Data processing: measuring – calibrating – or testing – Measurement system in a specific environment – Biological or biochemical
Reexamination Certificate
2000-01-20
2002-08-06
Marschel, Ardin H. (Department: 1631)
Data processing: measuring, calibrating, or testing
Measurement system in a specific environment
Biological or biochemical
C702S020000, C702S022000, C435S006120, C435S007100, C703S013000, C703S014000, C703S015000, C703S016000, C708S001000, C708S100000, C710S100000, C712S001000, C713S001000, C716S030000
Reexamination Certificate
active
06430511
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to computer architecture, and, more specifically to a computer that uses molecules as functional units (e.g. logic gates and memory cells).
BACKGROUND OF THE INVENTION
The investment by society in computer technology has been astonishing in both its rate of increase and its extent. In less than three decades, the personal computer, for example, has gone from experimental prototype to being an essential tool of business. Demands for computers with greater and greater capabilities to perform more and more tasks continue unabated. Heretofore, better computers have resulted from increased miniaturization, among other improvements.
Presently, however, we do not have a viable technology for our near future computer needs. Although Moore's law (an accurate empirical law at this time) predicts the doubling of computer power every 18 months, this trend cannot continue. Digital computers are presently based on silicon technology. More precisely, very large scale integration (VLSI) is a lithographic technology, and although silicon is indeed quite important, Moore's Law is essentially lithography driven. The law of diminishing returns will eventually conquer Moore's Law, perhaps by 2005, when the cost of an integrated-circuit fabrication factory will become exorbitant and spell the demise of the growth of VSLI-based computer systems.
Other than standard logic and memory tasks, a technique one might choose for an ideal post-VLSI computer system would be to utilize a non-lithographic approach to construction, e.g., a directed self-assembly, so there would either be a vast redundancy or the system should have fault tolerance through dynamic fault reduction. Another property for such a computer would be an ability to be self-reconfigurable, that is, it should be able to dynamically reconfigure interconnects in response to inputs, or else it will also fall victim to the incessant demands for interconnects. Finally, this post-VLSI computer would be innovative: able to reconfigure architectures not previously experienced.
There remains a need, therefore, for a new computer technology that provides the advantages of self-assembling construction and greater processing capability that can be reconfigured readily.
SUMMARY OF THE INVENTION
The present invention is a device having a multiplicity of input pins and a multiplicity of output pins on its external surface. The interior is composed of a self-assembled array of specifically selected and adapted molecules, called “moleware”, bridging the inputs and outputs. Initially, this present computer is in a blank state; that is, nothing is known about the electrical-signal-transferring relationship between the input and output pins. A voltage is applied to each of a series of bundles of inputs. The outputs are searched, also in bundles, to determine what outputs have signals running through them. Using a computer to do the searching, sets of bundles of inputs and outputs will eventually be identified that can be used to constitute a truth or partial truth table for the computer. Electrical or magnetic fields can also be applied across the container to increase configuration possibilities.
The present invention, in a preferred embodiment, may take the form, for example, of a one cubic centimeter box that contains 1000 metallic inputs (m), 1000 outputs (n), and 100 different learning inputs (1). The box will contain an intelligently self-assembled array of 10
20
pre-designed active and passive components (“moleware”) including molecular alligator clip-bearing 2-, 3-, and molecular 4-terminal wires, molecular resonant tunneling diodes, molecular switches, molecular controllers that can be controlled via external electrical or magnetic fields, massive interconnect stations based on single nanometer-sized particles, dynamic or static random access memory (DRAM or SRAM) components composed of molecular controller
anoparticle or fullerene hybrids. Moreover, the present molecular computer, once formed, can be further modified by new interconnect routes via electrochemically induced cross linking of the nanoparticles or by “burning out” components through large induced fields, analogous to the operation of a field programmable gate array (FPGA). Arrays of these molecular computers can be coupled using standard interconnect methods to form massive molecular computer arrays.
The present molecular computer has a number of advantages:
It can be teachable. The system can be trained by forcing the correct output for a given input by varying the operational inputs until consistency is achieved. The output as a function of input can be autoverified.
It can be reconfigurable. Either by successive retraining or by burning out specific functions, the system can be changed.
It can possess logic. The truth table for the molecular computer would not be known a priori, but would be determinable once it has been prepared or on the fly as self-assembly takes place.
It can possess memory. Semiconductor nanoparticles such as CdS or CdSe coated with molecular control elements, or C
60
surrounded by controller elements, or even nanometer-scale metallic particles sufficiently small so as to exhibit sizeable Coulomb blockade, will be preformed and then permitted or induced to assemble themselves as part of the network of the present molecular computer. In the situation where an electromagnetic field is applied, the controllers will open thereby permitting the semiconductive or metallic particle to store a charge, hence acting as a DRAM or SRAM component.
It will have an intrinsic extremely high fault tolerance because of the ability to use their terminals (input, outputs, and learning inputs) as rewiring control leads. Therefore, this embodiment takes advantage of the immense number of functions that can be make with their terminals. For example, for binary operation, 1000 inputs yield 2
1000
posible input combinations. Therefore with 1000 possible outputs yield a total of 10
32256000 . . . 0000
where the number of digits in the exponent is 304.
The system has dynamic fault reduction capabilities. Unlike VLSI, where the interconnect structure is rigid, the present molecular computer can be reprogrammed to eliminate undesired and inoperative fragments since the interconnect topology can be rewired at any time by applying sufficiently strong fields to regions of the molecular computer where reprogramming is desired.
The system lends itself to expansion. Arrays of molecular computers according to the preferred embodiment described above can be formed using standard interconnect methods to form massive computer arrays.
It is also possible to program the present computer in other than a binary digital mode, such as for example, an analog or multilevel (semi-analog) system. Perhaps programming becomes more formidable; however, data storage increases significantly and makes the programming challenges acceptable for certain applications.
The design of a molecular computer according to a preferred embodiment of the present invention takes into consideration that a very large percent of the fabricated moleware components may be defective. Hence, the testing will be more extensive and done in multiple stages in the fabrication and assembly. Molecular computers made according to the present it process will be tested individually and in place. Testing is performed preferably with a high-speed computer or supercomputer initially or, eventually, with the present molecular computer, due to the computational demands for rapidly sorting out the relationship between input and output pins.
An important feature of the present invention is that, unlike other approaches to making computers from molecules, specific placement of nanometer-sized molecular elements is not needed. Moleware is simply added in bulk in a dilute solution and given a chance to form links or “bridges” between inputs and outputs. The excess solvent is removed leaving the moleware components after self-assembly. Additionally, the moleware can be inserte
Allara David L
Reed Mark A
Seminario Jorge M
Tour James M
Weiss Paul S
Mann Michael A.
Marschel Ardin H.
Nexsen Pruet Jacobs & Pollard LLC
University of South Carolina
LandOfFree
Molecular computer does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Molecular computer, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Molecular computer will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2934395