Composition with Thermally-Treated Silica Filler for Performance Enhancement

[0001 ] FtELD OF THE I NVENTION

[0002] This invention relates to compositions containing a curable resin and thermally treated silica In particular, it relates to compositions containing cyanate ester and/or epoxy resins and thermally treated silica

[0003] BACKGROUND OF THE INVENTION

[0004] Silica fillers are used in numerous chemical compositions in the fabrication of semiconductor packages and microelectronic devices, and especially in underfill compositions The semiconductor packages and microelectronic devices contain a large number of eiectrical circuit components that are electrically connected to each other and to a earner or substrate One method for making these interconnections uses polymeric or metallic solder that is applied in bumps to the component or substrate terminals The terminals are aligned and contacted together and the resulting assembly is heated to reflow the metallic or polymeric solder and solidify the connection

[0005] During its normal service life, the electronic assembly is subjected to cycles of elevated and lowered temperatures including solder joint and reflow processing and post thermal cycling testing Due to the differences in the coefficient of thermal expansion (CTE) for the electronic component, the interconnect material, and the substrate, this thermal cycling can stress or warp the components of the assembly and cause it to fail To prevent warpage, the gap between the component and the substrate is filled with a poiymeric encapsulant, an underfill, to reinforce the interconnect and balance the CTE mismatch, thus reducing warpage

[0006] The ability of the underfill to prevent warpage is dependent on the properties of the materials used to formulate the underfill composition The underfill composition comprises a resin system filled with non-conductive filler, with silica being the filler commonly used

[0007] Various resin systems are used in underfill compositions, and these include epoxies, bismaleimides, and acrylates, and it would be desirable also to use cyanate ester compounds because cyanate esters are capable of high cross link density High crosslink density provides low moisture pickup and a high glass transition temperature

(Tg). The high glass transition temperature, usually higher than 125°C, maintains polymer integrity during thermal cycling.

[0008] High cross link density, however, contributes to warpage because the material is not as flexible due to the crosslinking. Warpage can be reduced by using a high loading of filler, which also serves to reduce the coefficient of thermal expansion, enhance material strength, and modify rheology. The problem with high filler loading is that it leads to increased viscosity and reduced flow rate.

[0009] In one method of applying underfill, a measured amount of the underfill is dispensed along one or more peripheries of the electronic assembly and capillary action within the component-to-substrate gap draws the underfill inward. The substrate may be preheated if needed to achieve the desired level of encapsulant viscosity for the optimum capillary action. After the gap is filled, additional underfill encapsulant may be dispensed along the complete assembly periphery to help reduce stress concentrations and prolong the fatigue life of the assembled structure. This underfilling or wicking process is time consuming and dependent on the flow properties of the underfill material. This makes a high viscosity material in general unsuitable because it slows down capillary flow action.

[0010] Thus, there is a need for a filled curable composition, particularly using cyanate ester, that can be used as an underfill material with good capillary flow action and that still provides low warpage.

[0011] BRIEF DESCRIPTION OF THE FIGURES

[0012] Figure 1 is a graph of viscosity versus temperature for a series of surface treated silica materials in a polyoctylmethyl silicone fluid. Figure 2 is a chart showing the variations in activation energy for some different filler types and resins. Figure 3 is a graph showing the dependence of activation energy on the combination of resin and filler. Figure 4 is a graph illustrating that aluminum nitride particles can exhibit near zero activation energy above a critical surface coverage. Figure 5 is a graph showing that the activation energy depends on filler loading. Figures 6A and 6B present the meaning of the normalized activation energy parameter, Q, and of the normalized interfacial interaction parameter, F, in graphic format. Figure 7 shows the range of F values where near zero Q values were observed for resin/filler pairs.

[0013] SUMMARY OF THE INVENTION

[0014] This invention is a curable composition comprising a curable resin and a thermally-treated silica filler. In one embodiment, the curable composition comprises

(a) a thermally treated silica filler, (b) a curable resin (c) an initiator, and (d) optionally, adhesion promoters and/or wetting agents. The curable resins can be cyanate ester resins, epoxy resins, maleimide resins, or acrylate or methacrylate resins.

[0015] In one aspect of the invention, the curable resin is a cyanate ester resin, or an epoxy resin, or a combination of cyanate ester resin and epoxy resin, and a thermally- treated silica filler, with initiator and optionally, adhesion promoters and wetting agents. The heat treatment of the siiica significantly reduces the concentration of hydroxyl groups on the surface resulting in improved compatibility with the resins, and particularly with cyanate ester, improved flow behavior, reduction in CTE and enhancement of modulus.. Chip packages underfilled with resins and thermally treated silica have reduced warpage because of these property characteristics.

