WO1998020549A1 - Utilisation d'une densite de perforation variable de la couche de cuivre pour commander le coefficient de dilatation thermique - Google Patents

Utilisation d'une densite de perforation variable de la couche de cuivre pour commander le coefficient de dilatation thermique Download PDF

Info

Publication number
WO1998020549A1
WO1998020549A1 PCT/US1997/019113 US9719113W WO9820549A1 WO 1998020549 A1 WO1998020549 A1 WO 1998020549A1 US 9719113 W US9719113 W US 9719113W WO 9820549 A1 WO9820549 A1 WO 9820549A1
Authority
WO
WIPO (PCT)
Prior art keywords
cte
chip
layer
layers
package
Prior art date
Application number
PCT/US1997/019113
Other languages
English (en)
Inventor
David A. Hanson
David B. Noddin
Mark F. Sylvester
Original Assignee
W.L. Gore & Associates, Inc.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by W.L. Gore & Associates, Inc. filed Critical W.L. Gore & Associates, Inc.
Priority to AU47596/97A priority Critical patent/AU4759697A/en
Priority to JP10521465A priority patent/JP2000503489A/ja
Publication of WO1998020549A1 publication Critical patent/WO1998020549A1/fr

Links

Classifications

    • HELECTRICITY
    • H01ELECTRIC ELEMENTS
    • H01LSEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L23/00Details of semiconductor or other solid state devices
    • H01L23/34Arrangements for cooling, heating, ventilating or temperature compensation ; Temperature sensing arrangements
    • H01L23/36Selection of materials, or shaping, to facilitate cooling or heating, e.g. heatsinks
    • H01L23/373Cooling facilitated by selection of materials for the device or materials for thermal expansion adaptation, e.g. carbon
    • H01L23/3733Cooling facilitated by selection of materials for the device or materials for thermal expansion adaptation, e.g. carbon having a heterogeneous or anisotropic structure, e.g. powder or fibres in a matrix, wire mesh, porous structures
    • HELECTRICITY
    • H01ELECTRIC ELEMENTS
    • H01LSEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L23/00Details of semiconductor or other solid state devices
    • H01L23/12Mountings, e.g. non-detachable insulating substrates
    • H01L23/14Mountings, e.g. non-detachable insulating substrates characterised by the material or its electrical properties
    • H01L23/142Metallic substrates having insulating layers
    • HELECTRICITY
    • H01ELECTRIC ELEMENTS
    • H01LSEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L23/00Details of semiconductor or other solid state devices
    • H01L23/34Arrangements for cooling, heating, ventilating or temperature compensation ; Temperature sensing arrangements
    • H01L23/36Selection of materials, or shaping, to facilitate cooling or heating, e.g. heatsinks
    • H01L23/367Cooling facilitated by shape of device
    • H01L23/3677Wire-like or pin-like cooling fins or heat sinks
    • HELECTRICITY
    • H01ELECTRIC ELEMENTS
    • H01LSEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L23/00Details of semiconductor or other solid state devices
    • H01L23/48Arrangements for conducting electric current to or from the solid state body in operation, e.g. leads, terminal arrangements ; Selection of materials therefor
    • H01L23/488Arrangements for conducting electric current to or from the solid state body in operation, e.g. leads, terminal arrangements ; Selection of materials therefor consisting of soldered or bonded constructions
    • H01L23/498Leads, i.e. metallisations or lead-frames on insulating substrates, e.g. chip carriers
    • HELECTRICITY
    • H01ELECTRIC ELEMENTS
    • H01LSEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L23/00Details of semiconductor or other solid state devices
    • H01L23/52Arrangements for conducting electric current within the device in operation from one component to another, i.e. interconnections, e.g. wires, lead frames
    • H01L23/538Arrangements for conducting electric current within the device in operation from one component to another, i.e. interconnections, e.g. wires, lead frames the interconnection structure between a plurality of semiconductor chips being formed on, or in, insulating substrates
    • HELECTRICITY
    • H01ELECTRIC ELEMENTS
    • H01LSEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L2224/00Indexing scheme for arrangements for connecting or disconnecting semiconductor or solid-state bodies and methods related thereto as covered by H01L24/00
    • H01L2224/01Means for bonding being attached to, or being formed on, the surface to be connected, e.g. chip-to-package, die-attach, "first-level" interconnects; Manufacturing methods related thereto
    • H01L2224/10Bump connectors; Manufacturing methods related thereto
    • H01L2224/15Structure, shape, material or disposition of the bump connectors after the connecting process
    • H01L2224/16Structure, shape, material or disposition of the bump connectors after the connecting process of an individual bump connector
    • H01L2224/161Disposition
    • H01L2224/16151Disposition the bump connector connecting between a semiconductor or solid-state body and an item not being a semiconductor or solid-state body, e.g. chip-to-substrate, chip-to-passive
    • H01L2224/16221Disposition the bump connector connecting between a semiconductor or solid-state body and an item not being a semiconductor or solid-state body, e.g. chip-to-substrate, chip-to-passive the body and the item being stacked
    • H01L2224/16225Disposition the bump connector connecting between a semiconductor or solid-state body and an item not being a semiconductor or solid-state body, e.g. chip-to-substrate, chip-to-passive the body and the item being stacked the item being non-metallic, e.g. insulating substrate with or without metallisation
    • HELECTRICITY
    • H01ELECTRIC ELEMENTS
    • H01LSEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L2924/00Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
    • H01L2924/10Details of semiconductor or other solid state devices to be connected
    • H01L2924/102Material of the semiconductor or solid state bodies
    • H01L2924/1025Semiconducting materials
    • H01L2924/10251Elemental semiconductors, i.e. Group IV
    • H01L2924/10253Silicon [Si]
    • HELECTRICITY
    • H01ELECTRIC ELEMENTS
    • H01LSEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L2924/00Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
    • H01L2924/15Details of package parts other than the semiconductor or other solid state devices to be connected
    • H01L2924/151Die mounting substrate
    • H01L2924/1515Shape
    • H01L2924/15151Shape the die mounting substrate comprising an aperture, e.g. for underfilling, outgassing, window type wire connections
    • HELECTRICITY
    • H01ELECTRIC ELEMENTS
    • H01LSEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L2924/00Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
    • H01L2924/15Details of package parts other than the semiconductor or other solid state devices to be connected
    • H01L2924/151Die mounting substrate
    • H01L2924/153Connection portion
    • H01L2924/1531Connection portion the connection portion being formed only on the surface of the substrate opposite to the die mounting surface
    • H01L2924/15311Connection portion the connection portion being formed only on the surface of the substrate opposite to the die mounting surface being a ball array, e.g. BGA
    • HELECTRICITY
    • H01ELECTRIC ELEMENTS
    • H01LSEMICONDUCTOR DEVICES NOT COVERED BY CLASS H10
    • H01L2924/00Indexing scheme for arrangements or methods for connecting or disconnecting semiconductor or solid-state bodies as covered by H01L24/00
    • H01L2924/30Technical effects
    • H01L2924/35Mechanical effects
    • H01L2924/351Thermal stress
    • H01L2924/3511Warping
    • HELECTRICITY
    • H05ELECTRIC TECHNIQUES NOT OTHERWISE PROVIDED FOR
    • H05KPRINTED CIRCUITS; CASINGS OR CONSTRUCTIONAL DETAILS OF ELECTRIC APPARATUS; MANUFACTURE OF ASSEMBLAGES OF ELECTRICAL COMPONENTS
    • H05K1/00Printed circuits
    • H05K1/02Details
    • H05K1/0271Arrangements for reducing stress or warp in rigid printed circuit boards, e.g. caused by loads, vibrations or differences in thermal expansion

