Uncovering the Nuclear Black Market: Working Toward
Closing Gaps in the International Nonproliferation Regime

By David Albright and Corey Hinderstein
Institute for Science and International Security (ISIS)

Prepared for the Institute for Nuclear Materials Management (INMM)45th Annual Meeting
Orlando, FL

July 2, 2004

In February 2004, Dr. Abdul Qadeer Khan, considered the father of Pakistan's nuclear bomb, admitted that he had transferred sensitive nuclear technology to Libya, Iran, and North Korea. Dr. Khan's statement was a significant breakthrough for international efforts to uncover a secret network involved in illegal trading of nuclear technology. His confession opened the door for US, IAEA, and other investigators to delve far deeper into which individuals, companies, countries, and specific technologies were involved in such sales. The network of sellers, middlemen, and manufacturers is very large, and it will take time uncover and break up the whole system. It is important to expose the full extent of the network so that it cannot continue on. In addition, a complete understanding of the way in which proliferators bought and sold equipment and information is vital to assessing flaws in current nonproliferation efforts, including safeguards and export controls. Such knowledge can lead to recommendations on how to improve the nuclear nonproliferation regime.


During the 1990s, Libya set out secretly to acquire a gas centrifuge plant and all the associated equipment and materials. Libya sought a large centrifuge plant, sufficient to produce enough highly enriched uranium (HEU) to make roughly 10 nuclear weapons each year.

Libyan nuclear officials had little trouble connecting with a shady, illicit international procurement network. This network was largely invisible to the world's leading intelligence agencies and the International Atomic Energy Agency (IAEA) and well-acquainted with ways to evade controls established by the Nuclear Suppliers Group (NSG) and national export laws. This black market network had already supplied Iran with gas centrifuge components and designs. The far-flung network emerged from Pakistan's multi-decade clandestine effort to acquire nuclear weapons by illegally shopping the world for key nuclear and nuclear-related items.

This network is commonly called the "Khan" network, although this name may be misleading by focusing too much on the role of one individual, namely Abdul Qadeer Khan, the father of Pakistan's gas centrifuge program. The network was international and relatively non-hierarchical. The key technology holders and several of its leaders were in Pakistan, including Khan. But many other leaders were spread throughout the world and located in Europe, Dubai in the United Arab Emirates, Turkey, South Africa, and Malaysia. The network also depended on a variety of unwitting manufacturing companies and suppliers on many continents.

The network succeeded in operating in secret for several years before it was exposed through a series of actions by the United States, Britain, and the International Atomic Energy Agency (IAEA). During the IAEA's inspections in Iran during 2003, strong evidence emerged that Pakistani scientists and intermediaries were important clandestine suppliers of centrifuge designs and components to Iran's secret gas centrifuge program.

The secret U.S. intelligence community penetration of at least one part of the Khan network led to the dramatic seizure in October 2003 of several containers of parts bound for Libya's secret centrifuge program on the ship BBC China. After Libya renounced nuclear weapons and other weapons of mass destruction in December 2003, U.S., British, and IAEA investigations learned many details about the activities of the network. These actions resulted in intensive pressure on the Pakistani government to conduct a thorough investigation. Once started, the Pakistani investigation led relatively quickly to Khan's confession that he supplied gas centrifuge items to Iran, Libya, and North Korea. IAEA and national investigations continue to seek a full understanding of the network and the history and procurement activities of Libya's, Iran's, and North Korea's secret gas centrifuge programs.

Turn-Key Centrifuge Plant and Bomb Designs

One of the biggest surprises about this network was its sheer audacity and scale. It intended to provide Libya a turn-key gas centrifuge facility, something typically reserved for states or large corporations in industrialized nations with full government support and knowledge. The plan called for the network to supply 10,000 centrifuges, piping to connect them together, detailed project designs for the centrifuge plant, electrical and electronic equipment, uranium feed and withdrawal equipment, the initial 20 tonnes of uranium hexafluoride, equipment to allow Libya to make more uranium hexafluoride, centrifuge designs, manufacturing equipment and technology to make more centrifuges indigenously, and on-going technical assistance to help Libya overcome any obstacles in assembling and operating the centrifuges in the plant.

