Glass Transition Temperature (Tg) of Epoxy Mold Compounds

Glass Transition Temperature (Tg) of Epoxy Molding Compounds

By def­i­n­i­tion, the glass tran­si­tion tem­per­a­ture (Tg) of an epoxy mold­ing com­pound (EMC) is the tem­per­a­ture at which the EMC changes from a hard, glassy sub­stance to a soft rub­bery one. Phys­i­cal mate­r­i­al prop­er­ties such as dimen­sion, vol­ume, and coef­fi­cient of ther­mal expan­sion (CTE) increase after this tran­si­tion region. CTE will be dis­cussed lat­er. Oth­er prop­er­ties such as mod­u­lus and hard­ness decrease beyond this region.

A good anal­o­gy to this behav­ior (though not entire­ly cor­rect) is the dash­board in your car. At 6:00 am the morn­ing of a hot sum­mer day, your dash­board will be stiff and hard. By mid­day, after the sun has beat down on it, it will have become soft­er, more flex­i­ble and have expand­ed some­what. How­ev­er, by 9:00 p.m. that evening, the dash­board has returned to a stiff, hard mate­r­i­al. So as long as the heat does not push the dash­board beyond its elas­tic lim­its, this cycle is repeat­able and non-destruc­tive. The same holds true with ther­moset plas­tics epoxy mold­ing compounds.

What the numbers don’t tell you

Unfor­tu­nate­ly, there’s a lot that the numer­i­cal val­ue for Tg does not tell you. As you can see in Fig­ure 1, this tem­per­a­ture is more accu­rate­ly defined as a region and not a sin­gle point. Depend­ing on the tests used and the report­ing method, the num­ber report­ed can vary wide­ly. Typ­i­cal­ly, the num­ber is giv­en as the mid­point of the tran­si­tion region, but what the val­ue doesn’t tell you is how wide a region this actu­al­ly is.

Fig­ure 1. Epoxy Mold Com­pound Prop­er­ties affect­ed by Glass Tran­si­tion Temperature

For exam­ple, say the Tg region of Epoxy Mold Com­pound A spans from 95°C to 115°C while that of Epoxy Mold Com­pound B spans from 98°C to 102°C, the Tg val­ues asso­ci­at­ed with them would be report­ed as 105°C and 100°C respec­tive­ly if the mid­point rule were fol­lowed. Look­ing strict­ly at the report­ed mid­point val­ues, one would be inclined to select Epoxy Mold Com­pound A with a Tg of 105°C, but it like­ly that the 100°C Tg of Epoxy Mold Com­pound B would make a bet­ter can­di­date based on the nar­row­er tran­si­tion window.

Also, as was men­tioned ear­li­er, the Tg of a mate­r­i­al is typ­i­cal­ly report­ed as the mid­point of the tran­si­tion region, though this nei­ther is a hard and fast rule. Tru­ly of inter­est to a design­er is the point at which mate­r­i­al prop­er­ties begin to change, not halfway through this tran­si­tion. In this case, the design­er would most like­ly pre­fer to have the Tg report­ed at the begin­ning of the tran­si­tion region.

Tak­ing the same exam­ple above, Epoxy Mold Com­pound A’s tran­si­tion region spans from 95°C to 115°C and the sup­pli­er decides to list the material’s Tg as the mid­point of the tran­si­tion region, 105°C. Epoxy Mold Com­pound B’s sup­pli­er lists the Tg as the begin­ning of the tran­si­tion phase, 98°C. Again, Epoxy Mold Com­pound A would like­ly win out in the eyes of an untrained designer. 

Con­verse­ly, two sup­pli­ers could pro­vide the same Epoxy Mold Com­pound A, the first report­ing the end­point of the tran­si­tion region, 115°C and the sec­ond report­ing the begin­ning of the tran­si­tion region, 95°C. Again, an untrained design­er would be inclined to select the first sup­pli­er, even though the mate­ri­als are identical.

