Characterization for Decommissioning: Yes, But to What Extent?

By Michele Laraia, Independent Consultant

Michele Laraia’s series of articles are intended to start a conversation and stimulate debate. We invite you to comment or ask questions below (or send us an email) regarding this article either using your name or anonymously, and we will print some of the comments/questions in an upcoming Weekly Digital Report, as well as in the article itself and on the forum. Laraia will also respond to any comments/questions from readers. Then, stay tuned next month for a new article and new debate.


A comprehensible and verifiable determination of the radiological inventory is essential for the decommissioning management with respect to safety, time and costs. For example, right from the start of the post operational phase, the radiological characterization has to enable the decision whether or not to perform a system decontamination, or to investigate waste disposal pathways. Furthermore, it is necessary to determine the relevant nuclides and their composition (nuclide vector) for the release of material and for sustaining the radiological health and safety at work (e. g. minimizing the risk of incorporation).

The following anecdote became popular years ago (I think Tom La Guardia made it up). “To the question: what are the three most important aspects of decommissioning, the right answer is: characterize; characterize; characterize.”

Characterization is essential to decommissioning planning and implementation: BUT…

  • Characterization is not a goal per se
  • Characterization is only as good as fulfills the decommissioning goals
  • Characterization calls for (not negligible) resources


  • Characterization is not normally a research project (unless somebody pays for it…)
  • Spending months and running costly codes prior to dismantling to differentiate whether a component is activated at 10 or 15 Bq/g is possibly a luxury…
  • ….when considering the whole plant will be eventually demolished. There will always be a chance of monitoring the waste containers before they are shipped to disposal.
  • Sampling the biological shield to the point it looked like Swiss cheese is a waste of time and money…
  • ….besides, any one more sample implies worker doses and generation of radioactive waste (not to mention the extra costs).
  • But characterization is essential. Taking samples from the bioshield ensures a good segregation of active from inactive parts, and as the bioshield may weigh say 800 t, there is a lot to gain from segregation and (rad) waste minimization.

The Trade-Off

Staying with the bioshield example, a good measure of conservatism would ensure that segregation of active from inactive parts is effective. For example, if the computer model—validated by a sufficient number of samples—shows that the activated depth is 50 centimeters from the inner side (out of 300 centimeters), but isolated samples show that 5-10 centimeters more can be activated, a conservative assumption can be that 60 centimeters are activated and should be disposed of as radioactive. Of course the degree of conservatism will depend on the risk—and the impact—that local hot spots are found in the supposedly clean zone of the bioshield.

The clearance regulations normally include averaging criteria. If a depth of few centimeters of concrete is slightly activated (say 2 Bq/g against a limit of 1 Bq/g) and is mixed in the disposal container with 20 centimeters’ depth of clean concrete resulting from bioshield demolition, the mix may (or may not, depending on your regulators) be acceptable as non-radioactive. Clear procedures and an early dialogue with the regulator will help here.

Beware! A lack of early dialogue between operator and regulator may turn a decommissioning project into a nightmare! See for example the role of averaging criteria and other statistics.

Another example is fingerprinting. Hard-to-detect (HTD) radionuclides are costly to sample and measure. Fingerprinting is a method to correlate concentrations of HTD nuclides to others that are chemically similar, but easier to measure. Co-60 can do this job for Ni-63 (which has a special role in decommissioning). This applies specifically to certain weak-beta emitters. Radiochemical analyses require a specific expertise and a number of separation steps.

However, fingerprinting requires the sampling and lab counting of a sufficient number of HTD nuclides to make up a statistically meaningful correlation factor. Conservativism and optimization play an important role here as well.

At older nuclear plants moving to a decommissioning phase, it is not rare that certain areas have not been surveyed for decades. Pre-decommissioning characterization is imperative to prevent unknowns potentially exposing workers unexpectedly during decommissioning: this is what characterization is about. But, when the area dose rate is well-known, not every single square-centimeter or gram should be individually measured.

One more trade-off case: would you send a team of mountaineers for a month to take radiological samples from the entire surface of a 20-meter-high ceiling when a week-long localized sampling campaign (supported by statistics) can do the job? Industrial safety should be balanced here with accuracy of measurements.

Inevitably a characterization campaign is constrained by factors such as costs, technical skills and instrumentation, clearness of objectives, timing, schedule, regulatory and stakeholder positions, etc. Sometimes, parameters affecting decisions for characterization can be reduced to a common monetary basis, and cost-benefit analysis is possible and desirable. In other cases, parameters belong to non-monetary categories and multi-attribute analysis should be seriously considered. In most cases, however, certain overwhelming factors are decisive and determine the trade-off.