Definitions

  • the present invention relates generally to the field of microelectronic fabrication and assembly and, more specifically, to methods and devices which reduce or eliminate bending moments and reduce stresses in an integrated circuit chip/package system which result from mis-matched coefficients of thermal expansion (CTE) for the various system components and/or from adhesive curing during assembly. More specifically, the present invention relates to assembly techniques for stabilizing and forming semiconductor chips/package systems, by matching the CTEs of the various components, by providing differential CTE control within the components, and/or by offsetting CTE-induced bending moments with nulling bending moments.
  • CTE coefficients of thermal expansion
  • the industrial significance of this technology is so great that continuous efforts are being made to improve chip performance, reliability and service life.
  • a typical chip is mounted on and is electrically coupled to a supporting substrate.
  • the substrate in turn, is secured to a printed circuit board.
  • the substrate not only serves the intermediate function of coupling electrical signals taken from conductors on the printed circuit board to the chip for processing, but also takes output signals from the chip and applies these output signals to other printed circuit board conductors for further processing.
  • one source of thermally related chip failure resides in the fact that the electrical characteristics of the circuit components that are diffused or otherwise impressed on the chip can vary markedly with changes in chip dimensions. These chip dimension changes caused, for example, by thermal expansion of the chip can make the expanded chip produce useless and, perhaps, damaging output signals. These undesired thermal expansion effects can also cause the central portion of a chip, secured at its margins to a substrate, to curve, bend or bow. This curving frequently causes at least some of the electrical connections between the chip and its underlying substrate to separate and disconnect from each other. Largely for these reasons, chip performance is degraded.
  • thermally induced curving include chip cracking and breaking.
  • chips crack and break in large measure because tensile stresses are established in the outermost surface of the chips as the chips are bent. These stresses, should they exceed the fracture strength of the chip, will cause the chip to crack or break.
  • Thermal effects are not limited to the chip, but also appear in the substrate and in other chip packaging materials.
  • a substrate is a structure that is assembled by stacking together two to fifteen or more layers of substrate materials, at least two of these layers being of different compositions.
  • the materials from which each of these layers are formed tend to be quite diverse, some layers, for instance, being metal (e.g., copper, nickel or gold), other illustrative layers being an epoxy resin and glass compound.
  • the CTE for these individual layers each being considerably different, invite an uncontrolled bending or thermally induced substrate surface distortion that is applied not only to the chip during circuit operation, but also to the substrate through the high temperatures that are required in substrate manufacture.
  • the substrate sur aces that support the chip and establish electrical contact between the chip and the printed circuit board should be “flat” in ail conditions of operation. Indeed, if the substrate itself is not sufficiently “flat”, it will prove impossible to establish electrical contact between the chip and substrate.
  • flatness for the purpose of chip mounting and packaging has been defined as the ratio of the maximum high to low deviation per unit area and has developed into an industry practice in which the maximum acceptable deviation from flatness is 2.5 ⁇ m.
  • warpage or loss in flatness for an entire chip package and the chip package components, in which warpage of more than 6 to 8 mils over the entire chip package is industrially unacceptable. This is a goal that is difficult to achieve, but it is a goal nevertheless, that the chip packaging industry must meet, in spite of the fact that thickness deviations in a given substrate layer can be as much as ⁇ 15%.
  • substrates that are essentially thin films, e.g., about 2 mils or less in thickness. These thin substrate films deform and absorb almost all of the compressive stresses, surface irregularities and the like, thus leaving the chip in an essentially flat, un-deformed condition, similar to the way in which "blister” or “shrink-wrap” packaging conforms itself to the shape of the packaged item.
  • the adhesive (usually an epoxy resin cement) bonds the chip to the substrate and reduces relative movement as a source of the soldered joint deterioration, the adhesive does produce some undesirable side effects. For instance, on curing or hardening, the adhesive becomes a further locus of undesirable structural stress. The cured adhesive shrinks, placing the soldered joints in compression and through these compressive forces thereby applies still another curving movement to the chip and the substrate.
  • Substrates in this intermediate range are too thin to force the chip to maintain a suitable degree of flatness. These intermediate range substrates are also too thick to absorb all of the bending moments, whatever the source, to enable the relatively thicker chip to structurally dominate the combination and maintain an all-important chip flatness. It has been found, in fact, that substrates in this intermediate range are excellent vehicles for transmitting bending stresses to their respective chips, thereby aggravating the chip bending stress difficulties summarized above. Nevertheless, in spite of these structural limitations, there is a significant commercial demand for chip substrates in this intermediate thickness range. To complete a chip package, the chip usually is mounted in the center of the substrate.
  • a ring to stiffen the chip and substrate combination often is bonded or secured to the substrate by means of an adhesive that is applied to the margin of the substrate, essentially enclosing the chip within the ring's center either before or after mounting the chip.
  • the ring forms a frame around the chip with the inner perimeter of the ring being spaced from the corresponding edges of the chip and the height of the ring being somewhat greater than that of the chip.
  • a lid, or cover, joined to that surface of the ring opposite to the ring surface that is bonded to the substrate is spaced above the corresponding die area of the chip, leaving a gap between the upper surface of the chip and the opposing, die area surface of the lid that later is filled with a thermally conducting material.
  • the present invention relates to assembly techniques and the resulting products which are thermally stable, have high structural integrity, and compensate for thermal stresses that occur between the various components of the package. This is accomplished, in-part, by designing the package so that the coefficient of thermal expansion (CTE) of a stiffening ring which is mounted on the package substrate matches the CTE of the substrate and optional lid. Further, the particular adhesives used to bond the stiffening ring are chosen to match their CTE to that of the substrate, ring and lid. Moreover, the substrate is designed so that its CTE, at least in-part, matches that of the chip, and also that of the stiffening ring.
  • CTE coefficient of thermal expansion
  • chip package bending can be significantly reduced by forming one or more slots or holes in the lid.
  • This slot, or slots establish fluid communication with the void space surrounding the chip that is formed in the die area between the chip, the substrate, the lid and the margin between the edges of the chip and the inner walls of the ring.
  • An adhesive is injected through the hole into this void space.
  • the chip is potted, or encapsulated, in the adhesive simultaneously with being underfilled. Bonding the chip to the entire support structure effectively integrates the structure of the chip into the more massive substrate, ring and lid combination.
  • the bonded chip is protected from bending under any one or more of the thermal and mechanical influences noted above.
  • one or more slots or holes also can be formed in the substrate, the ring, or both, also to communicate with the void space that surrounds the chip.
  • An adhesive introduced through a hole in the substrate or the ring alleviates the chip bending problem in the same manner as the adhesive admitted through the slot in the lid. It has been found preferable, however, to flow the adhesive into the void space only from one or two slots. Flowing the adhesive into the void space from more than two slots can produce bubbles in the adhesive, particularly under the chip, a most undesirable result.
  • a substrate can be formed by stacking from two to fifteen or more layers, at least two of these layers each being of different materials.
  • the dominant layers affecting warping occur at the outer surfaces of the substrate.
  • warping effects can be markedly reduced, e.g., the warpage in one instance has been reduced, through the practice of this feature of the invention from about 400 ⁇ m to under 150 ⁇ m.
  • the overall warping of the substrate is avoided to provide a substrate from which the undesirable bending phenomenon has been generally eliminated, at least within the range of temperatures that are ordinarily encountered in chip operation.
  • This principle of the invention has application not only to the substrate, but also can be applied to the ring and lid structures, as appropriate.
  • Ill UNIT AREA COMPOSITION CONTROL Recall that control of individual layer thickness can vary as much as ⁇ 15% from the desired thickness. These thickness irregularities create further bending in the substrate, or other chip package components as described above.
  • the various layers may contain differing compositions at different points in the layer. This might occur as a result of patterning of a metal layer to form discrete conductive pathways.
  • a chip package, manufactured in accordance with a process that further characterizes the invention, however, will provide valuable reductions in bending stress.
  • chip package components (of which the substrate is illustrative) are divided into small unit areas.
  • the unit area composition in each layer progressively toward the outer surface of the component is analyzed to determine if all of the portions of the layers that are equal in distance from the plane of symmetry in the substrate under each unit area contain essentially equal amounts of the same materials. Respective layers are adjusted and controlled to produce a chip package component in which quantities of the same materials that are in each opposing layer within a unit area are approximately equal in amount, thereby providing a structure that has a generally low warpage throughout.
