Jeffrey LewisIAEA Reports on Iran, Syria

Implementation of the NPT Safeguards Agreement and relevant provisions of Security Council resolutions 1737 (2006), 1747 (2007),
1803 (2008) and 1835 (2008) in the Islamic Republic of Iran
, GOV/2008/59 (November 19, 2008).

Implementation of the NPT Safeguards Agreement in the Syrian Arab Republic, GOV/2008/60 (November 19, 2008).

I think we beat our friends at ISIS, for a change. Enjoy!

(Or they put it up on their cool new Iran blog)


  1. Frank (History)

    May: 3970kg UF6
    Nov: 9750kg UF6

    need ~60,000 UF6 for a 15kg HEU device

    therefore will take another 50 months at this rate unless they add more centrifuged?

  2. Andreas Persbo

    I believe you did. I had ISIS in my inbox at 10:50 and you posted at 10:28, some 18 minutes (I have reason to suspect) after you received it. In other words, you’re a pretty fast textiler.

  3. Gump (History)

    I thought it was just too much 包子 with lunch making me sick until I realized it was just excess media spin .

  4. Allen Thomson (History)

    The Syria report said IAEA found “natural uranium particles.” It would be interesting and perhaps significant to know if those were metallic uranium or some compound like an oxide of uranium. Yongbyon-style reactors use natural U metal, no? (Maybe not — I’m not a reactor guy.)

    That they weren’t DU reduces the likelihood of the Syrian claim that they originated in the Israeli weapons. Doesn’t totally discredit it, but moves it down a bit.

  5. Hairs (History)

    So, pumping capacity – with appropriate power supply – sufficient for 25 MWth of cooling…

    Of course, this doesn’t prove that the water was intended for cooling (in principle it could be dilution water, flushing water, etc) but if we stick with the cooling hypothesis, and assume a reasonable figure for the temperature rise in the cooling circuit (e.g. from 20 deg C to 40 deg C) then this implies a pumping capacity of about 300 kg/s. That’s a lot of water for anything other than cooling.

    For my money 25 MWth is way too much to be accounted for by the cooling requirements of equipment (e.g. cooling of lubricating oil circuits, hydraulic circuits, transformers, compressors, etc). Additionally, any equipment that is generating 25 MWth of waste heat must be turning over a lot of MW in the main process; for an example of the scale involved, 25 MWe is the auxiliary demand (and hence the equipment’s waste heat) of a 1000 MW power station. Therefore if the pumping capacity was to deal with equipment’s waste heat, then the main process would have to be on the scale of 1000 MW or so, and there’s nothing in terms of incoming fuel lines, incoming power cables, or exhaust stacks to support that kind of energy.

    Conclusion: 25 MWth was the themal size of the main process.

    If it’s a chemical process then 25 MWth is equivalent to burning about 40 tonnes of diesel per day. For me this argues against any kind of missile work because if I allow 10 tonnes of propellant per missile it implies that they were going to turnover the energy equivalent of 4 missiles’ worth of propellant per day for an extended period. Similarly, if it was a chemical weapons plant – involving 4 strongly exothermic processes – it implies continuous processing of about 10 tonnes of product per day for an extended period.

    I’m trying very hard to keep an open mind, but unless our beloved leader has installed the mother of all jacuzzis in his swimming pool, this says “power station” at me.

    With no obvious external truck unloading facilities or storage tanks at the BOE, the fuel for such a power station could either come by pipeline or else it’s nuclear. I suppose if I stretch my imagination far enough, the “water treatment plant” some kilometres away could really be a fuel unloading facility in disguise, with transfer via the observed underground pipe. Alternatively there could be fuel unloading and storage inside the BOE. But in both cases, why?? If it’s just a conventionally-fired power plant, or even some conventionally-fired process (cement factory, anyone?), why complicate matters.

    25 MWth of cooling capability means that someone was expecting 25 MWth of heat generation, and that heat’s gotta come from somewhere. My bet is still on the BOE being an unfinished nuclear reactor i.e. structurally pretty much together but without moderator or fuel.

    Specifically on those two topics:

    If moderator (i.e. graphite) had been present during the bombing I suspect it should have either burnt like a torch (a la Chernobyl or Windscale), which would have glowed wonderfully in IR satellite imagery, or else been scattered to the four winds, in which case the IAEA should surely have detected at least a particle or two. Instead the IAEA report makes no mention of graphite.

