Method for forming complex electronic circuits by interconnecting groups of printed devices

Abstract

A programmable circuit includes an array of printed groups of microscopic transistors or diodes. The devices are pre-formed and printed as an ink and cured. The devices in each group are connected in parallel so that each group acts as a single device. In one embodiment, about 10 devices are contained in each group so the redundancy makes each group very reliable. Each group has at least one electrical lead that terminates in a patch area on the substrate. An interconnection conductor pattern interconnects at least some of the leads of the groups in the patch area to create logic circuits for a customized application of the generic circuit. The groups may also be interconnected to be logic gates, and the gate leads terminate in the patch area. The interconnection conductor pattern then interconnects the gates for form complex logic circuits.

Claims

1. A method for forming a circuit comprising: providing an ink containing a plurality of pre-formed, semiconductor electrical devices mixed in a solvent, wherein the devices formed on a wafer and then singulated into individual pre-formed devices, the devices comprising individual transistors; printing the ink, on a substrate, to form a plurality of separate groups of the pre-formed, semiconductor electrical devices, each group containing a plurality of substantially identical electrical devices forming a random two-dimensional arrangement of the electrical devices on the substrate; forming at least one conductor layer to connect the electrical devices in each group in parallel; interconnecting at least some of the groups, using an interconnection pattern, to achieve a specified electrical function; forming at least one input terminal connected to the interconnection pattern for receiving input signals for transformation by the specified electrical function to generate electrical output signals; and forming at least one output terminal connected to the interconnection pattern for receiving the electrical output signals.

2. The method of claim 1 wherein the step of interconnecting at least some of the groups comprises interconnecting at least some of the groups to create logic gates.

3. The method of claim 1 wherein the step of interconnecting at least some of the groups comprises interconnecting the groups in a designated patch area.

4. The method of claim 1 wherein the step of interconnecting at least some of the groups comprises interconnecting the groups using a first interconnection pattern to form logic gates and then interconnecting the logic gates in a designated patch area using a second interconnection pattern.

5. The method of claim 1 wherein the circuit is formed using a roll-to-roll process to form flexible circuits.

6. The method of claim 1 wherein the electrical devices have largest dimensions between 10-200 microns.

7. The method of claim 1 wherein the circuit forms a programmable gate array.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 is a cross-section of a loose monolayer of printed, microscopic vertical LEDs that may be formed using the assignee's prior art process.

(2) FIG. 2 is a top down view of the structure of FIG. 1, where FIG. 1 is taken across a horizontally bisected FIG. 2.

(3) FIG. 3 is a perspective view of a single 3-terminal transistor that has been singulated from a wafer, in accordance with one embodiment of the invention. The transistors are mixed into a solution to form an ink for printing on a substrate.

(4) FIG. 4 is a cross-section of a small portion of a printed layer of the transistors of FIG. 3 connected in parallel using three planes of conductor layers. About 10 of the parallel-connected transistors may be printed in each separate group, and an array of the groups is printed over the substrate.

(5) FIG. 5 illustrates how the transistor of FIG. 3 may be an npn bipolar transistor.

(6) FIG. 6 illustrates how the transistor of FIG. 3 may be a p-channel MOSFET.

(7) FIGS. 7 and 8 illustrate how some transistors may be incorrectly interconnected by improper orientations of the transistors when printing, where the interconnections do not adversely affect the functions of the properly oriented transistors.

(8) FIG. 9 is a chart identifying preferred functions of the top, bottom, and intermediate electrodes for MOSFETs and bipolar transistors.

(9) FIG. 10 illustrates interconnecting groups of transistors to form a logic circuit.

(10) FIG. 11 illustrates how the leads for the various groups of devices may be brought to a patch area of the substrate for interconnecting the groups. In another embodiment, the leads of logic gates (e.g., NAND gates) made from the groups may be brought to the patch area.

(11) FIG. 12 is a top down view of conductors being patterned using a hydrophobic mask.

(12) FIG. 13 is a cross-sectional view of one of the conductor lines formed using the mask of FIG. 12.

