Semiconductor Memories

Transcription

1 Digital Integrated Circuits A Design Perspective Jan M. Rabaey Anantha Chandrakasan Borivoje Nikolic Semiconductor Memories December 20, 2002

2 Chapter Overview Memory Classification Memory Architectures The Memory Core Periphery Reliability Case Studies

3 Semiconductor Memory Classification Read-Write Memory Non-Volatile Read-Write Memory Read-Only Memory Random Access Non-Random Access EPROM E 2 PROM Mask-Programmed Programmable (PROM) SRAM FIFO FLASH DRAM LIFO Shift Register CAM

4 Memory Timing: Definitions Read cycle READ Read access Read access Write cycle WRITE Data valid Write access DATA Data written

5 Memory Architecture: Decoders M bits M bits S 0 S 1 S 2 Word 0 Word 1 Word 2 Storage cell A 0 A 1 S 0 Word 0 Word 1 Word 2 Storage cell S N 2 2 Word N2 2 A K 2 1 Word N2 2 S N 2 1 Word N2 1 K 5 log 2 N Word N2 1 Input-Output (M bits) Input-Output (M bits) Intuitive architecture for N x M memory Too many select signals: N words == N select signals Decoder reduces the number of select signals K = log 2 N

6 Array-Structured Memory Architecture Problem: ASPECT RATIO or HEIGHT >> WIDTH 2 L 2 K Bit line Storage cell A K A K11 A L 21 Row Decoder Word line Sense amplifiers / Drivers M.2 K Amplify swing to rail-to-rail amplitude A 0 A K21 Column decoder Selects appropriate word Input-Output (M bits)

7 Hierarchical Memory Architecture Block 0 Block i Block P 2 1 Row address Column address Block address Control circuitry Block selector Global amplifier/driver Global data bus Advantages: 1. Shorter wires within blocks 2. Block address activates only 1 block => power savings I/O

8 Block Diagram of 4 Mbit SRAM Clock generator Z -address buffer X -address buffer Predecoder and block selector Bit line load Transfer gate Column decoder Sense amplifier and write driver CS, WE buffer I/O buffer x1/x4 controller Y-address buffer X -address buffer Digital Integrated Circuits 2nd [Hirose90] Memories

9 Contents-Addressable Memory I/O Buffers Data (64 bits) Commands Comparand Mask Control Logic R/W Address (9 bits) Address Decoder CAM Array 2 9 words 3 64 bits 2 9 Validity Bits Priority Encoder

10 Memory Timing: Approaches Address bus Row Address Column Address RAS CAS Address Bus Address Address transition initiates memory operation RAS-CAS timing DRAM Timing Multiplexed Adressing SRAM Timing Self-timed

11 Read-Only Memory Cells BL BL BL 1 WL WL V DD WL BL BL BL 0 WL WL WL GND Diode ROM MOS ROM 1 MOS ROM 2

12 MOS OR ROM BL[0] BL[1] BL[2] BL[3] WL[0] V DD WL[1] WL[2] V DD WL[3] V bias Pull-down loads

13 MOS NOR ROM V DD Pull-up devices WL[0] WL [1] GND WL [2] GND WL [3] BL [0] BL [1] BL [2] BL [3]

14 MOS NOR ROM Layout Cell (9.5λ x 7λ) Programmming using the Active Layer Only Polysilicon Metal1 Diffusion Metal1 on Diffusion

15 MOS NOR ROM Layout Cell (11λ x 7λ) Programmming using the Contact Layer Only Polysilicon Metal1 Diffusion Metal1 on Diffusion

16 MOS NAND ROM V DD Pull-up devices BL[0] BL[1] BL[2] BL[3] WL[0] WL[1] WL[2] WL[3] All word lines high by default with exception of selected row

17 MOS NAND ROM Layout Cell (8λ x 7λ) Programmming using the Metal-1 Layer Only No contact to VDD or GND necessary; drastically reduced cell size Loss in performance compared to NOR ROM Polysilicon Diffusion Metal1 on Diffusion

18 NAND ROM Layout Cell (5λ x 6λ) Programmming using Implants Only Polysilicon Threshold-altering implant Metal1 on Diffusion

19 Equivalent Transient Model for MOS NOR ROM Model for NOR ROM V DD WL r word C bit BL c word Word line parasitics Wire capacitance and gate capacitance Wire resistance (polysilicon) Bit line parasitics Resistance not dominant (metal) Drain and Gate-Drain capacitance

