Jeffrey LewisPatton on Pakistan’s U Supply

My colleague, Tamara Patton, has published her much anticipated master’s thesis on estimating the size of the Pakistani plutonium production reactors at Khushab in Science and Global Security  (Combining Satellite Imagery and 3D Drawing Tools for Nonproliferation Analysis: A Case Study of Pakistan’s Khushab Plutonium Production Reactors).  It’s one heck of a paper and available for free.

I asked Tamara to trace one of the policy implications that didn’t make the final cut of the paper — and she kindly agreed, writing this fascinating little piece about Pakistan’s shrinking supply of uranium:

Uranium Fuel Constraints for Pakistan’s Nuclear Weapon Complex

Tamara Patton, December 2012

 As construction on a fourth reactor at Pakistan’s weapons-grade plutonium production complex at Khushab continues apace, an important question is where the government plans to get the uranium needed to fuel its growing fleet of reactors.

The answer cannot be ‘Pakistan’ for much longer, at least not without severe difficulties. Pakistan is not a signatory to the Nonproliferation Treaty, which complicates the import of uranium. Pakistan has been able to secure Chinese LEU fuel assemblies for the Chasma Nuclear Power plants and a limited stock of safeguarded natural uranium fuel assemblies for the Karachi Nuclear Power Plant (KANUPP).  However, as Canada stopped supplying Pakistan with fuel assemblies for KANUPP in 1976, this stock is most likely gone by now, causing KANUPP to rely on domestic stocks of uranium in recent decades.  The weapons program, including military HEU production and fabrication of fuel for the reactors at Khushab, must also rely on domestic production. Pakistan’s Bagalchore mine was reportedly exhausted and closed by 2000, so uranium resources now only come from the Qabul Khel mine (opened in 1992), the Nanganai deposit (1996), and Taunsa deposits (2002), all using in situ leaching.  Current domestic production estimates from these sources stand at 40 tons of uranium per year.

A 2009 study by Mian, Nayyar, and Rajaraman estimates that when applied to the fueling of the Khushab fleet of reactors, the 40 tons per year amount alone can only support approximately 150 MWt of total reactor capacity operating at 70 percent efficiency and a low burnup of 1000 MWd/ton.  Forty tons would just barely support the first three reactors.  Today, there is a fourth.

In my recent paper, Combining Satellite Imagery and 3D Drawing Tools for Nonproliferation Analysis: A Case Study of Pakistan’s Khushab Plutonium Production Reactors, I sought to refine maximum thermal capacity estimates of the reactors based on 3D analysis of each reactor’s cooling towers (snapshot below).  Using these estimates, the table here shows how the completed four reactors at Khushab would operate at around a total of 200 MWt at 70 percent efficiency, which translates to a requirement of as much as ~70 tons of uranium per year.  The reactor capacity estimates in my paper are upper limits based on cooling capacity.  Seventy tons of uranium is therefore also an upper limit.  The reactors could be slightly smaller, with overdesigned cooling systems, or Pakistan may plan to operate the reactors at a lower capacity.  Still, Pakistan appears likely to run a uranium deficit, perhaps as much as 30 tons, that could exhaust uranium stocks and eventually the deposits themselves.


Building the Khushab reactors in Google SketchUp to estimate reactor thermal capacity and plutonium production capabilities. See Science and Global Security v.20 no. 2.

If Pakistan runs a consistent uranium deficit, the demand for fuel for the Khushab reactors may begin to starve Pakistan’s enrichment complex for HEU production.  The construction of a fourth plutonium production reactor in the face of limited uranium resources seems to signal a shift in Pakistan’s priorities from larger uranium-based to smaller plutonium-based nuclear warheads.  I estimate that the four reactors at Khushab could eventually produce between 60-70 kg of plutonium per year – enough for 10-20 warheads per year, depending on how plutonium is used in each warhead.  This is a significant capacity, and the quick pace of construction at Khushab is one indicator that Pakistan’s resources may be gravitating to this site.  Nevertheless, it’s too soon to rule out the continued production of uranium-based weapons in the future.

But where will Pakistan turn for more uranium if needed?  One possibility is that it can use as feed its accumulated depleted tails in addition to reprocessed uranium.  For every one ton of depleted uranium (with 0.2%U-235), Pakistan could re-enrich it to acquire approximately 0.27 tons of uranium with the 0.7% U-235 content of natural uranium needed to fuel the Khushab reactors. [Author’s correction: 0.2% U-235 is correct only for some PHWRs.  The number for Khushab should be around 0.6%.  See page 87 of this report for Alexander Glaser’s calculations on the isotopic composition of spent fuel from a low-burn-up 40 MWt heavy water reactor.]  Considering that Pakistan has been enriching uranium since around 1978, this is a not an inconsiderable amount.  Spent fuel from heavy water reactors like Khushab also contains about 0.2% U-235, and Pakistan may possess the capability to reprocess this material at its New Labs facility.

