Chernobyl — Reactor No. 4, 26 April 1986

The RBMK-1000 (Reaktor Bolshoy Moshchnosti Kanalny, "high-power channel reactor") is a low-enriched-uranium reactor (~2 % U-235), graphite-moderated and cooled by boiling light water. Designed in the 1960s at the Kurchatov Institute for the dual mission of electricity generation (1000 MWe) and military plutonium production — hence the absence of a leak-tight containment, incompatible with on-load refuelling. The core is a vertical cylinder 11.8 m in diameter and 7 m high, formed by a stack of 2,488 columns of graphite blocks (250 × 250 × 600 mm). At the centre of each column runs a vertical zirconium-niobium (Zr-2.5%Nb) pressure tube. Across the 1,661 fuel channels and 211 control rod channels, water rises from the bottom to the steam separators, then to the turbines. Four reactors of this type were operating at Chernobyl in 1986. Unit 4, commissioned in December 1983, incorporated the "2nd generation" RBMK-1000 improvements: control rods fitted with a boron carbide absorber and a graphite "displacer" at their lower end. This cosmetically benign modification would play a decisive role in the accident on 26 April 1986.

**Positive void coefficient (α_v).** In an RBMK, the moderator (graphite) does not move with the coolant (boiling water). Yet water, in addition to removing heat, absorbs a fraction of the thermal neutrons. When the steam fraction in the channels increases — through a pressure drop, more boiling, or merely a flow reduction — the share of neutrons captured by water drops, while moderation continues to be provided by the graphite. Net result: reactivity rises. At nominal power, INSAG-7 cites α_v = +4.7 β (β being the delayed neutron fraction, ≈ 0.007). At low power and with few rods in the core, this coefficient climbs to ≈ +5.5 β — a violently positive thermo-hydraulic feedback. **The graphite tip of the AZ-5 rods.** RBMK control rods comprise a 7 m B₄C absorber sitting above a 5 m water column, and preceded at the bottom by a 4.5 m graphite displacer. When a rod is almost fully withdrawn, most of the column above the core is water; below, graphite occupies the entire active height. The instant the rods are dropped (AZ-5 signal), the graphite displacer enters the bottom of the core, displacing the neutron-absorbing water. During the first seconds of the drop (~18 to 21 s for 7 m), graphite insertion adds reactivity (the "end effect"). Only beyond ~2 m of penetration does the absorber dominate. INSAG-7 estimates this effect at +0.2 β to +1 β depending on core state. These two flaws, individually manageable by an alert operator, become fatal when combined with a core depleted in rods and poisoned by xenon — which was the case at 01:23:40 on 26 April 1986.

On 25 April 1986, Unit 4 was scheduled for annual maintenance. The operator planned to use the shutdown to run an electrical test requested by the turbine manufacturer: verify that, in case of total loss of off-site power, the residual inertia of the turbine could power the main coolant pumps during the ~45 seconds needed to start the backup diesels. This "coastdown" test is purely electromechanical and should not involve neutronics. By 14:00, power had been brought down to ~1,600 MWth. The Kiev grid dispatcher then refused the full shutdown: the grid was short of generation. The reactor stayed at 1,600 MWth for 9 hours, in a configuration where the test could not happen. During those 9 hours, xenon-135 — a fission product with a very large capture cross-section (σ ≈ 2.6 × 10⁶ b at 0.025 eV) — accumulated in the core because its production from I-135 decay remained high while its destruction by neutron capture dropped (fewer neutrons at reduced power). This is "xenon poisoning". At 23:10, the test could resume. The setpoint was 700 MWth. The operator overshot the reduction, power collapsed to ~30 MWth — at 1 % of rated power, the reactor was in an unstable regime where the effective void coefficient is even more positive. To recover, the crew massively withdrew the control rods. At 01:22:30, the automatic SKALA program indicated that the operational reactivity margin (ORM, measured as equivalent fully-inserted rods) was 6-8 rods, while the procedure required an immediate shutdown below 15. The crew proceeded with the test anyway.

**01:23:04 — t₀.** The operator closes the fast-stop valve (MPA) of turbo-generator No. 8, isolating the steam flow. The coastdown test begins. The main pumps, now driven by the turbine's inertia, see their flow decay. The steam fraction in the channels rises, and with it reactivity (positive α_v). The automatic regulation system compensates by lowering a few rods slightly; power remains apparently stable around 200 MWth. **01:23:40 — AZ-5 button.** Shift engineer Akimov orders Toptunov to press the AZ-5 ("emergency shutdown") button. The 187 withdrawn rods begin their descent at 0.4 m/s. During the first 3 seconds, the graphite tip enters the bottom of the core and displaces the absorbing water. Combined with the positive void coefficient and the resulting power surge, the effect makes the core prompt-critical: power doubles every ~0.5 s. **01:23:43 — power peak.** In less than 4 seconds, power rises from 200 MWth to roughly 320 times rated power (≈ 1,000,000 MWth instantaneous peak). The DREG-3 recordings saturate; the value is reconstructed from irradiation indices of structures (γ, isotope ratios). UO₂ fuel vaporises locally, ruptures the zircaloy cladding and produces, within milliseconds, a violent steam expansion in the channels. **01:23:45 — first explosion (steam).** Channel pressure exceeds the strength of the pressure tubes. Simultaneous rupture of many tubes pressurises the reactor block and lifts the upper biological shield "Elena" (~2,000 t, plate E), which falls back across the vault. The steam manifolds are severed. This first explosion is mechanical, driven by superheated steam — not a nuclear explosion in the weapons sense. **01:23:48 — second explosion.** Two to three seconds later, a second explosion ejects incandescent graphite blocks and fuel up to ~100 m. INSAG-7 and NEA 1996 retain two hypotheses: (a) a chemical detonation of hydrogen produced by Zr-H₂O oxidation of cladding, mixed with air entering through the breach, or (b) a second localised neutron excursion in core fragments. The reactor building roof is destroyed. A graphite fire ignites — graphite, in air above 700 °C, burns. The fire was only mastered on 10 May 1986, after the helicopter delivery of ~5,000 t of materials (sand, boron, lead, dolomite, clay).

