See original online here

This is a guest post by my colleague Jamie Withorne. Jamie is a Research Assistant at the James Martin Center for Nonproliferation Studies. Her work focuses on emerging technology, North Korea sanctions evasion, and North Korea’s nuclear fuel cycle.

This is the third and final part of a three-part series on North Korea’s uranium conversion program. For Part One, see here. And for Part Two, see here.

Part Three: A Technical Assessment of North Korea Uranium Conversion Processes

As briefly discussed in parts one and two of this series, some steps of North Korea’s uranium conversion line changed processes and locations over time, while others remained consistent.

In addition to providing a historical review of North Korea’s uranium conversion program, it is also important to define what uranium conversion is and what specific processes North Korea might potentially be implementing. Conversion is a chemical process that converts uranium concentrate into uranium tetrafluoride (UF4 ) or uranium hexafluoride (UF6 ) in preparation for fuel fabrication or uranium enrichment. The World Nuclear Association defines conversion as the step in the nuclear fuel cycle between a mine (or a mill, to be more specific) and enrichment.[1]

Figure 1: Diagram of the nuclear fuel cycle [2]

Nuclear Fuel Cycle Diagram
Photo Credit: World Nuclear Association

There are several sub-processes that make up the uranium conversion process. First, uranium concentrate (e.g., yellowcake) is refined to uranium trioxide (UF3 ). The UF3 is then heated and reduced to produce uranium dioxide (UF2). Next, UF2 is reacted with hydrogen fluoride (HF) to produce UF4. This sub-process is often referred to as hydrofluorination. Finally, if necessary for fuel fabrication or weapons development, fluorine is reacted with UF4 to produce UF6, the feed material for uranium enrichment. This step is often referred to as fluorination. For purposes of clarity, this piece will refer to each of these sub-processes by their associated end-product (i.e., UO2 production, UF4 production, etc.).

Figure 2: Basic diagram of the uranium conversion process

Diagram of the uranium conversion process

Most steps in the uranium conversion process can either be wet, meaning the uranium product exists in a liquid state, or dry, meaning the uranium product exists in a solid or gaseous state. However, these qualifiers can be overly simplistic when defining uranium conversion sub-processes. In a similar vein, there is not one standard operating procedure for uranium conversion. The remainder of this piece will detail common conversion practices which can be defined as: practices most frequently detailed in publicly available English language literature by conversion experts and practitioners. Because North Korea has not explicitly declared all its uranium conversion processes in publicly available information, and there is a scarcity of recent technical information and literature specific to North Korea, it is unknown if any of the proceeding process descriptions accurately encompass processes employed historically or currently by North Korea.

Figure 3: Locations & operational summary of North Korea’s historic uranium conversion program

Maxar satellite image of the uranium fuel fabrication plant on April 21, 2021.
Satellite image: ©2021 Maxar Technologies

Wet UO2 Production and North Korea Contextualization

According to researchers at Oak Ridge National Lab, wet UO2 production processes are commonly employed by countries with nuclear programs,[3] though there are multiple factors that can affect process specifics including the size of the facility and the amount and quality of the uranium concentrate input.[4] In a wet UO2 production process at smaller sized facilities, uranium concentrate is first dissolved using nitric acid in a stirred, heated, corrosion-resistant vessel, and yields uranyl nitrate (UO2(NO3)²).[5] Next, the solution is purified, usually in a two-stage liquid-liquid solvent extraction. The first stage extracts uranium from the aqueous solution into an organic phase, and often uses tributyl phosphate dissolved in kerosene or dodecane. The second stage of this extraction, commonly referred to as scrubbing, removes impurities from the uranium-bearing organic stream. Equipment used in a purification process can include mixer-settlers, various types of columns (such as pulse columns), and centrifugal contractors.[6]

The next step is concentration, or uranyl nitrate solution processing. This step varies across different facilities. For example, large facilities tend to employ thermal denitration in concentration processes, whereas smaller facilities tend to utilize a precipitation process.[7] In a precipitation process, the liquid uranyl nitrate solution is converted into either ammonium diurnate (ADU) or ammonium uranyl carbonate (AUC), and then into UO2. Specifically, from the uranyl nitrate solution, either ADU is precipitated using ammonium hydroxide, or AUC is precipitated using ammonia and carbon dioxide or ammonium bicarbonate. The ADU or AUC is then calcined (i.e., heated) and reduced. ADU can be converted in a rotary furnace or fixed-bed furnace, and AUC can be calcined in a fluidized bed or rotary/fixed-bed furnace.