DSC vs. DMA vs. TMA

Equal­ly as impor­tant as the width of the glass tran­si­tion region is the test used to deter­mine the region. The three most pop­u­lar ways of mea­sur­ing this tem­per­a­ture are Dif­fer­en­tial Scan­ning Calorime­try (DSC), Dynam­ic Mechan­i­cal Analy­sis (DMA) and Ther­mo Mechan­i­cal Analy­sis (TMA).

Dif­fer­en­tial Scan­ning Calorime­try (DSC) is the fastest and eas­i­est way of deter­min­ing the Tg of a mate­r­i­al. It involves tak­ing a small sam­ple of Epoxy Mold­ing Com­pound, usu­al­ly 10–20 mg and heat­ing it in a pan at a con­stant rate of 10°C/min. The amount of heat absorbed is then mea­sured and plot­ted as a graph of heat flow vs. tem­per­a­ture (Fig­ure 2). The fig­ure is then ana­lyzed, and the first shift in base­line heat flow is defined as the DSC onset, the peak of this curve is defined as the DSC peak and (typ­i­cal­ly) the mid­point tem­per­a­ture of this tran­si­tion is defined as the Tg.

Fig­ure 2. Tg by Dif­fer­en­tial Scan­ning Calorime­try (DSC) Epoxy Mold­ing Com­pound Duroplast

Unfor­tu­nate­ly, this method of analy­sis is not very reli­able and results obtained by this method com­pared to those obtained by TMA and DMA often dif­fer by 10–15°C or more.

Ther­mo Mechan­i­cal Analy­sis (TMA) uses the change in coef­fi­cient of ther­mal expan­sion (CTE) of mate­ri­als to deter­mine Tg. For ther­moset plas­tics such as epoxy mold­ing com­pounds, this num­ber is often con­sid­er­ably high­er below the Tg than above the Tg. This test mea­sures the expan­sion of the EMC through­out a tem­per­a­ture range (Fig­ure 3). Below the Tg, the EMC expands at a low­er rate, mea­sured in parts per mil­lion per degrees Cel­cius (ppm/oC). Above the Tg, the mate­r­i­al expands at a greater rate. The deflec­tion in the curve show­ing the expan­sion vs. tem­per­a­ture indi­cates the Tg. In prac­tice, tech­ni­cians or soft­ware often join these two slop­ing lines and record the inter­sec­tion as the glass tran­si­tion temperature.

Fig­ure 3: Tg by Ther­mo Mechan­i­cal Analy­sis (TMA) of Epoxy Mold­ing Com­pound Duroplast

TMA also takes addi­tion­al prepa­ra­tion time. All the sam­ples must be as close in size as pos­si­ble and all read­ings need to be tak­en in the same place. This is because larg­er sam­ples will have dif­fer­ent cure stress­es than small­er ones, and for the same rea­son, read­ings tak­en at the cen­ter can be dif­fer­ent from read­ings tak­en towards an edge. Assum­ing that mea­sure­ments are always tak­en in the same place, this method pro­duces more repeat­able results. Ther­mo­me­chan­i­cal Analy­sis is also the way most Epoxy Mold Com­pound Man­u­fac­tur­ers report the Tg of their EMCs.

Dynam­ic Mechan­i­cal Analy­sis (DMA) involves oscil­lat­ing a load through a rec­tan­gu­lar bar of cured mate­r­i­al. This can be done using 3‑point bend­ing, dual can­tilever bend­ing or even a sin­gle bend­ing can­tilever as shown in Fig­ure 4. Stress is trans­ferred through the spec­i­men, and the rel­a­tive mod­u­lus of the mate­r­i­al is mea­sured as a func­tion of time and temperature.

Fig­ure 4: Tg by Dynam­ic Mechan­i­cal Analy­sis (DMA) of Epoxy Mold­ing Com­pounds Duroplast

Though this tech­nique is the most repeat­able of the three list­ed, the heat­ing rate and fre­quen­cy of oscil­la­tion used will great­ly influ­ence the result­ing Tg mea­sure­ments. Faster heat­ing rates will result in a high­er Tg due to ther­mal lag, and a high­er fre­quen­cy of oscil­la­tion will lead to a high­er Tg due to the inher­ent fre­quen­cy depen­dence of vis­coelas­tic mate­ri­als. Addi­tion­al­ly, the data obtained can be inter­pret­ed in sev­er­al ways. The sum­ma­ry of the con­sid­er­a­tions of each of these tests is list­ed in Table 1.