(13) FIG. 14 illustrates how the hydrophobic mask may be used for form conductor lines to interconnect groups of devices to form logic circuits.

(14) FIG. 15 illustrates how resistor values may be determined by the location of a conductor contacting printed resistive material.

(15) FIG. 16 illustrates how the devices may be printed on both sides of a substrate and interconnected by a through via.

(16) FIG. 17 illustrates how circuits may be printed on relatively small substrates and tested, then the small substrates are attached to a larger base substrate during a customizing step.

(17) FIG. 18 illustrates how electrodes on the small substrate of FIG. 17 may be bonded to electrodes on the base substrate.

(18) FIG. 19 illustrates a roll-to-roll process that may be used to form the circuits.

(19) Elements that are similar or identical in the various figures are labeled with the same numeral.

DETAILED DESCRIPTION

(20) The printed programmable circuits of the present invention may use any combination of passive devices (e.g., capacitors, resistors), 2-terminal inorganic semiconductor devices (e.g., diodes), and 3-terminal inorganic semiconductor devices (e.g., transistors). The most complex device to print and electrically connect to is a 3-terminal device. In some cases, a 3-terminal device, such a bipolar transistor, may be used as a diode by only using two of the terminals or connecting two terminals to the same conductor.

(21) The 3-terminal devices used in embodiments of the present invention may be less than the diameter of a human hair, rendering them essentially invisible to the naked eye when the devices are sparsely spread across a substrate. The sizes of the devices may range from about 10-200 microns across. The number of micro-devices per unit area may be freely adjusted when applying the micro-devices to the substrate. The devices may be printed as an ink using flexography, screen printing, or other forms of printing. Conventional designs for 3-terminal devices may be easily adapted for forming the micro-devices of the present invention. The precision of photolithography is well within the precision needed to form the micro-devices. Since many of the micro-devices will be operating in parallel, the efficiency of each micro-device is not critical.

(22) FIG. 3 is a perspective view of a 3-terminal device 40 that can be suspended in a solvent and printed as an ink on a substrate. The device 40 may be a bipolar transistor, a MOSFET, a JFET, a tri-MOS device, or any other 3-terminal device, generally including two current carrying terminals and a control terminal. The device 40 can be a lateral or vertical transistor, since the positions of the three electrodes do not dictate the locations of the semiconductor layers/regions or gates inside the device 40. The electrodes may contact any location in the device 40 using vias.

(23) The devices 40 are completely formed on a semiconductor wafer, including the electrode metallizations, by using one or more carrier wafers during the processing to gain access to both surfaces for metallization. Although the growth wafer may be silicon, the carrier wafer may be any material. The silicon wafer is affixed to the carrier wafer using an adhesive or other suitable material. The shape of each device 40 is defined by masking and etching. Various layers or regions may be doped using masked implantation or by doping the layers while being epitaxially grown. After the devices are formed on the wafer, trenches are photolithographically defined and etched in the front surface of the wafer around each device 40 down to the adhesive layer. A preferred shape of each device 40 is hexagonal. The trench etch exposes the underlying wafer bonding adhesive. The adhesive is then dissolved in a solution to release the devices 40 from the carrier wafer. Singulation may instead be performed by thinning the back surface of the carrier wafer until the devices 40 are singulated. The microscopic devices 40 are then uniformly infused in a solvent, including a viscosity-modifying polymer resin, to form an ink for printing, such as screen printing or flexographic printing.

(24) A similar technique may be used to form 2-terminal devices, such as a vertical diode, with one electrode on top and another electrode on the bottom. The diode may have a shape similar to that shown in FIG. 3 but without the middle electrode.

(25) Details regarding shaping vertical LEDs (2-terminal devices) in a wafer and then singulating the LEDs for printing as an ink are described in US application publication US 2012/0164796, entitled, Method of Manufacturing a Printable Composition of Liquid or Gel Suspension of Diodes, assigned to the present assignee and incorporated herein by reference. One skilled in the art may adapt such processes for forming the 3-terminal device 40 and non-LED diodes.