[0016] DETAILED DESCRIPTION OF THE INVENTION

[0017] Typically, silica filler is supplied by the manufacturer with the surface treated with silanes or containing a high concentration of hydroxyl groups. Some resins, for example, cyanate esters, are reactive with hydroxyl groups and the reaction releases volatile materials. When underfills cure, volatiles escape from the underfill composition creating voids, which can lead to eventual failure of the ultimate electronic device. Removal of the hydroxyl groups would correct this problem. In addition, the viscosity of a composition is determined by how filler and resin interact. By removing the hydroxyl groups, the surface energy of the silica is lowered, making the silica more compatible with the surface energy of particular resins, and especially cyanate esters, and consequently lowering the viscosity. This allows a higher loading of silica, which reduces warpage, without an increase in viscosity of the composition. Good compatibility between thermally treated silica and cyanate ester resins also affects the CTE and modulus of the composite and thus reduces warpage more significantly compared to untreated silica.

[0018] Cyanate esters suitable for use as underfill materials include those having the

generic structure which n is 1 or larger, and X is a hydrocarbon group. Exemplary X entities include, but are not limited to, bisphenol A, bisphenol F, bisphenoi S, bisphenol E, bisphenol O, phenol or cresol novolac, dicyclopentadiene, polybutadiene, polycarbonate, polyurethane, polyether, or polyester. Commercially available cyanate ester materials include; AROCY L-10, AROCY XU366, AROCY

XU371, AROCY XU378, XU71787.02L, and XU 71787.07L, available from Huntsman LLC; PRIMASET PT30, PRIMASET PT30 S75, PRIMASET PT60, PRIMASET PT60S, PRIMASET BADCY ₁ PRIMASET DA230S, PRIMASET MethylCy, and PRIMASET LECY, available from Lonza Group Limited; 2-allypheπol cyanate ester, 4- methoxyphenol cyanate ester, 2,2-bis(4-cyanatophenol)-1 ,1 ,1 ,3,3,3-hexafluoropropane, bispheπol A cyanate ester, diallylbisphenol A cyanate ester, 4-phenylpheno] cyanate ester, 1,1,1-tris(4-cyanatophenyl)ethane, 4-cumylphenol cyanate ester, 1,1-bis(4- cyanato-phenyl)ethane, 2,2, 3,4,4,5,5, 6,6,7,7-dodecafluoro-octanediol dicyanate ester, and 4,4'-bisphenol cyanate ester, available from Oakwood Products, Inc.

[0019] Other suitable cyanate esters include cyanate esters having the structure:

[0020] in which R ¹ to R ⁴ independently are hydrogen,

C ₁ - C ₁₀ alkyl, C ₃ -C ₈ cycloalkyi, C ₁ -C ₁₀ alkoxy, halogen, phenyl, phenoxy, and partially or fully fluorinated alkyl or aryl groups (an example is phenylene-1,3-dicyanate);

[0021] cyanate esters having the structure: in which R ¹ to R ⁵ independently are hydrogen, C ₁ - Cio alkyl, C ₃ -C ₈ cycloalkyi, C ₁ -C ₁₀ aikoxy, halogen, phenyl, phenoxy, and partially or fully fluorinated alkyl or aryl groups;

[0022] cyanate esters having the structure:

in which R ¹ to R ⁴ independently are hydrogen, C ₁ - C ₁₀ alkyl, C ₃ -C ₈ cycloalkyi, C ₁ -C ₁₀ alkoxy, halogen, phenyi, phenoxy, and partially or fully fluorinated alkyl or aryl groups; Z is a chemical bond or SO ₂ , CF ₂ , CH ₂ , CHF ₁ CHCH ₃ , isopropyl, hexafluoroisopropyl, C ₁ - C ₁₀ alkyl, O, N=N, R ⁸ C=CR ⁸ (in which R ⁸ is H, C ₁ to C ₁₀ alkyl, or an aryl group), R ⁸ COO, R ⁸ C=N, R ^S C=N-C(R ⁸ )=N, C ₁ -C ₁₀ alkoxy, S, Si(CH ₃ ) ₂ or one of the following structures:

[0024] (an example is 4,4' ethylidenebispheπyleπe cyanate having the commercial name AroCy L- 10 from Vantico),

