At the beginning of the 21st century, 1 Gb DRAMs will be in practical use, and sufficient in terms of memory capacity for most memory applications systems. The key technologies for multimedia systems include data compression, communication, storage, and human interfaces. Image data processing, ATM switch, and microprocessor in multimedia applications require the high data transfer rate from several 100 Mbits/s to Tbits/s. Storage systems, on the other hand, require the reduction of the price per bit to less than 10 cents/Mbytes. Application specific design approaches towards a system-on-chip are strongly needed for ULSI memories in the multimedia era.

Ever increasing demand for higher bandwidth memories, which is fueled by multimedia and 3D graphics, seems to be somewhat satisfied with various emerging memory solutions. This paper gives a review of these emerging DRAM architectures and a performance comparison based on a condition to let the readers have some perspectives of the future and optimized graphics systems.

As microprocessor units have become faster, DRAMs have also been required to become faster. One of the fast DRAMs is the synchronous DRAM, which transfers data at a high rate. We have developed a 100-MHz Synchronous DRAM using pipeline architecture and new high speed I/O lines method. This paper describes some features of the DRAM and its new pipeline architecture.

NAND-structured trench capacitor cell technologies for 256 Mb DRAM and beyond have been developed. The NAND-structured cell has four memory cells connected in series. The cell size can be reduced to 56% of the conventional cell. A substrate plate trench capacitor cell was adapted to this layout. The NAND-structured trench capacitor cell can achieve sufficient storage capacitance within the restricted capacitor area. A sufficient capacitance of 40 fF was achieved when the size and depth of trench were 0.5 µm and 5.0 µm, respectively. The most important point for realizing the NAND-structured trench capacitor cell is how to reduce the leakage current from the storage node. There are two main sources; one is the leakage current to the neighboring cells, the other is the leakage current to Pwell. These leakage currents have been investigated. An experimental 256 Mb DRAM with the NAND-structured cell was fabricated using the 0.4 µm design rule. The chip size is 464 mm², which is 68% of a conventional DRAM of the same design rule. This is the result of the reduction of the memory cell area by the NAND-structured cell and the introduction of the open-bit-line arrangement.

Low-power circuit techniques for size-configurable SRAM macrocells with wide range of operating frequency are presented. Synchronous specification is employed to drastically reduce the power dissipation for low-frequency applications. Dynamic circuits applied to bitliness and sense circuits contribute to the reduction of power dissipation. To enhance the high-end limitation of operating frequency, a latch-type fast sense circuit and an accurate activation-timing control technique for size-configurable memory macrocells are proposed, and a special CMOS-level input buffer is devised to enable the minimum cycle time of fast synchronous memory macrocells to be evaluated with conventional LSI-test systems. A memory macrocell using these techniques was fabricated with 0.5-µm CMOS technology. Its power consumption strongly depends on the operating frequency, and at 3-MHz suitable for codeless telephone applications is less than 5% that of an asynchronous SRAM designed with full-static CMOS circuits. Its maximum operating frequency at 3.3-V in 100-MHz.

PLL timing design techniques introduced here feature (1) a word-line resetting-equalization scheme for reducting the access time and the memory cell current, (2) a cyclic input buffer power-cutting scheme to reduce the excess power in low-frequency operation, and (3) a super-pipelined parallel test scheme which allows the evaluation of high-speed LSIs by low-speed LSI testers to reduce the test costs. These techniques successfully contribute to the development of a 7 ns GTL I/O 16 Mb BiCMOS SRAM LSI.

The stability of a high-resistivity load SRAM cell using thin-film SOI MOSFET's was investigated as compared with bulk-Si MOSFET's. In SOI MOSFET's back-gate-bias effect was suppressed by indirect application of back-gate-bias to the channel region through the thick buried oxide. The V_t shifts were reduced to be 10% and 14% of that in bulk-Si MOSFET's in partially and fully depleted devices, respectively. The reduction of back-gate-bias effect provides improvement of "high" output voltage and gain in the enhancement-enhancement (EE) inverter in a high-resistivity load SRAM cell, thereby offering improved cell stability. It was demonstrated by using partially depleted SOI SRAM cells that non-destructive reading was obtained even at a low drain voltage of 1.4 V without gate-potential boost, which was much smaller than the operation limit in a bulk Si SRAM with the same patterns and dimensions used as a reference. This implies that SOI devices can also offer low-voltage operation even in TFT-load cells used in up-to-date high-density SRAM's. These results suggest that thin-film SOI MOSFET's have a superior potential of low-voltage operation expected for further scaled devices and/or for portable systems in a forthcoming multimedia era.

An experimental 3 V-only 4 Mb NAND Flash memory with 65 ns access time has been developed using a new charge pump circuit and novel circuit techniques such as folded bit-line architecture. By adopting a new program verify technique, programming time is reduced to 11 µs/Byte.

A 3.3 V single power-supply 64 Mb flash memory with a DBL programming scheme has been developed and fabricated with 0.4 µm CMOS technology. 50 ns access time and 256 b erase/programming unit-capacity have been achieved by using hierarchical word- and bit-line structures and DBL programming scheme. Furthermore in order to lower operating voltage the HiCR cell is used. The chip size is 19.3 mm13.3 mm.