(26) The device 40 has two sections: a lower section 42 (or base portion) and an upper section 44. The upper section 44 is made relatively tall and narrow so that the devices 40 are rotated in the solvent, by fluid pressure, as they settle on the substrate surface. The devices 40 rotate to an orientation of least resistance. Over 90% like orientation has been achieved, although satisfactory performance may be achieved with over 75% of the devices 40 being in the same orientation.

(27) The lower section 42 should be shaped so that the device 40 sits flat on the substrate after the ink is cured. FIG. 4 illustrates three printed devices 40, where only two are printed with the correct orientation.

(28) The device 40 includes a metal top electrode 46, a metal intermediate electrode 48, and a metal bottom electrode (not shown in FIG. 3). The shape of the intermediate electrode 48 provides a large side surface area for good electrical contact with an intermediate conductor layer.

(29) The intermediate electrode 48 should be offset with respect to the middle of the device 40 so that an improper orientation of the device 40 after printing results in the intermediate electrode 48 not electrically contacting the intermediate conductor layer. In the example, the intermediate electrode 48 is below the middle of the device 40 (i.e., H2< H1).

(30) In FIG. 4, a starting substrate 50 is provided. The substrate 50 is preferably thin and flexible for light weight, low cost, good heat conduction to air or a heat sink, and ease of processing. The substrate 50 may be a suitable polymer, such as polycarbonate, PMMA, or PET, and may be flexible so as to be dispensed from a roll. The substrate 50 can be any size suitable for the ultimate product. The substrate 50 may be a conventional flex-circuit substrate, where metal (e.g., copper) traces have been already formed on the substrate 50 by conventional means prior to the below processing steps.

(31) If the substrate 50 does not already have metal traces formed on it as a flex-circuit, a conductor layer 52 (e.g., silver, aluminum, copper) is deposited on the substrate 50 such as by printing. Conductive vias 54 through the substrate 50 may be used to couple the conductor layer 52 to a metal layer 56 formed on the bottom surface of the substrate 50. In the various examples, the conductor layer 52 is printed as an array of circular spots on the substrate 50 (see FIG. 11). The spots are electrically isolated from one another to allow groups of the devices 40 to be interconnected in any manner to form logic circuits. Instead of circular spots, the conductor layer 52 may be printed in square spots or other shaped spots.

(32) The devices 40 are then printed on the conductor layer 52 such as by flexography or by screen printing with a suitable mesh to allow the devices 40 to pass through and control the thickness of the layer. Because of the comparatively low concentration, the devices 40 will be printed as a loose monolayer and be fairly uniformly distributed over the conductor layer 52. The printed locations of the devices 40 align with the locations of the printed spots of the conductor layer 52.

(33) The solvent is then evaporated by heat using, for example, an infrared oven. After curing, the devices 40 remain attached to the underlying conductor layer 52 with a small amount of residual resin that was dissolved in the ink as a viscosity modifier. The adhesive properties of the resin and the decrease in volume of resin underneath the devices 40 during curing press the bottom electrode 58 against the underlying conductor layer 52, making ohmic contact with it.

(34) A thin dielectric layer 60 is then printed to cover the conductor layer 52 and further secure the devices 40 in position. The dielectric layer 60 is designed to self-planarize during curing, by surface tension, so as to pull off of or de-wet the top electrode 46 and the intermediate electrode 48. Therefore, etching the dielectric layer 60 is not required. If the dielectric layer 60 covers the electrodes 46/48, then a blanket etch may be used to expose the electrodes 46/48.

(35) An intermediate conductor layer 62, aligned with the spots of the conductive layer 52, is then printed over the dielectric layer 60 to electrically contact the intermediate electrode 48 and is cured in an oven appropriate for the type of conductor being used. The various conductor layers may be metal (or contain metal) or be any other type of printable conductor layer.

(36) Another thin dielectric layer 64 is printed over the intermediate conductor layer 62 so as not to cover the top electrode 46.

