A Satellite Imagery Review of the Pyongsan Uranium Mill

The Pyongsan Uranium Concentration Plant and its associated mine are North Korea’s only publicly acknowledged source of uranium yellowcake. [1] From this site, uranium yellowcake is reportedly sent either to the Yongbyon nuclear complex for conversion into fuel elements for North Korea’s sole operational gas-graphite reactor or to a conversion facility in preparation for uranium enrichment. Although North Korea declared the Pyongsan Uranium Concentration Plant and its associated mine to the International Atomic Energy Agency (IAEA), and an IAEA team visited the facility in 1992, open-source researchers have struggled to understand the process flow at the site. 

During the 1992 IAEA visit, the North Koreans claimed that the Pyongsan mine supplies anthracite (also known as stone coal) from which they extract uranium, along with other byproducts including nickel and vanadium. Although the recovery of uranium from coal is unusual, it is not unheard of. The United States explored using another type of coal (lignite) as a uranium source in the early years of the US nuclear program before selecting more accessible and economically viable sources of uranium.[2] 

A review of satellite imagery of the Pyongsan Uranium Concentration Plant indicates that the site is active, and that North Koreans modified the extraction process at the plant some time before 2006, although the exact date is unknown due to limited commercial satellite imagery available from this time frame. The relatively unusual process for extracting uranium from anthracite, as well as changes to that process since the IAEA visit in 1992, complicate analysis of the facility.

To better understand how the Pyongsan mill operates, a brief overview of typical mining and milling processes will be useful for understanding North Korean practices. 

While each uranium mill is specifically designed to process the exact source ore being milled, the general overview remains fairly consistent. After mining, ores containing uranium pass through a circuit of processes that includes crushing and grinding, followed by a series of chemical treatments and filtration to remove undesirable materials and extract the uranium-laden content. Depending on the chemical composition of the source ore, other processes can be included to extract other materials of value that may be found in the ore. At the end of the milling process, after the uranium has been extracted, it is dried in the form of uranium oxide powder, known as yellowcake, which is then packaged and shipped to a secondary site for further processing. [3]

Without additional information, the “customized” nature of a uranium mill can make it difficult to understand the exact process flow at sites like Pyongsan. This analysis will attempt to identify specific processes at the Pyongsan mill, changes that have occurred over time, and any unique features that may have been previously overlooked.

DigitalGlobe/MAXAR – July 27, 2017

Pyongsan Uranium Mill

The Pyongsan uranium mill (also known as the Namchon Chemical Complex) is North Korea’s only declared location for uranium milling activities, excluding the reportedly decommissioned pilot mill at Pakchon. Most of what we know about the mill and its activities come from two sources: a North Korean defector named Kim Tae Ho, who worked at the mill’s wastewater treatment plant, and a video from an IAEA visit in 1992, led by IAEA director Hans Blix.

Kim Tae Ho originally worked at the Pakchon pilot mill during the mid-1980s. At some time during the late 1980s, Kim was transferred to the newly constructed mill located at Pyongsan, where he was put in charge of the wastewater treatment plant, which by his account was nearing operations in late 1989 or early 1990. In a book published in Japan in 2003, Kim wrote that the complex was “made up of up to 12 sections, such as uranium ore crushing workshop, precipitation workshop, extraction workshop, vanadium workshop, waste water treatment workshop, heat control workshop, exclusive line workshop, official business workshop, life’s essential goods workshop, and so on.” [4]

The wastewater workshop, where Kim worked, was reportedly where waste water was treated and further materials were extracted, including molybdenum, nickel, and radium.

Kim identified two different ore types used for uranium extraction, Ore no. 2 and Ore no. 3. Ore no. 2 was described as being a sort of limestone that was reportedly exhausted by 1987, while Ore no. 3 was described as being mostly composed of a “low-heat coal” with a concentration of 0.8% uranium, 1.4% vanadium, and the additional molybdenum and radium previously mentioned. [5]

The information in the 1992 IAEA visit video generally agrees with Kim’s statements, with one minor point of difference. According to the video of the visit, the source ore being used was described as anthracite, with the additional extraction of nickel and vanadium. [6] On the extraction of nickel and vanadium there is agreement. According to Kim, however, the ore being milled at Pyongsan, Ore no. 3, was a low-heat coal, whereas anthracite has the highest energy density of any grade of coal. The reason for this difference in reporting is not clear.

