Before instant film and digital flat-panels, radiographs were viewed on fluorescent screens and the patient’s data were scribbled on the dark-room door. In 2026 a German medical museum discovered that every X-ray tube manufactured between 1928 and 1942 carries its own hidden archive: the high-voltage strike that generated each exposure carved a micro-fusion channel into the lead-glass envelope, and the instantaneous tube current modulated the channel’s refractive index. Using femtosecond polarimetry and a plasma-physics inverse solver, researchers recovered 1,200 patient exposure logs from a single Siemens “Röntgenglas” tube—dose, kVp, even the radiographer’s pulse-coded patient ID—turning a cracked vacuum bulb into a write-once medical drive.
Lead-silicate glass (28 % PbO) is a favourite for early X-ray tubes because its high electron density stops stray radiation. When 80 kV surges across the anode-cathode gap, field emission ignites a 10 A, 100 µs plasma filament that locally melts the inner wall. During cool-down the glass vitrifies around the frozen shock wave, trapping birefringent stress lamellae whose slow-axis angle is proportional to the instantaneous current. Because the filament retraces the same path for thousands of exposures, the lamellae stack like tree rings, each 40–80 nm thick, encoding a chronological log of tube usage.
Reading the signal starts by sectioning the glass cylinder under argon to prevent hydration stress. A 5 × 20 mm strip is polished to λ/10 flatness and placed in a 790 nm Ti:sapphire pump-probe ellipsometer. The probe pulse (12 fs, 50 nJ) senses retardance with 0.02° sensitivity, mapping the slow-axis orientation at 200 nm lateral resolution. A full helical scan produces a 3-D tensor field; integrating along the radius yields a 1-D current trace sampled at 10 kHz—sufficient to resolve the 20 ms exposure window and the 50 Hz mains ripple that modulated the tube current.
Clock recovery exploits the hospital timetable. Radiographers worked 12-hour shifts; tube usage clusters at 09:00 and 21:00, visible as intensity spikes. Cross-correlation with archived ward rosters (kept in municipal archives) aligns the current trace to the calendar day. One anomalous 03:17 exposure coincides with a documented emergency tracheotomy on 14 March 1938, confirming temporal accuracy to within one minute.
Error correction utilises dose redundancy. Chest exams used 70 kVp, skulls 95 kVp; the birefringence magnitude scales with kVp². Segments whose spectral centroid deviate from known protocols are flagged as drop-outs and interpolated. After filtering, the RMS difference between recovered dose and archived film-badge records is 6 %, well within the ±10 % tolerance of 1930s ionisation chambers.
Storage density is modest but medically priceless. A 30 cm tube stores ~800 kB of exposure meta-data—enough for 1,200 high-quality radiographs. Across the estimated 60,000 pre-1945 tubes held in European hospitals, the potential archive is 48 GB of unique dose information, invaluable for retrospective epidemiological studies on early radiation risk.
Restoration is non-destructive; the glass strip is re-bonded with UV-curable epoxy, returning the tube to display-worthy condition. Legal ownership follows EU medical-device directives: the tube is waste, but the data, being non-tangible, are released 75 years after creation, ensuring patient privacy.
For medical historians the lesson is clear: every cracked X-ray bulb is a flight-data recorder of the clinic. Beneath the leaded glass and soot lies a crystallised vein where the spark of every diagnostic exposure still flickers, waiting for the right polarised pulse and the right plasma equation to step out of the vacuum and back into the ward ledger.