(37) A top conductor layer 66, aligned with the spots of the intermediate conductor layer 62, is then printed over the dielectric layer 64 to electrically contact the top electrode 46 and is cured in an oven appropriate for the type of conductor being used.

(38) A thicker metal layer 68 may then be printed over the conductor layer 66 for improving electrical conductivity and/or heat conduction. The intermediate conductor layer 62 extends out from the edge of the spot to form a terminal of the group of the devices 40.

(39) FIG. 4 illustrates that the only steps needed to form the structure of FIG. 4 are printing and curing steps 67 and 69. The random pattern of the devices 40 may resemble the pattern of LED 16 in FIG. 2.

(40) FIG. 4 illustrates that the rightmost device 40A is oriented in the opposite direction. However, the intermediate electrode 48 is left floating, so the device 40A does not operate and has no effect on the resulting circuit.

(41) The printed devices 40 are connected in parallel by the conductor layers. Suitable operating voltages and control voltages are applied to the conductor layers to operate the devices 40. In the example of FIG. 4, the top electrode 46 is the control electrode for the devices 40 (e.g., for gates or bases). The remaining two electrodes are the current carrying electrodes (e.g., source/drain, emitter/collector). Since the intermediate electrode 48 of the improperly oriented device 40A is floating, the device 40 remains off and an open circuit.

(42) FIG. 5 illustrates how the device 40 may be an npn bipolar transistors 40B, where the intermediate electrode 48 is the base electrode. The intermediate electrode 48 may be connected to any of the other semiconductor layers in the device 40B using a via.

(43) FIG. 6 illustrates how the device 40 may be a p-channel MOSFET 40C, where the intermediate electrode 48 is the source electrode. The intermediate electrode 48 may be connected to any of the other layers in the device 40C using a via.

(44) If the devices 40 are to be connected as diodes, only conductor layers 62 and 52 or 66 and 62 may be used. Accordingly, the effective polarity of the diode may be selected by which two conductor layers are used to contact the diodes. Alternatively, two conducting layers may be connected remotely to form a diode.

(45) Any number of the devices 40 may be connected in parallel in a group for handling a wide range of currents. In one embodiment, about 10 devices 40 are located in each group. The groups of the devices 40 are printed as a 2-dimensional array of groups, such as by using a pattern on a roller in a flexography print process or by using a mask on a screen print mesh, and the various conductor layers may be similarly patterned so that the devices 40 in each group are connected in parallel, but each group is electrically isolated from one another. Therefore, each group forms a separate component. The groups may then be selectively interconnected using programming conductor traces on the substrate 50 to form more complex circuits, such as logic circuits. A metal flex-circuit pattern on the substrate 50 may be used to interconnect the groups of devices 40 for form the logic circuits. In one embodiment, since each group may be as small as a millimeter per side or a millimeter in diameter, a 2-dimensional array of such groups may exceed several thousand groups. Groups within a small area may be interconnected to form logic gates, and the terminals of the gates may be interconnected during programming to perform any logic function.

(46) FIG. 7 illustrates how the improper orientation of the device 40A in FIG. 4 does not adversely affect the operation of the properly oriented devices 40 in the group connected in parallel. The device 40/40A is assumed to be an npn bipolar transistor with a top electrode 46 for the base, a bottom electrode 58 for the emitter, and an intermediate electrode 48 for the collector. Since the device 40A is undesirably oriented upside down during printing (shown in FIG. 4), its base is shorted to the emitter of the device 40 and its emitter is shorted to the base of the device 40. When the base/emitter junction of the device 40 is forward biased to turn on the device 40, the device 40A remains off and has no effect on the operation of the device 40. Note that, by using an intermediate electrode 48 that is offset from the middle of the device 40 (as shown in FIGS. 3 and 4), the intermediate electrode 48 of the device 40A would be floating, making its effect even more insignificant.

(47) FIG. 8 is similar to FIG. 9 but the devices 40 and 40A are MOSFETs.

(48) FIG. 9 is a table showing possible connections for the top, bottom, and intermediate electrodes of the device 40 formed as a MOSFET or a bipolar transistor so that improper orientations do not adversely affect the functions of the properly oriented devices 40 connected in parallel.