Another interesting possibility is the extraction of uranium from rock phosphate, which according to this report produced in Pakistan can contain anywhere from 0.005-0.04% U3O8, or yellowcake.  Phosphoric acid, obtained from phosphate rock, is the basic ingredient in di-ammonium phosphate (DAP), one of the most commonly used fertilizers in the world. According to these statistics by Pakistan’s National Fertilizer Development Centre, domestic production of DAP began in 1999.

Uranium should be removed from the phosphoric acid before it is converted to fertilizer, lest it end up in the final consumer products and pose a health hazard.  This can be done in a variety of ways, as described in this 1987 IAEA report.  Although less economical than mining, uranium extracted from phosphates prior to fertilizer production is potentially a significant source of uranium for a weapons program.  Not surprisingly, Syria was recently very interested in this process.

 Along with the fact that phosphoric acid is widely traded for the production of fertilizer, it is also not subject to heavy scrutiny through export controls.  Morocco in particular is a major exporter of phosphoric acid as it holds nearly 77% of worldwide phosphate rock reserves.  In recent years, Pakistan and Morocco have established a joint venture to ensure “uninterrupted supply” of phosphoric acid to Pakistan on a large scale. Export of phosphoric acid, a legitimate commodity, is not prohibited and there is no evidence that the joint venture is supporting Pakistan’s nuclear weapons program or engaged in any nefarious activities.  An important question is how much uranium may be inadvertently transported through the trade of phosphoric acid for DAP production.  This study shows how appreciable amounts of uranium present in processed phosphate rocks can pass into phosphoric acid and then to fertilizers.  Of all the materials tested in the study, DAP fertilizer is shown as possessing the highest end concentration of uranium: about 52 mg/kg.  Theoretically speaking, many factors could affect how much uranium Pakistan could acquire through the production of DAP.  Factors such as the uranium content of the original phosphate rock, the purity of the phosphoric acid received in Pakistan, and the method of extraction used to acquire the uranium would all affect the final amount of uranium.  Whatever the case may be, this is a potentially significant source of uranium for Pakistan, one that bears close scrutiny.

Reactor Thermal power estimate at 70% capacity factor (MWt) Operational days per year (days) Burnup (MWd/ton) U required per year (tons)
Khushab I 34 365 1000 12
Khushab II 46 365 1000 17
Khushab III 57 365 1000 21
Khushab IV 57 365 1000 21
                                                     Total U required per year 71

Estimated uranium fuel requirements for the Khushab reactors based on reactor thermal capacity estimates derived from limits of the cooling towers.




  1. Pavel (History)

    Free download is not available yet – S&GS posts free versions of its articles one year after publication.

    • Jeffrey (History)

      It’s listed as a “free .pdf” and works for me.

  2. David (History)

    I tried and have the same result as Pavel (above). If you are using the website inside the firewall of a university (or research institute), they may have paid access which is allowed based on the IP address of the computer or library.

    • Jeffrey (History)

      Must be. Although it shows up as free in mobile device from home …

    • Ano N. Ymous (History)

      Shows up as “Free PDF” for me too, but I can’t find a link that doesn’t lead to a “pay now” page.

  3. Robert Merkel (History)

    I can’t see a free download link either.

    One key question – as I understand it (from earlier publications) public estimates of Pakistan’s arsenal are in the order of 100 warheads. That seems rather a lot already.

    Are there any public analyses of how many warheads they want, and why? Is it simply an “as many as we can build and preferably more than India” approach?

  4. George William Herbert (History)

    Any reason they could not reprocess fuel for reuse?

    The US not doing so is well known, but others have done so succesfully. We don’t do it because it’s a proliferation enhancer.

    • cthippo (History)

      Makes sense. If they’re reprocessing the fuel to extract the plutonium, then by definition they’ve already extracted the uranium and it’s just one more step to precipitate it out and feed it back into the fuel cycle.

      Reactor fuel is usually considered “spent” not when all the U235 is burned up, but when the nuclear “ashes”, i.e. daughter particles, become so numerous that they interfere with the reaction. In plutonium production reactors such as the ones at Khushab the fuel is run for an even shorter period of time which is inefficient in terms of uranium usage, but makes for better plutonium. The fuel coming out of these reactors would tend to still have a lot of life left in it.

