SXUV photodiodes are radiation-hardened silicon detectors built by Opto Diode, a division of ITW, in Camarillo, California. What separates them from ordinary silicon is the entrance window: the platinum-silicide (PtSi) front window replaces the silicon-dioxide layer that causes UV-exposure-induced instability in common photodiodes, so an SXUV detector holds its calibration under fluences that visibly degrade oxide-windowed devices. That is why the family is used for dose and power monitoring in EUV lithography and on synchrotron beamlines, across the 1–190 nm range.
Source note: Silicon Photodiodes – SXUV Series with Platinum Silicide Front Entrance Windows (PDF). That document is the original, undated IRD application note and it predates the current SXUV catalog; it remains online because it holds the fluence-stability plots, the pulse-measurement method and the temperature data reproduced in summary below. Model names used in it that are not in the current line are noted where they appear. For specifications of a specific part, use the SXUV category page and the linked datasheets, which govern.
Why the entrance window is the whole story
A silicon photodiode collects charge generated within microns of its front surface, so whatever sits on that surface determines how much of the signal survives. In a conventional device that layer is silicon dioxide. The IRD note states the mechanism plainly: “the UV photon exposure induced instability of common silicon photodiodes is caused by the front silicon dioxide window.” Opto Diode’s UVG note traces that instability to the quality of the silicon–silicon-dioxide interface, where UV exposure creates interface states; near-surface collection efficiency falls away as a result.
SXUV devices replace that oxide with a platinum-silicide window, which the note credits with “eliminating the XUV exposure induced instability problem.” The trade is real and worth stating: SXUV responsivity in the deep UV is lower than a UV-optimized detector’s — the same note attributes the temperature behaviour of these parts to high surface recombination. SXUV is engineered for stability under dose, not for maximum responsivity. When the measurement needs raw UV sensitivity and the dose is modest, UVG is the better family; when calibration must survive heavy exposure, SXUV is.
What the stability testing actually showed
Stability testing was performed at NIST and LBL. Quantum efficiency did not change after exposure to 10 eV photons at a fluence of 1016 photons/cm², nor after 100 eV photons at 1022 photons/cm². From those results the note infers a radiation hardness of hundreds of Gigarads (Si). The series was developed in collaboration with NIST and NIH.
Three further results give the practical shape of that claim:
| Exposure | Result reported |
|---|---|
| 244 nm CW laser, 8.5 W/cm² over 4 days (2.94 MJ/cm² received) | Responsivity at 244 nm rose about 4.5% rather than falling; it had not decreased after five months’ storage in nitrogen. The note presumes surface-cleaning effects from the high-intensity beam |
| 100 eV photons, total fluence above 1022 photons/cm² | A line scan of the surface before and after showed no change in response within measurement uncertainty |
| 185 nm lamp, 6.53 mW/cm² for one month (720 hours) | The SXUVPS4C quadrant detector and a second SXUV part showed no significant change in responsivity; a UVG-series comparison diode of the same period degraded 25% over the same exposure. The current UVG line uses a radiation-hard oxynitride entrance window — see UVG principles |
The note states the pulsed guarantee twice: less than 3% variation in response at a total fluence of 1.6 × 105 J/cm² of 157 nm pulses for units in the field, and the same 3% at that same total fluence delivered at 100 mJ/cm² pulse energy density. Adding a directly deposited front metal filter does not significantly affect stability — a filtered device held its response after roughly 4 × 1020 photons/cm² of 60–110 eV exposure. The SXUV100TF135 is the current filtered part.
Pulsed measurement: the case for silicon over pyroelectrics
The SXUV series was developed specifically for high-flux sources — excimer lasers and third- and fourth-generation synchrotrons. The note’s comparison with pyroelectric detectors is direct: pyroelectrics offer only five or six orders of magnitude of dynamic range and show significant non-uniformity across the surface, whereas SXUV photodiodes provide more than eight orders of magnitude of dynamic range and better than 2% uniformity, with the accuracy, reliability, compact size and low cost of a solid-state part.
Measuring pulses brings its own constraints. Saturation can appear once source energy density exceeds about 1 µJ/cm². Applying reverse bias raises the saturation threshold and shortens the rise time, and the correct amount of bias is found empirically: increase it until the area under the voltage-time curve stops growing, which indicates the detector is operating in its linear region and all photo-generated charge is reaching the external circuit. Applied reverse bias must not exceed the detector’s breakdown voltage. A capacitively coupled bias tee with a DC blocking capacitor is used to inject that bias; the note records an approximate 5% signal loss in that bias tee that must be accounted for in absolute measurements. The note names a specific bias tee, and Opto Diode’s Proper Use of AXUV, SXUV, and UVG Detectors still recommends an external bias tee of that type for pulse-energy measurement. It is not a current catalog item — contact Opto Diode for a current recommendation.