The image above is of the colocated mine at Pyongsan – Image source: DigitalGlobe/MAXAR July 27th, 2017

Pyongsan Coal Mine

The first step in the Pyongsan process flow starts at the colocated coal mine to the north of the mill. From this mine, coal is moved to a primary coal processing plant (CPP). At this location, the mined coal is crushed to a size that allows it to be transported by piping to a secondary CPP inside the main milling complex. Over time, coal dust may be seen accumulating on the roof of the secondary CPP. The roof of this plant was cleaned between October 2013 and September 2014. Since then, the black coal dust has reappeared, most notably between images taken in 2015 and 2016, indicating continued operation.

Image source: DigitalGlobe/MAXAR July 27th, 2017

Rotary Kiln

Directly south of the secondary CPP is a large rotary kiln. A rotary kiln is an industrial structure that processes ores by subjecting them to high temperatures (roasting) to induce a chemical change in the input material (pyroprocessing). In the United States, the roasting of uraniferous lignite coal was necessary to remove excess carbon and moisture, whose presence would cause lower levels of uranium extraction and complications in the subsequent leaching process.[7] Roasting in the kiln converts coal into two different types of ash, bottom ash[8] and fly ash[9], which are then sent to the leaching phase of the uranium milling process. The use of coal ash as a source of uranium has been researched as a potential new source of uranium and as a way to remove radioactive contaminants from the coal ash produced at coal-fired power plants.[10]

Diagram of the use of a rotary kiln on uraniferous lignites for the extraction of uranium laden ash
Source: “Significance of Mineralogy in the Development of Flowsheets for Processing Uranium Ores”, Technical Reports Series no. 196, IAEA, 1980, www.iaea.org

Image source: DigitalGlobe/MAXAR July 27th, 2017

Until early 2006, the output end of the rotary kiln was connected to a conveyor belt that would carry the ash produced there to a now-defunct storage site in the northeastern corner of the complex. In the earliest imagery available, the input end of the kiln had no visible connection to the on site coal processing plant. There are at least two large entrances where a vehicle could bring material in for roasting making its servicing unique when compared to the rest of the site. The loading portion of the kiln also remained without a roof from earliest images available in 2003 until sometime between August 7th-10th of 2018. Indications of this building’s operation would notionally be seen in exhaust coming from the smoke stack to the east of the kiln but in the absence of this signature the structure may have been dormant during or before the removal of the conveyor belt in 2006.

The presence of a rotary kiln at the Pyongsan mill raises questions about the source material being mined for uranium extraction. In the previously discussed US example, the roasting of uraniferous lignite coal was necessary to improve the uranium extraction process. Lignites have a high moisture content along with other undesirable materials that would require roasting prior to continued milling. Anthracite has very low moisture content, but another possible benefit of roasting is elimination of carbon to improve solid-liquid separation. The kiln may have been built to facilitate the leaching process through pyroprocessing.[11] The changes in its connections to the complex at large may indicate either that the source material being milled has changed, or that the processes used to prepare ore for milling have advanced to a point where preroasting is no longer required. 

Probable Acid Leaching Tankhouse

Directly to the west of the secondary CPP is a large structure connected to it by piping. This building is also connected via piping to a second large structure described below, and possibly to the thermal plant located in the southern part of the complex. Its entire roof was replaced in 2014, with more recent roof repairs in certain locations as of April 2020. This section of the mill has also seen the addition of eight new ponds or cells to the north of the building, which appeared between 2003 and 2006, overlapping with the removal of the rotary kiln’s conveyor belt.

The purpose of the new cells remains unknown but could possibly be related to the disconnection of the rotary kiln. The cells are constantly in a state of flux in regards to the level of liquid each cell contains. Different colors can be observed at the bottom of the dried cells. These changes could serve to indicate whether or not the mill is operating, more precisely than the expansion of the tailings pond south of the mill, which shows significant changes only over a period of months.

