UCRL-ID-142036 CGSR-2001-001

Verifying the Agreed Framework

April 2001 The Center for Global Security Research The Center for International Security (CGSR), established in 1996, brings together and Cooperation (CISAC) is a multidiscipli- diverse expert communities to address com- nary community dedicated to research and mon challenges with significant policy implica- training in issues of international security. The tions. The Center studies how technology can Center, part of the Institute for International enhance national security and expand knowl- Studies (IIS) at Stanford University, promotes edge of the policy–technology interface. The a creative and collaborative environment in CGSR draws strength from its affiliation with which scholars, scientists, government offi- the Lawrence Livermore National Laboratory cials, military officers, and business leaders (LLNL), a Department of Energy national lab- can produce policy-relevant research to oratory managed by the University of address complex international problems. California. By supporting research that explic- itly considers both technical and policy factors Center for International Security and Cooperation involved in defense programs, arms control, Stanford University nonproliferation, peacekeeping, and related Encina Hall international issues, the Center contributes to Stanford, CA 94305-6165 national and international security. Phone: 650.723.9625 Fax: 650.723.0089 Center for Global Security Research Web: http://cisac.stanford.edu Lawrence Livermore National Laboratory P.O. Box 808 (L-189) Livermore, CA 94551 Phone: 925.422.6141 Fax: 925.422.5252 Web: http://cgsr.llnl.gov UCRL-ID-142036 CGSR-2001-001

Verifying the Agreed Framework

April 2001

Michael May, General Editor Chaim Braun George Bunn Zachary Davis James Hassberger Ronald Lehman Wayne Ruhter William Sailor Robert Schock Nancy Suski GLOSSARY

AF neutrons. The most common of these Low-enriched uranium (LEU) Agreed Framework, signed in 1994 materials are uranium-235 (235U) and uranium containing more than 0.71 between the US and the DPRK, to plutonium-239 (239Pu). Uranium-233 percent and less than 20 percent urani- negotiate nuclear issues on the Korean (233U) is also fissile. um-235 (235U). Most modern light- peninsula water power reactors use 3–5 percent Fuel-fabrication facility/plant LEU. LEU is insufficiently enriched in Agreement for Cooperation facility where nuclear materials (e.g., 235U to be used for nuclear explosives. what US law requires the US govern- enriched or natural uranium) are fabri- ment to negotiate with countries cated into fuel elements to be inserted Magnox (including the DPRK) before US into a reactor uranium nuclear fuel with magnesium- nuclear materials or US nuclear com- alloy cladding. Because this cladding ponents can be provided for reactors Fuel-grade plutonium corrodes easily, this type of fuel is diffi- for those countries, including the ROK plutonium containing less than 80 per- cult to store safely for a long period of reactors which KEDO plans to send to cent plutonium-239 (239Pu) and 7 to 18 time or dispose of in a geologic reposi- DPRK pursuant to the AF percent plutonium-240 (240Pu) tory. Typically, irradiated magnox fuel is reprocessed shortly after it is BOL Gas-graphite reactor removed from the reactor core. Beginning of Life, generally used with a nuclear reactor cooled by a gas and BOL fuel, the first cycle in a reactor moderated by graphite MW(e) Megawatt-electric, used in reference to Burnup Highly enriched uranium (HEU) a nuclear power plant, equals one mil- a measure of the thermal energy pro- uranium in which the percentage of lion watts of electricity duced per mass of fuel (usually meas- uranium-235 (235U) is raised ured in megawatts-thermal-days per (“enriched”) from a natural level of MW(th) tonne [MWth-d/t]) 0.71 percent to greater than 20 percent. Megawatt-thermal, one million watts All HEU can be used to make nuclear of heat Control rods explosives, although a very large rods of neutron-absorbing material quantity is required for HEU enriched Natural uranium inserted into the core of a reactor to to only 20 percent. uranium containing 0.71 percent urani- control its operations. Pushing in the um-235 (235U) rods decreases the rate of reaction, IAEA while removing the rods increases the International Atomic Energy Agency NPT rate of reaction. Non-Proliferation Treaty INFCIRC Core Information Circular, a series of IAEA Nuclear fuel cycle the central part of a nuclear reactor documents regarding safeguards, etc. the entire life cycle of nuclear fuel, containing the fuel rods, moderator, including the front end (initial mining, and control rods, where the nuclei of IRT reactor milling, conversion, enrichment and the fuel fission and release energy Soviet-designed research reactor fuel fabrication), reactor irradiation fueled with enriched uranium and (and resulting power generation), and DPRK moderated with water the back end (including spent-fuel Democratic People’s Republic of Korea, storage, reprocessing and recycling, commonly called North Korea KEDO and disposal) Korean Peninsula Energy Enrichment Development Organization ROK the process of increasing the concentra- Republic of Korea, commonly called tion of one isotope of a given element KSNP South Korea (in the case of uranium, increasing the Korea Standard Nuclear Plant concentration of uranium-235). Also, the Safeguards Agreement An agreement resulting concentration of that isotope. Kumho site of a country having nuclear activities location where the KEDO-supplied with the IAEA providing for safe- EOL reactors are being built in North Korea guards (nuclear material accounting End of Life, generally used with EOL and control plus periodic inspections) fuel, the third cycle in a reactor Light-water reactor (LWR) to assure that nuclear material is not a reactor that uses ordinary water as a diverted to nuclear explosives Fissile material moderator and coolant and low- material composed of atoms that fission enriched uranium as fuel when irradiated by slow (“thermal”) CONTENTS

Introduction: Verification and the Challenges to Constructive Engagement with North Korea . 1 Technical Questions, Strategic Implications ...... 1 Nuclear Cooperation: Two Sides of a Coin ...... 2 Means Versus Ends ...... 3 Program Management, IAEA Interaction, and Challenges to Preventive Diplomacy ...... 3 Dynamics of Technology–Policy Interaction ...... 4 By What Standard Shall We Judge, and When? ...... 5 Verifying the Agreed Framework: Executive Summary ...... 7 Safeguarding the Nuclear-Power Reactors Provided by KEDO ...... 7 Verification of the DPRK’s Declaration and of the Disposal and Dismantlement of Identified or Suspect Nuclear Facilities ...... 8 Verification of DPRK’s Declaration ...... 8 Verification of the Disposal and Dismantlement of Identified Yongbyon Facilities ...... 9 Verification Regarding Other Suspect Facilities ...... 10 Possible Adverse Developments ...... 10 Further Delays ...... 10 Non-Cooperation with the IAEA’s Verification of Compliance ...... 10 Need for an Amended DPRK Declaration ...... 11 Disagreements Over Material To Be Transferred from the DPRK ...... 11 Disagreements Over the Site of Ultimate Disposition ...... 12 Disagreements Over the Extent of Safeguards for KEDO Reactors ...... 12 Interference with Safeguards for KEDO Reactors ...... 12 Abrogation of AF or NPT after the KEDO Reactors Are Installed ...... 12 Executive Summary Conclusions ...... 13 Organization of Report ...... 13 Authors and Sponsors ...... 13 Chapter 1 A Brief History of the DPRK’s Nuclear Weapons-Related Efforts ...... 15 1.1 Early History ...... 15 1.2 Attempts To Restrain the DPRK From Making Nuclear Weapons ...... 16 1.3 Agreed Framework of October 1994 ...... 17 Notes to Chapter 1 ...... 20 Chapter 2 Currently Applicable Safeguards and Related Agreements ...... 23 2.1 The Existing IAEA–DPRK Safeguards Agreement ...... 23 2.2 New Information That the IAEA May Ask of All States Under INFCIRC 153 Safeguards . . . . 24 2.3 Additions to INFCIRC 153 Safeguards for States Willing To Agree to a New INCFIRC 540 Safeguards Protocol ...... 24 2.3.1 Safeguards on Existing ROK LWRs—Models for the DPRK LWRs ...... 25 2.4 Major Challenges Ahead to the Implementation of the Agreed Framework ...... 26 2.4.1 Completion of a “Significant Portion” of the First Reactor in the ROK and Reactor Buildings in the DPRK ...... 26 2.4.2 Financing ...... 26 2.4.3 Nuclear Liability ...... 26 2.4.4 US–DPRK “Agreement for Cooperation” and IAEA Inspection of Undeclared Facilities . . 26 2.4.5 Improving North Korea’s Electricity Distribution System ...... 27 Notes to Chapter 2 ...... 27 Chapter 3 The KEDO Reactors and Associated Facilities and Activities ...... 29 3.1 Introduction ...... 29 3.2 The KEDO Site ...... 29 3.3 Nuclear-Power Development Along the Shores of the East Sea ...... 30 3.4 Site Work to Date ...... 31 3.5 Nuclear-Fuel Shipments Into and Out of the Site ...... 33 3.6 Electricity Transmission Issues ...... 35 3.7 The ROK’s Energy Development Issues ...... 37 Notes to Chapter 3 ...... 39

iii Chapter 4 Safeguards on the KEDO Reactors ...... 41 4.1 Introduction ...... 41 4.2 Project Organization To Supply the KEDO Reactors ...... 41 4.3 Description of the KEDO Reactors ...... 42 4.4 Refueling Operation in the KEDO Reactors ...... 45 4.4.1 The Special Problem of the Beginning-of-Life and End-of-Life Fuel Discharges ...... 49 4.5 IAEA Safeguarding of Nuclear Fuel in the KEDO Reactors ...... 50 4.5.1 The Safeguards Goals for Quantity and Timeliness ...... 50 4.5.2 The Safeguards Accounting Process ...... 51 4.5.3 Material Balance Area in KEDO-type Reactors ...... 51 4.5.4 The Containment and Surveillance Process ...... 52 4.5.5 Safeguards Inspections During Routine Plant Operation ...... 53 4.5.6 The Physical Inventory Verification Type of Safeguards Inspection ...... 54 4.6 Assessment of Additional Measurements and Inspections ...... 55 4.6.1 Measures for Strengthened Safeguards Under INFCIRC 153: Environmental Sampling . . . . 55 4.6.2 Measures for Strengthened Safeguards Under INFCIRC 153: Remote Monitoring ...... 55 4.6.3 Measures for Strengthened Safeguards Under INFCIRC 153: Other Recommended Steps . 56 4.6.4 Satellite Remote Sensing ...... 56 4.7 Conclusions ...... 57 Notes to Chapter 4 ...... 58 Chapter 5 Diversion and Misuse Scenarios for Light-Water Reactors ...... 59 5.1 Scope and Intent ...... 59 5.2 Scenarios ...... 59 5.2.1 Scenario: Covert Diversion of Spent Fuel Produced During Normal Reactor Operations . 60 5.2.2 Scenario: Covert Materials Production ...... 61 5.2.3 Scenario: Short-Cycling the Reactor ...... 62 5.2.4 Scenario: Overt Reconfiguration of the Reactor for Materials Production ...... 63 5.3 Consequences ...... 64 5.4 Preventive Measures ...... 65 Notes to Chapter 5 ...... 66 Chapter 6 Known and Suspected Nuclear-Related Facilities in the DPRK ...... 67 6.1 Inspection, Dismantlement, and Disposal Requirements for Existing Nuclear Facilities . . . . . 67 6.1.1 “Full-Compliance” Inspections by the IAEA ...... 67 6.1.2 Disposal of Spent Fuel ...... 67 6.1.3 Dismantlement of Gas-Graphite Reactors ...... 67 6.2 The Facilities at Yongbyon ...... 67 6.2.1 The IRT-2000 Research Reactor ...... 68 6.2.2 Isotope Production Laboratory Near the IRT-2000 Reactor ...... 68 6.2.3 Probable Undeclared Waste Site South of the IRT-2000 Reactor ...... 68 6.2.4 Declared Waste Site ...... 68 6.2.5 Graphite-Moderated Reactor ...... 68 6.2.6 Fuel Reprocessing Facility ...... 69 6.2.7 Undeclared Waste Storage Building ...... 70 6.2.8 Other Unfinished Reactors ...... 70 6.3 Verification of the Initial DPRK Declaration ...... 70 6.4 Verification of Fuel Disposal and Facility Dismantlement ...... 74 6.4.1 Spent-Fuel Disposition ...... 74 6.4.2 Dismantlement ...... 74 6.5 How Much Verification Is Enough? ...... 76 Notes to Chapter 6 ...... 78 Chapter 7 Timeline for Verification and Safeguards ...... 79 Completion of a Significant Portion of the Project Site at Kumho, DPRK, and of the First KEDO Reactor in the ROK ...... 79 IAEA Declaration that the DPRK Is in Compliance with Its Safeguards Agreement ...... 79 Start of Delivery of Key Nuclear Components of the First KEDO Reactor to the Kumho Site Simultaneously with Start of Transfer of DPRK’s Spent Fuel from Yongbyon to Its Ultimate Disposition ...... 80 iv Simultaneous Completion of Yongbyon Spent-Fuel Transfer and of the First KEDO Reactor . . . . 80 Dismantlement of DPRK’s Graphite-Moderated Reactors and Related Facilities Begins ...... 81 Deliveries of the Nuclear Components for the Second KEDO Reactor in Parallel with Proportional Steps by the DPRK To Dismantle All Its Graphite Reactors ...... 81 Completion of Second KEDO Reactor at Kumho ...... 81 Shipping Spent KEDO Reactor Fuel When Appropriate Out of the DPRK ...... 81 Chapter 8 Potential Adverse Developments ...... 83 8.1 Benchmark Scenario: The DPRK Fully Cooperates with the IAEA Regarding Existing Agreements ...... 83 8.2 Scenario 1: The DPRK Maintains Its Present Level of Cooperation ...... 84 8.3 Scenario 2: The DPRK Attempts Covertly To Divert or Hide Nuclear Material ...... 85 8.4 Scenario 3: The DPRK Abrogates the Agreed Framework or Other Key Agreement ...... 85 8.5 Scenario 4: The US or the ROK Is Unable or Unwilling To Meet Its Commitments under the Agreed Framework Even Though the DPRK Does ...... 87 8.6 Predicting the Future ...... 87 Notes to Chapter 8 ...... 88 Verifying the Agreed Framework: Conclusions and Recommendations ...... 89 Safeguarding the KEDO Reactors ...... 89 Past DPRK Nuclear Activities ...... 89 Possible Adverse Developments ...... 90

FIGURES

Introduction Figure. Parallel interaction paths related to the Agreed Framework (AF) and the Democratic People’s Republic of Korea (DPRK) ...... 6 Figure 1-1. General map of the Korean peninsula ...... 15 Figure 3-1. General map of the DPRK ...... 29 Figure 3-2. Locations of nuclear reactors around the East Sea ...... 30 Figure 3-3. The vicinity of the Light-Water Reactor (LWR) project ...... 32 Figure 3-4. Artist’s conception of the future site for the KEDO reactors ...... 32 Figure 3-5. Scope of the grading efforts in the construction area...... 33 Figure 4-1. Organization chart for the supply of a Korean Standard Nuclear Plant (KSNP) reactor . . 41 Figure 4-2. Side view of a Korean Standard Nuclear Plant (KSNP) building...... 42 Figure 4-3. A top view of the Korean Standard Nuclear Plant (KSNP) reactor ...... 43 Figure 4-4. Schematic of the primary system layout in the containment building ...... 44 Figure 4-5. Internal layout of the Korean Standard Nuclear Plant (KSNP) reactor vessel ...... 45 Figure 4-6. Schematic of a fuel assembly ...... 47 Figure 4-7. Fuel transfer process ...... 48 Figure 4-8. The relationships between plutonium production, quality, and burnup ...... 48 Figure 4-9. Light-Water Reactor (LWR) layout showing the containment and surveillance process . . 52 Figure 4-10. Advanced camera–seal system implemented in the Republic of Korea (ROK) reactors. . 53 Figure 4-11.Additional monitoring measures that could augment existing safeguards activities at the KEDO reactors...... 57 Figure 5-1. Plutonium discharged from LWR operations ...... 64 Figure 6-1. Satellite image of the Yongbyon site in 2000 ...... 67 Figure 6-2. The buried and the new waste site near the uncompleted 50-MW(e) graphite reactor. . . 69 Figure 6-3. Satellite image showing the plutonium separation building and associated buildings and the suspect waste building...... 70 Figure 6-5. The measurements at the radiochemical lab ...... 72 Figure 6-4. Nuclear program at 5-MW(e) graphite reactor and radiochemical lab...... 72 Figure 6-6. The suspected route used to produce plutonium with the IRT-2000 research reactor . 73

v TABLES

Executive Summary Table. Envisaged time sequence of events ...... 11 Table 3-1. Status of nuclear power plants in South Korea...... 38 Table 4-1. Characteristics of the Korean Standard Nuclear Plant (KSNP) ...... 46 Table 4-2. Commercial remote-sensing platforms...... 57 Table 5-1. Accumulation of Light-Water Reactor (LWR) plutonium in spent fuel...... 64 Table 6-1. Visits and inspections by the IAEA of North Korean facilities ...... 71 Table 7-1. Summary of steps bearing on verification ...... 82

SIDEBARS

Spent Fuel and Plutonium ...... 48 Significant Quantities ...... 50 Plutonium and Weapons ...... 59

vi INTRODUCTION: VERIFICATION AND THE CHALLENGES TO CONSTRUCTIVE ENGAGEMENT WITH NORTH KOREA

RONALD F. L EHMAN,II

The Agreed Framework (AF) Technical Questions, Development Organization between the United States of Strategic Implications (KEDO). Procurement and con- America and the Democratic struction by KEDO of these reac- People’s Republic of Korea In this report, we examine sever- tors compels the major parties— (DPRK), signed in Geneva on al issues related to that cooperative the US, Japan, the Republic of October 21, 1994, has become the nuclear project, especially monitor- Korea (ROK), and the DPRK—to centerpiece of recent US efforts to ing and verification of nuclear- find ways to cooperate. Although reduce the threat of conflict with material production in the DPRK. each delay or setback has resulted North Korea. In particular, it seeks The safeguarding of the light-water in more public questioning of the to bring the DPRK into compliance reactors (LWRs) to be supplied to LWR procurement, as each step is with its obligations under the the DPRK is given particularly taken, commitment to the program nuclear Non-Proliferation Treaty careful scrutiny. In focusing on has grown stronger among those (NPT) not to acquire nuclear nuclear cooperation with North economic and political entities weapons. The AF document sets Korea, we are vividly aware that within the party states who see goals, outlines programs, initiates a success or failure in meeting the their stakes in the effort increase. US-led nuclear-power consortium, technical challenges posed by that and notes linkages. The AF refers project can have a wider impact. Provision of the LWR reactors, to a wider range of diplomatic and Compressed schedules, cultural however, is contingent upon international security initiatives, differences, and a limited history of Pyongyang resolving with the such as the NPT and the agreement cooperation with Pyongyang are International Atomic Energy on denuclearization of the Korean among the internal factors that Agency (IAEA) existing concerns Peninsula, and is meant to rein- could result in a failure to meet that the DPRK is developing force others, including those relat- program deadlines or objectives. nuclear weapons. The triggers ed to the reconciliation of the two External factors that could derail leading to the confrontation in Koreas. the process are perhaps even more early 1993 were apparent discrep- numerous and involve major con- ancies in Pyongyang’s initial The effect of each step taken or flicting interests. Thus, in support nuclear declaration to the IAEA. not taken under the AF will have a of our technical analysis of verifica- At a minimum, these suspect dis- significant impact broadly on tion issues, we have highlighted a crepancies must be resolved for North–South Korean relations, eco- number of scenarios involving reactor construction to proceed to nomic and humanitarian interac- delay or disputes that could come the installation of certain critical tion, East Asian security, US into play because of developments nuclear components, as agreed in national security and foreign poli- both within and outside the reactor the framework. cy, Western alliances, and nonpro- project. Such scenarios must influ- liferation in the region and around ence how we think about verifica- Even if a mutually acceptable the globe. The larger message of tion. Each, however, also has over- agreement on a declaration is the agreement is clear—as North arching strategic implications. reached with the IAEA, further Korea complies with international progress toward normalization of norms, relations with the outside We recognize the primacy of relations between the DPRK and world increasingly will be normal- these broader strategic considera- most of the rest of the world is, in ized, informally and formally. Aid tions even if our interest here is the reality, inevitably linked to greater and trade are to grow with the heart of the AF—its procurement confidence that North Korea has reduction of military threats and for the DPRK of two LWRs and the abandoned the rest of its nuclear- the expansion of political dialog linked, verified cessation and dis- weapons program. The two LWRs and contacts. As its engine, the AF mantlement of the DPRK’s nuclear are a major part of that confidence- sets in motion a remarkable joint weapons program. The new reac- building, but only part. Mere nuclear-energy project that is a cen- tors are being provided by an completion of the reactors cannot tral focus here. organization created for that pur- ultimately be the standard by pose, the Korean Peninsula Energy which success is judged. Still, the

1 INTRODUCTION nuclear-reactor project serves as a Despite some uncertain steps, how- powerful impact on other nations. major pacemaker and bellwether of ever, the US and its allies thus far This includes those seeking the status of the AF and its non- have traveled in the same general weapons, such as Iraq and Iran, proliferation objectives. direction, retaining their cohesion and those such as Japan, South in the face of both euphoria and Korea, and Taiwan who have thus Nuclear Cooperation: Two disappointment. far been willing to forgo nuclear Sides of a Coin weapons they could easily manu- The AF is mainly about facture. The precedents set with Inherently, the AF seeks to man- Pyongyang’s normalization with North Korea under the AF also age relations between adversaries the West. Russia and China may affect the standards applied by trying to find their way to safer share some of the West’s concern other nations in their nuclear trade ground in a changing world. The about proliferation on the Korean and technology transfer. US has been motivated in creating Peninsula, but they are likely more the AF as much by the risks of fail- ambivalent about the overall thrust Through the AF, we seek to use ing to act as by the prospects for of the AF. Nevertheless, successful nuclear and energy cooperation success. The AF has become our implementation of the AF may with North Korea to strengthen the most ambitious laboratory for benefit from engagement with global nonproliferation regime, yet defining and assessing “construc- China and, to a lessor extent, not every important actor has cho- tive engagement.” As the recent Russia and the nations of Western sen to interpret it that way. Mos- restatement of the compromise by Europe, as well as others. In some cow has already cited the AF as a former Secretary of Defense situations, carrots such as trade rationale for Russian nuclear-reac- William Perry and the related and recognition are involved. In tor deals with Iran, and New Delhi negotiations on North Korean bal- other instances, sticks such as suggests it is evidence of a double- listic missiles suggest, the AF is United Nations Security Council standard nuclear suppliers have subject to “re-invention.” Indeed, debate and action have been applied against India. Indeed, the AF itself is a reconfiguration of involved. These interactions also Indian hawks have asserted that the earlier agreements it references reflect the uncertainties of a world the respect given to North Korea on North–South reconciliation and in transition. What is a common because of its nuclear-weapons denuclearization of the Korean cause on one day becomes a source program further validates India’s peninsula. This variability in of tension on another. Not all of the own weapons program. Whether implementing mechanisms is issues involved are grave, but real or contrived, such assertions inevitable in any framework built some of the political concerns and remind us that our diplomacy in on a process in which a regime, security calculations of the key Northeast Asia is watched closely such as North Korea, frequently nations involved are serious, even elsewhere. tests whether it can gain more vital. Clearly, the outcome of the from delay, brinkmanship, and Korean engagement matters great- The game of carrot and stick even threats of war than from the ly to Chinese and Russian strategic with North Korea is meant to rein- reduction of tensions and expan- assessments of security in force international norms. It is built sion of cooperation. Northeast Asia, just as it does to upon the notion that abandoning the US and its allies in Asia. nuclear-weapons programs brings The AF also brings together Success in achieving reductions in tangible benefits, but this engage- allies balancing their collective and tensions on the Korean Peninsula ment could also suggest to those separate interests. For the United could aid significantly in the evolu- already receptive that nuclear States, Japan, and the ROK, the AF tion of relations with Russia and blackmail might pay. Success in tests their ability to work together China. Unfortunately, North Korea bringing about compliance with successfully under post-Cold War also has the potential to be a tragic the NPT in the near-term, however, conditions to reshape the Cold War spoiler in what had been substan- will still leave us with the classic legacy of a divided Korea. It tial reductions in the adversarial long-term question that can be remains a test they could still fail, psychology of the great military applied to many parties to the and the price of failure could well and economic powers. NPT: namely, will easier access to be fissures between friends as well technology under the NPT now as dangerous new confrontations Success or failure in stopping the facilitate future nuclear-weapons with Pyongyang. Our allies weigh North Korean nuclear-weapons efforts? Thus, the value of the AF constantly whether we are too program—one of the most difficult must be judged ultimately by the forceful or too restrained, a debate remaining challenges to an almost net contributions it makes to inter- echoed in the domestic debate in universal commitment to nuclear national security in the region and many capitals, including our own. nonproliferation—can have a around the world. The full

2 INTRODUCTION measure of its merit is to be found near-term danger posed by the and Nuclear Cooperation in that broader policy context. existing North Korean nuclear- Agreement are ahead. For example, weapons program would greatly under US law, a nuclear coopera- Means Versus Ends undermine the immediate nonpro- tion agreement must be reached, liferation objectives embodied in but North Korean noncompliance That the AF is bigger than just the AF, especially compliance with with the NPT presents legal and the freezing of the DPRK’s nuclear- the NPT. Here we build upon the political obstacles. The “when” and weapons program and the creation current approach. In undertaking “what” of the US nuclear coopera- of KEDO, with the provision of this study of the DPRK’s nuclear tion agreement is thus complicated two new safeguarded reactors to program, however, we are acutely by the hovering question of North Korea, deserves special men- aware that verification of the AF is “how?” Key business and budget tion up front. The bigger picture a means and not an end. strategies must await agreement on will also deserve further examina- liability, specific arrangements for tion after our analysis of the In this study of the interaction of fresh-fuel supply and spent-fuel prospects for verification of technology with policy—in this disposition, clarification as to par- nuclear-materials production and case, DPRK’s compliance with the ticipation in day-to-day reactor NPT compliance has been present- NPT—we reference the history of operations, availability of an inte- ed. Admittedly, this study has US interactions with North Korea grated infrastructure such as a dis- homed in on the prospects for veri- on the nuclear question. We do not tribution grid in the DPRK, com- fication of the dismantlement of intend here to re-open the debate pletion of more comprehensive the DPRK’s nuclear-weapons pro- over the wisdom of the AF or the training, etc. Thus, important ques- gram, especially the dismantlement process by which it was achieved. tions of “who” and “how much” of the reactors and reprocessing We begin our substantive analysis also remain. These implementation facilities at Yongbyon and the with the AF as it now stands. Our uncertainties all affect the confi- implementation of verifiable safe- objective is enhancing the prospects dence in verification and on the guards on the KEDO nuclear- that it could achieve its goals. We calculations of the risks and bene- reactor program. The core of our examine technical and program- fits of the AF, especially to the study is thus purposefully narrow, matic hurdles to be overcome in the degree that timing is viewed as a but understanding the implications implementation of the current critical factor. of each step or misstep from a IAEA and KEDO programs, and we broader perspective is even more explore means to ensure that verifi- Verification of initial declarations important. Completion of the cation milestones and standards and implementation of safeguards KEDO reactors would be a hollow can be met. We also seek to illumi- by the IAEA over the necessary achievement if nonproliferation nate the wider implications of suc- DPRK nuclear infrastructure create goals were not achieved. cess or failure at various stages. a number of “make or break” mile- Elimination of the Yongbyon facili- This inevitably raises questions stones. Delay in reaching these ties would be inadequate if the about alternative tactics and even moments of truth has not made DPRK finds another path to the exit strategies. If the implementa- them any easier. Indeed, the pas- procurement of fissile material or tion of the AF is delayed or sage of time may make some issues weapons. On the other hand, fail- derailed, will we still be able to more difficult to resolve. ure to complete the reactors might achieve our goals? And by what Recognition that acceptable confir- not be a waste if our international means? In some cases, our analysis mation of initial or even amended security objectives are otherwise may suggest the need for advance declarations may be problematical accomplished. consideration of options beyond has already reopened debate over the scope of this discussion. minimum acceptable requirements. A number of paths may reach Even the question of whether the our goal. For example, the provision Program Management, safeguards developed post-Iraq of non-nuclear electric power IAEA Interaction, and should be applied remains open plants may be cheaper, faster, and and under discussion in some more conducive to long-term non- Challenges to Preventive circles. proliferation objectives. Additional Diplomacy non-nuclear power plants undoubt- This question of the adequacy of edly will be necessary if normaliza- Implementing even the KEDO safeguards is but a subset of the tion brings with it extensive eco- portion of the AF is already behind larger question of whether mere nomic growth. On the other hand, schedule. Difficult program man- verification that material has not the provision of electricity by any agement, business, safety, and legal been diverted from indigenous means that does not address the decisions pertaining to the LWR reactors and facilities is sufficient.

3 INTRODUCTION

Certainly, the real nuclear threat Uncertainty about the DPRK’s with this problem. Such approach- environment is larger. Although it capability and intentions was great- es broaden the arena of engage- is true that the IAEA’s request for ly increased in the face of the ment by expanding the linkage of special inspections, which resulted provocative testing of long-range progress on verification and securi- in a confrontation in 1993, was missiles, including one launch over ty with progress on political nor- related to efforts to resolve discrep- Japan of a missile Pyongyang sub- malization and economic benefits. ancies in the DPRK’s declaration, sequently declared to be a space Such linkage is explicit in the AF the rejection called into question launch vehicle. Whatever it says just as it was explicit in the earlier the ability to verify that no broader about intentions, DPRK’s willing- efforts it refers to such as the 1991 nuclear-weapons program is ness to sell ballistic missiles to North–South Joint Declaration on underway by other means. In con- other troubled regions of the world, the Denuclearization of the Korean trast, South Africa was receptive to such as the Middle East and South Peninsula and other aspects of the special inspections that in turn Asia, has underscored the dangers North–South dialog. resolved discrepancies about its associated with Pyongyang’s possi- dismantled nuclear-weapons pro- ble two-way trade in Weapons-of- Such a comprehensive strategy gram. The United States govern- Mass Destruction (WMD) technolo- may reduce the tendency for both ment and others continue to gy with other potential prolifera- sides to become preoccupied with believe that North Korea has had tors. Money is highly fungible, and tactical leverage at the expense of underway an extensive effort to payments received from missile strategic advancement. More acquire nuclear weapons and has sales make it easier to support mili- importantly, it recognizes the fun- achieved much progress in that tary and WMD programs, especial- damental substantive relationship effort on its own. Pyongyang’s ly given the otherwise miserable between enhancing real security, emphasis on self-reliance has com- economic performance of the including verification, and broader plicated efforts to understand and DPRK. And what goods other than human interactions and openness. restrain the North Korean nuclear cash does North Korea get in return Emphasis on a more comprehen- program. for its missile and other sales? Has sive strategy, however, may also North Korea received inputs to its subject progress on the KEDO reac- We have long focused on the nuclear-weapons program from tor project and Yongbyon disman- impressive efforts by which outside its borders? To deal with tlement to a less buffered linkage Pyongyang has acquired an indige- missile launches and military trade, to the ups and downs of engage- nous capability to produce nuclear the US has been negotiating on ment and confrontation. Thus, weapons, especially in light of its terms to obtain a freeze on the issues such as family reunification political and economic isolation. In DPRK’s missile program. and high-level meetings such as this age of globalization of technol- Engagement has begun on this the recent summit by the heads of ogy, a deeper perspective may be issue, but the prospect of being con- state of the two Koreas may add to necessary. The “loose nukes” and fronted with one WMD or military the dynamics of the AF project “loose nuclear material controls” challenge after another is worrisome management and verification. associated with the breakup of the given the many types of weapons Soviet Union, and the emergence and means of delivery that might be of gray and black markets among deployed or marketed. Dynamics of nations of concern, such as those Technology–Policy described by the Rumsfeld Trade in WMD reminds us that Commission, suggest that indige- Pyongyang has continuously Interaction nous sources of nuclear capability sought new access to resources and are only a part of the problem. In new bargaining leverage. Whether The AF lies squarely at the begin- short, even as KEDO program slip- raising the specter of war, threaten- ning of a process that has been page puts off dismantlement of the ing to withdraw from the NPT or punctuated on both sides by— Yongbyon facilities, nuclear other agreements, or highlighting weapons-relevant activities outside the economic misery of its own 1. Delayed implementation, as well as inside these installations people, the DPRK has become also could undermine the AF or adept at identifying means to 2. Disagreement over compliance, stall its implementation. strengthen its negotiating posi- tions. It has been a “tit-for tat” 3. Near dissolution of fundamental Implementation and verification process and more. Movement agreements, and of the KEDO reactor program and toward a more comprehensive with it, the AF, face other large, approach as embodied in the Perry 4. Brinkmanship, including saber crosscutting issues as well. and Armitage Reports seeks to deal rattling.

4 INTRODUCTION

Prudence dictates analyses of different arenas for cooperation, • The ability to make judgments these and possible future crises. competition, and conflict simply about the meaning of obligations. This seems inherent in a process reflect progress or regression on involving efforts at cooperation by the nonproliferation front. In other • The centrality of provisions. parties deeply divided over politi- cases, they reinforce progress or cal, economic, and security objec- amplify setbacks. Thus, positive • The probabilities of cheating. tives and with fundamentally dif- feedback may result in a de facto ferent views on openness. hierarchy of security-building • The risks associated with non- arrangements, but negative feed- compliance. Verification issues take on greater back may amplify setbacks and significance in such an environment. encourage the widening of fissures. • The timeliness of warning Indeed, verification issues have been In some cases, strategic or tactical obtained. associated more or less with most of considerations involving either these past confrontations with international or domestic politics • The efficacy of redress or North Korea over the nuclear ques- may result in progress in one arena enforcement. tion. As we have examined verifica- being deliberately associated with tion of nuclear-material production stalemate or loss in another. • Other related costs and gains. and related activities, we have tried to keep this in mind. To understand Much has been written on the Not all of these elements are how verification might become motivations and negotiating tactics dealt with at length in this work. involved in a crisis again, one must of the DPRK, the ROK, the US and Nevertheless, risks and benefits look beyond the simple history of others involved with the North must be weighed against their implementation of IAEA obligations Korean nuclear debate. Although impact on the ends for which any and examine several paths that par- the dynamics of interaction fre- agreement is a means and against allel interaction between the IAEA quently repeat themselves, much the consequences and options if and the DPRK (Introduction Figure, disagreement exists among experts those ends are not met. Verification page 6). Lack of time and a more over interpretation. Is North Korea may never be perfect, but effective technical interest preclude our deal- afraid that resolving uncertainty verification can contribute signifi- ing with all of these interactions will reveal a secret program it cantly to confidence. Inadequate other than to mention their signifi- claims does not exist? Or, alterna- verification can lead to overconfi- cance. Notable among these parallel tively, is Pyongyang afraid to dence and miscalculation. paths are— reveal that the program is thus far unsuccessful, thus depriving it of In the language of the AF, the def- 1. The total nuclear programs of bargaining power? Both could be initions of clear success or obvious the DPRK (military and civilian). true at different times or even at failure are relatively easy to under- the same time. stand. Whatever the debate over the 2. Related bilateral and regional amount of potential weapons-grade efforts to engage Pyongyang on material that can be accumulated the nuclear question. By What Standard Shall We under various scenarios, if the Judge, and When? DPRK truly abandons its nuclear- 3. Other North Korean and region- weapons program, the central goal al military activities. Our primary interest in this will have been achieved. Effective report is the verification of nuclear- verification of the AF can help us 4. Broader NPT-related activity. material production and the reduc- know if this has happened, but that tion of risk associated with diver- knowledge is not guaranteed. If the 5. Wider defense and arms control sion or breakout. Fundamental to DPRK develops and manufactures developments. any such effort is the ability to or otherwise obtains or retains monitor activities, a main concern nuclear weapons—whether that is 6. The larger process of political of this study. Technology such as one weapon or many—the AF will change in the Koreas. instrumentation, sensors, sam- have failed to meet its explicit objec- pling, tagging, diagnostics, and tive. Verification measures may give 7. Related political events around communications contribute signifi- us earlier indications of this unde- the world. cantly here. Yet, verification has sired outcome, but again perfect always been a larger process than warning is not in the cards. If the These interrelated dynamics monitoring. Verification assess- DPRK uses the AF for continued work for and against confidence. In ments involve— leverage, verification measures may some cases, developments in the help us manage this dynamic more

5 INTRODUCTION effectively, but we must still recog- return to Cold War levels, confi- war is avoided, what has been lost? nize that these negotiations take dence in actual achievements of the If war is only delayed and ultimate- place against a complex geopolitical, AF will be reduced. Compliance ly made even more destructive, legal, and economic backdrop. disputes under the AF can be both a what has been gained? Will delay cause and an effect of such tensions. mean more or less chance that The AF offers important nonpro- nuclear weapons will be deployed liferation opportunities and verifi- Clear success and obvious failure or even used by North Korea? Will cation tools, yet, the AF aspires to may be easiest to understand, but in delay now mean more or less contribute more. If the military the history of negotiations with reform, and over what timeframe threat on the Korean Peninsula is North Korea, success and failure are and circumstances? Will delay affect greatly diminished and relations normally not so clear. As the above international cohesiveness in sup- among the nations involved are discussion of the dynamics of port of the IAEA or in the United normalized on a sound basis, the engagement on the Korean penin- Nations Security Council of US non- achievements of the AF will be even sula suggests, every success brings proliferation and other foreign-poli- more solid. Cooperation on verifica- with it a price. Every advance stim- cy objectives? These larger questions tion may facilitate normalization. ulates the grounds for new steps must be kept in mind as we exam- Indeed, it may facilitate political backward. The value of every spe- ine potential failure modes and change in North Korea. If tensions cific objective gets re-evaluated. If consider means to prevent failure.