A thermal denitration process yields UO3 without ADU or AUC, by means of calcining the uranium product after the initial extraction and purification.[8] The UO3 is then turned into UO2 through an oxide reduction process, which can occur by reacting UO3 with hydrogen gas.

Figure 4: Diagram of a uranium conversion process using wet UO2 production via ADU precipitation [9]

Argonne Uranium Conversion
Photo Credit: Argonne National Laboratory

As mentioned in Part One of this series, one of the earliest publicly available technical descriptions of North Korea’s uranium conversion process is provided by the dossier Solving the North Korea Nuclear Puzzle.[10] It should be noted that the conversion process description provided by the dossier is a wet UO2 production process. However, because North Korea has not officially confirmed its uranium conversion process, as previously discussed, it can only be assumed that available North Korea process information and context is a combination of the North Korean 1992 declaration to the IAEA, facts from the 1992 IAEA Yongbyon visit, and generic knowledge on conversion processes that is not necessarily specific to North Korea.

The dossier describes North Korea’s uranium conversion program as follows. In Building 1 (identified in Figure 5 below), yellowcake was “dissolved in hot nitric acid, purifying the nitrate solution by solvent-extraction with 30% volume of tributyl phosphate in kerosene, evaporating the product to precipitate hydrous uranyl nitrate, and then heated and dried to thermally produce UO2.”[11] The dossier states further reduction to UO2 (which would have also occurred in Building 1) is “normally performed in a two-stage fluidized bed at 600°C, with hydrogen decomposition of ammonia.” The wet UO2 production process described in the dossier includes precipitation via ADU. More recently, former IAEA officials have corroborated that North Korea’s UO2 production involved ADU.[12]

Figure 5: The process flow at North Korea’s Yongbyon Fuel Fabrication Plant [13]

Uranium conversion process flow at Yongbyon

As is described in Part Two of this series, Dr. Siegfried Hecker reported that in July 2007, five tons of UO3 were removed from Building 1.[14] While the provided description of an ADU precipitation process does not directly mention UO3, UO3 is an intermediary product in conversion and is generally more storage stable than  UO2.[15] Considering this, it is possible North Korea employs a wet ADU precipitation process to convert uranium concentrate into UO3 first, and then continues the process as needed by heating the stored UO3 and reducing it to UO2.

Wet and Dry UF4 Production and North Korea Contextualization

The step in the uranium conversion process that North Korea appeared to struggle with the most was UF4 production. North Korea likely intermittently operated: a wet UF4 production line from early 1990s until 2002, a small scale dry UF4 production line from end of 2002 until July 2007, and a full scale dry UF4 production line from early to mid-2009 until present.

While most available literature simply refers to UF4 production as hydrofluorination, researchers from the Nuclear and Energy Research Institute (IPEN) in Brazil have expertise in further distinguishing between wet and dry UF4 production processes.[16] This is notable as Brazil and North Korea potentially had similar uranium conversion facility sizes and outputs and might have also employed similar UF4 production processes. A paper describing the uranium cycle of IPEN states Brazil used a dry UF4 production method.[17] The paper further cites a description of UF4 production processes published by IPEN through the IAEA.[18][19] In this description, several methods for producing UF4 from UO2 that was precipitated via ADU are provided. These methods include one option for an aqueous production that relies on the precipitation of UF4 from sodium uranium solutions, and five different options for gaseous production of UF4, all of which involve reacting UO2 with fluorinated gas agents at high temperatures.