Programming and program-verification methods for low-voltage flash memories using the Fowler-Nordheim tunneling mechanism for both programming and erasure are described. In these memories, a great many memory cells on a selected word line, such as 512-bytes worth of cells, are programmed at the same time for high-speed programming. The bit-by-bit programming/verification method can precisely control threshold-voltage deviation of programmed memory cells on the selected word line for low voltage operation. By using an internal program-end detection circuit, the completion of program mode can be checked for in one clock cycle, without reading out 512-bytes of data from the memory chip to the external controller. Moreover, the variable pulse-width programming method reduces the total number of verifications.

We developed an on-chip memory tester macro using a microprogram ROM BIST circuit. Only slight modification of address buffers, data bus I/O circuits and control clock generators of the memory core circuits was required to implement this BIST macro. We fabricated a 1 Mb DRAM with the BIST, and experimental results showed that the measured shmoo plot of VCC versus the cycle time by the BIST closely agreed with that of a memory tester. Disagreement was caused by test address signal set-up time delay and V_OH/V_OL differences in both test conditions. The BIST macro will be especially useful for design-for-testability of embedded memories.

This paper describes the new α-particle induced soft error mechanism, the Minority Carrier Outflow (MCO) effect, which may seriously affect the reliability of the scaled DRAMs with three dimensional capacitors. The MCO chargge increases as the device size miniaturizes because of the three dimensional capacitor effect as below. As the device scales down, the storage node volume decreases which results in the higher minority carrier density in the storage node and larger outflow charge. Also as the device plan view miniaturizes, the stack capacitor height or trench depth does not scales down or even increases to keep the storage node capacitance, therefore the initially generated minority carrier becomes larger. A simple analytical MCO model is introduced to evaluate the MCO effect quantitatively. The model agrees well with the three dimensional device simulation. The MCO model predicts that the life time of the minority carrier in the storage node strongly affects the MCO charge, however, even when the life time is as small as the order of 100 ps, the MCO effect can be the major soft error mechanism.

There is general trend toward larger chip size and tighter layout due to customer requests of loading more and more functions on single chip. This trend makes yield difficult to be maintained high enough, since larger amount of defects are distributed on such large and tight-ruled chips. To overcome such a situation, RADLYS (RAnDom Logic Yield Simulator) and DD-TEG (Defect Density TEG) have been developed. DD-TEG extracts defect size distribution and its amount automatically, while RADLYS simulates defects on any layout and outputs yield based on the extracted defect size distribution. Critical layout from yield point of view can be found in this procedure. DD-TEG and RADLYS are used as a set of parameter extraction and simulation of the SPICE. In this paper, we introduce these tools and showed two application results. The predicted yield showed a good agreement with the actual yield in the first application (Optical Device A). Critical layout at the Local I/O portion was found in the second application (Random Logic portion of Memory Device B) and the layout was changed based on the RADLYS results.

This paper proposes a fully self-timing data-bus (FSD) architecture which includes a dual data-bus driven by the read-out data itself and a complementary output differential (COD) amplifier. The proposed COD amplifier achieves a high voltage gain and a high speed data transfer with low power consumption. The read-out data is transmitted from the COD amplifier to the output terminal without the timing control caused by the fluctuation of the device parameters. Therefore the proposed FSD architecture eliminates the timing delay and achieves a timing-free data transfer even in DRAMs with a small signal level at the sense amplifier and the data line. Applying this architecture to a 64-Mb DRAM, a fast column address access time of 16 ns and a RAS access time of 32 ns have been achieved.

This paper describes an 8:1 multiplexer and a 1:8 demultiplexer for fiber optic transmission systems. These chips incorporate new architectures having a smaller hardware and enabling the use of a lower supply voltage. The multiplexer and the demultiplexer are fabricated using 0.8 µm silicon-bipolar process with a double polysilicon self-aligned structure. The multiplexer operates at a bit rate of up to 3.0 Gb/s, while the demultiplexer operates at a bit rate of up to 4.1 Gb/s. The multiplexer consumes 272 mW and the demultiplexer consumes 388 mW under the power supplies of VEE=-4.0 V and VTT=-2.0 V. These values are the smallest so far above 2.5 Gb/s which is the standard of the Level-16 of the synchronous transfer mode (STM-16).

This paper describes analytical results of high-T_c superconducting asymmetric coplanar strip lines using the frequency-dependent finite-difference time-domain method. The propagation constants of the YBa₂Cu₃O_7-x asymmetric coplanar strip line fabricated on the LiNbO₃ substrate are reported. The effect of the SiO₂ buffer layer is also investigated.

The subject is the horizontal coil's temperature rise in DY for high frequency by being unavoidable for the tendency of more information on display monitor equipments. Writers made the temperature-balance model from a point of view that this temperature rise is coming from the heat rise and the conductivity, and we expressed the temperature rise of DY by using amount of the heat rise and conductivity characteristics of each element. Also, we indicated the method to decide about the selection of the wire size of coils, the section area and deflection sensitivity, with regard to reducing the temperature rise. We confirmed the effect of the temperature rise reduction by about 9 on products, under the condition of 64 kHz horizontal frequency.

A fiber optic temperature sensor using a conventional graded index multimode optical fiber is proposed. The multimode fiber is excited by two selected modes using a computer-generated holographic filter. A clear periodic signal created by interference between two modes is observed in the experiment.