20 Equivalent Transient Model for MOS NAND ROM Model for NAND ROM V DD BL r bit C L WL r word c bit c word Word line parasitics Similar to NOR ROM Bit line parasitics Resistance of cascaded transistors dominates Drain/Source and complete gate capacitance

21 Decreasing Word Line Delay WL Driver Polysilicon word line Metal word line (a) Driving the word line from both sides Metal bypass WL K cells (b) Using a metal bypass Polysilicon word line (c) Use silicides

22 Precharged MOS NOR ROM f pre V DD Precharge devices WL [0] WL [1] GND WL [2] GND WL [3] BL [0] BL [1] BL [2] BL [3] PMOS precharge device can be made as large as necessary, but clock driver becomes harder to design.

23 Non-Volatile Memories The Floating-gate gate transistor (FAMOS) Source Floating gate Gate Drain D t ox G n + Substrate p t ox n +_ S Device cross-section Schematic symbol

24 Floating-Gate Transistor Programming 20 V 0 V 5 V 10 V 5 V 20 V 2 5 V 0 V V 5 V S D S D S D Avalanche injection Removing programming voltage leaves charge trapped Programming results in higher V T.

25 A Programmable-Threshold Transistor I D 0 -state 1 -state ON DV T OFF V WL V GS

26 FLOTOX EEPROM Floating gate Source Gate Drain I nm n 1 Substrate p n 1 10 nm -10 V 10 V V GD FLOTOX transistor Fowler-Nordheim I-V characteristic

27 EEPROM Cell BL WL V DD Absolute threshold control is hard Unprogrammed transistor might be depletion 2 transistor cell

28 Flash EEPROM Control gate Floating gate erasure n 1 source programming p-substrate Thin tunneling oxide n 1 drain Many other options

29 Cross-sections sections of NVM cells Flash EPROM Digital Integrated Circuits 2nd Courtesy Intel Memories

30 Basic Operations in a NOR Flash Memory Erase cell array BL 0 BL 1 12 V G 0 V WL 0 S D 12 V 0 V WL 1 open open

31 Basic Operations in a NOR Flash Memory Write 12 V BL 0 BL 1 G 6 V 12 V WL 0 S D 0 V 0 V WL 1 6 V 0 V

32 Basic Operations in a NOR Flash Memory Read 5 V G 1 V 5 V BL 0 BL 1 WL 0 S D 0 V 0 V WL 1 1 V 0 V

33 NAND Flash Memory Word line(poly) Unit Cell Gate ONO Gate Oxide FG Source line (Diff. Layer) Digital Integrated Circuits 2nd Courtesy Toshiba Memories

34 NAND Flash Memory Select transistor Word lines Active area STI Bit line contact Source line contact Digital Integrated Circuits 2nd Courtesy Toshiba Memories

35 Characteristics of State-of of-the-art NVM

36 Read-Write Memories (RAM) STATIC (SRAM) Data stored as long as supply is applied Large (6 transistors/cell) Fast Differential DYNAMIC (DRAM) Periodic refresh required Small (1-3 transistors/cell) Slower Single Ended

37 6-transistor CMOS SRAM Cell WL V DD M 2 M 4 Q M Q M 5 6 M 1 M 3 BL BL

38 CMOS SRAM Analysis (Read) WL BL V DD M 4 Q = 0 Q = 1 M 6 M 5 BL V DD M 1 V DD V DD C bit C bit

39 V o l t a g e r i s e [ V ] CMOS SRAM Analysis (Read) Voltage Rise (V) Cell Ratio (CR) 2.5 3

40 CMOS SRAM Analysis (Write) V DD WL M 4 Q = 0 M 6 M 5 Q = 1 M 1 V DD BL = 1 BL = 0

41 CMOS SRAM Analysis (Write)

42 6T-SRAM Layout V DD M2 M4 Q Q M1 M3 M5 M6 GND WL BL BL

43 Resistance-load SRAM Cell V DD WL R L R L M 3 Q Q M 4 BL M 1 M 2 BL Static power dissipation -- Want R L large Bit lines precharged to V DD to address t p problem

44 SRAM Characteristics

45 3-Transistor DRAM Cell BL1 BL2 WWL RWL WWL M 3 RWL M 1 X M 2 X V DD 2 V T C S BL 1 V DD BL 2 V DD 2 V T DV No constraints on device ratios Reads are non-destructive Value stored at node X when writing a 1 = V WWL -V Tn

46 3T-DRAM Layout BL2 BL1 GND RWL M3 M2 WWL M1

47 1-Transistor DRAM Cell WL BL WL Write 1 Read 1 M 1 X GND V DD 2 V T C S BL V DD V DD /2 V sensing DD /2 C BL Write: C S is charged or discharged by asserting WL and BL. Read: Charge redistribution takes places between bit line and storage capacitance ΔV = VBL V PRE = V BIT V PRE C S C S + C BL Voltage swing is small; typically around 250 mv.

