
+ | | process 2 wafer type = Bulk | ||
+ | | process 2 wafer size = 300 mm | ||
+ | | process 2 transistor = FinFET | ||
+ | | process 2 volt = | ||
+ | | process 2 delta from = [[5 nm]] Δ | ||
+ | | process 2 fin pitch = 23 nm | ||
+ | | process 2 fin pitch Δ = 0.82x | ||
+ | | process 2 fin width = | ||
+ | | process 2 fin width Δ = | ||
+ | | process 2 fin height = | ||
+ | | process 2 fin height Δ = | ||
+ | | process 2 gate len = | ||
+ | | process 2 gate len Δ = | ||
+ | | process 2 cpp = 47 nm <!-- (after 4% optical bloat) --> | ||
+ | | process 2 cpp Δ = 0.88x/<!-- (drawn) -->0.92x <!-- (actual) --> | ||
+ | | process 2 mmp = 23 nm | ||
+ | | process 2 mmp Δ = 0.82x | ||
+ | | process 2 sram hp = | ||
+ | | process 2 sram hp Δ = | ||
+ | | process 2 sram hd = 0.0199 μm<sup>2</sup> | ||
+ | | process 2 sram hd Δ = 0.95x | ||
+ | | process 2 sram lv = | ||
+ | | process 2 sram lv Δ = | ||
+ | | process 2 dram = | ||
+ | | process 2 dram Δ = | ||
+ | | process 3 fab = [[TSMC]] | ||
+ | | process 3 name = N3E <info>N3 Enhanced</info>, N3P, N3X | ||
+ | | process 3 date = 2024, 2025, 2026 | ||
+ | | process 3 lith = EUV | ||
+ | | process 3 immersion = Yes | ||
+ | | process 3 exposure = SALELE | ||
+ | | process 3 wafer type = Bulk | ||
+ | | process 3 wafer size = 300 mm | ||
+ | | process 3 transistor = FinFET | ||
+ | | process 3 volt = | ||
+ | | process 3 delta from = [[5 nm]] Δ | ||
+ | | process 3 fin pitch = 23 nm | ||
+ | | process 3 fin pitch Δ = 0.82x | ||
+ | | process 3 fin width = | ||
+ | | process 3 fin width Δ = | ||
+ | | process 3 fin height = | ||
+ | | process 3 fin height Δ = | ||
+ | | process 3 gate len = | ||
+ | | process 3 gate len Δ = | ||
+ | | process 3 cpp = 48 nm | ||
+ | | process 3 cpp Δ = 0.94x | ||
+ | | process 3 mmp = 23 nm | ||
+ | | process 3 mmp Δ = 0.82x | ||
+ | | process 3 sram hp = | ||
+ | | process 3 sram hp Δ = | ||
+ | | process 3 sram hd = 0.0210 μm<sup>2</sup> | ||
+ | | process 3 sram hd Δ = 1.0x | ||
+ | | process 3 sram lv = | ||
+ | | process 3 sram lv Δ = | ||
+ | | process 3 dram = | ||
+ | | process 3 dram Δ = | ||
+ | <!-- Samsung --> | ||
+ | | process 4 fab = [[Samsung]] | ||
+ | | process 4 name = SF3E <info>3nm Gate All Around Early</info>, SF3 <info>3nm Gate All Around Plus</info> | ||
+ | | process 4 date = 2024 | ||
+ | | process 4 lith = EUV | ||
+ | | process 4 immersion = Yes | ||
+ | | process 4 exposure = SALELE | ||
+ | | process 4 wafer type = Bulk | ||
+ | | process 4 wafer size = 300 mm | ||
+ | | process 4 transistor = GAAFET (MBCFET) | ||
+ | | process 4 volt = | ||
+ | | process 4 delta from = [[5 nm]] Δ | ||
+ | | process 4 fin pitch = | ||
+ | | process 4 fin pitch Δ = | ||
+ | | process 4 fin width = | ||
+ | | process 4 fin width Δ = | ||
+ | | process 4 fin height = | ||
+ | | process 4 fin height Δ = | ||
+ | | process 4 gate len = | ||
+ | | process 4 gate len Δ = | ||
+ | | process 4 cpp = 48 nm | ||
+ | | process 4 cpp Δ = 1.0x | ||
+ | | process 4 mmp = 28 nm | ||
+ | | process 4 mmp Δ = 1.0x | ||
+ | | process 4 sram hp = | ||
+ | | process 4 sram hp Δ = | ||
+ | | process 4 sram hd = | ||
+ | | process 4 sram hd Δ = | ||
+ | | process 4 sram lv = | ||
+ | | process 4 sram lv Δ = | ||
+ | | process 4 dram = | ||
+ | | process 4 dram Δ = <br>. | ||