(49) FIG. 10 illustrates two groups, 72 and 74, of printed npn bipolar transistors (e.g., devices 40), where the transistors in each group are connected in parallel so that each group acts as a single transistor. The print patterns for the devices 40 and conductor layers form the groups as circular spots, but any shaped spot may be used. The interconnections of the groups in FIG. 4 cause the circuit to be an AND gate. The conductive traces 75 connect to the various conductor layers in FIG. 4 for each group. The two transistors (i.e., groups 72 and 74) are connected in series between the supply voltage terminals 76 and 78, the bases of the transistors are connected to the input terminals 80 and 82, and the output terminal 84 is connected to the emitter of the transistor formed by group 74. The various terminals may be near the edges of the substrate 50 or proximate to the groups.

(50) Resistors r1 and r2 are shown connected between the input terminals 80/82 and the bases for current control. Due to the simplicity of resistors, the resistive material may be directly patterned on the substrate with a patterned roller using flexography or a mask on a screen mesh used to print the resistive material. Either the shape of the resistive material may determine the resistance or the position of the connector along its length may determine the resistance. A resistor may also be included on each device 40. Capacitors may also be formed by printing the layers of the capacitors.

(51) The substrate 50 may contain hundreds or thousands of such AND gates, or other gates, and the gates may be interconnected to form more complex functions. In such a case, the gates are equivalent to a programmable gate array. For a more flexible circuit, the groups may be initially unconnected, and the programming mask for the interconnections may determine the final circuit. Three-dimensional programming may be used to allow the crossing over of traces. Any combination of gates and other logic circuitry may be created. Some groups may include transistors and other groups may contain other devices, such as diodes. Analog circuits may also be formed by interconnecting the various groups.

(52) Due to the random but substantially uniform distribution of the devices 40 in the ink, each group of the same area will have approximately the same number of devices 40. Minor differences in the number of devices 40 in a group will not affect the performance of a logic circuit. In one embodiment, there may be about 10 identical devices in each group due to the low currents required. The cost of the devices 40 in a single group, representing a single transistor, is about 0.143 cents. So the resulting circuit board may be made relatively inexpensively.

(53) As shown in FIG. 11, to simplify the programming of the groups, which may be printed in an ordered 2-dimensional array, conductive traces 85, originating at the conductor layers (FIG. 4) for all the groups, may terminate at a patch area 86 on the substrate 50, where the product is now a programmable circuit board 87. These traces 85 may be part of the standard design of the circuit board 87, which is then later customized for a particular use. This enables the printing process for forming the traces 85 to be optimized for connection to the conductor layers in the groups and the programming process to be optimized for interconnecting the ends of the terminals 88. For example, the programming process may be performed at a time after the circuit board 87 has been fabricated, and the programming step may be performed by special equipment under computer control. Further, the pattern of the interconnections may be much more complex than the traces 85 that electrically connect the transistor terminals to the patch area 86.

(54) In the example of FIG. 11, the programming in the patch area 86 forms the AND gate of FIG. 10. For more complex circuits, the programming traces 90 may need to cross, and multiple layers may be formed to avoid shorting of the traces.

(55) In another embodiment, the groups of devices 40 may be initially interconnected proximate to the groups to form separate logic gates, such as AND, NAND, NOR gates, and the leads for each gate terminate in the patch area 86 for later programming to customize the substrate for a particular customer. Accordingly, the generic circuit forms a programmable gate array.

(56) A plurality of spaced patch areas may be provided on the circuit board 87 to simplify routing of the interconnections. In one embodiment, the terminals for all the input signals are provided on one level in a patch area, and the output terminals are provided on another level.

(57) If the programming of the interconnections is complex, directly printing the interconnections in an X-Y plane on the substrate 50 may be insufficient. Direct printing of conductors on the substrate is limiting, since a minimum spacing between conductors is about 30 microns to avoid cross-bridging, and thin conductors have a tendency to break up by surface tension.

