Dielectric & Electric Epoxy Compound Traits — Dexter Technical Paper (1992)

Dr. H. W. Rauhut

Research Sci­en­tist, Dex­ter Elec­tron­ic Mate­ri­als Division

Tech­ni­cal Paper, April 1996

A SPECIAL POSITION

Epoxy mold­ing com­pounds hold a spe­cial posi­tion in elec­tri­cal and elec­tron­ic appli­ca­tions, pri­mar­i­ly because of their unusu­al­ly low melt vis­cos­i­ty in com­bi­na­tion with excel­lent mechan­i­cal and elec­tri­cal prop­er­ties. Besides elec­tri­cal insu­la­tion, they fea­ture elec­tron­ic com­pat­i­bil­i­ty. They are eas­i­ly trans­fer mold­ed at low pres­sure, typ­i­cal­ly at 1 Kpsi (70.3 Kg/sqcm) and 160–180°C. There­fore, their char­ter has been the encap­su­la­tion of elec­tri­cal and elec­tron­ic com­po­nents. Among oth­er ther­mosets, sil­i­cones still hold a niche in elec­tron­ic encap­su­la­tion. In the ear­ly his­to­ry of tran­sis­tor encap­su­la­tion, phe­no­lics were also com­mer­cial­ly used. But in gen­er­al, phe­no­lics and oth­er ther­mosets require trans­fer pres­sures at or above 10 Kpsi. One can imag­ine what such pres­sure would do to wire sweep in inte­grat­ed cir­cuits.

Con­sid­er­ing ther­mo­plas­tics, they are most­ly processed by injec­tion mold­ing at 200–375°C and pres­sures of 8–30 Kpsi. A few allow pro­cess­ing at 180–300°C, with trans­fer mold­ing at high pres­sure. There have been injec­tion over­mold­ing appli­ca­tions with ther­mo­plas­tics, where tran­sis­tor wire bonds were pro­tect­ed by an epoxy pow­der coat­ing, and of course ther­mo­plas­tics have served many elec­tri­cal appli­ca­tions oth­er than encap­su­la­tion. In the inor­gan­ic field, we find that high­ per­for­mance ceram­ic pack­ages for elec­tron­ic com­po­nents are very expen­sive. Thus, for the best bal­ance of eco­nom­ics and per­for­mance, the plas­tics pack­age is syn­ony­mous with epoxies.

Elec­tri­cal per­for­mance con­tributed to the epox­ies’ com­mer­cial suc­cess. The recent decade has seen a major pro­lif­er­a­tion of elec­tron­ic device types, com­bined with greater qual­i­ty aware­ness and expec­ta­tion. Accord­ing­ly, epoxy com­po­si­tions have changed, too. It is in order to look again at elec­tri­cal data under var­ied con­di­tions, even to revis­it elec­tri­cal ASTM test results. In addi­tion, dielec­tric analy­ses (DEA) have helped shed new light into epoxy com­pounds and applications.

COMPOUND COMPOSITION

Basi­cal­ly, epoxy mold­ing com­pounds con­sist of organ­ic and inor­gan­ic mate­ri­als. Most com­pounds are fused sil­i­ca (FS) filled. Some con­tain crys­talline sil­i­ca or oth­er filler selec­tions. By hard­en­er, epox­ies can be divid­ed into phe­no­lic novolac (PN) and anhy­dride cured com­pounds, the lat­ter for spe­cial, rel­a­tive­ly low­er vol­ume appli­ca­tions. A general

Descrip­tion of epoxy com­pounds is shown in Fig­ure 1, with fillers, epoxy resin and phe­no­lic hard­en­er as major weight (and vol­ume) frac­tions. These and oth­er mate­ri­als are main fac­tors in elec­tri­cal insu­la­tion and elec­tron­ic performance.

