In September 1987 two men scavenging in a partly demolished radiotherapy clinic in Goiania, Brazil, removed the source assembly from a caesium-137 teletherapy machine and sold it to a scrap dealer. The capsule was broken open, the caesium chloride inside it glowed blue in the dark, and fragments were handed around among family, employees and neighbors for more than two weeks before anyone recognized what it was. Four people died, roughly 112,000 were monitored, and several houses were demolished. Goiania remains the reference case for what happens when a licensed source stops being under anyone’s control.

September 1987: what was left in the abandoned clinic

A private radiotherapy institute in Goiania, capital of the Brazilian state of Goias, moved out of its premises in 1985 and left a caesium-137 teletherapy unit behind in the building. The circumstances of that abandonment were tangled up in a legal dispute over the property, and the accounts in the record describe correspondence with courts and with Brazil’s nuclear regulator, the Comissao Nacional de Energia Nuclear, along with a partial demolition of the structure and a guard who was not always present. The International Atomic Energy Agency published a detailed account of the accident in 1988 under the title The Radiological Accident in Goiania, and anyone who wants the primary narrative should read that report rather than take my summary as the record.

On 13 September 1987 two men entered the ruin looking for scrap metal and removed the rotating source assembly from the machine, moving it away in a wheelbarrow. Over the following days they worked on it with hand tools at a home nearby, trying to get at the metal, and in the process they breached the source capsule. Both men became ill within a short time with vomiting and diarrhea, which neither they nor the people who treated them connected to what they had been handling.

The IAEA account gives the activity of the source at the time of the accident as about 50.9 terabecquerels, which is roughly 1,375 curies of caesium-137. That is a quantity used to deliver therapeutic radiation doses to cancer patients under tight geometric control, inside a shielded head, on a machine operated by trained staff. Removed from the head and broken open, the same material sat on a workbench and later in a living room.

Five days after the assembly was taken, the pieces were sold to a scrap yard in the city. The owner of that yard noticed at night that the material inside was giving off a blue glow, and what happened next is the part that made this a public health emergency rather than an industrial accident with two casualties.

What an orphan source is, and how one gets made

The IAEA uses the term orphan source for a radioactive source that is not under regulatory control, whether because it never was, or because it has been abandoned, lost, misplaced, stolen or transferred without proper authorization. The definition is deliberately broad, because the practical problem is identical in every case: a quantity of material capable of injuring people exists somewhere, and nobody with authority knows where it is or is doing anything about it.

The IAEA also publishes a categorization scheme that ranks sources from Category 1, the most dangerous if not managed safely, down to Category 5. Teletherapy heads, large industrial irradiators, self-contained blood irradiators and radioisotope thermoelectric generators sit at the top of that scale, and industrial radiography sources are close behind. The categories are about the harm the source could do to a person who handles it, not about how the source is used, which is why a medical device and an industrial gauge can end up in the same risk bracket.

Sources rarely go missing dramatically. The common route is a facility that stops using a device and keeps it, because disposal costs money and there is no immediate pressure to spend it. Over the years the licensee’s staff turn over, the person who knew what was in the back room retires, a company is sold or goes bankrupt, and the paper record and the physical object drift apart. Theft happens too, usually for the metal value of the shielding rather than for the material inside, and armed conflict or the collapse of a regulatory authority orphans sources on a larger scale.

Goiania is not an isolated case. A cobalt-60 source that ended up in a Mexican scrap stream in 1983 and was distributed as contaminated rebar is covered separately on this site. A teletherapy head sold for scrap in Samut Prakarn, Thailand, in 2000 killed three people, and a cobalt-60 research irradiator auctioned from a university in Delhi and cut up in the Mayapuri scrap market in 2010 killed one man and injured several others. The pattern in each case runs from disuse to storage to sale to disassembly.

Blue powder, and contamination carried by ordinary contact

The caesium in the source was in the form of caesium chloride, a salt that is highly soluble and behaves as a fine powder when the ceramic matrix holding it is broken up. It also luminesces, which is why the material appeared to glow blue in a darkened room and why people found it interesting rather than frightening. The scrap yard owner brought pieces into his house, showed them to visitors, and gave fragments away to family members and friends over several days.