    In the case of (nuclear) fuel, the IAEA does indeed mention finding some uranium particles, and “…that the uranium is anthropogenic, i.e. that the material was produced as a result of chemical processing”. But surely these guys are smart enough to say nuclear processing / enrichment if that is what they mean; yet they don’t. “Chemical processing” to me means it’s not raw ore out the ground, but it’s not enriched either. Yes, it could be natural uranium-metal fuel, but the Syrian claim of debris from uranium weapons seems at least possible. Is it plausible, for example, for the Israelis to have attacked the concrete building with one munition, and then attacked an exposed (probably steel) pressure vessel with armour-piercing weaponry that uses uranium? At the risk of undermining myself, I’d be grateful if someone who understands weapons could put me right on that one…

  6. Ataune (History)


    What you make of NYT reports today then ?

  7. Cernig (History)

    Allen, the report doesn’t say whether or not it was DU, just some anon guy quoted in the WaPo.

    Ataune, that NYT story is a masterpiece of spin. There was enough LEU at 520kg, maybe even as little as 400Kg, yet Iran didn’t try a breakout at that threshold.

    A breakout- kicking out the inspectors, breaking seals and switching of cameras, would be a dead giveaway. It would take at least 2 months thereafter to enrich that LEU to weapons grade, and that’s to say nothing of actually building a bomb with it afterwards.

    “the Agency currently has no information
    — apart from the uranium metal document — on the actual design or manufacture by Iran of nuclear
    material components of a nuclear weapon or of certain other key components, such as initiators, or on
    related nuclear physics studies (GOV/2008/38, para. 21). Nor has the Agency detected the actual use
    of nuclear material in connection with the alleged studies.”

    Regards, C

  8. Yossi

    A few preliminary remarks:

    * ACW readers found that BoE was constructed between May 26, 2001 and September 5, 2002. Before the first date there is no structure and after the second there is a nice box. IAEA claims construction started between 26 April 2001 and 4 August 2001. The first date seems suspect.

    * The Syrians say no materials used in building nuclear reactors were found and IAEA don’t contradict them. This probably means they are right. In particular we understand that no trace of the supposed hundreds of tons of nuclear grade graphite bricks was found. If a Magnox reactor was built on the site the cleaners did a good job (the main structures were demolished by explosion and probably spewed a lot of debris).

    * IAEA says that in spite of the excellent cleaning job the environmental samples revealed a significant number of “natural Uranium particles”. The material’s type indicates they have been produced as a result of chemical processing.

    The Syrians says that one sample had 3 Uranium particles and 4 other samples had none although all were taken in a 30m radius. They maintain the only possible source is Israeli munitions. The IAEA seems to accept this is the only possible explanation although there are other alternatives. The IAEA intends to question Israel about its munitions but this is phrased obliquely. It’s not clear to me if Israel is to be questioned out of professional or diplomatic needs.

    The Syrians apparently don’t have a fuel production facility so near an alleged Magnox reactor we would expect any Uranium found to be accompanied by Magnesium and Zirconium, the ingredients of the cladding. These materials, like Uranium (yes I know it’s a controversial issue), are used also in some US aircraft munitions. If the IAEA had found them they don’t say a word so I guess they were not found which supports the non-nuclear theory.

    The term “natural Uranium” means it’s not enriched nor depleted. Note that the measurements involved require fairly large particles to be statistically reliable. Magnox fuel is usually called natural Uranium and is either the pure metal form or Uranium Dioxide (as a ceramic?). Basic Magnox technology probably don’t use UO3 or U3O8.

    It’s still not clear what Uranium compound was found so I’m not sure if the Homs yellowcake hypothesis is dead.

    * The IAEA confirms the pump station had sufficient capacity to cool a 25MWth reactor and enough electrical power to operate (contra to Syria initial claim). They don’t mention the similar pumping stations in the vicinity or the pipeline that connected BoE to the supposed Water Treatment Facility. They don’t mention if they checked the number of pipes between BoE and the pumping station.

    Maybe the IAEA didn’t even visit the alleged Water Treatment Facility? This could be a good check of another Syrian initial claim, that they don’t have available the required large quantities of treated water. Our Hairs gets here ten points for raising the treated water issue…

    * The IAEA got satellite images (not UAV images nor ground photos?) from “member states” that it will share with Syria. According to these images BoE external or internal features are similar to those “found in connection with a reactor site”. Do they mean the alleged reactor vessel?