  • This feature of the invention reduces relative differences in CTE between the substrate and the chip, thereby avoiding the differential in thermal expansion that produces bending stresses in the chip, solder joint shearing and fatigue.
  • the chip and the die areas of the substrate both enjoying generally the same CTE, expand and contract together as the temperatures change.
  • By expanding and contracting together relative movement between chip and die area of the substrate that in the prior art forced the chip to curve or to bend and applied shearing forces to solder joints is eliminated.
  • relative movement and the concomitant bending stresses and shearing forces between the chip package and the printed circuit board are also reduced.
  • V SELECTIVE CTE ADJUSTMENT To further cope with bending stresses of thermal origin, the invention also provides for an unusual technique that selectively adjusts the CTE of one layer of material to approximate that of another layer.
  • a structure of this character has two or more grooves, recesses, or holes, of any predetermined and desired shape, formed in a matrix layer. These holes, or loci, are filled with another material that has a CTE which is significantly different from the matrix layer CTE. On heating, the material in the holes expands at a rate and extent that is unlike the surrounding matrix in which the loci are formed.
  • the substrate layers can be provided with a selected number and distribution of holes in the die area. These holes, filled with materials that have different CTE's than the surrounding matrix, approximate the aggregate CTE of the die area matrix in order to approach the CTE of the overlaying chip.
  • a thermally conductive interface can be applied to the die area between the chip and the overlaying lid to conduct heat from the chip to the die area of the lid in order to spread heat generated in the chip over a large portion of the surface area of the lid.
  • a typical interface suitable for use with the invention is described in J.G. A een et al. United States Patent No.
  • chip package lids are formed from copper or aluminum.
  • a combination of aluminum and silicon carbide or copper and silicon carbide or other low CTE reinforcement could be used.
  • aluminum has a CTE of 23 PPM/°C. Consequently, by manipulating the ratio of aluminum to silicon carbide in different portions of the lid, any predetermined CTE in a spectrum that extends from 23 PPM/°C for pure aluminum to 3.7 PPM/°C for pure silicon carbide can be prepared.
  • a high silicon carbide and low aluminum concentration composition can be established in the die area for the lid in order to match the CTE of the chip that is bonded to the adjoining portion of the lid.
  • the marginal portions of the lid that are bonded to the ring are of a different aluminum/silicon carbide proportions.
  • the relative concentrations of aluminum and silicon carbide in the marginal portions of the lid are selected such that the average CTE of the lid matches the average CTE of the substrate/die combination.
  • the desired concentrations of aluminum and silicon carbide can be prepared in several ways.
  • One of these techniques that also characterize the invention provides a porous shape with the same dimensions as the outside dimensions of the ring.
  • the shape has a thickened central core of porous silicon carbide with dimensions about equal to the die area.
  • One or more peripheral recesses at the margin of the shape establish a concentration of silicon carbide which matches, in part, the CTE of the ring.
  • Molten aluminum is essentially dissolved in the porous silicon carbide matrix, the relative proportions of aluminum and silicon carbide varying over the span of the shape to match the respective CTE of each of the underlaying components.
  • the lid's peripheral CTE approximates a predetermined value (e.g., the CTE of the ring).
  • a lid is provided with very diverse thermal expansion characteristics, these characteristics matching the CTE of the lid die area to that of the chip and the peripheral CTE of the lid to that of the ring.
  • An additional feature of the invention counterbalances, or cancels, the bending moments that otherwise would be applied by the substrate to the chip, thereby eliminating relative movement between the substrate and the chip and thus avoiding the associated bending of the chip that this relative movement would cause.
  • electrically inactive components or passive electrical components e.g., capacitors, resistors and inductors
  • capacitors, resistors and inductors that have the same CTE or a CTE that is similar to that of the active chip are coupled to the exposed die area surface of the substrate on the side of the substrate that is opposite to the side to which the electrically active chip is coupled. Because the electrically active chip and the electrically passive or inactive elements both enjoy essentially the same CTE, the thermal expansions of both chips are about equal. Relative movements of these chips with respect to the substrate, although equal, are on opposite sides of the substrate, thereby effectively cancelling any chip-related bending moments ⁇ that otherwise would occur. In this situation, the electrically active chip remains suitably flat.
  • a further embodiment of this feature of the invention provides for the insertion of a stiffener in the die area of the substrate to prevent that portion of the substrate from bending relative to the chip mounted directly over that portion of the substrate, through the range of device operating temperatures. So mounted, the stiffener generally overcomes the undesirable substrate warping in the chip die area.
  • Figure 1 is a plan view of a chip/package system embodying principles of the invention
  • Figure 2 is a sectional view of the chip/package system of Figure 1 , taken along line ll-ll of Figure 1 ;
  • Figure 3 is a vertical sectional view of the package component of the chip/package system of Figure 1 ;
  • Figure 4 is a plan view of a portion of the package, showing a grid pattern for analyzing the material content within each grid of each layer of the package;
  • Figure 5 is a horizontal sectional view, taken along line V-V of Figure 3;
  • Figure 6 is a horizontal sectional view, taken along line VI-VI of Figure 3;
  • Figure 7 is a vertical sectional view of a layer used to form the package, and showing a plurality of pre-lamination holes;
  • Figure 8 is a vertical sectional view of the perforated layer in a pre-lamination stack with dielectric layers and outer conductive layers;
  • Figure 9 is a vertical sectional view of the stack of layers of Figure 8 after lamination, in which material from the dielectric layers flows into and fills the plurality of holes;
  • Figure 10 is a vertical sectional view of an alternative embodiment in which the layer is grooved instead of perforated, and filled with a filler material before lamination;
  • Figure 11 is a plan view of a perforated layer, with the holes formed in a desired pattern
  • Figure 12 is a side elevational view of a preform used to manufacture a lid having different regions of CTE;
  • Figure 13 is a sectional view showing the preform in a mold for pressure infiltrating molten metal into the ceramic preform of Figure 12;
  • Figure 14 is a side elevational view of the completed lid after molding
  • Figure 15 is a side elevational view of a lid having an opening in the middle for receiving an insert having a different CTE than the remainder of the lid;
  • Figure 16 is a vertical sectional view of the lid of Figure 15, taken along line XVI-XVI of
  • Figure 17 is a side elevational view of an insert used to complete the lid shown in
  • Figure 18 is a sectional view showing a false die attached to an underside of the package opposite the chip;
  • Figure 19 is a scanning electron micrograph (SEM) image showing the node-fibril infrastructure of an ePTFE matrix used in one embodiment of the present invention.
  • Figure 20 is a SEM image showing the node-fibril infrastructure of an ePTFE matrix used in another embodiment of the present invention.
  • a chip/package system 10 includes a package 12 having first and second opposite planar surfaces 14 and 16.
  • An integrated circuit chip 18 is connectable through solder bumps or balls 20 to the surface 14 of the package 12.
  • the solder balls 20 establish both a mechanical and electrical connection of the package 12 to the chip 18.
  • a constraining ring 22 having opposite surfaces 24 and 26 is bonded to the package 12 through an adhesive layer 28.
  • the constraining ring has a centrally located rectangular opening which defines a chip-mounting cavity 30.
  • the constraining ring 22 stiffens the package 12 to permit easier handling of the package prior to and after chip attachment and to reduce thermally-induced bending moments which result when the coefficient of thermal expansion (CTE) of the package 12 is mis-matched with that of the chip 18.
  • the thickness of the constraining ring 22 is slightly greater than the corresponding dimension of the chip 18.
  • a lid 32 having opposite surfaces 34 and 36 is bonded to the constraining ring 22 through an adhesive layer 38.
  • the inner surface 38 of the lid 32 is spaced from the upper surface of the chip 18.
  • the outer peripheral edges of the lid 32, constraining ring 22, and package 12 are co-existent and form a square.
  • the square measures 33 mm on each side. It is not necessary, however, for the package to be square shaped, although the square configuration is used frequently. Also, the 33 mm size is one embodiment, but other sizes can be used.
  • the chip/package system 10 When installed in an electronic device, the chip/package system 10 is mounted on a printed wiring board (PWB) 40 through solder balls 42. As with the solder balls 20, the solder balls 42 provide electrical and mechanical connection between the package 12 and the PWB 40.
  • PWB printed wiring board
  • both the package 12 and the PWB 40 are made of multi-layered laminations of alternatingly disposed dielectric and conductive layers.
  • the preferred dielectric materials which are organic, as well as the preferred conductive materials, will be discussed in greater detail below.
  • a problem in chip/package assembly arises when an adhesive underfill is applied between the lower surface of the chip 18 and the upper surface 14 of the package 12 to reinforce the mechanical connection between the package 12 and the chip 18.
  • the adhesive cures and shrinks, there is a tendency for the package 12 to warp, and since the chip 18 is connected to the package 12, a bending moment can be applied to the chip 18. This bending moment, if severe enough, can fracture the chip, disrupt circuits and components diffused in the chip, and/or compromise the solder ball connections between chip and the package.
  • One aspect of the present invention is to provide a counter or offsetting bending moment by filling the chip mounting cavity 30 with an adhesive which mechanically couples the lid 32 to the upper surface of the chip 18.