      Back in the 80’s (I think) Idaho National Engineering Lab demonstrated a closed cycle reprocessing process in which the fuel was unloaded from the reactor, the daughter particles were stripped out and the remaining fuel went straight back into the reactor without separation, all without leaving containment. the recycled fuel was still highly radioactive and therefore considered less proliferation sensitive than conventional reprocessing.

      The problem with reprocessing for fuel efficiency is economic and political, rather than technical. If you have lots and lots of uranium ore then it’s cheaper and easier to feed it through the reactor once then dispose of it (or let it sit in a pool forever). However, if, like Pakistan (and India until they got access to international markets), you are lacking in ore and are not concerned with the proliferation risks of your plutonium, then reprocessing for fuel makes total sense and should reduce the amount of ore you need annually.

      I think this is a case where it’s easy to forget that just because no one does something doesn’t mean that it can’t be done.

    • rwendland (History)

      There are various detailed problems which make reuse of RepU not as easy as might be imagined. The IAEA produced a report on it in 2009 “Use of Reprocessed Uranium: Challenges and Options” (NF-T-4.4):

      Among the issues are:

      * chemical impurities (eg chloride, sulphate, 99Tc and Pu)
      * minor uranium isotopes content about 6 times higher (though can blend with NatU to reduce problems), eg requires higher levels of enrichment for same fuel value (6% instead of 5%)
      * fuel product is more radioactive, increasing radiation doses for operation staff

      Biggest issue is the economics are very poor at current U prices, but perhaps that would not matter for Pakistan.

  5. Tamara (History)

    Robert, it’s a good question without a good answer. When talking about arsenal size ambitions, one measure we can consider is mentioned in Pakistani Ambassador Zamir Akram’s February 2010 statement to the CD. In the context of his argument as to why Pakistan cannot engage in negotiations toward an FMCT, he mentions that India is aiming for an arsenal of 400 nuclear weapons with a triad of delivery systems involving air, land and sea based capabilities (see here for full statement: ). While he does not state that Pakistan is also aiming for this number, from this statement, we at least have a better understanding of the perceived threat that Pakistan may be trying to address in developing its own force, not to mention India’s growing conventional capabilities.

    George, indeed, there is no good reason they could not already be reprocessing. Satellite imagery from 2009 showed significant expansions to Pakistan’s suspected reprocessing site (New Labs). Mian, Nayyar, and Rajaraman’s 2009 study goes into great detail on Pakistan’s reprocessing capabilities, as well as past indications that reprocessing may have occurred.

    Apologies to those who can’t access the article. As an author, I believe I’m allowed to share a few free copies, so please let me know if you’re interested.

    • Cthippo (History)

      Awesome article, Tamara, and congratulations on getting it done!

      I’m wondering about the assertion that the spent fuel contains only about 0.2% U235. Given that the reactor is optimized for Pu production and is therefore being run for relatively short periods on each fuel loading that seems like an awfully high burnup rate. From what I’m reading, it seems like second generation power reactors optimized for maximum efficiency struggle to achieve 6% burnup and the ones on the drawing board might get to 9%.

      Can you elaborate on where that number came from?

    • rwendland (History)

      That 0.2% U235 remaining claim does sound implausible.

      The UK Magnox RepU remaining (called Magnox depleted uranium or MDU in the UK) typically has 0.4% U235, and that is nearly all from a commercial cycle trying to get best burnup for economic reasons from NatU. [source is page 5 of the IAEA report I link to above]

      The UK has reused 17000 of its 41000 tonnes (as of 2009) of Magnox RepU for fuel. But has stopped doing that as it was uneconomic. Because of the relatively low burnup in Magnox, its RepU was much easier to reuse than from reactors that use LEU.

    • rwendland (History)

      To correct myself above, 0.2% U235 remaining is entirely plausible for PHWR. I should have read Tamara’s comment below! I’d failed to appreciate that PHWR can manage near twice the NatU burnup of GCRs like Magnox. Whoops.