(58) In situations where it is not desired for the conductor lines to be directly printed, a mask layer is first formed on the substrate followed by the deposition of the conductor ink over the mask layer as follows.

(59) One approach to pattern the interconnection traces, or to pattern any other traces on the circuit board 87, or to pattern the groups of devices 40, is to form hydrophobic masks. The masks may be deposited by printing (e.g., using a patterned roller or screen printing) or may be patterned by a photolithographic process (if printing cannot achieve the desired precision). One suitable masking substance is thoroughly cleaned diatomaceous earth particles infused in a solution as an ink. The ink is printed in a pattern that is negative to the desired wiring/device pattern. After curing, the resultant film is activated via a fluorination process, yielding a super-hydrophobic surface (i.e., it will not wet by the conductor ink or the device ink). The area of the substrate that is exposed by the film will be either mildly hydrophilic or super-hydrophilic (i.e., it will wet by the conductor ink or the device ink).

(60) For forming the traces, a hydrophilic conductive ink is prepared and deposited over the hydrophobic mask. The exposed substrate areas will be covered by the ink, and the conductive ink that has been deposited on the hydrophobic mask surface will pile up in the exposed areas. This yields a greater cross-sectional area of the conductive ink (for good conductivity and mechanical strength) and prevents cross-bridges.

(61) FIG. 12 is a top down view of the hydrophobic mask 94 defining areas 96 exposing the substrate. FIG. 13 is a cross-sectional view showing a single conductor 98 formed in one of the areas 96. Note that the conductor 98 is thicker than the mask 94. The height of the conductor 98 is determined by the amount of conductive ink deposited over the mask. For masks defining a large exposed area of the substrate, more conductive ink needs to be deposited to ensure that the exposed areas are fully covered by the ink. At termination areas of the traces, such as for connecting the ends of the traces to other conductors, an expanded pad area should be formed to ease alignment tolerances for a subsequent printed layer and to improve the resulting electrical connection.

(62) After curing the conductive ink, a dielectric ink is then deposited over the same mask, where the dielectric ink contains sufficient surfactant to cover the mask surface and the conductors and neutralize the hydrophobic effects of the mask. Additional mask and trace layers may be formed to create a 3-dimensional matrix of interconnections. Vertical vias may be used for interconnections between conductor layers.

(63) FIG. 14 illustrates the use of a hydrophobic mask 100 over the circuit board 87 of FIG. 11 when creating the interconnections between groups 72 and 74 to create an AND gate. In another embodiment, the mask 10 is only used in the patch area 86 for programming, and the traces 85 leading to the various groups are formed when printing the various conductor layers for the groups.

(64) This general masking process may also be used for patterning the groups of devices 40 and conductor layers. Groups of the same or different devices may be stacked to allow the formation of very complex circuits.

(65) The programming process may be inexpensively performed in a roll-to-roll process on a large number of the flexible circuit boards 87 after the standard features of the circuit boards 87 have been formed. After the final programming, the circuit boards 87 may be singulated from the roll. As seen, no vacuum processing or dangerous materials are used to fabricate the circuit board 87 and program it.

(66) FIG. 15 illustrates how the resistors R1 and R2 in FIG. 14 may be formed by flexography or screen printing the resistive material on the substrate, where a mask on the screen mesh defines the shapes of the resistive material. Other deposition techniques may be used. Either the shape of the resistive material (length, width, height) may determine the resistance or the position of the connector 102 or 103 along its length may determine the resistance. If the position of the connector determines the resistance, all the resistors may be formed identically. The resistance can also be selected by interconnecting resistors in series and/or parallel.

(67) FIG. 16 illustrates both a bottom view and a cross-sectional view of a circuit board 106 where both sides of the substrate 108 have printed thereon the groups of devices, such as groups 72 and 74 on the bottom and other groups 110 and 112 on the top. The traces 114 and 116 interconnect the groups. Through vias 118 connect the circuits on one side to the circuits on the other side. Prior to printing the interconnect layer, the via holes are punched in the substrate 108 and filled with a UV-cured hole-filling conductor, for example. If mirror images are formed, this simplifies the interconnection design, since the patch areas on both sides can be the same.