The described wet process involves reducing fluoride, chloride, or uranyl sulfate to a tetravalent state and precipitating UF4 by the addition of hydrofluoric acid. The described five dry UF4 production options include: 1) Fluorohydration[20] of UO2 2) reaction of UO3 with ammonia (NH?) and anhydrous hydrogen fluoride 3) reaction of uranium oxides with fluorinated hydrocarbons (e.g., Freon) at high temperatures 4) using metallic uranium or UH3 by fluorohydration at high temperatures and 5) reaction of UO2 with ammonium bifluoride. While it is out of the scope of this piece to go into the specifics of each of these different types of dry UF4 production, the IPEN descriptions convey that there are many possible pathways for producing UF4 using a dry method after UO2 has been produced by means of ADU precipitation. Given this, it is difficult to assess North Korea’s exact employed dry UF4 production processes from the available open-source information as will be further examined below.

More broadly, wet UF4 production methods are generally older than dry production methods and are conducted by precipitating uranium either by ion exchange or solvent extraction.[21] Solvent extraction is the more frequently discussed method in publicly available literature and essentially involves heating a solution of UO2 and hydrogen chloride (HCl), adding a reducing agent,[22] and then adding hydrofluoric acid as a precipitating agent.[23] Given North Korea likely used a wet UF4 process from around 1990 until 1992, it is possible it employed some variation of this solvent extraction process for UF4 production during that time frame.

On the other hand, most commonly described dry UF4 production processes tend to detail hydrofluorination, wherein UO2 powder is fed into a hot rotating kiln and reacted with gaseous hydrogen fluoride.[24] In its description of North Korea’s UF4 production, the Solving the North Korea Nuclear Puzzle dossier, published in 2000, states UF4 production “normally is performed by contacting UO2 in a two-stage fluidized bed at about 500°C with gaseous hydrofluoric acid.”[25] Notably, this process description is common to dry UF4 production. According to the provided timeline, at the time of the dossier’s publication, it is likely North Korea was employing a wet UF4 production line. Therefore, this assessment does not necessarily account for North Korea’s UF4 production process change following the end of the Agreed Framework, and the previously defined caveats regarding specific process assessments apply.

Anhydrous Hydrogen Fluoride Production and North Korea Contextualization

Hydrogen Fluoride (HF) is an important component of uranium conversion, and can be gaseous, anhydrous, or in the form of hydrofluoric acid. In its description of North Korea’s uranium conversion, the dossier Solving the North Korea Nuclear Puzzle, states anhydrous HF is used by North Korea and that “the source material [for HF] is probably an ore containing high concentrations of calcium fluoride.” Hecker’s description of the type of HF North Korea employs differs slightly, stating North Korea used hydrofluoric acid prior to the implementation of the Agreed Framework.[26] For clarity, it is worthwhile to briefly discuss this terminology.

A 2003 article published by Mark Hibbs reads “officials said anhydrous HF is produced in North Korea by treating calcium fluoride (CaF2) with a reagent such as sulfuric acid. The production of HF in the DPRK from CaF2 is monitored by Western intelligence agencies. The CaF2 is mined domestically by tapping considerable stratified Riphean-era deposits with ore grades above 30% CaF2.”[27] It is noteworthy that Hibbs uses the terms hydrofluoric acid and anhydrous HF interchangeably throughout the article. In a later article discussing Iran’s nuclear program, Hibbs describes “the term ‘anhydrous’ is just a fancy way of saying that hydrogen fluoride (HF), otherwise known as hydrofluoric acid, is more or less pure, having less than 400 ppm of water.”

[28] That is to say, anhydrous does not necessarily mean gaseous or solid. Based on this elaboration, it is likely that North Korea was using, and continues to use, anhydrous HF (AHF) for its UF4 production, though analysts sometimes use the terms AHF and hydrofluoric acid interchangeably.

As discussed above, calcium fluoride (CaF2) is likely required for North Korea’s AHF production, and therefore for its uranium conversion processes. CaF2 is also referred to as fluorite or fluorspar.[29] A declassified Central Intelligence Agency report from 1952 details North Korea’s historical operation of fluorspar mines, suggesting North Korea was able to produce large amounts of good quality CaF2 domestically at the time covered by the report.[30] For every one ton of uranium to be processed into UF4 and then UF6, one and two-third tons of CaF2 is required.[31] While it is out of the scope of this series, additional future research might be done to estimate how much CaF2 North Korea is presently producing, and moreover, how much of that CaF2 is directed to Yongbyon for AHF production. Future research could also examine other viable available sources for AHF production in North Korea beyond CaF2.