  • the adhesive is introduced into the chip mounting cavity 30 after the lid 32 has been bonded in place on the constraining ring 22.
  • the adhesive is applied in a liquid state through an opening 44, which is preferably an elongated slot formed parallel to one of the four sides of the chip/package system 10.
  • the liquid adhesive can be applied through a needle extending into the opening 44.
  • the tip of the needle is juxtaposed the solder ball region and the underfill adhesive is released from the tip while advancing the tip along the slotted opening 44.
  • the adhesive migrates by capillary action to cover the area between the lower surface of the chip 18 and the upper surface of the package 12.
  • the tip of the applicator or needle can be juxtaposed within the space between the lower surface 38 of the lid 32 and the upper surface of the chip 18, and as the needle is again advanced along the slotted opening, capillary action draws the adhesive into this space.
  • the entire chip mounting cavity can be filled with the adhesive, as shown in Figure 2, or simply the spaces between the chip 18 and surfaces of the lid 32 and package 12.
  • an opening 46 may be formed in the package 22. If the entire cavity 30 is to be filled, there is no requirement for either opening 44 or 46 to be aligned with any one of the outer edges of the chip 18.
  • Opening 46 would likely be a hole, and not a slot. Moreover, two openings are generally preferred when filling the cavity with adhesive to allow air to escape during adhesive filling.
  • the two holes can both be formed in the lid, or one can be formed in the lid and the other in the package, as depicted in Figure 2.
  • the package 12 and lid 32 are made of materials selected to approximately match the CTE of the two components so that the opposing bending moments are equal, but opposite.
  • the adhesive that is introduced into the cavity 30 is preferably thermally conductive to ensure that heat generated by the chip 18 passes from the cavity 30 to lid 32, which is likewise preferably made of a thermally conductive material.
  • the adhesive layers 28 and 38 can be made of any suitable adhesive material, such as, but not limited to epoxy adhesives, porous substrates impregnated with adhesives that are moisture resistant and are able to withstand temperatures in excess of 150°C, porous substrates that are impregnated with an adhesive-filler mixture that are moisture resistant and able to withstand temperatures in excess of 150°C, preferably in excess of 200°C, porous substrates that are impregnated with an adhesive-conductive particle mixture that is moisture resistant and able to withstand temperatures in excess of 150°C, preferably in excess of 200°O, e.g., Ablestik®.
  • the adhesive layers 28 and 38 may be made of the same material or different materials.
  • the overfill and underfill adhesives can be a liquid epoxy containing SiO 2 , e.g., Hysol 45260.
  • the overfill adhesives can also be solders or other metallurgical bonding agents. Other suitable materials for use as adhesive materials are described below.
  • the adhesive materials used for adhesive layers 28 and 38 are preferably in sheet form, while the adhesive filled in the cavity 30 is preferably in liquid form. However, the sheet adhesive layers could be replaced with liquid materials, and the liquid materials in the cavity 30 could be replaced with sheet materials, at least with respect to the area between the chip 18 and lid 32. II Selectively Stacked Substrate Layers
  • FIG 3 is a vertical sectional view of a laminated structure for the package 12, which may be used in the chip/package system of Figure 1.
  • each layer being about a 20 ⁇ m thick layer of metal, those materials that have the least departure from the manufacturer's thickness tolerance (usually the metal layers) and by progressively adding from the outer layers 48 and 50 toward the center or core layer 52 (e.g., copper, 35 ⁇ m thick) those materials in which manufacturing thickness tolerances become progressively greater (usually the dielectric layers), the overall substrate is balanced from a thermal expansion standpoint. The balance is such, moreover, that the curving, and its destructive effect on the chip (not shown in Figure 3), or the ability to establish electrical contact with the chips is largely overcome.
  • a combination of layers, to form a suitable substrate that does not exhibit significant thermal bending within the normal range of chip operation, as shown in Figure 3, would comprise, in order from the soldermask layer 48, inwardly toward the conductive copper center of core layer 52 through which a plane of symmetry passes: a) 20 ⁇ m Cu/Ni/Au layer 56, a conductor; b) 44 ⁇ m cyanate ester-epoxy-ePTFE (CE/E-ePTFE) layer 58, a dielectric; c) 9 ⁇ m Cu layer 60, a conductor; d) 50 ⁇ m bismalimide-triazine (BT)-Epoxy/Glass layer 62, a dielectric; e) 18 ⁇ m Cu layer 64, a conductor; and f) 44 ⁇ m CE/E-ePTFE layer 66, a dielectric.
  • CE/E-ePTFE cyanate ester-epoxy-ePTFE
  • the CE/E-epoxy-ePTFE layer 58 is prepared as follows: a varnish solution is made by mixing 5.95 pounds of M-30 (Ciba Geigy), 4.07 pounds of RSL 1462 (Shell Resins, Inc.), 4.57 pounds of 2, 4, 6-tribromophenyl-terminated tetrabromobisphenol A carbonate oligomer (BC-58) (Great Lakes Inc.), 136g bisphenol A (Aldrich Company), 23.4g Irganox 1010, 18.1 g of a 10% solution of Mn HEX-CEM in mineral spirits, and 8.40 kg MEK. The varnish solution was further diluted into two separate baths - 20% (w/w) and 53.8%
  • the two varnish solutions were poured into separate impregnation baths, and an e- PTFE web was successively passed through each impregnation bath one immediately after the other.
  • the varnish was constantly agitated so as to insure uniformity.
  • the impregnated web was then immediately passed through a heated oven to remove all or nearly all the solvent and partially cure the adhesives, and was collected on a roll.
  • the ePTFE web such as that shown in Figure 11 , may be any desired thickness, such as 25 ⁇ m, 40 ⁇ m, for example.
  • a 25 ⁇ m thick material has a weight per area of approximately 11.2 to 13.8 g/m 2 .
  • an identical, corresponding array of layers is provided between the outer layer 50 and the core layer 52.
  • CE/E-ePTFE which has a manufacturing thickness tolerance considerably greater than the copper layers
  • BT-epoxy/glass to meet the electrical needs of the apparatus, i.e., the circuit (which has a manufacturing thickness tolerance greater than CE/E-ePTFE) traces for the conductors and, as such, are not true copper layers, but are a pattern of copper conductors.
  • the principle of this aspect of the invention that the better controlled thickness materials are positioned toward the outside of the package 12, and the layers with lower thickness control positioned closer to the core layer 52 must be violated occasionally, as it can be seen with respect to the CE/E-ePTFE layer 66, to satisfy the primary electrical needs of the microchip package.
  • the thicknesses of the copper layers 54, 70 and 64 can be carefully controlled and the BT-epoxy/glass and CE/E-ePTFE layer thicknesses can also be controlled, but not as well as the copper layer thicknesses, the interposition of these copper layers among the B-T epoxy/glass and CE/E-ePTFE does not aggravate the CTE response of the substrate 12, but further serves to reduce the aggregate CTE of the substrate 12.
  • the principal feature of this aspect of the invention consequently, is characterized in the fact that the manufacturing tolerances, and thus, the potential for departing from a predetermined layer thickness and aggravating the overall CTE response of the substrate 12 is greater for the CE/E-ePTFE layer 66, which also has the highest CTE of any of the layers, than it is for any of the other layer materials in the substrate 12.
  • Other appropriate layer materials can be substituted for those which are described in connection with Figure 3, it only being necessary to arrange the layers according to the principles expressed above.
  • FR 4 a common dielectric material (or any suitable dielectric material) layer can be substituted for the BT-epoxy/glass layers shown in FIG. 3.
  • a greater or lesser number of layers forming the laminated package 12 can be employed.
  • compositions of the layers that form the substrate 12 can be changed through a substitution of other known substrate layer materials, as noted above, it being clear that the improved performance that is provided through the invention is attained by positioning those layers with the greatest manufacturing thickness tolerances (usually the dielectrics) as the innermost layers in the substrate.
  • the arrangement of individual layers in the laminated substrate or package 12 is selectively made to achieve dimensional control, increased flexural stiffness, and matching of CTEs within the substrate.
  • Another aspect of the present invention is to analyze the material composition of each layer to determine whether offsetting formations in opposing layers should be made to ensure that the CTEs of the two layers are matched.
  • conductive layer 60 is a patterned layer that may have areas where metal has been removed to form a circuit.
  • a corresponding area of symmetrically disposed conductive layer 74 may have a different metal content, resulting in localized differences in CTE between the two layers 60 and 74.
  • the differences in CTE can result in warping.
  • layers formed by deposition where more or less metal is deposited on the different but symmetrically opposed layers.
  • One aspect of the present invention is to analyze the material content of one layer, and then alter the opposing layer to match the material content.
  • substrate 12 is divided into an array of unit area squares 80. These squares measure, for example, 1 mm 2 .
  • the squares 80 extend in space through all layers of the package so that each square of one layer has a counterpart area in a symmetrically opposed layer.
  • compositions of each layer in the substrate 12 that are within the confines of each of the respective squares in the array of squares are analyzed to identify the materials and the concentrations of each of these materials in every layer of the unit area under consideration.
  • the copper conductors that comprise the "layer" 60 do not extend across the entire area of the substrate 12, but only occupy a small non-uniform or unequally distributed portion of the copper in that area, the BT-epoxy/glass layer 62 being exposed in those portions of the plan view area of the substrate 12 that are not covered by the copper conductors which comprise the layer 60.
  • Figure 6 shows the unequal or non-uniform disposition of the copper electrical conductors 74, and the somewhat larger concentration of copper in the conductors that form the layer 74 in comparison with respect to the quantity of copper in the layer 60.
  • the illustrative copper concentration per unit area in the layers can be developed.