What followed was contamination moving along the routes that ordinary social contact provides. People carried fragments home in their pockets and on the bus. Powder transferred to hands, then to food, then to mouths. It got onto clothing and into bedding, onto floors and furniture, into the money that changed hands at two scrap yards, and onto the ground where pieces were dropped. A six-year-old girl, Leide das Neves Ferreira, was reported to have eaten while handling the material, and she received a large internal dose that contributed to her death the following month.

This is the feature of an orphan source event that makes it an emergency management problem rather than only a regulatory one. A contaminated person is a moving source of contamination who feels well for hours or days, who has no reason to isolate, and who will visit relatives, ride public transit, go to work and go to a clinic in the meantime. The dispersal follows the social network of the people who touched the material first, so tracing it requires interviewing people about who they visited and what they touched, which is public health contact tracing rather than a radiation survey.

Contamination in Goiania was eventually identified at homes, at both scrap yards, on vehicles, on public roads and in soil, and the survey work had to be done across a populated urban area rather than at a single site with a fence around it. None of that would have been necessary if the source capsule had remained intact, and the difference between a sealed source found on the ground and a breached one is the difference between a cordon and a citywide operation.

Sixteen days to recognition, and the misdiagnoses along the way

People began presenting to medical care within days of the source being opened, with nausea, vomiting and diarrhea, and later with skin lesions on the hands and forearms. Those early presentations were attributed to food poisoning and to infectious and tropical illnesses, and some patients were sent to a hospital for tropical diseases where the skin injuries were treated as an allergic reaction. Acute radiation effects look like a great many other things in the first few days, and there was no reason for anyone treating those patients to have radiation in the differential.

The break came on 29 September 1987, sixteen days after the assembly was removed from the clinic. Maria Gabriela Ferreira, the wife of the scrap yard owner, had concluded that the material was making people ill, retrieved what remained of it, and took it in a bag on a city bus to a public health facility. A medical physicist who was in the area obtained a suitable instrument, established that the readings were extraordinary, and intervened to keep the bag from being carried away again. Both the local authorities and the national regulator were engaged within about a day, and the national response began.

The sixteen days matter because everything that made the response expensive happened during them. The material was fragmented and distributed, doses accumulated, contamination reached houses and streets, and the population of people who needed to be found and monitored grew steadily larger. There is nothing in the record to suggest anyone acted maliciously, and the delay is entirely explained by the fact that the hazard was invisible, the material looked valuable, and the illnesses it produced resembled common local diseases.

Maria Gabriela Ferreira died in October 1987 of the injuries she received while the material was in her home. Her decision to carry the remains to a health authority is the single act that ended the exposure, and it came from a member of the public rather than from any part of the regulatory or medical system.

Monitoring a city, and the casualty figures

Brazilian authorities set up a monitoring operation at the Olympic stadium in Goiania because a stadium was the only structure available with the throughput, the parking and the queuing space to handle a self-referring population. The IAEA account gives the number of people monitored as approximately 112,000, with some sources citing 112,800, out of a metropolitan population then around a million. Aerial and vehicle-mounted gamma survey work was used to find contaminated locations across the city, which is how several contaminated sites were located after the initial interviews had run out of leads.

Of those monitored, the IAEA account gives 249 people as contaminated internally or externally. Internal contamination was identified in 129 of them, and 20 people were sick enough to require hospital treatment, with bone marrow depression developing in 14 of those admitted. Prussian blue was administered to accelerate the excretion of caesium from those with significant internal contamination. Four people died in the weeks after recognition: Maria Gabriela Ferreira, the six-year-old Leide das Neves Ferreira, and two young men who had worked at the scrap yard. Some published tallies count later deaths among the exposed population as attributable, and that attribution is contested, so treat four as the accepted acute toll and treat anything higher as a claim that needs its source named.

The ratio is the part I would ask any emergency manager to sit with. Roughly 112,000 people presented for monitoring and 249 were contaminated, which is about 450 people screened for every person found. That was not irrational behavior by the public, because nobody in the city could know whether they had been in a contaminated house or on a contaminated bus, and the only way to find out was to be checked. Any radiological incident in a populated area generates that demand, and the monitoring operation, not the medical treatment, is where the bulk of the personnel, the instruments and the days go.

