Liquid Anhydride Hardeners for Electrical Casting: MNA, MTHPA and MHHPA

Electrical casting and impregnation put three constraints on an anhydride hardener at once, and they pull in different directions. The resin has to flow into a winding or a coil and out again under vacuum, which wants low viscosity. The cured part has to hold its mechanical and dielectric properties at class temperatures, which wants a stiff backbone. And the part has to look the same after ten years in service, which wants no unsaturation left anywhere in the network.

There are three liquid anhydrides in ordinary supply that a formulator actually chooses between for this work: methyl nadic anhydride (MNA, 25134-21-8), methyl tetrahydrophthalic anhydride (MTHPA, 11070-44-3) and methyl hexahydrophthalic anhydride (MHHPA, 25550-51-0). They differ by about 6% in molecular weight, by a factor of five in viscosity and by a great deal in everything the casting shop cares about. This page is about choosing between them for electrical work; for the chemistry of tetrahydro versus hexahydro in general, and for blends and single isomers, see the linked pages.

The Three Candidates, Side by Side

Property MNA (25134-21-8) MTHPA (11070-44-3) MHHPA (25550-51-0)
Formula / MW C10H10O3 / 178.18 C9H10O3 / 166.17 C9H12O3 / 168.19
Skeleton Bicyclic norbornene + methyl Single ring, one double bond Single ring, fully saturated
Viscosity at 25 deg C 170-300 mPa.s 30-50 or 50-70 mPa.s 50-70 mPa.s
Crystallisation point -15 deg C max -15 or 0 deg C, by grade -15 deg C max
Working AEW 181.8 at 98% 175.7 at 41.0% anhydride 171.6 at 98%
Anhydride equivalents per kg 5.50 5.69 5.83
phr at EEW 185, r 0.90 88.5 85.4 83.5
Retained double bond Yes (norbornene) Yes (ring) None
Colour stability in hot hold Fair Fair Best
Heat resistance of cured network Highest Middle Middle

Constraint 1: Viscosity and Impregnation

Impregnation is a flow problem before it is a chemistry problem. A vacuum-pressure cycle has to pull resin into a winding, hold long enough for the air to leave, and then drain or leave the resin where it is wanted. Viscosity at the process temperature sets how fast that can happen, and the difference between the three products is large: MTHPA at 30 to 50 mPa.s and MNA at 170 to 300 mPa.s are not the same plant.

Process What the hardener has to do What that points to
Vacuum-pressure impregnation of a winding Penetrate fine wire bundles, then drain cleanly Low viscosity: MTHPA at the 30-50 end
Vacuum casting of a filled coil Carry silica without settling through a long fill Higher viscosity: MNA, or MTHPA at the 50-70 end
Drip or trickle impregnation Wet out and stay in place Viscosity controlled by temperature, plus a gel-time match
Large casting with slow gel Stay homogeneous for hours Viscosity and filler settling both matter; MNA earns its cost here

The practical point is that MNA’s high viscosity is a feature in the second and fourth rows and a problem in the first. Where it is a problem, warming to 50 to 70 deg C before mixing brings it into the range a tetrahydrophthalic anhydride works cold – but a warm mix shortens the pot life, and a plant that warms the hardener has to shorten its day plan to match.

Constraint 2: Heat Resistance

Heat resistance of an anhydride-cured epoxy comes mostly from the stiffness of the anhydride backbone, and the three products are not equal. MNA carries a bridged bicyclic ring, which is the stiffest skeleton in the family; MTHPA and MHHPA carry a single six-membered ring, which can flip between conformations and gives a more flexible segment between crosslinks.

That is the whole reasoning behind the class H impregnating varnish tradition: a rigid hardener raises the temperature at which the cured network starts to lose modulus, at the cost of toughness and of a hotter cure. It is also why the choice is not made on a single temperature figure. Ask for the heat deflection temperature of the actual formulation – resin, hardener, filler ratio, cure schedule and accelerator all move it – rather than for a hardener property, because the hardener does not have one on its own.

What can be stated from the molecules: on the same resin, at the same r, MNA cures stiffer and more heat resistant than MTHPA, which is close to MHHPA. The step from MNA to the single-ring products is the large one; the step between MTHPA and MHHPA is small.

Constraint 3: Colour Stability

The opposite pressure comes from the retained double bond. Both MNA and MTHPA keep one carbon-carbon double bond, and that bond is an oxidation site: a mix held hot and open will yellow, and the yellowing is faster at a higher hold temperature. MHHPA is fully hydrogenated and has no such site, which is why it is the hardener specified for visible castings, optical parts and any application where the cured colour is on the drawing.

For a casting that is potted, painted or hidden inside a steel tank, this is worth nothing and should not be paid for. For a casting that is inspected, or seen, or used in a light-coloured assembly, it is often the deciding argument – and it is the reason a plant will run MHHPA on one line and MTHPA on the next.