  • the adjusted concentrations of copper in common unit areas 80 in each of the layers 60 and 74 are not identical because of electrical circuit requirements that are imposed on these layers.
  • the electrically inactive copper strip 60A in the layer 60 and the electrically unnecessary copper portion 74A that is removed from the layer 74 nevertheless do establish an approximate balance between the two copper electrical circuit layers 60 and 74.
  • Warping can be reduced considerably in the substrate 12 by balancing the concentrations of the same substrate layer materials. Balancing, for this purpose, is defined as establishing equal concentrations, or void spaces on opposing pairs of layers that are at equal distances from the plane of symmetry.
  • a strip of electrically inactive copper 60A is added to the surface of the BT-epoxy/glass layer 62.
  • the relative copper concentrations also can be balanced by removing an equivalent amount of electrically unnecessary copper 74A from the electrical conductors that form the layer 74.
  • perfect balance can not be achieved in actual practice. Balance, for the purpose of this invention, is reached to the extent that the electrical requirements of the circuit are satisfied while nevertheless establishing as close an approximation to an equality between the respective copper concentrations in each of the layers as the physical structure will permit.
  • the imbalance may be compensated for by purposely imbalancing a second pair of layers, e.g., 64 and 70, so as to create an equal and opposite bending moment.
  • the CTE of the opposing layers is matched locally.
  • the areas overlying each other meaning a common zone, has an approximately similar amount of material, either by altering one or the other, or both, of the two layers.
  • the alternation can be in the form of material additions, such as by deposition, or material subtractions, such as by etching.
  • the effect is the same: the package reliability is enhanced by accounting for CTE gradients across the length and width of the individual, symmetrically matched layers.
  • an integrated circuit chip such as chip 18, is typically made of a material, such as silicon, that has a substantially different coefficient of thermal expansion (CTE) as compared to an underlying package made of non-ceramic materials, such as the package 12.
  • CTE coefficient of thermal expansion
  • PWB printed wiring board
  • the portion 12A of the package will have a lower CTE than the surrounding area of the substrate.
  • the center area could have a higher CTE than the surrounding area.
  • Creating a different CTE for the middle, chip-mounting area 12A of the package 12 can be accomplished in a variety of ways.
  • the combination of layered materials for the portion 12A of the package 12 that is under the chip 18, e.g., copper, molybdenum or Invar, are selected to produce an overall CTE in the region 12A that is similar to but generally greater than the CTE of the chip 18 through the normal range of operating temperatures for both the chip 18 and the package 12. In this manner, undesirable relative movement between the chip 18 and the substrate 12 is largely eliminated.
  • the area 12A of the package 12 under the chip 18 should enjoy a CTE as close to the range of 2-3 PPM/°C as is possible, illustratively 6 to 10 PPM/°C.
  • the aggregate or average CTE of the package 12 also should be in the 16-17 PPM/°C range in order to approach the printed circuit board CTE.
  • the chip mounting region does not have to be centrally located but can be disposed virtually in any position on the package.
  • the CTE can be varied by physically altering the layered structure that forms the package 12.
  • a core layer 82 is drilled, etched or otherwise worked to include a plurality of transversely oriented holes 84 which extend through the core layer from one surface 86 to the opposite surface 88.
  • the core layer 82 is one that would correspond to the core layer 52 of Figure 3, and as with layer 52, the core layer is usually made of copper.
  • the core layer 82 is shown stacked between two dielectric layers 90 and
  • a lamination assembly of layers will additionally include conductive layers 94 and 96 disposed respectively on opposite sides of the dielectric layers 90 and 92.
  • the dielectric layers are made of an organic material which just before pressing is in a b-stage, the dielectric material flows into and fills the holes 84, as shown in Figure 9.
  • the core layer 82 thus becomes a composite which includes a copper matrix having dielectric material dispersed therethrough.
  • the dielectric material has a higher CTE than copper and thus, the CTE of the core layer is increased where the holes 84 are present.
  • the holes 84 could be filled with any material prior to lamination to achieve a desired effect on the CTE of the package.
  • a filler having a lower CTE could fill the holes 84 as a liquid and solidified before passing the core 82 on the lamination step.
  • a core layer 98 can be machined, etched or otherwise worked to form one or more grooves 100 which are filled with a filler material 102 which is selected to achieve a desired CTE change.
  • a lower CTE filler will lower the CTE of the core 98, and a higher CTE will raise the core CTE.
  • the core can be worked with any number of known techniques, such as laser drilling, machining, punching, etching, etc.
  • the formations are not necessarily made to the core layer of a laminated package, but could be made in any layer, whether dielectric or conductive.
  • the symmetric counterpart layers When making formations in layers beyond the core layers, the symmetric counterpart layers must be similarly modified.
  • the same pattern of holes and same filler must be used in the same position of the symmetrically opposed conductive layer 70.
  • the holes 84 can be arranged in a grid-like pattern in the core layer 82 in a relatively tightly packed arrangement where localized changes in CTE are required. For example, to match the CTE of the package 12 in an area under the chip 18
  • the core layer 82 can be patterned with holes in the chip mounting area that overlies the core layer. The holes are then filled with a material that increases or decreases the CTE of the package in that area.
  • the chip 18, which can be on the order of 1cm 2 to 4cm 2 generates an unusual amount of heat, on occasion as much as 100 watts.
  • lids usually were formed from either aluminum or copper or a composite material, e.g., by adding either aluminum or copper to silicon carbide or some other low CTE reinforcement matrix.
  • Aluminum for instance, has a CTE of 23 PPM/°C and pure silicon carbide has a CTE of 3.7 PPM/°C. Consequently, the potential exists for the lid
  • the lid should be provided with different regions of CTE to match the different components of the chip/package
  • the lid 32 can be made of a metal matrix material which includes silicon carbide and aluminum.
  • concentration of materials can be varied to achieve a desired two-region CTE in the lid, and in particular, the central region of the lid, corresponding positionally to the chip 18, will be configured to have a lower CTE than the surrounding region of the lid 32, which preferably is CTE matched to the higher CTE constraining ring 22.
  • a two-region CTE lid can be constructed in the following manner. Referring to Figure
  • a preform 104 is made of silicon carbide powder or whiskers.
  • the preform can be pressed to achieve a desired shaped, or can be machined or cut to form its desired shape.
  • the desired shape is one where the middle region 106 has a thicker dimension than the outer regions 108 and 110.
  • the preform 104 can be made using standard powder metallurgy techniques.
  • the preform 106 is next placed in a mold and pressure infiltrated with molten aluminum. As seen in Figure 13, the mold die 112 and 114 approximately abut the middle region of the preform, and form voids around the outer regions. After pressure infiltration, and as shown in Figure 14, a finished lid 116 has the overall shape desired to fit the package 12, but since the middle region has a higher concentration of silicon carbide, the middle region will have a lower CTE. Specifically, the lower CTE is designed to approximately match that of the chip, which is about 2.6, while the outer regions having higher concentrations of higher
  • the peripheral portion of the lid that is bonded to the constraining ring should have a high aluminum and low silicon carbide concentration to approach the CTE for the ring, in absorbing the aluminum in the peripheral portion of the silicon carbide that forms the lid, relatively more aluminum is added to this peripheral portion than to the central portion of the lid.
  • the average CTE of the lid matches the average CTE of the package 12 and the different average CTE of the chip 18. This manipulation of relative aluminum and silicon carbide concentrations in the lid permits the microchip package and components to remain essentially flat while reducing the stress applied to the die or the adhesive interface.
  • the large central portion of the silicon carbide preform which is essentially porous, establishes a ratio of aluminum to silicon carbide in the different regions of the finished lid. As noted, a greater concentration of aluminum relative to the silicon carbide is required in the vicinity of the ring. These concentrations are established by dissolving aluminum in the preform and filling the voids above and below the outer regions of the preform. In this way, a concentration of aluminum and silicon carbide producing a CTE that approaches the CTE of the ring in one region and the chip in another region is provided.
  • silicon carbide/aluminum systems are described above, other materials may be used to form lids exhibiting different CTEs in different regions. It is also possible to use a single powdered material, ceramic or metal, and vary the CTEs in different regions by applying different consolidation forces. For example, the middle region could have a greater or lesser theoretical density than the outer regions, which could provide a sufficient difference of CTE in some cases. In this case, the powdered materials can be selected and/or mixed attain the desired differential in CTE.
  • a lid 118 is made to include a central opening 120.
  • An insert 122 is sized to fit in the opening 120.
  • the insert 122 is made of a material that has a CTE that is different than that of the material used to form the lid 118. In particular, if a lower CTE is desired for the middle region of the lid, as when the middle region overlies the chip, the insert can be made of material having a lower CTE
  • the lid 118 can be made as a single piece and the central opening 120 can be machined out, or the opening can be formed integrally, as when the lid is made from powdered metal or ceramic.
