Jeffrey Lewis has stirred up the question of whether apparatus photographed in Burma/Myanmar (the original article is now gone; this link is from the Wayback Machine, 2/2/15) is useful for uranium and rare earth chemistry. I’ve provided a quick overview based on my experience. But there’s another source of information on uses of equipment for both uranium and rare earth chemistry.

The Silmet plant in Sillamäe, Estonia, produced uranium compounds through the 1980s and rare earth compounds since the 1970s. The full story is still not available, although much of it is covered in the first two parts of Historical Survey of Nuclear Non-Proliferation in Estonia, 1946-1995, by Ello Maremäe and Hain Tankler. I have also consulted several chapters in Turning a Problem into a Resource: Remediation and Waste Management at the Sillamäe Site, Estonia. I am also drawing early Silmet brochures in my possession, my experience as a research chemist, and my observations at Sillamäe.

The Maremäe and Tankler report (which I will abbreviate M&T) is the more recent, published in 2003, and is a voluntary supplement to Estonia’s  State Declaration under its  Additional Protocol to the Nuclear Nonproliferation Treaty. Along with other sources, Maremäe and Tankler used a roomful of reports left at the Silmet plant by the Soviets, many of them with classification markings. More documentation is believed to exist at the Institute of Chemical Technology in Moscow and at other archives, but the Russians have not been willing to make that documentation available.

The Silmet plant is now owned by Molycorp. I have used material from their website on the current operation of the plant. The name “Silmet” came into use in 1997, when the plant was privatized.

Silmet has processed both uranium and rare-earth compounds, so it is reasonable to look at Silmet to try to understand whether the apparatus photographed in Burma indicates a nuclear weapons program. I’ll consider the uranium and rare-earth processing Silmet has done, along with its tantalum and niobium production, and then draw some conclusions about the equipment photographed in Burma.

Uranium Processing at Sillamäe

A plant was built at Sillamäe in 1946 to supply uranium to the Soviet atomic bomb project. Its source material changed over the years. Uranium processing ended as the Soviet Union destabilized in the late 1980s. Rumors have long circulated that uranium enrichment took place at the plant, but neither the documents nor an examination of the facilities show any evidence of that. Three very large reinforced-concrete buildings, possibly suitable for centrifuge facilities, were constructed in the 1980s but never completed; they were rubblized for use in the remediation of the tailings pond.

As World War II ended, the Soviet Union was ramping up its nuclear program. It faced a severe shortage of uranium deposits. Northeastern Estonia had long mined its very rich oil shale for fuel, and part of that formation was a black shale containing uranium. The Sillamäe plant was initially designed to recover uranium from that black shale. A flow chart for the process is given on page 21 of M&T. The product was yellowcake (U3O8). We can deduce the types of equipment from that chart and the text:

Ore grinders
Gerreshof kilns
Vessels and piping for handling ore and processing liquids
Drum filters
Plate-and-frame filters
Vessels and piping for precipitation
Electric drying furnaces
A chemical laboratory was also necessary for analyzing raw materials, process solutions, and product.

Black shale is the most difficult uranium ore from which to produce uranium, and in 1952 black shale was discontinued as the ore, although research on it continued into the sixties (p. 24, M&T). Starting in 1952, richer ores were brought in from other parts of the Soviet Union and its satellites (p. 19, M&T). The use of these ores would have eliminated the Gerreshof kilns and some of the ore-treating equipment.

Increasing purity of the product yellowcake was required through the 1950s and 1960s. First, multiple precipitation steps (flow chart on p. 25) were introduced and later ion exchange. The timing coincides with the introduction of centrifuges into the Soviet enrichment complex; impurities in UF6 are more harmful to centrifuges than to the gaseous diffusion process that the centrifuges replaced. The best place in the process to remove those impurities is during production of U3O8.

In the 1970s, the raw material was changed from ore to concentrate, which would have lessened the number of or eliminated the ore grinders. Improvements continued on the ion-exchange process, but this would not have changed the equipment significantly. In 1979, automated monitoring and control was introduced; again, probably little change in equipment (p. 28, M&T).

In 1983, additional automatic control was put into operation. In 1985, pulsating ion-exchange columns replaced the older columns.

An additional production line started up in 1982 to reclaim rejected fuel-element UO2. The starting material was technical grade U3O8 enriched to 2.0 – 3.6% in U-235. It came from the Machine Building Plant in Elektrostal, Ulbinsky Metallurgical Plant in Ust-Kamenogorsk, and the Novosibirsk Chemical Concentrate Plant, where rejected UO2 pellets were oxidized to U3O8. The process flow diagram is on page 32 of M&T. Most of the steps are similar to the other unit processes and would require similar equipment. Solvent extraction and a furnace with hydrogen flow would have been added.

“Granulated microfuel” with uranium enrichment of 21 – 90% was also produced at Sillamäe (pp. 31-33, M&T), requiring ball and rolling mills and sintering furnaces. “Spherical fuel elements” were also produced from the microfuel, probably requiring additional mixing equipment and furnaces. These operations were at a pilot scale and were discontinued in 1987.