Lady Liberty Germany Soviet Union Asian currency Clinton meets GW Bush meets Related demonstrations Reunited collapses crisis DPRK general Kim Dae Jung, In Beijing orders review World Gorbachev & Ceausescu Moscow Armitage & Perry reports executed Putin visits Political peristroika coup Pyongyang Events Berlin Wall Carter in US provides food aid falls Gulf War Pyongyang & lifts some sanctions US allies begin normalization DPRK boycotts 2 Koreas in UN Kim Jung-Il ROK permits Nobel to Kim Dae Jung Seoul Olympics DPRK TV broadcasts Political Kim Young Sam Sec State in DPRK Change Kim Dae Jung North-South Kim Jung Il Agreement on N–S Summit visits Shanghai in Korea Roh Tae Woo Reconciliation Floods & Famine in DPRK 3 N-S Ministerials Kim Il-sung dies Sunshine Policy

Related Presidential Nuclear US nuclear test US, ROK, & Japan US debate on Initiatives moratorium establish TCOG ballistic missile Defense/ CFE & CSBMs defenses in Europe CWC Arms MTCR INF Nunn–Lugar Control established Treaty START I START II CTR CTBT NPT extension, 187 Parties NPT, NP, Chernobyl UN summit on Rumsfeld & IAEA Nonproliferation Commission around DPRK SA, Uk, Kaz, Indo-Pak. Pakistan globe signs NPT UNSCOM & Belarus nuclear launches rollback tests DPRK missile Team Spirit US military buildup Four Party Talks Japan fires on Military cancelled & evacuation planning DPRK ship Activity Evidence of DPRK nuclear weapons DPRK DPRK US–DPRK missile talks relevant program reported DCI fears I or 2 weapons from submarine missile test ROK sinks to Korea material diverted in ‘89 infiltration over Japan DPRK ship

North-South N-S Inspection Regional Denuclearization regime talks falter Agreement Kumchang-ni North-South Nuclear US–USSR AGREED tunnel defense ministers Restraint talks on NP JNCC meets FRAMEWORK inspection meeting & DPRK US withdraws DPRK threatens tactical nukes NPT withdrawal

IAEA IAEA IAEA Agreement, Activity Safeguards discrepancies, inspection Agreement not confirms reprocessing, IAEA verify declaration? in DPRK in place for 20 suspect site inspection MW(th) demanded

50-MW(e) and DPRK 5-MW(e) 5-MW(e) Possible Pu 200 MW(e) Nuclear reactor reactor Crisis KEDO organization, reprocessing reactors Delayed LWR LWR nuclear 1st LWR Program begins defueling under over fuel rods Site preparation, operation plant w/o IAEA? construction begins components? operational? & KEDO located construction w/material for ~6 bombs & financing Projects

1985 1987 1989 1992 1994 1996 1998 2000 2002 2004 2006

Introduction Figure. Parallel interaction paths related to the Agreed Framework (AF) and the Democratic People’s Republic of Korea (DPRK).

6 VERIFYING THE AGREED FRAMEWORK: EXECUTIVE SUMMARY

Under the 1994 Agreed accomplished and what is needed risks of diversion of nuclear materi- Framework (AF) between the to accomplish it. Given the US goal al for nuclear weapons. United States and the Democratic of nuclear non-proliferation on the People’s Republic of Korea Korean peninsula, central questions Safeguarding the Nuclear- (DPRK), the US and its allies will include the following: How verifi- Power Reactors Provided provide two large nuclear-power able are the provisions of the AF reactors and some other benefits to and ancillary agreements such as by KEDO the DPRK in exchange for an the Supply Agreement? How well agreement by the DPRK, inter alia, can it be verified that the DPRK has KEDO is a consortium spon- no access to nuclear weapon-usable sored by the US but mostly funded • To declare how much nuclear- material? What is the potential by the ROK. The two large nuclear- weapon material it has produced, impact of delays, disagreements, power reactors to be provided by and lack of cooperation on verifica- KEDO under the AF are of a rela- • To identify, freeze, and eventual- tion? This report is devoted to tively well-understood type built ly dismantle specified facilities answering those questions. and operated by the ROK called for producing this material, the Korea Standard Nuclear Plant Our general answer is that veri- (KSNP). In what follows, we refer • To remain a party to the nuclear fication can be accomplished to a to them as KEDO Reactors 1 and 2. Non-Proliferation Treaty (NPT) satisfactory degree of accuracy on and allow the implementation of most counts, but that special effort The fresh nuclear fuel for these its safeguards agreement. will be needed from the reactors contains no nuclear International Atomic Energy weapon-usable material. On the The AF and associated agree- Agency (IAEA), as well as support other hand, like all current nuclear- ments are now being carried out from the US and the Republic of power reactors, these reactors pro- according to a complex and cur- Korea (ROK). Furthermore, cooper- duce reactor-grade plutonium rently delayed schedule. Benefits ation and openness from the DPRK along with energy for electricity. have been provided to the DPRK, are essential for accurate and com- This plutonium, albeit not ideal, the site for the two nuclear-power plete verification of the AF. can be used to make nuclear reactors has been largely prepared, weapons. Significant amounts of and construction has begun on Our more specific answers are this plutonium are left in the highly some components. For its part, the presented under three headings: radioactive spent fuel when that DPRK has declared some nuclear- one dealing with the nuclear-power fuel is taken out of the reactors. The weapon material and has identified reactors to be provided by the main objective of the IAEA safe- and frozen some facilities for pro- Korean Energy Development guards is to detect the diversion or ducing this material. Organization (KEDO), one dealing clandestine production of nuclear with known or suspected nuclear- material in a timely fashion. As evidenced by the current dis- materials production facilities in the cussion of the AF, including DPRK, and one dealing with possi- KSNPs have a well-developed President Bush’s statement at the ble adverse developments affecting and effective safeguards package White House on March 7, 2001, ver- verification and safeguards. These agreed to between ROK and the ification is an essential part of any answers provide the basis for some IAEA.* A similar safeguards agreement with North Korea, analysis of what the US and its package (INFCIRC 403) has been including the AF. In what follows, allies will and will not know under agreed to between the IAEA and we provide an assessment of various, possible future circum- the DPRK. These safeguards have whether such verification can be stances, and how to minimize the been developed in accordance with

* The IAEA is charged by the NPT with verifying that the nuclear facilities of non-nuclear-weapon NPT parties are used for peaceful purposes and not to make nuclear weapons. The actual measurements and accounting practices used for such verification are called “safeguards.”

7 EXECTUIVE SUMMARY a model agreement known as INF- Using these technologies, if spent-fuel pools under video sur- CIRC 153. They comprise measure- inspectors are promptly available veillance by the IAEA, until its ments, inspections, accounting pro- and allowed to investigate any radioactivity has decayed suffi- cedures and other measures, exam- questionable event, the safeguards ciently to allow its being placed ined in detail in the report.* package should be able to detect into dry storage casks or removed attempts at covert diversion of to a storage facility away from the The safeguards applicable to the plutonium in quantities smaller reactor. Provisions for that part of KSNP reactors are adequate to than eight kilograms (down per- the process, which are some years detect in a timely fashion (defined haps to the amount contained in a in the future, have not been speci- by IAEA as three months) attempts single fuel pin, a few tens of fied or agreed to in detail. Under at covert diversion of “significant grams) in a shorter time than three the AF, the DPRK must allow the quantities” of nuclear material, in months (depending on inspector removal of this spent fuel at the this case plutonium (defined by the availability). request of KEDO. In the case of IAEA on the advice of nuclear- the DPRK, it may be desirable to weapons member states as eight We considered overt and covert remove spent fuel from the coun- kilograms). Effective safeguards diversion and misuse scenarios. We try as soon as is practicable. This require that monitoring and data- could not find credible covert is particularly true for fuel transmission equipment be proper- diversion and misuse scenarios removed early in the reactor life- ly installed and maintained and under the conditions stated above time. The highest quality plutoni- kept secure. In the case of remote for the KEDO reactors. On the um exists in the very first reactor monitoring, the transmission of other hand, overt action to break cycle and it will be the first to lose data must also be secure and unin- out of the safeguards agreement is radioactivity. We note that final or terrupted. Equally important are always possible, as discussed long-term storage of spent fuel properly trained inspectors and the under the “Possible Adverse away from the reactors is a prob- timely review of all safeguards-rele- Developments” heading later on in lem that has not been resolved vant information. Cooperation and the Executive Summary. anywhere and which is more openness on the part of the DPRK political than technical. on these matters are essential. In reaching the above conclu- sion, we emphasize two conditions Verification of the DPRK’s Safeguards are not fixed for all that must be met: Declaration and of the time but are improved as time goes on. Advanced technologies within 1. The IAEA will continue to Disposal and Dismantle- INFCIRC 153 include environmen- require technical and financial ment of Identified or tal sampling of air, water, soil, and assistance from the US and other Suspect Nuclear Facilities vegetation near inspection sites, member states to implement an and secure remote monitoring in effective safeguards regime at Verifying the DPRK’s declara- real time of key entry points, locks, the KEDO reactors. tion and its disposal and disman- seals, power output, and other tlement of identified or suspect facets of reactor operations. All 2. The DPRK must fully commit to nuclear facilities is a complex task. require reliable means of transmis- the terms and conditions of its Neither the scope of activities to be sion, as well as an automated safeguards agreement. This will verified nor the history of the review and analysis of acquired require providing the IAEA DPRK’s nuclear program is fully data. Several enhancement meas- inspectors full and prompt known at this time. The task may ures, now in demonstration stage, access to nuclear facilities for the be broken down and analyzed as could be applied to the data-trans- purposes of performing inspec- follows. mission program. These technolo- tions and equipment installation gies are currently being tested in and maintenance, as well as pro- Verification of DPRK’s the ROK and in other NPT parties viding all requested information Declaration with INFCIRC 153 safeguards. The in a timely fashion. implementation in the DPRK of The DPRK has declared the exis- such improved safeguards systems After being taken out of the tence of the following facilities as would add significantly to their reactor, the highly radioactive, subject to inspections, six at effectiveness. plutonium-containing spent fuel Yongbyon and one at Taechon: usually spends several years in

* Note that IAEA safeguards are not required positively to establish the physical diversion of material. To establish a violation of an INFCIRC 153-type safe- guards agreement, it is enough for the IAEA Board of Governors to conclude that “the Agency is not able to verify that there has been no diversion of nuclear material required to be safeguarded under the Agreement to nuclear weapons or other nuclear explosive devices.”

8 EXECUTIVE SUMMARY

• Two reactors that were in opera- The DPRK must provide unhin- objectives could work with the tion, one of which is subject to dered access for the purpose of IAEA to develop plans and the freeze. measurements and inspections at resources adequate to this task, all identified and suspect DPRK anticipating and preparing for • One reactor that remains nuclear facilities wherever located. the specific verification problems incomplete. to be encountered. According to the IAEA, the path • A fuel-fabrication facility. forward to determine the correct- 2. Information provided by third ness and completeness of the initial parties not limited to the US has • A radioisotope laboratory. declaration has been developed, been and will continue to be an including planning for contingen- essential support to the IAEA. • A waste site. cies. It includes plans for the cases where large amounts of radioactive 3. The US and other states support- • A reactor under construction at wastes are discovered in previous- ing nuclear nonproliferation Taechon that also remains ly hidden waste sites. There is no objectives, including especially incomplete. plan to attempt to verify the accu- the ROK and Japan, must sup- racy and completeness of the initial port the IAEA in maintaining the Plutonium-containing fuel has declaration unless access to the highest standards of verification. been withdrawn from one of the suspect waste sites is granted, if Yongbyon reactor cores and is the Agency stays with the recom- Verification of the Disposal and stored in cans in a nearby pool. mendations of former Director- Dismantlement of Identified General Hans Blix. The details of Yongbyon Facilities The DPRK further declared that the IAEA’s plan are not public it had separated a small quantity information. With DPRK coopera- Plans for disposing of existing (less than 100 grams) of plutonium. tion, the process is estimated to DPRK spent fuel have not been take 2–4 years. finalized. So long as the spent fuel The DPRK’s initial declaration to and separated material are continu- the IAEA identifying facilities and A review of the methods avail- ously monitored by the IAEA, they quantities of nuclear material sub- able to the IAEA indicates that do not pose a serious verification ject to safeguards appears to be where the full panoply of IAEA problem. incomplete. At least one unde- measurements and inspections are clared waste site has been identi- brought to bear, there will be rea- While it is understood the fuel fied, probably containing addition- sonable confidence (i.e., as good will leave the DPRK, the destination al plutonium-containing wastes, confidence as there is with respect for the fuel has not been determined. and there is evidence for more fuel to other states that have ended Only the Sellafield plant in the removal and more plutonium-sep- their nuclear weapons activities United Kingdom currently has facili- aration activity than the DPRK has and joined the NPT as non-nuclear ties specifically for handling the declared. While estimates of pluto- weapons states) that past nuclear Magnox type of fuel used in the nium production vary, there is activities have been generally iden- Yongbyon reactors.* According the strong evidence that the DPRK tified. The exact quantity of pluto- AF timetable, arrangements must be separated more than its original nium separated can only be made by the time the first KEDO declaration of less than 100 grams. approximately determined. reactor is completed. Disagreement An amended declaration will very Depending on the reactor operat- over the destination for this spent likely be required for the IAEA to ing history available, there may not fuel could result in significant delays complete its verification activities. be high confidence in the exact in completing the KEDO reactors. With such an amended declaration, number of kilograms separated. Perhaps of most concern is the access the US and other interested parties We note three points: the DPRK would have to the would have more complete and weapons-usable material in the reliable knowledge of the DPRK’s 1. The IAEA is likely to need addi- spent fuel should it decide to abro- nuclear materials and facilities, and tional resources to prepare for gate the AF. more complete and ongoing safe- and carry out its verification guards over such material and activities in the DPRK. The US After the first KEDO reactor has facilities would be possible. and other states supporting been completed and the DPRK’s nuclear nonproliferation spent fuel has been removed from

* Magnox fuel readily corrodes and is unsuitable for long-term storage or geologic disposal.

9 EXECTUIVE SUMMARY

Yongbyon, the DPRK will be Possible Adverse declaration could decay to an extent required to dismantle its frozen Developments that depends on the techniques to nuclear facilities. The IAEA has be used—and that we cannot fully compiled considerable information Adverse developments are use- assess. A lessening of the IAEA’s regarding decommissioning that fully considered in the framework ability to reconstruct the reactor’s should help the DPRK in the dis- of the time sequence of events operating history could be impor- mantling effort. The major disman- envisaged by the AF. This time tant in estimating the amount of tlement challenge will be dealing sequence, together with its implica- plutonium separated. On the other with the radioactivity in the reactor tions for verification and safe- hand, verification that the DPRK and other facilities. While the basic guards, is summarized in the does not have a program at industrial processes for decommis- Executive Summary Table (next Yongbyon is not affected by delays sioning and dismantling nuclear page). The time sequence present- so long as the Yongbyon site contin- facilities are reasonably well ed is not complete. A number of ues to be verifiably frozen. In understood, this is an expensive other important steps are linked to essence, delays at this stage mean and complex undertaking. The cost the time sequence. The steps sum- that the situation remains frozen to dismantle these facilities, based marized in the Table are those with no additional source of nuclear on past experience, is likely to be which require verification or safe- material provided to the DPRK. at least a few hundred million dol- guards or which bear on the timing lars. Some of the crucial pipes and of verification and safeguards. The Delays after either or both the special equipment will have to steps in the left-hand column are KEDO reactors are completed and be removed or destroyed early in shown in time order. As actual delays in the DPRK’s allowing spe- the process to make the dismantle- dates have slipped and are likely to cial inspections or additional safe- ment verifiably irreversible. continue to slip, no dates are guards measures could have more shown. Only the sequence of serious consequences, and are dis- Verification Regarding Other events is used. cussed below. Suspect Facilities We note, in connection with the Non-Cooperation with the There have been and may even- sequence of events in the table, the IAEA’s Verification of tually again be facilities other than crucial link between Steps 2 and 3, Compliance those at Yongbyon that come under i.e., the IAEA declaring the DPRK suspicion of being used for nuclear- to be in compliance, the delivery of Limited or lack of cooperation material production, storage, or nuclear components to KEDO on the part of the DPRK has much other nuclear activities. These could Reactor 1, and the necessary the same effect as delays at this pose different verification problems Agreement for Cooperation with stage of the AF. Non-cooperation than the Yongbyon facilities. Some the US. No Congressional review of at Yongbyon and other possible additional methods beyond INF- an Agreement for Cooperation and sites of past activities has taken CIRC 153, notably broader environ- no provision of nuclear components the form of not allowing access by mental sampling away from to the DPRK is possible until the IAEA inspectors to undeclared declared sites and easier inspections IAEA is satisfied that the DPRK’s suspect facilities, limiting IAEA at undeclared facilities would be reports on all its nuclear materials measurements where inspectors useful—and could be necessary—to and facilities are accurate and are allowed, and preventing off- detect these facilities. Satellite moni- complete. site environmental sampling. toring and other third-party infor- Continuation of this behavior, mation are needed as well. With Further Delays either in the form of outright these tools, the uncertainty associat- denial or undue delay, would ed with the detection of undeclared Delays have occurred and are have the effect, under the AF, of activities would be diminished. likely to continue to occur for a preventing delivery of key variety of legal, financial, and politi- nuclear components for the KEDO At the same time, “false posi- cal reasons. A number of these rea- Reactor 1 to the DPRK. Citing the tives,” such as the supposed but sons are noted in the report. Delays time necessary to complete the now known to be non-existent reac- before any KEDO reactor is com- verification exercise, the IAEA’s tor at Kimchang Ni, can pose prob- pleted puts off the eventual IAEA Director-General has repeatedly lems for both IAEA and US credi- declaration of DPRK’s compliance requested the DPRK to get started bility and for effective cooperation. and the attendant knowledge of as soon as possible. They have Being overly suspicious poses prob- what material and facilities the given no indication that they are lems for effective verification just as DPRK actually has. The ability of ready to proceed. not being suspicious enough. the IAEA to verify the initial DPRK

10 EXECUTIVE SUMMARY

Executive Summary Table. Envisaged time sequence of events.

Step Verification Issue Possible Problems Partial completion of KEDO Reactor 1 in None but the IAEA wants to start the next Financial and legal delays cause some the ROK and partial preparation of Kumho step early (2-4 years needed) loss of data at Yongbyon site in the DPRK

IAEA declaration that the DPRK is in com- Verification of accuracy and completeness 1. DPRK does not open suspect sites to pliance with its agreements of the DPRK’s initial declaration on all the IAEA. nuclear materials in the DPRK, at 2. The IAEA’s activities are interfered with. Yongbyon, and possibly elsewhere 3. Initial declaration is wrong—can it be amended?

Delivery of KEDO Reactor 1’s key nuclear Safeguards for KEDO Reactor 1 are 1. Disagreements over the extent of components starts. Transfer of Yongbyon installed. Transfer of Yongbyon spent fuel safeguards. spent fuel (and other material?) to “ulti- (and other material?) to “ultimate disposi- 2. Disagreements over site of disposition. mate disposition” starts. tion” verified. 3. Disagreements over what is to be transferred.

Simultaneous completion of previous Safeguards on KEDO Reactor 1 opera- Same as previous, plus interference with steps tional. Disposition site monitored. KEDO Reactor 1 safeguards

Dismantlement of Yongbyon facilities in Safeguards for KEDO Reactor 2 are 1. Disagreements over extent of parallel with delivery of KEDO Reactor 2’s installed. Dismantlement verified. safeguards. key nuclear components 2. DPRK abrogation. 3. US or ROK non-compliance with AF.

Simultaneous completion of previous Safeguards on KEDO Reactor 2 1. Interference with safeguards. steps operational 2. DPRK abrogation. Disposition of KEDO spent fuel. Monitoring disposition site(s). Disagreement over site of disposition.

Need for an Amended DPRK more complete ongoing safeguards nevertheless that could occur under Declaration over such material and facilities pretext of safety or environmental would be possible. As noted earlier, considerations. In that case, or in When the IAEA is allowed to however, there is an irreducible any case of a disagreement over the carry out inspections and measure- uncertainty in verification of the transfer of nuclear material, the ments at Yongbyon and at such amount of plutonium separated. IAEA would have to monitor the other sites as may be needed, the material in the DPRK on a continu- IAEA is likely to conclude that Disagreements Over Material To ing basis. If the DPRK were to abro- more plutonium-containing fuel Be Transferred from the DPRK gate the AF in the future, it would and/or more separated plutonium regain access to this material. exists in the DPRK than the DPRK A verification issue would arise Such a disagreement could bring originally reported. In that case, it if the DPRK took a narrow view of the AF to an end. Because, howev- would become an issue for negotia- what it was obligated to allow to er, a transfer takes place simultane- tion between the US and the DPRK transfer out of the country under ously with the installation of key whether an amended DPRK decla- the AF, for instance, restricting the nuclear components in KEDO ration would be presented and transfer to spent fuel from specified Reactor 1, it is desirable that an allowed under the AF. From a veri- Yongbyon facilities rather than explicit agreement be reached fication point of view, the situation from all relevant facilities. If these regarding what materials are to be would be better than it is now: the facilities had been identified and transferred, including material that US and other interested parties inspected, and nuclear material to may be found in suspect but not would have more complete and be safeguarded found there, it identified facilities. reliable knowledge of the DPRK’s seems unlikely that the DPRK nuclear material and facilities, and would not allow its transfer, but

11 EXECTUIVE SUMMARY

Disagreements Over the Site of some of them can give early warn- earlier, remote monitoring coupled Ultimate Disposition ing of illicit activity, significant with prompt readout of data at the disagreement over cooperation on IAEA and the availability of So long as the sites of ultimate this issue should be taken quite inspectors in the region could disposition for the DPRK’s nuclear seriously. reduce the warning time to that material are outside the DPRK and period. Lack of cooperation with otherwise acceptable to the US, no Interference with Safeguards for such measures should be consid- verification issue is likely to arise KEDO Reactors ered significant. from disagreements over this issue, except that such disagreements Once safeguards have been Abrogation of AF or NPT after could prolong the time that the agreed to and are operational, the KEDO Reactors Are material remains in the DPRK and interference with their operation Installed that the IAEA has to monitor it. in any form (e.g., denial or delay Again, such delays would mean of needed access, late or incom- No safeguards can prevent that, if the DPRK were to abrogate plete records, interference with overt acts such as the abrogation the AF during that period, it would transmission of data, unreliable of agreements and the expulsion regain access to this material. It is power supply, or interference with of inspectors. In the case of abro- highly desirable that full agree- updating of equipment) could gation, the US would know how ment be reached regarding the site seriously affect the assurance that much potential nuclear-weapon of ultimate disposition before the nuclear material, in particular material is in the DPRK initially. KEDO reactors come on line. spent fuel, is not being diverted. It would, as now, externally moni- The seriousness would depend on tor such large-scale activities as Disagreements Over the Extent of the details of the particular situa- reactor operations and the con- Safeguards for KEDO Reactors tion, but it would be particularly struction of facilities, and it damaging to verification if it would to some extent identify As reviewed above, there are occurred after the initial refueling, major activities. It would be pos- useful advanced technologies that when the initial load of spent fuel sible to estimate how much pluto- the IAEA can use over and above contains plutonium that is particu- nium is made in reactor opera- the measures in place at most reac- larly attractive for weapons use. tions subsequent to abrogation. tors, but still within the INFCIRC The fuel initially discharged con- Handling the radioactive spent- 153 package of safeguards. Though tains enough plutonium to make fuel rods and separating the plu- that package of safeguards has 10–20 nuclear-explosive devices, tonium are major operations, proven its worth, the new meas- for instance, of the type tested in requiring facilities recognizable ures, together with the enhance- 1945 at Trinity. Spent fuel must by national technical means. The ment measures for the data trans- normally spend several years in facilities, however, might be hid- mission, would add significantly to cooling ponds under safeguards den underground. assurance of compliance with the before it can be moved. Once it is NPT. The specific package to be removed from the pool building, it Once the facilities are built and implemented at the KEDO reactors must remain the subject of contin- the procedures practiced (without has to be negotiated, and disagree- uing safeguards in the DPRK until actually having light-water reactor ments could occur over application it is removed from the DPRK, at (LWR) spent-fuel assemblies avail- of these safeguards. Depending on which time safeguards appropri- able for realistic tests), the time the outcome of the negotiations, ate to the new location are needed to separate several bombs’ the effectiveness of safeguards imposed. This location is likely to worth of material might be only could be less than optimal. be the ROK and removing spent days or weeks if all went accord- fuel from the DPRK is thus likely ing to plan. In view of others’ Successful negotiations for to depend on having storage facil- experience, however, the time installing these measures in the ities in the ROK. needed is likely to be much DPRK will depend on similar longer. The IAEA’s Standing measures being implemented into The IAEA safeguards are nor- Advisory Group on Safeguards the ROK. In general, the role of the mally aimed at detecting the diver- Implementation has estimated that ROK is crucial in the safeguards sion of eight kilograms of weapon- the time required to convert pluto- area as well as in all other aspects usable material within a period of nium in spent fuel into a weapon of the KEDO program. Given that three months. It would be desir- is one to three months, compared the new measures are increasingly able to have the ability to detect to seven to ten days for metallic being tested and adopted in many the diversion of smaller amounts plutonium. countries, however, and given that within a shorter period. As noted

12 EXECUTIVE SUMMARY

It is possible to remove plutoni- nuclear non-proliferation Chapter 4 provides a detailed um-containing spent fuel from the objectives, including especially description of the KEDO reactors reactor site sooner than normal the ROK and Japan, must sup- and the safeguards expected to practice. Such early removal may be port the IAEA in maintaining the be applied under the present desirable in the case of the so-called highest standards of verification. agreements. “Beginning of Life” fuel, which is more desirable for weapons use and The parties to the AF are faced Chapter 5 describes diversion which will be the first to cool (i.e., with both opportunities and chal- and misuse scenarios for light-water lose radioactivity) sufficiently to be lenges. The opportunities for the reactors, such as the KEDO reactors, handled. An agreement to do this DPRK include not only the provi- and their possible consequences. would best be reached ahead of sion of electricity at concessionary time. Covert removal of such fuel, rates,* but also the opportunity to Chapter 6 provides a description or later fuel, by the DPRK would be become fully cooperative and open of known and suspect nuclear- difficult as the casks needed to in an important area of internation- related facilities in the DPRK, remove the fuel are large enough to al concern. Obviously, these oppor- specifically at Yongbyon and have satellite signatures, and as the tunities are also challenges, both Taechon, and of the methods that fuel would have to be replaced in for the DPRK and for the US, ROK, can be used to assess how much the pool covertly. Removing spent and other KEDO members. The plutonium has been produced and fuel from the DPRK as soon as is challenges include the challenge of separated. practical minimizes the advantage verification that has been taken up of overt fuel diversion after a possi- in this report. Chapter 7 summarizes the ble abrogation. sequence of activities envisaged by We conclude, based on the con- the AF and notes what there will Executive Summary siderations of this report, that the be to verify and what measures of Conclusions challenges of verifying the Agreed verification will take place as the Framework can be met, under the agreement is carried out. While more complete conclusions conditions outlined in this report and recommendations are present- and summarized here. With these Chapter 8 examines the process ed at the end of the report, briefly conditions met, verification is of implementation under assump- the bottom lines are as follows: robust under most scenarios. tions of varying degrees of DPRK cooperation or obstruction. 1. With adequate preparation and Organization of Report support of the IAEA, and with The Conclusions and Recom- full cooperation of the DPRK Chapter 1 provides a brief histo- mendations section provides a brief with all measures called for in ry of the DPRK’s nuclear program summary of the principal conclu- the AF and in the safeguards and its interactions with the IAEA sions and recommendations made packages (standard and addi- and other countries. A summary in the report. The recommenda- tional) applicable to the KEDO- and analysis of the AF and related tions are not separated from the type reactors, the AF can be ade- agreements that are now or will be conclusions. quately verified. in force is given. Authors and Sponsors 2. For this to take place, the DPRK Chapter 2 summarizes the safe- must reach a decision to make guards and how they are applied The work presented in this its operations in the nuclear under existing agreements. report was undertaken at the sug- area, past and present, actual Possible additional agreements that gestion of former Secretary of and suspected, fully open. Such the IAEA has reached with other Defense and now Stanford openness will be readily evident countries are also discussed. Professor William J. Perry. The at the working inspectors’ level. sponsoring organizations were the Chapter 3 describes the KEDO Center for International Security 3. The IAEA is likely to need addi- reactor site, nuclear reactor activi- and Cooperation of the Institute for tional resources to prepare for ties in East Asia, and related issues International Studies at Stanford and carry out its verification such as fuel shipment and electrici- University, directed by Professor activities in the DPRK. The US ty provision. Scott Sagan and Professor Chris and other states supporting Chyba, and the Center for Global

* Such provision depends on a number of other developments, such as improvements in the DPRK electrical grid.

13 EXECTUIVE SUMMARY

Security Research at Lawrence Every author contributed to the parts of the report. Errors and Livermore National Laboratory, Conclusions and Recommendations, omissions, of course, are solely the directed by Dr. Ronald Lehman. and also reviewed the other authors’ responsibility of the authors. The project was carried out under chapters. Everyone also contributed the leadership of Professor Michael to the Executive Summary. In this report, the authors do May of Stanford University. not take a position on the desir- In the process of preparing this ability of the Agreed Framework, The Stanford authors were report, the authors convened two the future of the DPRK, or the Professor Michael May (general workshops, one at Stanford desirable US policy in these mat- editor, Executive Summary, University in September 2000, and ters. It is likely that views on Chapters 7 and 8), Professor one at the Lawrence Livermore these matters differ among the George Bunn (Chapters 1 and 2), National Laboratory in individual authors. The report is Dr. Chaim Braun (Chapters 3 and January–February 2001. During the solely an assessment of safe- 4), and Dr. William Sailor course of these workshops, the guards and verification. (Chapters 6 and 8). authors received invaluable sug- gestions, information, and criti- The report does not represent the The Lawrence Livermore con- cisms from a number of people views of Stanford University or any tributors were Drs. Zachary Davis who had had experience with the of its institutes or centers, nor those (Chapter 8), James Hassberger AF, the IAEA, the DPRK, nuclear- of the Lawrence Livermore National (Chapters 4 and 5), Ronald reactor safeguards, and other rele- Laboratory, the Department of Lehman (Overview), Wayne vant topics. The authors greatly Energy, the US Government, or the Ruhter (Chapter 5), Robert Schock, profited from these and are grate- University of California. and Nancy Suski (Chapter 8). ful to all those who took the time to read and critique early versions The authors are grateful to the of this report and to contribute to Carnegie Corporation of New York the workshops. Special thanks are for their support of this work at owed Richard Hooper, who Stanford. contributed significantly to several

14 CHAPTER 1 A BRIEF HISTORY OF THE DPRK’S NUCLEAR WEAPONS-RELATED EFFORTS

This chapter is not a complete This reactor is located, along • a 50-MW(e), partially built gas- history but only highlights those with most of the other known graphite reactor. Its construction aspects of the history and of the nuclear facilities, at Yongbyon in has halted, and it has never been agreements entered into by the central North Korea (Fig. 1-1). Also operated. Democratic People’s Republic of at Yongbyon are— • a small highly enriched uranium Korea (DPRK) that bear on the (HEU) research reactor that has safeguards and verification of • a radiochemical laboratory for been decommissioned. nuclear facilities and activities. We plutonium separation. note and summarize agreements relevant to safeguarding the Vladivostok Figure 1-1. General Liaoyuan Onsong Korean Energy Development Yanji map of the Korean RUSSIA peninsula. Organization (KEDO) reactors and CHINA Unggi verifying the frozen or dismantled Najin Fushun Puryong status of nuclear weapons-related Hunjiang Shenyang Tonghua Chongjin facilities. Chasong Hyesan Kapsan 1.1 Early History Kanggye Kilchu Kuandian Chosan Songjin The Soviet Union began training Dandong Changjin Pukch’ong Huichon Hamhung North Koreans on nuclear matters Sinuiju Taechon Sinpo Hungnam in the early 1950s. In 1965, the Yongbyon Anju Soviets provided a small, Yonghung Korea NORTH Choson 2-MW(th), light-water-moderated, Bay KOREA Man Wonsan research reactor that burned highly P’yongyang Namp’o enriched uranium: the IRT-2000 Kosong research reactor subsequently Sariwon SEA OF JAPAN upgraded to 4 MW. Then, to Haeju Chorwon Kaeson Yanggu reduce its reliance on outside assis- Chunchon Kangyang tance, the DPRK began mining or Seoul producing uranium and graphite, Inch’on Samchok began experimenting with plutoni- Chingju um separation, and built a YELLOW SOUTH KOREA SEA Chongju Andong Yondok graphite-moderated reactor that Sangju burned natural (unenriched) urani- Taejon Kunsan Kumch’on Puhang um. It was similar in design to the Chonju Taegu reactors used by Great Britain to make the plutonium best suited for Kwangju Chinju Pusan weapons. It was a 5-MW(e), Masan Mokp’o 20-MW(th)* reactor used to irradi- i t r a Sasuna S t ate fuel rods from which the DPRK e a r o later extracted plutonium. This K Shimonoseki Ube reactor can probably produce Cheju haehyop Kita-Kyushu enough plutonium for approxi- 1 Cheju 0 50 100 Miles mately one bomb per year. 0 50 100 150 KilometersJAPAN

* While the DPRK usually designates its reactors by their electrical rating, that rating is the product of the reactor’s thermal power and the efficiency of the rest of the plant at converting that power into electricity. From the point of view of assessing a plutonium-making capability, the thermal rating—usually three to four times the electric rating—is more relevant. Typically, every megawatt of thermal power will generate about half a gram of plutonium per day (actual numbers vary with the type of fuel, reactor design, etc).

15 A BRIEF HISTORY OF THE DPRK’S NUCLEAR WEAPONS-RELATED EFFORTS

• buildings, tunnels, and other came the Bush–Gorbachev graphite reactor’s irradiated fuel facilities that have not been announcements of withdrawals of than the DPRK had reported to the declared by the DPRK but that American and Soviet tactical IAEA.6 When the IAEA requested may have been used to store weapons from other countries.3 In that special inspections take place undeclared spent fuel for a plu- December 1991, President Roh Tae at two undeclared sites in the tonium recovery program. Woo announced publicly, and in Yongbyon area, which satellite the presence of President Bush, photographs suggested might be The entire set of facilities at that there were no US tactical places where the DPRK had hid- Yongbyon is analyzed in detail in nuclear weapons in the ROK.4 den the products of unreported Chapter 6. South Korea announced a plan for plutonium separation, the DPRK a nuclear-weapon-free Korean resisted strenuously. The DPRK also has a partially peninsula and engaged the North built, 200-MW(e) gas-graphite in negotiations that produced a After the IAEA’s Board of reactor located at Taechon, and a general declaration to this end at Governors insisted on these special sub-critical nuclear facility at a the beginning of 1992. This Joint inspections, the DPRK announced university in Pyongyang.2 Declaration on North–South it was withdrawing from the NPT Construction on the 200-MW(e) Denuclearization not only called and gave the required 90-day reactor has also been halted. for a nuclear-weapon-free peninsu- notice. In the IAEA discussions la but also prohibited both the that followed, China opposed In 1985, the Soviet Union per- North and the South from possess- going to the U.N. Security Council suaded the DPRK to join the Non- ing facilities for enriching uranium for an order imposing economic Proliferation Treaty (NPT) by or for separating plutonium from sanctions on the DPRK if it with- promising two light-water reactors spent reactor fuel. Moreover, it pro- drew from the NPT. As a result, the (LWRs) that were better for pro- vided for reciprocal inspections— IAEA’s Board of Governors simply ducing electric power than of the DPRK by the Republic of reported to the Security Council graphite reactors but not as good Korea (ROK) and vice versa.4 what had happened. The Board for producing weapons-grade also used its authority to terminate plutonium. Later, the DPRK said it Despite lengthy negotiations, the all IAEA technical assistance to the was unable to meet its commit- two countries could not agree on DPRK. Subsequently, the DPRK ment to pay for even the site- the sites in each country that ended its membership in the IAEA. survey conducted by the Soviet would be inspected by the other. This, of course, did not affect the Union at a new location in the area Later in 1992, however, the DPRK DPRK’s obligation to comply with now proposed for the new LWRs accepted the NPT-required safe- the NPT or the safeguards agree- (see Fig. 1-1). At the end of 1991, guards agreement for inspection of ment made pursuant to the NPT. after the Soviet Union dissolved all its nuclear installations by the The Security Council then called into its former republics, Russia IAEA (INFCIRC 153 safeguards). upon the DPRK to permit the withdrew the Soviet promise to In accordance with this agreement, IAEA inspections but did not order provide LWRs to the DPRK. the DPRK provided the IAEA with sanctions if the DPRK refused, a report that was supposed to probably because of the likelihood 1.2 Attempts To Restrain declare all its nuclear material and of a Chinese veto.7 the DPRK From Making facilities for IAEA inspection. When the IAEA’s Director-General China urged negotiations with Nuclear Weapons Hans Blix visited these facilities in the DPRK. In consultation with 1992, he reported among other South Korea and Japan, the US Though a party to the NPT from things that the DPRK had a contin- renewed such negotiations. On the 1985, the DPRK did not accept the uing interest in securing LWRs. At last day of the 90-day notice period comprehensive International that time, however, South Korea for withdrawal from the NPT— Atomic Energy Agency (IAEA) rejected the idea of helping the June 11, 1993—the DPRK safeguards agreement covering all DPRK acquire them.5 announced that it would remain a its nuclear activities required by party to that treaty, at least for the the NPT for seven years. In 1991, Before the end of 1992, informa- time being. In a joint statement the first Bush administration tion from IAEA inspectors based with the US, the DPRK said it had developed a program to restrain on environmental samples they “decided unilaterally to suspend as the DPRK from making nuclear had taken at the Yongbyon facili- long as necessary the effectuation weapons. The program relied on ties and from US satellite photo- of its withdrawal from” the NPT. South Korean and Soviet efforts as graphs of the area suggested that Both governments “expressed sup- well as those of the US. In 1991 the DPRK had probably separated port for the North–South Joint more plutonium from its small Declaration of the Denuclearization

16 A BRIEF HISTORY OF THE DPRK’S NUCLEAR WEAPONS-RELATED EFFORTS of the Korean Peninsula in the Meeting again in August 1993, Efforts to agree on a U.N. interest of nuclear non-prolifera- the DPRK rejected a US proposal to Security Council sanctions resolu- tion goals.”8 Both sides agreed to freeze its nuclear program in tion failed to gain China’s assent. general principles about the appli- exchange for conventional power At the same time, US satellite pho- cation of IAEA safeguards and the stations but agreed “to replace its tographs showed the North Korean need for a “fundamental solution graphite-moderated reactors and military forces moving to a war of the nuclear issue on the Korean related facilities with light-water footing. The Pentagon estimated peninsula.” But, there was no reactor (LWR) power plants,” and that a war would result in 300,000 agreement by the US to provide the US agreed “to make arrange- to 500,000 military casualties in the LWRs to the DPRK and no agree- ments for the provision of LWRs of first 90 days. No estimates were ment by the DPRK to freeze the approximately 2,000 MW(e)” to the given for civilian deaths. Both the operation of its small graphite reac- DPRK. The US also agreed to make ROK and Japan opposed going to tor or the construction of its two interim energy arrangements so war with the DPRK.13 bigger graphite reactors.9 long as the graphite-moderated reactors were not operated. DPRK Then, in a June 1994 visit to Further DPRK–US negotiations agreed to “freeze construction” of DPRK, sanctioned by President produced an agreement in July of the two such reactors still under Clinton, former President Jimmy 1993 on the following joint statement: construction, to “forego reprocess- Carter met with the DPRK’s leader ing” of spent fuel and to “seal” the Kim Il Sung and reported to the “Both sides recognize radiochemical laboratory where it press afterwards that the crisis was the desirability of the said reprocessing had taken place. over. He said that the DPRK had DPRK’s intention to It also promised to remain a party agreed not to reprocess the spent replace its graphite- to the NPT and to allow IAEA fuel from the small graphite reactor moderated reactors and inspections. In addition, it declared that had been shut down, to accept associated nuclear facili- its continuing willingness to imple- IAEA inspection of its reactors and ties with light-water- ment the North–South Joint other facilities declared by it to the moderated reactors Declaration on Denuclearization.12 IAEA, and to freeze its existing (LWRs). As part of the nuclear program. In return, the final resolution of the But, the DPRK continued to dis- DPRK expected assistance in secur- nuclear issue, and on the agree with the IAEA over the scope ing LWRs and an end to US efforts premise that a solution of IAEA inspections, with the ROK to impose sanctions if IAEA inspec- related to the provision of and US over whether the annual tors were denied access to other LWRs is achievable, the Team Spirit ROK–US military exer- locations. Thereupon, the ROK USA is prepared to sup- cise would be conducted in South announced that it was prepared to port the introduction of Korea that year, and with the ROK provide technology and major LWRs and to explore over inspections pursuant to the financing to the DPRK for two with the DPRK ways in North–South Joint Declaration on LWRs.14 It may be noted that this is which the LWRs could be Denuclearization. The DPRK shut essentially the same agreement as obtained.”10 down its small graphite reactor, the one from the previous year, but which by then may have contained the DPRK had now prevented the Both sides also agreed to the enough unseparated plutonium for IAEA from making possibly key “full application” of IAEA safe- at least one bomb. It soon began measurements on its spent fuel, guards to the DPRK’s nuclear facil- unloading the fuel rods from the and had delayed the US from ities. The DPRK promised to begin reactor without waiting for IAEA imposing sanctions if the DPRK consultations with the IAEA on inspectors to test samples of mate- prevented the IAEA from inspect- safeguards as soon as possible. rial from them as they were taken ing suspect facilities. Both re-affirmed the importance of out. Such tests could have deter- implementing the North–South mined which fuel rods had been in 1.3 Agreed Framework Joint Declaration on Denucleariz- the reactor for a long time and of October 1994 ation, and the DPRK said it was which had not, and thereby prepared to begin North–South improve the estimate of how much When US and DPRK negotiators talks on nuclear and other issues. plutonium had been made within met again, they produced the Finally, the US and the DPRK the fuel rods. This DPRK move Agreed Framework (AF) of October agreed to meet soon to resolve prompted fear that the DPRK 1994 that gave more detail to the remaining questions including intended to acquire the plutonium basic agreement announced in the those “relating to the introduction for weapons. August 1993 joint statement and in of LWRs.”11 President Carter’s press conference.