48 DRAM Cell Observations 1T DRAM requires a sense amplifier for each bit line, due to charge redistribution read-out. DRAM memory cells are single ended in contrast to SRAM cells. The read-out of the 1T DRAM cell is destructive; read and refresh operations are necessary for correct operation. Unlike 3T cell, 1T cell requires presence of an extra capacitance that must be explicitly included in the design. When writing a 1 into a DRAM cell, a threshold voltage is lost. This charge loss can be circumvented by bootstrapping the word lines to a higher value than V DD

49 Sense Amp Operation V BL V(1) V PRE DV(1) Sense amp activated Word line activated V(0) t

50 1-T T DRAM Cell Capacitor Metal word line Poly n + n + Inversion layer Poly induced by plate bias Cross-section SiO 2 Field Oxide Diffused bit line Polysilicon gate Layout Polysilicon plate M 1 word line Uses Polysilicon-Diffusion Capacitance Expensive in Area

51 SEM of poly-diffusion capacitor 1T-DRAM

52 Advanced 1T DRAM Cells Word line Insulating Layer Cell plate Capacitor dielectric layer Cell Plate Si Capacitor Insulator Refilling Poly Transfer gate Storage electrode Isolation Storage Node Poly 2nd Field Oxide Si Substrate Trench Cell Stacked-capacitor Cell

53 Static CAM Memory Cell Word Bit CAM Bit Bit CAM Bit Bit M4 M8 M6 M9 M7 M5 Bit Word CAM CAM Word Match S M3 int M2 S M1 Wired-NOR Match Line

54 A d d r e s s D e c o d e r H i t L o g i c CAM in Cache Memory CAM ARRAY SRAM ARRAY Input Drivers Sense Amps / Input Drivers Address Tag Hit R/W Data

55 Periphery Decoders Sense Amplifiers Input/Output Buffers Control / Timing Circuitry

56 Row Decoders Collection of 2 M complex logic gates Organized in regular and dense fashion (N)AND Decoder NOR Decoder

57 Hierarchical Decoders Multi-stage implementation improves performance WL 1 WL 0 A 0 A 1 A 0 A 1 A 0 A 1 A 0 A 1 A 2 A 3 A 2 A 3 A 2 A 3 A 2 A 3 A 1 A 0 A 0 A 1 A 3 A 2 A 2 A 3 NAND decoder using 2-input pre-decoders

58 Dynamic Decoders Precharge devices GND GND V DD WL 3 V DD WL 3 WL 2 V DD WL 2 WL 1 WL 0 V DD WL 1 WL 0 V DD φ A 0 A 0 A 1 A 1 A 0 A 0 A 1 A 1 φ 2-input NOR decoder 2-input NAND decoder

59 2 - i n p u t N O R d e c o d e r 4-input pass-transistor based column decoder BL 0 BL 1 BL 2 BL 3 A 0 S 0 S 1 S 2 A 1 S 3 Advantages: speed (t pd does not add to overall memory access time) Only one extra transistor in signal path Disadvantage: Large transistor count D

60 4-to-11 tree based column decoder BL 0 BL 1 BL 2 BL 3 A 0 A 0 A 1 A 1 Number of devices drastically reduced Delay increases quadratically with # of sections; prohibitive for large decoders Solutions: buffers progressive sizing combination of tree and pass transistor approaches D

61 Decoder for circular shift-register V DD V DD V DD V DD V DD V DD WL 0 R f f f f R WL 1 f f f f R WL 2 f f f f V DD

62 Sense Amplifiers t p = C ΔV I av make ΔV as small as possible large small Idea: Use Sense Amplifer small transition s.a. input output

63 Differential Sense Amplifier V DD M 3 M 4 y Out bit M 1 M 2 bit SE M 5 Directly applicable to SRAMs

64 Differential Sensing SRAM V DD PC V DD BL EQ BL y M 3 V DD M 4 V DD 2 y WL i x M 1 M 2 2 x x 2 x SE M 5 SE SRAM cell i SE x Diff. Sense 2 x Amp V DD y Output Output (a) SRAM sensing scheme SE (b) two stage differential amplifier

65 Latch-Based Sense Amplifier (DRAM) EQ BL BL V DD SE SE Initialized in its meta-stable point with EQ Once adequate voltage gap created, sense amp enabled with SE Positive feedback quickly forces output to a stable operating point.