GoogleEarth – MAXAR – September 28, 2019

The specific areas where roof damage has occurred help to identify the purpose of this building, which appears to be an acid leaching facility

Much of what we know about the milling process relies on Kim Tae Ho’s 2003 book, in which he refers to the facility as the Namchon Chemical Complex.[12] Kim states that the mill used an acid leaching process for ore leaching. The use of acid leaching at this site is also mentioned in a 1993 report produced by the South Korea’s Korean Atomic Energy Research Institute (KAERI), which appears to rely on IAEA information.[13] A pair of research papers by a single North Korean author, dated 1989, also specifically describe processes for the handling of “sulfuric acid leaching liquors of uranium ore.” [14] 

Buildings that conduct open-vat acid leaching require appropriate ventilation to remove dangerous and corrosive acid vapors produced during the leaching process. While there are many different ways to ventilate a building that serves this purpose, the one used at this building in the Pyongsan complex is a natural draft design.

Source: J.A. Murray, M.R. Nees and P.W. Krag, “Acid mist containment in electrowinning using Bechtel’s electrode cap,” (1996), https://www.911metallurgist.com/electrowinning-tankhouses-air-quality/. 

The type of rust and damage that can be expected at a structure with this type of roofing ventilation can be seen in the image below at a copper electro-winning plant.

Gunnison Copper Extraction Plant, United States. Source GoogleEarth

As discussed in a previous report on this building, the roof over this structure has repeatedly experienced localized damage. Repairs have taken place periodically, including as recently as this year.

The continued appearance of localized roof rust and repairs suggests that acid leaching is conducted in this building, possibly along with thickening before solvent extraction. Being directly adjacent to and connected via pipes to the secondary coal processing plant also make it a strong candidate for this portion of the milling process.

Probable extraction facility 

Directly to the south of the probable acid leaching tankhouse and connected to it by piping is another building with the same type of natural draft ventilation design. It does not appear to suffer from the same type of damage that repeatedly appears on the roof of the probable acid leaching tankhouse. As it is the last of the large buildings in the circuit, it is likely that this building contains the last major step in the process: extracting yellow cake from the feed fluids. Piping from this structure leads south to the suspected waste-water treatment plant where additional materials are supposedly extracted, prior to its final destination in the tailings reservoir.

Southern half of the complex

The southern half of the facility includes two railheads, a thermal energy plant, a wastewater plant, and a large warehouse. The first railhead, located at the center of the milling complex, probably facilitates the removal of milled materials (including yellowcake) from the Pyongsan mill. A variety of rail cars can usually be observed here. The second railhead, located at the south end of the complex, is used for supplying coal to the thermal plant that powers milling operations. Hopper cars associated with moving coal are usually seen at this second railhead.

Kim Tae Ho described working at the wastewater treatment plant, which extracts excess materials, including molybdenum, nickel, and radium. With the main loop of piping at Pyongsan connecting to one building with a CCD, in turn looping south and out over the adjacent river to the tailings pond, the suspected location for that portion of the milling process likely occurs in the building labeled in the image above.

Tailings pond

After passing through the wastewater treatment plant, the mill’s wastewater is deposited in a tailings pond located across the river from the complex. The tailings pond is a valley that was dammed off for use as a storage location for the mill’s waste tailings. The tailings pond has slowly grown since operation started sometime around 1989-90. This growth has noticeably continued, with the ponds tailings pile expanding between 2019 and 2020.


While this report will be updated as new observations are made, the uranium mill at Pyongsan continues to be operational, indicating that North Korea continues to produce yellowcake. As satellite imagery analysis demonstrates, the facility remains in good repair with the tailings pile continuing to grow on the southern end of the complex. While the exact milling process can only be speculated on with the information available in the open-source, North Korea appears to have upgraded some portion of the original milling process at the facility some time before 2006.

As the site continues to operate, continued monitoring of Pyongsan will remain vital to understanding North Korea’s overall nuclear capabilities.

[1] North Korea also declared a second uranium mill, near Pakchon, in 1992 – but this mill was apparently no longer active by this time.