Decontamination, demolition, and the waste

The remediation work in Goiania ran through the last months of 1987 and into 1988 and covered houses, yards, streets, public places and the two scrap yards. Figures commonly cited from the IAEA account are that 85 houses were found to be significantly contaminated, that 41 of these were evacuated, and that 7 were demolished outright because decontamination was not practical. Roofs, walls, floors, furniture and personal effects were removed, and topsoil was excavated where the material had reached the ground.

The waste figures published for this event use different measures and do not reconcile cleanly. The IAEA account describes a volume on the order of a few thousand cubic meters, and the number most often quoted from it is 3,500 cubic meters. Tonnages in the thousands of tonnes appear in other published accounts, and those are not the same measurement of the same thing, so if you need a figure for a briefing, state the volume, name the IAEA report as the source, and say plainly that tonnage figures elsewhere are measuring differently.

The waste was moved to a site near Abadia de Goias, some tens of kilometers from the city, where a repository was built to hold it. Caesium-137 has a half-life of about 30 years, so the material placed there in the late 1980s has passed roughly one half-life and remains a managed radiological waste site requiring institutional control well beyond the working life of anyone involved in the original response. A single teletherapy source, opened by hand, produced a disposal obligation that outlasts the careers of everyone who cleaned it up.

The demolitions are worth dwelling on for a different reason. Deciding to knock down an occupied family’s house because it cannot be cleaned is a housing and mass care decision, made under a radiological constraint, and it carries all the displacement, compensation and community relations problems that any condemnation carries, on top of a population that is frightened of a hazard it cannot see.

Stigma, economic loss, and the long tail

Within weeks of the accident becoming public, people and products from Goias were being refused elsewhere in Brazil. Reported effects included agricultural goods rejected by buyers in other states, cancelled hotel bookings, and residents of Goiania being turned away or treated with hostility when they traveled. The burial of one of the victims drew protests from residents near the cemetery who feared the grave, and that reaction happened in a city that had just spent weeks watching its own neighborhoods surveyed with radiation instruments.

The exposed population also carried the stigma personally, in employment, in relationships and in medical care, and Brazilian follow-up programs registered and monitored people who had been exposed or contaminated. A victims’ association was formed and pursued recognition and compensation over many years. The long-term medical follow-up of that cohort is published in the scientific literature, and I would point anyone who needs the health outcome data to that literature rather than to secondary accounts, because the doses varied enormously between individuals and any summary statement flattens that.

Economic damage estimates for the event circulate widely and I am not going to publish one, because I have not found a figure I can attribute to a specific study with a specific methodology and a date. What can be said without qualification is that the costs fell into categories most emergency plans do not budget for at all: mass monitoring of a metropolitan population, demolition and replacement of housing, long-term waste custody, lost agricultural sales across a state, and multi-decade medical surveillance of a defined cohort.

Source security, tracking and reporting today

The international framework built after Goiania and after the Soviet collapse centers on the IAEA’s Code of Conduct on the Safety and Security of Radioactive Sources, whose current form dates from the early 2000s, together with supplementary guidance on the import and export of sources. It is not a treaty, and it works through national commitments, national registries of the higher-category sources, and cradle-to-grave accountability that ties a source to a licensee from manufacture through disposal. The IAEA also maintains an incident and trafficking database to which member states report losses, thefts and recoveries.

In the United States, radioactive materials are regulated by the Nuclear Regulatory Commission and by the Agreement States, which regulate most materials licensees within their borders under agreement with the NRC. The NRC operates a National Source Tracking System for the higher-category sources, imposed additional security requirements on Category 1 and 2 licensees in the years after 2001, and maintains a database of material events including losses and thefts. Licensees are required to notify the NRC or the Agreement State by telephone when licensed material is lost or stolen, with the required promptness depending on the quantity involved, and you should verify the current requirement and the current phone number with your state radiation control program rather than from an article.