Epoxy com­pounds have been 70- 74 % FS filled. New­er types con­tain a high­er (up to 85%) fused sil­i­ca lev­el, and/or spe­cial epoxy resins (1). Most com­pounds are still based on ECN (Epoxy Creso­lic Novolac). For some advanced appli­ca­tions, biphenyl epoxy, Tris-epoxy, or oth­er epoxy resins are employed. Our per­for­mance study below focus­es on high­ly filled fused sil­i­ca sys­tems with var­ied resins, all phe­no­lic novolac cured and postcured 4 hours @ 175°C, unless indi­cat­ed otherwise.

ELECTRICAL AND DIELECTRIC PERFORMANCE BY ASTM TESTS

Elec­tri­cal insu­la­tion fea­tures may be described by such diverse para­me­ters as Dielec­tric Strength (DS), Arc Resis­tance (AR), Vol­ume (VR) and Sur­face Resis­tiv­i­ty (SR), Dielec­tric Con­stant (K) and Dis­si­pa­tion Fac­tor (DF), ref­er­ence (2).

Dielectric Strength and Arc Resistance

The high-volt­age dielec­tric strength test fol­lows ASTM D149, by our test­ing at a thick­ness of about 20 mils. Test results sig­nif­i­cant­ly decrease with increas­ing spec­i­men thick­ness, for instance by about 70% at 200 mils (5 mm). OS per mil at any thick­ness can be cal­cu­lat­ed with rea­son­able accu­ra­cy from a mea­sured OS val­ue and spec­i­men thick­ness, accord­ing to ref­er­ence (3) and the fol­low­ing formula:

In oth­er words, cal­cu­lat­ed OS (y) equals the mea­sured DS, times the square root of its spec­i­men thick­ness, divid­ed by the square root of a dif­fer­ent thick­ness T (y) in mils. DS decreas­es with increas­ing tem­per­a­ture (for instance by 25% between 25 and 150°C). It also decreas­es by chem­i­cal, phys­i­cal, long-time elec­tri­cal stress, and spec­i­men voids.

Fig­ure 2 shows dielec­tric strength data @ 23°C on var­i­ous ECN com­pounds over a spread of 15% filler, up to about 85% FS. There is an increas­ing OS trend with increas­ing FS lev­el, which appears rea­son­able con­sid­er­ing FS’s very high dielec­tric strength of 20 KVolts per mil. Fig­ure 2 also shows a sim­i­lar arc resis­tance trend line. AR is mea­sured by ASTM 0495. Evi­dent­ly, an increased FS lev­el does not hurt an epoxy’s resis­tance to a high­ volt­age, low-cur­rent arc, which may form a con­duct­ing sur­face path.

Volume and Surface Resistivity

Vol­ume resis­tiv­i­ty is an epoxy’s most impor­tant insu­lat­ing char­ac­ter­is­tic. Both VR and SR are mea­sured by ASTM 0257, at 500 Volts by our test­ing. VR and SR at room tem­per­a­ture and 100 Volts tend to be slight­ly low­er than man 500 V data.

Fig­ure 3 shows sim­i­lar vol­ume resis­tiv­i­ties for var­ied resin sys­tems after post-cure 6 hours @ 175°C. At 100°c, VR is typ­i­cal­ly one decade low­er than at RT. The very high VR of FS filler con­tributes to VR data of high­ly filled com­pounds. ECN resin sys­tems at RT with­out post-cure showed VR data up to one decade low­er than with postcure, and also low­er SR data. No sig­nif­i­cant VR and SR changes were observed after extend­ed cure times (up to 3 min­utes) in the mold. Through mois­ture expo­sure, VR of the high­ly filled ECN and Biphenyl resin com­pounds declined less than that of pre­vi­ous ref­er­ence com­pounds, Fig­ure 4. An anhy­dride-cured com­pound’s VR showed the great­est mois­ture sensitivity.

Room tem­per­a­ture sur­face resis­tiv­i­ty of epox­ies is 10–100 x 10′ Ohms. Typ­i­cal SR100 data ver­sus tem­per­a­ture is shown in Fig­ure 5.