+ | }} | ||
+ | |||
+ | |||
+ | === 3 nm process nodes === | ||
+ | {{see also|Intel 7|Intel 4|Intel 18A}} | ||
+ | {| class="wikitable" cellpadding="3px" style="border: 1px solid black; border-spacing: 0px; width: 100%; text-align:center;" | ||
+ | ! | ||
+ | ! colspan=2 | [[Samsung]] <ref>{{cite book |url=https://fuse.wikichip.org/news/6932/samsung-3nm-gaafet-enters-risk-production-discusses-next-gen-improvements/ |title=Samsung 3nm GAAFET Enters Risk Production; Discusses Next-Gen Improvements |website=WikiChip Fuse |date=5 July 2022}}</ref><ref>{{cite book |url=https://ieeexplore.ieee.org/abstract/document/10185353 |title=https://ieeexplore.ieee.org/abstract/document/10185353 |website=IEEEexplore |date=16 June 2023}}</ref> | ||
+ | ! colspan=4 | [[TSMC]] <ref>{{cite book |url=https://fuse.wikichip.org/news/7375/tsmc-n3-and-challenges-ahead/ |title=TSMC N3, and Challenges Ahead |date=27 May 2023}}</ref><ref>{{cite book |url=https://www-prod.techinsights.com/blog/apple-apl1v02-a17-pro-processor-tsmc-3-nm-finfet-process-digital-floorplan-analysis?utm_source=direct&utm_medium=website |title=Apple APL1V02 A17 Pro Processor TSMC 3 nm FinFET Process Digital Floorplan Analysis |date=27 Dec 2023}}</ref><ref>{{cite book |url=https://www.techinsights.com/blog/introducing-tsmc-n3e-power-behind-apples-m4-soc?utm_source=direct&utm_medium=website |title=Introducing TSMC N3E |date=15 April 2024}}</ref> | ||
+ | ! [[Intel]] <ref>{{cite book |title=Intel Delivers Leading-Edge Foundry Node with Intel 3 Technology; on Path Back to Process Leadership |url=https://community.intel.com/t5/Blogs/Intel-Foundry/Systems-Foundry-for-the-AI-Era/Intel-Delivers-Leading-Edge-Foundry-Node-with-Intel-3-Technology/post/1607454 |date=18 June 2024 |website=community.intel.com}}</ref> | ||
+ | |- | ||
+ | ! Process name | ||
+ | | 3GAE (SF3E) | ||
+ | | 3GAP (SF3) | ||
+ | | N3 (N3B) | ||
+ | | N3E | ||
+ | | N3P | ||
+ | | N3X | ||
+ | | [[Intel 3]] | ||
+ | |- | ||
+ | ! Transistor type | ||
+ | | colspan=2 | [[GAAFET]] ([[MBCFET]]) | ||
+ | | colspan=5 | [[FinFET]] | ||
+ | |- | ||
+ | ! Transistor density <br>(MTr/mm<sup>2</sup>) | ||
+ | | 150 | ||
+ | | 190 | ||
+ | | 197 | ||
+ | | 216 | ||
+ | | colspan=2 | 224 | ||
+ | | - | ||
+ | |- | ||
+ | ! Transistor gate <br>pitch (nm) | ||
+ | | 40 | ||
+ | | - | ||
+ | | 45 | ||
+ | | 48 | ||
+ | | - | ||
+ | | - | ||
+ | | 50 | ||
+ | |- | ||
+ | ! Interconnect <br>pitch (nm) | ||
+ | | 32 | ||
+ | | - | ||
+ | | - | ||
+ | | 23 | ||
+ | | - | ||
+ | | - | ||
+ | | 30 | ||
+ | |- | ||
+ | ! SRAM bit-cell <br>size (μm<sup>2</sup>) | ||
+ | | - | ||
+ | | - | ||
+ | | 0.0199 | ||
+ | | 0.021 | ||
+ | | - | ||
+ | | - | ||
+ | | - | ||
+ | |- | ||
+ | ! Release status | ||
+ | | 2022 H1 risk <br>production <br>2022 H2 <br>production <br>2022 shipping | ||
+ | | 2024 Q1 risk <br>production <br>2024 H2 <br>production | ||
+ | | 2022 <br>risk production <br>2023<br>production | ||
+ | | 2024 <br> | ||
+ | | 2025 <br> | ||
+ | | 2026 <br> | ||
+ | | 2024 | ||
+ | |- | ||
+ | |} | ||
+ | |||
+ | |||
+ | === 2 nm process nodes === | ||
+ | {| class="wikitable" style="text-align:center" | ||
+ | ! | ||
+ | !colspan="4" | [[Samsung]] | ||
+ | ! colspan="3" | [[TSMC]] | ||
+ | ! colspan="2" | [[Intel]] | ||