The Ministry of Medium Machine Building sent instructions to Sillamäe in June 1989 to end all uranium processing and send uranium materials to other plants in the Soviet nuclear complex.

Rare Earth Processing at Sillamäe

Not as much information is available on rare earth processing. Production of rare earth compounds began in the early 1970s, and they quickly became the major project. Production of niobium metal began in the mid-seventies (p. 10, M&T). An undated Silmet brochure, probably from 1992, lists as products

Lanthanum oxide concentrate
Lanthananum and cerium oxide concentrate
Cerium dioxide concentrate
Didymium (a mixture of praseodymium and neodymium) carbonate concentrate
Concentrate of oxides of middle group of rare-earth elements
Oxides of rare-earth elements of the cerium group
Optical polishing powder
Metallic niobium ingots and powders
Nickel-based rare-earth master alloys (for steel and other alloys)
Titanium oxide
Titanium and calcium filler
Nitrogen-containing fertilizers
Polymeric filtering materials

A 1999 brochure lists

Tantalum hydroxide
Niobium pentoxide
Metallic niobium ingots and powders
Cerium subgroup rare earth oxides (carbonates)
Cerium subgroup rare earth fluorides
Fluorine-containing polishing powders (optical and technical)
Cerium subgroup rare earth nitrate solution
Cerium oxide (carbonate) concentrate
Lanthanum oxide (carbonate) concentrate
Neodymium oxide (carbonate) concentrate
Neodymium fluoride concentrate
Middle group rare earths oxide (carbonate) concentrate
Samarium-gadolinium oxide (carbonate) concentrate
Nickel-based rare earth master alloys (for steelmaking)

In Turning a Problem into a Resource: Remediation and Waste Management at the Sillamäe Site, Estonia, V. Petrenko and A. Siinmaa (pp 47-56) describe the process as “primarily hydrometallurgical, including dissolution, precipitation, filtration, and liquid extraction processing. The final stages of production include drying and calcination in electric drum furnaces.” They also provide a flow sheet.

The process for production of niobium and tantalum from columbite concentrate includes grinding, dissolution, filtration and washing of solids, extraction of tantalum and niobium by TBP in mixer-settlers, washing, and re-extraction of the two metals as hydroxide. The niobium hydroxide is calcined to oxide.

In the same book, V. D. Kosynkin and V. J. Nikonov of the All-Russia Research Institute of Chemical Technology proposed recovery of scandium from the loparite ore raw material as well (pp. 57-61), but this suggestion has not been followed up on.

Molycorp Silmet now lists a similar suite of products to the 1997 list. The most significant additions for our purposes are niobium metal and niobium and tantalum hydrides.

The equipment required for these processes overlaps with that required for the uranium processing at the plant. Production of the metals requires a reduction step, but probably not the bomb-type vessel used for uranium, because niobium and tantalum metals are stable in air, whereas uranium metal is not. Uranium metal was never produced at Sillamäe. A specialized furnace may be necessary for production of the hydrides.

Looking at Burma Again

Molycorp Silmet has provided a very nice video that shows a lot of equipment. It’s about four and a half minutes long, and apparently non-embeddable. Watch the whole thing to get a sense of the plant before you read further. There’s a little about the town, too, and I’ll have more about that if you read all the way to the end! I’ll say more about that big grassy area too.

Here’s one example from the Burmese photos (link from Wayback Machine).

Bomb vessels Kelley

Or from the textbook Jeffrey quotes.


There’s no comparison. Silmet is a full-scale metals plant. It is possible that Silmet is larger than a weapons uranium plant might be; check out the quantities of product at 2:47 in the video. But the equipment in the Burmese photos and the Indian text is much smaller. Some of it is similar to equipment that can be bought commercially. The famous (to chemists anyway) Parr bombs look a lot like the vaunted Burmese bomb reactor.

Could the Burmese equipment be used for a developmental program on producing uranium metal? Yes. Could it be used for many other things? Yes.

The Burmese equipment, and the equipment in the Indian text, is for research, not production. At most, some of it might be considered pilot plant scale.

When Sillamäe produced uranium, it was yellowcake, oxide rather than metal. But it has produced niobium and tantalum metals (video, 0:55 and 2:03). Because of chemical differences between uranium and the other two, most notably that uranium catches fire (reoxidizes) when it’s heated in air, the process and equipment for reduction of uranium to metal are different from those for Silmet’s metals.

However, most of the equipment for refining ores or concentrates is dual-use; it can be used for uranium or commercial metals. It’s not clear how much equipment at Sillamäe actually was repurposed from uranium to rare earths; rather, the capabilities of the workers for similar unit processes allowed the changeovers, probably with both addition and repurposing of equipment.