Dielectric constant (K) and Dissipation Factor DF (DF)

These para­me­ters are mea­sured accord­ing to ASTM 0150. The dielec­tric con­stant or per­mit­tiv­i­ty reflects the capac­i­tive nature of an insu­la­tor. A mate­r­i­al between the pow­er and sig­nal lay­ers of a print­ed cir­cuit board con­trols the cir­cuit speed by its dielec­tric con­stant (4). In gen­er­al, an elec­tron­ic design engi­neer prefers low dielec­tric con­stants (e.g., in the case of epoxy com­pounds, room tem­per­a­ture K‑values of about 4 @ 1 KHz), and also low dis­si­pa­tion fac­tors (e.g., OF @ 1 KHz and RT of 0.002–0.003). DF is a key insu­la­tion para­me­ter, like vol­ume resis­tiv­i­ty and dielec­tric strength. In an alter­nat­ing elec­tri­cal field, ener­gy is lost as heat with increas­ing DF. In our study, var­i­ous resin sys­tems showed sim­i­lar­ly low K, Fig­ure 6, and sim­i­lar DF with­in the range of pre­vi­ous ECN ref­er­ence com­pounds, Fig­ure 7.

Frequency and Temperature Effects on K and DF

Epoxy dielec­tric con­stants slight­ly decrease with increas­ing fre­quen­cy and with decreas­ing tem­per­a­ture. K and OF changes are caused by dielec­tric polar­iza­tion in the mate­r­i­al. From high­est to low­est fre­quen­cies, atomic/ elec­tron­ic, then mol­e­c­u­lar dipole, then mate­r­i­al inho­mo­gene­ity polar­iza­tions add up to a maxi­mum K at the low­est fre­quen­cy. Typ­i­cal K data by ASTM 0150 is exhib­it­ed in Fig­ure 8, and will be reviewed again below on the basis of dielec­tric analy­ses. Typ­i­cal OF data is shown in Fig­ure 9. Both K and OF assume high­er val­ues with­out postcure, as indi­cat­ed in Fig­ures 8–9 @ 150°C.

Batch-to-Batch Comparison of K and DF

Although epoxy com­pounds are made by batch process­es, they dis­play sim­i­lar batch-to-batch dielec­tric con­stants and dis­si­pa­tion fac­tors, Fig­ures 10–11.

Material Effects on K and DF

Com­pound raw mate­ri­als influ­ence K and DF (5). Shown here are only com­par­isons of fused and crys­talline sil­i­ca (CS), with FS pro­vid­ing low­er dielec­tric con­stants (K) and gen­er­al­ly some­what low­er dis­si­pa­tion fac­tors (DF) than CS, depen­dent on fre­quen­cy, Fig­ures 12–13.

Since fused sil­i­ca plays such a dom­i­nat­ing role in micro­elec­tron­ic com­pounds, its K and DF val­ues are reviewed in Fig­ure 14. Over a wide range, up to giga­hertz fre­quen­cies, fused sil­i­ca’s dielec­tric data remain low and will not cause a rise of K or DF at very high frequencies.

Moisture Effects on K and DF

Both dielec­tric con­stants and dis­si­pa­tion fac­tors increase sig­nif­i­cant­ly with absorbed mois­ture, Fig­ures 15–16. Fused sil­i­ca, with its inher­ent­ly low K, DF and mois­ture absorp­tion, shows in its com­pounds small­er K and DF increas­es than crys­talline sil­i­ca sys­tems. But mois­ture effects on K and DF of FS-filled com­pounds are still sig­nif­i­cant. One prac­ti­cal con­se­quence relates to dielec­tric pre­heat­ing if the mois­ture con­tent of micro­elec­tron­ic epoxy pow­der is not con­trolled. Because of an increased dis­si­pa­tion fac­tor, gen­er­at­ed heat may be high­er than desired for opti­mum molding.