+ | |- | ||
+ | ! Process name | ||
+ | | SF2 || SF2P || SF2X || SF2Z | ||
+ | | N2 || N2P || N2X | ||
+ | | 20A || 18A | ||
+ | |- | ||
+ | ! Transistor type | ||
+ | | colspan="4" |[[MBCFET]] | ||
+ | | colspan="3" |[[GAAFET]] | ||
+ | | colspan="2" |[[RibbonFET]] | ||
+ | |- | ||
+ | ! Transistor density (MTr/mm<sup>2</sup>) | ||
+ | | 231 || - || - || - | ||
+ | | 313 || - || - | ||
+ | | - || 238 | ||
+ | |- | ||
+ | ! Transistor gate <br>pitch (nm) | ||
+ | | - || - || - || - | ||
+ | | - || - || - | ||
+ | | - || - | ||
+ | |- | ||
+ | ! Interconnect <br>pitch (nm) | ||
+ | | - || - || - || - | ||
+ | | - || - || - | ||
+ | | - || - | ||
+ | |- | ||
+ | ! SRAM bit-cell <br>size (μm<sup>2</sup>) | ||
+ | | - || - || - || - | ||
+ | | 0.0175 μm² || - || - | ||
+ | | - || 0.021 μm² | ||
+ | |- | ||
+ | ! Release status | ||
+ | | 2025 volume production || 2026 volume production || 2026 volume production || 2027 volume production | ||
+ | | 2024 H2 risk production <br>2025 H2 volume production || 2026 H2 <br>volume production || 2026 H2 <br>volume production | ||
+ | | 2024 H1 risk production <br>2024 volume production <br>Canceled 2024 || 2024 H2 risk production <br>2025 H1 production | ||
+ | |- | ||
+ | |} | ||
+ | |||
+ | == 3 nm Microprocessors== | ||
+ | * [[Apple]] | ||
+ | ** [[apple/ax|A17 Pro]] | ||
+ | ** [[apple/ax|A18]] | ||
+ | ** [[apple/ax|A18 Pro]] | ||
+ | ** [[apple/mx|M3]] | ||
+ | ** [[apple/mx|M3 Pro]] | ||
+ | ** [[apple/mx|M3 MAX]] | ||
+ | ** [[apple/mx|M3 Ultra]] | ||
+ | ** [[apple/mx|M4]] | ||
+ | ** [[apple/mx|M4 Pro]] | ||
+ | ** [[apple/mx|M4 Max]] | ||
+ | |||
+ | * [[Qualcomm]] | ||
+ | ** {{qualcomm|Snapdragon 8}} Elite | ||
+ | * [[MediaTek]] | ||
+ | ** {{mediatek|Dimensity}} 9400 | ||
− | |||
{{expand list}} | {{expand list}} | ||
− | == 3 | + | == 3 nm Microarchitectures == |
+ | |||
+ | * [[AMD]] | ||
+ | ** {{amd|Zen 5|l=arch}}c | ||
+ | ** {{amd|Zen 6|l=arch}} | ||
+ | |||
+ | === 4 nm Microarchitectures === | ||
+ | * [[AMD]] | ||
+ | ** {{amd|Zen 5|l=arch}} | ||
+ | |||
+ | * [[ARM]] • [[Cortex]] | ||
+ | ** {{armh|Hayes|Cortex-A520|l=arch}} | ||
+ | |||
+ | * [[Neoverse]] V2 | ||
+ | ** [[AWS Graviton4]] | ||
+ | |||
{{expand list}} | {{expand list}} | ||
− | [[ | + | == References == |
+ | * Kinam Kim, President of Semiconductor Business, announced MBCFET for the node after [[5 nm]], May 24, 2017 | ||
+ | |||
+ | [[category:lithography]] |
Latest revision as of 19:15, 27 June 2025
The 3 nanometer (3 nm or 30 Å) lithography process is a technology node semiconductor manufacturing
- process following the 5 nm process node.
Commercial integrated circuit manufacturing using 3 nm process is set to begin some time around 2023.
The term "3 nm" is simply a commercial name for a generation of a certain size and its technology, and
- does not represent any geometry of the transistor.
Contents
Industry[edit]
Intel[edit]
P1278 Intel's 5-nanometer (renamed as Intel 20A) process node is expected to ramp around the 2H 2024/2025 timeframe.
TSMC[edit]
N3 technology will offer up to 70% logic density gain, up to 15% speed improvement at the same power and up to
- 30% power reduction at the same speed as compared with N5 technology (According to TSMCs website).
If this holds true we could see 300+ MT/mm2.