If Libya had continued with its nuclear ambitions and the network had not been exposed, Libya could have succeeded in about 4-5 years in assembling its centrifuge plant and operating it to produce significant amounts of HEU. If Libya managed to operate all 10,000 centrifuges, it would have had a plant with a capacity of up to 50,000 separative work units (SWU).

Armed with the HEU, Libya would have known how to turn that HEU into nuclear weapons. The network provided Libya with information to build a workable nuclear weapon. Libya received almost all of the detailed nuclear weapon component designs, component fabrication information, and assembly instructions for a nuclear weapon. The few missing designs could have been provided by the network, or Libya could have developed them on its own in the several years prior to the start of the centrifuge plant.

The weapons documents appear to be information Pakistan received from China in the early 1980's, including hand-written notes from lectures given in China. The design is reportedly for a Chinese warhead that was tested on a missile, has a mass of about 500 kilograms, and measures less than a meter in diameter. Although this design would not have fit on Libyan SCUD missiles, it could have been air-dropped or intended for a more advanced missile Libya may have sought. The design would have fit on Iranian and North Korean missiles.

As part of Libya's abandonment of its nuclear weapons program, it voluntarily gave the United States, Britain, and the IAEA centrifuge design documents, centrifuge components, and nuclear weapons documents. Libya sent the gas centrifuge items and nuclear weapons documents to the United States for safekeeping.

Although questions remain about the exact involvement of Pakistani government officials in the network and the extent of senior Pakistani government awareness of the activities of A. Q. Khan and his associates in this network, the Pakistani government was not directing this network. It was essentially a criminal operation, a more disturbing and dangerous operation than if it had been a secret government-controlled effort.

Getting at all the Nodes of the Network

As of early July 2004, much remains to be discovered about this network before its operations are fully understood or its complete demise can be celebrated. It is necessary for governments and the IAEA to be persistent in investigating this case. If these investigations are not done thoroughly, the risk will be greater that a similar network could rise again from the remnants of the disbanded Khan network.

Identifying all the Key Players and Their Activities

The network was the creation of A. Q. Khan and his associates who sought to capitalize on the elaborate, highly successful illicit procurement network they had created to supply the Pakistani gas centrifuge program beginning in the 1970s. It had many manifestations over the years and involved many people from a variety of countries. Some international suppliers and middlemen had long working relationships with Khan and remained committed to the network for two or more decades. Some of these individuals had undergone official investigations and prosecutions in past years, yet they continued working clandestinely in this network.

The latest manifestation of the network, which focused on providing Libya with gas centrifuges, was coordinated by B. S. A. Tahir, a Sri Lankan based in Dubai and Malaysia. It represented the capstone of this multi-decade illicit procurement effort centered in Pakistan.

There was also a familial aspect to the network. Europeans who were involved in the 1970s or 1980s had sons that became involved with them in the 1990s.

Many legal investigations of members of the network have started. Momentum may in fact be increasing around the world to prosecute the key players in this network. But a critical task is determining all the players and their activities. The ultimate goal should be prosecuting them as fully as possible under both export control laws and laws banning exports to terrorist states, such as Libya. This process will likely take years.

Determining the Network's Customers

Although Khan has admitted that he provided centrifuge items to Libya, North Korea, and Iran, little has been reported about other recipients of centrifuge or nuclear weapon assistance. Many details of Khan's and his associates' assistance remain unknown, particularly in regards to North Korea.

Although considerable information now exists about gas centrifuge assistance to Iran and Libya, far less information is available about the gas centrifuge assistance to North Korea. This deficiency has resulted from North Korea's denial that it has a gas centrifuge program, the lack of IAEA inspections in North Korea, and Pakistan's perceived reluctance to provide information about its nuclear dealings with North Korea.

A key question is whether Iran or North Korea also received nuclear weapon information. There is evidence of what appears to have been an early Khan effort in late 1990 to offer centrifuge help and nuclear weapons design information to Iraq. Iraq was not able to pursue this offer other than deciding to request a sample of the offer just prior to the 1991 Persian Gulf War, the start of which ended that effort.

Questions remain whether Syria was a customer for centrifuge or nuclear weapon assistance, although Khan has denied selling anything to Syria. There is also speculation about other customers, including Saudi Arabia, Burma, and al Qaeda or other terrorist organizations that were based in Afghanistan during the rule of the Taliban.