[0025] cyanate esters having the structure

[0026] in which and n is a number from 0 to 20, R ⁶ is hydrogen or d- C ₁₀ alkyl and X is CH ₂ or one of the following structures

(examples include XU366 and XU71787 07, commercial products from Vantico),

[0028] cyanate esters having the structure N≡C-O-R -O~C≡N _and

[0029] cyanate esters having the structure N=C-O-R _ _m which R ⁷ ιs a non- aromatic hydrocarbon chain with 3 to 12 carbon atoms, which hydrocarbon chain may be optionally partially or fully fluorinated

[0030] Suitable epoxy resins include bisphenol, naphthalene, and aliphatic type epoxies Commercially available materials include bisphenol type epoxy resins (EPICLON 830LVP ₁ 830CRP, 835LV, 850CRP) available from Dainippoπ Ink & Chemicals, lnc , naphthalene type epoxy (EPICLON HP4032) available from Dainippon Ink & Chemicals, Inc., aliphatic epoxy resins (ARALDITE CY179, 184, 192, 175, 179) available from Ciba Specialty Chemicals, (Epoxy 1234, 249, 206) available from Dow Corporation, and (EHPE-3150) available from Daicel Chemical Industries, Ltd

[0031] Other suitable epoxy resins include cycloaliphatic epoxy resins, btsphenol-A type epoxy resins, bisphenol-F type epoxy resins, epoxy novolac resins, biphenyl type epoxy resins, naphthalene type epoxy resins, dicyclopentadienephenol type epoxy resins

[0032] Suitable maleimide resins include those having the generic structure

which π is 1 to 3 and X ¹ is an aliphatic or aromatic group. Exemplary X ¹ entities include, poly(butadienes), poly(carbonates), poly(urethanes), poly(ethers), poly(esters), simple hydrocarbons, and simple hydrocarbons containing functionalities such as carbonyl, carboxyl, amide, carbamate, urea, ester, or ether. These types of resins are commercially available and can be obtained, for example, from Dainippon Ink and Chemical, Inc.

[0033] Additional suitable maleimide resins include, but are not limited to, solid aromatic bismaieimide (BMI) resins, particularly those having the structure

[0034] in which Q is an aromatic group. Exemplary aromatic groups include:

[0044] in which n is 1 - 3,

[0046] Bismaleimide resins having these Q bridging groups are commercially available, and can be obtained, for example, from Sartomer (USA) or HOS-Technic GmbH (Austria).

[0047] Other suitable maleimide resins include the following:

which C ₃₆ represents a linear or branched hydrocarbon chain (with or without cyclic moieties) of 36 carbon atoms;

[0050]

[0052] Suitable acrylate and methacrylate resins include those having the generic

or aliphatic group. Exemplary X ² entities include poly(butadienes), poly-(carbonates), poly(urethanes), poly(ethers), poly(esters), simple hydrocarbons, and simple hydrocarbons containing functionalities such as carbonyl, carboxyl, amide, carbamate, urea, ester, or ether. Commercially available materials include butyl (meth)acrylate, isobutyl (meth)acrylate, 2-ethyl hexy! (meth)acrylate, isodecyl (meth)acrylate, n-lauryl (meth)acrylate, alkyl (meth)-acrylate, tridecyl(meth)-acrylate, n-stearyl (meth)acrylate, cyclohexyl(meth)-acrylate, tetrahydrofurfuryl-(meth)acrylate, 2-phenoxy ethyl(meth)- acrylate, isobornyl(meth)acrylate, 1 ,4-butanediol di(meth)acry]ate, 1 ,6- hexanedio! di(meth)acrylate, 1 ,9-nonandiol di(meth)acrylate, perfluorooctylethyl (meth)acrylate _T 1 ,10 decandiol di(meth)-acrylate, nonylphenol polypropoxylate (meth)acrylate, and polypentoxylate tetrahydrofurfury] acrylate, available from Kyoeisha Chemical Co., LTD; polybutadiene urethane dim ethacry late (CN302, NTX6513) and polybutadiene dimethacrylate (CN301 , NTX6039, PRO6270) available from Sartomer Company, Inc; polycarbonate urethane diacrylate (ArtResin UN9200A) available from Negami Chemical Industries Co., LTD; acrylated aliphatic urethane oligomers (Ebecryl 230, 264, 265, 270,284, 4830, 4833, 4834, 4835, 4866, 4881 , 4883, 8402, 8800-20R, 8803, 8804) available from Radcure Specialities, Inc; polyester acrylate oligomers (Ebecryl 657, 770, 810, 830, 1657, 1810, 1830) available from Radcure Specialities, Inc.; and epoxy acrylate resins (CN 104, 111 , 112, 115, 116, 117, 118, 119, 120, 124, 136) available from Sartomer Company, Inc. In one embodiment the acrylate resins are selected from the group consisting of isobornyl acrylate, isobornyl methacrylate, lauryl acrylate, lauryl methacrylate, poly(butadiene) with acrylate functionality and poly(butadiene) with methacrylate functionality.