Table 1: Con­sid­er­a­tions of DSC vs. DMA vs. TMA
MethodTyp­i­cal Run TimeSam­ple Prepa­ra­tionRepeata­bil­i­tyDepend­abil­i­tyCom­ments
DSC20 min­utesEasyGoodMar­gin­alMany mate­ri­als do not show a clear transition
TMA40 min­utesMedi­umFairGoodDepen­dent on sam­ple preparation
DMA120 min­utesDif­fi­cultExcel­lentGoodRate of heat­ing, oscil­la­tion and inter­pre­ta­tion can lead to dif­fer­ent values

Now that it has been estab­lished that the Tg is a range, and not just a numer­i­cal val­ue, and that there are sev­er­al stan­dard ways of report­ing this range, let us explore the method by which a mate­r­i­al reach­es its Tg.

The chem­istry of epox­ies is known as addi­tion chem­istry. This means that in order to obtain spe­cif­ic mate­r­i­al prop­er­ties, part A (known as the resin) is mixed with part B (known as the hard­en­er). The resul­tant mixed chem­istry has cer­tain prop­er­ties includ­ing vis­cos­i­ty, hard­ness, CTE and Tg.

With this type of chem­istry, it is very dif­fi­cult to obtain a Tg that is much high­er than the cure tem­per­a­ture, espe­cial­ly with short cure times. For exam­ple, if the epoxy mold com­pound is cured at 80°C, the Tg of the EMC will not typ­i­cal­ly be much high­er than 85°C – 90°C. An epoxy matrix will gel rapid­ly until the Tg reach­es the cure tem­per­a­ture, then the cure rate slows con­sid­er­ably. In the case of many high-tem­per­a­ture cure epoxy mold com­pounds with mul­ti­ple cure sched­ules such as 5 min­utes at 165°C, the prod­uct will appear cured in approx­i­mate­ly 2–3 hours at 100°C, but will not reach the high Tg (typ­i­cal­ly 130°C) stat­ed on the tech­ni­cal data sheet. So although there are longer, low­er-tem­per­a­ture cure cycles for these high-tem­per­a­ture cure EMCs, the high Tg list­ed on the tech­ni­cal data sheet will not be reached by these cycles. High­er tem­per­a­ture cures are required to achieve the high Tg list­ed by high Tg epoxy mold­ing com­pounds.

But sim­ply expos­ing these epox­ies to high­er tem­per­a­tures is also not suf­fi­cient. Time must be allowed for a heat cure epoxy to “soak” at these tem­per­a­tures. This addi­tion­al soak­ing time allows the matrix to com­plete its cure. A case study was per­formed on a heat cure epoxy with an ulti­mate Tg list­ed on its tech­ni­cal datasheet of 125°C. Fol­low­ing the rec­om­mend­ed cure tem­per­a­ture of 150°C yield­ed the ulti­mate Tg in 2 hours. A low­er cure tem­per­a­ture of 120°C yield­ed a Tg of 100°C in 1 hour, but an addi­tion­al 2.5 hours was required to reach the ulti­mate Tg of 125°C. Fur­ther reduc­ing the cure tem­per­a­ture to 100°C yield­ed a Tg of 100°C in 2.5 hours, but a full 8 hours was required to reach the Tg of 125°C list­ed on the tech­ni­cal data sheet. These results are sum­ma­rized in Table 2 and a more detailed expla­na­tion can be found in our arti­cle, “How Epoxy Mold Com­pounds Cure”.