66 Charge-Redistribution Amplifier V ref V L M 1 V S M 2 M 3 C large C small Concept Transient Response V in V S V 1.0 V L time (nsec) V ref 5 3V 3.00

67 Charge-Redistribution Amplifier EPROM V DD SE M 4 Load Out V casc M 3 C out Cascode device WLC M 2 C col Column decoder BL WL M 1 C BL EPROM array

68 Single-to to-differential Conversion BL Cell WL x Diff. S.A. 2 x 1 2 V ref Output How to make a good V ref?

69 Open bitline architecture with dummy cells EQ L L 1 L 0 V DD R 0 SE R 1 L BLL BLR C S C S C S SE C S C S C S Dummy cell Dummy cell

70 V V V DRAM Read Process with Dummy Cell BL BL 1 BL 1 BL t (ns) t (ns) reading 0 reading 1 3 EQ WL 2 SE t (ns) control signals

71 Voltage Regulator V DD M drive V DL V REF Equivalent Model V bias V REF - + M drive V DL

72 Charge Pump V DD 2V DD 2 V T CLK A M1 B V B V DD 2 V T 0 V C pump M2 V load C load V load 0 V

73 DRAM Timing

74 m u x / d e m u x n e t w o r k RDRAM Architecture Bus Clocks Data bus k k3 l memory array Column Row demux demux packet dec. packet dec.

75 Address Transition Detection V DD A 0 DELAY t d ATD ATD A 1 DELAY t d A N 2 1 DELAY t d

76 Reliability and Yield

77 V, V, C, Q, C s m a x D S S D D Sensing Parameters in DRAM 1000 C D(1F) V smax (mv) 100 C S(1F) Q S(1C) 10 V DD (V) Q S 5 C S V DD /2 V smax 5 Q S /(C S 1 C D ) 4K 64K 1M 16M 256M 4G 64G Memory Capacity (bits /chip) From [Itoh01]

78 Noise Sources in 1T DRam BL substrate Adjacent BL C WBL a-particles WL leakage C S electrode C cross

79 Open Bit-line Architecture Cross Coupling EQ BL WL 1 WL 0 C WBL WL D WL D WL 0 WL 1 C WBL BL C BL C C C Sense Amplifier C C C C BL

80 Folded-Bitline Architecture WL 1 WL 1 WL 0 WL 0 WL D C WBL WL D BL C BL x y C C C C C C EQ Sense Amplifier BL C BL x y C WBL

81 Transposed-Bitline Architecture BL 9 BL BL BL 99 C cross (a) Straightforward bit-line routing SA BL 9 BL BL BL 99 C cross SA (b) Transposed bit-line architecture

82 Alpha-particles particles (or Neutrons) a-particle WL V DD BL n SiO 2 1 Particle ~ 1 Million Carriers

83 Yield Yield curves at different stages of process maturity (from [Veendrick92])

84 R o w D e c o d e r Redundancy Redundant columns Redundant rows Memory Array Row Address Fuse : Bank Column Decoder Column Address

85 Error-Correcting Codes Example: Hamming Codes with e.g. B3 Wrong 1 1 = 3 0

86 Redundancy and Error Correction

87 Sources of Power Dissipation in Memories V DD CHIP I DD 5SC i DV i f1s I DCP C PT V INT f nc DE V INT f selected m mi act I DCP n ROW DEC non-selected ARRAY m(n 2 1)i hld PERIPHERY mc DE V INT f COLUMN DEC V SS Digital Integrated Circuits 2nd From [Itoh00] Memories

88 ( A ) Data Retention in SRAM 1.30u I leakage 1.10u 900n 700n 500n Factor m CMOS 300n 100n 0.18 m CMOS m V DD SRAM leakage increases with technology scaling

89 Suppressing Leakage in SRAM sleep V DD low-threshold transistor V DD V DDL V DD,int sleep V DD,int SRAM cell SRAM cell SRAM cell SRAM cell SRAM cell SRAM cell sleep V SS,int Inserting Extra Resistance Reducing the supply voltage