17 A BRIEF HISTORY OF THE DPRK’S NUCLEAR WEAPONS-RELATED EFFORTS

The relevant terms of the AF and 3. In the AF, the DPRK promised is evidence, discussed in Chapter the subsequent Supply Agreement that it would “freeze” not only 6, that measurements on these that implements much of it are as the three graphite reactors but products would show that the follows: also the related facilities, which DPRK removed from the reactor it had declared to the IAEA. In more irradiated fuel than it report- 1. The DPRK and the US would the meantime, the US and KEDO ed to the IAEA—presumably to “cooperate to replace the agreed to make arrangements separate out the plutonium.18 DPRK’s graphite-moderated for periodic delivery of oil to the reactors and related facilities DPRK to offset the energy fore- 5. “Key nuclear components” with light-water reactor (LWR) gone due to the freeze on the include items such as nuclear power plants.” Operation of the graphite reactors.17 material, reactors, the equipment small graphite reactor is to be attached to reactors other than “frozen” and subject to continu- 4. DPRK agreed to remain a party the turbine generators, the ing inspection by the IAEA. to the NPT. It also agreed to equipment that controls the level Construction of two larger “allow implementation of its of power in reactor cores, and unfinished graphite reactors is safeguards agreement under the any other components “which also to be frozen. Later, all three Treaty [NPT].” However, the US normally contain or come in are to be dismantled. Spent fuel promised the DPRK that the direct contact with or control the from the small reactor has been IAEA would not be permitted at primary coolant.” These include, removed by the DPRK as indi- first to verify the accuracy of the for example, reactor pressure cated. Canning of all accessible DPRK’s initial report on all its vessels, as well as reactor control spent fuel rods was completed nuclear material. This verifica- rods, pressure tubes, etc.19 in April 2000, and the rods tion could be carried out by, e.g., remain stored in a cooling pond inspecting the two undeclared 6. The Supply Agreement states near the reactor. Pursuant to and disputed sites where prod- that “the provision of the LWR future US–DPRK negotiations, ucts of earlier separations are project and the performance the fuel will be disposed of “in believed by the IAEA and US to steps…are mutually condition- a safe manner that does not be hidden. Inspection for this al.”20 As a result, KEDO’s deliv- involve reprocessing in the purpose would come later: ery schedule is “integrated” DPRK.”15 There is no require- with the DPRK’s performance ment that the DPRK’s small “When a significant schedule in several ways. This Yongbyon research reactor or its portion of the LWR proj- integration is to be accomplished sub-critical research facility in ect is completed, but as follows: Pyongyang be dismantled. before delivery of key nuclear components, the First, the DPRK’s acceptance of 2. The US would make arrange- DPRK will come into full IAEA inspections at the disputed ments for provision of the LWRs compliance with its safe- and undeclared sites is not required to the DPRK through “an inter- guards agreement with until “a significant portion of the national consortium” which the the IAEA (INFCIRC 403), LWR project is completed but US would form and for which it including taking all steps before delivery of key nuclear com- would be the principal point of that may be deemed nec- ponents.” This language brought contact. This consortium, which essary by the IAEA, fol- objections from the IAEA because it became known as the Korean lowing consultations postpones the “full-scope” inspec- Energy Development Organiz- with the Agency with tions necessary to determine ation (or KEDO), would enter regard to verifying the whether the DPRK has separated into a “Supply Contract” for the accuracy and complete- more plutonium than it reported— LWRs with the DPRK.16 At the ness of the DPRK’s initial as is suspected. Major construction end of 1995, the DPRK and report on all nuclear of reactor buildings at Kumho (the KEDO signed the Supply material in the DPRK.” proposed LWR site) and delivery of Contract for the LWRs. In addi- non-nuclear components including tion to the US, KEDO executive Thus, before delivery of “key turbine generators—a “significant board members include South nuclear components,” the DPRK portion” of the LWR project—will Korea, Japan, and the European must permit the IAEA to inspect likely have taken place first.21 On Union. More than a dozen other the two sites where the US and the other hand, the DPRK is not members are in KEDO, includ- the IAEA believe has the hidden entitled to any key nuclear compo- ing Australia, New Zealand, products of its small graphite nents if it does not permit these Canada, Poland, and Argentina. reactor’s earlier operation. There IAEA inspections. Some, including

18 A BRIEF HISTORY OF THE DPRK’S NUCLEAR WEAPONS-RELATED EFFORTS the IAEA Deputy Director for 7. The model for the LWRs will be submitted to Congress. Under Safeguards, believe that the agree- “selected by KEDO” but was recent legislation, an agreement for ment will likely break down at this promised to be “the advanced cooperation indeed cannot be sub- point because the DPRK will resist version of US-origin design and mitted at this time. the inspections necessary to bring it technology currently under into compliance with its safeguards production.”26 In fact, the reactors Recent legislation requires that agreement.22 Others experts will be manufactured in South the President, when submitting believe, even if the DPRK cooper- Korea using as models two exist- such an agreement with the DPRK, ates with the IAEA to permit these ing ROK reactors based on a US- certify that IAEA inspections— inspections, the whole process will origin Combustion Engineering such as those required by the AF— require so much investigative work design but including refinements establish the accuracy and com- and materials testing that it will of that design. pleteness of the report the DPRK likely take two years or more, thus made to the IAEA when inspec- delaying provision of the nuclear 8. The DPRK must eventually repay tions began in 1992. For example, components and completion of the KEDO for the LWRs “on a long- the inspections must establish that program.23 term, interest-free basis.”27 KEDO there has been no clandestine pro- will provide nuclear fuel for the duction of plutonium.31 Thus, Second, when delivery of the initial loading of the LWRs. The IAEA “full-scope” inspections that “key nuclear components” for the DPRK promises to use the reac- may well take two to four years first LWR begins, the transfer of tors, technology, and nuclear must precede congressional action spent fuel from the small graphite material involved “exclusively for on an agreement for cooperation. reactor must begin. The transfer is peaceful, non-explosive purpos- to be completed when this first es.” In addition, it promises not to The basic nonproliferation compli- LWR is completed. Again, if the reprocess or increase the enrich- ance requirements of the agreements DPRK delays delivery of spent ment level of any nuclear material referred to above are as follows: fuel, KEDO can delay delivery of acquired pursuant to the agree- LWR components.24 ment, and not to transfer any 1. In the NPT, the DPRK has agreed nuclear material, equipment, or “not to manufacture or otherwise Third, “when the first LWR is technology acquired pursuant to acquire” nuclear weapons. In the completed [at Kumho], the DPRK the agreement outside the territo- NPT, it has also agreed to accept will begin dismantlement of its ry of the DPRK except for the IAEA safeguards on its nuclear frozen graphite-moderated reactors spent fuel transfer already activities for the purpose of ful- and related facilities, and will com- described.28 filling this obligation. plete such dismantlement when the second LWR is completed.” Thus, 9. The AF provided that the DPRK 2. The DPRK’s safeguards agreement KEDO must complete the first LWR and the US would conclude “an (IAEA INFCIRC 403) says that the before any of the graphite reactors agreement for cooperation in the IAEA shall “verify” the “findings” are dismantled but can schedule peaceful uses of nuclear energy” of the DPRK’s accounting and deliveries for the second LWR as necessary.29 Later, the Supply control system showing that “there based upon the DPRK’s perform- Agreement said “In the event has been no diversion of nuclear ance of its dismantlement obliga- that US firms will be providing material from peaceful uses to tions. The deliveries of the nuclear any key nuclear components,” nuclear weapons.” The agreement components for the second LWR such an agreement would be is based on the IAEA’s 1972 model can be staged in parallel with pro- negotiated “prior to the delivery agreement for NPT safeguards, the portional steps by DPRK to dis- of such components.”30 These only model available at the time mantle all its graphite reactors. provisions entail a complication. (IAEA INFCIRC 153). Details of safeguards implementation for the Thus, the Supply Agreement con- Agreements for cooperation on reactors now in the DPRK appear templates reciprocal steps of per- peaceful uses of nuclear energy in a confidential “subsidiary formance to assure each side that with other nations are authorized arrangement” agreement with the the other is doing its part. An agree- by the US Atomic Energy Act IAEA. Details for LWR safeguards ment that has not been made public (Section 123). Before they can enter implementation will be negotiated provides more detail. And the into force, they must lay over in with the IAEA in another “sub- Supply Agreement calls for negotia- Congress for 90 days of continuous sidiary arrangement.” tion of still another protocol defin- session without passage of a joint ing the reciprocal steps further.25 resolution of opposition. No agree- 3. The AF goes beyond the basic ment for cooperation has yet been NPT obligation “not to

19 A BRIEF HISTORY OF THE DPRK’S NUCLEAR WEAPONS-RELATED EFFORTS

manufacture or otherwise or will award contracts to four 4. Mitchell Reiss, Bridled Ambition: acquire nuclear weapons.” For other South Korean companies for Why Countries Constrain Their example, in it the DPRK has nuclear components. Delays have Nuclear Capabilities, (Woodrow agreed to “take steps to imple- resulted because of difficulties on Wilson Center, Washington, ment the North–South both sides, and completion is now D.C., 1995), p. 239; Foreign Declaration on Denuclearization hoped for by 2007 rather than the Broadcast Information Service, of the Korean Peninsula.” This 2003 date targeted in the AF. FBIS-EAS-91-243 (December 18, contains provisions prohibiting Further delay is likely given the 1991), p. 14. either the DPRK or the ROK anticipated problems described. from enriching uranium or sepa- 5. Mazaar, pp. 84, 86. rating plutonium. The AF also Notes to Chapter 1 requires that the DPRK remain a 6. Mazaar, pp. 94-95; Reiss, pp. party to the NPT. Finally, as 1. David Albright, and Kevin 246–49. already indicated, the AF O’Neill, eds., Solving the North requires that the DPRK “come Korean Nuclear Puzzle (Institute 7. Reiss, pp. 250–53. into full compliance with its for Science and International safeguards agreement” before Security, Washington, D.C., 8. Joint Statement of the DPRK the delivery of “key nuclear 2000), Chapter 1, pp. 15–29.; and the USA, New York, June components.” Full compliance M.J. Mazaar, North Korea and the 11, 1993. includes taking all steps that Bomb (MacMillan, London, may be deemed necessary by the 1995), pp. 25, 29, and 36; V.I. 9. Joint Statement of the DPRK IAEA “with regard to verifying Denisov, “Nuclear Institutions and the USA, New York, June the accuracy and completeness and Organizations in North 11, 1993. of the DPRK’s initial report on Korea,” in The North Korean all nuclear material in the Nuclear Program, J.C. Moltz and 10. Agreed Statement between the DPRK.” This report is the one A.Y. Mansourov, eds. USA and the DPRK, Geneva, that the 1991-92 IAEA measure- (Routledge, New York, 2000), June 11, 1993. ments and environmental moni- pp. 21–26; A. Zhebin, “A toring suggested may be inaccu- Political History of 11. Agreed Statement between the rate because it does not report Soviet–North Korean Nuclear USA and the DPRK, Geneva, all the plutonium the IAEA sus- Cooperation,” in Moltz and June 11, 1993. pects the DPRK separated before Mansourov, pp. 27–34. it submitted the report. 12. Agreed Statement between the 2. D. Albright, F. Berkhout, and USA and the DPRK, Geneva, The AF and Supply Agreement W. Walker, Plutonium and August 12, 1993. are being implemented by KEDO Highly Enriched Uranium 1996: and the DPRK with considerable World Inventories, Capabilities, 13. Reiss, pp. 259, 265–71. success so far, though at a slower and Policies (Oxford University rate than the AF contemplated. The Press, New York, 1997), pp. 14. Mazaar, pp. 162–63; Reiss, pp. survey for the Kumho site has been 282–84; R.W. Jones and M.G. 271–72, p. 275; Leon Sigal, completed and major construction McDonough, Tracking Nuclear Disarming Strangers (Princeton to make roads and build buildings Proliferation (Carnegie University Press, Princeton, for the LWRs and the personnel Endowment, Washington, D.C., 1998), pp. 155–62. who will operate them has taken 1998), pp. 158–59. place. KEDO has awarded a “turn- 15. Agreed Framework, pars. I, 1 key contract” for the manufacture 3. G. Bunn, Arms Control by and I, 3. of the LWRs to KEPCO, a South Committee (Stanford University Korean company, which in turn has Press, Stanford, California, 16. Agreed Framework, par. I, 1. 1992), pp. 251–52. 17. Agreed Framework, par. I, 2.

20 A BRIEF HISTORY OF THE DPRK’S NUCLEAR WEAPONS-RELATED EFFORTS

18. Agreed Framework, par. IV. 22. Statement of Pierre 26. Supply Agreement, Art. I. Goldschmidt to William Sailor 19. The Supply Agreement refers at the International Nuclear 27. Supply Agreement, Art. II. for its definition of “key Materials Manangement nuclear components” to the (INMM) Annual Meeting in 28. Supply Agreement, Art. XIII “Trigger List” developed by the July 2000. Nuclear Suppliers’ Group to 29. Agreed Framework, Art. I, 1. implement the export controls 23. David Albright, required by Art. III, par. 2 of “Inconsistencies in North 30. Supply Agreement, Annex 3, the Non-Proliferation Treaty. Korea’s Declaration to the par. 4. See M.M. El Baradei, E.I. IAEA” and “How Much Nwogugu, and J.M Rames, The Plutonium Did North Korea 31. North Korean Threat Reduction International Law of Nuclear Produce?” Chaps. IV and VI in Act of 1999, Sect. 821-23 of P.L. Energy (Nijoff, Dordrecht, Albright and Kevin O’Neill, 103-113. See Henry Sokolski, Netherlands, 1993), pp. eds., Solving the North Korean “Implementing the DPRK 1,520–33. Nuclear Puzzle (Institute for Nuclear Deal: What US Law Science and International Requires,” Nonproliferation 20. Non-Proliferation Treaty, Art. Security, Washington, D.C., Review (Fall–Winter, 2000), pp. III, par. 1. 2000). 146, 148.

21. Supply Agreement, Art. III and 24. Supply Agreement, Annex 3. Annex 4. 25. Supply Agreement, Art. III, 3; Reiss, pp. 276–77.

21 A BRIEF HISTORY OF THE DPRK’S NUCLEAR WEAPONS-RELATED EFFORTS

22 CHAPTER 2 CURRENTLY APPLICABLE SAFEGUARDS AND RELATED AGREEMENTS

2.1 The Existing areas, what goes out, and what is control equipment, apply sur- IAEA–DPRK Safeguards there at the time of measurement.3 veillance and containment meas- Agreement ures (e.g., TV monitoring cam- For the IAEA staff to designate eras focused on the reactors and The IAEA–DPRK Safeguards these points and areas, decide seals on the reactors that, when Agreement is a standard agree- where sensors and cameras broken, show the reactor could ment based on the IAEA’s 1971 should be placed, and perhaps have been opened when the “model safeguards agreement,” even suggest minor changes in inspectors were not present). It INFCIRC 153.1 Among other design that would help in safe- can also use other unspecified things, it provides— guarding, the papers showing methods “which have been the design of the reactor should demonstrated to be technically 1. The DPRK (Democratic People’s be submitted to the IAEA well in feasible.”5 Republic of Korea, or North advance of the reactor’s installa- Korea) must establish and tion—although this IAEA sug- 5. The Safeguards Agreement calls maintain a system of accounting gestion is sometimes ignored by for one routine inspection per and control that will enable the reactor builders. year for small facilities and for International Atomic Energy material balance areas outside all Agency (IAEA) to verify the 3. The DPRK is required to provide facilities’ annual throughput of DPRK’s accounting for its the IAEA with two kinds of nuclear material not exceeding nuclear material at Yongbyon, accounting reports for each five kilograms. For facilities with Kumho, or any other location. It material balance area: “invento- an annual throughput exceeding must permit the IAEA to make ry change reports” and “material five kilograms, the “number, “independent measurements balance reports.”4 The inventory intensity, duration, timing, and and observations” to assure the change reports are to be sent to mode of routine inspections… “timely detection of diversion the IAEA within 30 days after shall be no more intensive than of significant quantities of the end of the month in which is necessary and sufficient to nuclear material from peaceful the change occurred. The materi- maintain continuity of knowl- nuclear activities to the manu- al balance reports are to show edge of the flow and inventory facture of nuclear weapons or of the balance based on a physical of nuclear material….” The other nuclear-explosive devices inventory of nuclear material in Agreement goes on to specify or for purposes unknown, and the material balance area and are standards for determining the deterrence of such diversion by due within 30 days after the “maximum routine inspection the risk of early detection.”2 inventory has been taken. More effort” in such cases.6 This is the standard for the details for the light-water reac- Presumably an IAEA–DPRK IAEA’s INFCIRC 153 safeguards tors (LWRs) should appear in negotiation will determine this agreements. IAEA–DPRK “subsidiary effort at a later time. arrangements” for the LWRs. 2. From the design of a proposed These arrangements have not yet At the time the IAEA–DPRK reactor, the IAEA will select “key been negotiated. Safeguards Agreement went into measurement points,” called effect (1992), the Agreed “strategic points,” to be used to 4. The IAEA has the right to con- Framework (AF) had not been measure the nature and quanti- duct “routine inspections” to negotiated and the safeguards ties of nuclear material, and to verify the information supplied contemplated in the Safeguards determine its flow and inventory. by the DPRK’s accounting sys- Agreement were for graphite- These strategic points will deter- tem. It can examine the DPRK’s moderated reactors. The changes mine the “material balance areas” accounting records, make inde- required will not be made in the where the DPRK (subject to pendent measurements, verify Safeguards Agreement itself but checks by IAEA inspectors) will the functioning and calibration in the subsidiary arrangements measure what comes into the of measuring instruments and for the LWRs.

23 CURRENTLY APPLICABLE SAFEGUARDS AND RELATED AGREEMENTS

6. Under the Safeguards requested in the past, was included records predating the entry into Agreement, the IAEA also has in INFCIRC 153, the model safe- force of an INFCIRC 153 safe- the right to conduct “special guards agreement upon which the guards agreement. (Given the inspections” if, for example, it DPRK’s agreement is based.8 length of time the DPRK operat- decides that the routine inspec- [Neither the DPRK nor the ed nuclear facilities before it tions and the information and Republic of Korea (ROK, or South agreed to its INFCIRC 153 safe- explanation provided by the Korea) has agreed to accept the guards agreement, this could DPRK “is not adequate for the new IAEA model safeguards proto- produce more IAEA questions.) Agency to fulfill its responsibili- col resulting from the 93+2 pro- ties under this Agreement.” As ceedings.] Some of the Part 1 infor- • Information on other locations summarized in Chapter 1, the mation that could be asked of the containing nuclear material. (This IAEA secretariat tried to institute DPRK for IAEA use in safeguard- should include the disputed, a special inspection in 1992 to ing the new reactors is as follows: undeclared sites.) determine if the DPRK had sepa- rated more plutonium than it • Responses to a detailed ques- The provisions of the AF and had reported. The IAEA may tionnaire showing the DPRK’s Supply Agreement (summarized in again seek special inspections system of accounting and con- Chapter 1) that deal with these when the time comes under the trol for nuclear material includ- issues are thus reinforced by the AF for the IAEA to inspect sites ing the scope and timing of the IAEA Board’s decision that safe- not declared as nuclear by the DPRK’s own inspections and guards agreements like the one DPRK to determine whether the related activities relevant to applying to the DPRK require the DPRK’s original declaration of its IAEA safeguards. DPRK to supply information of nuclear materials and facilities is this kind. Moreover, in adopting its accurate and complete. The IAEA • Information on past nuclear 93+2 decisions, the IAEA made is more likely to simply insist activities relevant to assessing clear that remote monitoring from that the DPRK comply with the the DPRK’s declarations of pres- a distant, safeguarded reactor to AF provision requiring it to take ent nuclear activities, including IAEA headquarters in Vienna “all steps that may be deemed the completeness and correctness using communications equipment necessary by the IAEA…with of its initial report on nuclear attached to cameras and other con- regard to verifying the accuracy material. (There is already a trol and surveillance devices and completeness of the DPRK’s major dispute between the installed at the reactor could be initial report on all nuclear mate- DPRK and the IAEA over requested by IAEA inspectors rial in the DPRK.” accounting for the nuclear mate- under an INFCIRC 153 safeguards rial in the DPRK’s initial report. agreement. Acceptance by the gov- 2.2 New Information That The IAEA suspects that some of ernment where the reactor was the IAEA May Ask of All the fuel rods, after radiation in located of a 93+2 strengthened pro- the small graphite reactor, were tocol for its safeguards agreement States Under INFCIRC 153 reprocessed to extract plutoni- was not necessary. Safeguards um, after which the plutonium and the wastes were hidden 2.3 Additions to INFCIRC As we have seen, “when a sig- from IAEA inspectors.) 153 Safeguards for States nificant portion of the LWR project is completed but before delivery of • Information on decommissioned Willing To Agree to a New key nuclear components,” the nuclear facilities, and on other INCFIRC 540 Safeguards DPRK must provide information locations previously containing Protocol showing where all the nuclear nuclear material that had hot material in its original inventory cells or where activities relating The IAEA “93+2” decision deals has gone.7 Under “Part 1” of the to conversion and reprocessing with subjects beyond the INFCIRC IAEA’s “93+2” decisions to of fuel fabrication took place. 153 safeguards agreements, sub- strengthen safeguards, the DPRK (This may produce IAEA ques- jects requiring negotiation of an could also be asked by the IAEA to tions on the DPRK radiochemi- amendment or “protocol” to exist- provide certain information about cal lab or on a smaller chemical ing agreements. In its 1996 report its past dealings with nuclear separation facility that the IAEA on 93+2 to a conference of its mem- material. In describing the 93+2 Director-General Blix suspected bers, the IAEA described the new safeguards requirements, the IAEA the DPRK once had.) information that would be required legal staff concluded, and the as part of what it called the “Part Board accepted, that Part 1 infor- • Access to existing historical 2” amendments to be incorporated mation, though not always accounting and operating in a new Protocol to the standard

24 CURRENTLY APPLICABLE SAFEGUARDS AND RELATED AGREEMENTS

INFCIRC 153 safeguards agree- pleteness of the DPRK’s initial language quoted in the two pre- ment.9 The model agreement for report on all nuclear material in the ceding paragraphs from the AF this new protocol is in IAEA INF- DPRK.” may be helpful in producing CIRC 540. The new protocol DPRK cooperation. requires, for example, information Even under an INFCIRC 540 not involving nuclear material protocol, however, a government One of the promises of provid- including: official can prevent IAEA inspec- ing INFCIRC 540 information is tors from going to places they that it will permit closer coopera- • nuclear research activities and have not inspected before. The tion between the IAEA, the opera- future plans for nuclear develop- main difference may in practice be tors, and the government officials ments; a difference in the “threshold responsible for nuclear activities. showing” necessary to justify This will result in many fewer • all activities related to the enrich- going beyond routine inspection inspections for countries like ment of uranium, reprocessing of sites. For special inspections of Canada or Japan where a tremen- spent fuel, or treatment of non-routine, undeclared sites dous share of the total IAEA nuclear waste before the activities under INFCIRC 153 safeguards inspections costs has gone in the involve nuclear material; when a government asks for justi- past. In general, adoption of INC- fication, inspectors must argue FIRC 540 protocol would have, it • activities in buildings on sites of that they simple can’t do their jobs was hoped, produced a major nuclear facilities; without the special inspection. For reduction in the number of inspec- visits to non-routine, undeclared tions in countries that accepted the • acquisition of a long list of sites under INFCIRC 540 safe- new safeguards. IAEA safeguards equipment and non-nuclear guards, when the host govern- experts question, however, material related to the operation ment asks why, the inspectors whether that is an objective worth of nuclear facilities; should only have to show that a pursuing with the DPRK. Some question or inconsistency has aris- believe that having more inspec- • information on materials con- en that necessitates an inspection tions in the DPRK pursuant to taining uranium or thorium too to resolve. In either case, the coop- DPRK’s existing Safeguards low in concentration to be eration of the host government is Agreement will be more valuable “nuclear material” within the likely to be the most important than attempting to apply the INC- IAEA definition. factor. Clearly, if the DPRK does FIRC 540 protocol to the DPRK.10 not cooperate by taking “all steps The INFCIRC 540 protocol’s that may be deemed necessary by 2.3.1 Safeguards on Existing requirements were also intended to the IAEA,” the language quoted at ROK LWRs—Models for the permit access to a greater range of the end of the preceding para- locations than under INFCIRC 153, graph, it will violate an important DPRK LWRs for example, of sites not declared requirement of the AF. by the facility’s government. These The DPRK is to receive two pres- had been available for inspection INFCIRC 540 protocols also per- surized LWRs of 1,000-MW(e) in the past only through “special mit “environmental monitoring” capacity each, based on a US inspections,” which produced at other locations than those Combustion Engineering Company much controversy between the declared by the operator. The design that has been refined by IAEA and the DPRK. If the IAEA IAEA has been doing “environ- South Korean reactor designers. interprets the “special inspection” mental monitoring” at some rou- Two reactors of this design are provision as originally intended tine inspection sites in other coun- operating in South Korea. They are and as interpreted in its 1993 case tries. Presumably, it will do this at both subject to INFCIRC 153 safe- involving the DPRK, inspections at the DPRK’s routine inspection guards and are good models for other than the declared reactor sites (if it has not already). The the safeguards anticipated for the sites should be possible with the hope of INFCIRC 540 is that this two DPRK LWRs. Safeguards on DPRK’s cooperation without form of monitoring will become large LWRs of 1,000-MW(e) power amending the DPRK’s safeguards much more common. But, the host or more do not vary a great deal agreement to add the INFCIRC 540 government’s cooperation will from one reactor to another.11 Thus, protocol requirements. Indeed, the still be necessary to permit the the IAEA safeguards on the KEDO AF requires that the DPRK take inspectors to go to non-routine reactors in Kumho will likely be “all steps that may be deemed nec- locations to take their samples— very similar to the safeguards on essary by the IAEA…with regard from the environment, the air, the the ROK reactors expected to be to verifying the accuracy and com- leaves, the ground, etc. Again, the the models for the KEDO reactors.

25 CURRENTLY APPLICABLE SAFEGUARDS AND RELATED AGREEMENTS

2.4 Major Challenges AF to accept comprehensive IAEA But, looking to the future, this Ahead to the inspections. If the DPRK refuses could become a major problem Implementation of the inspections before this obligation again. Agreed Framework comes into force, as it has so far, and if, based on IAEA expecta- 2.4.4 US–DPRK “Agreement for tions, the IAEA inspections and Cooperation” and IAEA Most of this report is dedicated review take two or more years, to the technical means for verify- major additional costs to KEDO Inspection of Undeclared ing that the plutonium produced for the reactor will be caused by Facilities in the new LWRs at Kumho is not the delay. The reactor is being built diverted to weapons purposes; under a fixed-cost contract that This problem was discussed in that the graphite reactors and assumes only a few months will be Chapter 1. Negotiating and imple- other nuclear facilities at needed for the IAEA’s inspection menting a US agreement for coop- Yongbyon and Taechon are dis- and review. eration with the DPRK permitting mantled as required by the AF; the export of necessary nuclear and that the spent fuel now at 2.4.2 Financing components by American manufac- Yongbyon is not diverted but is turers seems tightly tied to satisfy- delivered to KEDO when the AF South Korea carries the major ing the IAEA that the DPRK has and Supply Agreement require it burden of financing now. The US declared and reported all its nuclear to be. Other major challenges, has paid for most of the heavy fuel materials and facilities—that the however, lie ahead that can pre- oil delivered to the DPRK to fuel DPRK has not, for example, sepa- vent or delay implementation of its electricity supply, but has not rated more plutonium than it has the AF. paid for much else. Japan has declared. Even if the DPRK cooper- made major contributions to the ates fully in the IAEA special or 2.4.1 Completion of a already large costs of construction. other inspections, the process of “Significant Portion” of the First Australia and the European Union reconstructing what happened to all have made smaller ones. nuclear materials in the DPRK from Reactor in the ROK and Completing construction at Kumho the time it joined the NPT to the Reactor Buildings in the DPRK and building the reactors in North present is likely to take two to four Korea will be very costly, and dis- years. No Congressional review of This step has been delayed so putes continue over how much an agreement for cooperation and far by about four years for many each of the interested parties no provision of nuclear components reasons. There is potential for should pay. to the DPRK appears likely until more delay from the challenges this process is completed and the listed below, including the liability 2.4.3 Nuclear Liability IAEA is satisfied that the DPRK’s problem. Completion is not now reports on all its nuclear materials expected before 2004, a year after This applies, for example, to and facilities are accurate and com- the original target for completion potential liability if one of the plete. By one estimate, the IAEA of the entire project. After comple- LWRs has a major accident during would have to complete its inspec- tion of this “significant portion” operations, and some people tions and make a decision favorable but before “delivery of key nuclear receive high doses of radioactivity. to DPRK within about 30 months if components,” the DPRK must Congress has prohibited the US the export licensing of a US nuclear “come into full compliance with from agreeing to indemnify a US component or components is not to its safeguards agreement…with manufacturer that provides major delay completion of the Kumho regard to verifying the accuracy components for the DPRK reac- project under the present time and completeness” of its 1992 tors. As a result, General Electric schedule—which has already been report to the IAEA on “all nuclear has turned down an offer to pro- delayed several years beyond the material in the DPRK.” This major vide the turbine generators for the original the original 2003 target verification problem is discussed DPRK’s reactors. Two Japanese date.12 in Chapter 6. It holds the potential firms, who seem somewhat less for causing substantial further concerned about the liability prob- If the IAEA takes around 24 delay and expense. KEDO esti- lem, will now supply them. South months for its inspections and mates provide only a few months’ Korea wants KEDO, the European appraisal, the inspections should time for IAEA inspection after a Union, Japan, and the US to begin by mid-2001. But, under the significant portion of the first reac- assume this liability. Negotiations AF, the time has not yet come for tor is built in the ROK and the have not yet produced an agree- the IAEA inspections to begin, as DPRK becomes obligated by the ment to share this liability risk. Chapter 1 shows. Implementation

26 CURRENTLY APPLICABLE SAFEGUARDS AND RELATED AGREEMENTS of pertinent provisions of the AF and what has been done so far. Annex II, pp. 5–7. may well be delayed for years beyond the four it has already been Notes to Chapter 2 9. Report of the IAEA Director- delayed, unless the DPRK agrees to General to the General earlier inspections. And, of course, 1. The DPRK Safeguards Conference, Strengthening the if the IAEA inspections and Agreement is IAEA INFCIRC Effectiveness and Improving the appraisal do not produce the 403, May 1992. Efficiency of the Safeguards required result, implementation of System, GC(40)/17 (August 23, the AF pursuant to its present 2. DPRK Safeguards Agreement, 1996), Annex II, pp. 7–15. terms could end. In Chapter 8, Art. 7. after more complete discussions 10. Personal communication in and assessments of verification and 3. DPRK Safeguards Agreement, August 2000 from James safeguards procedures, we return Arts. 46, 98 (m) and (s). Larrimore, former IAEA safe- to the question of how various guards expert. Rich Hooper, the eventualities could affect verifica- 4. DPRK Safeguards Agreement, IAEA official most closely asso- tion and safeguards and vice versa. Art. 63 (b). ciated with leading the devel- opment of the 93+2 revisions, 2.4.5 Improving North Korea’s 5. DPRK Safeguards Agreement, believes these revisions should Electricity Distribution System Arts. 72, 74. be used in the DPRK if the DPRK’s consent to doing so can Before it will be safe to turn the 6. DPRK Safeguards Agreement, be obtained. Personal commu- LWRs on, major expensive Arts. 79, 80. nication in September 2000. improvements in the DPRK’s elec- trical distribution system are neces- 7. See Chap. 1, par. 3 and DPRK 11. Personal communication from sary. North Korea does not appear Safeguards Agreement, Arts. 8, James Larrimore, August 2000. to have the resources to pay for 49, 63–68. these improvements, and neither 8. See the report by the IAEA 12. Henry Sokolsky, KEDO nor any of its participants Director-General to the IAEA “Implementing the DPRK has agreed to pay for them. This General Conference, Nuclear Deal: What U.S. Law problem is discussed in Chapter 3, Strengthening the Effectiveness Requires,” Nonproliferation which also describes the are where and Improving the Efficiency of Review (Fall–Winter, 2000), the reactors for the DPRK will go the Safeguards System, p. 149. GC(40)/17 (August 23, 1996),

27 CURRENTLY APPLICABLE SAFEGUARDS AND RELATED AGREEMENTS

28 CHAPTER 3 THE KEDO REACTORS AND ASSOCIATED FACILITIES AND ACTIVITIES

3.1 Introduction A description of the KSNP reactors along the DPRK’s East Sea coast- (the portions relevant to the safe- line, a distance of slightly more This chapter includes a discus- guards program), a discussion of than 160 kilometers (100 miles) sion regarding the location of the the proposed safeguards measures, from the ROK’s border to the Korean Energy Development and possible additions to them are south, and about 288 kilometers Organization (KEDO) reactors in provided in Chapters 4 and 5. (180 miles) from the Russian bor- the Democratic People’s Republic Other chapters address issues relat- der to the northeast. The nearest of Korea (DPRK, or North Korea) ed to the smaller graphite-moder- town to the site is Kumho, and we as well as the location of other ated reactors installed or under have chosen to refer to the site as nuclear-power plants along the construction before the AF at the Kumho site so as to be consis- shores of the East Sea. The discus- Yongbyon and elsewhere. tent with other publications that sion is limited to the large-sized, also refer to the Kumho site.1 The power-producing reactors that 3.2 The KEDO Site location of the Sinpo village is KEDO will provide to the DPRK shown in Fig. 3-1. The Kumho site under the terms of the Agreed The KEDO reactors will be built is located in a rural area away from Framework (AF). Further items dis- on the shores of the East Sea, some- the main population centers cussed here are site-work progress time referred to as the Sea of Japan. around the Pyongyang area to the to date, proposed shipments of (We will use the term East Sea to west and around Congjin close to nuclear fuel into and out of the avoid the clash of perceptions as to the Russian border to the north. reactor site, transmission of the the proper name for that sea.) The The climate at the site is dry and generated electricity out of the site, KEDO reactors will be located 30 cooler than in Pyongyang or Seoul, and related Republic of Korea kilometers (19 miles) north of the though the seaside location moder- (ROK, or South Korea) energy- village of Sinpo about midway ates the winter temperatures. development issues. All of these Vladivostok Figure 3-1. General issues bear on our main topic of CHINA North Korea Liaoyuan Songhua Onsong Hu Toudao Yanji map of the DPRK. safeguards and verification and RUSSIA 0 50 kilometers A what will be known under various H 0 50 miles S Unggi I Najin circumstances. A B

Fushun G Puryong Shenyang Hunjiang N Tonghua A Chongjin H The proposed power reactors to Yalu C Hun Chasong Hyesan be built at the Kumho site are often K E jin A g n Kapsan referred to interchangeably as the a h M Kanggye C Kilchu Kuandian Shuifeng Sk Chosan N KEDO reactors (for their provider A Changjin Songjin S Pujon Maokui shan Res. Res. organization), the Korean Standard Pukch’ong Dandong M Changjin

I Kumho Site Nuclear Plant (KSNP) reactors (for Sinuiju Huichon Taechon N Hamhung Sinpo Yongbyong G Hungnam their model name), or the Kumho N

Anju A Yonghung Pressurized-Water reactors (PWRs) N K orea Sinmi Bay NOR TH Choson for their general design category. KOREA Man We attempt to use these acronyms P’yongyang Wonsan Namp’o Nam where they fit best, but we remind Kosong Sariwon EAST SEA Imjin the reader that they refer to the (SEA OF JAPAN) Haeju Chorwon same two PWR-type reactors to be Kaeson Yanggu built at Kumho under KEDO’s Chunchon Kangyang sponsorship based on the KSNP Ha ng Ullung do Inch’on Seoul a n Samchok model. We remind the reader also g that PWRs, as well as Boiling- Chingju YELLOW SOUTH KOREA Water Reactors (BWRs) are types of SEA Chongju Andong reactors within the general catego- Anmyon-do Yondok Sangju ry of Light-Water Reactors (LWRs).