DIELECTRIC PERFORMANCE BY DEA TEST

Dielec­tric analy­sis (DEA) mea­sures capac­i­tance and con­duc­tance ver­sus time, tem­per­a­ture and fre­quen­cy. The results are cor­re­lat­ed to chem­istry, rhe­ol­o­gy and mobil­i­ty of ions or dipoles in a poly­mer sys­tem. DEA gen­er­ates dielec­tric data and their tem­per­a­ture and fre­quen­cy depen­den­cies in a par­tic­u­lar­ly effi­cient way, includ­ing dielec­tric con­stant (gen­er­al­ly called per­mit­tiv­i­ty), loss fac­tor (loss index), dis­si­pa­tion fac­tor (Tan Delta), and ion­ic con­duc­tiv­i­ty. Data below was gen­er­at­ed on a DuPont 2970 DEA ana­lyz­er using a ceram­ic par­al­lel plate sen­sor and is an exten­sion of pre­vi­ous­ly pre­sent­ed DEA infor­ma­tion (6).

Dielectric Constant, Loss and Dissipation Factors

Fig­ures 17–20 show data of a high­ly FS-filled EC /PN com­pound. As the tem­per­a­ture increas­es, the var­i­ous dielec­tric respons­es increase at a giv­en fre­quen­cy. In an alter­nat­ing elec­tri­cal field @ 0.3 Hz, dipoles can fol­low the field and thus con­tribute to very high dielec­tric con­stants (K) at ele­vat­ed tem­per­a­tures, shown in Fig­ure 17 only up to 130°C At high­er fre­quen­cies, the K‑versus­ tem­per­a­ture data tend to straight­en out. How­ev­er, on scru­ti­niz­ing 10–100 KHZ data in Fig­ure 18, one observes a K‑peak tem­per­a­ture around 220°C, due to the system’s rhe­o­log­i­cal and dielec­tric cor­re­la­tion. Fig­ures 19–20 show tem­per­a­ture and fre­quen­cy effects on the same sys­tem’s loss and dis­si­pa­tion fac­tors. Obvi­ous­ly, the loss fac­tor becomes numer­i­cal­ly sim­i­lar to the dis­si­pa­tion fac­tor (DF) at high­er fre­quen­cies, par­tic­u­lar­ly in the low tem­per­a­ture region.

Dielectric Data by DEA versus ASTM

Fig­ure 21–22 allow a com­par­i­son of dielec­tric DEA and ASTM data on the com­pound described above. Although just based on avail­able mea­sure­ments, gen­er­al­ly a close cor­re­la­tion was found at low test tem­per­a­tures, which sug­gests that DEA may be con­sid­ered as an alter­nate test method in place of ASTM D150. Fig­ure 23 shows a pro­jec­tion of dielec­tric con­stants by DEA into high­er fre­quen­cy regions, since then con­firmed by oth­er means.

DEA Ionic Conductivity 

The dielec­tric loss fac­tor con­sists of two com­po­nents, one due to ion­ic con­duc­tiv­i­ty, the oth­er one due to dipole ori­en­ta­tion in the epoxy. At low fre­quen­cies, the loss fac­tor is dom­i­nat­ed by ion­ic con­duc­tiv­i­ty. We mea­sured @ 0.3 Hz. Below 150°C, ion­ic con­duc­tiv­i­ties are very low, even in terms of minute pmho/cm units. There is a sig­nif­i­cant crosslink effect on ion­ic con­duc­tiv­i­ty (6), shown by postcure ver­sus no postcure data in Fig­ure 24, since ion mobil­i­ty is hin­dered in more vis­cous or tight­ly crosslinked (postcured) struc­tures. For the same rea­son, high­er glass tran­si­tion tem­per­a­ture (TG) means also low­er ion­ic con­duc­tiv­i­ty, fig­ure 25.

Dielectric Material Effects

DEA data also reflects mate­r­i­al effects of epoxy com­po­si­tions, as dis­cussed ear­li­er (6). Fig­ures 26–27 show oth­er exam­ples of an addi­tive, which helps decrease the dielec­tric con­stant, while slight­ly increas­ing the ion­ic con­duc­tiv­i­ty of an epoxy. Fig­ure 28 com­pares the inher­ent­ly low ion­ic con­duc­tiv­i­ty of anhy­dride­cured epoxy with that of two typ­i­cal novolac-cured com­pounds. In terms of dielec­tric con­stants, these com­pound types were found more sim­i­lar. Many oth­er exam­ples exist that rec­om­mend DEA as a unique ana­lyt­i­cal tool in ther­moset applications.