  • the insert can be formed by the same techniques and may for example comprise a silicon carbide preform infiltrated with aluminum. At any rate, the CTE of the insert can be customized to match that of the chip while the lid 118 CTE can be customized to match that of the constraining ring. Also, a higher concentration of silicon carbide in the center region of the lid can be accomplished by removing the center region of a low density porous silicon carbide preform which has the same dimensions of the desired lid.
  • a chip 124 is shown mounted on a package 126 through solder balls 128.
  • a constraining ring 130 is mounted to the upper surface of the package 126. Due to the difference in CTE between the chip 124 and the package 126, a bending moment is generated that acts upon the chip 124 and can potentially compromise the flatness of the chip. The differences in CTEs are largely due to the material differences in that the chip 124 is silicon and the package is one which utilizes thin organic dielectric layers made of the materials described herein. According to the invention, a false die 132 is bonded to the lower surface of the package 126 prior to bonding the chip to the package.
  • the false die 132 is CTE matched to the chip 124 and tends to oppose the bending moments that are generated by the chip/package attachment, particularly as the underfill adhesive is cured. Thus, the false die 132 counterbalances the moments during chip attach with opposing bending moments.
  • the false die 132 can be made of silicon but other materials exhibiting CTEs similar to the chip 124 can be used. However, it should be positioned directly opposite to the chip 124. Also, materials having different CTEs over the silicon chip could be used, if made thicker to resist the bending moments. For example, aluminum has a CTE of about 9, but could be used if made thick enough to get the same effect.
  • a further aspect of this invention is to adhesively bond the false die to the package at the same time the underfill adhesive is applied to the region between the chip 124 and the package 126.
  • the co-curing of both adhesives will ensure that the false die bending moments offset bending moment generated by adhesive shrinkage at the underfill.
  • the false die 132 can be a passive stiffener, it can also take the form of an electrical component such as a capacitor formed on the package. Also, the false die could be a real chip that failed quality control or it can be one or more circuit components that are frequently referred to as "passive" components, e.g., capacitors, resistors and inductances. The significant point being, however, that the passive component, throughout the range of anticipated operating temperatures for the microchip package, should provide thermal expansion caused bending moments that are essentially equal to, but opposite in direction to, those created by the chip.
  • each of the embodiments of the invention described above significantly contribute to an improved microchip package through the essential elimination, control or reduction of relative movement among the individual components that form the microchip package, without limitation to the source of this relative movement.
  • each of these sources of relative movement among the microchip package components and the undesirable consequence of such movement is alleviated to a significant extent.
  • adhesives 28 and 38 can be a variety of adhesive materials.
  • adhesives include a fluoropolymer material, such as, is a porous polytetrafluoroethylene (PTFE), and especially an extruded and/or expanded PTFE, such as that taught in U.S. Pat. No. 3,953,566 to Gore.
  • PTFE porous polytetrafluoroethylene
  • the material is usually porous expanded
  • PTFE which has been stretched at least 2 to 4 times its original size in accordance with U.S. Pat. No. 3,953,566. This stretching created pores that act as natural air reliefs when the filled material is urged between two components.
  • Suitable interface compositions include PTFE with about 50 to 60% by volume of the solid components of ZnO, BN, or any other good thermally but electrically non-conductive filler.
  • the final product may be expanded in ratios of 4:1 or 3:1 or 2:1 to achieve the desired degree of conformability.
  • the presence of the pores created from the expansion process is responsible for the conforming nature of the finished product and aids in the relief of trapped air when this material is placed between two parallel plates and then are urged together.
  • These materials may be formed into any suitable shape, such as thin tapes having thicknesses in the 5 to 15 mil (0.127 to 0.381 mm) range.
  • Another suitable composition for use in the present invention involves filling the PTFE with a metal powder, such as copper or nickel, having particle sizes in the 1 to 40 micron range. Bimodal and trimodal distributions can increase the loading of this material, such as providing particles in the 1 to 5 micron range mixed with particles in the 40 to 45 micron range. This allows greater packing density, with a subsequent increase in thermal conductivity without sacrificing conformability.
  • the total volume percent (including air) of metal to finished filled PTFE is in the 20 to 90% range.
  • the finished material 17 and 56 can be a material that is a porous matrix system which is imbibed or impregnated an adhesive- filler mixture.
  • the adhesives 17 and 56 are a porous matrix that is a non-woven substrate imbibed with high quantities of filler and a thermoplastic or thermoset adhesive, as a result of the initial void volume of the substrate, heated to partially cure the adhesive and form a B-stage composite.
  • Substrates include fluoropolymers, such as the porous expanded polytetrafluoroethylene material of U.S. Patent Nos. 3,953,566 and 4,482,516, each of which is incorporated herein by reference.
  • the mean flow pore size (MFPS) should be between about 2 to 5 times or above that of the largest particulate, with a MFPS of greater than about 2.4 times that of the filler being particularly preferred.
  • suitable composites can be prepared by selecting the ratio of the mean flow pore size to average particle size to be greater than 1.4. Acceptable composites can also be prepared when ratio of the minimum pore size to average particle size is at least above 0.8, or the ratio of the minimum pore size to the maximum particle size is at least above 0.4.
  • the MFPS to particle size ratios may be determined with a Microtrak® Model FRA Particle Analyzer device.
  • Another mechanism for gauging relative pore and particle sizes may be calculated as the smallest pore size being not less than about 1.4 times the largest particle size.
  • porous expanded polyolefins such as ultra high molecular weight (UHMW) polyethylene, expanded polypropylene, polytetrafluoroethylene made prepared by paste extrusion and incorporating sacrificial fillers, porous inorganic or organic foams, or microporous cellulose acetate, can also be used.
  • the porous substrate has an initial void volume of at least 30%, preferably at least 50%, and most preferably at least 70%, and facilitates the impregnation of thermoset or thermoplastic adhesive resin and particulate filler paste in the voids while providing a flexible reinforcement to prevent brittleness of the overall composite and settling of the particles.
  • the filler comprises a collection of particles when analyzed by a Microtrak® Model
  • FRA Particle Analyzer device which displays a maximum particle size, a minimum particle size and an average particle size by way of a histogram.
  • Suitable fillers to be incorporated into the adhesive include, but are not limited to,
  • Especially preferred fillers are SiO 2 , ZrO 2 , Ti0 2 alone or in combination with non-conductive carbon.
  • Most preferred fillers include filler made by the vapor metal combustion process taught in U.S.
  • Patent No. 4,705,762 such as, but not limited to silicon, titanium and aluminum to produced silica, titania, and alumina particles that are solid in nature, i.e., not a hollow sphere, with a uniform surface curvature and a high degree of sphericity.
  • the fillers may be treated by well-known techniques that render the filler hydrophobic by silylating agents and/or agents reactive to the adhesive matrix, such as by using coupling agents.
  • Suitable coupling agents include, silanes, titanates, zirconates, and aluminates.
  • Suitable silylating agents may include, but are not limited to, functional silylating agents, silazanes, silanols, siloxanes.
  • Suitable silazanes include, but are not limited to, hexamethyldisilazane (Huls H730) and hexamethylcyclotrisilazane, silylamides such as, bis(trimethylsilyl)acetamide (Huls B2500), silylureas such as trimethylsilylurea, and silylmidazoles such as trimethylsilylimidazole.
  • Titanate coupling agents are exemplified by the tetra alkyl type, monoalkoxy type, coordinate type, chelate type, quaternary salt type, neoalkoxy type, cyclo-heteroatom type.
  • Preferred titanates include, tetra alkyl titanates, Tyzor® TOT ⁇ tetrakis(2-ethyl-hexyl) titanate, Tyzor® TPT ⁇ tetraisopropyl titanate ⁇ , chelated titanates, Tyzor® GBA ⁇ titanium acetylacetylacetonate ⁇ , Tyzor® DC ⁇ titanium ethylacetacetonate ⁇ , Tyzor® CLA ⁇ proprietary to DuPont ⁇ , Monoalkoxy (Ken-React® KR TTS), Ken-React®, KR-55 tetra (2,2 diallyloxymethyl)butyl, di(dithdecyl)phosphi
  • Suitable zirconates include, any of the zirconates detailed at page 22 in the Kenrich catalog, in particular KZ 55- tetra (2,2 diallyloxymethyl)butyl, di(ditridecyl)-phosphito zirconate, NZ-01- neopentyl(diallyl)oxy, t neodecanoyl zirconate, NZ-09-neopentyl-(diallyl)oxy, tri(dodecyl)benzene-sulfonyl zirconate.
  • the aluminates that can be used in the present invention include, but are not limited to Kenrich®, diisobutyl(oleyl)acetoacetylaluminate (KA
  • thermosetting matrix adhesives such as, cross-linked vinylic polymers, e.g., divinylbenzene, divinyl pyridine or a sizing of any of the disclosed thermosetting matrix adhesives that are first applied at very high dilution (0.1 up to 1.0% solution in MEK) can be used.
  • certain organic peroxides such as, dicumylperoxide can be reacted with the fillers.
  • the adhesive itself may be a thermoset or thermoplastic and can include polyglycidyl ether, polycyanurate, polyisocyanate, bis-triazine resins, poly (bis-maleimide), norbornene- terminated polyimide, polynorbornene, acetylene-terminated polyimide, polybutadiene and functionalized copolymers thereof, cyclic olefinic polycyclobutene, polysiloxanes, poly sisqualoxane, functionalized polyphenylene ether, polyacrylate, novolak polymers and copolymers, fluoropolymers and copolymers, melamine polymers and copolymers, poly(bis phenycyclobutane), and blends or prepolymers thereof. It should be understood that the aforementioned adhesives may themselves be blended together or blended with other polymers or additives, so as to impact flame retardancy or enhanced toughness.
  • mean flow pore size and minimum pore size were determined using the Coulter® Porometer II (Coulter Electronics Ltd., Luton UK) which reports the value directly.