29 The KEDO Reactors and Associated Facilities and Activities

The Kumho site had already has inherited the original Russian • A hard-rock site with adequate been selected for three Russian plant site as this was the only loca- bedrock area at grade level, thus 1,000-MW(e) VVER-1000 reactors tion in the DPRK that had been sur- allowing the entire reactor build- and had been cleared as a veyed for the construction of a ing to be constructed above prospective nuclear-power plant large civilian power plant. grade as the Russian design site. Russia (then the Soviet required. Union) had promised those reac- There were several advantages tors to the DPRK as a part of the for originally choosing the Kumho • Mid-way location between the international nuclear electrifica- site as a prospective nuclear site: DPRK load center at the tion program of the Comecon Pyongyang area and the Russian Organization and in the expecta- • A seaside site that can use sea- load center at Vladivostok. tion of transmitting a part of the water for condenser cooling. The electric output to the Vladivostok seaside location and availability On the other hand, the distance area. In fact, two similar reactors of barge-docking facilities at from the load centers requires the (though of a more updated Sinpo and on site permit the construction of high-voltage trans- design) are now being constructed shipment of heavy equipment mission lines to transmit the large in the People Republic of China items as well as fresh and spent amount of electricity generated on (PRC), in Tinwan, Jiangsu fuel, rather than relying on the site to the ultimate consumers. Province.2 Early site work for the dilapidated road and rail net- Russian reactors project in Kumho works. 3.3 Nuclear-Power began in 1990 and ceased with the Development Along the collapse of the Soviet Union and in • Location in a low-density rural face of the DPRK’s apparent inabil- area, thus minimizing potential Shores of the East Sea ity to pay the cost of this large proj- routine or accidental radiation ect with its own resources. KEDO exposure. The KEDO reactors at the Kumho site will join the large num- ber of other nuclear facilities locat- ed along the shores of the East Sea. The four littoral states—the ROK, DPRK, Russia, and Japan—have Tomari Tsuruga turned the East Sea into one of the Fugen most heavily ‘nuclearized’ areas of 3 Monju the world. The ROK and Japan Mihama Higashidori have located several large plant clusters along the shores of the East Takahama Ohi Sea (Fig. 3-2). The ROK has located Onagawa three of its four nuclear sites on the eastern part of the peninsula, start- Fukushima I ing with Kori—the first nuclear- Kashiwazaki- Fukushima II power station of 2,940 MW(e) net Kariwa Tokai capacity—extending to Wolsung, the Korean–Canadian CANDU Shika reactors station of 2,800 MW(e) Sinpo East Sea (Sea of Japan) capacity, and to the Ulchin station Taechon Tokyo Yongbyon with 3,900 MW(e) installed capacity DMZ Shimane that includes the first of the series Pyongyong Hamaoka of the standardized KSNP plants, Seoul Ulchin two of which are planned for the KEDO site. Two KSNP reactors, Taejon Wolsung Pacific Ocean each one similar in size to the Yellow Sea KEDO station, are planned for the Kori Operation Construction Yeonggwang Kori and the Ulchin sites. Genkai Ikata Symbol Additionally, a new nuclear site Korea 16 4 + 2 will be opened up in Bonj I1, just Japan 52 3 Sendai north of the Wolsung station. That site could accommodate four KSNP Figure 3-2. Locations of nuclear reactors around the East Sea.

30 The KEDO Reactors and Associated Facilities and Activities reactors or four advanced CANDU national fuel-cycle center in power is to grow at the announced reactors, and has been re-designat- Rokkasho Mura, located at the levels and provide the energy ben- ed in early 2001 as a KSNP site.4,5 northern tip of the main island of efits promised. Honshu along the bay connecting Should relations between the the East Sea and the Pacific Ocean, 3.4 Site Work to Date DPRK–ROK continue improving— includes a uranium-enrichment and should the ROK’s nuclear pro- plant, a French-designed fuel-repro- KEDO initiated preliminary gram suffer from lack of nuclear cessing plant of 800 tonnes per year work at the Kumho site shortly sites, as discussed later—then it is capacity now under construction, a after its inception. Major work at possible to assume that additional large-sized spent-fuel storage pool, the site was initiated only after the ROK reactors will be built on DPRK and the prospective Higashidori signing of the Preliminary Works sites, probably starting at the nuclear-power plant site. The Contract (PWC) between KEDO Kumho site, to generate power for Kashiwazaki-Kariwa nuclear-power and Korea Electric Power transmission to the ROK. This will station, with five 1,100-MW(e) Corporation (KEPCO) in August result eventually in additional BWR-5 reactors and two 1,350- 1997. The PWC has been incorpo- nuclear stations located on the DPRK MW(e) advanced BWRs, is the rated into the main Turn Key section of the East Sea coastline. largest nuclear station in the world Contract (TKC) between KEDO with more than 8,000 MW(e) of and KEPCO, signed in December The Russian Navy has major installed capacity. The Japan Atomic 1999, and it is now included in the nuclear facilities southeast of the Power Company (JAPCO) has new overall KEPCO scope for the Vladivostok area, including prototypes station in Tsuruga that KEDO reactors. The full value of Chazhma Bay (nuclear submarine includes, among others, the 150- the PWC site-related work scope is refueling and defueling), Bolshoy MW(e) plutonium-burning Fugen estimated at $93 million, out of a Kamen (defueling and dismantel- Advanced Thermal Reactor (ATR) total KEPCO work scope under the ment of nuclear submarines), and and the 250-MW(e) Monju experi- TKC, estimated at $4,080 million. Pavlovsk Bay (operational subma- mental Liquid Metal Fast Breeder Information on the work progress rine base), all on the shores of the Reactor (LMFBR). And lastly, three to date can be obtained from the East Sea. A large part of the nuclear- PWR stations of the Kansai Electric KEDO Annual Reports.6 As part of powered surface ships and sub- Power Company are clustered at the its activities, KEDO has opened an marines of Russian Navy’s Pacific Mihama, Takahama, and Ohi sites office in Kumho to act as a con- Fleet are home-ported at these har- with a total installed net capacity sular office for staff from KEDO bors. Naval facilities in this area exceeding 9,200 MW(e). member countries without diplo- include storage of spent fuel and matic relations with the DPRK, and low-level waste from Pacific Fleet The profusion of nuclear-power to interact with the DPRK General ships and submarines. Spent-fuel stations and fuel-cycle facilities Bureau (GB) for the LWR Project on and radioactive-waste storage sites located on the shores of the East Sea, project- and site-related issues. are near Dunai on the Shkotovo to which the KEDO reactors will be KEDO also cooperates with the Peninsula, and a floating liquid- the latest addition, may well require DPRK regulatory authority—the radioactive-waste filtration plant the littoral states to reach agree- State Nuclear Safety Regulatory (the Landysh) at Bolshoy Kamen. ments about regulating water efflu- Commission (SNSRC) on matters All these facilities need upgrading ents and air emissions from these related to the safety evaluation of to Western radiation-exposure and facilities. Such agreements may well the KSNP reactors and to the train- emissions-control standards. The have to include provisions for emis- ing of nuclear-plant operators and Russians have long-term plans for sions monitoring, effluent control maintenance crews. the construction of the Primorski standards and cross boundary pol- Kray nuclear-power station north of lutant damages, and third-party lia- Site activities to date have concen- Vladivostok, and discussions have bility in cases of nuclear accidents. trated on three areas (Fig. 3-3), been held since 1995 with both In fact, it may be to the advantage of including the construction housing Russian and Canadian organizations these states to harmonize and and warehouses area, the reactors to that purpose. improve their nuclear-insurance reg- site, and the process-water intake ulations to meet current standards, site, some 16 kilometers away from The extensive, peaceful nuclear- at least at the levels proposed by the the reactors site.7 KEDO has also power program in Japan has result- International Atomic Energy interconnected all three sites with a ed in a large number of nuclear sites Agency (IAEA). Implementing road system considered among the located on the shores of the East Sea. such measures will be essential for best in that region of the DPRK. The Among the more important facili- the well-being of the citizens of all housing area now includes facilities ties, we should mention four. The the four states involved, if nuclear for several hundred construction

31 The KEDO Reactors and Associated Facilities and Activities workers and visitors. It includes medical, dining, banking, and recre- ational facilities. Construction and design offices were also built, as well as warehouses for the construc- Power plant construction area tion equipment and supply items. (3,108,000 m2) KEDO has also established inde- Aggregate/water source area pendent supplies of reliable electric- (1,654,000 m2) ity, potable water, communications system, and constructed environ- Community area mental monitoring facilities and a sanitary-waste treatment facility.

A site view7 of the future KEDO reactors is shown in Fig. 3-4. In terms of work to date, KEPCO under the PWC has concentrated Seohochon on site-grading and removing a Yanghwa Port large hill from the area where the reactors are to be constructed. The Figure 3-3. The vicinity of the Light-Water Reactor (LWR) project. scope of the grading effort is graphically depicted in Fig. 3-5 where the dimensions of the grad- ed area are shown in relation to the size of the future KSNPs.7 In total, about 4 million cubic meters of rock and soil have been removed down to the bedrock. The exposed bedrock at grade level now forms the ‘platform’ over which the base mats for the two reactor plants will be laid. The excavated rock materi- als are used to create the cooling- water intake and discharge struc- tures, and to build the breakwater for the docking harbor on site, which will be used to barge-trans- port heavy equipment items to the construction site.

With the signing of the major TKC, the direct responsibility for the KSNP reactor construction has devolved to KEPCO as the KEDO General Contractor. KEDO itself continues to negotiate with the Figure 3-4. Artist’s conception of the future site for the KEDO reactors. From a DPRK authorities various other KEPCO–KEDO brochure supplied by J. Mulligan, KEDO, June 2000. contracts that will govern the con- struction work and the reactors’ commissioning and operations signed. Substantial progress was 2001. KEDO is discussing with the phases. A Memorandum of achieved by July 2000 on a Protocol DPRK the strengthening of its reg- Understanding was signed with dealing with Quality Assurance ulatory agency—the SNSRC. As a the DPRK on September 1999 and LWR Equipment Warranties. part of its extensive safety-related regarding environmental protec- KEDO has negotiated with the discussions with the DPRK, KEDO tion and indemnification. An IAEA to conduct a Standard has transferred to the DPRK copies Operators’ Training Protocol was Design Safety Review of the pro- of the ROK regulatory agencies’ completed on April 2000 and is posed Kumho reactors by mid- Safety Review Guidelines, the ROK

32 The KEDO Reactors and Associated Facilities and Activities

Atomic Energy Act and Enforcement Decrees, and a number of codes and standards adopted in the ROK’s nuclear energy program. KEDO has also conducted detailed discussions and seminars with the DPRK SNSRC and GB regarding the application and implementation of those documents. A Preliminary Safety Analysis Report (PSAR) for the KEDO reactors has been submit- ted to the DPRK regulatory agency as a part of the application for a construction permit. 3.5 Nuclear-Fuel Shipments Into and Out of the Site Fresh nuclear fuel to the KEDO Master Site Grading Plan reactors and spent nuclear fuel dis- charged from the reactors will be Cross Sectional View of the Site shipped in and out of the Kumho site by barge to the ROK because road and rail links between the DPRK and the ROK have not yet been established. A trial opening of the Seoul– Pyongyang rail line occurred in October 2000; however, commercial rail service between the two countries is speculative. Furthermore, the status of road and rail overpasses, underpasses, and junctions that need to be traversed, as well as the general condition of the transportation routes may not allow bulky, overweight spent-fuel (and even the lighter fresh fuel) transporters. A heavy spent-fuel- cask rail car may weigh about 100 tons. It is not clear that the DPRK rail network could safely carry cars Master Site Grading Plan of this weight. Should a transporta- • Power Block Area tion accident happen due to the – Graded elevation: EL 10.0 m (above MSL) inadequate maintenance of the – Graded area: about 220,000 m2 road or rail links and various junc- • Temporary Facilities Area tions along the routes, the accident – Graded elevation: EL 6.0 m (above MSL) consequences may be severe, and – Graded area: about 780,000 m2 the fuel insurers may not even • Total Volume/Area allow such transportation. – About 6,000,000 m2/1,000,000 m2 From a non-proliferation Figure 3-5. Scope of the grading efforts in the construction area (according to a KEDO perspective, once spent fuel is brochure). moved out of the Kumho site, the shorter the time period as it moves through DPRK territory, the less chance for attempted fuel diver- sion, or for delaying the spent-fuel

33 The KEDO Reactors and Associated Facilities and Activities transport and using it as a bargain- Taejon (see Fig. 3-2). KNFC fabri- At this point, KEPCO has not yet ing chip in some future negotia- cates all the fuel assemblies for the developed its long-range, spent- tions. From all the perspectives nuclear plants in Taejon, and it fuel management plan for its mentioned above, moving the fuel will likely fabricate the fuel assem- KSNP reactors. KEPCO may well in and out of the Kumho site by blies for the KEDO reactors just as decide to keep all old spent KSNP barge is the feasible and desirable it fabricates fuel assemblies for all fuel discharge in concrete dry-stor- transportation mode. other KSNP reactors. From Taejon, age casks at each site, rather than the fabricated fuel assemblies will remove the spent fuel to a central As discussed in Chapter 4, each be trucked onto a barge docked in storage facility. At this time, no of the two KSNP reactors on site at the designated ROK harbor and ROK-centralized, away-from-reac- Kumho will be refueled every 15 the barge will carry the nuclear- tor, spent-fuel storage facility has months, as is the practice with all fuel shipment to the Kumho barge been developed and built by the other KSNP reactors operated harbor, whence the fresh fuel will KEPCO. Thus, the DPRK authori- by KEPCO. In general, it is be tracked to the reactor fuel ties may well decide, all other KEDO’s intention, as much as pos- buildings. The DPRK operator things being equal, to keep the sible, to build the Kumho reactors will take custody of the fresh fuel KEDO reactors’ old discharged fuel like all other KSNP reactors operat- at the barge-docking facility and in the spent fuel pools in the reac- ed in the ROK. This issue is dis- move it to the fuel-storage facility tor buildings and then in dry stor- cussed next. in the plant. age casks on site at Kumho, until KEPCO’s spent-fuel disposal plans A typical KSNP reactor requires Unlike the fresh-fuel-shipment are firmed, assuming the spent fuel 60 new assemblies of fresh fuel process, which is quite straightfor- will be disposed of in the ROK. loaded into the core every 15 ward, the issue of spent-fuel ship- Alternately, if by mutual agree- months. In parallel, 60 assemblies ment is more complicated. Spent ment between the DPRK and other of spent fuel are discharged from fuel discharged from the reactor is countries involved, the spent fuel the reactor and stored in the spent highly radioactive. Current industry is to be removed from the Kumho fuel storage pool, located in the practices call for keeping the spent- site as quickly as feasible, a storage fuel building adjacent to the reactor fuel assemblies in the spent fuel facility in the ROK or a third coun- containment building. Each assem- storage pool in the fuel building on try has to be so designated, and, bly weighs slightly more than half site until their residual radioactivity assuming it is a new storage site, a ton (1,000 pounds). A fresh fuel has been sufficiently reduced that facility has to be designed, load then weighs less than 35 tons. through the natural decay of the licensed, and built. That process The fresh-fuel assemblies are only short-lived fission products before may require 20 years or more due slightly radioactive and can be han- shipping the assemblies in a heavily to potential public opposition to dled by hand, with most of the shielded container. Usually, spent storing spent fuel from a different alpha activity of the low-enriched fuel is kept in wet storage on site country. If the KEDO reactors’ fuel uranium (LEU) being contained in for at least ten years. Most spent is shipped to an existing storage the zirconium oxide-clad material. fuel storage pools can be further re- facility, e.g., in the UK, France, or Thus, the fresh fuel shipment into racked and equipped with neutron the PRC, high storage fees may be the Kumho site does not pose any absorber plates to increase their imposed, assuming the spent fuel transportation difficulties and can storage capacity to more than 20 will not be reprocessed. Thus, the be carried out by two commercial years’ worth of discharge, while fate of the KEDO reactors’ spent trucks. avoiding potential nuclear criticality fuel is yet quite uncertain, though accidents. Under these conditions, adequate time remains before the It is likely, though not yet estab- and assuming the KEDO reactors spent-fuel storage pools on site run lished, that the DPRK will pur- reach commercial operation in 2008, out of storage capacity. chase its natural uranium in the it may not be necessary to ship world markets and ship it to one spent fuel out of the Kumho site Once a decision is made to of the major uranium enrichment prior to 2020, unless so required by remove the spent fuel from the vendors, such as the United States non-proliferation considerations or Kumho site, it is necessary to bring Enrichment Corporation (USEC), by special contract conditions. It is a heavily shielded storage cask to for enrichment. The LEU produced more likely that with proper stan- the site. Such a cask, depending on by USEC will be converted to ura- dardized re-racking measures, the the number of spent fuel assem- nium oxide in the US and then KEDO reactors’ spent fuel will not blies it carries, can weigh up to shipped to the ROK’s fuel-fabrica- have to be moved out of the storage 100 tons and is usually rail-mount- tion plant, operated by KEPCO pools in the reactor’s fuel buildings ed. The cask has to be barge- Nuclear Fuel Company (KNFC) in prior to 2030. shipped into the site and then

34 The KEDO Reactors and Associated Facilities and Activities moved by rail from the barge har- when both reactors reach commer- Of these three, Russia is in most bor on site to the fuel building at cial operation. In fact, the total gen- immediate need for power import each reactor. The spent fuel is eration from both reactors by 2010 due to severe energy shortage in loaded into the cask in a special will nearly equal the total com- the Vladivostok and the Primorsk section of the spent fuel pool inside mercial electricity generation in the regions. Russia may, however, lack the reactor building (discussed in DPRK in 2000.9 Hayes’ low esti- the hard currency required for Chapter 4). The loaded and sealed mate of the DPRK total generation such energy export. Furthermore, cask is then lowered into its rail may itself be an overestimation. It the transmission links from the carriage and moved into the barge, was reported in December 2000 Kumho site to the Russian border which carries it to its destination in that the DPRK has demanded are the longest, thus requiring the ROK or beyond. There, the direct electricity supply from the more investments in rebuilding spent fuel is removed and the cask ROK as a pre-condition to a contin- and upgrading the DPRK trans- readied for another shipment. The ued North–South dialog.10 This mission network. Nearer-term logistics and the financial arrange- would indicate that available fuel interconnections may include ment for spent-fuel disposition stockpiles at the DPRK’s power ROK–DPRK, PRC–DPRK or have not yet been made. plants are lower than expected. ROK–DPRK–PRC links. The DPRK may barely be able to 3.6 Electricity Transmission provide on its own even its most Shipping the KEDO reactors’ Issues essential power supplies. This situ- power to market via the ation, while increasing the leverage ROK–DPRK transmission link has As important as the fuel ship- of the US and the ROK in dealing been costed and is the most techni- ment into the site is the issue of with the DPRK, will also create sev- cally feasible and potentially near- electricity transmission out of the eral problems, the four most impor- term option. Even now there exist site. The US Energy Information tant ones being the disposition of limited diurnal and seasonal elec- Administration estimates the total the generated electricity, network tricity interchanges between the electricity-generating capacity of stability, reliability of power trans- ROK and the DPRK. Transmitting the DPRK by January 1998 as mission, and the supply of high- the KEDO reactors output to the 10,000 MW(e),8 with a total genera- quality power in-house. ROK would expand the existing ad tion in 1998 of 32.0 billion KWh. hoc arrangements several-fold. Hydroelectric plants provide 50 The electricity generation from Given the high-growth rate of elec- percent of the installed capacity the KEDO reactors beyond 2010 tric demand in the ROK—now that and 70 percent of generation will likely significantly exceed the the country’s economy is emerging nation-wide, and the rest is provid- total demand of the DPRK, assum- from the 1998 financial crisis—hav- ed by a mix of coal- and oil-fired ing other existing generating ing access to the KEDO reactors will thermal power plants. Peter Hayes plants are brought again into com- defer a new power-plant project by of the Nautilus Institute9 provides a mercial operation once economic several years, which will be benefi- similar estimate for total installed recovery is underway. Eventually, cial considering the paucity of new capacity—9.75 GWe. According to the DPRK will have to earn hard plant sites available in the ROK. Hayes, the DPRK energy crisis of currency to pay its obligations to The DPRK will benefit from the 1996 resulted in only 2-3 GWe of the KEDO participants, principally energy payments received. KEPCO installed capacity in operation by the ROK and Japan, and return at has by now evaluated the feasibility the year 2000 due to the lack of least a portion of their investments and the cost of creating such trans- commercial fuel, cancellation of in the Kumho reactors, and to pro- mission interlinks and will most subsidized oil supply from Russia, vide a cash-flow stream to support likely be ready to implement such a and drought seasons. Total genera- plant operations and maintenance. project once international financing tion by 2000 is estimated at only Both supply-and-demand consid- is available and the political 15.0 billion KWh. The KEDO reac- erations, as well as hard-currency approvals are obtained. The main tors will inject a 2,000-MW(e) requirements, imply that a large problem with this concept is the capacity increment into the DPRK’s portion of the generated power political dimension. Connecting the power grid, with each reactor gen- will have to be dedicated for ROK and the DPRK transmission erating about 7.0 billion KWh per exports out of the DPRK. The only networks will require integrated year (assuming annual operation at nearby countries that can take the operation of both networks, most 80 percent capacity). Evidently, the KEDO reactors output are the likely under the technical control of KEDO project will become a major ROK, the PRC, or Russia, though the more advanced ROK side. This generating center in the DPRK, pro- an underwater cable in the East dependency on uninterruptible viding about a third of all genera- Sea that connects the KEDO reac- transmission flows could lead to a tion from the Kumho site by 2010, tors to Japan is being considered. loss of political freedom of action

35 The KEDO Reactors and Associated Facilities and Activities unacceptable to the DPRK leader- concerns relate to the stability of the assuring its stability after introduc- ship. Thus, technical and political existing DPRK transmission grid, ing the large-sized KEDO reactors feasibility considerations seem to the hurdles of interconnection to into their national grid. While these operate in different directions as other grids, and the interaction of issues can be resolved using related to the large-scale the grid with the KEDO reactors. advanced electrical engineering ROK–DPRK transmission link. In equipment and software tools, these December 2000, the DPRK request- Network stability problems are not the problems that the DPRK ed electricity supplies from the ROK. occur when a large-capacity gener- planners have encountered to date, Should this prove politically and ation plant or load center appears and they may face many new chal- technologically feasible, it will in the transmission network, thus lenges, including some possible resolve part of the problems creating a large electricity source in reactor-safety issues. In this context, involved in eventually transmitting the network, which if lost due to a it is noteworthy that in early 2001, electricity from the KEDO reactors technical problem, may result in several large European engineering out of the DPRK. high electric currents in the grid companies such as Asea Braun that may exceed its carrying capac- Boveri Group (ABB) of Switzerland Another option would be to con- ity. As a general rule, a single gen- and Siemens AG of Germany nect the KEDO reactors to the eration node should not exceed 7 expressed interest in supplying elec- PRC’s Northeast Power Network. to 10 percent of the total system tric transmission and distribution Such an arrangement would bring generation for stability and reliabil- equipment to the DPRK in order to additional power supply to the fast- ity. Because the KEDO generating upgrade the entire country trans- expanding PRC energy market and plant will exceed the limits of this mission grid.11 These initiatives would defer the need to construct network stability–reliability rule, it were launched at the behest of the new, mainly coal-fired plants in the will be necessary to transmit European Union, which is interest- northeastern provinces by several power out of the site through sev- ed in upgrading its involvement in years. Deferring additional coal- eral high-voltage transmission resolving the Korean Peninsula dis- fired capacity would reduce region- lines that overlay the current lower pute. It is not clear, however, who al air pollution and may qualify the voltage transmission system and will pay for these grid upgrades, PRC for carbon-emissions reduction interconnect to it at several differ- and how will the DPRK earn the credits under the Clean ent nodes through step-down hard currency to pay for the new Development Mechanism of the transformers. Furthermore, electric transmission equipment. Kyoto Protocol. The major problem because the DPRK’s total generat- with the DPRK–PRC transmission ing capacity once the KEDO reac- As noted, the high-voltage link is the longer transmission lines tors are in operation will exceed (345 kV or 500 kV) transmission required from the Kumho site to the demand for years to come, it may lines out of the Kumho site will Chinese border—500 kilometers as be necessary to back down (or, in form a separate network that will compared with the transmission fact, defer restarting) some of the overlay the existing mainly 220-kV lines from the site to the ROK— existing fossil thermal plants to transmission grid and interconnect about 200 kilometers.9 make room for the newer, more with it near load centers. The efficient, and more cost-effective DPRK’s electric planners favor the A more stable, long-term solution nuclear plants. 500-kV lines, while the backbone of might be a three-party transmission the much larger ROK transmission interlink connecting the ROK net- Nuclear plants are generally grid is the 345-kV high-voltage work to the DPRK network at the designed to operate at base load lines. Assuming each line from the Kumho site and then further con- and are not well optimized to meet site carries the electrical output of necting to the PRC’s Northeast a fluctuating load. Thus, the one KSNP reactor, then at least two Power Network. Such an arrange- DPRK’s electric system planners high-voltage lines have to be con- ment could benefit all three parties, may choose to utilize their hydro- structed. From reliability and but connecting the currently decrepit electric plants at full capacity dur- redundancy perspectives, three DPRK transmission network to ing the good hydro months and lines will take care of potential out- another country, let alone integrating their nuclear plants as base load ages in either of the two other it to two different networks, would units. However, some of the ther- high-voltage lines or their step-up be difficult. Before discussing region- mal plants cannot shut down due to and step-down transformers. The al transmission schemes, it is neces- local network stability problems in three lines option is more expen- sary to consider some of the practi- sections of the transmission net- sive than the two lines option, both cal concerns related to integrating work. The DPRK’s system planners in terms of direct investments, the KEDO reactors into the DPRK’s will have an intricate job of balanc- right-of-way requirements, and transmission network. These ing the transmission network and line maintenance expenses, though

36 The KEDO Reactors and Associated Facilities and Activities it avoids the need for using other, “reactor completion” means the high quality of the external power less optimal reliability measures. point in time when the reactors can supply (minimizing voltage and generate as per the KEDO Supply frequency fluctuations) is the chal- The DPRK does not have the Agreement, or whether the com- lenge here. Should external power resources to finance construction of pletion date is the time when the within the acceptable voltage and these high-voltage lines. Financing plant’s electric output reaches the frequency ranges not be provided, has to be provided by international ultimate consumers. KEDO may hypothetically have to lending institutions or from contri- install its own power supply on butions by third parties. Extending The cost and completion date site in the form of a gas turbine, the high-voltage transmission lines issues become particularly pointed additional emergency diesel gener- to the ROK or PRC grids, as dis- when the need for the KEDO reac- ators, or solid-state rectifiers, or cussed above, could bring in the tors to receive external power from AC–DC–AC converters to improve hard currency required to pay for the grid is considered. The two the shape and quality of the grid- the transmission lines, as well to KEDO reactors consume in-house supplied power. It is not clear that pay back the investment costs in about 60 MW(e) per unit, or a total these options will be acceptable to the reactors, which the DPRK is of 120 MW(e). This load is required the safety regulator. All these obligated to do. to operate the four large primary options would further increase the coolant pumps in each reactor, the cost of the project and might delay Another problem affecting the feedwater pumps, and all other completion date, with associated integration of the KEDO reactors lighting, air conditioning, and vari- verification consequences. KEDO, into the DPRK grid as well as any ous equipment. The in-house power however, will not ship nuclear fuel interconnection with the ROK or supply needs to be high quality and to the Kumho site unless accept- the PRC grids is that the DPRK stable so as not to interrupt, slow able sources of off-site power are network operates at an effective down, or shut down any continu- established. frequency of 50 Hz, whereas both ously operated, electrically driven the ROK and PRC networks oper- equipment. If, due to grid problems 3.7 The ROK’s Energy ate at 60 Hz. The Kumho area unrelated to the reactors, the net- Development Issues transmission network operates at a work voltage or frequency sags 60-Hz frequency, while other below pre-determined safety mar- The KEDO project is of value to regions of the DPRK’s grid operate gins, some electrical apparatus may the ROK for obvious non-prolifera- at 50 Hz. The KEDO reactors, need to be shut down. This could tion and political considerations, being similar to other ROK KSNP start a sequence of events leading to discussed elsewhere in this report. reactors, are designed to operate at a full-plant shutdown, further exac- Additional energy development 60 Hz. In addition, both voltage erbating the initial grid problem. In issues, however, provide the ROK and frequency on the DPRK grid the extreme, disruption of off-site with further justification for the fluctuate around the nominal val- power could, if no other counter- project. These issues are briefly ues at wider amplitudes than are measures are available, lead to vari- reviewed here because they could acceptable from stability considera- ous nuclear accident chains. In the affect judgments regarding timing tions in the more modern and more likely event, a prompt shut- and verification issues. mature ROK and PRC grids. At the down of a reactor due to inadequate interlink points, it may be required off-site power supply could lead to First, the ROK is embarked on a to install AC–DC–AC converters to a major electric grid operations dis- significant nuclear-power expan- eliminate voltage and frequency ruption, and, in the extreme, to a sion program, but there are not fluctuations in the DPRK grid. large-scale network collapse. enough sites available in the ROK Similar measures may be needed in which to build the proposed when providing in-house power It is evident that just as the new number of nuclear power plants. from the DPRK grid to the KEDO plant affects grid operations and Table 3-112 includes a list of poten- reactors. While these problems can stability, so does grid stability tial new ROK plants, several of be resolved by standard electric affect the operation of the new which do not yet have a site desig- transmission equipment, these will plant. The DPRK has agreed to nation.* It may make sense, with increase the cost of the entire AF provide 220-kV commissioning the general warming of relations project. These difficulties could be power from the DPRK’s 220-kV between the ROK and the DPRK, relevant to the question whether network. Maintaining the required and as a part of a long-term plan

* The KSNP program now extends over the Yeonggwang 3 and 4 plants—the pre-series, Ulchin 3 and 4 plants—the reference units, Yeonggwang 5 and 6, Ulchin 5 and 6, the New (Shin)-Kori 1 and 2, and the KEDO reactors. Further nuclear-plant construction in the ROK may be based on the newer Korea Next-Generation Reactor plants for which no sites have yet been designated. While two (or four) more units may ultimately be built in the New Kori site, and the New (Shin)-Wolsung site may contain four new KSNP or KNGR-type reactors, no other nuclear sites have been dedicated in the ROK.

37 The KEDO Reactors and Associated Facilities and Activities

Table 3-1. Status of nuclear power plants in South Korea (l = in operation, w = under construction, m = in planning). Capacity Commercial Plant Unit Rx.Type (Mwe) Reactor Manufacture TG Operation Remarks Kori Unit 1 PWR 587 WH GEC 1978.4 G Turnkey Kori Unit 2 PWR 650 WH GEC 1983.7 G Turnkey Kori Unit 3 PWR 950 WH GEC 1985.9 G Non-Turnkey Kori Unit 4 PWR 950 WH GEC 1986.4 G Non-Turnkey Yeonggwang Unit 1 PWR 950 WH WH 1986.8 G Non-Turnkey Yeonggwang Unit 2 PWR 950 WH WH 1987.6 G Non-Turnkey Yeonggwang Unit 3 PWR 1,000 KHIC/CE KHIC/GE 1995.3 G Non-Turnkey Yeonggwang Unit 4 PWR 1,000 KHIC/CE KHIC/GE 1996.1 G Non-Turnkey Yeonggwang Unit 5 PWR 1,000 KHIC/CE KHIC/GE 2001.6 ◗ Non-Turnkey Yeonggwang Unit 6 PWR 1,000 KHIC/CE KHIC/GE 2002.6 ◗ Non-Turnkey Ulchin Unit 1 PWR 950 Framatome Alstome 1988.9 G Non-Turnkey Ulchin Unit 2 PWR 950 Framatome Alstome 1989.9 G Non-Turnkey Ulchin Unit 3 PWR 1,000 KHIC/CE KHIC/GE 1998.6 ◗ Non-Turnkey Ulchin Unit 4 PWR 1,000 KHIC/CE KHIC/GE 1999.6 ◗ Non-Turnkey Wolsung Unit 1 PHWR 678.7 AECL NEI/Parsons 1983.4 G Turnkey Wolsung Unit 2 PHWR 700 AECL/CE KHIC/GE 1997.6 ◗ Non-Turnkey Wolsung Unit 3 PHWR 700 AECL/CE KHIC/GE 1998.6 ◗ Non-Turnkey Wolsung Unit 4 PHWR 700 AECL/CE KHIC/GE 1999.6 ◗ Non-Turnkey New Unit 1 PWR 1,000 Unspecified Unspecified 2003.6 ❍ New Unit 2 PWR 1,000 Unspecified Unspecified 2004.6 ❍ New Unit 3 PWR 1,000 Unspecified Unspecified 2005.6 ❍ New Unit 4 PWR 1,000 Unspecified Unspecified 2006.6 ❍ New Unit 5 PHWR 700 Unspecified Unspecified 2006.3 ❍ to interconnect the infrastructures of nuclear facilities located on the allow the KEDO project to use GE- of the two countries, for the ROK East Sea, as discussed previously. designed turbines, as they were to build some of its future nuclear used in all other ROK KSNP reac- plants in the DPRK, ship some of Second, KEPCO is embarked on tors, could deal a blow to the the power south, and use the elec- large-scale, nuclear-plant standardi- KEPCO standardization program in tricity sales revenues to pay the zation programs, starting with the its first application beyond the ROK construction costs in the DPRK KSNP program and extending to borders. GE is concerned with the sites. Should such a program the 1,300-MW(e) Korea Next- lack of adequate regulation govern- materialize, the KEDO project Generation Reactor, now in the ing third-party nuclear liability would be the first of its kind, with design and licensing stages. KEPCO issues in the DPRK, as discussed in other similar projects (built under would like each of its nuclear plants Chapter 2. Because these issues bilateral ROK–DPRK agreements) built in the DPRK to be identical to were not resolved, KEDO is now to follow. The creation of the spe- its existing plants in the ROK, save negotiationg with the ROK turbine cial ROK–DPRK currency-clearing for minimal site-specific modifica- manufacturer, Hanjung, and with a mechanism, which applies to tions. A standardization program consortium of the Japaneese electri- financial settling of bilateral trans- would ease the licensing burden on cal-equipment vendors, Toshiba actions between the two countries the regulatory agencies, reduce and Hitachi, to provide the turbine only, may ease inter-Korean joint plant capital cost, simplify opera- generators for the KEDO reactors, projects, so long as all payments tions maintenance and refueling thus making them somewhat differ- are ultimately guaranteed by the procedures, and reduce the annual ent from all other ROK-located ROK. Such a construction program electricity production costs. In this KSNP reactors that use the GE tur- of ROK reactors in DPRK sites context, the General Electric bine, manufactured by Hanjung would further increase the number Company (GE) decision not to under license in the ROK.

38 The KEDO Reactors and Associated Facilities and Activities

Hanjung (Korea Heavy Notes to Chapter 3 9. Peter Hayes, Nautilus Institute Industries and Construction) is the for Security and Sustainable design and manufacturing contrac- 1. David Albright and Kevin Development, “DPRK Energy tor for the Nuclear Steam Supply O’Neill, eds., Solving the North Dilemmas and Regional System and the Turbine Generator Korean Nuclear Puzzle (Institute Security,” Presentation to the (TG). Because GE prohibits of Science and International Asia/Pacific Research Center, Hanjung from using its technology Security, Washington D.C., Stanford University, Stanford, for the KEDO TGs, Hanjung will 2000). California, October 10, 2000. need to license new TG technology from another vendor, e.g., the 2. A. Nechaev and S. Onufrienko, 10. Ji-Ho Kim, “Minister Denies Japanese manufacturers Toshiba or “Full Speed Ahead for China’s ‘Backstage Accord’ on Power Hitachi. Both Toshiba and Hitachi First VVER-1000s Modern Supply to North Korea,” The design and build TGs for GE- Power Systems,” Korea Herald, “ Seoul, Korea, designed BWRs built under license Atomenergoproject, February December 21, 2000. in Japan. While Toshiba and 2000, pp. 41–43. Hitachi have close working rela- 11. Jay Solomon, “Europe tions with GE and are licensors of 3. Byung-Koo Kim, Director, Engineers Wait for U.S. Move GE technologies, they have not Technology Center for Nuclear to Offer Energy Help to North built PWR-type TGs that operate at Control, Korea Atomic Energy Korea,” The Wall Street Journal, different pressure and temperatures Research Institute (KAERI), March 21, 2001. than BWR turbines. Should the TG “KEDO LWR Project for technology finally chosen for the International Cooperation and 12. Korea Electric Power KEDO reactors be different from Nonproliferation,” Presentation Corporation, Korean Standard that used for other KSNP plants, at the Monterey Institute for Nuclear Power Plant (KSNP), this will cause significant changes International Studies, Monterey, Seoul, Korea, 1996. in the Balance of Plant (BOP) California, July 24, 2000. design. These may include redoing the entire BOP heat balance, sizing 4. Nuclear Engineering up of equipment items, and chang- International, 1997 World ing the dimensions and the design Nuclear Industry Handbook of the TG building. The changes (Wilmington Business required in the BOP may delay Publishing, Ruislip, United completion of the entire KEDO Kingdom, 1997). Project and cause further delays in meeting some of the DPRK’s non- 5. Central Research Institute of proliferation obligations, tied to the Electric Power industry reactors’ completion dates. (CRIEPI), Nuclear Power Stations in Japan 1995, Tokyo, Japan, KEPCO has targeted various 1995. East Asia nuclear markets, particu- larly the PRC as potential opportu- 6. Korean Peninsula Energy nities to export the standardized Development Organization, KSNPs. The KEDO project is per- KEDO at Five–Annual Report ceived by KEPCO to be the first 1999/2000, New York, NY, example of building standardized September 2000. KSNPs outside the DPRK, thus serving as a reference plant for future third-party sales. Should dif- 7. J.B. Mulligan, Director for ferent turbines be used in the Project Operations, KEDO, per- KEDO reactors, as compared with sonal communication, New the ROK KSNPs, the value of the York, NY, September 2000. KEDO project as a demonstration or reference plant for further 8. Lowell Feld, US Energy potential exports, as well as serv- Information Administration, ing as a lead project for other North Korea Country Analysis KSNPs to be built in the DPRK, Brief, Washington, D.C., June would diminish. 21, 2000.