    • RAJ47 (History)

      Please send me a copy at [my gmail address]
      Thanks in anticipation.
      I have seen your presentations earlier too, they are simply impressive.
      Pakistan has reopened Baghlachur in a big way around 2008/9.
      The Google Earth will provide all the details.
      ISIS has a lot of satellite imagery analysis on the subject of doubling nuke production.
      Nellore too has an additional reactor which most analysts have not taken into account.
      New mines at Baghlachur @
      30 03 10 70 21 56 – Between 2004-2009
      30 03 23 70 22 14 – Between 2004-2009, closer to 2009
      30 03 26 70 22 23 – Between 2009-2011
      The security has been beefed up with double fencing.
      If you like, I can send a kmz file too.

  6. Mansoor (History)

    Pakistan currently fields nine different types of ballistic and cruise missile systems. Hypothetically, each of these delivery systems may be allocated 10 warheads. In addition, the air force would also have its own claim on a few dozen warheads. More importantly however, the introduction of TNWs ought to be taken into account while estimating Pakistan’s fissile material requirements and possible future production trajectories.

    In this context, Pakistan barely has sufficient fissile material (especially plutonium) for its current delivery systems and TNWs. As for the number of warheads, if India goes for a triad-based force of 400 warheads, Pakistan will probably quantify its own force goals that keeps its minimum deterrent credible, but does not necessarily have to be a force as large.

    The development and induction of cruise missiles (land, air and sea versions) would also need more fissile material as Pakistan appears to be diversifying its counter-force targeting options.

    Having said that, there are limits to Pakistan’s fissile material production as mentioned above and the country is unlikely to compromise on its plutonium production capability until it begins to run out of indigenous natural uranium, unless fresh reserves are found within the country. This is evident from the likely completion and anticipated commissioning of the commercial-scale Chashma reprocessing plant. So Pakistan’s position on FMCT will probably remain unchanged for foreseeable future for strategic and domestic-political reasons.

  7. MK (History)

    Can’t yet access the article, but congratulations are clearly in order.
    Have you written about the possibility that Kahuta’s future role might change from helping to build HEU-based devices to servicing a plutonium-based arsenal?

  8. Tamara (History)

    Thanks for the kind words, Cthippo, and thanks also for the very important correction you mention. Typical spent fuel from PHWRs contains about 0.2-0.3% U-235 (CANDU, for instance, puts out about 0.23%), which is why I mentioned that number in the context of Khushab. However, as you correctly pointed out, when optimized for Pu-239 production, the shorter running periods and lower burn-up of the reactor will make for a higher remaining percentage of U-235. The number mentioned by Mian, Nayyar, and Rajaraman – 0.616%, makes much more sense in this case. It’s an important point, and I think it underlines the notion that reprocessing spent fuel from the Khushab fleet would make sense for Pakistan if in serious need of more uranium.

    MK, unfortunately I don’t have more to offer on the future role of Kahuta at this point. The expansion at Khushab despite limited uranium resources certainly seems to signal a shift in priorities, but I think it’s too soon to rule out continued production of uranium-based warheads. For one, Pakistan has a lot more experience with them, and they may turn out to be the more reliable option. As another consideration, we may be underestimating the amount of uranium Pakistan has access to, especially given possibilities like extraction from phosphates, which could allow Pakistan the flexibility to produce both HEU and Pu-based warheads. It’s certainly an interesting question though, and I think watching to see how the relevant facilities develop over time will be important.

    • Rajaram (History)

      Congratulations. Good analysis, but would like to read the full article. I and my colleagues have been wondering how much confidence Pakistan may be putting in the Pu based weapons, given the fact that they have not tested them. The reliance on HEU weapons may continue?

  9. tamara (History)

    Thank you, Rajaram. And yes, I think we’re all wondering the same thing… It’s certainly interesting to consider the possibility that Islamabad may not have a high level of confidence in smaller Pu-based warheads when it is so clearly sending a message to the world that it is. With ISPR sending out regular press releases with statements that missiles seemingly requiring smaller Pu-warheads are nuclear capable, we can at least be sure that Pakistan wants its adversaries to believe it is confident…deterrence is more about perceived capabilities than actual capabilities, after all. In my view, the major resource shift toward Khushab is perhaps the biggest indicator that Islamabad has “enough” confidence in a plutonium implosion device. Cannibalizing Kahuta doesn’t seem like a move one would make on a hunch. Nevertheless, as I mentioned in another comment, we need to think more about possible peripheral uranium resources like phosphates to get a better understanding of whether Pakistan could afford to produce both HEU and Pu-warheads in the longer term.

    Jeffrey started a good conversation related to this in December last year:

  10. saf (History)

    I have posted extracts from your article on the website above.
    I have mentioned you as the author.
    The article was posted for public discussion.
    Feel free to join in.
    If you want the article to be removed,please let me know.