Table 2: Effect of Var­i­ous Cure Sched­ules on Tg
Cure Sched­uleUlti­mate Tg
2 hours at 150°C125°C
1 hour at 120°C100°C
3.5 hours at 120°C125°C
2.5 hours at 100°C100°C
8 hours at 100°C125°C

With the above under­stand­ing, let’s exam­ine a com­mon exam­ple where an untrained design­er might come up with some erro­neous con­clu­sions. For demonstration’s sake, let’s say that the design­er selects the above-men­tioned epoxy mold com­pound with a Tg list­ed on its data sheet of 125°C. In order to cure this epoxy, he selects the alter­nate cure sched­ule of 2.5 hours at 100°C. Then, sat­is­fied that he has thor­ough­ly cured the epoxy, he runs a DSC test to ver­i­fy his con­clu­sions. As explained above, a typ­i­cal DSC test takes a sam­ple of EMC and heats it a con­stant rate of 10°C/min. Since he is expect­ing to get a val­ue around 125°C, he will like­ly run the test to about 150°C. Low and behold, his expec­ta­tions are met, as the curve indi­cates that the Tg is around 125°C. What he hasn’t real­ized is that the test itself has post-cured the epoxy, push­ing the sample’s Tg above the 100°C that it would have oth­er­wise achieved.

This exam­ple rais­es two impor­tant questions:

  • Do the high­er cure tem­per­a­tures need­ed to increase a material’s Tg need to be per­formed in a sin­gle stage?
  • What is the effect of an under-cured epoxy mold compound?

Post-bake cure cycles for Epoxy Mold Compounds

As demon­strat­ed in the above exam­ple, a future, sec­ondary-stage cure cycle can be used to increase the Tg of an epoxy mold com­pound. This means that a pri­ma­ry cure cycle can be used to effec­tive­ly “gel” the mate­r­i­al in place.

In fact, many com­pa­nies — par­tic­u­lar­ly capac­i­tor man­u­fac­tur­ers — do not post mold cure their epoxy mold com­pounds at all. The in-mold cure gives suf­fi­cient mechan­i­cal prop­er­ties, and any increase in per­for­mance of EMCs can be obtained dur­ing operation.

Effect of an under-cured epoxy mold compound

The short answer is that an under-cured epoxy has not max­i­mized its full poten­tial in terms of envi­ron­men­tal resis­tance, flex­ur­al strength and mod­u­lus, and glass tran­si­tion tem­per­a­ture. A bet­ter answer will reveal that ful­ly max­i­mized prop­er­ties may not always be required for all giv­en appli­ca­tions. Although longer, hot­ter cure sched­ules will give clos­er to 100% cure, low­er or short­er cure cycles will often give prop­er­ties that are 90% — 95% those of their ful­ly cured counterparts. 


The glass tran­si­tion tem­per­a­ture, Tg of epoxy mold­ing com­pounds is bet­ter described as a range than a sin­gle val­ue on a tech­ni­cal data sheet. An EMCs Tg can be test­ed in sev­er­al ways, includ­ing DSC, TMA and DMA. Each of these meth­ods has its advan­tages and dis­ad­van­tages, and each of them leaves room for inter­pre­ta­tion. Under­stand­ing Tg and the var­i­ous meth­ods of test­ing for it in an epoxy mold com­pound are impor­tant weapons in the arse­nal of a semi­con­duc­tor mold engi­neer and designer.

Please vis­it us at to learn more about our whole range of epoxy mold com­pounds (EMC) includ­ing our semi­con­duc­tor-grade epoxy mold com­pounds, fiber­glass rein­forced indus­tri­al-grade mold com­pounds, and opti­cal­ly clear epoxy mold­ing com­pounds for opto­elec­tron­ics appli­ca­tions. If you have any oth­er ques­tions about epoxy mold­ing com­pound pel­let sizes, please feel free to leave a com­ment below or don’t hes­i­tate to con­tact us.

About Chris Perabo

Chris is an energetic and enthusiastic engineer and entrepreneur. He is always interested in taking highly technical subjects and distilling these to their essence so that even the layman can understand. He loves to get into the technical details of an issue and then understand how it can be useful for specific customers and applications. Chris is currently the Director of Business Development at CAPLINQ.

3 thoughts on “Glass Transition Temperature (Tg) of Epoxy Molding Compounds

Leave a Reply

Your email address will not be published. Required fields are marked *