[2] Edward C. Murphy, “Uranium in North Dakota”, Geological Investigations No. 184, North Dakota Geological Survey, 2015, dmr.nd.gov / George W. Moore, Robert E. Melin, and Roy C. Kepferle, “Uranium-Bearing Lignite in Southwestern North Dakota”, Trace Elements Investigations Report No. 463, United States Department of the Interior Geological Survey, June 1954, pubs.usgs.gov / Holger Albrethsen Jr., Frank E. McGinley, “Summary History of Domestic Uranium Procurement Under U.S. Atomic Energy Commision Contracts – Final Report”, A Facsimile Report, United States Department of Energy, September 1982, doe.gov

[3] “Nuclear Fuel Cycle Information system: A Directory of Nuclear Fuel Cycle Facilities 2009 edition”, IAEA-TECDOC-1613, pp. 9-10, International Atomic Energy Agency, April 2009, iaea.org

[4] Kim Tae Ho, Watashi ga mita Kitachōsen kaku kōjō no shinjitsu [The truth about the North Korean nuclear facility I saw], translated by Kim Chan (Tokuma Shoten, 2003), pp. 130-135.

[5] Kim Tae Ho, Watashi ga mita Kitachōsen kaku kōjō no shinjitsu, p. 9-21.

[6] Visit by IAEA Director General Hans Blix to DPRK, May 11-16 1992, video at https://www.youtube.com/watch?v=qfr0PEf60xE. See 10:10 through 10:45.

[7]  “Environmental Assessment of Remedial Action at the Uraniferous Lignite Processing Sites at Belfield and Bowman, North Dakota” – DOE 1989 – https://www.osti.gov/servlets/purl/6302456 – Edward C. Murphay, Uranium in North Dakota, Geological Investigations Mo. 184, North Dakota Geological Survey 2015, dmr.nd.gov

[8] Bottom ash is the ash residue that remains on the bottom of an object after burning

[9] Fly ash are the particles the that result from burning a material that do not settle to the bottom but instead flow upwards with the smoke of a burnt material, often caught with filters.

[10] “Rising from the ashes”, The Economist, April 8th, 2010, economist.com / “Sparton produces first yellowcake from Chinese coal ash” World Nuclear News, October 16, 2007,world-nuclear-news.org

[11]  “Uranium extraction Technology”, Technical Reports Series No. 359, IAEA, 1993, www-pub.iaea.org

[12] ] Kim Tae Ho, Watashi ga mita Kitachōsen kaku kōjō no shinjitsu [The truth about the North Korean nuclear facility I saw], translated by Kim Chan (Tokuma Shoten, 2003)

[13] “Study on the Status of Nuclear Development and Utilization in North Korea”, KAERI, 1993, iaea.org

[14]  Han Hi Jong, “Investigation of extractive separation of uranium and vanadium from sulfuric acid leaching liquors of uranium ore(1): Extractive property of HDEHP,” KAERC 502/89, Research Center for Atomic Energy, The Ministry of Atomic Energy, Pyongyang, 1989; Han Hi Jong, “Investigation of extractive separation of uranium and vanadium from sulfuric acid leaching liquors of uranium ore(2): Reduction of Fe(III) from sulfuric acid,” KAERC 503/89, Research Center for Atomic Energy, The Ministry of Atomic Energy, Pyongyang, 1989. Both papers appear in the holdings of the IAEA library.

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.


  1. Cheryl Rofer (History)

    Great work, Dave!

    This answers my question about why the ore is so black. I was thinking volcanic rock, but, of course, it’s impossible to tell from overhead photos in visual wavelengths.

    As you note, coal is often a source of uranium. Water containing uranium in solution (6+ oxidation state) can be reduced by organic matter to the insoluble 4+ oxidation state and thus concentrated in underground flows. I’ve heard that partially petrified logs in the Grants (NM) uranium mining district had to be mixed with lower-grade ore in the mill to keep the radiation levels down.

    So it’s not surprising that an anthracite deposit might have a significant uranium content. I agree that a roasting step to oxidize carbon out before extracting the ash would be wise.

    The modifications could be to accommodate ore from another site. The mill in Sillamäe, Estonia, was originally built to handle a local ore source, but eventually received ore and concentrate from as far away as central Asia. It’s expensive to ship all that way, but North Korea wouldn’t be dealing with those distances, and transportation may be cheaper than building another mill. Or there may be other reasons. Autocratic regimes don’t necessarily worry about economics.