Calculating energy per pulse
The photo-generated charge Q is proportional to the area under the voltage-time curve, so with an oscilloscope that integrates the trace:
| Step | Relation | Terms |
|---|---|---|
| Charge from the trace | Q = ∫V(t)dt → Q = Av/R | Av is the area of the time-integrated voltage signal; R is the shunting resistance — the scope input impedance, or the feedback resistance if an operational amplifier is used |
| Energy per pulse | Energy/pulse = Q · Ep / QE | Ep is the photon energy in eV; QE is quantum efficiency in electrons generated in the external circuit per incident photon. Q in coulombs with Ep as a number in eV gives energy in joules |
This is the method the note prescribes for absolute pulse-energy measurement, and it is why the 5% bias-tee loss above matters: it enters directly into Q.
Temperature behaviour
Two temperature dependencies are characterized. Shunt resistance falls steadily as temperature rises — the note reports a factor-of-two decrease for every 6 °C — which is the ordinary behaviour of a silicon junction and the reason low-light measurements are kept near room temperature. Responsivity at 254 nm decreases by approximately 0.1% per degree Celsius; the note ascribes that decline, presumably, to an increase in surface recombination velocity at higher temperatures — SXUV photodiodes have high surface recombination, which the note says is indicated by their low responsivity.
Uniformity is quoted as within a couple of percent when scanned with a 0.1 mm × 0.5 mm beam. For absolute measurement, keep the whole beam inside the active area: Opto Diode’s Proper Use note warns that an overfilled or flooded diode puts light on the periphery, where quantum efficiency is lower and unknown.
Where SXUV sits among the families
Opto Diode builds three silicon detector families for the ultraviolet and shorter wavelengths, and they are not interchangeable:
| Family | Optimized for | Range | Principles |
|---|---|---|---|
| SXUV | Holding response after EUV and UV exposure — dose and power monitoring | 1–190 nm | This page |
| AXUV | Windowless broadband response and 100% internal quantum efficiency | 0.0124–190 nm | AXUV principles |
| UVG | UV-enhanced silicon for DUV and UV measurement | 190–400 nm | UVG principles |
The AXUV vs SXUV selection guide walks through that choice in more detail. Current SXUV models are listed on the SXUV category page, with dedicated specification pages for the SXUV100, the filtered SXUV100TF135, the high-speed SXUV20HS1 and the SXUVPS4 and SXUVPS4C quadrant detectors. Before handling any windowless part, review Handling Precautions and Proper Use of AXUV, SXUV, and UVG Detectors.
Frequently asked questions
What makes an SXUV photodiode radiation-hardened?
+
The platinum-silicide entrance window. UV-exposure-induced instability in common silicon photodiodes is caused by their front silicon-dioxide window; the PtSi window replaces it, eliminating that instability mechanism. NIST and LBL testing found no change in quantum efficiency after 10 eV photons at 10 to the 16th photons per square centimeter or 100 eV photons at 10 to the 22nd, which implies a radiation hardness of hundreds of Gigarads in silicon.
Why is SXUV responsivity lower than a UV-enhanced detector’s?
+
Because the two families are optimized for different things. SXUV trades absolute responsivity for stability under dose; the source note attributes its temperature behaviour to high surface recombination. If your measurement needs maximum UV sensitivity and the accumulated dose is modest, a UVG detector is the better choice. If calibration has to survive heavy exposure, SXUV is.
How do I measure pulse energy with an SXUV photodiode?
+
Integrate the voltage-time trace to get charge, dividing the integrated area by the shunting resistance, then convert to energy using the photon energy and the quantum efficiency. Apply enough reverse bias that the area under the curve stops increasing, which shows the detector is in its linear region, and never exceed its breakdown voltage. If a bias tee is used, account for roughly 5 percent signal loss in it when making absolute measurements.
At what point does an SXUV detector saturate on pulses?
+
Saturation can become apparent once the source energy density exceeds about 1 microjoule per square centimeter. Reverse bias raises that threshold and also reduces rise time, which is why pulsed applications bias the detector rather than running it at zero volts.
Do filters affect SXUV stability?
+
Not significantly. The source note reports that adding directly deposited front metal filters does not meaningfully change the stability of SXUV devices; a filtered detector held its response after roughly 4 times 10 to the 20th photons per square centimeter of 60 to 110 eV exposure. The current filtered part is the SXUV100TF135, which integrates a thin-film filter with a 12 to 18 nm detection range covering the 13.5 nm EUV wavelength.
Specifying an EUV dose or power monitor, or deciding between SXUV, AXUV and UVG? Opto Diode typically responds to online requests within 1–2 business days.
Reviewed and updated August 2026.