Due to the variable nature of the systems being decommissioned, the sampling procedures are often based on ad hoc principles. The number of samples needed is determined by the historical characterization of the systems. For systems where existing knowledge is scarce, more samples are generally needed earlier in the decommissioning process. The following section describes typical circumstances relevant to optimization in pre-decommissioning characterization.

Pros and Cons of Sampling/Analysis vs Non-Intrusive Methods

Non-intrusive methods (e.g. use of gamma spectrometers) are cheaper and more convenient to use than sampling and analysis methods. For example, a medium-size component can be analysed by gamma spectrometry within tens of minutes and, depending on the materials comprising the component, the assay methodology and counting time, there should be a considerable degree of confidence in the accuracy of the assay. Once the gamma spectrometer has been procured and commissioned, running costs are not substantial. For individual samples, laboratory analyses provide the highest degree of precision. But in the overall context of the whole characterization campaign, sampling and analysis methods may end up being less accurate, mainly because of the difficulties involved in obtaining representative sampling.

When formulating a decommissioning characterization plan, there are a number of considerations to be addressed. One major consideration is the identification of which radionuclides need to be characterized. For example, some waste streams have only small concentrations of Cl-36 (which is difficult to measure), but it can be a key radionuclide in terms of disposal and therefore it needs to be assayed. Analyzing each sample for this radionuclide would be prohibitively expensive, therefore the fingerprint methodology should be used. If a correlation has been established between this radionuclide and say Co-60, both measurements need to be made in conjunction with each other. Thus, both intrusive and non-intrusive assay techniques would be required.


This paper has focussed on the optimization of the radiological characterization with respect to the requisite extent during the decommissioning process. For example, what additional information, besides the history of operation, is essential to determine the relevant nuclides and their compositions? How much characterization is enough to ensure safe and cost-effective decommissioning? What methods and tools should be used?

What do you think? Please leave your comments and contributions below, or email them to us.


LaraiaAbout the Author:

Michele Laraia is a chemical engineer by background. An Italian citizen, he gained his first degree at the University of Rome in 1973. In 1975, he received a post-graduate degree as nuclear engineer. From 1975 to 1991, he worked at Italy’s Regulatory Body (ENEA/DISP), firstly in the capacity of reviewer of radioactive waste management systems, and since 1982 as licensing manager of decommissioning projects. During the 1982-1991 period, under his management seven small research reactors and other nuclear fuel cycle facilities were totally dismantled in Italy and their sites returned to other uses. In other plants, modifications to license conditions were implemented in preparation to decommissioning.

From July 1991 to 2011, Laraia worked at the International Atomic Energy Agency, Waste Technology Section, as Unit Leader responsible for decontamination and decommissioning of nuclear installations, closeout of uranium mining and milling sites, and environmental restoration.
The objectives of the work were to provide advice to Member States on the planning and implementation of adequate methodologies and technologies for decommissioning of nuclear and radiological installations and site remediation, to collect and disseminate information on good practices for safe and cost-effective decommissioning, and to provide direct assistance (through the Technical Co-operation  Programme) to Member States in the implementation of their programes and establishment of the required infrastructure for decommissioning and site remediation, and to strengthen their technical capabilities. His tasks included the drafting of technical publications, organization of international conferences and seminars, and the management of technical cooperation projects with developing countries, either on a national or regional scale. Some 50 technical reports and other documents for dissemination to the international community were prepared by Laraia as the Scientific Secretary, and another few dozen coordinated with him. His publications (journals, conference proceedings) amount to over 100. Over 30 less-developed countries received direct IAEA assistance with Laraia in the role of leading Technical Officer.

Laraia retired from the IAEA in 2011. Since then, he has offered consulting services in nuclear decommissioning including lecturing, training, reviewing decommissioning plans, and drafting of worldwide overviews and topical reports for a number of national/international organizations.

  • Paul Dinner

    A well written and thought provoking article by Michele. It raises in the first few paragraphs the issue of averaging of waste contents for disposal. This leads to the question: “can I add some higher activity waste to this container”, or “this partially filled container looks like it’s going to exceed our WAC, find us some lower activity waste so we can mix it in”. Unlikely this was the intent of permitting averaging over the volume in the first place! I would be interested in opinions as to what constitutes ” best practice” in this regard.


    Knowing characterization objectives and constraints is definitively the key for characterization then decommissioning successes.
    By quantifying spatial prediction uncertainties, geostatistics provides efficient tools for radwaste categorization. It naturally integrates outputs from numerical models (activation, migration, diffusion…) and correctly addresses change-of-support issues (decision volume unit). In addition, geostatistics advantageously optimizes the sampling effort: number of points, locations, balance between HTD nuclides, gamma emitters, count (or dose) rates and prior knowledge…