39 The KEDO Reactors and Associated Facilities and Activities

40 CHAPTER 4 SAFEGUARDS ON THE KEDO REACTORS

4.1 Introduction Corporation (KEPCO) in December consulting contract with the US 1999, this organization chart applies firm of Sargent & Lundy (S&L). This chapter begins with a equally to the KEDO reactors’ proj- S&L was the original non-Korean description of the Korean Standard ect with the following modifications: A/E of Yeonggwang 3 and 4 (the Nuclear Plant (KSNP) reactors pre-KSNP plants) and (in a deployed in the Republic of Korea 1. KEPCO, who does not own the reduced capacity) the remainder of (ROK) and proposed for the Korean KEDO reactors, has assumed the the KSNP reactors. KOPEC has Energy Development Organization role of the general contractor for carried increasing A/E responsibil- (KEDO) project at the Kumho site in the Kumho Site Nuclear Project, ities in the ROK’s KSNP plants. the Democratic People’s Republic of with KEDO being the overall Korea (DPRK). The parts of the owner for whom KEPCO man- 3. Hanjung (Korea Heavy Industries reactor plant relevant to the safe- ages the project. This arrange- and Construction) remains the guards program are presented so as ment is different from all other design and manufacturing con- to provide the proper context for KSNP projects where KEPCO is tractor for the Nuclear Steam the safeguards discussion that fol- both the owner and the general Supply System (NSSS) and the lows. Most of this chapter describes contractor. Turbine Generator (TG). However, in detail the International Atomic the TG to be used in the KEDO Energy Agency (IAEA) safeguards 2. KOPEC-A/E (Korea Power reactors will not be based (under program as applied to a KSNP-type Engineering Company) has license) on the General Electric of nuclear-power plant. In this dis- assumed sole Architect/ (GE) turbines like the rest of the cussion, we assume that the safe- Engineering (A/E) responsibility KSNP plants because GE did not guards program now applied to the for the KEDO reactors subcontract- obtain the measures of relief it ROK’s nuclear plants will equally ing to KEPCO and has canceled its sought from nuclear liability. apply (at the least) to similar reac- tors in the DPRK. The description of the safeguards program stresses Overall Project Management both the accounting process and the measurement process during nor- KEPCO mal operation and during refueling outages. We conclude the chapter Architect/ Equipment Construction Start-up with a discussion of additional Engineering measurements and inspections pos- sible under the IAEA safeguards Nuclear Steam Turbine Balance of Construction Fuel agreement now in force with the Design Supply System Generator Plant Management Review DPRK, together with a brief assess- Yellow Cake, Start-up and Design Design ment of other verification measures. Conversion, Procurement Supervision Approval Review Review Enrichment 4.2 Project Organization To Design/ Design/ Installation Supply the KEDO Reactors Design Fabrication Manufacture Manufacture Manufacture and Erection KOPEC Hanjung Hanjung KNFC Local/Foreign Local Co. Figure 4-1 is an organization chart for the supply of a standard KSNP Design reactor in the ROK.1 This figure KAERI depicts the contractors’ organization Fuel Design S/L ABB-CE G/E assembled to supply a KSNP reactor KAERI in the ROK. Since the signing of the Turn Key Contract between KEDO Figure 4-1. Organization chart for the supply of a Korean Standard Nuclear Plant (KSNP) and Korean Electric Power reactor. KOPEC stands for Korea Power Engineering Company. KNFC stands for Korean Nuclear Fuel Corporation.

41 SAFEGUARDS ON THE KEDO REACTORS

4. The Korean Atomic Energy Nuclear Plant. The first two units of the energy-conversion part of the Research Institute’s (KAERI) the KSNP series of standardized plant but not related to the nuclear responsibilities for system design plants were the Ulchin 3 and 4 fuel or to safeguards arrangements. have been transferred to the sys- plants, the reference plants of the The nuclear reactor itself is in the tems design department of series. Four other plants of this lower part of the containment build- KOPEC (KOPEC-SD). series are under construction in the ing. During routine operations, the ROK; two additional ones were reactor pressure vessel is closed and 4.3 Description of the KEDO announced and the two KEDO the entire vessel is immersed in a Reactors reactors, planned as identical to large water pool. other KSNP units, are further The KEDO reactors are the KSNP extensions of this ROK series. During routine operations, access design, a modification and scale- to the containment building is limit- down of the System 80 design of A description of the KSNP plants is ed and tightly controlled. The two Combustion Engineering Company available in References 1-3 (see the major entries into the containment (formerly ABB/CE, now Notes at the end of this chapter) as building are through personnel Westinghouse Electric Corporation, well as in many other nuclear-engi- access port(s), camera-monitored a unit of British Nuclear Fuel Ltd.). neering publications. The discussion and controlled by the site-security The System 80 design, in turn, is here is limited to the parts of the plant force, and through the large equip- based on an improved design of the relevant to the refueling operations ment hatch. The equipment hatch is 1,300-MW(e) reactors built in the and to which safeguards arrange- regularly tightly sealed to complete- Palo Verde plant in Arizona. ments apply. A side cut of a KSNP ly isolate the containment building KEPCO has built two plants of the building is shown in Fig. 4-2.1 The from the outside for safety consider- System 80 design—the containment building is in the center, ations. Additionally, the IAEA Yeonggwang 3 and 4 units. This with the fuel building on the right. places its own seal on the equip- design was enhanced by KEPCO, The left side of the figure includes the ment hatch and installs a remotely adjusted to Korean conditions, and TG building and other Balance of monitored camera in the reactor renamed the Korean Standard Plant (BOP) facilities—essential for building surveying the equipment

Polar crane 230'-0"

Dessicator & storage tank 211'-8"

202'-2" Refueling machine assembly CEA change platform Spent-fuel handling machine 101'-6" 195'-0" Steam generator 195'-0" 164'-0" Reactor 165'-0" 173'-0" Generator L.P. Turbine H.P. Turbine vessel 135'-0" 144'-0" 142'-0" 125'-0" 125'-0" 122'-0" 100'-0" 100'-4" 100'-6" 92'-6" 77'-0" 77'-0" 73'-0" 58'-0" 58'-0" 53'-0" 58'-0" Condenser 58'-0" 34'-0" 34'-0"

Turbine building Primary aux. building Containment building Fuel building

Figure 4-2. Side view of a Korean Standard Nuclear Plant (KSNP) building (adapted from Reference 1).

42 SAFEGUARDS ON THE KEDO REACTORS hatch area. These two measures are safeguards concern. In case of the bridge. Any surreptitious attempt to intended to ensure that no large- KEDO rectors, assuming commer- move spent fuel through the transfer scale equipment, such as spent-fuel cial operation by 2010, a detailed canal from the reactor to the fuel storage casks, can be brought into plan for disposing of old and dis- building, or to remove fuel from the the containment building to surrep- charged spent fuel will have to be fresh or spent-fuel storage pools in titiously remove fuel elements from in place no later than 2030. the fuel building, will be detected. the core. A top-down (footprint) view of a Figure 4-4 is a schematic of the of Fresh fuel elements are brought KSNP reactor is shown in Figure 4-3.1 the primary system equipment lay- into the nuclear plant and spent fuel As seen, the parts of the plant of con- out in the containment building.2 elements are taken out of the plant cern to the safeguards program—the The equipment arrangement of the through the fuel building. The fuel containment building, the fuel build- KSNP reactor is unique among building contains the receiving area ing, and the fuel-transfer canal are a Pressurized-Water Reactor (PWR) for the fresh-fuel-element contain- relatively small part of the overall designs. This design is built on two ers, the water pool in which the plant buildings’ area. The major part large steam generators, each connect- fresh fuel elements are stored prior of the plant is taken up by the auxil- ed through a “hot leg” pipe to the to their insertion into the reactor iary building around the containment reactor vessel. The discharge from core during the refueling operation, building, and by the TG building. each steam generator is divided and the transfer canal from the fuel sent through two large primary building to the containment build- A particular point of concern to pumps through two large pipes, “the ing, the spent-fuel storage pool, and the safeguards program is the hori- cold legs” back into the reactor ves- the spent-fuel cask loading area. zontal fuel-transfer canal located in sel. Most 1,000-MW(e) PWRs are the lower part of the containment built on a three-loop design, with By far, the largest part of the fuel building and connected to the bot- each of the three loops containing its building is the spent-fuel storage tom part of the spent-fuel storage own steam generator. The System 80 pool, sized to contain the routine pool in the fuel building. The IAEA design with only the two “hybrid” discharges of at least 10 years of installs a seal on the moveable loops is also distinguished by having operation, plus additional capacity bridge on top of the fuel-transfer two very large steam generators. for a full-core discharge in case all canal. A remotely monitored camera Because the equipment hatch is sized the nuclear fuel from the reactor’s is installed by the IAEA on the fuel to pass through the largest equip- pressure vessel must be removed to building’s wall facing the fuel-trans- ment item in the containment build- maintain or repair the vessel or its fer canal and the IAEA seal on the ing, the System 80 design has a large internal components. It is possible through re-racking and installing of neutron-absorbing plates to increase the storage capacity of the spent-fuel storage pool to about 20 Access Secondary auxiliary control building years’ worth (maintaining the full- building core discharge capability) while Fuel-transfer canal avoiding criticality concerns. Once Primary auxiliary the capacity of the spent-fuel stor- building age pool is reached, the spent fuel Fuel already in the pool must be building removed either to storage pools at Containment other reactors, to an away-from- Turbine building building reactor storage facility, or to dry storage in steel or concrete casks at the reactor site or at a remote cen- tralized facility. Because the operat- Heater bay ing life of a KSNP is expected to be at least 40 years and could well be extended to 60 years, the disposi- Figure 4-3. A top view of the Korean Standard Nuclear Plant (KSNP) reactor. tion of its spent fuel outside of the containment building is a matter of

43 SAFEGUARDS ON THE KEDO REACTORS

Development of Korean PWRs hatch capable of passing through a Kori 1Kori 3 Yeonggwang 3 KSNP large steam generator and is certainly Gross capacity capable of passing through even the (MWe) 587 550 1000 1000 largest spent-fuel storage cask. Comm op. date 1978.4 1985.9 1995.3 Figure 4-5 shows the internal lay- out of the KSNP reactor vessel.2 The General reactor core, in which the fuel ele- layout ments are contained, is located in the lower half of the pressure vessel below the two outlet nozzles of the two hot legs and the four inlet noz- zles of the cold legs. The water in Containment ID ID ID ID the primary system (the reactor ves- building sel, the tubes side in the two steam generators, the four primary pumps, and the pressurizer) enters Type Steel containment Pre-stressed concrete Pre-stressed concrete Pre-stressed concrete the reactor vessel through the inlet ID (m) 32 39.6 43.9 43.9 Design pressure nozzles, flows down around the cir- (lb/in2 (g)) 43 60 54 57 cumference of the vessel, enters the Free volume (m3) 41,000 61,000 76,000 76,000 reactor core area from the bottom part and flows upwards through the core volume inbetween the fuel Reactor coolant elements. Water is heated on its pas- loops sage through the core and then exits the reactor vessel through the two Loops 2 3 2 2 hot legs on its way to the steam No. of fuel generators. The steam generators assemblies 121 157 177 177 produce steam in the secondary sys- Reactor ID (cm) 132 157 162/396.9 162/396.9 tem (the vessel side of the two Inlet/outlet steam generators, the TG, condens- temperature 282.6/320.2 291.9/326.9 295.8/327.3 295.8/327.3 er, and the feedwater heating sys- ° ( C) tem) that rotates the TG machinery to generate electricity. Table 4-1 Pressurizer summarizes design data for the KSNP reactor.2 Steam generator Steam generator The upper part of the reactor-ves- sel internals contains the control Reactor rods’ guide tubes and the portions coolant pump of the control rods withdrawn from the reactor core to achieve criticality in the core and maintain a stable nuclear chain reaction. The upper part of the vessel is also filled with primary system water, which pro- vides extra water volume in the vessel for emergency cooling pur- Reactor poses, increasing the safety margins of the entire reactor system. The control rods are connected though Figure 4-4. Schematic of the primary system layout in the containment building. the top of the reactor vessel with the Control-Rod Drive Mechanisms (CRDMs) that govern the vertical movements of the rods in and out of the reactor’s core.

44 SAFEGUARDS ON THE KEDO REACTORS

4.4 Refueling Operation in Generally, the reactor is refueled 3. Shuffling the remaining in-reac- the KEDO Reactors by— tor fuel to different positions within the core. This is done to Of particular interest from a pro- 1. Shutting the reactor down, level the reactivity and power liferation perspective is the time allowing it to cool, and remov- distribution in the core. required to access the reactor core ing the pressure vessel top (or and its fuel assemblies. This takes at head) and the associated equip- 4. Inserting the fresh fuel. The fresh least two to four days with a rela- ment. fuel is inserted into the outer tively experienced crew and may periphery of the core. The higher take longer. First, it takes about two 2. Removing the oldest fuel from reactivity of the fresh fuel compen- days for the reactor to cool and the reactor. This is the fuel that sates for the lower neutron flux depressurize and for the intense has resided in the core for three and the higher neutron leakage radioactivity to subside to tolerable full cycles, and therefore has rate from the core at the periphery. levels. Second, the vessel head and achieved its nominal design burn- CRDM mechanisms must be up. This fuel is generally in the 5. Reinstalling and bolting the removed. This takes an additional center of the core. This fuel is sent pressure vessel head and recon- one-half to two days. to the spent-fuel storage pool. necting the CRDMs.

The time between refueling (also called the cycle time) depends on a CRDM Figure 4-5. Internal number of variables, including the layout of the Korean Standard Nuclear Closure head nozzle design burnup (a measure of the Plant (KSNP) reac- total energy supplied by a given CEA extension shaft Reactor vessel tor vessel. mass of fuel), and the design of the closure head fuel and reactor. Most PWRs operat- assembly (Class 1) ing today are transitioning from a one-year cycle to an eighteen-month cycle and some to a two-year cycle. These cycle times are “nominal” and can vary in practice with power demand or to take advantage of personnel availability. Current PWRs of the KSNP vintage are designed to operate at a design bur- Inlet nozzle Outlet nozzle nup of 45 to 50 MWd/kg, and there is interest in increasing these bur- nup levels still further to about 70 MWd/kg. The transition to high- Upper guide er design burnup levels and longer structure residence times in the core (longer assembly Fuel assembly cycles) means that over time a smaller amount of spent fuel is gen- Surveillance Core shroud erated that has to be stored, moni- capsule tored, and disposed of. Core support barrel Lower One-third of the core is replaced support Reactor vessel structure (Class 1) at each refueling. Fuel will normal- ly reside in the reactor’s core for Flow skirt three full cycles, four and one-half In-core Core stop years, before being discharged as instrumentation spent fuel. In practice, both the nozzle refueling schedule and the fraction Bottom head nozzle of the core replaced each time can vary between cycles depending on the operating history of the plant.

45 SAFEGUARDS ON THE KEDO REACTORS

6. Restarting the reactor. Table 4-1. Characteristics of the Korean Standard Nuclear Plant (KSNP). Reactor In the US, the shortest reactor Type PWR refueling operation has been accom- Thermal output 2815 MW(th) plished in just 16 days (the current Coolant flow rate 121.5 x 106 lb/h US record), although the average is Design pressure 2,500 Ib/in2 approximately 34 days. The actual Operating Pressure 2,500 lb/in2 time required to refuel a reactor Design temperature 650°F varies depending on the type of Inside diameter at shell 162 in reactor, the skill and experience of Overall height 48 ft the operators, and the maintenance Fuel (both scheduled and unexpected) Number of fuel assemblies 177 required during the outage. Number of UO2 fuel rods per assembly 236 (16 X 16) Fuel weight 188,609 lb As seen in Table 4-1, the reactor Core height (active) 150.0 in core contains 177 assemblies. Each Core diameter (equivalent) 123.0 in one-third-core refueling results in 59 Clad material Zircaloy-4 Clad thickness 0.025 in assemblies withdrawn from the core as spent fuel and 59 fresh fuel Reactor coolant system assemblies inserted into the core. Number of loops 2 Figure 4-6 is a schematic of a fuel Hot leg/cold leg 42/30 in assembly.3 Each assembly weighs Reactor inlet temperature 564.5°F slightly over 600 kilograms (more Reactor outlet temperature 621.2°F Total coolant volume 11,315 ft3 than half a ton) and contains 236 long- and small-diameter fuel rods Control rods arranged in a square 16 x 16 lattice Number of control assemblies 73 with an empty “water hole” in the Number of rods per assembly 4 or 12 center. Each fuel rod contains a Material (full/part strength) B, C/Inconel number of cylindrical fuel pins or Steam generators pellets stacked in a column and held Type, number of units Vertical U-tube, 2 in a zircalloy tube. Steam flow per steam generator 6.354 X 106 lb/hr Steam pressure at full power 1,070 lb/in2 During refueling, each fuel Steam temperature at full power 550.5°F assembly is inspected before being Maximum moisture 0.25% inserted or reinserted into the reac- Feedwater temperature 450°F tor. The level of detail to which each Reactor coolant pumps assembly is inspected can vary. For Number 4 example, fuel that has been in the Motor/type AC Induction/ Vertical, centrifugal reactor for two cycles may be Design capacity 82,500 gal/min inspected more thoroughly than Design head 340 ft fuel in for only one cycle. If there Containment are indications of problems (such as Type Prestressed cylindrical concrete with steel liner detection of fission-product gases Inside diameter 144 ft released from a leaking fuel pin), a Height 216 ft 6 2 more rigorous inspection may Free volume 2.73 x 10 ft ensue. Fuel that is damaged or oth- Liner thickness 0.144 in erwise suspect is not reinserted into Turbine the reactor, but sent to the spent- Number 4 (high 1, low 3) fuel storage pool. Depending on the Type Serial 6 flow arrangement type and extent of damage, the RPM 1,800 rejected fuel assembly may be Generator Number, type 1, 4 poles (1,800 RPM) placed in an intermediate sleeve Voltage 22 kV, 3 phases (and the entire assembly placed in Frequency 60 Hz Net electrical output 1,000 MW(e) the spent-fuel storage pool) to mini- mize contamination of the spent- Condenser fuel storage pool. Alternately, a Number, type 3, once-through sea water cooling Pump type Vertical, Centrifugal

46 SAFEGUARDS ON THE KEDO REACTORS leaking fuel rod can be removed from an assembly (underwater) and Upper end Upper a new rod inserted so that the fitting end cap repaired assembly can be reinserted into the core. Spring A reactor operator normally has a supply of fresh fuel on hand. Al O Immediately prior to a refueling, CEA 2 3 spacer there is at least one-third of a core of guide tube fresh fuel available. This stock is assembly accumulated over the course of Fuel some months, as it was not likely pellets delivered in a single shipment. 15.719" Because of the possibility of defec- (399.27 mm) tive fuel, reactor operators normally 9 spaces keep a small stock of additional 178.250" (4527.55 mm) (beyond that required for the next Lower zircalloy refueling) fresh-fuel assemblies on spacer grid hand. Partial refueling may occur between normal fueling outages. 163.550" Although unusual, this can occur if, 150" (4154.17 mm) Fuel (3810 mm) for example, failed fuel is detected cladding Active during operation (usually by detec- fuel tion of fission-product gases such as length iodine). 16.719" (424.66 mm) Fuel rod The spent fuel assemblies dis- charged from the reactor are moved Inconel spacer by the charge/discharge machine grid away from the pressure vessel and placed in a horizontal position on a Lower end trolley that carries them through a fitting transfer canal (a tunnel) leading from 4.312" Al2O3 the containment building to the fuel (109.52 mm) spacer building. Figure 4-7 is a schematic of 2 Lower the fuel transfer process. Once in the end cap fuel building side of the canal, the fuel assembly is brought again to an upright position and is carried under- Fuel Rod Assembly water to its storage position in the spent-fuel storage pool. The position and identity of each assembly in the storage pool are recorded and verified Bottom view by the IAEA safeguards inspectors on their periodic inventory-verification Figure 4-6. Schematic of a fuel assembly. visits to the plant.

47 SAFEGUARDS ON THE KEDO REACTORS

1. CSB storage stand 2. Alignment guide pin 3. Pool seal 4. Upending machine 5. Transfer tube valve 6. Spent-fuel handling machine 7. New fuel shipping container 8. Transfer system control console 9. Hydraulic power package 10. Spent-fuel shipping cask loading 11. Refueling machine 12. CEA change platform 13. CEA storage racks 14. Transport container 16 15. Reactor vessel head assembly 7 10 21 16. Missile shield 17. CRDM cable trays 6 18. Upper guide structure lift rig 9 17 24 19. Upper guide structure 8 20. Fuel transfer tube 15 22 11 23 21. Transfer system winch 4 22. New fuel elevator 20 5 23. New fuel storage 1818 1 24. Spent-fuel storage pool 12 2 4 3

14

13 19

Figure 4-7. Fuel transfer process.

SPENT FUEL AND PLUTONIUM

Light-Water Reactors (LWRs) (measured as the fraction that is bility of a weapon containing such fueled with Low-Enriched Uranium 239Pu) decreases with increasing isotopes difficult. (LEU) (essentially all of them) burnup. As the reactor continues to inescapably produce plutonium as operate (and burnup increases), For the plutonium in spent fuel to a byproduct. The quantity and quali- some of the plutonium itself cap- be used in a weapon, it must first ty of the plutonium are both affect- tures additional neutrons. Some of be extracted from the spent fuel. ed by plant operations, with the it then fissions, contributing a signif- Although the basic chemistry for principal variable being the total icant amount to the total energy extracting plutonium from spent irradiation of the fuel in the reactor. generated by the reactor, and some LWR fuel is well known (the so- Fuel irradiation is referred to as bur- is converted to the so-called “higher called “PUREX” process), the nup and is a measure of the total isotopes” of plutonium. The higher process is significantly complicated energy produced by a unit mass of even-isotopes of plutonium, in par- by the existence of both nuclear fuel. Fresh LEU fuel has no plutoni- ticular, do not fission efficiently in a radiation and the heat output of the um and the total amount of plutoni- thermal reactor and thus accumu- spent fuel. Both the radiation and um in the fuel increases with late (at the expense of the 239Pu). heat exacerbate handling, influence increasing burnup. Other nuclear reactions also con- safety, and complicate the chemical tribute to the production of trouble- processing. Both slowly decay with Plutonium is produced primarily some isotopes (such as 238Pu), time so that these complications by neutron capture in 238U, the with the fraction of these isotopes are somewhat less onerous for most abundant isotope of uranium also increasing with increasing bur- “older” spent fuel (spent fuel dis- in LEU fuel. This produces primarily nup. These other isotopes of pluto- charged from the reactor several 239Pu, the isotope best suited to nium may themselves be used in a years past) than for “younger” spent weapons applications. However, the weapon, but they make design, fuel (recently discharged). quality of the plutonium produced manufacturing, handling, and relia-

48 SAFEGUARDS ON THE KEDO REACTORS

4.4.1 The Special Problem of the Similarly, at the second refueling, than nominal burnup. The earli- Beginning-of-Life and the spent fuel removed has seen est discharged BOL fuel has seen End-of-Life Fuel Discharges only two cycles, so the quantity and only one cycle of operation and quality of its plutonium lies roughly will only have 12 to 15 MWd/kg A reactor can be considered to mid-way between that of one-cycle burnup (assuming a “nominal” have three distinct phases in its and three-cycle or full-burnup spent 45 MWd/kg design burnup). The overall life cycle: beginning of life fuel. Roughly, the BOL portion of first cycle discharges spent fuel (BOL), equilibrium, and end of life the reactor lifetime can be consid- containing approximately 100 (EOL). Most of the reactor’s lifetime ered to last for the first three cycles. kilograms of plutonium, of which is in the equilibrium part of the life over 80 percent is 239Pu. The one- cycle where spent fuel is put in the A similar situation occurs at the cycle EOL spent fuel will have reactor, remains for three cycles, end of the reactor life. At the next- similar characteristics. and then is removed as spent fuel. to-the-last refueling, the fresh fuel inserted into the reactor will only 2. The second refueling discharges Early in the reactor’s operating remain in the reactor for two cycles, fuel at approximately life (the BOL), much of the fuel in and that inserted during the last 30 MWd/kg, yielding some the core is very fresh, without the refueling will remain in the reactor 175 kilograms of plutonium con- buildup of fission-product poisons, for only a single cycle. The implica- taining roughly 65 percent 239Pu. plutonium, and other elements that tions of the EOL part of the lifecycle The two-cycle EOL spent fuel will affect the reactivity of the fuel. For are similar to those at the BOL: also have similar characteristics. this reason, some of the BOL fuel is some of the EOL spent fuel will supplied with lower enrichment have seen only one or two cycles, 3. Each equilibrium refueling than fuel used later in the equilibri- and thus has plutonium of higher (45 MWd/kg) discharges fuel um part of the lifecycle. In the “ini- quality, albeit lesser quantity. containing about 288 kilograms tial core,” the lowest enrichment of plutonium with about 57 per- fuel goes into the center zone of the Thus, three “types” of spent fuel cent 239Pu content.* core, medium-enrichment fuel goes are generated as a result of normal into the intermediate zone, and plant operations: Due to their higher 239Pu quality nominal enrichmenƒt fuel is placed (and thus increased attractiveness for in the outer zone. This arrangement 1. Of greatest concern is the spent weapon use), both the BOL and EOL mimics the reactivity distribution of fuel discharged during the BOL discharged fuel assemblies merit spe- a core later in its life. and EOL, which has much less cial attention and verification efforts.

At the first refueling, the central 20 100 zone of the core is removed and sent to the spent fuel pool, just as normal spent fuel. However, this fuel has 16 80 been in the reactor for only one cycle, and therefore has only one- 12 60 third the “nominal” burnup. As Fig. Pu 4-8 shows, reduced burnup intro- 239 duces two effects that have impor- 08 40 % tant implications for proliferation: %239Pu • The total quantity of plutonium Total Pu Production

Pu produced (kg/tonne HM) 04 20 in that spent fuel is reduced because of the reduced burnup, 0 0 but 0204060 80 Net burnup (MWd/kg) • The quality of the plutonium for Figure 4-8. The relationships between plutonium production, quality, and burnup. At very weapon use is higher than for 239 full-burnup spent fuel. low burnup, very little plutonium is produced but that produced is primarily Pu. At higher burnup, the total amount of plutonium produced becomes substantial, but the fraction of 239Pu decreases significantly.

* The values cited in this paragraph, as well as in Fig. 4-8, were developed from a variety of data from different LWR reactor designs, using various initial enrich- ments, design burnups and operating cycles. They are intended for illustration only, and are not representative of any particular reactor or operating scenario.

49 SAFEGUARDS ON THE KEDO REACTORS

It is important to note that this the Agreed Framework are dis- materials within specified time peri- low-burnup fuel issue is not unique cussed in Reference 8, and the ods between periodic inspections of to the BOL and EOL parts of the IAEA safeguards requirements the LWR. The inspection goal for reactor life cycle. At all times, the related to the DPRK nuclear pro- each LWR is attained if all the quanti- reactor core itself will contain a mix gram are discussed in Reference 6. ty and timeliness criteria specified in of lower, medium, and higher bur- the Facility Agreement signed nup material. What differentiates 4.5.1 The Safeguards Goals for between the IAEA and the host this in-core material is that the Quantity and Timeliness nation of the specific LWR are met. lower and medium burnup in-core fuel is not normally discharged The standard IAEA approach to In terms of the timeliness criteri- from the reactor, and therefore con- LWR plant safeguards for signato- on, the IAEA defines the frequency sidered somewhat less available. ries to the Non-Proliferation Treaty of inspections as: (NPT) is the IAEA “Model 4.5 IAEA Safeguarding of Safeguards Agreement,” INFCIRC • one year for fresh Low-Enriched Nuclear Fuel in the KEDO 153 of 1971, and its specific appli- Uranium (LEU) fuel (FF), cation to the DPRK is documented Reactors in IAEA INFCIRC 403 of May • three months for reactor core fuel 1992. In general, the purpose of containing bred plutonium (CF), The following discussion of LWR-type safeguards agreements IAEA safeguards aims at describ- is to verify that nuclear materials • three months for spent LWR fuel ing the actual implementation of are not diverted to produce outside the reactor core (in the safeguards agreements related to nuclear weapons or other nuclear- spent-fuel storage pool) and con- nuclear-power plants such as the explosive devices. The verification taining plutonium (SF). KEDO reactors. The application of process is achieved through the safeguards measures to commercial inspection measures described The choice of the time period LWRs and particularly to 1,000- next. The purpose of these inspec- between inspections is predicated on MW(e)-class PWRs such as the tions is two-fold: first, the the IAEA’s assumptions regarding the KSNP reactors is described in Quantity Component is meant to time required from diversion through References 4 and 5, both of which provide assurance that there has fissile-material refining to the manu- relate to safeguarding nuclear- occurred no diversion of a facturing of a completed nuclear- power plants in the Korean Significant Quantity (SQ) of vari- explosive device. Based on the details Peninsula. We assume that the ous nuclear materials over a of the specific Facility Agreement, the safeguard measures applied to the Materials Balance Period (MBP); IAEA must perform interim visits at ROK’s KSNP reactors will apply as and second, the Timeliness the frequencies mentioned above to a minimum to the KEDO reactors. Component is meant to provide ensure that no diversion of nuclear In fact, the IAEA might well insist timely detection against abrupt materials has occurred during the on further enhancements within its diversion of a SQ of nuclear period since the last inspection. agreements with the DPRK, and some of these are described in Section 4.6 and Reference 6. Until Significant Quantities (SQs) of nuclear materials are defined as the the appropriate Facility approximate quantities of materials required to manufacture a nuclear- Agreements are signed between explosive device. The International Atomic Energy Agency (IAEA) has the IAEA and the DPRK, we will defined SQ based on direct use of fissile materials as— not know what these will be. An IAEA perspective on the general • 8 kilograms of plutonium, type of IAEA safeguard measures • 25 kilograms of 235U contained in Highly Enriched Uranium (HEU), applied to typical LWRs is present- • 8 kilograms of 233U. ed in References 6 and 7. An ROK perspective on developing a lead The IAEA further defined the SQ for uranium requiring additional pro- project on safeguards enhance- cessing (indirect use) as— ments that will equally apply in the ROK and the DPRK, and which • 75 kilograms of 235U contained in Low-Enriched Uranium (LEU), could in fact apply to all other • 10 tons of natural uranium, LWRs is presented in Reference 5. • 20 tons of depleted uranium. Some problem areas in implement- ing safeguard measures related to These various forms of uranium require further enrichment to extract adequate amount of fissile 235U for direct use in weapons production.

50 SAFEGUARDS ON THE KEDO REACTORS

The IAEA meets its safeguarding borders and narrows down to the • Examination of the facility’s obligations through item accounting specific plant and specific Material operating records such as the and independent measurement, and Balance Areas (MBAs) within each operations logbook, assembly through Containment and plant. At the country level, the history cards, core fuel maps, Surveillance (C&S) measures. Item accounting process may include pool fuel maps, etc. accounting includes record check- unprocessed or partially processed ing at various locations, physical materials such as natural, enriched, • Reconciliation of the accounting identification, counting, non- or depleted uranium. At the power- records, the operating records, destructive measurements, and plant level, the accounting process and the results of the last inven- examinations to verify the integrity deals with specific numbers of tory inspection. (over time) of different items. In a assemblies at different locations in nuclear plant, the items inspected differing time periods. • Comparison of the facility and verified are the fuel assemblies, records with the country reports and in few cases of leaking pins— To carry out the accounting and specific notifications, to individual fuel rods. Inspections are process, the IAEA relies first on allow vertical reconciliation. performed using non-destructive examinations of records. These procedures. C&S measures comple- include both country reports and • Preparation of the summary ment accounting and inspection by specific facility records. The IAEA results of the inventory inspec- providing seals at critical points in further: tion that will become the baseline the reactor plant to prevent unau- for the next inspection. thorized entry, and by operating • checks domestic and internation- surveillance systems to detect unde- al materials transfers, The above activities are supple- clared movement of nuclear materi- mented by physical inspections als or attempts to tamper with the • attempts to confirm that no and verification examinations of containment system or the IAEA unrecorded production of direct- the FF, CF, and the SF, to make sure seals and safeguard devices. The use materials (measured in SQ) that the facility records correspond C&S system is particularly applica- has taken place, to the results of the actual inspec- ble to highly radioactive nuclear tions on site. At the power plant, materials, which cannot be inspect- • confirms the absence of borrowed this means that the office records ed from nearby and thus are nuclear materials, related to the number of assemblies remotely monitored. In practice, the at various locations in the plant C&S system applies mostly to the • correlates all the above with the coincide with the results of the plutonium-bearing CF in the con- data from its last interim inspec- plant inspections and assembly tainment building and the SF in the tion, counting on site. spent-fuel storage pool in the fuel building. Both the accounting and • and, out of these data, recreates 4.5.3 Material Balance Area in the C&S measures, which are the an updated material balance that KEDO-type Reactors heart of the safeguards regime, are accounts for the whereabouts and described in the next section. disposition of all nuclear materi- A KEDO-type, 1,000-MW(e) PWR als at the time of the specific includes three MBAs during which 4.5.2 The Safeguards inspection. an inventorying process is undertak- Accounting Process en between each periodic inspection This material-balancing process and the next. The MBAs include: The IAEA safeguards regime is occurs both at the national and at basically a large-inventory account- the facility levels with the records • FF Storage with Key ing system, systematically applied from each level aggregating (or dis- Measurement Point A. This bal- to account for all nuclear materials aggregating, as the case may be) to ance is applicable to LEU only. It at all the declared nuclear facilities the other level. is based around the FF pool in the in each country. Full inventory (sta- fuel building and the reactor core tistically complete) is taken at At the facility level, in particular, in the containment building (at appropriate frequencies based on the IAEA’s records examination refueling outages). This material the timeliness criterion to ensure may include several activities: balance considers new FF brought that no diversion which could have into the plant, FF in storage at the led to the manufacture of nuclear- • Examination of the facility’s FF pool in the fuel building, and explosive devices has taken place accounting records such as the FF loaded into the reactor core since the last inspection and inven- general ledger, receipt and ship- during a refueling outage. tory. In each country, the accounting ping records, inventory docu- process starts from the country’s mentation, etc.

51 SAFEGUARDS ON THE KEDO REACTORS

• CF with Key Measuring Point B. • SF removed from the storage placing seals at strategic locations This balance is updated during pool for dry-cask storage on site to prevent access to nuclear materi- scheduled refueling outages, or to another SF storage pool, als and by installing surveillance which occur in a KEDO-type cameras to check for suspicious reactor every 18 months. Special • SF removed from the pool to a movements in the containment or inspections are also held if for centralized away-from-reactors fuel buildings. The IAEA standard any reason the reactor was shut storage facility, method of installing C&S equip- down, the pressure vessel top ment is shown schematically in was removed, and core fuel was • SF removed for direct disposal. Fig. 4-9.4 The IAEA installs a seal discharged. The assembly balance and a surveillance camera in the is held between the numbers of The SF material balance has to containment building, and another assemblies in the core as of the consider both the plutonium content seal and a camera in the fuel build- last inspection (CF), the number in the various SF assemblies and the ing. Some variation on this basic of spent-fuel assemblies removed residual 235U remaining in the SF. scheme involving additional seals from the core (SF), and the num- also exists, as seen in Fig. 4-9. An ber of fresh fuel assemblies In each periodic inspection, additional temporary surveillance added to the core (FF). Core-fuel described next, a material balance is camera is installed by the IAEA in inventorying has to consider the carried out on each MBA, and it is the containment building during a both enriched 235U in the FF and required that the each of the bal- refueling outage. CF (as well as the unfissioned ances properly close and that all 235U in the SF), the bred plutoni- three balances correspond to the The seal in the containment build- um growth and fissioning in the facility records, taking into account ing is placed on the large equipment CF, and the discharged plutoni- the plant’s operating history. Given hatch, and the camera is pointed at um in the SF. this multi-layered, inventory-taking the hatch. The purpose of observing process, it is now necessary to dis- the equipment hatch is to make sure • SF with Key Measurement Point cuss the C&S process, and following it is not opened surreptitiously to C. The spent-fuel storage pool is that, the inspection process. admit a heavy shielded cask into the the most sensitive part of the containment building to remove safeguarding process, consider- 4.5.4 The Containment and core fuel or spent-fuel assemblies ing that at each 18-month inter- Surveillance Process from the building. Both the CF and val, an additional 59 SF assem- SF assemblies, being highly radioac- blies are added to the pool, and The C&S measures implemented tive, require large-sized holding that the short-lived fission prod- by the IAEA are aimed at comple- casks that cannot pass through the ucts decay at a fast rate. After 18 menting the accounting process by narrow personnel access ports and years of operation, close to 600 assemblies will have accumulated in a KEDO-type reactor’s SF Underwater Fuel Transfer Channel Reactor Double-Containment Building Overhead pool, all of which have to be Canal Gate (possible sealing) accounted for, based on the time- (Polar) Crane liness criterion, every three SV Unit 1 Transfer Channel Pit Transfer Reactor Pit months. As detailed earlier, the (observes Channel Seal on only exit Walking Racks for spent Pit Cable TTray/ray/ first core (BOL) assemblies with a hatch) Bridges fuel assemblies Reactor high 239Pu fraction are particular- Blocks SV Exit Spent Fuel Pond Reactor Refueling ly sensitive, especially after its Unit Hatch Temporary Dome Bridge (relatively lower) activity has 1 Flask Reactor decayed for 18 years. Loading Base Rack Decontamination area View Angle Temporary Inaccessible Seal The material balance on the SF SV Unit 3 pool includes: Accessible Seal Hatch • the inventory as of the last SV = surveillance Steam Generator inspection, Racks Limited • additional SF discharged from the Access Seal Temporary core in the latest refueling outage SV Unit 2 Equipment Door SV Unit 4 (assuming one has occurred since (if additional hatch exists) the last inspection), Figure 4-9. Light-Water Reactor (LWR) layout showing the containment and surveillance process.