Summary

In this study, dielec­tric and elec­tric epoxy com­pound data were ana­lyzed using ASTM tests and Dielec­tric Analy­sis (DEA) to exam­ine var­i­ous fac­tors such as tem­per­a­ture, fre­quen­cy, batch-to-batch vari­a­tions, mois­ture, mate­ri­als, and the absence of postcure effects. The results were com­pared to data obtained from dom­i­nant fused sil­i­ca (FS) filler. The find­ings indi­cat­ed that dielec­tric strength and arc resis­tance of the new com­pounds slight­ly increased with high­er lev­els of FS. The vol­ume resis­tiv­i­ties, dielec­tric con­stants, and dis­si­pa­tion fac­tors of dif­fer­ent resin sys­tems were sim­i­lar, but decreased with high­er tem­per­a­tures or expo­sure to mois­ture. Fre­quen­cy and tem­per­a­ture had an impact on the dielec­tric data, with high­er val­ues observed in com­pounds with­out postcure or with crys­talline sil­i­ca and moist FS fillers com­pared to ref­er­ence com­pounds or con­di­tions. The dielec­tric prop­er­ties of ECN-based com­pounds remained con­sis­tent across dif­fer­ent batch­es. Dielec­tric con­stants, loss, and dis­si­pa­tion fac­tors increased with tem­per­a­ture, par­tic­u­lar­ly at low­er fre­quen­cies, and exhib­it­ed a peak at around 220°C due to rheological/dielectric cor­re­la­tion. Dielec­tric con­stants declined steadi­ly at low tem­per­a­tures towards the giga­hertz range. Ion­ic con­duc­tiv­i­ties increased with tem­per­a­ture and were influ­enced by postcure or glass tran­si­tion tem­per­a­ture. The study con­clud­ed that dielec­tric analy­ses effec­tive­ly described mate­r­i­al effects of epoxy com­po­si­tions. The DEA method, which involves effi­cient tem­per­a­ture and fre­quen­cy scans, was rec­om­mend­ed for gen­er­at­ing dielec­tric epoxy data when ASTM D150 is not specif­i­cal­ly required and could be explored fur­ther with oth­er thermosets.

CAPLINQ spe­cial­izes in the dis­tri­b­u­tion and tech­ni­cal exper­tise of spe­cial­ty chem­i­cals and mate­ri­als, includ­ing but not lim­it­ed to ther­mal inter­face mate­ri­als, epoxy mold­ing com­pounds (EMCs) and coat­ing pow­ders, and adhe­sives. In par­tic­u­lar, CAPLINQ rep­re­sents EMCs from Hysol Huawei Elec­tron­ic Co. Ltd. in Europe, Amer­i­ca, and Asia while we also man­u­fac­ture our own line of prod­ucts.

You may find it use­ful to browse our over­all semi­con­duc­tor-grade epoxy mold­ing com­pounds prod­uct cat­a­log. We also have Die attach mate­ri­als, Ther­mal inter­face mate­ri­als, and Encap­su­la­tion mate­ri­als that can pro­vide solu­tions to you and your team. 

Please also refer to this CAPLINQ Prod­uct Port­fo­lio pre­sen­ta­tion for detailed infor­ma­tion about our prod­uct offerings.

Should there be addi­tion­al spec­i­fi­ca­tions required, please con­tact us to assist and pro­vide you with addi­tion­al recommendations.