(68) Instead of vias, a wrap-around connector can be used.

(69) Since the substrate 108 may be a very thin and flexible film (like a flex-circuit), the resulting circuit board 106 may be folded to reduce its size. Flexible conductors formed by inks are commercially available. There may be special areas on the substrate 108 that define where the circuit board 106 may be folded without damage to the circuitry.

(70) To improve reliability and flexibility in the usage of the circuit boards, a base circuit board 120 (FIG. 17) may be fabricated to have certain basic features and connection terminals. After the circuit board 120 has been tested and approved, additional circuit boards 122 and 124 may be electrically attached to the base circuit board 120 to customize the performance for a particular application.

(71) In the embodiment of FIG. 17, the tested and approved circuit boards 122 and 124 have an adhesive applied to its surface that will be adhered to the base circuit board 120. The terminals 126 of the circuit boards 122 and 124 align with terminals on the base circuit board 120. Those terminals 126 are coated with a conductive adhesive. The circuit boards 122 and 124 are then aligned with the base circuit board 120 and adhered to the surface of the base circuit board 120. In one embodiment, the device side of the circuit boards 122 and 124 face the device side of the base circuit board 120. By forming various functional units separately, the pass rate for each unit during testing will be higher, and the functional units can be connected in various combinations to add more functional possibilities.

(72) In one embodiment, the circuit boards 122/124 are formed in a roll-to-roll process and, after testing, the adhesives are applied at the final station. The circuit boards 122/124 may have test tabs that are cut during singulation. After singulation, the circuit boards 122/124 are adhered to the base circuit board 120. As an arbitrary example, one circuit board 122 may be an A/D converter and the other one may be a D/A converter.

(73) FIG. 18 illustrates another technique for mounting the circuit board 122 to the base circuit board 120. In FIG. 18, the circuit board 122 is perforated at the electrical connection locations at areas 128. The bottom side (non-device side) of the circuit board 122 is then coated with a dielectric adhesive, and the circuit board 122 is adhered to the base circuit board 120 so the perforations are over the connection terminals on the base circuit board 120. A conductive adhesive 130 is then deposited through the perforations to connect the terminals of the base circuit board 120 to the top terminals of the circuit board 122. For example, the traces 132 and 134 are connected by the conductive adhesive 130.

(74) This technique can also be used with the double-sided circuit board of FIG. 16.

(75) Using massively redundant arrays of devices (e.g., devices 40 in FIG. 11), along with the standard passive devices (e.g., resistors R1-R3 in FIG. 11) in the patch areas 86, allows the circuit boards to have a very high pass rate and creates programmable circuit boards that can be later programmed to make unique devices on an as-needed basis.

(76) For higher densities of the groups of the devices, multiple insulated layers of the groups may be printed to form a 3-dimensional structure. Vertical vias may be used to gain access to the various layers. Groups of the devices may be connected in series using vertically aligned groups.

(77) FIG. 19 schematically illustrates one possible assembly line for manufacturing the circuits by printing in a roll-to-roll process. The roll 136 contains the substrate material, and the roll 138 is a take-up roll. The various stations are labeled. The process sequentially prints the various layers and cures the layers. Flexography is preferred for printing using a roll-to-roll process. The number of layers depends on the complexity of the circuits and the devices printed. The roll-to-roll process may produce the unprogrammed circuit boards, and a separate system may be used for the final programming step, depending on a particular customer need.

(78) The various directional attributes used herein, such as bottom, top, and vertical, are not to be construed to convey absolute directions relative to the Earth's surface but are used to convey orientations relative to the enclosed figures when the drawing sheets are held upright. In an actual embodiment, such terms still apply to the product regardless of the absolute orientation of the product relative to the Earth's surface.

(79) While particular embodiments of the present invention have been shown and described, it will be obvious to those skilled in the art that changes and modifications may be made without departing from this invention in its broader aspects and, therefore, the appended claims are to encompass within their scope all such changes and modifications as fall within the true spirit and scope of this invention.