Every diode leaks a little current in reverse, and the number on the datasheet is almost never the number your circuit experiences. A datasheet line reading “10 µA at 25 °C” quietly assumes a specific reverse voltage, a 25 °C junction, and a room-temperature measurement; run that same diode at 125 °C and the leakage can be a thousand times larger. For battery-powered, standby, or high-temperature designs, reverse leakage is not a datasheet footnote — it is a real current drain and a real heat source. This article defines what IR actually means, gives you the temperature heuristic to estimate hot leakage on paper, contrasts PN and Schottky magnitudes, and shows how to measure leakage on the bench without fooling yourself.
What the Data Sheet Means by “IR at VR”: Conditions Are Everything
Reverse leakage current, written IR on a datasheet, is the current that flows through a diode when it is reverse-biased. The value is quoted at a stated reverse voltage (VR) and temperature, and those conditions are the entire meaning of the number. A leakage figure without its test voltage and temperature is not a number you can design against; it is a point on a curve.
The two conditions that move IR the most are the reverse voltage and the junction temperature. Leakage rises as the applied reverse voltage approaches the breakdown region, and it climbs steeply with temperature — for many junctions, roughly an order of magnitude for every 25–30 °C. So a “10 µA at 25 °C” figure measured at a low reverse voltage says almost nothing about the same diode at 125 °C and its rated VR. Reading the leakage column without reading its test conditions is how undersized battery and standby budgets happen.
The practical habit is to treat IR as a function with two inputs — temperature and reverse voltage — and to always quote both when you carry a leakage number into a design. When a design article says “this diode leaks 10 µA,” the honest version adds “at 25 °C and 20 V reverse.” The forward-voltage hub is the mirror of this discipline on the forward side; the same test-condition rigor applies on both directions of the junction.
The x10-per-25C Heuristic: Estimating Hot Leakage on Paper
The single most useful estimate in leakage work is the temperature multiplication rule: for a typical junction, reverse leakage grows roughly tenfold for every 25 °C rise in junction temperature. It is an approximation with real error bars, but it turns “I do not know the leakage at 125 °C” into a defensible back-of-envelope number.
Work the arithmetic. Start from 10 µA at 25 °C. At 50 °C, multiply by 10 → 100 µA. At 75 °C, ×100 → 1 mA. At 100 °C, ×1000 → 10 mA. At 125 °C, ×10,000 → 100 mA. The progression makes the Schottky weak spot obvious: a part whose 25 °C leakage looks harmless can carry a tenth of an amp of reverse current in a hot junction box, dissipating real power and heating itself further.
The error bars are worth stating. The 10×-per-25 °C factor is a mid-range estimate; the true coefficient varies with material, structure, and the specific curve, and Schottky devices often show an even steeper climb near their temperature limit. So the heuristic is a screening tool, not a substitute for the datasheet’s own IR-versus-temperature curve. When leakage at temperature matters, read the actual curve; the heuristic tells you whether the curve is worth finding in the first place.
A concrete example makes the temperature budget concrete. A battery-powered sensor keeps a diode reverse-biased across a 3.6 V lithium cell for months of standby. The datasheet lists IR at 30 µA, 25 °C. The product sits in a 55 °C enclosure in summer, roughly a 30 °C rise — one step of the heuristic — so the standby leakage is on the order of 300 µA, not 30. Over a month that is roughly 0.2 Ah drained through the diode alone, a measurable slice of a small cell’s capacity. The fix is not to debate the 30 µA headline but to select a part whose leakage at the real ambient, not at 25 °C, fits the battery budget.
PN vs Schottky Leakage: Two Orders of Magnitude Story
The physics of the junction decides how much leakage a family carries, and the difference between PN and Schottky is often two orders of magnitude or more at the same voltage and temperature.
A silicon PN diode is a p-n junction whose reverse leakage is limited by the small number of minority carriers that can cross the junction. That number is small and grows predictably with temperature, which is why general-purpose silicon rectifiers carry leakage in the nanoamp-to-microamp range at room temperature. The structure is inherently low-leakage, and it stays low relative to Schottky at high temperature.
A Schottky diode, by contrast, uses a metal-semiconductor barrier whose reverse leakage is dominated by thermionic emission over a lower barrier. The trade that gives it a low forward drop — the low barrier — is the same thing that lets more current leak in reverse. A Schottky can leak hundreds of times more than a silicon PN part of the same voltage class, and the gap widens with temperature. That is why a Schottky is the right low-drop choice on a high-current rail and a poor choice for a high-temperature standby load that must hold reverse voltage. The trade between the two families is exactly the low-drop material comparison — low VF bought at the price of leakage.
Measuring Leakage with a DMM in µA Mode: Setup and Pitfalls
Leakage is measurable on the bench, but the setup matters more than the reading. Set the meter to the current mode’s microamp range, connect the diode in reverse bias, and read the small current the junction passes. A clean measurement needs three things: a stable reverse voltage source, a meter with enough resolution, and an awareness of what else is in the circuit.
The pitfalls are the ones that turn a good reading into a bad design decision. First, a TVS or Zener in parallel with the diode participates in the reading — a protection part’s own leakage adds to the number and looks like the diode’s. Second, the meter’s own burden and the leakage of the probe/test fixture can mask a genuine nanoamp figure. Third, the temperature must be controlled and recorded, because a warm bench inflates leakage exactly as the heuristic predicts. A leakage measurement taken on a hot afternoon and compared to a 25 °C datasheet line will always “look bad” — the same measurement trap the forward-voltage hub warns about, mirrored on the reverse side.
The cleanest setup is to test the diode out of circuit, with the reverse voltage set near the intended operating point and the temperature recorded. If the measured leakage falls near the datasheet curve at that voltage and temperature, the part is behaving; if it is orders of magnitude high, the junction is degraded or the measurement has a parallel path.
One more measurement caution keeps the bench honest: leakage is tiny and easily corrupted, so a reading of “zero” should be treated with as much suspicion as a high reading. Many meters display zero on the µA range when the true leakage is below their resolution, and that zero is not evidence the part is leak-free — it is evidence the meter cannot see it. When the budget depends on the exact standby figure, use a meter with adequate resolution or a dedicated leakage test, and record the setup with the number so the field can reproduce it.
When Leakage Becomes a System Problem: Battery, Standby, and Thermal Runaway
Leakage stops being a datasheet number when it becomes a system cost, and three scenarios turn it into one.
The battery scenario: in a battery-powered product, a diode that leaks continuously is a permanent drain. A 10 µA leakage across weeks of standby is a real fraction of a small battery’s capacity; a degraded part leaking a milliamp is a battery killer. Standby power budgets are exactly where the test-condition discipline pays for itself, because a diode chosen on its 25 °C leakage can leak a hundred times more in a warm, powered-off product.
The high-temperature scenario is the thermal runaway loop. A Schottky at high reverse voltage and high temperature leaks; that leakage dissipates power as heat; the heat raises the junction temperature; and the higher temperature raises leakage further. The loop is the mechanism behind the thermal runaway cases seen in hot enclosures, and it is covered in depth by the Schottky thermal runaway explainer. This article’s contribution is the measurement and the on-paper estimate that catch the loop before it starts.
The design response in all three cases is to budget leakage at the worst-case temperature, not the 25 °C headline. A battery design reads the leakage at its worst ambient, a standby design reads it at the powered-off temperature, and a hot Schottky design reads it against the thermal loop. The general rectifier category and the low-leakage families provide parts chosen with this budget in mind, and the datasheet comparison method applies the same discipline to the reverse side of the sheet.