[0053] Suitable silica particles are spherical with average particle sizes of from 0.1 μ, to 10μ and can be obtained, for example, from Admatechs as products SOE5 and SOE2. In one embodiment, mixtures of sizes can be used for increasing the total filler loading level and at the same time keeping the viscosity at a manageable level.

[0054] The thermal treatment procedure is conducted by placing aliquots, typically 10g to 50g, of silica powder as received from the supplier into ceramic crucibles and heating them in an oven preheated to a temperature or a range of temperatures between 400°C and 450°C for a period of 72 hours. A suitable furnace is a Thermolyne F-A1740 muffle furnace. The crucibles are removed from the furnace and allowed to cool to 80°C under dry air or nitrogen. The treated silica is transferred hot (80 ⁰ C) from crucibles to polymer based storage containers with an air tight seal (for example, 250ml amber Nalgene bottles) under a dry atmosphere. Containers filled with treated silica can be stored in evacuated packaging (such as food-saver bags) for additional protection from ambient moisture and long term storage.

[0055] EXAMPLES

[0056] EXAMPLE 1. FLOW RATE.

[0057] Compositions were prepared and tested for flow capability using a resin composition and silica filler treated as described in the detailed description. The resin composition consisted of 60% by weight of cyan ate ester (obtained as PRIMASET

LECY from Lonza LTD) and 40% by weight Bis-F-epoxy (obtained as EPON 828

Resolution Performance Products). The silica filler had a maximum particle size of 5μ

(obtained as SE5050 silica from Admatechs) and was loaded into the resin system at

40% and 70% by weight. A control had no filler.

[0058] Filled samples were prepared by combining the resin and filler in polypropylene cups with 3mm glass or zirconia milling media. The cups were placed into a Flaktek Speed Mixer and processed in three sequential 30 second mixing cycles at 1800 RPM, 2100 RPM and 2700 RPM respectively.

[0059] The flow velocity of the filled cyanate ester/epoxy resin compositions were measured with a custom designed instrumentation consisting of a rectangular capillary channel, a temperature controlled flow chamber, and an optical configuration and digital imaging system for flow front measurement. The channel is constructed of an upper and lower substrate bonded together along their lengths on both sides and separated by some distance. Results reported here were obtained using two glass slides (75x25 mm) and a separation of 50μ.

[0060] The temperature controlled flow chamber is designed to hold the channel described above and provide a uniform temperature across the entire sampie and test vehicle. Flow behavior in the range of 20 ⁰ C to 280°C can be investigated with this system. One end of the chamber allows access to the beginning of the channel so that the sample can be introduced. Samples were introduced via syringe, and 0.05 to 0.1.2 ml of material was deposited. The top of the chamber is designed so that the sample can be viewed as it flows down the channel. The flow of the material in the channel is captured by a digital image acquisition system. Flow front position, velocity and estimated fill-time are obtained from digital image analysis.

[0061 ] The results are reported in Table 1 and show the flow velocity of the un-filled and filled cyanate ester/epoxy resin at 90°C determined using the above described instrumentation and method.

Table 1.

[0062] The flow rate shows minimal dependence on the filler treatment at low filler loading levels. However, as the concentration increases past the percolation point, approximately 35% by weight, the filler surface properties begin to contribute significantly to the viscosity and therefore the flow rate. To meet thermal expansion and modulus performance requirements silica needs to be present at a level of least 40%. The table above shows at least a five-fold improvement in the flow rate at 70% filler loading using the thermally treated filler. The thermally treated filler can enable high loading without compromising flow rate in the cyanate ester/epoxy system.

[0063] Example 2. Warpage

[0064] Warpage is measured as the height difference between the highest point and the lowest point of the underfilled package, relative to the substrate. Compositions were prepared and tested for warpage using a resin system and thermally treated silica filler as described in the detailed description. The resin formulations consisted of 55.8% by weight of cyanate ester (obtained as PRIMASET LECY from Lonza Ltd.), 37.2% by weight of Bis F-epoxy resin (prepared in-house by National Starch and Chemical Company), and 7% by weight of 3,3' diamino diphenyl sulfone (obtained from Aldrich).