In any case, this part of the investigation is not finished. Determining all the customers and what they received will take considerable time.

Taking Stock of the Manufacturers of the P2 Centrifuge Parts

The network sold what the Pakistanis have called the P1 and P2 centrifuges, the first two centrifuges deployed in large numbers by Pakistan's gas centrifuge program. The P1 centrifuge uses an aluminum rotor, and the P2 centrifuge uses a maraging steel rotor.

The components for roughly 500 P1 centrifuges that went to Iran in the mid-90s were, according to reports, from centrifuges that Pakistan had retired from its main centrifuge program. Members of the network were able to remove them in secret and sell them to Iran.

Libya received 20 of its P1s in that manner. Libya also bought about 200 P1 centrifuges from the wider network. At least some, if not all, of the components of the additional 200 P1s were made outside Pakistan at workshops under contract with companies in the network. Aluminum rotors, for example, were made in Malaysia.

The P2 centrifuges, which are more advanced machines, reportedly left Pakistan in much smaller numbers. The two that were sent to Libya, for example, were samples or demonstration models. One of the P2s that went to Libya was not suitable for enrichment with uranium hexafluoride gas. It did not have the final surface coating necessary to prevent corrosion by uranium hexafluoride gas.

In the case of Libya, the network focused on making P2 components outside Pakistan. The Libyans have stated that they placed an order for 10,000 P2 machines. Since each centrifuge has roughly 100 different components, this order translates into a total of about one million components, a staggering number of parts given the sophistication of gas centrifuge components. The network was assembling an impressive cast of experts, companies, suppliers, and workshops to make all these components.

The workshops that contracted to make components for the network typically imported the necessary items, such as metals, equipment, or subcomponents. After they made the item, they would then send it - either assembled or as a finished centrifuge component - to Dubai under a false end-user certificate. Then it would be repackaged and sent off to Libya. According to Mohamed ElBaradei, Director General of the IAEA, "Nuclear components designed in one country could be manufactured in another, shipped through a third (which may have appeared to be a legitimate user), assembled in a fourth, and designated for eventual turn-key use in a fifth."

Based on information found in Libya, roughly half a dozen key workshops have been identified as making or doing final assembly of the centrifuge components. The network selected a workshop based on the type of centrifuge component needed and the materials and equipment involved in making those particular components. It remains unclear, however, if these are all the workshops involved in making components for Libya and other recipients.

The most well known workshop was located in Malaysia at a company called SCOPE. The parts seized on the BBC China were from SCOPE. SCOPE was also near the company that made the aluminum rotors for the 200 P1 centrifuges that Libya imported from the network.

The network contracted with SCOPE to make thousands of 14 different high precision aluminum centrifuge components for Libya's order. The contract with SCOPE involved up to about 15 percent of the total number of components sought by the network for Libya.

Workshops in Turkey made the centrifuge motor and frequency converters used to drive the motor and spin the rotor to high speeds. These workshops imported subcomponents from Europe and elsewhere, and they assembled these centrifuge items in Turkey. Under false end-user certificates, these components were shipped to Dubai for repackaging and shipment to Libya. A container of at least some of these items was also on the BBC China, but it was not discovered by investigators after its interception. Eventually, the container arrived in Libya and was declared to the United States and the IAEA.

SCOPE did not make the P2 centrifuges' maraging steel parts, which comprise the bulk of the rotating components in a centrifuge and are more difficult to make than the aluminum parts. A mystery is which workshop, if any, was contracted to make the sensitive maraging steel rotor and bellows. The network appears to have experienced trouble in finding a workshop to make these components, although one workshop may have tried unsuccessfully to make the maraging steel rotors. In the end, the network may have been unable to find a workshop to make these components, and Libya may have needed to make them itself. Libya may have initially intended to buy all the components overseas, but it may not have been able to do so in the end.

Libya also ordered from the network a sophisticated manufacturing center, code-named Workshop 1001, to make centrifuge components. The original plan called for this center to make additional centrifuges after the network delivered the first 10,000 centrifuges, either to replace broken ones or add to the total number of centrifuges. However, if the network had difficulty making a component for the original 10,000 machines, this center may have had to make that particular component.