  9. Allen Thomson (History)

    On the weaponeering part, I agree with someone upthread that non-penetrating iron bombs, probably with JDAM kits or maybe with Paveway laser-guided ones, are the the most likely to have been used.

    Cracking the core with penetrators would have been unnecessary: destroying the various other machinery — pumps, heat exchangers, etc. — around the core in the putative reactor building would have done the job equally well.

    Not cracking the core, of course, leaves its contents, such as graphite, in place. But then where would the uranium particles have come from? Unloaded fuel elements? If so, as already asked, what about cladding?

  10. Josh

    Magnox fuel is made with green salt (UF4). The process is described here:

    The process appears to have been pretty similar in North Korea, too.

    There are a few conversion steps between yellowcake and green salt. If the uranium traces were Magnox-related, then they could have been either U3O8 (yellowcake), UO3, UO2, or UF4. These four substances reflect the three process lines at Yongbyon.

  11. Yossi

    A few more remarks on the Syria report:

    * Since it was considered an almost finished Magnox reactor (i.e. the ground photos), BoE was probably attacked with high penetration bombs. The attackers surely wanted to break through the upper concrete biological shield and destroy the core. Destroying only the peripherals and sparing the core is not what we would expect from an hysterical attacker on a mission to save his country from an imminent nuclear annihilation.

    We would expect some graphite to be blown off and if the great mass (hundreds of tons!) caught fire to find un-burnt particles carried up by the convective draft all over the region. It seems no graphite was found.

    * The fuel in NK Magnox technology is natural Uranium metal. This is a pyrophoric material that burns at a very high temperature. USIC claimed the Syrian reactor was not loaded with fuel so the connection with the Uranium found by the IAEA is not clear. If there was fuel inside the reactor or stored outside we would expect to find traces of either natural Uranium metal or an oxide. The “produced as result of chemical processing” terminology of IAEA certainly applies to the metal form but some Uranium oxides can be found in nature.

    The ingredients of Magnox fuel cladding, Magnesium and Zirconium, also burn nicely and we would expect to find traces of either the metal form or the oxides. It seems none were found.

    * Hairs analysis of BoE thermal economy is very interesting and important but a CW facility is still a viable option. The loophole in the analysis is that the pumped water were not necessarily used on a continuous basis. Hairs himself raises the possibility of “flushing water” and drops it without sufficient reason.

    The high pumping capacity (probably similar to the other pumping stations in the vicinity) could be an emergency measure in case of a CW accident. The large concrete cylinder housed the CW warhead together with a few barrels of Sodium Hydroxide. In an emergency the cylinder was quickly filled with water, dissolving the NaOH and neutralizing the nerve agent.

    * The Uranium used in military munitions is probably also in the metal form. For this purpose natural Uranium is as good as DU. Natural Uranium could be used instead of DU in the case DU is not readily available (a country without significant enrichment facilities) or to facilitate plausible deniability. Because it’s pyrophoric we would expect Uranium from military munitions to be found in the oxide form.

  12. Yossi

    Sorry, but I have some more remarks on the Syria report.

    * Another reason that BoE attackers probably tried to maximize damage is that the 2007 strike could well be their only opportunity. If the Syrians moved there some state of the art Russian SAMs (e.g. Pantsyr, Tor, S-300) a second strike would be very dangerous. Note there were then rumors on Syria acquiring Pantsyrs and some sources said Israel was in a hurry to attack before they arrived. In Digital Globe images there was what seems like a single SAM launching post in the vicinity.

    To maximize damage the core had to be destroyed. I don’t believe the attackers would have settled for destroying the heat exchangers, spent fuel pond, workshops etc. Suppose they missed and the damage was repairable? The threat was too severe for such gambling.

    * In the probably fake “construction photo” we see a massive concrete superstructure above the “reactor hall”. A convenient way to get access to the “reactor hall” would be to use penetrators and pierce the superstructure. Another way was to lobe bombs into the “reactor hall” from the sides but then you probably couldn’t destroy the core.

    In the post-strike UAV images you can see the “reactor hall” walls fairly complete. The superstructure had collapsed inside and covered it with debris. This is not what you would expect from large non-penetrating bombs but from penetrators. Note that the signs of a high-penetration bomb are difficult to see in such case as it burrows deep into the earth and the entrance hole may collapse and certainly was covered with the superstructure debris.

