Alicyclic and Aromatic Dianhydrides: Choosing the Polyimide Backbone

A polyimide is built from a dianhydride and a diamine, and buyers usually choose the dianhydride on equivalent weight. That is the wrong lever. Two dianhydrides can differ by 3% on equivalents and produce films that differ in colour, dielectric loss and coefficient of thermal expansion by margins no formulation can close. What actually decides those properties is the backbone attached to the anhydride rings – aromatic, alicyclic, or bridged – and this article is about choosing it deliberately.

Two independent variables in one molecule

Every dianhydride does two jobs at once, and they are separable:

  • Stoichiometry – how many anhydride equivalents per gram, which sets the molar charge against the diamine and appears in the phr arithmetic.
  • Backbone – the ring system between the anhydride groups, which sets colour, dielectric behaviour, chain stiffness, solubility and hydrolysis sensitivity.

The first is a number on a certificate. The second is a decision nobody can check for you after the fact, because the same certificate can describe two materials with completely different service behaviour.

The five backbones that matter

Dianhydride CAS Formula / MW AEW Backbone class Defining trade
PMDA 89-32-7 C10H2O6 / 218.12 109.06 Aromatic, fused Highest rigidity and heat resistance at the lowest cost; amber film
HPMDA 2754-41-8 C10H8O6 / 224.17 112.09 Alicyclic, fused Colourless film, low dielectric loss; lower Tg and higher CTE
BPDA 2420-87-3 C16H6O6 / 294.22 147.11 Aromatic, biphenyl Rigid rod, high modulus, high heat resistance
Methylenebis(phthalic anhydride) 4200-79-7 C17H8O6 / 308.24 154.12 Aromatic, CH2 bridged Flexible hinge and improved solubility; pale amber
Ethylidenebis(phthalic anhydride) 22197-42-8 C18H10O6 / 322.27 161.14 Aromatic, CH(CH3) bridged One carbon longer, less crystalline, more soluble
ODPA 1823-59-2 C16H6O7 / 310.21 155.11 Aromatic, ether bridged Melt processability, high cost
BPADA 38103-06-9 C31H24O8 / 524.5 262.25 Aromatic, bisphenol A Thermoplastic PI, very high cost

Read the first three columns as one group and the last three as another. Within the aromatic family, the difference between PMDA, BPDA, ODPA and the bridged bis-anhydrides is largely a stiffness-versus-processability trade at a similar colour. Between PMDA and HPMDA the trade is different in kind, because the backbone stops being aromatic.

The equivalent weight trap: 3% apart, worlds apart

PMDA and HPMDA are the cleanest illustration of why AEW alone is not a specification.

Parameter PMDA HPMDA Difference
Molecular weight 218.12 224.17 +2.8%
AEW (two anhydride rings) 109.06 112.09 +2.8%
Equivalents per kg 9.17 8.92 -2.7%
Molar charge for the same diamine 1.000 1.028 Under 3% – a balance check, not a reformulation
Film colour after imidisation Amber to brown Colourless to pale yellow Not comparable
Dielectric loss at high frequency Higher Lower Not comparable
Coefficient of thermal expansion Low (about 20 ppm/K typical) High (about 40 to 60 ppm/K typical) Two to three times
Solubility of the polyamic acid Limited in common solvents Better Practical, not marginal

The first four rows say the charge barely moves. The last four say the product is not the same product. Anyone who substitutes one for the other on the strength of the AEW match will find out in the first film they cure.

Where the colour comes from

The yellow-to-brown cast of an aromatic polyimide is not a purity problem and not a thermal degradation product. It is a ground-state charge-transfer interaction: the aromatic dianhydride is an electron acceptor, the aromatic diamine is an electron donor, and where they stack along the chain the pair absorbs in the visible. The effect is strongest with the most electron-poor dianhydride (PMDA) and a strongly donating diamine (ODA), which is exactly the classical PMDA-ODA film that everyone recognises by colour.

There are three ways to suppress it, and they cost different things:

  • Remove the aromatic acceptor. An alicyclic dianhydride such as HPMDA cannot form the charge-transfer pair with the dianhydride side, so the absorption band weakens and the film stays pale. This is the mechanism behind alicyclic CPI chemistry, and it is why HPMDA exists at all despite costing more than PMDA.
  • Weaken the donor. Fluorinated diamines (TFMB and relatives) reduce donation and are the usual partner for a colourless formulation, at a significant cost penalty.
  • Interrupt the stacking. Bulky substituents and non-planar bridges – the ethylidene bridge, for example – reduce chain packing and blunt the interaction, at the cost of Tg.