Disposal pathways exist specifically to keep disused sources from becoming orphans. The Off-Site Source Recovery Program, run through Los Alamos National Laboratory under the National Nuclear Security Administration, recovers and dispositions certain unwanted sealed sources, and the Conference of Radiation Control Program Directors has run a source collection program that helps licensees dispose of disused sources at reduced cost. The NNSA’s radiological security work also funds voluntary physical security upgrades at hospitals, universities and industrial licensees, and has funded replacement of caesium blood irradiators with x-ray devices that present no orphan risk at all.

On the response side, the technical authority in most states is the state radiation control program, backed by the Department of Energy’s Radiological Assistance Program teams, which are regionally based and can deploy with survey and identification capability. The FBI takes the lead where diversion is criminal, National Guard Civil Support Teams carry radiological detection and identification equipment, and REAC/TS in Oak Ridge provides 24-hour medical consultation on radiation injury and contamination. Scrap yards, landfills and steel mills with portal monitors are, in practice, one of the main places orphan sources are found in this country, which makes their alarm response procedures part of the public safety system whether anyone wrote that down or not.

What to do at your agency

• Have the training officer pull every personal radiation detector and survey instrument off the apparatus this month, confirm batteries and a documented function check, and put that check on the same recurring schedule as the gas meter bump test so it is not a special event.
• Have the emergency manager request from the state radiation control program a current list of the licensed radioactive materials facilities in the county, and separately identify every scrap metal recycler, landfill scale house and steel facility in the jurisdiction that operates a radiation portal monitor.
• Ask the operations manager at your largest scrap or metals facility for a copy of their portal monitor alarm response procedure and the phone number they are told to call, and add one line to the dispatch protocol so that a call from that address gets the right response and the right notifications the first time.
• Build a single notification card listing the state radiation control program’s 24-hour number, the DOE Radiological Assistance Program regional contact, the NRC Operations Center, the FBI field office duty number and REAC/TS, verify each number by calling it during business hours, and place copies in the EOC, in dispatch and in the battalion chief’s vehicle.
• Put population monitoring on the agenda of the next scheduled exercise design or LEPC meeting, name a specific building with parking and queuing space, and have the planner write down the assumed screening throughput per hour so the plan can be tested against a realistic self-referring crowd rather than against the expected number of contaminated people.
• Have the communications officer confirm that the state radiological response team and the DOE RAP team have a usable interoperability channel or talkgroup in your system, write the assignment into the communications annex, and test it once on a scheduled interoperability drill rather than discovering the gap on the day.
• Ask the EMS medical director to confirm with the receiving hospital’s emergency department that a radiation injury consultation number is posted where the charge nurse can find it, and that the department knows contamination is a decontamination problem that does not delay life-saving care.

Takeaways

• The Goiania accident began on 13 September 1987 when two scrap scavengers removed a caesium-137 teletherapy source assembly, reported by the IAEA at about 50.9 terabecquerels or roughly 1,375 curies, from a radiotherapy clinic that had been abandoned since 1985.
• The source material was caesium chloride, a soluble salt that behaved as a powder once the capsule was breached, which is why the contamination spread through households, scrap yards, buses and money rather than staying at one site.
• Sixteen days passed between the removal of the source and its recognition on 29 September 1987, and the early illnesses were attributed to food poisoning and to infectious and allergic causes because acute radiation effects resemble common conditions.
• Approximately 112,000 people were monitored, with 249 found to be contaminated and 129 of those internally contaminated according to figures in the IAEA’s 1988 report, and four people died in the weeks that followed.
• The remediation involved 85 significantly contaminated houses, 41 evacuations and 7 demolitions in the commonly cited counts, and produced a waste volume most often given as about 3,500 cubic meters, now held at a repository near Abadia de Goias, while tonnage figures published elsewhere are measuring something different.
• Consequences that emergency plans rarely budget for dominated the aftermath, including mass population monitoring, housing demolition, statewide rejection of agricultural products, stigma against exposed individuals, and decades of medical follow-up and waste custody.
• Sources become orphans mainly during disuse, when a device goes out of service and disposal is deferred until staff turnover and lost records separate the object from the paperwork, which makes prompt disposal the control that actually prevents these events.
• In the United States, orphan sources are most often found by portal monitors at scrap and waste facilities, and the working response chain runs from the state radiation control program through DOE Radiological Assistance Program teams, with the FBI involved where diversion is criminal.