To understand the challenges and opportunities that states face in acquiring nuclear weapons, it helps to know a little about what nuclear weapons are, how they originated, why few countries were able to develop them in the 20th century, and how that could change in the 21st.

Nuclear weapons were first invented in the United States during World War II after an intensive and secretive research and development project now known as the Manhattan Project. The concept of a nuclear weapon by means of fission chain reactions was only 2 years old when the project started, and it was far from certain that it could ever work. Despite discovering fission, the Germans, led by the theoretical physicist Werner Heisenberg, reportedly miscalculated the critical mass needed for an atomic weapon and deemed the concept impossible.

The idea of a nuclear chain reaction was conceived six years before the discovery of fission, which made the atom bomb possible. The Hungarian American physicist Leo Szilard had proposed a mechanism for this chain reaction in 1933 after the discovery that bombarding the stable isotope lithium-6 with protons produced alpha particles and released energy greater than the energy of the initial proton. Szilard’s idea was to use the energy produced by the reaction to initiate another reaction and employ it as a means to generate thermal power, which could then be converted into electricity. But the concept was not ultimately realizable, as the kinetic energy of the initial proton was only a small part of the total energy needed to accelerate it, especially with the primitive accelerator technology of the time.

The Germans reportedly miscalculated the critical mass needed for an atomic weapon and deemed the concept impossible.

It was nearly a decade later when the Italian physicist Enrico Fermi put together the first sustained nuclear chain reaction in the Chicago Pile-1 (the world’s first artificial nuclear reactor) by means of fission in natural uranium moderated by graphite. Unlike protons, neutrons are uncharged particles that do not need to overcome electrostatic barriers to interact with nuclei. Fission can happen at any neutron energy in fissile materials without the need for acceleration. It was the discovery of fission of uranium by neutrons in Germany in late 1938 that prompted Szilard to write a letter in 1939, signed by Albert Einstein, to President Roosevelt—warning that the Nazis might develop nuclear weapons and advising the United States to seek its own.

Fission reactions are a type of nuclear reaction that occur when a neutron is absorbed by unstable nuclei, such as that of uranium-235 or plutonium-239. It results in the release of multiple neutrons and kinetic energy after the nucleus disintegrates into fission products. The released neutrons may interact and induce additional nuclear reactions, or they may leak and leave the system. Achieving a sustained chain reaction requires maintaining a minimal critical amount of matter, which varies depending on the material composition and density of its constituent nuclides, temperature, shape, and reflection. Below the critical mass (known as the subcritical state), the neutron population decreases in the absence of a constant source, and the chain reaction comes to an end.

A nuclear weapon is a mechanism through which fission chain reactions can be sustained at high reactivity long enough to produce a target energy yield. A basic conceptual nuclear weapon involves fissile materials stored in a subcritical state and a system for transition from the subcritical state to a supercritical state at target reactivity.

There are different types of nuclear weapons that vary in their size, energy yield, and mechanisms used. In a discussion about the proliferation of nuclear weapons, it suffices to consider the challenges that potential proliferator states face prior to weapon design and assembly (a classified subject, but arguably of less importance to new proliferators than it is to the technical development of existing nuclear weapons capabilities).

Simply put, the pursuit of nuclear weapons is slow, expensive, and requires a large industrial complex, with access to the necessary materials and trained expertise. The success of the 20th century’s nonproliferation regime was therefore largely due to inherent physics, in addition to policy. These challenges effectively formed a barrier to potential proliferator states that was and remains difficult to overcome without detection from other states. But technological advances in the 21st century have eased the acquisition of special nuclear materials, and will likely make the pursuit of nuclear weapons cheaper and faster. The role of advanced detection technology and effective international policy have therefore become more critical.

The primary challenge for a 21st-century state seeking nuclear weapons is the production of fissile materials with the necessary quantity and quality. It is widely recognized that once a state has acquired enough fissile materials with adequate purity, nuclear weapons production is only a question of energy yield and reliability, both of which can be improved with time and experience. North Korea’s nuclear weapons in 2006, for example, produced low yields of only 0.7 kilotons of TNT (20 times less than the Hiroshima bomb). By 2017, Pyongyang managed to boost the yield of its weapons to between 70 and 280 kilotons. We should therefore focus more narrowly on the challenges of acquiring fissile material, and on the evolving technologies that make 20th-century nonproliferation strategies increasingly ineffective.

Fissile materials, which are basically materials that can undergo fission at any neutron energy, can be extracted naturally, or produced through the conversion of abundant “fertile” material by neutron capture. We all have at least two reasons to thank nature: First, fissile materials in nature are not available in a form that is directly useful for weapons purposes. Second, for the purposes of fissile material production, it is difficult to release neutrons from nuclei in large quantities and to efficiently slow them down to targeted neutron energies.