The resin components were mixed together before the filler was added. The silica filler had a maximum particle size of 5μ (obtained as SOE2from Admatechs) and was loaded into the resin system at 40% and 60% by weight. The resin and fillers were mixed by a high speed centrifuge mixer at 3000 rpm for five minutes and then degassed to remove trapped air in the formulations.

[0065] The warpage tests were conducted using flip chip test vehicles. The die was 15X15mm with lead free solders and the substrate was a BT laminate 42X42mm. The die and BT substrate were jointed together through a standard lead free reflow process. Prior to underfilling, test vehicles were heated in a165°C oven for two hours to remove absorbed moisture. The test vehicles were then kept at 110 ⁰ C during underfilling. The underfill formulations were dispensed from a syringe with 0.33mm needle by hand only along one side of the die. The filled packages were then cured at 165°C for 90 minutes.

[0066] The warpage of the test vehicles were measured using a laser profilometer (Cobra 3D, Optical Gaging Product) on the backside of the substrate along the two diagonal directions. The average height of the two curves was calculated as the warpage of the package. The warpage measurements were taken before underfilling and immediately after the cured packages being cooled to 25°C. The warpage increase after cure was used to compare the effect of filler on the warpage. The results are reported in Table 2.

Table 2.

Filler loading Warpage increase of Warpage increase of % change SOE2-MRCE (micron) treated.SOE2-MRCE

(micron)

40%wl 23.2±3.0 19.7±2.0 -15%

60%wt 38.6±4.2 28.1 ±2.0 -27%

[0067] The results show that using thermally treated silica, the warpage of the composition used in underfilled packages decreased. The reduction increases as the amount of filler used increases. At 60%wt loading, the warpage was reduced by 27% after cure.

[0068] DISCUSSION OF PREDICTION OF VISCOSITY CHANGES WITH TEMPERATURE. [0069] The rate at which a material changes its viscosity influences performance, such as, bleed out, sag, and flow. This effect has practical application in the design of capillary underfill in electronics packaging, printable inks, and adhesive materials, for which viscosity change can be either a wanted or unwanted characteristic. Therefore, it would be a benefit to be able to predict the viscosity behavior of a filled resin system.

[0070] Filled resin systems do not all react to temperature at the same rate as the unfilled resin. Some filled resins thin at a much lower rate than the unfilled resin and even approach constant viscosity with temperature. Figure 1 shows a graph of viscosity versus temperature for a series of surface treated silica materials in a polyoctylm ethyl silicone (POMS) fluid. The surface treatments were conducted with phenyl trialkoxy silane (phenyl in the Figure), phenyl amino trialkoxy silane (phenyl amino), and dimethyl silicone oligomer end terminated with a trialkoxy sϋane group (silicone). The filler size and loading ievel are identical for this series. The graph shows that with increasing temperature, the viscosity decreases, but at different rates for the different fillers. [0071] With respect to the effect of temperature on viscosity, the behavior of filled resin systems is an energy activated process. The activation energy can be experimentally determined from the slope of the line resulting from a plot of the logarithm of viscosity versus reciprocal temperature. POMS for example has an activation energy (Ea) of 2000K at a shear rate of 1 rad/s while a typical Bis-F epoxy is 9190K at the same rate. [0072] Activation energies (Ea) were determined for silica and alumina fillers with a series of surface treatments in epoxy, poly(phenylmethyl)silicone (PPMS) and POMS. The variations in activation energy for some different filler types and resins are shown in Figure 2 (activation energy values obtained from a plot of the natural log of the viscosity versus MT). A wide range of behaviors are exhibited; some systems mimic the activation energy of the unfilled resin; some remain invariant with temperature; and others thicken slightly with temperature.

[0073] In another example of the dependence of activation energy on the combination of resin and filler, a cyanate ester resin (PRIMASET L-10) was loaded with silica having the following surface modifications: treated with hexamethyldisilazaπe (HMDZ), treated with epoxy silane, thermally treated (calcinated), and untreated. The untreated, epoxy silane treated and calcinated filler materials all thin with temperature, but the HMDZ treated material initially thins and then thickens with temperature. See Figure 3. This sort of behavior is not easily anticipated or even expected.

