Active solid-state devices (e.g. – transistors – solid-state diode – Field effect device – Having insulated electrode
Reexamination Certificate
2002-06-27
2004-05-18
Whitehead, Jr., Carl (Department: 2813)
Active solid-state devices (e.g., transistors, solid-state diode
Field effect device
Having insulated electrode
C257S308000, C257S309000
Reexamination Certificate
active
06737699
ABSTRACT:
BACKGROUND
1. Field
An integrated circuit decoupling capacitor and more specifically, an on-chip decoupling capacitor.
2. Background
The operation of low power, high speed integrated circuits can be affected by the electrical noise generated by the continuous switching of the transistors located in the circuit. It is well known that the inductive noise of an integrated circuit can be reduced by connecting decoupling capacitors to the circuit. Decoupling capacitors placed on power-consuming circuits are able to smooth out voltage variations with the stored charge on the decoupling capacitor. The stored charge is used as a local power supply to device inputs during signal switching stages, allowing the decoupling capacitor to mitigate the effects of voltage noise induced into the system by parasitic inductance.
Typically, a decoupling capacitor is placed in the same package as the chip. Unfortunately, this arrangement is costly to manufacturer, and the long lead lines from the power-consuming circuit to the capacitor electrodes contributes a substantial inductance. Such off-chip decoupling capacitors, however, are not sufficient for very high speed microprocessor applications. The voltage drop across an inductor may be described by the relationship L di/dt, where L is inductance and di/dt represents the change in current through a circuit over a period of time. Implicit in the di/dt is a frequency component (omega), so as frequency goes up, inductance becomes more and more a factor for power distribution. The frequency dependent L di/dt voltage drop makes the off-chip capacitors generally unusable with gigahertz switching circuits unlike low frequencies for which voltage drops are dominated by resistance.
Some efforts have been made to integrate decoupling capacitors as part of the gate dielectric processing step. A portion of the active silicon area is used to deposit the gate dielectric for use as a decoupling capacitor. An advantage to this is that there are no additional processing steps involved with it. Disadvantages include the decoupling capacitor takes up high-value real estate on the chip as the capacitors compete for valuable chip area that could be used for building additional circuits. Also the capacitor made with a gate oxide designed for very high transistor performance generally has a great deal of leakage. These gate decoupling capacitors also generally have associated parasitic resistance from the relatively low silicon conductivity that entails an additional resistance-capacitance (RC) time constant for charge extraction. Although it is possible to integrate gate capacitors within the chip's circuit elements, due to the limited area in which to build these capacitors, the overall capacitive decoupling that they provide is also limited.
Another approach to decoupling capacitor fabrication is a decoupling capacitor that may be fabricated between metal layers in an integrated circuit. In a representative chip architecture having seven metal layers, for example, the capacitor may be fabricated between metal six (n−1) layer and metal seven (n) layer. Advantages to this embodiment include that there is little no additional real estate (area) on the chip consumed for fabrication of the decoupling capacitor, and the decoupling capacitor directly bridges the on-chip power grid being representatively up to 10 microns, according to current technologies, from the integrated circuit element it is supporting with a very low inductance per unit length due to the tight spacing of the power and ground lines. One problem with locating the decoupling capacitor between metal layers is that, in the example given, approximately 30 percent of the metal six layer is devoted to power supply. This limits the total amount of decoupling capacitance that can be provided on-chip per layer, because total capacitance available is generally a function of the total area dedicated to capacitor plates.
REFERENCES:
patent: 5021842 (1991-06-01), Koyanagi
patent: 5142639 (1992-08-01), Kohyama et al.
patent: 5153813 (1992-10-01), Oehrlein et al.
patent: 5240871 (1993-08-01), Doan et al.
patent: 5455204 (1995-10-01), Dobuzinsky et al.
patent: 5550080 (1996-08-01), Kim
patent: 5696395 (1997-12-01), Tseng
patent: 5701264 (1997-12-01), Shrivastava et al.
patent: 5710233 (1998-01-01), Meckel et al.
patent: 5789267 (1998-08-01), Hsia et al.
patent: 5835337 (1998-11-01), Watanabe et al.
patent: 5843822 (1998-12-01), Hsia et al.
patent: 5891772 (1999-04-01), Hsu
patent: 5907774 (1999-05-01), Wise
patent: 6022772 (2000-02-01), Watanabe et al.
patent: 6027968 (2000-02-01), Nguyen et al.
patent: 6064085 (2000-05-01), Wu
patent: 6268995 (2001-07-01), Beuerman et al.
patent: 6294420 (2001-09-01), Tsu et al.
patent: 6344392 (2002-02-01), Liaw
patent: 6346455 (2002-02-01), Thakur et al.
patent: 6373092 (2002-04-01), Okumoto
patent: 6407423 (2002-06-01), Okumoto
patent: 6555433 (2003-04-01), Liaw
Block Bruce A.
List Richard Scott
Blakely , Sokoloff, Taylor & Zafman LLP
Intel Corporation
Jr. Carl Whitehead
Schillinger Laura M
LandOfFree
Enhanced on-chip decoupling capacitors and method of making... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Enhanced on-chip decoupling capacitors and method of making..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Enhanced on-chip decoupling capacitors and method of making... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3267148