Uranium-235, a fissile isotope of uranium, represents only 0.72% of natural uranium, and is chemically inseparable from the uranium-238 isotope which represents 99.27% of uranium. Increasing the concentration of uranium-235 (“uranium enrichment”) is a slow and energy-expensive process that gradually separates uranium isotopes through physical means, as in gas centrifuges. Centrifuges allow the separation of uranium isotopes as the heavier isotope (uranium-238) gains more momentum than the lighter isotope, and becomes relatively concentrated in the outer parts of the centrifuges. Thousands of centrifuges are typically used in enrichment cascades as dictated by the achievable stage separation factors, even for the enrichment levels required for peaceful applications, such as nuclear reactors for power production.

While criticality could be achieved using natural uranium moderated in graphite (as in the Chicago Pile), or in heavy water moderation (as in Canada Deuterium Uranium, or CANDU reactors), the very large critical mass required makes it impossible to construct a deployable weapon with any significant yield, even compared to conventional weapons using moderated natural uranium. For example, the explosion energy yield in the Chernobyl disaster, which was by far the most severe accident involving low-enriched uranium in history, was estimated to be 10 tons of TNT. This is an insignificant amount of energy, considering that just one RBMK-1000 reactor was fueled by nearly 200 tons of 2% enriched uranium. Uranium enrichment, an expensive and slow process, is necessary for any uranium-235-based nuclear weapons.

Uranium-235 is the only fissile material present in nature in appreciable quantities. Plutonium-239 and uranium-233 are present in nature in trace quantities and are normally made in nuclear reactors through the conversion of uranium-238 and thorium-232, respectively. The challenge with creating fissile materials through the conversion of fertile materials such as uranium-238 and thorium-232 is the enormous number of neutrons necessary to obtain a sufficient amount of fissile matter. Based on simple math, to obtain 10 kilograms of plutonium-239, you need to convert 25 trillion trillion uranium-238 atoms. The fact that neutron interactions are probabilistic, and that there are different competing interactions and different materials in a system, means that far more neutrons are required to achieve that number of conversions (without even accounting for the plutonium-239 that would be lost in the process through fission or conversion to plutonium-240).

The open academic literature is rich with extensive discussions on the risk of nuclear weapons proliferation posed by different technologies that can extract neutrons from nuclei. Fortunately, however, large-scale extraction of neutrons for fissile material production requires sophisticated, expensive infrastructure like nuclear reactors or particle accelerators, which eases attempts to employ safeguards for the prevention of weapons proliferation.

Scott Kemp, the director of the MIT Laboratory for Nuclear Security and Policy, published a thorough review of the role of particle accelerators in the proliferation of nuclear weapons by comparing different accelerator designs. He estimated that a proliferator state would need tens of accelerators to acquire enough fissile material for one weapon in two years.

Particle accelerators, especially isochronous cyclotrons, may be attractive to potential proliferator states because of the lack of international safeguards. But as Kemp notes, the accelerators that can achieve high enough charged particle energies for fissile material production are extremely technologically sophisticated. While the main concept is simple, the physics and engineering details involved in such accelerators are cumbersome, and few countries have the technical capability or expertise to build ones with capability for large-scale production of special nuclear material. That said, the increase in legitimate uses of cyclotrons in research and in medical isotope production may ease the acquisition and transfer of such technology.

Particle accelerators are neither the most productive nor the cheapest path to nuclear weapons, but the proliferation risk they pose should not be ignored. High-energy particle accelerators could be used to produce uranium-233 from thorium, which is widely abundant in monazite sands. Unlike natural uranium, which can be used to fuel a reactor and produce fissile material, thorium is useless for the production of fissile material in the absence of other fissile material. In accelerators, however, thorium can be used to produce uranium-233 without other fissile material. Thorium is not inherently proliferation-resistant because of the uranium-232 contaminant, as some proponents of the thorium fuel cycle suggest. Uranium-232 results from interactions that only occur at high neutron energies and may be controlled through moderation of the target.

The threat of proliferation through fissile material production from nuclear reactors, on the other hand, has been recognized since the dawn of nuclear power. Nuclear reactors are systems in which fission chain reactions can be sustained at constant rates but constrained by limits on the structural materials. In nuclear reactors, the neutron population can grow exponentially, subsequent to steady state operation. Exponential growth is what allows nuclear reactors to generate a lot of power and large numbers of neutrons, some of which can be used for fissile material production.

Different types of nuclear reactors vary in plutonium production capability and in the quality of the plutonium produced. Fast spectrum reactors, and reactors whose energy spectrum shifts toward uranium-238 resonance absorption energies, tend to produce larger quantities of plutonium-239. Modern civilian reactors are typically designed to reduce plutonium production. Safeguards are also emplaced to ensure that plutonium is not withdrawn with isotopic purity that would allow for weapons use.

The isotopic purity of plutonium is an important factor in safeguarding plutonium produced in nuclear reactors. Spontaneous fission in plutonium-240 practically sets limits on the purity of plutonium that could otherwise be used in weapons with any significant yield, depending in part on the mechanisms that may be employed for transition to supercriticality. The ratio of plutonium-240 to plutonium-239 in a reactor increases with time due to the conversion of some of the plutonium-239 to plutonium-240 and fissions in plutonium-239. Safeguards ensure that reactors are refueled in long-enough time intervals to prevent the withdrawal of weapons grade plutonium.