52 SAFEGUARDS ON THE KEDO REACTORS that can only be brought into the other in the fuel building. The world, but is also quite economical containment building through the unique aspect of this scheme is that in terms of conserving scarce IAEA equipment hatch. all cameras and seals are connected resources. Installing this system in via to a server that records the sen- the KEDO reactors, in parallel with The seal in the fuel building is sors’ data and transmits them via its installation in all the ROK’s installed on the bridge above the an Internet uplink or a satellite nuclear plants, would significantly transfer canal from the reactor to the uplink to the IAEA in Vienna every enhance the KEDO reactors’ safe- fuel building. The purpose of this three minutes. The server is located guards, particularly if the system is seal is to reveal any unauthorized in the fuel building and is placed installed on three camera-seal sys- use of the transfer canal, for instance, inside a secured box to prevent tems. Until there is considerable to remove CF or SF fuel from the tampering. This remote-control positive experience of cooperation containment building to the fuel data-gathering allows a near-real- with the DPRK, however, the sys- building. The wide-angle-lens cam- time inspection of the status of the tem should not be used as a reason era on the wall of the fuel building equipment hatch and the fuel pools to reduce actual visual inspections covers the entire fuel pools area, areas by the IAEA safeguards by trained inspectors. including the fresh and spent fuel experts, thousands of miles from pools, the cask-loading and unload- the actual nuclear plants in the 4.5.5 Safeguards Inspections ing areas, and the fuel-transfer canal. Korean Peninsula. If this proves to During Routine Plant Operation This camera tracks and records all work effectively—and all indica- movements within the fuel building tions are that it will—the IAEA will The safeguards accounting and can record any unauthorized consider this system as an equip- process and the verification of the transfers of fuel in or out of the ment enhancement package that integrity of the containment meas- building or the storage pools. matured with new technologies, ures and surveillance devices are and which should be implemented carried out by the IAEA safeguards In some cases, additional seals at all reactors under safeguards. inspectors during plant inspections. are placed on the cable-tray bridge The data-gathering and transmis- The purpose of the inspections is to connecting the reactor pressure ves- sion systems should themselves be confirm the operators’ recorded sel top and the side of the contain- hardened and monitored so that inventory of nuclear materials and ment building. These seals reveal they cannot be spoofed or inter- to reconcile the records with the access to the reactor top, which has fered with without warning. IAEA’s independent measurements. to be opened to divert CF or SF Each inspection is conducted by from the reactor core. The IAEA This experimental program not MBAs and is considered valid for a does not routinely install a third only provides an immediate view specified Material Balance Period. seal and camera in the plant, within of the sensitive parts of a nuclear Two types of Inventory Verification the reactor building and facing the plant located anywhere in the inspections are carried out: top of the reactor vessel and CRDM assembly, and the cable-tray bridge leading to it. During outages, the 5-wire VACOSS cable 7-wire IAEA does rig a temporary camera multi-core data + power in the containment building facing VACOSS seal 2 Camera 3 cable (80 m) the open top of the pressure vessel External equipment Steam Truck 7-wire multi-core hatch VACOSS generator and records the entire sequence of access for RS-232 (195 m) VACOSS seal 1 the refueling operation for later area Canal gate review. Routine installation of a Camera 2 third seal and camera would height- en confidence that the reactor vessel is not being tampered with and is Spent-Fuel Pool desirable in this case. Steam Personnel FF storage generator access hatch 5-wire multi-core data + power Even more desirable is an cable (COM1) (30 m) Containment advanced seal–camera system Camera 1 7-wire multi-core data + power cable penetration box Server routed through auxiliary building implemented on a trial basis in Main power (120 V/60 Hz) (COM2) (265 m) three ROK reactors, shown schematically in Fig. 4-10.5 This IAEA scheme is based on two camera–seal pairs, one installed in Figure 4-10. Advanced camera–seal system implemented in the Republic of Korea (ROK) the containment building and the reactors.

53 SAFEGUARDS ON THE KEDO REACTORS

• Physical Inventory Verification (from a safeguards perspective) by during an IIV and includes replacing (PIV). This inspection coincides one PIV inspection, one PIV-equiva- the seals on the equipment hatch and with the physical inventory lent inspection, and four IIVs. on the transfer-canal bridge gate, and taken by the operator, most inspecting the operating status of the importantly during refueling 4.5.6 The Physical Inventory surveillance camera in the contain- outages. The detailed fuel Verification Type of Safeguards ment building. The inspection activity accounting possible during such for the SF is identical to that carried inspections allows closing of the Inspection out during an IIV. It includes verify- Material Balance Period. ing the status of the C&S measures The most comprehensive type of installed in the fuel building, item • Interim Inventory Verification a safeguards inspection is the PIV counting of the SF assemblies in the (IIV). This inspection is carried inspection. The PIV is carried out pool, and NDA inspection of the SF out during the periods inbetween nominally once every calendar year assemblies using Cherenkov radia- two PIVs. Its purpose is to verify to correspond to the timeliness cri- tion-viewing devices. the status of CF and SF materials terion for LEU in the fresh fuel. The within the plant for meeting the maximum allowed time between The most complex type of PIV timeliness component of the safe- two consecutive PIVs should not occurs during a refueling operation, guards goals. Alternatively, the exceed 14 months, except if the PIV when the IAEA inspectors have to IIV is used to re-establish the schedule coincides with a refueling coordinate their inspection activities continuity of knowledge of the outage. PIVs are carried out at year- with the refueling work of the plant status of the nuclear-materials ly intervals even for reactors operat- operators and with the physical inventory following a safeguards ing on an 18 months’ refueling inventory undertaken by the opera- breakdown episode, e.g., a failure cycle. The PIV is carried out based tors for completion of the plant’s of the surveillance system. on the nuclear-materials inventory own records. The IAEA activities records provided by the reactor when a PIV coincides with a refuel- IIV inspections are carried out operator. The IAEA inspects the ing outage can be divided into three four times a calendar year, with the plant records, reconciles them with distinct phases. Prior to the outage, maximum time between two consec- the country statement, conducts its the IAEA carries out Pre-PIV activi- utive inspections being no longer own inventory of the fuel in the ties, which include removing the than three months and three weeks. plant, and further reconciles its own seals from the equipment hatch, the The period between inspections is measurements with the plant fuel-transfer canal gate, and the top governed by the timeliness criterion records. If the IAEA can close the of the reactor core (if applicable), for inspecting plutonium-bearing MBAs within the plant, and the installation of a third temporary, fuels (CF and SF), which is three total plant balance, the Material camera in the containment building months. Because the reactor pressure Balance Period can be closed. facing the pressure vessel, and inven- vessel is closed during an IIV and torying the fresh fuel in the fresh fuel the plant is in routine operation, the Two types of PIV are carried out, pool prior to insertion into the reac- scope of the inspection is limited to depending on the plant’s refueling tor. At the end of the core refueling verifying the C&S measures and outage schedule: a PIV during a rou- and before the closing of the pressure inspecting the spent fuel. CF inven- tine plant operation with a closed vessel, the IAEA conducts its own tory is verified by checking the core (called a PIV-equivalent inspec- independent inspection of the fuel in integrity of the seal(s) and the sur- tion) and a PIV during a refueling the core. For the fresh fuel (both in veillance camera in the containment outage. A PIV-equivalent inspection the core and remaining in the FF building. The SF inventory in the is similar to an IIV, except that for pool), this includes item counting of fuel building is verified by checking completion of the material balances, assemblies, identification of individ- the status of the C&S measures in the FF inventory is also determined. ual-assembly serial numbers, and the fuel building, by item counting The fresh fuel on hand is a part of the NDA. For the core fuel, this includes of SF assemblies in the spent-fuel FF inventory buildup toward the item counting of assemblies in the storage pool, and by carrying out next refueling outage. The inspection core, serial-number identification of non-destructive examination of SF activity for the FF assemblies FF assemblies in the core, and inspec- assemblies using Cherenkov radia- includes item counting, serial-num- tion of the permanent surveillance tion-viewing devices. Because ber identification of individual camera in the containment building. refueling outages in KEDO-type assemblies and comparison with the For the spent fuel, this includes PWRs occur nominally every 18 plant records, and non-destructive assembly item counting and inspec- months, the period between two assay (NDA). The inspection activity tion of the surveillance equipment in refueling outages will be covered for CF is similar to the one carried out the fuel building.

54 SAFEGUARDS ON THE KEDO REACTORS

After the pressure vessel has sites but not at undeclared places implemented in safeguard systems. been closed, the IAEA conducts without getting permission or In addition to the camera–seal sys- Post-PIV activities. These include requesting a “special inspection.” tem, other sensors can also be placing new seals on the equipment The utility and practicality of envi- remotely monitored: hatch, the spent-fuel transfer canal ronmental sampling in and around gate, and the vessel head assembly nuclear facilities have been validat- • Portal monitors, video, and bridge; removing the temporary ed through field trials at the invita- TESA-type locks9 that record camera from the rector hall; re- tion of a number of member states. entry and exit information (per- inspecting the spent-fuel storage Samples can be collected from sonnel, date, and time), and mon- pools including SF assemblies item water, air, soil, and vegetation. Each itor nominal reactor operations, counting; inspecting all irradiated has its own features. Water sam- the movement (or non-move- assemblies or parts of assemblies pling has been shown to be an ment) of nuclear material, and using Cherenkov and gamma-ray effective and relatively low-cost detect interference with contain- detectors; and verifying the status technique for both short- and long- ment or tampering with IAEA of the C&S equipment in the fuel range detection of nuclear activi- safeguards devices, samples or building. At the end of these inspec- ties, but a clandestine facility can data. tion activities, the IAEA compares prevent water-borne effluents from its own inventory records with the reaching the sampled bodies of • Reactor-emission sensors to pro- plant operators’ records and recon- water. Air-sampling techniques can vide facility data from operating ciles the differences. be applied to both airborne gases, reactors. such as 85Kr, and airborne radionu- 4.6 Assessment of clides associated with particulates, • Tamper-resistant electrical power 129 106 Additional Measurements such as I and Ru. Some monitors that would continually radioisotopes can only be detected monitor and transmit to remote and Inspections locally, some as far away as 100 locations the current and voltage kilometers from the site of emis- in an electrical transmission line, The package of safeguards sion. In some cases, highly sensitive electrical substation, or power described in Section 4.5 has been modern analytic capabilities are line entering and leaving the adequate to prevent any covert needed, e.g., for sampling of emit- facility. diversion from safeguarded power ted particulates. Signatures in soil reactors to date. The international and vegetation generally cannot be • Neutron spectrum and fluence community has not yet encountered detected at ranges greater than monitors to indicate reactor per- a premeditated NPT breakout approximately 10 kilometers and formance between inspections. attempt by a determined and a would be useful in the neighbor- resourceful country relying on the hood of declared, and, in a special This additional monitoring has yet nuclear materials accumulated in a inspection, suspect facilities. While to be tested. For several of the sen- LWR plant. As discussed in Chapter environmental sampling cannot sors, new procedures and portable 2, the currently applicable IAEA provide 100 percent certainty of instrumentation must be developed. safeguards agreement with the detecting covert or illegal activities, DPRK (that follows INFCIRC 153) it faces a prospective diverter with Monitoring systems require reli- allows for additional measurements a significant chance of detection, able and sure means of data trans- and inspections beyond the meas- particularly if diverted material is mission to the analyzing organiza- urements and inspections at the not simply transported and stored, tion. The DPRK (or power reactor reactors that have just been but is involved in some industrial vendors) will also probably require described. In this section, we dis- process. The latter would have to secure data transmissions. Three cuss these additional measurements take place if spent fuel is to be separate requirements (reliability, and inspections briefly together reprocessed to yield its plutonium. surety and security) can be provid- with their utility. ed by e-commerce communication 4.6.2 Measures for Strengthened technologies.10 4.6.1 Measures for Strengthened Safeguards Under INFCIRC 153: Safeguards Under INFCIRC 153: Remote Monitoring Reliability of the data provides Environmental Sampling for recoverable data in the case of As discussed in Section 4.5.4, communication errors, potential Under INFCIRC 153, the IAEA field trials of remote monitoring of partial signal jamming, possible can carry out environmental sam- camera–seal systems are being suc- component failure, and power sys- pling only at places to which it oth- cessfully completed, though this tem vagaries. Surety is the portion erwise has access to at declared measure has not been routinely

55 SAFEGUARDS ON THE KEDO REACTORS of the data system that encapsulates 4.6.3 Measures for Strengthened prevent. A systems approach to the the signal and metadata (time- Safeguards Under INFCIRC 153: overall package of safeguards must stamps, identification, and status Other Recommended Steps be taken to ensure that the particu- records) to ensure an unambiguous lar additional measures are those data source and accuracy of that add the most assurance for the attached metadata. Security is com- Other recommended safeguard money invested. The authors of munication and data encapsulation measures include: this report have not had either the that ensures the message contents time or the specific data needed to are not usefully disclosed to a third • A Safeguards Center, a dedicated carry out such an approach, which party. Reliability is provided by room for safeguards equipment is rather the province of the IAEA. redundancy, error correction and and activities, is a recommended backup components, and systems. security consideration. 4.6.4 Satellite Remote Sensing Surety is provided by authenticated timestamps, digital signatures, and • A redundant stable, uninterrupt- Satellite-based remote sensing is checksums. Security is provided by ible power supply. Protection is not a part of the IAEA safeguards encryption, tamper-detection sys- needed against loss of power for package, but if carried out by the US tems and secure communications safeguards equipment. or other member state, is a useful channels. addition to safeguards. Satellite-based • Adequate training for the DPRK remote sensing may prove particular- Another necessary feature of any situation, including red-teaming ly applicable to detecting an “intent- unattended monitoring system is the particular physical and opera- to-abrogate” from requirements of automated review and analysis of tional circumstances. the NPT treaty and IAEA agreements. acquired data. The classic method is the review of video-monitoring data • Updating equipment and training Reliable and informative remote at video review stations after the on a regular basis. sensing suffers from a number of dif- fact. Typically, surveillance tapes are ficulties. The weather does not collected from the sealed monitor- Safeguards—no matter how well always cooperate, the observed party ing units by inspectors and designed—are only as good as the may conceal facilities, equipment, returned to a “home” facility for training, maintenance, and support and evidence from overhead view, review. The tapes contain time-lapse systems that implement them. and current commercial technology images of the items being moni- Given the DPRK’s record of limited (and marketability) limits GSD tored. Existing video review sta- cooperation, inspectors will need to (Ground Sample Distance) to a mini- tions allow the review of imagery at be trained to look beyond the con- mum of one meter, so that only items increased speeds. Yet, even with ventional monitoring and inspection the size of automobiles or larger can time compression of the monitoring points for unexpected activities. All be clearly distinguished.* Never- system and review speed, each hour of these recommended steps necessi- theless, remote sensing can be the of human review time can currently tate that the IAEA and its member only source of timely information only evaluate a little over a day of states pay special attention to the when trying to monitor broad areas, surveillance data from a single sen- funding additional inspectors and restricted facilities, or suspect sites in sor. This makes clear the very labor- equipment needed for safeguarding an uncooperative state. In particular, intensive nature of current video the KEDO reactors. the following can be done: review efforts. An automated assessment tool for the video or In conclusion, the additional What can be monitored? High- other signal data streams that detect measures and inspections resolution imaging systems can and assess changes and conditions described would clearly lower the monitor specific items and changes indicating safeguards significant probability of covert diversion at a nuclear site. Synthetic aperture events is needed. Otherwise it may from a safeguarded reactor. Absent radar can evaluate conducting items be weeks or months before informa- any instance of such covert diver- (fences, transmission lines, tion is reviewed. In certain cases, sion from a safeguarded reactor, it pipelines, railways, equipment, and this may be too late. is not possible to make this conclu- some industrial structures). sion quantitative. Red-teams could bring out some covert diversion Shipping casks. Individual fuel possibilities that these additional shipping casks are large enough to measures and inspections would be imaged from space, but the casks

* High-contrast items smaller than the GSD limit can be imaged, but they cannot be unambiguously identified.

56 SAFEGUARDS ON THE KEDO REACTORS need identifying marks to be site to be monitored. In the case of imagery supply of a particular tracked from satellites. The state of a SAR, weather conditions are not country. Hyper-spectral imagery at cask’s contents cannot be deter- important. Commercial satellites in high spatial resolution, not yet mined from space, but if infrared polar orbits revisit exact positions in developed, might identify specific imagery can determine the surface 15–26 day cycles. More frequent chemical species on surfaces and in temperature of a shipping cask, it imaging of locations can be per- gas and liquid plumes. This devel- might be possible to infer something formed by systems with the capabil- opment might provide significant about spent-fuel contents. Satellite ity to acquire images at angles other new monitoring capabilities. imagery could also be used to count than perpendicular to the earth’s casks at a dry storage facility. surface. This additional capability Figure 4-11 shows how the vari- allows images of a site to be ous additional monitoring measures Cooling units, associated acquired as frequently as once considered in this section could pumps, and exchangers. The ther- every 3–4 days. augment existing safeguards activi- mal state of elements of an industri- ties at the KEDO reactors. al site can be evaluated by infrared Remote-sensing resources. Table imagery. The thermal output of a 4-2 lists some commercial, space- reactor can be estimated from tem- based remote-sensing platforms.* A Table 4-2. Commercial remote-sensing platforms. perature changes in exposed cool- number of additional imaging sys- ing water flows and cooling struc- tems are planned for the next System Nominal Revisit resolution frequency tures. The operational status of site decade. The yet-to-be-deployed sys- (m) (days) facilities can be evaluated by moni- tems will generally not add signifi- toring the temperature of exposed cantly new technical capabilities in IKONOS 1 14 heat exchangers, steam pipes, the near term. Additional systems SPOT 10 4 HVAC equipment, power transmis- will provide greater revisit frequen- IRS 5 5 sion components, and gas-exhaust cy, and perhaps an independent LANDSAT 15 ? plumes.

Facility construction at declared facilities. High-resolution imagery of sites clearly identifies outdoor changes associated with construction. Electric Satellite power motor Earthmoving, grading, plowing, dis- surveillance turbances from heavy vehicles, build- ing construction, and security perime- ters are typically easily distinguished. Portal and Change-detection techniques allow rooms monitor IAEA analysts to focus quickly on changes KEDO Neutron spectrum Data LWR fluence monitor in sites between images acquired on Collection different dates, thus highlighting the effects of construction changes. Determining the nature of the con- Water Air struction and purpose of facilities are samples samples more difficult, but satellite imagery can direct the planning and conduct- Reactor Environmental engineering sampling ing of ground-based inspections.

What is the response time scale of satellite remote sensing sys- Figure 4-11. Additional monitoring measures that could augment existing safeguards activi- tems? The time scale for monitoring ties at the KEDO reactors. by satellite is determined by the orbital dynamics of the resources used and the local conditions at the

* LANDSAT is a US government system, but the images are available commercially.

57 SAFEGUARDS ON THE KEDO REACTORS 4.7 Conclusions Notes to Chapter 4 7. O.J. Heinonen, International As the foregoing discussion indi- 1. Korea Electric Power Atomic Energy Agency, cates, the IAEA safeguards under Corporation, Korean Standard “Additional Protocol and the applicable agreements—if car- Nuclear Power Plant (KSNP), Safeguards Agreements—The ried out with adequate funding, Seoul, Korea, 1996. New Non-Proliferation cooperation of the host country, and Standard,” Paper presented at preferably with augmentation from 2. Byung Ryung Lee, Keun Sun the 12th Pacific Basin Nuclear national technical means—provide Chang, and Jun Seok Yang, Conference, Seoul, Republic of high assurance that no nuclear- Korean Standard Nuclear Plants: Korea, October 30, 2000 materials diversion has taken place Safer, Simpler, and Easier to Build since the last inspection. Indeed, the (Korea Atomic Energy Research 8. Daniel A. Pinkston, Center for global record to date suggests that Institute, Nuclear Engineering Nonproliferation Studies, no fuel diversion from a civilian International, August 1992), pp. Monterey Institute of LWR under safeguards has taken 29–35. International Studies, place. Both the quantity and the “Implementing the Agreed timeliness criteria on which the 3. Korea Heavy Industries & Framework and Potential entire accounting and inspection Construction Co. Ltd. (Hanjung), Obstacles,” Paper presented at system are based have proven ade- Nuclear Power Plant, Seoul, the 12th Pacific Basin Nuclear quate so far. Including the addition- Korea, October 2000. Conference, Seoul, Republic of al measures permitted under the Korea, October 30, 2000. applicable agreements and dis- 4. Yousry Abushady, International cussed above, safeguards in our Atomic Energy Agency, 9. TESA Entry Systems, P.O. Box opinion will give high assurance “Inspection Activities at Light 620138, Atlanta, GA 30362-2138. that covert programs at or near any Water Reactors,” Paper present- inspected locale are not going on. ed at the 2000 Workshop on 10.D.R. Manatt, R.B. Melton, C.E. IAEA Safeguards, NTC/KAERI, Smith, and S.M. Deland, The international community Taejon, Republic of Korea, “International Safeguards Data has not yet encountered an overt August 9, 2000. Authentication,” Lawrence premeditated NPT breakout by a Livermore National Laboratory, determined and a resourceful 5. Byung-Koo Kim, Korea Atomic Paper presented at the 37th country relying on the nuclear Energy Research Institute, Annual International Nuclear materials accumulated in a LWR “KEDO LWR Project for Materials Management meeting, plant. That option cannot be com- International Cooperation and Naples, Florida, July 28–August pletely be ruled out but only pro- Non-Proliferation,” Paper pre- 1, 1996 (UCRL-JC-124825). tected against. In such a case, IAEA sented at the Monterey Institute safeguards—supplemented by of International Studies, other indications of diversion dis- Monterey, California, July 24, cussed in Chapter 5—should pro- 2000. vide timely warning, which is to say, warning time shorter than the 6. Neil Harms and Perpetua time required from diversion Rodriguez, International Atomic through fissile-material refining to Energy Agency, Safeguards at the manufacturing of a completed Light Water Reactors: Current nuclear-explosive device. Practices, Future Directions, IAEA Bulletin 38/4, pp. 1-6, Vienna, Austria, September 29, 2000.

58 CHAPTER 5 DIVERSION AND MISUSE SCENARIOS FOR LIGHT-WATER REACTORS

5.1 Scope and Intent PLUTONIUM AND WEAPONS This chapter briefly describes a few scenarios by which a Light- While all plutonium isotopes can be used in a weapon, not all iso- Water Reactor (LWR) could con- topes are equally desirable. So-called “weapons-grade” plutonium has ceivably be exploited for the 90 percent (or more) of the isotope 239Pu and little of the other iso- purposes of— topes. All the “even” isotopes cause problems because they emit neu- trons, which can lead to premature detonation and impaired reliability. 1. diverting spent fuel to support The isotope 240Pu is the most common of these. 238Pu is also prob- weapons acquisitions, or lematic because it produces a great amount of heat in addition to neu- trons. Even the “odd” isotope, 241Pu, is somewhat less desirable than 2. misusing the reactor to improve 239Pu because of its higher neutron and heat production rates. the quality of plutonium found Because all these other isotopes of plutonium are less desirable than in spent fuel. 239Pu, we can approximate the weapons-usable “quality” of reactor plutonium as indicated by the concentration of 239Pu. Exploitation of an LWR for pro- liferant purposes requires bypass- ing safeguards, either covertly or tion, either in the LWR or in other proliferator has the facilities and overtly (by abrogating treaty obli- types of reactors (assuming the capabilities for further enriching gations). Access to the reactor and cross-section libraries for other uranium. Other vulnerabilities to the spent fuel produced as a types of reactors are available). associated with LWR plants are also result of reactor operations is large- Such indirect scenarios are not dis- relatively limited. Misuse of ancil- ly limited to the operations cussed here. lary capabilities (such as glovebox- involved in refueling the reactor. es) or the application of skills, This chapter also looks briefly at 5.2 Scenarios knowledge, and expertise to a major signatures and other indica- weapons-development program tors that might serve as early The most important vulnerabili- offer little of value to potential pro- warnings of potential proliferant ties with LWR nuclear-power liferators and are capabilities rela- activities. We also look at some of plants are associated with plutoni- tively easy to obtain. the technical options that could um: either that generated normally reduce the consequences of diver- during reactor operations and The fact that spent LWR fuel is sion or misuse and those that can bound up in spent fuel, or that not ideal for weapons leads to the improve the reliability and/or potentially generated as the result second concern: the intentional decrease the response time of safe- of improper use of the reactor. misuse of the reactor for producing guards measures. Because plutonium is created in the plutonium with improved isotopic LWR fuel as it is being used, the composition. Two scenarios are There are scenarios by which an major proliferation concern is the offered that could be used by pro- LWR power plant can “indirectly diversion of spent fuel. As already liferators to covertly produce such support” the development of noted, although the plutonium in plutonium. A proliferator could nuclear weapons. For example, spent LWR fuel (so-called “reactor- substitute so-called “target” assem- LWR construction and operation grade” plutonium) is not ideal for blies (assemblies specially designed can justify the development of an nuclear weapons, it is considered to produce plutonium) in place of underlying infrastructure and usable, and thus an attractive target normal fuel assemblies, or could serve as a cover for a nuclear- for theft or diversion of spent fuel arrange to have some fuel removed weapons program. Some technolo- resulting from normal reactor oper- prematurely from the reactor by gies associated with LWR opera- ations. Vulnerabilities associated making it appear to be “failed tions, such as core physics comput- with fresh fuel are minimal, as fuel.” A common feature of these er codes, could be used for calcu- fresh LEU fuel has no value to a scenarios is that the material to be lating weapon-materials produc- potential proliferator unless that diverted in either case ends up in

59 DIVERSION AND MISUSE SCENARIOS FOR LIGHT-WATER REACTORS the spent-fuel storage pool. Thus, little plutonium is in an individual the fuel-pin sections into a shorter both of these scenarios share many fuel pin [a single Beginning of Life but larger diameter cask would also features. (BOL) pin contains only about exacerbate heat removal. 7 grams of plutonium, a normal The last two scenarios discussed burnup fuel pin contains about 21 Because of the difficulties associ- will be the overt misuse of the grams of plutonium). Thus, many ated with this variant, and espe- reactor to produce “weapons- (several hundred to a thousand) cially in light of the small amount grade” plutonium, either through would have to be diverted (proba- of plutonium in each fuel pin, this “short cycling” the reactor or bly over a long period of time) to appears an unlikely scenario for through modifications to the reac- accumulate a large-scale diversion. However, tor core design. While these two significant stock of plutonium. this scenario may serve as an scenarios could be attempted attractive prelude to abrogation, by covertly, the complexity and signa- Second, the diverted rods would providing material to perform ini- tures associated with these scenar- have to be replaced with fuel pins tial testing of a covert reprocessing ios are of sufficient magnitude that that matched those removed, both plant and associated weapons detection is essentially assured and in terms of appearance and possi- material infrastructures and abrogation of the Democratic bly radiation characteristics. If only designs. People’s Republic of Korea’s one or a few pins were removed (DPRK) international commitments from a given fuel assembly, match- 5.2.1 Scenario: Covert Diversion is required. We also briefly discuss ing the radiation characteristics of Spent Fuel Produced During the other options that the DPRK may not be required, as the signa- may have should it choose to abro- ture from the total fuel assembly Normal Reactor Operations gate its obligations. would change only a few percent, probably within the errors inherent The diversion of spent fuel The reactors for the DPRK use a in safeguards instrumentation. requires overcoming three obstacles: fuel assembly designed to be safeguards, radiation, and heat. “reconstituted,” that is, they are Third, the operations would designed to facilitate the replace- have to be carried out without First, to covertly divert signifi- ment of individual fuel pins by alerting the safeguards inspectors. cant quantities of spent fuel suc- the plant operators. This is done The fuel-reconstitution station is in cessfully, the safeguards surveil- to maximize fuel utilization in the the spent-fuel storage pool visible lance systems monitoring the event that a few fuel pins become to the safeguards cameras. While spent-fuel storage pool and access defective or damaged during the cameras can probably identify portals have to be compromised. operation. This fact leads to the that operations are proceeding at Current IAEA safeguards are possibility of a variant of the that station, they may not discern specifically tailored to address this spent-fuel diversion scenario, the the details of the operation. Thus, issue, and R&D continues on ways potential diversion of individual while a proliferator will need to to improve the reliability of spent- spent-fuel pins. justify fuel-reconstitution opera- fuel surveillance methods. tions on a specific fuel assembly, The main advantage of this vari- the extent of those operations may Second, even the least radioac- ant to a potential proliferator is not be observed. tive spent fuel requires transport in that a single spent-fuel pin, or even a heavily shielded container, and a few spent-fuel pins, might be Finally, even though it would be the activities associated with more easily removed from the easier to shield and transport a few removing spent fuel from storage, spent fuel building without detec- spent fuel pins, it is not trivial to do and preparing and loading it into tion. The radiation from a few such so without risk of detection. The the transport casks require suffi- pins can be shielded with a smaller fuel pins are long, so a shield cask cient manpower that much of the cask (several centimeters of lead would also be quite long, and operating staff would have to be pipe might be sufficient shielding,1 therefore likely visible to surveil- involved in the operation. and the heat load from a few pins lance cameras. The fuel pins could can be more readily dealt with by be cut or chopped into shorter Third, “young” spent fuel (spent filling the cask with water for short lengths, but that would require spe- fuel discharged less than several periods of time. cialized equipment, and even if years) produces a great amount of accomplished with minimal equip- heat and must be cooled to prevent There are several difficulties ment, it would release considerable damage or even melting (especially (from the proliferator’s perspec- radioactive fission gasses and likely of very young spent fuel.) tive) to this variant. First, relatively trigger radiation monitors. Packing

60 DIVERSION AND MISUSE SCENARIOS FOR LIGHT-WATER REACTORS

The combination of the latter two and advanced notice of spent-fuel feasible to introduce a small num- obstacles places tough requirements storage facility activities involving ber of target assemblies specially on the design of the shipping cask. the movement of spent fuel is nor- designed to improve the production Shortcuts in the handling and trans- mally required. Related signatures of weapons-quality plutonium. Lu, port of young spent fuel greatly include evidence of intrusion or et al. estimated that as much as 8 increase the risk of fuel damage and illicit activities detected directly by kilograms of plutonium/year could poses significant personnel hazards. the monitoring equipment and evi- be produced in a Pressurized-Water dence of manipulation of the moni- Reactor (PWR) by populating nor- Both the radiation and heat pro- toring equipment (perhaps mally empty control-rod guide duced by spent fuel decays slowly observed as failures in the monitor- tubes with special target pins, and with time, so older spent fuel can ing equipment). Raising spent fuel that this might be accomplished represent a greater diversion risk from the storage pool increases the without substantially affecting reac- than fresher spent fuel. Because the radiation level in the facility and is tor operating characteristics.1 The oldest spent fuel (the BOL spent readily detectable. physics and cycle length of the LWR fuel) is also the spent fuel with the core make it unlikely that a target highest quality plutonium, this ten- Because the diversion of spent assembly could produce high-quali- dency is reinforced, and there is a fuel requires the transport of the ty plutonium if left in the core for desire (from a nonproliferation material away from the reactor site, the entire 4.5-year life of a normal view) to preferentially relocated the a transport cask is needed. Even a fuel assembly. However, if left in the oldest spent fuel and place it under cask designed to bare minimum core for only a single cycle, the plu- more effective international control. requirements will be large and tonium produced by special target bulky, and likely observed from assemblies could have as much as Spent-fuel storage pools do not surveillance satellites, although the 90 percent 239Pu. normally have enough capacity to cask may be on site for a relatively store all the spent fuel discharged short time (on the order of hours.) During refueling, LWR operators during the life of a plant. As spent inspect fuel assemblies before rein- fuel ages, and the heat and radiation For spent-fuel diversion to serting them back into the core; generated by the spent fuel lessen, remain undetected, the diverted they inspect more carefully if fuel the spent fuel can safely be stored in spent fuel must be replaced with a damage is indicated. It is feasible dry casks at the reactor site. These dummy that mimics the appear- that a potential proliferator could casks are very large, bulky, and diffi- ance and characteristics of the arrange to have some fuel or target cult to transport without specialized diverted fuel assemblies. assemblies modified to appear equipment. Removing fuel from the Safeguards practices include peri- defective as a justification for casks is a difficult procedure owing odic surveys of spent fuel to verify removing them at relatively low to the high radiation field. The casks both the appearance and the burnups. have tamper-proof seals to detect nuclear characteristics of the mate- any attempt to open the casks. rial. A dummy fuel assembly with The possibility of defective fuel Safeguards of dry-cask storage areas such characteristics presents simi- presents additional safeguards rely on surveillance monitoring and lar radiation hazards to normal implications. Any fuel (either on periodic inspections of inventory spent fuel. This in turn requires failed or not) removed before and cask-seal integrity. The size of remote handling during the manu- achieving nominal burnup will the casks makes them detectable facture, necessitating potentially have a plutonium isotopic compo- from satellites, providing additional observed facilities. Transport sition similar to BOL fuel, and thus surveillance certainty. Thus, diver- requires a similar (or same) ship- represent a more attractive diver- sion attempts on dry-cask storage ping cask as normal spent fuel. sion target. Thus, there is some areas incur difficulties similar to Movement of the dummy assembly incentive for verifying that only diversion attempts from spent-fuel into the spent-fuel storage area truly failed fuel is removed prema- storage pools, but they could also be entails the same observables as turely from the reactor. The fuel detected from surveillance satellites. moving the real spent fuel out. designed for these reactors is expected to be “reconstituted,” Diversion of spent fuel has a 5.2.2 Scenario: Covert Materials that is, in the event of fuel failures, number of signatures, especially in a Production individual fuel pins can be once-through fuel cycle. First, the replaced and the bundle reinserted movement of the spent fuel itself is The ability to covertly misuse an into the reactor. On the one hand, observed via the installed safe- LWR to produce better quality plu- this increases the number of guards measures. Cameras and por- tonium than that found in normal accountable items because individ- tal seals verify the security of access, spent fuel is limited. It is technically ual fuel pins become accountable.

61 DIVERSION AND MISUSE SCENARIOS FOR LIGHT-WATER REACTORS

On the other hand, it reduces the An advantage (to the prolifera- 5.2.3 Scenario: Short-Cycling accumulation of low-burnup spent tor) of this scenario is that it pro- the Reactor fuel because only those failed fuel vides a mechanism for producing pins remain in the spent-fuel spent fuel of higher plutonium A reactor can be operated for storage pool. quality within the normal operat- less than its normal cycle, reducing ing procedures of the reactor. burnup and improving the quality Such a scenario requires several Conversely, the use of reconstitut- of the plutonium in the spent fuel. steps. First, an expectation of failed ed fuel assemblies significantly This “short-cycling” can be either fuel needs to be established to pre- reduces the amount of low-burnup covert or overt. Covert short- pare the inspector to accept the spent fuel that could be accumulat- cycling of the reactor is limited to unusual fuel change-out. This ed under such a scenario. premature discharge of only a few requires spoofing the reactor failed- fuel assemblies, perhaps as “failed fuel monitoring equipment. Such Such a scenario has several fuel” as discussed previously. an expectation also serves to mask observables. First, utilization of spe- Premature discharge of more than the true intent of the mis-identifica- cial target assemblies to optimize one or two such assemblies is so tion of the failed fuel. Alternately, a plutonium production requires unusual that detection is essential- fresh fuel element might be covert- obtaining such assemblies and ly certain and the scenario must be ly “pre-damaged” prior to the ini- introducing them into the normal considered overt. There are few tial insertion into the reactor so that fuel supply. Arranging for fuel “legitimate” reasons for short- it would be a truly “failed” fuel ele- assemblies to fail (or to appear cycling a reactor, and these are ment at the next refueling. Because failed) requires similar efforts. generally for safety-related issues fresh fuel will be imported into the Accomplishing such a feat requires such as component failure (includ- DPRK, pre-damaging it would fooling of safeguards, either by fal- ing severely failed fuel), system require cooperation from the coun- sifying records and/or swapping leaks, and control problems. Most try of origin, presumably the target assemblies for real fuel safety- and control-related reactor Republic of Korea (ROK), or the assemblies, as well as avoiding shutdowns do not necessitate fuel would have to be covertly detection during fuel handling and opening the reactor, let alone damaged at the plant site. fuel inspection. Safeguards practices handling fuel. are specifically designed to detect Second, the “failed” fuel would the introduction of extraneous fuel Thus, if potential proliferators have to be “identified,” likely assemblies into the fuel supply. wanted to short-cycle a reactor to through spoofing of fuel inspection covertly produce plutonium of procedures or by contaminating (or The second observable occurs in higher quality, they need to con- otherwise camouflaging) the fuel the need to manufacture the target coct a relatively elaborate scheme to appear failed. assemblies, which would require to make the early shutdown the acquisition of a number of appear justified, especially one Third, the evidence of failed fuel items and materials subject to vari- that requires even partial unload- would have to be sufficient to con- ous export and nuclear-supplier ing of the reactor core. Even if vince the on-site refueling inspector. restrictions. The target assemblies such a covert scenario were fol- have to be indistinguishable from lowed, it would be limited to the From there, the “failed” fuel the real fuel assemblies, have the diversion of very few fuel assem- would be placed in the spent-fuel serial numbers of the real fuel blies, owing in part to the necessi- storage pool. Generally, failed or assemblies, and the real fuel ty to have replacement fresh fuel suspect fuel is first placed in a spe- assemblies being replaced have to available. Normally, a reactor cialized storage cask (or sleeve) to be covertly removed from the site. operator maintains a small stock isolate the failed fuel and minimize of fresh fuel in the event that a potential for contaminating the The third observable in the sce- fuel element requires replacement, spent-fuel storage pool. Once in nario requires some justification for but fuel is relatively reliable, and the storage pool, the fuel is no removing the special assemblies the necessary stock of surplus fuel more or less vulnerable to diver- before the end of the normal fuel is small. sion than any other spent fuel, life, most likely after a single cycle. with the slight possible complica- In addition, once the short-cycle tion of (perhaps) needing to Finally, even if such assemblies fuel was removed from the reactor, remove the fuel assemblies from are successfully irradiated, they it would eventually have to be the failed-fuel storage cask. end up in the spent-fuel storage diverted from the spent-fuel stor- pool and must be covertly age pool, as in Scenario 1. removed.