    An interesting speculation is that the deep hole seen in the last stage of BoE razing may have been dug by the Syrians who wished to remove remnants of the penetrating bombs. Maybe they were worried the bombs contained Uranium and it would be used to incriminate them. It would be ironic if the Uranium traces found by the IAEA were dug out by the Syrians themselves to avoid framing.

    * Syria is a poor and backward country whose scientific infrastructure is not only insufficient for a military nuclear program but even to defend itself against such an accusation. A technical debate with a global superpower wouldn’t be a fair fight. If the US picked on someone its own size, say Russia or China with such dubious evidence it would probably have been defeated humiliatingly.

    The GWB policy of marking countries as enemies then blindly hitting them again and again had failed. It wouldn’t be wise if small Israel adopted a policy that a global superpower couldn’t carry out just because of gratitude to her great friend. There are many other ways to express this very deep gratitude. It would serve no one if the saga of the alleged Syrian reactor continue rolling in the IAEA corridors and poison the regional peace talks. The outcome could be either humiliation for Syria or disgrace for the Israeli IC, both bad for Israel’s interests. The best way is to have Israel un-officially help Syria get out of this mess and earn a good will point in Damascus on the way.

  13. Hairs (History)


    You’re quite right – I should have given some justification for not considering water use other than for cooling.

    River water is (comparatively) filthy stuff, and I’ve rarely seen it used directly for anything except for cooling purposes. I think it is unlikely that the Syrians would use raw river water directly as part of a chemical process because of the contamination it would cause. I added the “in principle” qualifier (weasel words, perhaps!?) because I supposed it was possible that they’d want to dilute some discharge before returning it to the river, although that then begs the question “Where does / did the bulk high-quality water come from?”.

    I agree that the water need not be pumped on a continuous basis, but I would argue that an intermittent water consumption actually makes the alternatives to a reactor even more unlikely. For example, let’s suppose that the water was to absorb the heat liberated by destroying a missile’s worth of propellant. If the propellant is oxidised continuously then the you’ll need cooling capacity equal to the heat liberation rate. But if you oxidise in batches (say 12 hours on, 12 hours off) then you’ll need double the cooling capacity because the heat liberation rate is doubled. Now I know all this talk of destroying a missile’s propellant is probably nonsense (I’m pretty ignorant about misslies), but the point is that your maximum-cooling-capacity is minimised by releasing the heat continuously. Therefore instead of putting in 25 MWth’s worth of pumping for an intermittent demand, the Syrians could have installed just, say, 5 MWth’s worth of pumping and created a continuous (but lower) cooling demand. I can’t imagine why they’d want to batch process something, at the expense of increased cooling demand (something that would be far more likely to show up on infra red imagery), rather than continuously process it at a much lower cooling rate.

    The general point about whether the pumps were designed for intermittent or continuous demand is certainly a valid one. In principle one could get an indication of what was intended by looking at the switchgear feeding the motor, checking the valve arrangements, etc. Certainly if the pump’s min-flow bypass was scoured clean by river silt it would strongly suggest intermittent, or very low flow, operation. (Sometimes I wish I could join an inspection team for a few hours to satisfy my own curiosity on certain points!).

    With regard to the pump being part of an emergency system, I think this would be even less likely. For example, if a digger cuts through the power cable, or if lightning strikes the switchgear, then you’ve got no emergency system. A typical emergency system for something as dangerous as chemical weapons (e.g. emergency flooding systems in refineries, or emergency core cooling systems in nuclear reactors) tends to rely on the energy for delivery already being present (e.g. rods dropping under gravity or water being injected by the pressure of a continuously maintained inert, high-pressure gas system) and not on the – perhaps unreliable – delivery of electrical energy just when the system is needed. If the Syrians were using the pump for some emergency purpose then I think the minimum we could expect is to see 3 or 4 pumps, each with 100% capacity: thus, one in operation, one or two on hot standby, and one down for maintenance / testing.

  14. Yossi

    Hairs thanks.

    What do you think about the following speculation?

    There is near BoE’s southern side a large structure (30m X 10m X 5?m) partly covered with earth that may have been an above ground water tank (~1500 cubic meters, a little more than the volume of Yongbyon’s pressure vessel). It looks like a concrete cast and survived the strike and BoE dismantling. It can be seen very clearly in the post-strike UAV images.

    How about the riverside pump station filling this container with water, to be used immediately in case of emergency, while the pump itself stands as backup?