Note what does not work: adding a UV absorber or a optical brightener to the resin. The absorption is in the polymer backbone, and a stabiliser that has to survive 350 degC imidisation is neither cheap nor reliable.

Dielectric behaviour and the high-frequency case

For high-frequency laminates and antenna layers, two numbers matter: dielectric constant (Dk) and dissipation factor (Df). Aromatic conjugation raises both, because the delocalised system responds to the applied field. Removing it – alicyclic dianhydride, alicyclic or fluorinated diamine – lowers both, which is why alicyclic CPI formulations dominate the 5G substrate literature.

Two practical cautions. First, the resin is only half the answer: water uptake raises Df sharply, so a hydrophobic formulation matters as much as the backbone. Second, published Dk and Df figures are measured at a stated frequency and humidity and often on a specific film thickness – treat any number quoted in a datasheet as a starting point to be measured on your own film before it goes into a laminate specification.

Thermal expansion: the real price of colourlessness

An aromatic backbone is planar and stiff, so a PMDA-based film expands little with temperature. A cyclohexane ring is not planar and rotates, so an HPMDA-based film expands more – typically in the 40 to 60 ppm/K range against about 20 ppm/K for a rigid aromatic polyimide, depending on the diamine and the imidisation history.

That matters the moment the film is bonded to something else. A film that expands more than the copper it is laminated to will curl, and a film under constraint will develop stress. The usual answers are rigid co-monomers (a portion of BPDA or a rigid diamine), a controlled imidisation ramp, and thicker adhesive layers that absorb the mismatch. All three add cost or process time, which is why the colour decision is really a system decision.

Hydrolysis: the same backbone, a different appetite for water

Both families hydrolyse, and the arithmetic is worth carrying on the bench. HPMDA at 224.17 g/mol consumes two moles of water per mole to open both rings:

  • 1 g of water deactivates 6.22 g of HPMDA (224.17 / 36.03).
  • In a 20% polyamic-acid solution, a solvent carrying 1,000 ppm water therefore destroys about 0.31% of the dianhydride solids before the reaction even starts; at 5,000 ppm it is past 1.5% and the molecular weight distribution shifts visibly.
  • The first symptom is not a failed batch but a viscosity that stops climbing, because an opened ring caps the chain end instead of extending it.

Aromatic dianhydrides are also moisture sensitive, but the practical difference is that an alicyclic dianhydride is usually dissolved at higher solids in a more polar solvent, so there is more water in the pot to begin with. Dry the solvent, blanket the tank and keep the powder sealed.

Choosing between them: a four-question protocol

Question Answer points to What it rules out
1. Does the part have to be pale or optically clear? Alicyclic only (HPMDA) or aromatic plus a fluorinated diamine PMDA, BPDA for a colourless target
2. Is the service temperature above 300 degC, or is dimensional stability the primary requirement? Rigid aromatic (PMDA, BPDA) HPMDA as the sole dianhydride
3. Must the polyamic acid stay in solution at high solids, or must the resin melt-process? Bridged or ether dianhydrides (methylene, ethylidene, ODPA), or BPADA for thermoplastic behaviour PMDA alone in a solution process
4. Is dielectric loss the specification driver? Alicyclic dianhydride plus low-uptake diamine Aromatic backbone as the main component

Most real formulations answer “yes” to more than one, and the resolution is a blend. A common pattern is a majority alicyclic dianhydride for colour with 10 to 30 mol% of a rigid aromatic dianhydride to pull the CTE back into range – the same logic that makes blended anhydride hardeners work in epoxy, applied to a different property set.

The one-page summary

If the requirement is Primary dianhydride Watch
Colourless or transparent film HPMDA CTE two to three times the aromatic figure; price
Maximum heat resistance and modulus PMDA or BPDA Amber colour; limited solubility
Solubility and processability at high solids Methylene- or ethylidene-bridged bis-anhydride Slightly lower Tg than a rigid aromatic
Low dielectric loss HPMDA with a low-uptake diamine Moisture, not the backbone, sets the final Df
Thermoplastic behaviour BPADA Very high cost, high melt viscosity
Epoxy curing, not polyimide Any of them, on AEW HPMDA at 54.5 phr is brittle alone – blend it

The table above is the decision in one place. Everything else in this article is the reasoning behind it, and the reason the equivalent-weight match between PMDA and HPMDA is the least useful fact in the comparison.

Related reading

Related Products

Scroll to Top