There are different levels of nuclear proliferation safeguards, which may include inspections, satellite monitoring of activity, and active on-site surveillance. History has shown that safeguards applied to dual-use technologies do not prevent proliferation in the long term. North Korea used its Yongbyon facility to acquire plutonium for multiple nuclear weapons, even though it was claimed to be a “research facility” and was subject to IAEA inspections in the 1990s. “Research” is often a deliberately vague term, especially when the technology in question is mature and much of the relevant knowledge is accessible in open literature. The low-power graphite moderated reactor that the Yongbyon facility featured had little use other than for the production of plutonium and medical isotopes. Iran has also claimed that its venture into heavy water reactor technology is intended for peaceful purposes like medical isotope production and “research.”

Heavy water reactors, fueled by natural uranium, may be designed to produce large quantities of plutonium for weapons, or they may be used peacefully to generate electricity and medical isotopes (as in CANDU reactors). Iran’s Arak plant was designed to produce 40 megawatt (thermal), which is not useful for the generation of significant amounts of electricity. Since the 1980s, Iran’s interest in dual-use heavy water moderated reactor technologies, as opposed to purely civilian reactors, makes it clear that its intent is not medical isotope production. While there is little doubt that the Iranian nuclear program is primarily intended to develop nuclear weapon production capability, there is far more debate on how to stop it from acquiring nuclear weapons (or weapons-usable materials).

Under the Joint Comprehensive Plan of Action (JCPOA), the Arak reactor should be redesigned to reduce plutonium production, and the spent fuel should be exported for the lifetime of the reactor. The deal could in principle block a primary potential route to plutonium-based nuclear weapons in the foreseeable future. Critics of the deal rightfully argue, however, that it nevertheless provides upfront benefits to the Iranian regime by supporting its industrial and military infrastructure in exchange for a temporary suspension of elements of its nuclear program—all without a permanent commitment from Iran not to acquire weapons usable materials. Under the JCPOA, Iran is essentially required to scale down its uranium enrichment infrastructure for 10 to 15 years, but is permitted to maintain the requisite technology, produce centrifuges, and expand the program again after 10 years. The deal also allows for limited research on advanced centrifuges with a gradual lifting of restrictions after 8.5 years. It’s reasonable to speculate that the deal will scale up and boost Iran’s uranium enrichment infrastructure after the restrictions end. Highly enriched uranium, unlike plutonium, does not require sophisticated weapons systems; acquisition of enough material is by itself nearly equivalent to producing a nuclear weapon.

The technological challenges in isotope separation and large-scale fertile material conversion do not form an ultimate barrier that can prevent the spread of nuclear weapons—they merely form a resistance that slows down the acquisition of nuclear weapons, allowing time for political action and response from other states. With the evolution of technology and industrialization around the world, the time needed for entry-level proliferators to acquire nuclear weapons will likely be much shorter in the 21st century than it was in the 20th.

Technology transfer in the 20th century was limited and necessitated formal collaboration between different countries. Pakistan, for instance, was able to develop nuclear weapons in the 1970s as a result of Chinese technical assistance on various nuclear technologies. India obtained nuclear weapons in the mid-1970s by exploiting information and technology transferred during the Atoms for Peace program and the Canadian-supplied CIRUS reactor. It was India’s acquisition of nuclear weapons that demonstrated how technical information and hardware transferred for civilian applications such as nuclear power could be exploited to develop nuclear weapons. With the advent of the internet, it is no longer difficult for researchers from any country to access information from the large body of academic literature on nuclear reactor physics and technology. Open-source software applications for numerical simulations are also widely available and may be used on low-end supercomputers to develop civilian reactors for peaceful applications or dual-use reactors for plutonium production (the physics involved is the same).

The supply of raw uranium has traditionally been subject to control efforts to limit the supply of raw materials to potential proliferators. Recent advances in uranium extraction from seawater have substantially increased yields to over 3 kilograms of uranium per ton of adsorbent. A kilogram of natural uranium extracted from seawater in offshore adsorbent fields today costs only twice that from a mine. While natural uranium extracted from seawater cannot directly be used to build a nuclear bomb, the IAEA has long safeguarded uranium mines to verify countries’ activities. New technology for extraction of uranium from seawater may allow for a greater diffusion of activity and make verification more difficult.

The nonproliferation regime does not rely on a single process or a single layer. But technological advances in uranium extraction, and the potential for distribution of activity rather than centralization, call for more investment in research and development of detection technology.

In the 21st century, therefore, we should expect to rely at least as much on policy as on technology to prevent the proliferation of nuclear weapons. While science and technology can inform policy, establish safeguards, and enhance the detection of clandestine or treaty-violating activity, the tendency of technological advancement is to reduce the time and resources needed to build nuclear weapons. Technological capability alone, however, does not necessarily motivate states to acquire them. There are many countries today that have the industrial capabilities to produce nuclear weapons but choose to adhere to international treaties. Limiting proliferation starts with effective policy, conflict resolution, and security assurances that reduce the demand for nuclear weapons in the first place.

Khaled Talaat is a doctoral candidate in nuclear engineering and a researcher at the University of New Mexico. He has conducted interdisciplinary research at the intersection of nuclear engineering and bioengineering.