Most of the machine tools, furnaces, and other equipment for the center came from Europe, particularly from or through Spain and Italy. The equipment was not on the nuclear dual-use list, but it was still adequate for use in a centrifuge manufacturing program, particularly because the network also supplied detailed manufacturing information for almost all the parts. The bulk, if not all, of this equipment was sent to Libya via Dubai.

Investigations of the supply chain of the network are unfinished. It is not known if all the key workshops and companies have been identified. Components may have been made but not delivered to Libya. Components may have also been made and delivered to customers other than Libya.

Retrieving Centrifuge Designs and Manufacturing Instructions from the Network

The key to the success of this network was its virtual library of centrifuge designs and detailed manufacturing manuals. A key task is to track down the members of the network with this kind of sensitive centrifuge information, prosecute them, and try to retrieve as much of this information as possible.

Although tracking down this kind of information can be difficult, the network may have helped by tightly controlling this information and not widely disseminating it. For example, the February 2004 Malaysian police report describes the activities of the Swiss national Urs Tinner at the SCOPE factory. Urs is the son of Friedrich Tinner, who earlier supplied Pakistan's and Iraq's secret centrifuge program in the 1980's. Urs carefully controlled the centrifuge design information. He took it to SCOPE to program the machines to make the centrifuge components, but he made sure that others could not remove this sensitive information. When he finished working at SCOPE, he removed the information and took it with him.

The fact that the information may be in the hands of just a few members of the network gives hope that it can be retrieved. However, retrieving all the centrifuge information may not be possible, since copies can be made and hidden for years if desired. Thus, even if the retrieval effort is reasonably successful, this centrifuge information may form the core of a future network aimed at secretly producing or selling gas centrifuges.

Findings and Recommendations

A priority is preventing illicit networks like the Khan network and other nuclear smuggling using less elaborate methods. In addition, steps are needed to increase the probability of discovering undeclared nuclear efforts before they are as advanced as the programs in Libya and Iran.

The first priority is fully investigating and dismantling the Khan network. Investigations need to continue and intensify in a range of states, including Malaysia, Switzerland, Britain, France, Italy, Spain, UAE, Germany, South Africa, and Turkey. More information is needed from states that benefited from this network, particularly Libya, Iran, and eventually North Korea. In addition, Pakistan's cooperation is critical. The Pakistani government has provided useful information to the IAEA and other governments, and it appears committed to providing more information. However, the Pakistani government should permit the IAEA, and perhaps other governments, direct access to A. Q. Khan and his associates involved in the network.

The successes of the Khan network should shatter any complacency about the effectiveness of national and international nuclear-related export controls to stop or sound an alarm about illegal nuclear or nuclear-related exports. The network was masterful in identifying countries that had inadequate national export laws yet adequate industrial capability for the network's purposes. These countries were both inside and outside the NSG. Although many suppliers to the network did not know the actual purpose of the materials they provided or the parts they were contracted to make, they were often in countries where the authorities were unlikely to carefully scrutinize exports or encourage curiosity about the actual end use of an item. The companies themselves had little motivation -- either from conscience or threat of punishment -- to confirm the explanations they were given. The network also knew how to obtain for its illicit endeavors necessary subcomponents, materials, machine tools, and other manufacturing equipment from countries in Europe with stringent export control systems.

NSG efforts to stop illegal exports should be improved. The NSG should be expanded in membership to include states that, while not normally considered actual or potential suppliers of nuclear items, have advanced industrial and manufacturing infrastructures that can be exploited for the production of direct-use nuclear items, such as centrifuge components. The network operated in countries outside the NSG such as Malaysia that had little knowledge of nuclear technology or effective export controls but had high-tech industries eager to make direct-use nuclear items apparently oblivious or indifferent to the actual nature of the item. By expanding NSG membership, more countries will improve their export control systems and receive help from more experienced members that will reduce the chance of cases like SCOPE happening again.

In addition, NSG members should share more information about actual procurements among themselves and with the IAEA. Currently, NSG members share any denials of an export but not the approvals of key items. The IAEA does not even routinely receive the denial information. NSG members should share key approvals, and the IAEA should receive both the denials and these key approvals. Sharing of this information would help members of the NSG and the IAEA develop a better picture of a potential proliferant's overall nuclear capability, potentially leading to significantly earlier warnings of undeclared nuclear activities.