  • Average particle size and largest particle size were determined using a Microtrak® light scattering particle size analyzer Model No. FRA (Microtrak Division of Leeds & Northup, North Wales, PA, USA).
  • the average particle size (APS) is defined as the value at which 50% of the particles are larger.
  • the largest particle size (LPS) is defined as the largest detectable particle on a Microtrak® histogram. Alternatively, the largest particle size is defined at the minimum point when the Microtrak FRA determines that 100% of the particulate have passed.
  • the method for preparing the adhesive-filler composite involves:(a) expanding a polytetrafluoroethylene sheet by stretching a lubricated extruded perform to a microstructure sufficient to allow small particles and adhesives to free flow into the void or pore volume; (b) forming a paste from polymeric, e.g., thermoset or thermoplastic material and a filler; and (c) imbibing by dipping, coating, pressure feeding, the adhesive-filler paste into the highly porous scaffold, such as expanded polytetrafluoroethylene.
  • Table 1 shows the effect of the relationship of the substrate mean flow pore size (MFPS) and particulate size.
  • MFPS mean flow pore size
  • a fine dispersion was prepared by mixing 281.6 g TiO 2 (Tl Pure R-900, Du Pont Company) into a 20% (w/w) solution of a flame retarded dicyanamide/2-methylimidazole catalyzed bisphenol-A based polyglycidyl ether (Nelco N-4002-5, Nelco Corp.) in MEK. The dispersion was constantly agitated so as to insure uniformity. A swatch of expanded PTFE was then dipped into the resin mixture. The web was dried at 165°C for 1 min. under tension to afford a flexible composite.
  • the partially-cured adhesive composite thus produced comprised of 57 weight percent TiO 2 , 13 weight percent PTFE and 30 weight percent epoxy adhesive.
  • a fine dispersion was prepared by mixing 386 g Si0 2 (HW-11-89, Harbison Walker Corp.) which was pretreated with phenyltrimethoxysilane (04330, Huls/Petrarch) into a manganese catalyzed solution of 200 g bismaleimide triazine resin (BT206OBJ, Mitsubishi Gas Chemical) and 388 g MEK.
  • the dispersion was constantly agitated so as to insure uniformity.
  • a swatch of 0.0002" thick expanded PTFE was then dipped into the resin mixture, removed, and then dried at 165°C for 1 min. under tension to afford a flexible composite.
  • EXAMPLE 3 A fine dispersion was prepared by mixing 483 g SiO 2 (HW-11-89) into a manganese- catalyzed solution of 274.7 g bismaleimide triazine resin (BT2060BJ, Mitsubishi Gas Chemical) and 485 g MEK. The dispersion was constantly agitated so as to insure uniformity. A swatch of 0.0002" thick expanded PTFE was then dipped into the resin mixture, removed, and then dried at 165°C for 1 min. under tension to afford a flexible composite. Several plies of this prepreg were laid up between copper foil and pressed at 250 psi in a vacuum-assisted hydraulic press at temperature of 225°C for 90 minutes then cooled under pressure.
  • the resulting dielectric thus produced comprised of 57 weight percent SiO 2 , 4 weight percent PTFE and 39 weight percent adhesive, displayed good adhesion to copper, dielectric constant (at 10 GHz) of 3.2 and dissipation factor (at 10 GHz) of 0.005.
  • a fine dispersion was prepared by mixing 15.44 kg TiO 2 powder (Tl Pure R-900, DuPont Company) into a manganese-catalyzed solution of 3.30 kg bismaleimide triazine resin (BT206OBH, Mitsubishi Gas Chemical) and 15.38 kg MEK. The dispersion was constantly agitated so as to insure uniformity. A swatch of 0.0004" TiO 2 -filled expanded PTFE (filled according to the teachings of Mortimer U.S. Patent No. 4,985,296, except to 40% loading of TiO 2 and the membrane was not compressed at the end) was then dipped into the resin mixture, removed, and then dried at 165°C for 1 min. under tension to afford a flexible composite.
  • the partially cured adhesive composite thus produced comprised of 70 weight percent TiO 2 , 9 weight percent PTFE and 21 weight percent adhesive.
  • Several plies of this prepreg were laid up between copper foil and pressed at 500 psi in a vacuum-assisted hydraulic press at temperature of 220°C for 90 minutes then cooled under pressure. This resulting dielectric displayed good adhesion to copper, dielectric constant of 10.0 and dissipation factor of 0.008.
  • a fine dispersion was prepared by mixing 7.35 kg SiO 2 (ADMATECHS SO-E2, Tatsumori LTD) with 7.35 kg MEK and 73.5 g of coupling agent, i.e.,3-glycidyloxypropyltri-methoxysilane (Dynasylan GLYMO (Petrach Systems).
  • SO-E2 is described by the manufacture as having highly spherical silica having a particle diameter of 0.4 to 0.6 mm, a specific surface area of 4- 8m 2 /g, a bulk density of 0.2-0.4 g/cc (loose).
  • This dispersion was subjected to ultrasonic agitation through a Misonics continuous flow cell for about 20 minutes at a rate of about 1-3 gal./minute.
  • the fine dispersion thus obtained was further diluted to an overall bath concentration of 11.9% solids (w/w).
  • the Frazier number relates to the air permeability of the material being assayed.
  • Air permeability is measured by clamping the web in a gasketed fixture which is provided in circular area of approximately 6 square inches for air flow measurement.
  • the upstream side was connected to a flow meter in line with a source of dry compressed air.
  • the downstream side of the sample fixture was open to the atmosphere. Testing is accomplished by applying a pressure of 0.5 inches of water to the upstream side of the sample and recording the flow rate of the air passing through the in-line flowmeter (a ball-float rotameter that was connected to a flow meter.
  • the Ball Burst Strength is a test that measures the relative strength of samples by determining the maximum at break.
  • the web is challenged with a 1 inch diameter ball while being clamped between two plates.
  • the Chatillon, Force Gauge Ball/Burst Test was used.
  • the media is placed taut in the measuring device and pressure affixed by raising the web into contact with the ball of the burst probe. Pressure at break is recorded.
  • the web described above was passed through a constantly agitated impregnation bath at a speed at or about 3 ft./min, so as to insure uniformity.
  • the impregnated web is immediately passed through a heated oven to remove all or nearly all the solvent, and is collected on a roll.
  • Example 4 The physical properties of the particulate filler used in Example 4 and Example 7 are compared below.
  • EXAMPLE 6 An ePTFE matrix containing an impregnated adhesive filler mixture, based on SiO 2 prepared from the vapor combustion of molten silicon is prepared as follows. Two precursor mixtures were initially prepared. One being in the form of a slurry containing a silane treated silica similar to that of Example 5 and the other an uncatalyzed blend of the resin and other components. Mixture I
  • the silica slurry is a 50/50 blend of the SO-E2 silica of Example 5 in MEK, where the silica contains a coated of silane which is equal to 1 % of the silica weight.
  • a five gallon container 17.5 pounds of MEK and 79 grams of silane were added and the two components mixed to ensure uniform dispersion of the silane in the MEK. Then, 17.5 pounds of the silica of Example 5 were added.
  • Two five gallon containers of the MEK-silica-silane mixture were added to a reaction vessel, and the contents, i.e., the slurry, was recirculated through an ultrasonic disperser for approximately one hour to break up any silica agglomerates that may be present. The sonication was completed and the contents of the reaction vessel were heated to approximately 80°C for approximately one hour, while the contents were continuously mixed. The reacted mixture was then transferred into a ten gallon container.
  • the desired resin blend product is an MEK based mixture containing an uncatalyzed resin blend (the adhesive) contains approximately 60% solids, where the solid portion is an exact mixture of 41.2% PT-30 cyanated phenolic resin, 39.5% RSL 1462 epoxy resin, 16.7% BC58 flame retardant, 1.5% Irganox 1010 stabilizer, and 1% bisphenol A co-catalyst, all percentages by weight.
  • the desired product is a mixture of the silica treated with a silane, the uncatalyzed resin blend, and MEK in which 68% by weight of the solids are silica, and the total solids are between 5% and 50% by weight of the mixture.
  • the exact solids concentration varies from run to run, and depends in part on the membrane to be impregnated.
  • the catalyst level is 10 ppm relative to the sum of the PT-30 and RSL1462.
  • mixtures I and II were determined to verify the accuracy of the precursors and compensate for any solvent flash that had occurred. Then mixture I was added to a ten gallon container to provide 12 pounds of solids, e.g., 515 solids content, 23.48 pounds of mixture I. Then mixture II was added to the container to provide 5.64 pounds of solids, e.g., 59.6% solids, 9.46 pounds of mixture II. the manganese catalyst solution (0.6% in mineral spirits), 3.45 grams, was added to the mixture of mixture I and mixture II and blended thoroughly to form a high solids content mixture.
  • the bath mixture for impregnating an ePTFE matrix 28% solids mixture, was prepared by adding sufficient MEK to the high solids content mixture to a total weight of 63 pounds. Thereafter, an ePTFE matrix was impregnated with this bath mixture to form a dielectric material.
  • a fine dispersion was prepared by mixing 26.8 grams Furnace Black (Special Schwarz 100, Degussa Corp., Ridgefield Park, New Jersey) and 79 grams of coupling agent (Dynaslan GLYMO CAS #2530-83-8; 3-glycidyloxypropyl-trimethoxysilane (Petrach Systems). The dispersion was subjected to ultrasonic agitation for 1 minute, then added to a stirring dispersion of 17.5 pounds SiO 2 (SO-E2) in 17.5 pounds MEK which had previously been ultrasonically agitated. The final dispersion was heated with constant overhead mixing for 1 hour at reflux, then allowed to cool to room temperature.
  • an adhesive varnish was prepared by adding the following: 3413 grams of a 57.5% (w/w) mixture of Primaset PT-30 in MEK, 2456 grams of a 76.8% (w/w/) mixture of RSL 1462 in MEK, 1495 grams of a 53.2% (w/w) solution of BC58 (Great Lakes, Inc.) in MEK, 200 grams of 23.9% (w/w) solution of bisphenol A (Aldrich Company) in MEK, 71.5 grams Irganox 1010, 3.21 grams of a 0.6% (w/w) solution of Mu HEX-CEM (OMG Ltd.) in mineral spirits, and 2.40 kg MEK.
  • An adhesive varnish was prepared by adding the following: 3413 grams of a 57.5% (w/w) solution of Primaset PT-30 (PMN P-88-1591)) in MEK, 2456 grams of a 76.8% (w/w) solution of RSL 1462 in MEK, 1495 grams of a 53.2% (w/w) solution of BC58 (Great Lakes, Inc.) in MEK, 200 grams of 23.9% (w/w) solution of bisphenol A (Aldrich Company) in MEK, 71.5 grams Irganox 1010, 3.21 grams of a 0.6% (w/w) solution of Mn HEX-CEM in mineral spirits, and 2.40 kg MEK. ln a separate container, 1328 grams of the adhesive varnish described above, 42.3 pounds MEK, 6.40 grams of Furnace Black (Special Schwarz 100, Degussa Corp., Ridgefield,

