2015
DOI: 10.1587/elex.12.20152002
|View full text |Cite
|
Sign up to set email alerts
|

Review of wafer-level three-dimensional integration (3DI) using bumpless interconnects for tera-scale generation

Abstract: The prospects of three-dimensional (3D) integration for Terabyte large scale integration using bumpless interconnects with low-aspect-ratio TSVs and ultra-thinning are discussed. Bumpless (no bump) interconnects between wafers are a second-generation alternative to the use of microbumps for Wafer-on-Wafer (WOW) technology. Ultra-thinning of wafers down to 4 µm provides the advantage of a small form factor, not only in terms of the total volume of 3D ICs, but also the aspect ratio of ThroughSilicon-Vias (TSVs).… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1

Citation Types

0
7
0

Year Published

2016
2016
2024
2024

Publication Types

Select...
7
1

Relationship

1
7

Authors

Journals

citations
Cited by 22 publications
(7 citation statements)
references
References 18 publications
0
7
0
Order By: Relevance
“…As for the I/O power consumption, the first target of the bumpless HBM was one-thirtieth that of the current HBM2 [28], as shown in Figure 38. Figure 39 compares HBM2 with bumps [31][32][33] and bumpless HBM with respect to their data bandwidth and I/O buffer power, according as the number of I/O's [28,29,34,92,93]. Figure 39 compares HBM2 with bumps [31][32][33] and bumpless HBM with respect to their data bandwidth and I/O buffer power, according as the number of I/O's [28,29,34,92,93].…”
Section: Competitive Bbcube Dram Structurementioning
confidence: 99%
See 2 more Smart Citations
“…As for the I/O power consumption, the first target of the bumpless HBM was one-thirtieth that of the current HBM2 [28], as shown in Figure 38. Figure 39 compares HBM2 with bumps [31][32][33] and bumpless HBM with respect to their data bandwidth and I/O buffer power, according as the number of I/O's [28,29,34,92,93]. Figure 39 compares HBM2 with bumps [31][32][33] and bumpless HBM with respect to their data bandwidth and I/O buffer power, according as the number of I/O's [28,29,34,92,93].…”
Section: Competitive Bbcube Dram Structurementioning
confidence: 99%
“…Figure 39 compares HBM2 with bumps [31][32][33] and bumpless HBM with respect to their data bandwidth and I/O buffer power, according as the number of I/O's [28,29,34,92,93]. Figure 39 compares HBM2 with bumps [31][32][33] and bumpless HBM with respect to their data bandwidth and I/O buffer power, according as the number of I/O's [28,29,34,92,93]. Bumpless HBM can achieve an ultra-high data bandwidth by increasing the I/O number to 1 K, 10 K and 100 K, and can lower the I/O buffer power to 1/2 or 1/4 by reducing the I/O pin frequency with a four-phase shielded I/O scheme, as illustrated in Figure 40.…”
Section: Competitive Bbcube Dram Structurementioning
confidence: 99%
See 1 more Smart Citation
“…In Si-LSI (LSI) and 3D-LSI, significant miniaturization of the Cu interconnect has continued to progress, [1][2][3][4][5][6] where reduction of the interconnect resistance and improvement of the electromigration (EM) resistance are strongly required. 7) Under these circumstances, the realization of Cu(111) orientation, which is the most resistant to EM, is strongly desired.…”
mentioning
confidence: 99%
“…Copper films have low resistivity and excellent electromigration (EM) resistance, and they are the most commonly used interconnect material for silicon-large scale integration metallization and/or three-dimensional LSI such as throughsilicon-via technology. [1][2][3][4][5][6] However, because Cu films diffuse easily into SiO 2 and/or Si at low temperatures of less than 200 °C, 7,8) a diffusion barrier is required for reliable Cu interconnects. [9][10][11] Simultaneously, to further improve the EM resistance, an underlying material that can control the orientation and grain sizes of the Cu film is required.…”
mentioning
confidence: 99%