Conclusions and Series Summary

This three-part series aimed to provide an in-depth look at North Korea’s uranium conversion capabilities through historical and process-oriented lenses. It remains unknown where and how North Korea presently conducts all sub-processes of its uranium conversion program. However, based on the provided information, it is possible North Korea continues to employ wet UO2 production via ADU precipitation somewhere at the Yongbyon Fuel Fabrication Plant (FFP). It is also possible that North Korea has employed an anhydrous UF4 production line since 2009, though this line’s exact process specifications and location remain unknown. Similarly, specific process and location information on North Korea’s UF6 production line remains unknown.


[1] “Conversion and Deconversion,” World Nuclear Association, September, 2020, <https://www.world-nuclear.org/information-library/nuclear-fuel-cycle/conversion-enrichment-and-fabrication/conversion-and-deconversion.aspx>.

[2] “Nuclear Fuel Cycle Overview,” World Nuclear Association, May, 2020, <https://www.world-nuclear.org/information-library/nuclear-fuel-cycle/introduction/nuclear-fuel-cycle-overview.aspx>.

[3] According to the World Nuclear Association, only the United States employs a dry UO2 production process. For more, see: “Conversion and Deconversion,” World Nuclear Association.

[4] “Model of a Generic Natural Uranium Conversion Plant – Suggested Measures to Strengthen International Safeguards,” Oak Ridge National Laboratory and Brazilian Nuclear Energy Commission, November 2009, <https://info.ornl.gov/sites/publications/Files/Pub13143.pdf>.

[5] “Conversion and Deconversion,” World Nuclear Association.

[6] “Module 3.0: Uranium Conversion,” Nuclear Regulatory Commission, <https://www.nrc.gov/docs/ML1204/ML12045A005.pdf>.

[7] “Model of a Generic Natural Uranium Conversion Plant – Suggested Measures to Strengthen International Safeguards,” Oak Ridge National Laboratory and Brazilian Nuclear Energy Commission.

[8] For more on denitration, see:  Allen Bakel and George Vandergrift, “Equipment and Method Choices for Concentration and Denitration of the Uranium Product from UREX,” Argonne National Lab, August, 2013, <https://publications.anl.gov/anlpubs/2013/08/76448.pdf>.

[9] “Conversion of Yellow Cake to UF?,” Depleted UF6 Management Information Network, <https://web.evs.anl.gov/uranium/guide/prodhand/sld006.cfm>.

[10] David Albright and Kevin O’Neill, Solving the North Korean Nuclear Puzzle, (Washington, DC: Institute for Science and International Security, 2000).

[11] Ibid., p. 146.

[12] Jack Liu, Olli Heinonen, Peter Makowsky, and Frank Pabian, “North Korea’s Yongbyon Nuclear Center: Additional Activity at the Radiochemical Laboratory and Uranium Enrichment Plant,”

[13] This graphic is a reproduction of Figure 9 from Chaim Braun, Siegfried Hecker, Chris Lawrence, Panos Papadiamantis, “North Korea Nuclear Facilities After the Agreed Framework,” Center for International Security and Cooperation at Stanford University, May 27, 2016 <https://fsi-live.s3.us-west-1.amazonaws.com/s3fs-public/khucisacfinalreport_compressed.pdf>.

[14] Siegfried Hecker, “Report of Visit to the Democratic People’s Republic of North Korea (DPRK),” Center for International Security and Cooperation at Stanford University, March 14, 2008, <https://fsi-live.s3.us-west-1.amazonaws.com/s3fs-public/HeckerDPRKreport.pdf>.

[15] S.M. Thein and P.J. Bereolos, “Thermal Stabilization of UO2, UO2, and U3O8,” Oak Ridge National Laboratory, July 2000, <https://info.ornl.gov/sites/publications/Files/Pub57536.pdf>.