  15. Hairs (History)

    Hmmm… sounds intriguing, Yossi. Do you have an address / link where I could see it, please?

    My immediate thoughts are that something of that size would be about right for a seal pit. It’s probably twice the size strictly necessary from a flow point of view, but perhaps there’s also an aspect of trying to treat / cool the water before return to the river.

    A seal pit is used to create a siphon effect in the cooling water flow – basically the suction of water flowing “down” towards the river helps to pull incoming water “up” from the pumps, which reduces the required pumping power.

    Normally the seal pit would be located fairly close to the final point of discharge, which happens to be near the pumps. However, in the case of the BOE there is about 70m difference in elevation between the pumps at the river and the BOE itself. This means that the Syrians couldn’t put a seal pit near the intake pumps. This is because a vacuum can’t hold a water column much taller than 10m, thus the water would drop out of the reactor towards the river, leaving a vacuum (or, actually, a gas lock) just where you need the cooling. There are three main options for preventing this:

    1) Pump the water around so forcefully that the flow resistance keeps the pipe full. If it helps, imagine the downspout of normal household guttering; at a low flow the downspout cannot be kept full, instead the water just drops out of it. However, at a high enough flow (e.g. if you put a fire-hose to the top!) then the water can’t escape as quickly as it’s coming in, and consequently the pipe remains full.

    2) Put a discharge valve at the outlet (in the analogy, put it at the bottom of the downspout) in order to hold the water in.

    3) Let the water discharge into a seal pit that’s not more than 10m lower than the highest point of the piping.

    Option 1 increases the required pumping power massively, and also restricts operation to maximum capacity only. Option 2 wastes pumping capacity because a lot of your pumping power is going towards overcoming the flow resistance (i.e. the discharge valve) that you put in yourself! Which leaves option 3 as the most efficient way of reducing pumping power.

    From a seal pit the water can return to the river under gravity, thus seal pits are a cheap and simple way of minimising pumping power for cooling.

    Coming back to your point about the structure being used as an emergency system, I still have doubts even without seeing it. If the 1500m3 of water is being held at ground level (or even below ground level if the tank has been set into the ground) then unless the facility you wish to flood is at an even lower elevation you still face the problem of getting the water into the facility. Discounting gravity, the most common system for guaranteeing that you flood the required volume is to inject the water using compressed gas. In this case you’d have a discharge line at the bottom of your tank, and inject high pressure gas (usually “inert” gas such as nitrogen) over the top of the water. The resulting pressure will drive the water out of the discharge to where it’s got to go.

    If a gas-driven discharge is used then the tank would probably be steel (concrete’s not so good under tension, so it’s not very popular for vessels that will pressurised internally) and I’d expect to see some sort of gas facility e.g. bundles of nitrogen bottles or a separate gas tank in the vicinity. A typical nitrogen bottle is around 30 litres at 200 bar, and a VERY gross appoximation says this will expand to 600 litres at 10 bar (the sort of injection pressure you’d want for a fast, fail-safe response). So to displace 1500m3 of water you’d be looking at something around 2500 bottles. Even if we drop the injection pressure to 2 bar we’re still talking 500 bottles, so let’s say something of the order of 1000 bottles for a gas-driven injection system. In reality 1000 bottles represents too many sources of leaks – I mean, the guys would be running around all day doing nothing but leak testing connections and changing bottles – so more likely a gas-driven system would rely on a gas tank (or two for redundancy). But tanks are much harder to pressurise because they have to be proportionately stronger than a bottle. So if tanks were used at, say, 50 bar the 1000 bottles would equate to something like 2 × 6000 litre tanks. I know 6m3 isn’t particularly large, but I’d guess if two 6m3 tanks were installed at the time of the photo then they ought to be visible (unless they were inside the main building for some reason).

    Lastly, I still can’t get over the disadvantages of using raw river water for any kind of emergency system. Besides the risk of corroding / fouling the emergency system itself, there’s simply no way to know what horrible gunk (fertiliser run-off, discharge from an upstream factory, oil slick, etc) might happen to be in the water just when you want to use it. If I was in the Syrians’ situation the last thing I’d want to do in a chemical emergency, be it rocket propellants or chemical weapons, is to start flooding or spraying with water containing unknown contaminants. Even basic fire-fighting water is treated to be just short of drinking quality; for emergency chemical applications I’d want to have nothing less than good quality demin water, so that I’d know exactly what reactions may or may occur once it starts flowing.