Samsung[edit]
On May 24, 2017 Samsung announced they will be switching to a transistor they call Multi-Bridge-Channel FET
- (MBCFET), an extension of a Gate-all-around (GAA) FET.
This is planned for somewhere after the 5 nm node but the exact timeline or specification is currently unknown.
Specifications[edit]
Process Name | |
---|---|
1st Production | |
Litho- graphy |
Lithography |
Immersion | |
Exposure | |
Wafer | Type |
Size | |
Tran- sistor |
Type |
Voltage | |
Fin | Pitch |
Width | |
Height | |
Gate Length (Lg) | |
Contacted Gate Pitch (CPP) | |
Minimum Metal Pitch (MMP) | |
SRAM bitcell |
High-Perf (HP) |
High-Density (HD) | |
Low-Voltage (LV) | |
DRAM bitcell |
eDRAM |
Intel | TSMC | TSMC | Samsung | ||||
---|---|---|---|---|---|---|---|
P1276 (CPU/Chipset) | N3(B) | N3E N3 Enhanced , N3P, N3X |
SF3E 3nm Gate All Around Early , SF3 3nm Gate All Around Plus
| ||||
2024 | 2023 | 2024, 2025, 2026 | 2024 | ||||
EUV | EUV | EUV | EUV | ||||
Yes | Yes | Yes | Yes | ||||
SE EUV+DUV SAPQ | SALELE | SALELE | SALELE | ||||
Bulk | Bulk | Bulk | Bulk | ||||
300 mm | 300 mm | 300 mm | 300 mm | ||||
FinFET | FinFET | FinFET | GAAFET (MBCFET) | ||||
Value | 4 nm Δ | Value | 5 nm Δ | Value | 5 nm Δ | Value | 5 nm Δ |
30 nm | 1.0x | 23 nm | 0.82x | 23 nm | 0.82x | ||
narrower | |||||||
taller | |||||||
1.0x | |||||||
50 nm | 1.0x | 47 nm | 0.88x/0.92x | 48 nm | 0.94x | 48 nm | 1.0x |
30 nm | 1.0x | 23 nm | 0.82x | 23 nm | 0.82x | 28 nm | 1.0x |
0.0300 μm2 | 1.0x | ||||||
0.0240 μm2 | 1.0x | 0.0199 μm2 | 0.95x | 0.0210 μm2 | 1.0x | ||
. |
3 nm process nodes[edit]
Samsung [1][2] | TSMC [3][4][5] | Intel [6] | |||||
---|---|---|---|---|---|---|---|
Process name | 3GAE (SF3E) | 3GAP (SF3) | N3 (N3B) | N3E | N3P | N3X | Intel 3 |
Transistor type | GAAFET (MBCFET) | FinFET | |||||
Transistor density (MTr/mm2) |
150 | 190 | 197 | 216 | 224 | - | |
Transistor gate pitch (nm) |
40 | - | 45 | 48 | - | - | 50 |
Interconnect pitch (nm) |
32 | - | - | 23 | - | - | 30 |
SRAM bit-cell size (μm2) |
- | - | 0.0199 | 0.021 | - | - | - |
Release status | 2022 H1 risk production 2022 H2 production 2022 shipping |
2024 Q1 risk production 2024 H2 production |
2022 risk production 2023 production |
2024 |
2025 |
2026 |
2024 |
2 nm process nodes[edit]
Samsung | TSMC | Intel | |||||||
---|---|---|---|---|---|---|---|---|---|
Process name | SF2 | SF2P | SF2X | SF2Z | N2 | N2P | N2X | 20A | 18A |
Transistor type | MBCFET | GAAFET | RibbonFET | ||||||
Transistor density (MTr/mm2) | 231 | - | - | - | 313 | - | - | - | 238 |
Transistor gate pitch (nm) |
- | - | - | - | - | - | - | - | - |
Interconnect pitch (nm) |
- | - | - | - | - | - | - | - | - |
SRAM bit-cell size (μm2) |
- | - | - | - | 0.0175 μm² | - | - | - | 0.021 μm² |
Release status | 2025 volume production | 2026 volume production | 2026 volume production | 2027 volume production | 2024 H2 risk production 2025 H2 volume production |
2026 H2 volume production |
2026 H2 volume production |
2024 H1 risk production 2024 volume production Canceled 2024 |
2024 H2 risk production 2025 H1 production |
3 nm Microprocessors[edit]
- Qualcomm
- Snapdragon 8 Elite
- MediaTek
- Dimensity 9400
This list is incomplete; you can help by expanding it.
3 nm Microarchitectures[edit]
4 nm Microarchitectures[edit]
This list is incomplete; you can help by expanding it.
References[edit]
- Kinam Kim, President of Semiconductor Business, announced MBCFET for the node after 5 nm, May 24, 2017