62 DIVERSION AND MISUSE SCENARIOS FOR LIGHT-WATER REACTORS

Short-cycling a reactor creates a reload) would contain approxi- reactor core to achieve significant number of observable signatures mately 50 kilograms of plutonium increases in the quantity and quali- related to the loss of power. First is with about 90 percent 239Pu. In ty of discharged plutonium. the loss of power to the grid that such a case, the primary observable may be detected if the power grid would be the premature shutdown Such modifications are expen- has “external” connections (espe- of the reactor. sive and time-consuming and cially if significant power is nor- require significant reactor-design mally exported, or is monitored. We note again that the previous capabilities. Moreover, the modi- Second is the loss of “dump heat” two paragraphs apply to overt fied reactor would require fuel sub- that can be remotely detected by diversion, in which case the DPRK stantially different from that nor- the loss of infrared signals.* Both abrogates the Agreed Framework mally used, probably requiring signatures should be detected (AF) and withdraws effectively much higher enrichments than nor- promptly, and certainly within the from the Non-Proliferation Treaty mal LWR fuel. Thus, besides the time needed to cool the reactor (NPT). design and manufacturing efforts prior to removing the head and associated with the core reconfigu- gaining access to the fuel (which 5.2.4 Scenario: Overt ration, a source of fuel and enrich- normally takes a few days). Reconfiguration of the Reactor ment capabilities also has to be provided. Diverted short-cycle spent fuel for Materials Production must be replaced if the reactor is to The modifications necessary to remain operating. Thus, replace- While LWRs can be short-cycled achieve these goals could not go ment fresh fuel has to be obtained to improve the isotopic composi- unnoticed under any inspection (probably in advance of the reactor tion of plutonium found in the regime, and an overt abrogation of shutdown) and this has to occur spent fuel, such an approach is treaty responsibilities would many months sooner than for nor- costly—both in terms of operating appear necessary. While design mal reactor operations. The early costs and lost revenues. Weapon- and construction of the modifica- accumulation of fresh fuel would material production through the tions might proceed before any thus indicate possible reactor use of special target assemblies or overt display of intent, the lack of short-cycling. through the ruse of failed fuel known enrichment capabilities in offers very limited capabilities for the DPRK means that an external In the event of overt short- material production. To provide source of fuel would be an essen- cycling of the reactor with a goal to significant increases in material tial element of such an approach. obtain plutonium with 90 percent quality and quantity requires the 239Pu, burnup would be limited to introduction of a large number of In a worst-case scenario, one in no more than about 7 MWd/kg, target assemblies. Target assemblies which all modifications were com- limiting the reactor cycle time to using depleted uranium, arranged patible with the existing core inter- approximately 9 months. as a blanket surrounding the core nal structures and all materials Assuming an industry average of (a region of reduced neutron flux) (including both modified fuel 40 days per refueling outage and would help extend the cycle time of assemblies and target assemblies) refueling of the entire core, the the reactor, while helping to were available prior to modifying reactor could produce as much as improve the quality of the plutoni- the reactor core, the substitution 150 kilograms of plutonium (con- um produced. However, such an could not proceed without notice taining approximately 90 percent approach significantly reduces the by the inspectors. On the other 239Pu) every 10 months.† total amount of reactivity in the hand, it may be feasible that the core (a measure of how much modifications could proceed in If the goal is to obtain a limited “active” fuel is in the core), signifi- approximately the same time as quantity of 90 percent 239Pu with a cantly affects the thermal and required for a normal refueling minimum of observables, then the nuclear performance of the reactor, (perhaps as little as 15 days, but reactor could simply be shut down and dramatically affects the per- more likely 30–60 days). This within about 9 months of the last formance of the reactor control sys- means that the rest of the world refueling. At equilibrium, one-third tems. These effects lead to the need could expect at least 15 days to of the core (the last fresh fuel to substantially reconfigure the respond to an abrogation before

* Only about one-third of the power produced in the reactor is turned into electricity, the remaining two-thirds is wasted and dumped to the environment. The dumped-heat infrared signature of an operating reactor is easily detected from satellites. † These data are obtained from the rough approximations of plutonium production in LWRs described earlier. Plutonium production rates vary broadly with fuel enrichment, and using lower fuel enrichments would likely reduce cost and optimize plutonium production. Significant additional analysis is required to determine more accurate estimates of fuel requirements and of material production and plutonium isotopic contents at these lower burnups.

63 DIVERSION AND MISUSE SCENARIOS FOR LIGHT-WATER REACTORS reactor restart (and some additional Table 5-1. Accumulation of Light-Water Reactor (LWR) plutonium in spent fuel. time—on the order of several Discharged Discharged Discharged Cumulated weeks—before significant plutoni- Time from First Low-Burnup Medium-Burnup High-Burnup Total Pu um could be accumulated.) Startup (years) Pu (kg) Pu (kg) Pu (kg) (kg) 1.25 100 0 0 100 5.3 Consequences 2.5 — 175 0 275 The primary consequence of 4 ——275 550 diverting spent fuel from an LWR 5.5 ——288 838 is tied to the quality and quantity —— — —— of the plutonium contained in the 40 100 175 288 7,737 spent fuel. Normal reactor opera- tions result in the discharge of Total Discharges 200 350 7,187 7,737 nearly 300 kilograms of reactor- grade plutonium every 18 months, with lesser discharges of somewhat higher quality plutonium during 9000 the beginning of the reactor life. 8000 Cum. Low BU Pu The spent fuel discharged during Cum. Med BU Pu 7000 the beginning and end of reactor Cum. Hi BU Pu life, while still not considered 6000 Cum. Total Pu “weapons-grade” would be some- what less difficult to design and 5000 fabricate into a weapon than the 4000 higher burnup equilibrium fuel. Because all plutonium produced as Plutonium (kg) 3000 a result of LWR operations is con- 2000 sidered “weapons-usable,” these differences must be considered dif- 1000 ferences in degree, not differences 0 in kind. An approximation of the 0 5 10 15 20 25 30 35 40 projected accumulations of spent Time from startup (years) fuel is summarized in Table 5-1 and Fig. 5-1.* Figure 5-1. Plutonium discharged from LWR operations (BU = burnup).

In addition to these spent-fuel of low-burnup plutonium, 100 kilo- potential for such attempts. Thus, discharges, there is also partially grams of which is normally dis- production of substantial quanti- spent fuel in the reactor at all charged. Thus, should the DPRK ties of low-burnup plutonium and times. At any refueling, there decide to abrogate its obligations at subsequent diversion of it likely remains in the reactor one-third the end of the first cycle, there requires an overt abrogation of core of 1-cycle fuel (containing would be approximately 300 kilo- treaty obligations. approximately 100 kilograms of grams of low-burnup plutonium low-burnup plutonium) and one- available. Exploitation of the plutonium third core of 2-cycle fuel (contain- produced in an LWR requires ing approximately 275 kilograms As already discussed, overt reprocessing to extract the plutoni- of medium burnup plutonium). At short-cycling of an LWR could pro- um. Although reprocessing of the end of the reactor lifetime, this duce as much as 150 kilograms of spent LWR fuel is similar to the “in-core” fuel is discharged (shown essentially “weapons-grade” pluto- reprocessing of spent fuel from in Table 5-1 and Fig. 5-1). nium roughly every 10 months, graphite reactors (of which the and there some potential that DPRK has experience), there are This in-core fuel is of most con- smaller amounts could be accumu- some notable differences. First, cern during the reactor’s BOL peri- lated through some of the scenar- LWR fuel is clad with zircalloy, a od. As previously noted, at the end ios discussed here. Safeguards very durable alloy that complicates of the first cycle, the core fuel con- have been designed to help detect dissolution of the fuel (as com- tains approximately 300 kilograms such activities and minimize the pared with magnesium-clad metal

* The values for accumulated plutonium shown in Table 5-1 and Figure 5-1 refer to the plutonium content of freshly discharged spent fuel and do not account for the decay of plutonium, mostly associated with the decay of the short-lived isotope 241Pu.

64 DIVERSION AND MISUSE SCENARIOS FOR LIGHT-WATER REACTORS fuel that is easily dissolved.) The nature, are quite robust against fuel, reduces safety, and compli- LWR fuel must be mechanically misuse or modification for weapon- cates transport, inspection, and chopped into small pieces, and the materials production. As has been handling. Moreover, such an toughness of the alloy makes this a discussed, such misuse is expen- approach deals only with the pro- difficult process. Second, the com- sive, time-consuming, and involves liferation risk of fresh LEU fuel, a plexity and size of the LWR fuel sufficient, easily detected observ- risk already considered very low. assembly (hundreds of very long ables. Because the DPRK reactors As such, it is unlikely that such pins as opposed to the single short are copies of existing designs, there measures would be willingly slug of a graphite reactor fuel ele- are few, if any, substantial modifica- accepted by any plant operator. ment) further complicates this tions to the reactors or the plants process. Third, the combination of that could significantly reduce the The issue of the higher-quality the zirconium and oxide fuel risk of misuse, a risk that appears spent fuel resulting from BOL makes the chemical dissolution relatively low to begin with. operations can be eliminated by step itself slower than dissolution Remote monitoring of plant-operat- initially fueling the reactor with of the magnesium-clad metal fuel ing characteristics, both directly partially irradiated fuel, such that used in the graphite reactors. (through telemetry of plant param- all BOL spent fuel would be Finally, spent LWR fuel has higher eters) and indirectly (through exposed to full design burnup on radiation levels than graphite remote surveillance), discussed in discharge. That approach intro- reactor fuels, due both to its high- Chapter 4, is an area of technology duces a number of technical and er burnup and the fact that the that could improve the ability of operational challenges and is out- fuel assemblies themselves are the safeguards community to detect side the bounds of current reactor much larger. In short, even if an and reduce the attractiveness of operations. Partially irradiated fuel active reprocessing capability for any of the scenarios discussed. rods are much more prone to dam- the graphite reactor fuel is avail- age during handling and transport able, a new front-end to the repro- Increasing the burnup capability than fresh fuel and introduce sig- cessing facility must be provided of LWR fuel is the most promising nificant risks for fuel failures on re- to prepare the spent fuel for repro- near-term approach to enhancing irradiation. The operational chal- cessing. This front-end needs to be the intrinsic proliferation resistance lenges include, among others (1) co-located with and likely con- of the LWR fuel cycle, including identifying a source of the partially tiguous to the rest of the repro- those in the DPRK. In addition to irradiated fuel, (2) licensing the cessing facility, as the dissolved reducing the quality and overall partially irradiated fuel following fuel is too radioactive to be rea- quantity of plutonium resulting the additional handling and trans- sonably shipped. from LWR operations, it can reduce portation involved, and (3) ensur- overall fuel cycle costs and thus is ing personal protection and facility 5.4 Preventive Measures an attractive option for the reactor safety when introducing partially operator, and could be implement- irradiated fuel into the reactor As just described, reducing the ed within the AF. Research is operations normally designed to proliferation risk associated with underway, both in the US and handle fresh fuel. None of these potentially weapons-usable pluto- abroad, to extend the burnup capa- steps are even remotely practical or nium generated by reactor opera- bility of LWR fuel, and this option economic. tions fall into several broad areas: could become available in time. Diversion of individual fuel pins 1. Reducing access to plutonium- Other technologies have been can be impeded by eliminating the containing materials, proposed to further enhance the ability to reconstitute fuel. This is proliferation resistance of LWR technically easy to accomplish; 2. Reducing the quantity or quality plants and associated fuel cycles. indeed, not all LWR fuel is designed of the plutonium in spent fuel, Most, if not all, of these options fall to be reconstituted. However, sever- far outside the bounds of current al issues need to be solved before 3. Reducing the desirability of mis- LWR practice and are likely unac- implementing such an option. First using the LWR for weapon- ceptable within the guidelines of is licensing. The current licensed material production. the AF. One example is the addi- fuel design for these reactors is tion of substances to fresh fuel to reconstitutable, and any new design For practical purposes, all covert make fresh fuel radioactive as a would have to be licensed. Second, scenarios for misusing an LWR for deterrent to theft and diversion. is the issue of reciprocity. Because weapon-material acquisition can be Such an approach, while feasible, the ROK reactors use reconstitutable reduced to the issue of spent-fuel introduces complications that sig- fuel, the DPRK is unlikely to accept diversion. LWR power plants, by nificantly increase the cost of the different treatment. Third is the

65 DIVERSION AND MISUSE SCENARIOS FOR LIGHT-WATER REACTORS cost. A new fuel design would be Section 4.6. These methods, under Notes to Chapter 5 more costly (mainly due to licens- the conditions recommended, pro- ing costs), but would also be slight- tect against covert scenarios so that 1. M.S. Lu, R.B. Zhu, and M. ly more costly to operate. Fuel is other methods of attaining a Todosow, Unreported Plutonium designed to be reconstituted to gain weapon capability, other than Production in Light-Water maximum life out of each fuel covert diversion from the KEDO Reactors, Brookhaven National assembly. A failed single pin in a reactors, would probably be chosen Laboratory, Upton, New York, non-reconstitutable assembly can by any proliferator. Overt diver- ISPO-282, TS-88-1 (February lead to premature discharge of the sion, the abrogation of agreements, 1988). entire assembly. cannot be prevented by safeguards, though they can be warned about In summary, the best practicable and damage can be limited. This methods for preventing the diver- overt scenario is discussed in sion scenarios outlined are the Chapter 8. ones described in Chapter 4,

66 CHAPTER 6 KNOWN AND SUSPECTED NUCLEAR-RELATED FACILITIES IN THE DPRK

6.1 Inspection, the completion of construction of Kumho, remains to be negotiated. Dismantlement, and the LWR buildings and the deliv- The technical problems do not Disposal Requirements for ery of the turbine generator are still appear to be major, but how the dis- a year or more ahead. mantlement will take place and Existing Nuclear Facilities where the dismantled parts of facili- 6.1.2 Disposal of Spent Fuel ties will go remains to be decided. 6.1.1 “Full-Compliance” When the transfer of key nuclear Inspections by the IAEA 6.2 The Facilities at components to Kumho for the first Yongbyon* LWR takes place, the DPRK is obli- The Agreed Framework (AF) pro- gated by the AF to begin transfer of The main site for the North vides that the Democratic People’s the spent fuel stored in the cooling Korean nuclear program was the Republic of Korea (DPRK) must pond near the small graphite reactor Yongbyon Nuclear Center, about come into “full compliance with its at Yongbyon. Transfer will presum- 100 kilometers north of safeguards agreement with the ably continue during the period of Pyongyang, on the Kuryong River International Atomic Energy delivery and installation of key (Fig. 6-1). Many of the facilities at Agency (IAEA)” when a “signifi- nuclear components for the second the site have been “frozen” as a cant portion of the Light-Water LWR at Kumho because the spent Reactor (LWR) project is completed, fuel is all to be transferred to “ulti- but before delivery of key nuclear mate disposition” by the time that components.” This means comple- LWR is completed at Kumho. The tion, to a stage still to be agreed on, AF does not specify where the spent of the buildings for the turbine gen- fuel is to go but says that DPRK and erators for the first LWR at Kumho the US will “cooperate… to dispose and delivery of its turbine generator of the fuel in a safe manner that but not the LWR itself or key does not involve reprocessing in the nuclear components for it. Thus, DPRK.” Thus, where the spent fuel before IAEA inspection is required will go and how it will be transport- of facilities other than those ed still remains to be decided. “declared” to the IAEA by the DPRK when the DPRK accepted its 6.1.3 Dismantlement of Gas- safeguards agreement, the Korean Graphite Reactors Energy Development Organization (KEDO) must complete major con- When the first LWR is completed struction of buildings at Kumho at Kumho, the dismantlement of the and deliver much of the non- three existing gas-graphite reactors nuclear equipment for electric gen- and “related facilities” at Yongbyon, eration. The specific stage of con- Taechon, and elsewhere must begin. struction and delivery required is to This dismantlement must be com- N be agreed in a “delivery protocol.” pleted by the DPRK when the LWR project (including installation of the Believing that the inspections second LWR) is completed at and analysis necessary to show Kumho. More agreement than this Figure 6-1. Satellite image of the Yongbyon “full compliance” will take two to on how this dismantlement will be site in 2000 (used with permission from three years, the IAEA has been scheduled, including how it will be Space Imaging). The facility extends on pushing the DPRK to permit these coordinated with steps toward both sides of the river throughout essen- inspections to begin even though installation of the second LWR at tially the entire photograph.

* Much of Sections 6.2 and 6.3 is taken from David Albright and Kevin O’Neill, eds., Solving the North Korean Nuclear Puzzle, Institute for Science and International Security Press, Washington, D.C., 2000.

67 KNOWN OR SUSPECTED NUCLEAR-RELATED FACILITIES IN THE DPRK result of the Agreed Framework 6.2.1 The IRT-2000 Research site is commonly associated with (AF) of 1994, but the site as a Reactor IRT reactors. Iraq, for example, has whole is still active. an almost identical facility at the In 1965, the IRT-2000 reactor was Tuwaitha Nuclear Research Center. The site includes facilities for commissioned in Yongbyon. The It has two dry wells for storage of fuel manufacture, three nuclear reactor is light-water-cooled and irradiated fuel elements, plus two reactors, at least one hot-cell facili- moderated, uses enriched uranium cylindrical tanks for storage of liq- ty, at least one spent-fuel repro- fuel, and was designed to operate uid radioactive waste. The site at cessing facility and several waste at a power level of 2 MW(th) [later the Yongbyon facility was covered sites. There is also a high-explo- the power level was increased to with dirt in August 1992, and the sives testing area, which has been 4 MW(th) and then 8 MW(th)]. road leading to the site was hidden inactive since at least 1992. Soviet nuclear reactor specialists by freshly planted trees. Some of departed the DPRK after the reac- the trees then died within a few The North Koreans reported to tor was completed, but continued months, and more vegetation was the IAEA that the site was dedicated to cooperate on agreements to sup- planted. The IAEA has not been to the pursuit of peaceful nuclear ply the enriched fuel.1 The reactor allowed access to the waste stored energy and there was no weapons was placed under IAEA safeguards at this site, but it is suspected that program; however, inspections, in 1977. This reactor is not frozen further waste from processing IRT measurements, and satellite photo- and is not scheduled to be disman- reactor targets is stored there. graphs have shown suspicious tled under the AF. activities. As of this writing, it has 6.2.4 Declared Waste Site not been determined exactly which 6.2.2 Isotope Production of the Yongbyon facilities were for Laboratory Near the IRT-2000 Nearby, there is a declared waste weapons-manufacturing and which site that did not exist until mid-1992. (if any) were for energy production. Reactor The North Koreans, nevertheless, It is possible that entire site was a stated that this site has been active dedicated weapon facility that used In the northern part of since 1977, holding solid waste in 28 a peaceful cover story. Given the Yongbyon, there is a set of labora- steel-lined storage pits (upgraded to activities observed, it is unlikely that tory buildings including one desig- 42 pits in 1990). The deliberate con- the site was dedicated entirely to nated as the Isotope Production cealment of the older waste site, peaceful purposes. Laboratory. In this building, there together with the false history of the are hot cells for handling radioac- declared site, make it appear that Uranium mining, milling, and tive material. This is where North the IRT-2000 reactor produced sig- refining operations were carried Korea first produced plutonium, nificant amounts of plutonium, and out at other sites in North Korea. generated in uranium targets in the that the wastes have been concealed. The DPRK has sufficient domestic IRT-2000 reactor. They have admit- uranium resources to be self-suffi- ted to the production of less than 6.2.5 Graphite-Moderated cient. Estimates of the capacity of of gram of the material by this Reactor North Korea’s uranium mining means. This laboratory is not and milling operations in 1992 frozen under the AF. In the early 1980s, construction of range from 300 tonnes to 1,800 a nominally 5-MW(e), 20-MW(th) tonnes per year. There are no 6.2.3 Probable Undeclared graphite-moderated power reactor known uranium isotopic-enrich- Waste Site South of the was started in Yongbyon. A decision ment facilities in the DPRK. The to seek nuclear weapons could reactor program was to operate IRT-2000 Reactor already have been made at this with fuels of natural enrichment, time, and the graphite-reactor tech- with the exception of the Soviet- Near the (uncompleted) 50- nology would be easily adapted to supplied IRT-2000 research reactor. MW(e) graphite reactor is an unde- a dual-use role. The reactor design clared waste site that was probably is said to be based on the Calder In the rest of this section, we in operation from the 1970s until Hall reactors, originally built to pro- focus on the main facilities, declared August 1992. From satellite duce electricity and support the and suspected, at Yongbyon. imagery (Fig. 6-2), it was deter- British nuclear-weapons program. mined that the undeclared waste site contains two cylindrical tanks, The Yongbyon design uses urani- each about 5.8 meters in diameter, um-alloy (99.5 percent uranium, in addition to a rectangular 0.5 percent aluminum) fuel rods in arrangement. This type of waste magnesium-alloy (99.5 percent Mg,

68 KNOWN OR SUSPECTED NUCLEAR-RELATED FACILITIES IN THE DPRK

0.5 percent Zr) tubes. The rods are thick but there are longitudinal fins, The reactor core consists of a set 2.9 centimeters in diameter and 52 bringing the outer diameter to of large graphite blocks, with 812 centimeters long and contain 6.24 about 5 centimeters. Cooling is by vertical holes (“channels”) bored in kilograms of the uranium alloy. The forced convection with carbon diox- them for the fuel and coolant. The tube walls are roughly a millimeter ide gas under 6 bars of pressure. channels are 6.5 centimeters in diameter. There are 10 fuel rods per New waste site channel, stacked vertically. The diameter of the core is 6.4 meters and the height, not including reflec- tors, is 6 meters. While the core is about half the physical size of the Calder Hall units, the maximum 50-MW50-MWee thermal power is only about one- reactor tenth that of the Calder Hall units.

6.2.6 Fuel Reprocessing Facility During the late 1980s, a fuel- reprocessing facility was construct- Buried ed. The facility is housed in a waste site building 192 meters long and six stories tall, quite visible in satellite photographs (Fig. 6-3). In 1994, the facility had a nominal capacity to reprocess 220–250 tonnes of Magnox spent fuel per year, using two PUREX processing lines, if the lines are operated 24 hours per day, 300 days per year.*

The spent fuel from the Magnox reactor would be transported to the south end of the reprocessing facili- ty by truck in buckets within shield- ed casks. The buckets and fuel rods would then be moved to the north side of the building by remotely operated vehicle and the processing occurs in the southerly direction.

TunnelTunnel Along the eastern side of the building is a complex set of storage tanks within shielded vaults that contain liquid processing wastes of low- and intermediate-radiation levels. At the southeastern end of the building is a vault containing two storage tanks for highly radioactive fission-products. About 120 meters east of the reprocessing Bridge building is an L-shaped building associated with the storage of high- ly radioactive waste. There are four Figure 6-2. The buried and the new waste site near the uncompleted 50-MW(e) graphite tanks just south of this building in reactor. Only the new waste site was declared and the IAEA has not yet been allowed to an underground vault that contain inspect the buried waste site (used with permission from Space Imaging).

* The PUREX process, or Plutonium Uranium Reductive Extraction, is used ubiquitously to reprocess spent fuel. See Benedict, Pigford, and Levi, Nuclear Chemical Engineering (McGraw-Hill Book Company, New York, 1981).

69 KNOWN OR SUSPECTED NUCLEAR-RELATED FACILITIES IN THE DPRK

with the reprocessing facility) and the Taechon reactor was to be for electrical production only. Suspect waste building Cooling tower 6.3 Verification of the Initial DPRK Declaration The IAEA inspections were Stack (shadow extends NW) meant to verify the correctness and completeness of the initial declara- tion of nuclear materials that the DPRK provided to the IAEA on Spent fuel May 4, 1992. The declaration reception includes statements as to—

1. The nuclear-material inventory of seven facilities declared by the DPRK to the IAEA as subject to safeguards.

2. Design information of those Plutonium seven facilities. separation building 3. A list of locations of nuclear materials outside these facilities.

Figure 6-3. Satellite image showing the plutonium separation building and associated build- 4. A list of nuclear facilities under ings and the suspect waste building, the so-called Building 500 where nuclear waste from construction or planned. undeclared reprocessing campaigns is suspected to be hidden. The waste could include 50 tons of uranium and 30 kilocuries of 137Cs (used with permission from Space Imaging). 5. A list of scientific institutions. most of the volume of the declared from undeclared reprocessing cam- 6. A list of nuclear facilities related reprocessing high-level waste. paigns in 1989-91. Inspectors who to the nuclear industry. visited this building during the 6.2.7 Undeclared Waste third ad hoc inspection of The first IAEA visit occurred on Storage Building September 1992 were told incorrect- May 11–16,1992, and was followed ly that this building had no base- by six ad hoc inspections to deter- An undeclared waste storage ment, and that it was a workshop mine the correctness and complete- building (sometimes called for military vehicles. ness of the initial declaration. These Building 500) is located about 300 inspections and some meeting and meters east of the main reprocess- 6.2.8 Other Unfinished Reactors correspondence dates associated ing building. This building, built with those inspections are listed in primarily in 1991, is 18 meters high Construction began on two other Table 6-1. Attempts to inspect two (including the basement), 24 meters graphite reactors, one a nominally undeclared facilities at the wide, and 67 meters long. The 50-MW(e) reactor at Yongbyon and Yongbyon site in early 1993 under basement has four large pits for liq- another a 200-MW(e) reactor at “special inspection” authority were uid-waste storage tanks and six Taechon. Neither of these two reac- blocked. The DPRK then smaller compartments for storage tors was completed. The reactor announced that it was withdrawing of containerized solid wastes. The graphite had been installed at the from the Non-Proliferation Treaty basement is covered with concrete 50 MW(e) reactor but not at the (NPT). Although its withdrawal slabs for shielding. Trenches were 200 MW(e) reactor at the time of was suspended, the DPRK has dug and pipes were installed from the freeze. never allowed IAEA special inspec- the main reprocessing building in tions. The IAEA began conducting the winter of 1991-92. It is suspect- It has been speculated that the inspections at the 5-MW(e), ed that these pipes pumped aque- 50-MW(e) reactor at Yongbyon was 20-MW(th) gas-graphite reactor and ous radioactive wastes containing to have been the main producer of the other two unfinished graphite fission products and spent uranium weapon materials (in conjunction

70 KNOWN OR SUSPECTED NUCLEAR-RELATED FACILITIES IN THE DPRK reactors and related nuclear facili- in the 5-MW(e) graphite reactor the reactor had been reprocessed ties declared by the DPRK starting and that only a few defective fuel and several kilograms of plutoni- shortly after the AF was signed. rods had ever been removed. To um removed. The nature of the dis- the contrary, evidence leads to the crepancy is shown schematically in The declaration claimed that the suspicion that beginning in 1989 Fig. 6-4, which shows a suspect very first core-load of fuel was still large amounts of spent fuel from path employing the 5-MW(e) graphite reactor. Table 6-1. Visits and inspections by the IAEA of North Korean facilities. From satellite imagery of the Dates Purpose and accomplishments steam emitted from the reactor’s May 11–16, 1992 Initial official visit by Director-General Hans Blix and cooling tower, a rough power his- delegation. tory of the reactor was determined. May 26–June 5, 1992 First ad hoc inspection to verify the initial declaration. The design of the reactor does not require the reactor to be shut down July 8–18, 1992 Sampling of radiochemical lab includes swipes in and to remove some of the fuel. around five glove boxes that compose the plutonium- However, it was determined that production area of the lab. significant outages had occurred September 11–14, 1992 Visits to radiochemical lab and fuel-fabrication complex. during operation, which would Acting on suggestions made by the US, there are visits to allow large fractions of the core the high-explosive facility and to Building 500, which inventory to be removed, depend- appears to be a single-story building under military ing on the speed of the refueling control. machines available. It has been November 2–13, 1992 IAEA provides DPRK with details of the inconsistencies estimated that during a 70-day out- and requests an explanation. Answers are not satisfactory age in 1989, half the core could to IAEA, who has been shown US satellite photos of the have been removed with a single construction of Building 500 and the trenches leading to it. refueling machine, or the entire core could have been removed December 22, 1992 Hans Blix requests extraction of samples around with two refueling machines. basement of building in question. Based on these rough estimates, it January 5, 1993 DPRK’s Atomic Energy Minister Choi rejects Blix’s is likely that up to 50 tons of fuel request. was processed in the radiochemical January 1993 Agency task force summarizes inconsistencies and makes laboratory, which then may have recommendations for a set of measurements on reactor produced as much as 8.5 kilograms fuel. of plutonium. If this were the case, there would be tens of kilocuries of January 20–22, 1993 El Baradei of IAEA visits Pyongyang to request special radioactive fission products, and inspections of the two undeclared waste sites. Response tens of tons of uranium sludge in is that the sites are military and non-nuclear—there will the basement of Building 500. be no inspections. January 26–February 6, 1993 Sixth ad hoc inspection. Extensive meetings and The North Koreans declared that discussions between IAEA team and DPRK officials the reactor operated with only one discussing isotopic inconsistencies found in swipe core load of fuel from 1986 until samples from radiochemical lab. 1994, when the entire core invento- February 22–23, 1993 IAEA Board of Governors meets and discusses findings of ry was removed and put in the inconsistencies. Satellite photos are shown. Board spent-fuel storage pool. Currently, decides to support Blix, and declares that inspection of about 8,000 rods are sealed in cans additional sites is essential to ensure verification of in the fuel pool, representing this compliance with the agreement. one full core load. Only about a 100 rods were otherwise removed March 12, 1993 DPRK announces its intention to withdraw from the NPT. from the core because they had May 10–14, 1993 Seventh ad hoc inspection team visits Yongbyon and per failed during reactor operation. Of forms maintenance and replacement of safeguards those rods, they were able to take equipment at reactor. 86 from the spent-fuel storage pool June 11, 1993 DPRK announces suspension of NPT withdrawal, based and transfer them to the radio- on negotiations with US Assistant Secretary of State chemical lab for “hot tests” of the Gallucci. facility. The other rods, they claimed, were too badly damaged

71 KNOWN OR SUSPECTED NUCLEAR-RELATED FACILITIES IN THE DPRK to be removed from the pool and 86 rods are still in the sludge in the bottom 5-MW(e) + 8,000 rods Fuel of the pool. For “cold tests” of the reactor pool radiochemical lab, they claim to Building 500 have used 172 fresh fuel rods. The 86 rods basement “radiochemical lab” (reprocessing facility) was therefore declared to Radiochemical lab Tunnels 172 rods (fresh) have only processed 86 irradiated rods and 172 fresh fuel rods. It was claimed that the processing Figure 6-4. Nuclear program at 5-MW(e) graphite reactor and radiochemical lab. Declared occurred in a single campaign, quantities are shown as numerical values. It is suspected that as much as one additional core was processed without declaration and the wastes could be stored in Building 500, consisting of three batches, in an undeclared waste site. 1990. The amount of plutonium declared was a uniform metal sample of 62 grams. 86 rods (hot) Fission Uranium Plutonium As was mentioned in Table 6-1, Dissolution product 172 rods (fresh) extraction extraction the IAEA inspectors visiting the extraction radiochemical lab in July of 1992 took a variety of samples, includ- ing swipe samples in the plutoni- um area of the lab. Some of the information from those measure- 62 grams ments is shown in Fig. 6-5. In this figure, the PUREX process is bro- ken down into its four stages. The Waste inventory in mixer-settlers and in-process waste in the tanks in storage tanks each of the stages was analyzed for (240Pu content averaged about 2.25%) Plutonium the plutonium isotopic content. nitrate The fraction of the plutonium that 240Pu content was 240Pu was different in the tank about 2.44% inventory than in the plutonium metal sample shown the IAEA, Figure 6-5. The measurements at the radiochemical lab, shown broken down into the four which could indicate the wastes stages of the PUREX process. The declared input and output processing quantities are that were in the tanks were not the shown. wastes resulting from the manufac- ture of the plutonium sample. This were undeclared processing cam- means by which the reactor could result in itself is not a “smoking paigns. Additionally, the ratio of have been used to produce pluto- gun,” however, because there may 241Am to 241Pu was measured for nium is described schematically in have been very different extraction these dust particles. 241Pu has a Fig. 6-6. The reactor used enriched efficiencies in the three batches and 14-year half-life, and decays to uranium oxide fuel, not of suffi- the three batches may have had 241Am. The ratio 241Am/241Pu cient enrichment for use in different irradiation histories. should therefore indicate the weapons. The fresh fuel was sup- amount of time elapsed since the plied by the Soviet Union, and Swipe samples taken in the plu- plutonium was separated, as long spent fuel was stored in a storage tonium area of the facility provid- as the initial separation was clean canal near the reactor. It is sus- ed more evidence of undeclared and the sampling and analysis pected that targets made of natu- reprocessing campaigns. Using were performed without the intro- ral uranium were irradiated in the sophisticated techniques, the frac- duction of bias. This measurement neutron flux in the core of the tional content of 240Pu was deter- indicated that reprocessing had reactor, and then transferred to the mined for individual dust parti- occurred in three separate cam- “Isotope Production Laboratory” cles. This fraction was found to paigns, in 1989, 1990, and 1991. or other facility near the IRT for vary from particle to particle, clus- plutonium removal. The total tered in three groups, whereas the Because of the hidden waste site amount of plutonium that could plutonium metal sample shown apparently associated with the IRT be made by this route is probably the inspectors was uniform. This is reactor, there is also an issue asso- less than four kilograms. a very unlikely result, unless there ciated with that reactor. The

72 KNOWN OR SUSPECTED NUCLEAR-RELATED FACILITIES IN THE DPRK

Enriched Storage additional facility (for plutonium uranium fuel canal storage) may have to be declared to the IAEA. IRT Hot-cell Declared reactor facility at IRT waste site Natural uranium targets Even if there is openness on the part of the DPRK, there may still Buried be ambiguity due to the impreci- undeclared sion of the assay methods or the site permanent loss of operating records. As an example, suppose that an assay of the suspect waste Figure 6-6. The suspected route used to produce plutonium with the IRT-2000 research sites determines that a “missing” reactor. 8 kilograms of separated plutoni- um should be accounted for. It is not clear that the building lost when the DPRK refused to Suppose that by this time North identified to the IAEA inspectors as permit IAEA inspectors to be pres- Korea has changed their declara- the Isotope Production Laboratory ent and to sample materials from tion, claiming that 6 kilograms of is the facility that would have been the rods from each of a number of plutonium was produced and they used for this process. The North key locations in the reactor when show this amount to the IAEA. Koreans may plausibly have had a the DPRK removed the rods. Even if the DPRK were to open pilot reprocessing plant some- Allowing such sampling could many of the facilities at Yongbyon where that enabled them to scale have allowed a more accurate to inspection and grant interviews up to the large scale used in the reconstruction of how much pluto- with workers, the discrepancy Radiochemical Lab. They have nium was produced and when. between 6 and 8 kilograms may denied there was a pilot facility never be fully resolved. and none has been found. It will be important to perform measurements on the fuel along The size of the IAEA-allowed The path forward for the IAEA with isotopic depletion calculations discrepancy would be partially to determine the correctness and to verify the reactor operation his- based on the estimated accuracy of completeness of the initial declara- tory. Spontaneously emitted neu- the assay (see Section 6.5). Time is tion has been developed, includ- tron rate measurements can be a factor, because a more detailed ing planning for contingencies. used to estimate the 240Pu content assay will require more time. A This plan, which is not public of the fuel. From the 240Pu content realistic goal may be ~25 percent of information, will be presented to per unit mass of fuel, the plutonium the amount of plutonium inferred the IAEA Board of Governors production can be estimated. It will by the waste site inspections, sometime in the future. It includes be difficult to infer accurately the which in this particular example plans for the cases where large reactor’s operational history with would be 2 kilograms. If full coop- amounts of radioactive wastes are this information alone, so that fur- eration had been given to the discovered in previously hidden ther analysis or data may be IAEA, they may accept that the waste sites. There is no plan to required. It would be helpful to verification process is complete, attempt to verify the accuracy and have the reactor operator’s log provided that there were no more completeness of the initial declara- books for the years when the reac- facilities or sites they felt that they tion unless access to the two sus- tor operated. Perhaps more infor- needed to inspect. It is unlikely pect waste sites is granted, if the mation could be obtained by that any discrepancy over a few agency stays with the recommen- gamma-ray spectral analysis of kilograms would be allowed if dations of former IAEA Director- structural materials within the there were signs that North Korea General Hans Blix. empty reactor. was still concealing information. In addition, any discrepancy larger One core load of spent fuel If an assay of the suspect waste than one weapon’s worth of pluto- remains in the pool near the sites is found to contain significant nium (~8 kilograms) would proba- 5-MW(e) graphite reactor. To pre- amounts of fission products or ura- bly not be tolerated by the US vent corrosion, the fuel has been nium sludge, there will be a need Congress, regardless of the appar- placed in an argon atmosphere for the DPRK to amend its initial ent degree of DPRK cooperation. within stainless-steel cans kept declaration. It would then have to underwater for cooling. The infor- reveal the appropriate amount of mation as to where each fuel rod separated plutonium, which may was located within the reactor was be as much as 10 kilograms. An

73 KNOWN OR SUSPECTED NUCLEAR-RELATED FACILITIES IN THE DPRK

6.4 Verification of Fuel The best place to send the fuel clad fuel from past US operations Disposal and Facility for technical reasons is the at Savannah River and a large vari- Dismantlement Sellafield plant in the UK, where ety of fuels from INEEL. Some of Magnox fuel is routinely the aluminum-clad fuel at SRS will reprocessed. For example, the probably be chemically processed 6.4.1 Spent-Fuel Disposition Tokai-1 nuclear power plant in into other forms before ultimate Japan was a Magnox reactor that disposal. Additionally, over 2,000 The spent fuel at the 5-MWe operated from 1965–1998. During tons of aluminum-clad spent fuel is graphite reactor is kept in canisters this time, the fuel was sent by ship being put in dry storage at the manufactured by NAC to Sellafield for reprocessing. The Hanford site and is tentatively Corporation of Atlanta, Georgia UK still operates a dozen or more being packaged for disposal with- constructed of stainless steel, filled Magnox reactors; the Magnox out further processing.3 with inert gas, and equipped with a reprocessing facility at the pressure relief valve. The canisters Sellafield plant is scheduled to If one of the sites in the US is were designed so as to fit within a operate until at least 2011. If chosen for the spent fuel, there is shielded shipping cask, also manu- Sellafield were chosen, it may have certain to be some opposition from factured by NAC Corporation. No to store the resulting reprocessing state governments. A lesson from agreement has been worked out wastes, which is unusual, because several cases of such opposition in between the DPRK and the US on the wastes are ordinarily returned the US, however, is that state gov- exactly how the canisters will be to the customer. ernments may be willing to con- shipped. It is known that the US sider limited and well-defined government paid the cost of the Other locations, which could in shipments of spent nuclear fuel of canisters to begin with, which indi- theory process the fuel, include overriding national security inter- cates the shipping will probably Tomsk in Russia, Sichuan in China, est. There are also some legal bar- also be paid for by the US. If the and the Savannah River Site (SRS) riers to sending the fuel to either destination for the fuel is Russia or in Aiken, South Carolina. These site that would have to be over- China, it could be transported over sites have facilities specifically for come. The current program accepts land. Otherwise, ship transporta- processing uranium metal fuel and fuel irradiated in foreign research tion will be required, which will are preferable for technical reasons reactors, but it only covers fuel increase cost. to sites that process commercial that is of US origin, and the uranium-oxide fuels. The compli- Yongbyon fuel is certainly not of The fuel is in poor condition and cation of the magnesium cladding US origin. The existing there may be some significant safe- could introduce significant pre- Environmental Impact Statement ty issues when shipping these treatment costs. While the Russian (DOE/EIS-218F) would also have materials. The magnesium and Chinese locations would entail to be modified for this fuel. cladding was extensively corroded lower transportation costs, it is an before the fuel was containerized. open issue whether safety stan- 6.4.2 Dismantlement In some canisters, apparently, ura- dards at those plants would be nium hydride has formed from the considered adequate. Some important verification reaction between uranium metal issues in the dismantlement and water. This reaction releases Both the SRS and the Idaho process are called for in the AF. oxygen and has caused overpres- National Environmental and After the first LWR has been com- sure. The sludge in the bottom of Engineering Lab (INEEL) now pleted, the DPRK will be required the spent-fuel storage pool, which accept reactor spent fuel (for stor- to dismantle its frozen nuclear may contain as much as a kilogram age) from other countries as part of facilities. If this part of the AF actu- of plutonium, should also be the National Spent Nuclear Fuel ally takes place, the DPRK will canned and shipped. Wherever the Program.2 Of these two sites, the have already come into compliance fuel is shipped, it will have to INEEL site has more diversity in its with its INFCIRC 153 IAEA undergo sorting into a multitude operations and accepts more differ- requirements and will have of waste streams, including bare ent types of fuel than the SRS, allowed all of its spent fuel to be uranium rods, rods with intact which only accepts aluminum-clad removed from Yongbyon. While it cladding, and fuel sludge. The uranium-metal fuel. In any case, is the responsibility of the DPRK to sludge and cladding can probably the fuel is ultimately to be buried dismantle its own facilities, it may be converted directly into waste as waste in the National Repository be in the interests of US national forms, but the uranium metal may (currently thought to be Yucca security to assist the DPRK in its require processing before it is in a Mountain in Nevada), along with effort. The IAEA can also provide suitable waste form for disposal. thousands of tons of aluminum- some help (if the DPRK re-joins it).