REFERENCES

1. H.W. Rauhut, New Types of Micro­elec­tron­ic Epoxy com­pounds, SPE, Ther­moset RETEC, Chica­go 1994
2. R. See­berg­er, Capac­i­tance and Dis­si­pa­tion Fac­tor Mea­sure­ments, IEEE Elec­tri­cal Insu­la­tion Mag­a­zine, Vol­ume 2, No. 1, Jan­u­ary 1986, pp. 27–36
3. E.R. Salmon, Dielec­tric Strength of an Insu­la­tion Material­ ls it a con­stant?, IEEE Elec­tri­cal Insu­la­tion Mag­a­zine, Vol­ume 5, No. 1, January/February 1989, pp. 36–38
4. J. Brauer, IBM, Design and Mate­ri­als, Elec­tron­ic Mate­ri­als Hand­book, Vol­ume 1, Pack­ag­ing ASM Inter­na­tion­al, 1989, Sec­tion 1, page 1
5. H. Lee and K. Neville, Hand­book of Epoxy Resins, McGraw-Hill Book Com­pa­ny, New York, 1967
6. H.W. Rauhut, Dielec­tric Analy­ses of Micro­elec­tron­ic Epoxy Com­pounds, SPE, ANTEC, Vol­ume 1, San Fran­cis­co, 1994, pp. 941–950

List of Definitions

Arc Resis­tance	the abil­i­ty of an insu­la­tor sur­face to with­stand con­duc­tive bridg­ing (car­boniza­tion) by high volt­age and low cur­rent. Break­down Unit: Second.
Dielec­tric Analy­sis (DEA)	mea­sures capac­i­tance and con­duc­tance ver­sus time, tem­per­a­ture, fre­quen­cy. DEA data are cor­re­lat­ed to chem­istry, rhe­ol­o­gy and poly­mer mobility.
Dielec­tric Breakdown	an insu­la­tor’s com­plete fail­ure at high volt­age by dis­rup­tive elec­tri­cal dis­charge. Unit: V/cm.
Dielec­tric Constant	the ratio of the capac­i­tance of an insu­la­tor to the capac­i­tance of vac­u­um at a giv­en elec­trode configuration.
Dielec­tric Loss	elec­tri­cal ener­gy dis­si­pat­ed as heat in an insu­la­tor, due to ion and dipole motion/ fric­tion in an alter­nat­ing elec­tri­cal field.
Dielec­tric Strength	a mate­ri­al’s abil­i­ty to with­stand volt­age, i.e., max­i­mum volt­age required to break down a cer­tain thick­ness of insu­la­tion. Unit Volts per mil (VPM) or V/mm.
Dis­si­pa­tion Fac­tor (DF, loss tangent)	the ratio of the loss index to its per­mit­tiv­i­ty (dielel­tric constant).
Ion­ic Conductivity	the sum of ions per insu­la­tor vol­ume, times their charges and mobil­i­ty; deter­mined by loss fac­tor mea­sure­ments at low frequency.
Loss Index or Loss Factor	the prod­uct of per­mit­tiv­i­ty and dis­si­pa­tion fac­tor. Ion­ic con­duc­tiv­i­ty and dipole ori­en­ta­tion are the com­po­nents that con­tribute to the Loss Fac­tor. At very low fre­quen­cy, the loss fac­tor is com­plete­ly dom­i­nat­ed by Ion­ic Conductivity.
Per­mit­tiv­i­ty	the ratio of insu­la­tor capac­i­tance to the capac­i­tance of vac­u­um (bet­ter term than dielec­tric con­stant, which is not a con­stant). It is due to dipole alignment.
Pow­er Fac­tor (PF)	the ratio of ener­gy in watts (dis­si­pat­ed in a mate­r­i­al) to the prod­uct of effec­tive volt­age mul­ti­plied with the cur­rent, also expressed as cosine of phase angle.
Sur­face Resistivity	a mate­ri­al’s abil­i­ty to resist pas­sage of an elec­tric cur­rent along the sur­face of an insu­la­tor. Unit: Ohm.
Vol­ume Resistivity	a mate­ri­al’s abil­i­ty to resist pas­sage of an elec­tric cur­rent through the cross-sec­tion of an insu­la­tor. nit: Ohm-cm.