The Dose Difference Is Small; the Process Difference Is Not

At EEW 185 and r 0.90 the three doses are 88.5, 85.4 and 83.5 phr. The spread from MNA to MHHPA is 6% and from MTHPA to MHHPA is 2.2%. Those numbers are small enough that a substitution looks harmless on the mixing sheet, and that is the trap:

Change Dose step at EEW 185, r 0.90 What else changes
MTHPA to MHHPA -2.2% Virtually nothing in dose; better colour, slightly lower heat resistance
MTHPA to MNA +3.6% Viscosity roughly five times higher; hotter, slower cure; stiffer network
MNA to MHHPA -5.6% Large drop in viscosity; better colour; lower heat resistance
Any of the three from solid THPA +13% to +20% A different plant – liquid versus melt

Read the last line as the warning it is: the substitution that actually goes wrong is not between the liquids, it is between a liquid and a solid, where the dose moves 15% and the equipment changes completely. Between the three liquids, a reformulation is still needed, but for processing and performance reasons rather than for stoichiometry.

Fillers, Settling and the Long Cast

Most electrical castings are filled, with silica or alumina at loadings high enough that settling becomes a real risk over a two- or three-hour fill. Settling rate depends on the viscosity of the continuous phase and on the particle size of the filler, and the hardener sets the floor of that viscosity: an MNA system starts thick and stays thick, an MTHPA 30-50 system starts thin and depends on the filler to build viscosity back.

That is the argument for MNA in a large filled casting and against it in a thin impregnation. It also explains why the two grades of MTHPA exist – 30 to 50 mPa.s for penetration, 50 to 70 mPa.s for filled work – and why a plant that casts both will keep both.

Cure Schedules and Accelerators

All three are heat-cured systems. A typical casting schedule starts with a gel hold around 90 to 120 deg C, moves to a cure hold at 140 to 160 deg C and finishes with a post-cure above that where the heat deflection temperature has to be developed. What changes between the products is where in that band the schedule wants to sit: MNA’s rigid skeleton generally wants more heat, and its higher viscosity means a filled mix reaches the same flow at a higher temperature.

Accelerators are routine. Benzyldimethylamine, imidazoles and quaternary ammonium salts are the common families, dosed from a fraction of a percent to about two parts per hundred of resin, and the exact figure is a property of the whole system. Two rules survive from the plant: an accelerator changes the time, not the stoichiometry – do not adjust the hardener dose when you add one – and acceleration shortens pot life in a filled tank as well as gel time in the mould.

Selection Table

Situation Reasonable starting point Why
Class H motor or generator impregnation MNA, or MTHPA low-viscosity grade Heat resistance is the specification; viscosity decides which of the two
Visible or optically inspected casting MHHPA No retained double bond, so no yellowing in the hold
Large filled transformer casting MNA Filler stays in suspension; heat resistance is a bonus
Thin section, fast fill, cost-sensitive MTHPA 30-50 Lowest viscosity at the lowest equivalent cost in the group
Mixed plant casting several part types MTHPA as the common hardener One hardener, two viscosity grades; MHHPA or MNA added only where justified
Winter storage without heating Any of the three, but check the grade MTHPA is also sold at a 0 deg C crystallisation point – that grade is not the winter one

Qualification Checklist

  • Viscosity at the actual process temperature, not only at 25 deg C, and stated per lot.
  • Crystallisation point, read against where the drum will be stored, not against the certificate’s convenience.
  • Free acid or moisture, converted into a water budget: 0.1% free acid is roughly 0.1 kg of water per tonne, which is why the tight grades are sold with a nitrogen blanket.
  • Colour on the basis you will judge by – Pt-Co on the neat liquid, specified rather than reported, if the part is visible.
  • Heat deflection temperature of the finished formulation, measured on your own resin, filler ratio and cure schedule; no hardener certificate answers it.
  • Accelerator compatibility, because the cure schedule and the pot life both live there.

FAQ

Which of the three gives the highest heat resistance?

MNA, because of the bridged bicyclic skeleton. It is the whole reason MNA exists next to the cheaper single-ring products. The trade is viscosity, cure temperature and colour stability.

Can MTHPA and MHHPA be swapped without reformulating?

The dose moves only 2.2%, so the stoichiometry survives the swap. What changes is colour stability, heat resistance and – if the grades are different – viscosity. Treat it as a minor reformulation rather than a like-for-like substitution.

Why is MNA so much more viscous?

The bridged ring resists rotation and the commercial product is an isomer mixture, both of which raise viscosity. 170 to 300 mPa.s against 30 to 70 for the single-ring products. Warming to 50 to 70 deg C brings it into a pumpable range.

Do all three need an accelerator?

None of them cures at room temperature, and all of them benefit from an accelerator when the oven cycle is the constraint. Whether one is needed is a question about the part and the oven, not about the hardener.

Which one is specified for visible castings?

MHHPA, because it has no retained double bond and therefore no oxidation site for yellowing. If the casting is not visible, that property is worth nothing and the cheaper single-ring grade does the job.

References and Further Reading

Contact PolyBlueChem for a current certificate of analysis, a sample of any of the three hardeners, and technical data on viscosity, free acid and ion content.

Scroll to Top