[0074] The temperature dependence and independence of filled resins is a general phenomenon dependent on the interfacial interactions of the resin and filler particles. A series of surface treated aluminum nitride materials containing various levels of hydrophobic nano-silica (prepared via a dry coating method called magnetically assisted impact coating, MAIC) were dispersed into an epoxy resin. Figure 4 illustrates that these nano-silica treated aluminum nitride particles exhibit near zero activation energy above a critical surface coverage. This indicates that activation energy, and consequently

viscosity behavior relative to temperature, can be manipulated by altering the level of surface modification.

[0075] In another example, various filler loadings were made to a filled resin system containing silicone filler (TREFILL E-600) in epoxy resin (EPON 826). The results are shown in Figure 5, namely, that the activation energy (Ea) depends on the filler loading, progressing from the activation energy of the polymer toward zero as the loading is increased.

[0076] Relating the surface properties of the resin and filler to the change in viscosity due to temperature can be approached by defining a normalized activation energy (Q), in which Q is the ratio of the activation energy of a filled resin to an unfilled resin, and by defining a normalized interfacial interaction (F) ₁ in which F is the ratio of the Work of Adhesion (W _sL ) to the total surface energy of the filled system. That is, the normalized

activation energy parameter is and the normalized interfacial

interaction parameter is ^{v / L J} . The Work of Adhesion is calculated as the surface tension of the resin times (1 plus the cos of the contact angle between the filler particle and the resin). The two parameters F and Q normalize the rheology and interactions to the unfilled resin properties. This allows for direct comparison between resin systems allowing isolation of the resin-filler interfacial effects. Figures 6A and 6B present the meaning of F and Q in graphical representation. [0077] In practice, engineering this effect can have benefit for fillet shape control or resin-bleed out control in underfill applications in electronic packaging. An underfill with a near zero (Ea) will not thin and spread as the temperature of a device is raised in a reflow or curing oven. In capillary underfill applications a material is dispensed at room temperature and is required to thin and flow under a die as the temperature is raised. Resin and filler combinations need to be selected so that the filled resin thins with temperature sufficient enough to flow, but not to the level where it bleeds out from the location where it should reside.

[0078] In a further example, two underfill formulations (JKL and MNO) were prepared containing POMS as the resin and silica as the filler in an amount of 35% by volume. Formulation JKL silica was surface treated with silicone and MNO silica was surface treated with phenyl silane. The activation energy is positive for JKL and near zero for MNO. These formulations were deposited between a silicon wafer and a glass substrate, and equilibrated at room temperature for several hours. Both formulations

filled the volume under the wafer completely before heating and had a similar fillet shape and width. The formulations were placed on a hot plate with a surface temperature of 12O ⁰ C for V-i hour, after which it was observed that the MNO formulation with the near- zero Ea (activation energy) remained under the wafer while JKL bled out. The Ea determined from the Theological behavior clearly manifested itself in terms of the electronic materials relevant performance characteristic of bleed-out control. [0079] The interfacial & Theological properties of the filled POMS resins had the following characteristics:

[0080] The range of F values where near zero Q values were observed were mapped out to provide insight into the types of resin filler pairs that exhibit this behavior. See Figure 7. Two zones have been annotated in the figure where the near-zero behavior has been experimentally observed. One zone exists for F values greater than 2.5 and the other for F less than 1.5. The high-F zone indicates that the work of adhesion is sufficiently greater than the liquid surface tension. This can occur when the filler is higher in energy than the resin. When F=2 the contact angle is equal to zero. The high- F zone corresponds to excellent wetting conditions. The low-F zone occurs when the work of adhesion decreases and formation of a filler-resin interface is comparably favorable to the resin maintaining an interface with itself. This situation occurs when the filler is lower in energy than the liquid phase. The filler is simply not energetic enough to completely interrupt the liquid interfacial behavior.

[0081 ] Many of the silicone-filler pairs fail into the high-F zone due to the low energy of the silicone resin materials (~25 mN/m). Formulation MNO is an example of such a

system, low energy silicone high energy silica (~50 mN/m). Conversely, the high energy cyanate ester resin (~45 mN/m) in combination with low energy filler treatment such as HMDZ (-20 mN/m) falls into the low-F zone. Epoxy, propylene glycol and aliphatic BMI materials fall in between these two zones and are therefore predicted to thin with temperature. Thus, the inventors discovered that the data correlation of F and near zero Q can be used to help predict and design resin/filler systems that should exhibit the near-zero behavior. For example, the Q versus loading level was mapped for a mafeimide resin and a cydoaSiphatic epoxy resin as a function of filler ioading and surface treatment, and the correlation predicts that these two materials should thin with temperature for all loading levels.