Atmospheric releases spanned 10 days (26 April – 6 May 1986). The official UNSCEAR 2008 total is ≈ 14 EBq (14 × 10¹⁸ Bq), including noble gases. Notable nuclides: noble gases Xe-133 and Kr-85 (near-total core release), iodine I-131 (≈ 1,760 PBq, ≈ 50 % of core), caesium Cs-134 and Cs-137 (≈ 85 PBq Cs-137, ≈ 33 % of core), strontium Sr-90 (≈ 10 PBq), and transuranics including Pu-239 (≈ 0.03 PBq). Weather conditions — southeasterly wind then rotation to northwest, fire-driven updrafts — lifted the plume to 1,500 m, sparing the immediate surroundings but carrying it across Belarus, the Baltic states and Scandinavia. On 28 April, monitors at Sweden's Forsmark plant detected contaminated shoes on incoming staff: this is how the world learned of the accident, the USSR having said nothing. Three Cs-137 soil contamination zones were defined by the 1990 EU/CIS-CE convention: > 1,480 kBq/m² (mandatory evacuation), 555-1,480 kBq/m² (voluntary resettlement), 37-555 kBq/m² (strict control). The 30-km exclusion zone covers ~2,600 km². 29 villages were buried.

**Acute effects.** 134 plant staff and liquidators developed confirmed acute radiation syndrome (ARS). 28 died within 4 months of the accident, plus 2 additional trauma victims from the explosions. UNSCEAR 2008 thus retains 30 direct deaths. **Thyroid cancers.** Iodine-131, with a half-life of 8.02 d, concentrates in the thyroid via diet (milk above all). Among children and adolescents (< 18 in 1986), thyroid cancer incidence in Belarus, Ukraine and Russia rose statistically significantly from 1991 onwards. UNSCEAR 2008 documents ~6,000 excess cases over 1991-2005, including 15 deaths. Nearly all of this epidemic could have been prevented by prophylactic distribution of potassium iodide in the days following the accident. **Other cancers and stochastic effects.** Beyond thyroid cancers, no statistically significant rise in cancer incidence or leukaemia attributable to Chernobyl has been demonstrated in exposed populations outside the ~530,000 liquidators. Long-term mortality estimates rest on applying the linear no-threshold (LNT) model to low collective doses — a method the IAEA and UNSCEAR caution against below 100 mSv. WHO 2006 cites a range of 4,000 additional deaths among the 600,000 most-exposed individuals (liquidators + 30-km zone + severely contaminated areas); other models extrapolate up to 16,000 across Europe — both figures must be handled with the methodological caveats their authors recommend. **Non-radiological effects.** WHO 2006 highlights the psychosocial impact: depression, anxiety, PTSD, alcoholism, social rupture from forced relocation. In public health terms, this is the broadest and least anticipated effect.

The original "sarcophagus" (Object Shelter), built in 206 days between May and November 1986 by 90,000 workers, is an emergency concrete-and-steel structure poured over the ruins of Reactor Building 4. Its rated lifetime was 30 years, extended by successive sealing campaigns. It encloses ~200 t of dispersed nuclear fuel, part of which exists as "FCMs" (fuel-containing materials, the Chernobyl lavas — molten silicates loaded with U, Zr, Pu). The New Safe Confinement (NSC, or "arch"), built by the Novarka consortium for €1.5 billion, is a steel arch enclosure 108 m high, 162 m long, 257 m span and 36,200 t. It was assembled 180 m from the building then slid on rails over the sarcophagus in November 2016 — the largest structure translation ever performed. Its rated lifetime is 100 years. It will enable progressive decommissioning of the sarcophagus and FCM extraction. Final extraction and conditioning of the FCMs has not yet begun as of 2024 — the war in Ukraine has interrupted international cooperation on site. The volumes involved (~200 t of highly radioactive mixtures) impose a realistic horizon of several decades.

**On RBMKs.** The 11 RBMKs still in service in 1986 were modified: enrichment raised from 2.0 to 2.4 then 2.8 % U-235 (reduces the fissile plutonium share and thus the void coefficient), control rods fitted with additional absorbing tips (eliminates the end-effect), minimum ORM raised to 30 rods, rod drop speed doubled. No RBMK was built after 1986. The last Ukrainian RBMK (Chernobyl 3) was shut down in December 2000. The last operating worldwide is Smolensk-3 (Russia, operating until ~2030). **On safety generally.** Creation in 1989 of WANO (World Association of Nuclear Operators) for cross-border operating experience sharing. Convention on Nuclear Safety (1994) mandating peer review. INES scale adopted in 1990. Defence in depth doctrines reformulated (INSAG-10, 1996; INSAG-12, 1999) with seven levels and symptom-based rather than event-based EOPs (Emergency Operating Procedures). **On crisis communication.** Initial Soviet opacity (official announcement 36 h after the accident, Pripyat evacuation 36 h after) remains the archetypal counter-example. Post-Chernobyl Western reactors publish their INES within 24 h. Fukushima Daiichi (March 2011, INES 7) was announced publicly the same day.