Let us look specifically at some of the claims about Burma.  The possible cold trap or UF6 collector looks poorly made, certainly not a product of the numerically controlled machine tools reported to have been acquired, or not utilizing their capabilities fully. The design seems strange too; the center well, presumably the liquid nitrogen reservoir, is too big, and it’s not clear why it is square. For a collector, you want lots of cold surface area, and the spiral objects might provide that. But the large square well requires a significant amount of liquid nitrogen where it isn’t really helping to cool surfaces.

Burma cold trap

Conclusion: This apparatus could be a UF6 collector, but its capacity is small and it appears poorly designed and made. It is not recognizably unique for uranium work.

Although there has been some highlighting of the use of Inconel, Kelley and Fowle recognize that it is not diagnostic for nuclear work. Inconel is used where corrosion resistance is needed. Rare earth processing at Silmet uses chloride and fluoride in its processes, both of which are likely to require the use of Inconel. Inconel has many uses.

Conclusion: The use of Inconel is in no way diagnostic of uranium chemistry.

What is not shown in the Kelley and Fowle report is anything resembling ion exchange columns or solvent extraction. At Sillamäe, both have been used for uranium purification. Solvent extraction is essential to rare earth purification because many of the rare earth elements are similar enough to each other in their chemistry that ion exchange separates them poorly. Both of these are exacting technologies; not as exacting as centrifuge operation, but more so than the simple fluid handling systems and tube furnaces shown in the Burmese photographs. Solvent extractors are shown at 2:25 in the Silmet video.

350Molycorp Silmet video solvent ext

Both separation technologies require significant knowledge of chemistry. An enormous number of ion-exchange resins and other media are available. The proper ones must be chosen according to the anions present and the pH of the solution, along with what cations are to be separated. The solvent exchange processes used at Sillamäe have used TBP – tributyl phosphate – as the complexing medium. It works well for both uranium and rare earth solutions. But temperature, flow rate, and solution properties (of two solutions!) must all be controlled. Hence the chemistry laboratory, which has other functions as well.

Conclusion: Without ion exchange or solvent extraction, there is no way to prepare a pure product. Nothing relating to these technologies appears in the Burmese photographs. Although these technologies could be used for rare earths as well as uranium, either program is incomplete without them. There is also no indication of a chemical laboratory in the Burmese photographs. Little development can be done without one.

Overall conclusion: The Burmese photographs show nothing that unambiguously relates to uranium processing. The equipment shown is of a laboratory or pilot plant scale, not a production scale. An important capability that is not shown in the photographs is purification of whatever is being processed. Ion exchange and solvent extraction are commonly used for this purpose.

Appendix I: The Molycorp Silmet Video

This is a nicely done video, with some cool photos of equipment. An annotation identifying equipment and parts of the plant and town follows. I think the narrator is misprouncing “Sillamäe”. I have a hard time with it, too. Some parts of Estonian are hard for English speakers.

Beginning: Mere Puiestee (Ocean Boulevard), one of my favorite places to walk.
0:23  The town from the remediated tailings pond
0:32  The plant from the remediated tailings pond; tanks to the right are part of the port.
0:39  Main processing building
0:55  Metals production (probably tantalum or niobium)
1:14  Another view of the plant
1:41  The port is in the background. The large grassy area is the remediated tailings pond.
2:03  Metals production
2:13  A furnace
2:25  Bank of solvent extractors
2:47  Clearly the production is larger than would be needed for a uranium plant for nuclear weapons.
3:13  The port from the remediated tailings pond
4:10   Sillamäe town hall and park

Appendix II: Myanmar or Burma?

In my earlier post, I referred to the country as Myanmar rather than Burma. I was under the impression that that was the current US State Department usage. This says different, so I’m using Burma.

Appendix III: What Makes You Think You Know So Much About Sillamäe, Cheryl?

I have been visiting Sillamäe regularly for more than a decade. On my last visit, in the summer of 2011, I posted some photos and explanations. If there are differences between those posts and this one, this one is more likely to be correct.
Sillamäe Today
Molycorp Comes to Sillamäe
More On Molycorp and Sillamäe

You may also consult Turning a Problem into a Resource: Remediation and Waste Management at the Sillamäe Site, Estonia, which is the report of a NATO Advanced Research Conference I co-chaired with Tõnis Kaasik in 1998. It contains additional background on Sillamäe’s problems, many of which are being fixed, in pretty much the ways we recommended in that conference. Molycorp Silmet’s website lists some of them as “Acheivements,” [sic] but they were done by others. The biggest achievement was the remediation of the tailings pond, a kilometer long and half a kilometer wide, which included stabilization of the dam, isolation from groundwater, and an engineered cover to prevent rainwater from leaching through the tailings. That was done by Ökosil Ltd. The remediation allowed a port to be developed, which continues to expand.

On this, I’m not objective. I’m proud of my Estonian friends and what they’ve done, and my small contributions. I think that my lack of objectivity on this subject didn’t affect my analysis of equipment, though.

facebooktwittergoogle_plusredditpinterestlinkedinmailby feather

Leave a Reply