[16] Ivan Santos, Alcidio Abrao, Fatima M.S. Carvalho, and Jamil M.S. Ayoub, “Decommissioning of an Uranium Hexafluoride Pilot Plant,” International Nuclear Atlantic Conference, September, 2009, <https://inis.iaea.org/collection/NCLCollectionStore/_Public/41/115/41115745.pdf?r=1&r=1>.

[17] Alcidio Abrao, “O Ciclo Do Uranio No Ipen,” Instituto de Pesquisas Energeticas e Nucleares, September, 1994, <https://www.ipen.br/biblioteca/ipen/IPEN_PUB_398.pdf>.

[18] Adelino Filho and Alcidio Abrao, “Tecnologia para a preparacao de tetrafluoreto de uraniuo por fluoridetacao de UO2 obtido de diuranto de amonio,” Instituto de Energia Atomica, January, 1975, <https://inis.iaea.org/collection/NCLCollectionStore/_Public/08/286/8286362.pdf?r=1>.

[19] Primary sources from IPEN are written in Portuguese and were translated directly by the author.

[20] This is a translation from Portuguese and it can be assumed this term is synonymous with “hydrofluorination.”

[21] H.G Petrow, et al., “Preparation of Dense, Metal-Grade UF4 from Ores and Concentrates,” US Atomic Energy Commission, May 15, 1958, <https://www.osti.gov/servlets/purl/4314844>.

[22] Reducing agents can include chlorides such as SnCl2 and FeCl2. For more, see: E.U.C Frajndlich, et al., “Alternative Route for UF6 Conversion Towards UF4 to Produce Metallic Uranium,” Instituto de Pesquisas Energeticas e Nucleares, October, 1998, <https://inis.iaea.org/collection/NCLCollectionStore/_Public/35/040/35040239.pdf>.

[23] J.V Opie, “The Preparation of Pure Uranium Tetrafluoride by a Wet Process,” US Atomic Energy Commission, April, 1946, <https://www.osti.gov/biblio/4335919>.

[24] “Hex Business,” Westinghouse Nuclear, <https://www.westinghousenuclear.com/Portals/2/Documents/Brochure%20-%20Hex%20Businessl.pdf>.

[25] Albright and O’Neill, “Solving the North Korean Nuclear Puzzle,” p. 146.

[26] Chaim Braun et al., “North Korea Nuclear Facilities After the Agreed Framework.”

[27] Mark Hibbs, “DPRK poised to embark upon UF6 production at Yongbyon,” Nuclear Fuel, Vol. 28, No. 18, September 1, 2003, p. 3.

[28] Mark Hibbs, “Iran’s Quest for the F6 in Its UF6,” Arms Control Wonk, January 2, 2012, <https://www.armscontrolwonk.com/archive/1100585/irans-quest-for-the-f6-in-its-uf6/>.

[29] Hobart King, “Fluorite,” Geology.com, <https://geology.com/minerals/fluorite.shtml>.

[30] “The Fluorspar Industry in the Soviet Bloc,” Central Intelligence Agency, May 20, 192, <https://www.cia.gov/readingroom/document/cia-rdp79r01141a000100040002-3>.

[31] “Canadian Minerals Yearbook 1970,” Department of Energy, Mines and Resources Ottawa, 1972, <https://ftp.geogratis.gc.ca/pub/nrcan_rncan/publications/STPublications_PublicationsST/247/247713/gid_247713.pdf>.

This work is part of an ongoing joint project between the Verification Research, Training and Information Centre (VERTIC), the James Martin Center for Nonproliferation Studies (CNS) at the Middlebury Institute of International Affairs (MIIS), and the Royal United Services Institute (RUSI). The project seeks to update and systematize analysis of North Korea’s nuclear complex, using remote sensing data and industry-standard nuclear fuel cycle modelling to generate independent models of plausible scenarios for that complex, in the past, present and future, and to use those results to assess priorities for control and verification. The project is generously funded by Global Affairs Canada.


*Original article online at https://www.armscontrolwonk.com/archive/1212293/north-koreas-uranium-conversion-the-history-and-process-part-3/