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Condensed Matter Physics & Semiconductors (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Hardware Design (AREA)
  • Microelectronics & Electronic Packaging (AREA)
  • Power Engineering (AREA)
  • Chemical & Material Sciences (AREA)
  • Materials Engineering (AREA)
  • Wire Bonding (AREA)
  • Structures Or Materials For Encapsulating Or Coating Semiconductor Devices Or Solid State Devices (AREA)

Abstract

Selon un procédé permettant de modifier le coefficient de dilatation thermique (CTE) dans au moins une zone d'un substrat stratifié, on enlève un premier matériau ayant un premier CTE d'une couche dudit substrat, de façon à former au moins un vide, puis on ajoute un second matériau ayant un second CTE au vide de la première couche.
PCT/US1997/019113 1996-11-08 1997-10-17 Utilisation d'une densite de perforation variable de la couche de cuivre pour commander le coefficient de dilatation thermique WO1998020549A1 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
AU47596/97A AU4759697A (en) 1996-11-08 1997-10-17 Use of variable perforation density in copper layer to control cte
JP10521465A JP2000503489A (ja) 1996-11-08 1997-10-17 熱膨張率をコントロールするための銅層に形成された可変的穿孔密度の使用

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
US74703296A 1996-11-08 1996-11-08
US08/747,032 1996-11-08

Publications (1)

Publication Number Publication Date
WO1998020549A1 true WO1998020549A1 (fr) 1998-05-14

Family

ID=25003387

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/US1997/019113 WO1998020549A1 (fr) 1996-11-08 1997-10-17 Utilisation d'une densite de perforation variable de la couche de cuivre pour commander le coefficient de dilatation thermique

Country Status (3)

Country Link
JP (1) JP2000503489A (fr)
AU (1) AU4759697A (fr)
WO (1) WO1998020549A1 (fr)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1901350A1 (fr) * 2005-07-07 2008-03-19 Kabushiki Kaisha Toyota Jidoshokki Dispositif dissipateur de chaleur et module de puissance
EP2975639A1 (fr) * 2014-07-18 2016-01-20 Thales Dispositif d'interface thermique avec joint microporeux capable d'empecher la migration de graisse thermique
US10559512B2 (en) 2015-11-16 2020-02-11 Hewlett-Packard Development Company, L.P. Circuit package

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE2826252A1 (de) * 1977-06-29 1979-01-04 Semi Alloys Inc Vorgefertigte, waermeuebertragende platteneinrichtung aus verbundmetall
DE2855494A1 (de) * 1978-12-22 1980-07-17 Bbc Brown Boveri & Cie Elektrisch isolierende waermeableitscheibe
DE3017514A1 (de) * 1979-05-08 1980-11-13 Norman Hascoe Vorgefertigte verbundmetall-waermeuebertragungseinheit
DE3144759A1 (de) * 1980-11-21 1982-06-24 General Electric Co., Schenectady, N.Y. "waermespannungen beseitigende bimetallplatte"
DE3314996A1 (de) * 1982-04-27 1983-10-27 Compagnie d'Informatique Militaire Spatiale et Aéronautique, 75008 Paris Zusammengesetztes substrat mit hoher waermeleitung und verwendung desselben fuer gehaeuse von halbleiter-schaltanordnungen
EP0365275A2 (fr) * 1988-10-21 1990-04-25 Texas Instruments Incorporated Membre refroidisseur d'un matériau composé pour un élément semi-conducteur et procédé pour sa fabrication
DE4100145A1 (de) * 1990-01-10 1991-07-11 Murata Manufacturing Co Substrat fuer die montage von integrierten schaltkreisen und es umfassendes elektronisches bauteil
EP0437656A1 (fr) * 1988-12-23 1991-07-24 Inco Limited Structure composée comprenant un coéfficient de dilatation thermique spécifique

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE2826252A1 (de) * 1977-06-29 1979-01-04 Semi Alloys Inc Vorgefertigte, waermeuebertragende platteneinrichtung aus verbundmetall
DE2855494A1 (de) * 1978-12-22 1980-07-17 Bbc Brown Boveri & Cie Elektrisch isolierende waermeableitscheibe
DE3017514A1 (de) * 1979-05-08 1980-11-13 Norman Hascoe Vorgefertigte verbundmetall-waermeuebertragungseinheit
DE3144759A1 (de) * 1980-11-21 1982-06-24 General Electric Co., Schenectady, N.Y. "waermespannungen beseitigende bimetallplatte"
DE3314996A1 (de) * 1982-04-27 1983-10-27 Compagnie d'Informatique Militaire Spatiale et Aéronautique, 75008 Paris Zusammengesetztes substrat mit hoher waermeleitung und verwendung desselben fuer gehaeuse von halbleiter-schaltanordnungen
EP0365275A2 (fr) * 1988-10-21 1990-04-25 Texas Instruments Incorporated Membre refroidisseur d'un matériau composé pour un élément semi-conducteur et procédé pour sa fabrication
EP0437656A1 (fr) * 1988-12-23 1991-07-24 Inco Limited Structure composée comprenant un coéfficient de dilatation thermique spécifique
DE4100145A1 (de) * 1990-01-10 1991-07-11 Murata Manufacturing Co Substrat fuer die montage von integrierten schaltkreisen und es umfassendes elektronisches bauteil

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP1901350A1 (fr) * 2005-07-07 2008-03-19 Kabushiki Kaisha Toyota Jidoshokki Dispositif dissipateur de chaleur et module de puissance
EP1901350A4 (fr) * 2005-07-07 2010-03-10 Toyota Jidoshokki Kk Dispositif dissipateur de chaleur et module de puissance
US8198539B2 (en) 2005-07-07 2012-06-12 Kabushiki Kaisha Toyota Jidoshokki Heat radiator and power module
EP2975639A1 (fr) * 2014-07-18 2016-01-20 Thales Dispositif d'interface thermique avec joint microporeux capable d'empecher la migration de graisse thermique
FR3023975A1 (fr) * 2014-07-18 2016-01-22 Thales Sa Dispositif d'interface thermique avec joint microporeux capable d'empecher la migration de graisse thermique
US10559512B2 (en) 2015-11-16 2020-02-11 Hewlett-Packard Development Company, L.P. Circuit package
US11183437B2 (en) 2015-11-16 2021-11-23 Hewlett-Packard Development Company, L.P. Circuit package

Also Published As

Publication number Publication date
AU4759697A (en) 1998-05-29
JP2000503489A (ja) 2000-03-21

Similar Documents

Publication Publication Date Title
US6127250A (en) Method of increasing package reliability by designing in plane CTE gradients
US6014317A (en) Chip package mounting structure for controlling warp of electronic assemblies due to thermal expansion effects
US5900312A (en) Integrated circuit chip package assembly
US5868887A (en) Method for minimizing warp and die stress in the production of an electronic assembly
US6183592B1 (en) Method for minimizing warp in the production of electronic assemblies
US5983974A (en) Method of making a lid for a chip/package system
US5970319A (en) Method for assembling an integrated circuit chip package having at least one semiconductor device
US6011697A (en) Constraining ring for use in electronic packaging
US6143401A (en) Electronic chip package
US5879787A (en) Method and apparatus for improving wireability in chip modules
WO1998020538A1 (fr) Traitement par espacements entre impulsions multiples pour ameliorer la formation des entrees des trous d'interconnexion a 355 nm
WO1998020549A1 (fr) Utilisation d'une densite de perforation variable de la couche de cuivre pour commander le coefficient de dilatation thermique
US5853517A (en) Method for coining solder balls on an electrical circuit package
WO1998019863A1 (fr) Procede et appareil d'alignement et de stratification de substrats sur des raidisseurs, dans des circuits electriques
WO1998020546A1 (fr) Cavites a haute tolerance dans des boitiers de puces
WO1998020713A1 (fr) Materiau de masquage de soudures de dimensions stables et procede d'application
WO1998020545A1 (fr) Procede de fabrication de cavites de grande tolerance dans des boitiers de puces

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): AL AM AT AU AZ BA BB BG BR BY CA CH CN CU CZ DE DK EE ES FI GB GE GH HU IL IS JP KE KG KP KR KZ LC LK LR LS LT LU LV MD MG MK MN MW MX NO NZ PL PT RO RU SD SE SG SI SK SL TJ TM TR TT UA UG UZ VN YU ZW

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): AT BE CH DE DK ES FI FR GB GR IE IT LU MC NL PT SE

121 Ep: the epo has been informed by wipo that ep was designated in this application
REG Reference to national code

Ref country code: DE

Ref legal event code: 8642

122 Ep: pct application non-entry in european phase
122 Ep: pct application non-entry in european phase