Two recent initiatives aim to correct shortcomings in the current national and international export control systems. However, they do not go far enough.

The Proliferation Security Initiative (PSI), which is credited with the seizure of the centrifuge components on the BBC China, usefully complements existing export control systems. PSI has improved the chance of interdicting WMD, ballistic missiles, and the means to produce them while such items are in transit to and from states and non-state actors of proliferation concern. However, it cannot alone fix the fundamental weaknesses of the current export control system exposed by the discovery of the Khan network.

The April 2004 UN Security Council Resolution 1540 fills a serious gap in existing WMD non-proliferation regimes that target only states and omit non-state actors. It is expected to make it harder for nuclear smuggling to occur. An immediate improvement is that all 191 UN member states must establish, review, and maintain appropriate effective nuclear and nuclear-related export controls. By applying to almost all states, this resolution helps fix some of the problems posed by the voluntary, limited membership of the NSG. However, the resolution has several problems. Some states may resist key provisions of this resolution, believing that the law should have been established through treaty negotiations. The law will likely be applied unevenly among states. Without extensive assistance, many key states will experience difficulties in enacting effective export control legislation, implementing it, and enforcing it.

Missing in the current system is an aggressive, intrusive verification and investigation organization that can provide greater confidence that states are implementing effective export controls, flag deficiencies in the export control practices, and detect illicit nuclear and nuclear-related procurements.

What is needed is a universal treaty-based system controlling nuclear export activities that is binding on states and includes a means to verify their compliance. Under such a treaty, countries would implement a set of nuclear and nuclear-related export control laws and criminalization procedures, similar in nature to those required by UNSC Resolution 1540. The agreement, however, would also mandate an organization to verify compliance, ensure the adequacy of those laws, and investigate illicit procurement activities. Signatories would inform this organization of all sensitive nuclear or nuclear-related exports, and it would have the mandate and legal rights to verify that the transactions are indeed legal. The organization would verify that a country's declaration about its nuclear or nuclear-related exports or imports is accurate and complete.

The IAEA is a logical choice to undertake this role. It is already pursuing investigations of illicit procurement activities by Iran and Libya as part of its safeguards responsibilities under the Nuclear Non-Proliferation Treaty (NPT). These investigations include taking inventories of all centrifuge equipment and components in Iran and Libya and verifying their accuracy and completeness, determining the suppliers and manufacturing activities of the network, cooperating with a range of governments on the activities of the network, receiving supplier information from member states, and meeting with Pakistani investigators. These investigations involve receiving more information from states than required by the Additional Protocol.

Libya, Iran, Pakistan, and other states have provided the IAEA specific information about nuclear and nuclear-related procurements. Armed with this information, the IAEA has more effectively conducted investigations into the Khan network and the procurement activities of Iran and Libya that are directly related to its safeguards responsibilities. In particular, the IAEA is in a much stronger position to make a determination about Iran's and Libya's compliance with the NPT and the steps necessary to develop confidence about the absence of undeclared nuclear activities or materials in these two countries.

Expanding the IAEA's mandate by requiring states to report on a wider variety of exports would be a logical extension of current safeguards which place great emphasis on developing a broader picture of a state's nuclear and nuclear capable infrastructure. This new mandate should be seen as part of implementing credible safeguards.

The Additional Protocol does require reporting of exports of direct-use nuclear items, but in practice nuclear smuggling networks would export those items illegally. States would often not know the true nature of those items and thus could not report them. Dual-use items with legal export licenses that often would be more likely on a routine basis to reveal undeclared activities are not reported to the IAEA under the Protocol.

In addition, a treaty-based system of export controls and verification would impose new requirements on all states, even those that have not implemented the Protocol. In any case, state declarations submitted under the protocol or as part of a new treaty regime should be broadened to include the export of dual-use items.

By more formally linking its safeguards system with export control verification and monitoring, the IAEA would be in far better position to assure the absence of undeclared nuclear activities and detect cheating in a timely manner. By performing a task that governments have been unable to do, the IAEA under such a treaty-based system would significantly increase U.S. and international security.