90 Data Retention in DRAM I ACT Current (A) I DC I AC Cycle time : 150 ns T 5 75 C,S 5 97 mv/dec M 64M 255M 1G 4G 15G 64G Capacity (bit) Operating voltage (V) Extrapolated threshold voltage at 25 C (V) Digital Integrated Circuits 2nd From [Itoh00] Memories

91 Case Studies Programmable Logic Array SRAM Flash Memory

92 PLA versus ROM Programmable Logic Array structured approach to random logic two level logic implementation NOR-NOR (product of sums) NAND-NAND (sum of products) IDENTICAL TO ROM! Main difference ROM: fully populated PLA: one element per minterm Note: Importance of PLA s has drastically reduced 1. slow 2. better software techniques (mutli-level logic synthesis) But

93 Programmable Logic Array Pseudo-NMOS PLA GND GND GND GND V DD GND GND GND V DD X 0 X 0 X 1 X 1 X 2 X 2 f 0 f 1 AND-plane OR-plane

94 Dynamic PLA f AND GND V DD f OR V DD f AND X 0 X 0 X 1 X 1 X 2 X 2 f 0 f 1 GND f OR AND-plane OR-plane

95 Clock Signal Generation for self-timed dynamic PLA f f Dummy AND row f AND f AND t pre t eval f AND Dummy AND row f OR f OR (a) Clock signals (b) Timing generation circuitry

96 PLA Layout V DD And-Plane Or-Plane φ GND x 0 x 0 x 1 x 1 x 2 x 2 Pull-up devices f 0 f 1 Pull-up devices

97 4 Mbit SRAM Hierarchical Word-line Architecture Global word line Sub-global word line Local word line Local word line Block group select Memory cell Block select Block 0 Block 1 Block select Block 2...

98 Bit-line Circuitry Bit-line load Block select ATD BEQ Local WL Memory cell B/T B/T CD I/O line CD Sense amplifier CD I/O I/O

99 Sense Amplifier (and Waveforms) Address I/O I/O SEQ BS Block select ATD ATD BEQ SA BS SA SEQ Vdd I/O Lines GND SEQ SEQ DATA SEQ De i SEQ Vdd SA, SA GND DATA BS Data-cut

100 1 Gbit Flash Memory 512Mb Memory Array BL0 BL1 BL Mb Memory Array BL16996 BL16897 BL33791 Word Line Driver SGD WL31 WL0 SGS Block0 Block1023 BLT0 BLT1 Sense Latches ( ) 3 8 Word Line Driver Word Line Driver Block0 Block1023 Bit Line Control Circuit Sense Latches ( ) 3 8 Word Line Driver Data Caches ( ) 3 8 I/O I/O Data Caches ( ) 3 8 Digital Integrated Circuits 2nd From [Nakamura02] Memories

101 N u m b e r o f c e l l s R e a d l e v e l ( 4. 5 V ) Writing Flash Memory Verify level V Word-line level V Number of memory cells 0V 1V 2V 3V 4V Vt of memory cells Result of 4 times program V 1V 2V 3V Vt of memory cells 4V Evolution of thresholds Final Distribution Digital Integrated Circuits 2nd From [Nakamura02] Memories

102 125mm 2 1Gbit NAND Flash Memory 10.7mm Charge pump 2kB Page buffer & cache 32 word lines x 1024 blocks bit lines 11.7mm Digital Integrated Circuits 2nd From [Nakamura02] Memories

103 125mm 2 1Gbit NAND Flash Memory Technology 0.13μm p-sub CMOS triple-well 1poly, 1polycide, 1W, 2Al Cell size 0.077μm2 Chip size 125.2mm2 Organization 2112 x 8b 8b x page x 1k 1k block Power supply 2.7V-3.6V Cycle time 50ns Read time 25μs Program time 200μs // page Erase time 2ms // block Digital Integrated Circuits 2nd From [Nakamura02] Memories

104 Semiconductor Memory Trends (up to the 90 s) Memory Size as a function of time: x 4 every three years

105 Semiconductor Memory Trends (updated) Digital Integrated Circuits 2nd From [Itoh01] Memories

106 Trends in Memory Cell Area Digital Integrated Circuits 2nd From [Itoh01] Memories

107 Semiconductor Memory Trends Technology feature size for different SRAM generations