AKKN ElectronicsReplacing a wet aluminum electrolytic capacitor with a solid polymer type is often treated as a direct upgrade: lower ESR, higher ripple current, longer life. The "longer life" claim needs qualification. The two chemistries wear out by different mechanisms, and the lifetime model that applies to one does not transfer to the other. This article compares the aging physics, gives the derating numbers that matter, and lists design rules for cross-replacement.
Two Chemistries, Two Wear-Out Mechanisms
A wet aluminum electrolytic capacitor contains a liquid electrolyte impregnated in the anode foil. The electrolyte gradually evaporates or is consumed by reactions with the oxide layer. The rate follows an Arrhenius relationship: life roughly doubles for every 10 °C reduction in core temperature. End-of-life is typically defined as a 20 % capacitance drop, a 2× ESR increase, or leakage current exceeding the datasheet limit — whichever comes first.
A solid polymer capacitor replaces the liquid with a conductive polymer, usually PEDOT. There is no liquid to evaporate, so the classic drying-out mechanism is absent. Degradation appears instead as gradual oxidation of the polymer at high temperature, which raises ESR, and as sensitivity to moisture and mechanical stress. Polymer capacitors still carry load-life ratings — typically 2 000 to 10 000 h at rated temperature, depending on series — but the end-of-life criteria differ: common thresholds are an ESR increase of about 50 % or a capacitance shift of ±20 %. The progression is slower and less exponential than the wet type's wear-out.
The voltage capability also differs. Wet aluminum electrolytics are available up to 450 V and beyond. Most aluminum polymer series stop at 100–125 V, and polymer tantalum at 25–50 V. For high-voltage DC-link or PFC stages, the wet type remains the practical choice; polymer parts fit low-voltage, high-ripple outputs.
Derating by Temperature and Ripple: The Numbers
For wet electrolytic capacitors, the core temperature is the sum of the ambient temperature and internal heating from ripple current:
T_core = T_ambient + I_rms² × ESR × R_th
where R_th is the thermal resistance from core to ambient. The life estimate then follows:
L = L0 × 2^((T0 − T_core) / 10)
with T0 being the rated core temperature. A part rated at 105 °C that runs at 85 °C core has roughly 4× the rated life; at 65 °C, roughly 16×.
Ripple current is the main driver of internal heating. Consider a 470 µF / 25 V capacitor in a 10 mm × 10 mm can: a wet type might have an ESR of about 80 mΩ at 120 Hz, a polymer type about 10 mΩ at 100 kHz. At 1 A rms ripple, the wet part dissipates 80 mW internally; the polymer part dissipates 10 mW. The wet part also develops a positive feedback loop: as ESR rises with age, the same ripple produces more heat, accelerating further degradation.
Typical ripple derating curves for a 105 °C-rated wet part allow roughly 100 % of rated ripple at 105 °C ambient, about 1.7× at 85 °C, and about 2.5× at 65 °C, because more internal heating can be tolerated at lower ambient temperatures. Polymer parts are limited by the same thermal budget, but their lower ESR allows a much higher ripple current for the same temperature rise. Verify the limit with I_max = sqrt(P_max / ESR) rather than assuming a fixed multiplier.
Voltage derating also differs. Wet aluminum can operate at full rated voltage when ripple and temperature are within limits. Aluminum polymer should be derated to no more than 80 % of rated voltage; polymer tantalum to 50–70 % for reliable operation, because its surge voltage margin is narrow.
Operating Conditions vs. Expected Life
| Capacitor type | Core / case temperature | Applied ripple | Voltage derating | Expected life and end-of-life criteria |
|---|---|---|---|---|
| Wet Al, 105 °C rated | 105 °C | 100 % of rated | 100 % of rated | 2 000–3 000 h; end at C −20 % or ESR ×2 |
| Wet Al, 105 °C rated | 85 °C | 100 % of rated | 100 % of rated | 8 000–12 000 h; same criteria |
| Wet Al, 105 °C rated | 65 °C | 100 % of rated | 100 % of rated | 32 000–48 000 h; same criteria |
| Al polymer, 105 °C rated | 105 °C | 100 % of rated | ≤80 % of rated | 5 000–10 000 h; end at ESR +50 % or C ±20 % |
| Al polymer, 105 °C rated | 85 °C | 100 % of rated | ≤80 % of rated | No electrolyte wear-out; ESR drift is small; random defects dominate |
The wet part has a predictable, exponential life curve; the polymer part's life is less temperature-driven and more dependent on voltage stress, moisture, and mechanical integrity. In a low-voltage, high-ripple output stage running at moderate temperature, the polymer part usually gives the lower failure rate. In a high-voltage, low-ripple bulk capacitor application, the wet type remains the appropriate choice.
Design Rules for Cross-Replacement
- Compute the core temperature for wet parts before estimating life: T_core = T_ambient + I_rms² × ESR × R_th. Apply the 10 °C rule only after this calculation.
- Do not apply the 10 °C doubling rule to polymer capacitors; their wear-out model is not Arrhenius-dominated.
- When replacing wet with polymer, compare ESR at the operating frequency (typically 120 Hz for wet, 100 kHz for polymer), ripple current rating at the application ambient, and surge voltage withstand.
- Derate voltage: ≤80 % for aluminum polymer, 50–70 % for polymer tantalum; wet aluminum can run at full rated voltage within its ripple and temperature limits.
- Monitor different end-of-life indicators: capacitance loss for wet parts, ESR rise for polymer parts. An ESR increase without capacitance change points to polymer degradation; a capacitance drop with ESR increase points to electrolyte loss.
- Check ripple at the actual switching frequency. Polymer ESR stays low from 100 kHz to 1 MHz; wet ESR rises with frequency, so a part selected at 120 Hz may be under-rated at the converter frequency.
- When sourcing a replacement across brands, compare the lifetime model parameters — rated temperature, ESR limit, ripple current at temperature — rather than the brand name. Two parts with the same voltage and capacitance can differ by 3× in ripple rating.