74 KNOWN OR SUSPECTED NUCLEAR-RELATED FACILITIES IN THE DPRK

There are three stages of decom- with similar facilities. A verified It is to be remembered that the missioning as far as the IAEA is cocooning of the graphite reactor North Koreans have no experience concerned4: could be performed relatively with nuclear decommissioning. It quickly compared to the green- therefore may be advisable to hire Stage 1 (“safe storage”): The field approach. The process is a private company to do the proj- outer contamination barrier is kept deemed “irreversible” if the reac- ect management and provide tech- as it was during operation, but the tor is taken to a state where it nology. Obviously, the North mechanical opening systems are would be cheaper for the DPRK to Koreans may press that their own permanently blocked and sealed begin an entire new reactor con- labor will be used and may (valves and plugs, etc.). The con- struction project rather than re- attempt to engage the US in pay- tainment building is kept in a state activate the old reactor. Emphasis ing for costs not directly related to appropriate to the remaining haz- would be placed on removing and reactor entombment, such as site ard and the atmosphere inside the destroying critical reactor compo- cleanup and waste disposal. A building is subject to appropriate nents such as control rod drives. decision will have to be made as to control. Access to the building is The steam generators could be the scope of US involvement in allowed, subject to monitoring and destroyed in situ by filling them those other activities. surveillance procedures. with concrete. The empty core of the reactor could also be filled Handling the radioactive Stage 2 (“cocooning”): The with concrete. The reactor graphite core blocks is time-con- outer contamination barrier is graphite may be poisoned by the suming and presents a radiological reduced to a minimum size and all introduction of a boron-contain- risk to workers because of the parts easily dismantled are ing spray or resin that would ren- debris and dust. This is part of the removed. The sealing of that barri- der it permanently useless for reason why greenfield costs esti- er is reinforced by physical means nuclear purposes. mates are so much higher than the and the biological shield in a reac- actual costs at Hanford.* It is prob- tor is extended if necessary so that The US has recently obtained ably undesirable for these reasons it completely surrounds the barri- extensive experience cocooning its to take the Yongbyon reactor to the er. After decontamination to own graphite reactors, many of final state of decommissioning. acceptable levels, the containment which are located at the Hanford building and the nuclear ventila- Reservation in Washington. These The Yongbyon reactor should tion system may be modified or old reactors, built to produce plu- easily be brought to a safe storage removed if they are no longer tonium for weapons, were nearly state within a three-year period. required for radiological safety. identical to one another and were Because of the special security each rated at 425 MW(th). Five are issues surrounding this reactor, Stage 3 (“greenfield”): All mate- being cocooned now, and one is some critical pipes and possibly the rials, equipment, and parts of the complete. As experience has been reactor vessel itself could be cut plant in which activity remains sig- gained, the cost per reactor has with a saw to provide assurance nificant despite decontamination gone down. The typical cost is that the Stage 1 decommissioning are removed. In all remaining about $25 million each (130 man- will not be reversed. The steam parts, contamination has been years labor each). The key to the generators, the refueling machines, reduced to acceptable levels. The low cost is the use of the existing and control rod drives should be plant and site are released for unre- outer concrete shield that sur- removed and destroyed. stricted use. From the point of rounds each reactor as a major part Cocooning could begin as soon as view of radiological protection, no of the cocoon. Stage 1 is complete. further surveillance, inspections, or tests are necessary. The Yongbyon reactor is much The other two graphite reactors, smaller and is also less radioactive one located at Yongbyon and the The first of two 2 major dis- than the Hanford reactors. Its outer other at Taechon, could be taken to mantlement challenges at concrete shield wall is similar in more advanced stages of decom- Yongbyon is the 5-MW(e) graphite thickness to those of the Hanford missioning. Of some concern is the reactor, which operated for sever- reactors. In theory, therefore, it disposition of the nuclear-grade al years and should be quite should not be significantly more graphite from the 50-MW(e) reac- radioactive. Fortunately, there is difficult to entomb. tor, which should be treated in some international experience some way so that it cannot be used to reconstruct the reactor.

* The final greenfield decommissioning costs for the Hanford reactors has been estimated to be in the billions of dollars.

75 KNOWN OR SUSPECTED NUCLEAR-RELATED FACILITIES IN THE DPRK

The second dismantlement chal- most time-consuming components declared and presented to the lenge is the Radiochemical Lab at to dismantle. It is to be remem- IAEA is consistent with what could Yongbyon. Experience obtained at bered that far more fuel in total have been produced. the Eurochemic Reprocessing Plant was processed at Eurochemic than in Belgium (and at other reprocess- at Yongbyon. The cost and man- Cooperation between the IAEA ing plants) should provide some power estimates for the and a state is necessary for the suc- assistance. This facility was con- Eurochemic plant are therefore rea- cessful implementation of safe- structed in the early 1960s and was sonable upper limits for the corre- guards in any context. The level of owned by a 13-nation consortium. sponding numbers at Yongbyon. cooperation essential to the process It went into active operation in of verifying the completeness of an July 1966 as a demonstration plant Again, it may be sensible to initial declaration goes beyond that with a 60-ton annual capacity. engage a Western company for required to implement the safe- Between 1966 and 1974, about 180 project management and technolo- guards agreement or even an tons of low-enriched uranium fuel gy know-how. For instance, mod- Additional Protocol (INFCIRC 540) and 30 tons of high-enriched urani- ern remotely operated machines to that agreement. It can be argued um fuel were reprocessed. After its could cut some of the more that the state is obliged to provide shutdown in January 1975, the radioactive pipes and remove some any existing facility-operating plant was decontaminated during of the more contaminated equip- records to the extent they are perti- the years 1975-79 for keeping it in ment. In any case, the assistance nent to assessing the completeness a safe, standby condition. should focus on the rapid, verifi- of present declarations. However, Decommissioning began in 1991 able, and irreversible destruction of the process may require access to and the plant has been completely components and subsystems of individuals (e.g., knowledgeable decommissioned as of today, greatest concern to US national facility personnel) and locations except for the outer structure and security, such as remote manipula- (e.g., locations beyond declared the floor. tors and leaded-glass windows. nuclear sites) that the state is not Help provided to the North legally obliged to provide (at least, In 1987, the first decommission- Koreans for activities that do not without resort to a request for a ing phase was estimated to require directly affect verification process, special inspection). Further, the a team of 20 dismantlement opera- such as general site cleanup, geo- verification process does not pro- tors and 17 technical and safety logical waste disposal, and shallow vide certainty and while a high assistance to work for 11 years. land burial of wastes should be level of openness regarding a Total cost was estimated at $172 provided only if it furthers US state’s past and current nuclear million. In 1992, the cost estimate national interests. If the DPRK re- activities is a necessity, it may not was revised upward to $235 mil- joins the IAEA, some assistance be sufficient depending on the lion, reflecting increased waste dis- can be provided for additional site level of ambiguity judged political- posal costs. Progress reports pub- cleanup through that mechanism. ly acceptable at the time. Technical lished over the last few years have aspects of the verification exercise confirmed that the 1992 estimate is 6.5 How Much Verification are important determinants, but correct to within about 25 percent. Is Enough? the final judgement regarding how Waste management and disposal much verification is enough will costs are more than a quarter of the Verification of the completeness likely be political in nature. total cost. A computerized data- of a state’s initial report, particular- management system has been set ly for states that have, or are sus- At the point when the process of up to keep detailed records of pected of having, produced verifying the completeness of the costs, hours worked, and wastes weapons-usable nuclear material DPRK’s initial declaration was produced to enable the OECD prior to entry into force of their stopped, the DPRK was not in countries to make estimates for safeguards agreement, is complex. compliance with their Safeguards future decommissioning activities.5 The process involves a detailed Agreement. There were a number review of facility-operating histo- of major inconsistencies between The Yongbyon plant had a total ries, a comparison of facilities and the DPRK’s declarations and the capacity of the plant in 1994 of nuclear material types and IAEA’s inspection results. If any- about 220 tons/year (with two amounts with other information thing, the situation has worsened lines). One of those lines has never available to the IAEA and the reso- in that the possibilities of recon- been tested or contaminated. The lution of any resulting inconsisten- structing and in some sense verify- radioactive waste tanks and con- cies. The objective is a high level of ing historical activities degrades taminated concrete will be the assurance that the nuclear material with time. There is no metric or

76 KNOWN OR SUSPECTED NUCLEAR-RELATED FACILITIES IN THE DPRK quantitative measure of a state’s as comprehensive and detailed pic- that state’s nuclear material are intent; however, when the process ture of past activities as possible complete as a continuing feature in of verifying the completeness of through collection and evaluation verifying compliance with the NPT the DPRK’s initial declaration of information from open sources commitments. At this time, 55 resumes, the DPRK’s intentions to and third parties, together with states, including Japan, the be open (or not open) should IAEA inspection data and DPRK Republic of Korea (ROK) and the become visible quickly. The DPRK declarations. The Agency should US, have signed Additional knows the actions necessary to also anticipate and prepare specific Protocols to their Safeguards come into compliance with their verification measures including Agreement with the IAEA. Safeguards Agreement. It is aware measurement methods and devel- Nineteen, including that for Japan, that its initial declaration needs to op agreed standards that at least have entered into force. be amended. It is aware that the bind the ambiguities. There is no legal obligation for a IAEA has and will continue to state to accept the Additional receive and make use of informa- The support provided to the Protocol, but KEDO principals tion from third parties that could IAEA is a real issue. Most impor- would not be asking the DPRK to result in IAEA requests for access tantly, the IAEA is struggling today accept a more intrusive verification to locations not identified in their to maintain the routine implemen- regime that they, themselves, have initial declaration. tation of safeguards. Resources will not indicated a willingness to be stretched very thin when they accept. The presence of an Several parties, working in other prepare for and effectively respond Additional Protocol is not likely to areas, have indicated that, with to the DPRK verification problem. be a determining feature in the ver- patience and perseverance, they The IAEA also needs information ification of the DPRK’s initial have obtained cooperation from provided by third parties and report (either the DPRK provides the DPRK necessary to make states other than the US. The IAEA the required cooperation or they progress. This has not been IAEA’s also needs political support to deal don’t), but it could simplify things. experience. When inspections first with the likely pressure to accept a Under an Additional Protocol, got underway in late May 1992, the lower verification standard (i.e., a IAEA inspectors would have access IAEA enjoyed a high level of coop- higher level of ambiguity). to any location on a nuclear site eration from the DPRK. This coop- (e.g., the Yongbyon site), other eration disappeared quickly as Finally, the evolving NPT verifi- locations involved in the produc- problems developed. Today, the cation framework provides another tion or storage of source material, DPRK accepts the continuous pres- avenue that could be pursued with and any other location in the DPRK ence of IAEA inspectors to monitor the DPRK. Beginning in the early for the purpose of collecting envi- the freeze as prescribed in the AF 1990s and continuing through the ronmental samples. The DPRK has and they accept safeguards on decade that followed, the IAEA has good reason to be familiar with the declared nuclear material per the been involved in an extensive effort power of environmental sampling Safeguards Agreement. They have to improve the effectiveness and and this might encourage them to steadfastly refused any actions efficiency of its safeguards system. be more forthcoming regarding the requested by the IAEA that would IAEA’s Programme 93+2, the cor- whole of their past and current have the effect of improving the nerstone of this effort, concluded in nuclear program. Certainly, should IAEA’s position to deal with the May 1997 when the IAEA’s Board the verification of the DPRK initial verification of the DPRK’s initial of Governors approved the “Model declaration come to a successful declaration once that exercise gets Protocol Additional to the conclusion, an Additional Protocol underway again. Agreement(s) Between State(s) and would be very helpful in assuring the International Atomic Energy the DPRK’s continuing compliance Citing the time necessary to Agency” (referred to as the with their NPT commitments. complete the verification exercise, Additional Protocol and published the IAEA’s Director-General has as INFCIRC 540). The Additional repeatedly requested the DPRK to Protocol is designed, through great- get started as soon as possible. ly expanded nuclear openness and They have given no indication that inspector access, to improve the they are ready to proceed. Still, the IAEA’s capability to detect unde- IAEA can do a number of things to clared nuclear material and activi- prepare now. The first is to develop ties and thus increase assurances

77 KNOWN OR SUSPECTED NUCLEAR-RELATED FACILITIES IN THE DPRK

Notes to Chapter 6 2. National Spent Nuclear Fuel 4. International Atomic Energy Program web site, Agency, State of the Art 1. G. Kaurov, “A Technical History of http://nsnfp.inel.gov. Technology for Decontamination Soviet-North Korean Nuclear and Dismantling of Nuclear Relations,” in J.C. Moltz and A.Y. 3. M. Gerber, “Removing the K- Facilities, Technical Reports Mansourov, eds. The North Korean Basins Fuel: Downpayment on Series No. 395, Vienna, 1999. Nuclear Program: Security, Strategy Protecting the Columbia River,” and New Perspectives from Russia Radwaste Solutions Magazine, 5. L. Teunckens, P. Lewandowski, (Routledge Press, 2000), pp. 17–18. American Nuclear Society, and E. Trauwaert, “Taking it March/April 2001. Apart: The Eurochemic Reprocessing plant,” Nuclear Engineering International, August, 1994.

78 CHAPTER 7 TIMELINE FOR VERIFICATION AND SAFEGUARDS

In this chapter, we present a 5. Dismantlement of DPRK major building construction at simplified timeline that focuses on graphite-moderated reactors and Kumho. verification and safeguards issues. related facilities at Yongbyon • Completion of the nuclear-plant The chapter contains no new mate- begins. design for the LWRs by the rial. Rather, the material relevant ROK. to the Agreed Framework (AF) 6. Deliveries of the nuclear compo- timeline from earlier s is brought nents for the second KEDO reac- • Delivery of the turbine genera- together and the implications for tor in parallel with proportional tors for the first LWR, with verification and safeguards of each steps by the DPRK to dismantle other delivery details still to be step in the timeline are noted. all its graphite reactors. agreed on. Based on this timeline, Chapter 7 presents an analysis of what can 7. Completion of second KEDO As noted earlier, it would be go right and wrong with the AF reactor at Kumho. advantageous to South Korea, the from the viewpoint of verification major nuclear-reactor supplier and and safeguards. In addition, another step is not provider of money, for the IAEA explicitly linked to the timeline but and the DPRK to begin negotia- Seven major time-linked steps must be carried out if verification tions soon on the special expand- in the AF and related agreements that the DPRK does not have ed inspections and to get them may be identified, of which all but nuclear weapons-usable material is started before too much is invest- the first require verification or to be complete: ed in site preparation, construc- safeguards: tion, and manufacturing. But, 8. Transfer of KEDO spent fuel when there is no DPRK obligation to 1. Completion of a significant por- appropriate out of the DPRK. permit these IAEA inspections tion of the project site at Kumho, prior to those investments, and the DPRK (Democratic People’s We review briefly what informa- DPRK has not volunteered so far Republic of Korea, or North tion is expected from each step, to cooperate on this matter. If it Korea), and of the first Korean how long each step will take, and does not, the inspections of unde- Energy Development what must take place before the clared facilities cannot begin for Organization (KEDO) reactor in step can be taken. 2–3 years at least, more if the cur- the Republic of Korea (ROK, or rent problems holding up delivery South Korea). Completion of a Significant of the turbine generators are not Portion of the Project Site resolved in that time. 2. International Atomic Energy Agency (IAEA) declaration that at Kumho, DPRK, and of IAEA Declaration that the the DPRK is in compliance with the First KEDO Reactor in DPRK Is in Compliance its safeguards agreement. the ROK with Its Safeguards Agreement 3. Start of the delivery of key Before the DPRK is obligated by nuclear components of the first the AF to permit the IAEA to After the step above, the DPRK KEDO reactor to the Kumho inspections beyond the DPRK’s is obligated to come into full com- project site simultaneously with declared facilities, a “significant pliance with its safeguards agree- the start of the transfer of DPRK portion of the Light-Water Reactor ment, including taking all meas- spent fuel from Yongbyon to its (LWR) project” at the Kumho site ures that may be deemed necessary ultimate disposition. must be completed. The steps to be by the IAEA “to verify the accuracy completed have been spelled out in and completeness of the DPRK’s 4. Completion of Yongbyon spent- Chapter 1. They include— initial declaration [to the IAEA] on fuel transfer and of the first all nuclear materials in the DPRK.” KEDO reactor simultaneously. • Completion by KEDO of site These measures, and the informa- preparation, excavation, and tion they will yield, include—

79 TIMELINE FOR VERIFICATION AND SAFEGUARDS

• Visual inspections of all declared IAEA has told both the IAEA mate disposition site is outside the and undeclared suspect facilities General Conference and the U.N. DPRK and under the IAEA’s juris- at Yongbyon (detailed in General Assembly that he believes diction and control. If the material Chapter 6), including possible it will take 3–4 years for the IAEA were to remain in the DPRK, con- undeclared waste sites, to identi- to complete this task, presumably tinuing verification and accounta- fy facilities used for plutonium assuming DPRK cooperation. bility would be required. production and separation, waste storage, radiochemistry, If insufficient cooperation is The time needed to ship spent and any other potential nuclear forthcoming, attempts to complete fuel and any separated plutonium weapon-related activity. this step could bring about an identified out of the DPRK is prob- indefinite delay. Furthermore, if the ably measured in months once all • Measurements, including swipe DPRK declaration and the IAEA preparations, both operational and and particle assays, of all equip- findings cannot be brought into diplomatic, have been made. Those ment, surfaces, and other material agreement, the AF could end. We preparations, however, could take at these facilities that could con- discuss the implications of those much longer. Shipping appropriate tain or have contained any prod- two outcomes in Chapter 8. to the transport of spent fuel must uct associated with the produc- be made available. Possible sites tion and storage of plutonium. The AF is silent on what for disposal are discussed in should be done if spent fuel, Chapter 6, Section 6.4. • Analysis of at least some of the waste, or separated plutonium is spent fuel from the reactor at found outside the ponds where Simultaneous Completion Yongbyon, and of any waste or the declared spent fuel is stored of Yongbyon Spent-Fuel other relevant material recovered. at Yongbyon. Thus, the outcome Transfer and of the First of this key step is both crucial KEDO Reactor • Identification and visits to sites, and indeterminate. Whatever the other than those at Yongbyon, outcome, at this point, no nuclear Upon completion of the spent- suspected of being used for fuel or key nuclear component fuel transfer from Yongbyon, the nuclear materials-related activi- would have been delivered to the first KEDO reactor can be complet- ties. If identified, measurements KEDO reactor site. ed. All safeguards for that reactor similar to those at Yongbyon have (discussed in detail in Chapter 4) to be made. Identification of sites Start of Delivery of Key must be installed and operational outside of Yongbyon may require Nuclear Components of the before operations can begin. Until some use of national technical First KEDO Reactor to the operations begin, there will be no means of surveillance as well as Kumho Site Simultaneously nuclear weapon-usable material in cooperation from the DPRK. with Start of Transfer of the reactor. About two years of DPRK’s Spent Fuel from prior training in maintaining the These measures should yield an safeguards and in material estimate of the amount of nuclear- Yongbyon to Its Ultimate accountancy are needed before the weapon material made and of the Disposition reactor can begin operation. The forms it may be stored in, together time needed after the completion with an estimate of the capability After the DPRK comes into com- of the first two steps noted above, of the DPRK to make more such pliance, and simultaneously with and after the training has taken material. They could also yield the start of delivery of key nuclear place, has been estimated at about information regarding what other components to the first KEDO reac- a year, depending on the detailed parts of a nuclear-weapon program tor, the DPRK is obligated to begin circumstances. Delays could be the DPRK has carried out. They the transfer of spent fuel from the incurred owing to technical and could lead to a need to modify the cans in the small graphite reactor legal problems associated with the original DPRK declaration, which pools to “its ultimate disposition,” reactor itself and its safeguards, or in turn, could jeopardize the AF. outside the DPRK if that is what to problems external to the reactor KEDO wants. The spent fuel is to be installation, such as the lack of an Verification of accuracy and completely transferred by the time adequate electrical grid to accept completeness of the DPRK’s decla- the first KEDO reactor is completed. the power. These problems have ration is needed for the IAEA to been mentioned previously. declare that the DPRK is in compli- If identification of the material ance. The time needed to complete to be transferred is complete and The information expected from this verification is difficult to esti- reliable, verification of this step is this step will continue to flow as mate. The Director-General of the straightforward so long as the ulti- the reactor operates. As noted in

80 TIMELINE FOR VERIFICATION AND SAFEGUARDS

Chapter 4, Section 4.5, the global • The possibilities that nuclear The spent fuel from the LWR record to date suggests that no fuel facilities not covered by the AF reactors provided by KEDO will be diversion from a commercially but implicitly covered by the high-burnup fuel, not nearly as operated LWR has taken place. It 1992 ROK–DPRK Joint suitable for weapons use as the seems to us highly unlikely that Denuclearization declaration, to fuel from the Yongbyon reactors such a diversion could be carried which the two parties recently would have been, but nevertheless out undetected if the safeguards recommitted themselves, exist. a potential proliferation risk. Such described are adequately imple- fuel is left in cooling ponds at the mented. As noted at the end of • The question of who pays for the reactor site for a period of years. Section 4.5, the additional meas- dismantling. After the radioactivity in the fuel ures and inspections described assemblies has decayed sufficiently could be useful in further lowering Auxiliary agreements and proto- to permit handling the assemblies the probability that covert diver- cols covering those points are on dry land, the assemblies can be sion could take place from a safe- needed. Absent knowledge of these placed in casks and kept in dry guarded reactor. Further, as noted agreements and protocols, esti- storage for an indefinite period of in Section 4.6, satellite monitoring, mates of time and verification time. Such dry storage is in use in which is not a part of IAEA safe- beyond the ones made in Chapter the US for example, but so far it guards, could—in case of some 6, which are based on analogy with has not been practiced in Asia. The intent to abrogate—give informa- plans for dismantling similar but assemblies continue to become less tion as to the time of the diversion larger facilities, cannot be made. and less radioactive over the years and the amount and form of the According to those estimates, dis- and must continue to be monitored material involved. mantlement would take several and accounted for until ultimately years owing to the radioactivity in disposed of. Dismantlement of DPRK’s some of the buildings, but irre- Graphite-Moderated versible changes could be brought If the plutonium-containing Reactors and Related about sooner. spent fuel from the KEDO reactors Facilities Begins is not to be left indefinitely in the Completion of Second DPRK, facilities for ultimate dis- KEDO Reactor at Kumho posal or at least long-term dry stor- Deliveries of the Nuclear age must be found. This problem is Components for the The same verification and safe- not unique to the DPRK’s KEDO Second KEDO Reactor in guards comments can be made reactors and must be solved for all Parallel with Proportional here as were made in connection Asian and other countries that use Steps by the DPRK To with the completion of the first nuclear power, although the DPRK Dismantle All Its Graphite KEDO reactor. case is particularly sensitive. Reactors Discussions about disposal and Shipping Spent KEDO long-term storage for spent nuclear The timing of these parallel pro- Reactor Fuel When fuel in Asia have been going on for cedures is complex and may be Appropriate Out of the years inconclusively. The existence contentious. Only some of the DPRK of spent DPRK fuel may add to the agreements between KEDO and incentives for resolving them. the DPRK have been made public. According to the Supply Among possible points of con- Agreement (Art. VIII, par. 3), Table 7-1 summarizes in a sim- tention particularly relevant to ver- “KEDO and the DPRK shall coop- plified way the steps outlined in ification are— erate to ensure the safe storage and this chapter. It does not contain all disposition of the spent fuel from the linkages and details described • The precise meaning of “dis- the LWR plants. If requested by above but may serve as a useful mantlement” and of “ultimate KEDO, the DPRK shall relinquish reminder of the overall pattern of disposition” of the dismantled any ownership rights over the activities bearing on or associated parts. Chapter 6, Section 6.4 out- LWR spent fuel and agree to the with verification. lines some technical possibilities, transfer of the spent fuel out of its but they are suggestions only. territory as soon as technically pos- sible after the fuel is discharged, • The nature and extent of “relat- through appropriate commercial ed facilities” at Yongbyon and contracts.” possibly elsewhere.

81 TIMELINE FOR VERIFICATION AND SAFEGUARDS

Table 7-1. Summary of steps bearing on verification. Step Verification Issue Possible Problems Partial completion of KEDO None but the IAEA wants to Financial and legal delays Reactor 1 in the ROK and start the next step early (2-4 cause some loss of data at partial preparation of Kumho years needed) Yongbyon site in the DPRK

IAEA declaration that the Verification of accuracy and 1. DPRK does not open DPRK is in compliance with completeness of the DPRK’s suspect sites to the IAEA. its agreements initial declaration on all 2. The IAEA's activities are nuclear materials in the interfered with. DPRK, at Yongbyon, and 3. Initial declaration is possibly elsewhere wrong—can it be amended?

Delivery of KEDO Reactor Safeguards for KEDO 1. Disagreements over the 1's key nuclear components Reactor 1 are installed. extent of safeguards. starts. Transfer of Yongbyon Transfer of Yongbyon spent 2. Disagreements over site spent fuel (and other fuel (and other material?) to of disposition. material?) to "ultimate "ultimate disposition" 3. Disagreements over what disposition" starts. verified. is to be transferred.

Simultaneous completion of Safeguards on KEDO Same as previous, plus previous steps Reactor 1 operational. interference with KEDO Disposition site monitored. Reactor 1 safeguards

Dismantlement of Yongbyon Safeguards for KEDO 1. Disagreements over facilities in parallel with Reactor 2 installed. extent of safeguards. delivery of KEDO Reactor Dismantlement verified. 2. DPRK abrogation. 2's key nuclear components 3. US or ROK non- compliance with AF.

Simultaneous completion of Safeguards on KEDO 1. Interference with previous steps Reactor 2 operational safeguards. 2. DPRK abrogation.

Disposition of KEDO spent Monitoring disposition site(s) Disagreement over site of fuel disposition

82 CHAPTER 8 POTENTIAL ADVERSE DEVELOPMENTS

We present a benchmark cooper- evidence that the DPRK separat- evaluating differences in ation scenario plus four other sce- ed more than its original declara- shipper/receiver measurements; narios that among them bracket tion of less than 100 grams. If so, taking physical inventory; evaluat- possible degrees of cooperation of an amended declaration will be ing accumulations of unmeasured the Democratic People’s Republic required for the IAEA to certify inventory and unmeasured losses; of Korea (DPRK) with safeguarding the DPRK’s plutonium holdings. and providing reports to the and verification efforts. These sce- Agency in accordance with its safe- narios do not capture variations in • Permit IAEA inspections, meas- guards agreement. These safe- other dimensions—such as finan- urements, and environmental guards measures will be applied to cial, regulatory, diplomatic—to monitoring at all sites where all nuclear material subject to safe- which they are nevertheless linked. nuclear activities subject to safe- guards, including the IRT research They serve as a rough framework guards may have occurred. reactor, the light-water reactor for evaluating potential failure (LWR) power plants, and their modes of the Agreed Framework • Begin early training of its per- related facilities. (AF) and their implications for sonnel in nuclear-material verification and safeguards. accountability. Evidence to date indicates that the DPRK has not implemented 8.1 Benchmark Scenario: • Participate positively in negotia- such rigor in its nuclear-materials The DPRK Fully tions aimed at taking spent fuel accounting activities in the past, out of the DPRK (the DPRK may nor has it operated nuclear facilities Cooperates with the IAEA not be the main problem in these on the scale of the Korean Energy Regarding Existing negotiations, but its early coop- Development Organization (KEDO) Agreements eration would facilitate them). reactors. Perhaps implicit in the AF and the formulation of KEDO is the This scenario is given to provide Fully satisfactory implementa- provision of necessary training and a success benchmark for compari- tion of the DPRK’s safeguards technical support to the DPRK to son. Under this scenario, the DPRK agreement with the IAEA will assure that the DPRK is prepared to would— require that the DPRK cooperate accept this responsibility. As we with the application of advanced have learned from other coopera- • Allow the International Atomic safeguards technologies as they tive threat-reduction programs, Energy Agency (IAEA) to carry evolve. As discussed in Chapter 2, increased openness with respect to out inspections at suspect even without the implementation the peaceful use of nuclear material facilities soon. of the “Additional Protocol” (INF- is key to building a sustainable CIRC 540), the IAEA has the right nonproliferation regime. • Allow measurements on the to utilize remote and unattended Yongbyon spent fuel and pro- monitoring systems, request ad hoc Even with full cooperation from vide reactor logbooks and other and special inspections, and review the DPRK, verification of past activ- documentation to help recon- the data collected by the DPRK’s ities is likely to take years. Early struct the operating history. national system of materials con- DPRK cooperation with such meas- trol and accounting. The DPRK ures as described above, especially • Cooperate with the IAEA in issu- will be required to make accurate pertaining to verification activities ing a modified declaration that measurements of several key at Yongbyon, would relieve pres- would reconcile the discrepan- parameters, such as the quantities sure on the time schedule and cies between its previous decla- of nuclear material received, pro- allow for earlier operation of the ration and IAEA measurements duced, shipped, lost or otherwise KEDO reactors. Reconstruction of (see Chapter 6) and allow verifi- removed from inventory, and the past activities at inspected sites cation of that declaration. While quantities of material in inventory. (declared and suspect) could be as estimates of past plutonium Procedures must be implemented accurate, and knowledge that production vary, there is strong for identifying, reviewing, and nuclear-weapons activities have

83 POTENTIAL ADVERSE DEVELOPMENTS ceased at these facilities could be as on the presently anticipated time IAEA in making a determination of assured as was the case for South scale in two areas: “completeness” with respect to Africa.1 The safeguarding of the declared nuclear facilities and KEDO reactors would be positively • Reconciling the data taken and materials. These include: affected: there would be greater to be taken by the IAEA regard- assurance that inspections, meas- ing activities at Yongbyon with • Provision and verification of urements, materials accounting, the DPRK declaration of these design information, and data transmission would be activities. The consequence of a carried out as specified in the safe- failure to certify the DPRK’s dec- • Notification of new facilities or guards agreement. laration will be a function of modifications to existing facili- when and how this failure ties, and This scenario would represent a occurs. Given the history of departure from past experience. It interactions with the DPRK, • Access for routine, ad hoc, and would speed progress on the meas- issues of uncertainty in the special inspections. urable requirements of the AF, IAEA measurements may arise. mainly contained in Articles I and There is no predetermined stan- • In particular, “special” inspec- IV having to do with nuclear activi- dard by which “acceptable” tions may be requested in cases ties and energy supplies. It could uncertainty is measured. In fact, where the information provided ease working toward normalization this is often a function of specific to the IAEA is not adequate for of economic and political relations, measurement technologies the Agency to fulfill its responsi- and toward peace and security, as employed and the openness of bilities. Special inspections may called for in Articles II and III. Such the inspected party to the IAEA. be requested at “undeclared” increased DPRK cooperation could As with so many things in the facilities. lessen any incipient strain among nonproliferation regime, it is a KEDO members stemming from subject of negotiation. It was the request for a “special” the members’ different relative pri- inspection of two undeclared waste ority among these goals. • Assuring that no further sites at Yongbyon that precipitated nuclear-weapons-related work is the DPRK to withdraw from the 8.2 Scenario 1: The DPRK going on in the DPRK, and that NPT. It seems likely that the IAEA Maintains Its Present Level no separated undeclared pluto- will require access to these unde- nium is stored anywhere. The clared sites to verify the DPRK’s of Cooperation AF does not freeze nuclear facili- declarations. A full understanding ties or activities other than those of the operation of the IRT research Under this scenario, further frozen at Yongbyon. Although reactor (discussed in Chapter 6) delays can be expected. In this the question of plutonium pro- and the disposition of nuclear event, one must consider the rami- duction at these facilities is in materials associated with its opera- fications of continued delay on the forefront of most verification tion are among steps necessary to security planning in the ROK, US, efforts, providing a determina- provide confidence that the DPRK and throughout the region, a con- tion of the “completeness” of the has no undeclared nuclear pro- sideration that goes beyond what DPRK declaration is paramount grams. In addition, there may be this report has undertaken to do. to the continued movement of other special inspections required As time passes, there is continued the DPRK towards membership to fully determine that the DPRK opportunity for improved relations in good standing in the Non- has no undeclared nuclear facili- between the North and South, sig- Proliferation Treaty (NPT), ties, nuclear-material inventory, or nificant investment in KEDO by which is required by the AF. activities. This would require better member states, as well as possibly DPRK cooperation than has taken an opportunity for the DPRK to As discussed in Chapter 2, place, and probably cannot be tamper with materials and records though it has rarely occurred, done perfectly. The best that could needed for complete verification. access to “undeclared” sites or to be done would be to identify and Continued delays could lead to an locations suspected of containing make measurements of the type inability to certify the DPRK decla- “undeclared” nuclear material is described in Chapters 4 and 6 on rations and possibly result in a possible under the model IAEA any facility where plutonium is breakdown of the cooperation safeguards agreement (INFCIRC thought to have been made and between the ROK, US, and Japan. 153). Several provisions under the separated. Some of these measure- In particular, verification could DPRK’s safeguards agreement ments could become difficult or almost surely not be accomplished (INFCIRC 403) could assist the impossible as time goes on.

84 POTENTIAL ADVERSE DEVELOPMENTS

8.3 Scenario 2: The DPRK between the North and South, veri- discharged from the KEDO reac- Attempts Covertly To Divert fication problems could evoke dif- tors will most probably be US-obli- or Hide Nuclear Material ferent responses from the KEDO gated, which will require that the members with respect to continua- US approve any movement or While declared facilities at tion of the AF. If the DPRK were to retransfer of the fuel. Yongbyon are verifiably frozen, refuse to allow access to facilities overt diversion or hiding of materi- and information needed by the We note parenthetically that the al from the suspect but undeclared IAEA to verify declarations, it is DPRK previously (in 1994) gave a facilities at Yongbyon and else- difficult to imagine that the US 90-day notice of intent to with- where in the DPRK is feasible so would continue with its part of the draw from the NPT, which it then long as the extensive IAEA inspec- AF. It is less certain what position “suspended.” This may leave the tions and measurements discussed KEDO member states might take, length of notice that may be given above have not taken place, as especially in light of domestic, of any future intent to withdraw noted in Chapter 6. Early inspec- bilateral, and regional political ambiguous. tion of suspect undeclared facilities pressures. and measurements of what is In the case of abrogation, the found there could help identify These comments pertain to outside world would know how hidden material and prevent future material and facilities at Yongbyon much potential nuclear-weapon diversions. Without such inspec- and possibly elsewhere in the material there is in the DPRK at tions and measurements, it cannot DPRK but not to the KEDO reac- least as well as it knows now, bet- be known whether or not nuclear tors. Covert diversion or hiding of ter if inspections and removal material additional to the declara- nuclear material generated in the operations at Yongbyon have been tion remains hidden. The tech- course of operating the KEDO carried out. It would, as now, be niques discussed in Chapter 4, reactors would be far more diffi- able to externally monitor such Section 6, particularly environmen- cult, as discussed in Chapters 4 large-scale activities as continued tal monitoring at sites other than and 5. So long as the IAEA and its reactor operations, construction of the declared sites, would go some member states are successful in facilities, and, to some extent, iden- ways toward reducing the likeli- obtaining full enforcement of safe- tification of major activities, as dis- hood of successful covert diversion guards, we believe that the proba- cussed in Chapter 4, Section 6. It or hiding. bility of covert diversion from the would not be possible to know KEDO reactors is very low. The accurately how much plutonium is Continued assistance from the main covert diversion or hiding made in reactor operations subse- US and other member states will be problem, at least in the first many quent to abrogation, should those required for the IAEA to provide years of the AF, is connected with occur, nor how much is separated, assurance that the DPRK is not the earlier DPRK activities at or how it is being used. engaging in clandestine nuclear Yongbyon and possibly elsewhere. activities. Early warning of poten- It is to be remembered that no tial illicit activities through the use 8.4 Scenario 3: The DPRK nuclear components will be deliv- of satellite imagery and other tech- Abrogates the Agreed ered to the KEDO site until a bilat- nical means is critical. In addition, eral nuclear cooperation agreement it is important to have a well-exe- Framework or Other Key is signed, which will first require cuted and coordinated approach Agreement that the DPRK is in full compliance with respect to investigating per- with its IAEA safeguards obliga- ceived clandestine activities. If the DPRK were to abrogate its tions. For this to occur, there must agreements, and for instance, expel first be a dramatic change in the Attempts by the DPRK to divert the IAEA inspectors, it would have openness of the regime, as stated or hide nuclear material would be control over any nuclear material above. The LWR reactors would incompatible with the AF. Delays left at Yongbyon and other possible begin operation only after the in carrying out the inspections and sites (and over spent fuel left at the DPRK has verifiably given up its measurements discussed above at KEDO reactor site if abrogation Yongbyon program and its store of Yongbyon and elsewhere could occurs after the KEDO reactor(s) spent fuel. lead to increased suspicions that have begun operation). This is, of material was diverted from inspec- course, an argument for removing Once these factors are consid- tions and hidden. In particular, if such material from the DPRK as ered, there is still the chance that the DPRK were to delay effective soon as is practicable. Fuel can be relations with the North could IAEA inspections while continuing safely, and more or less routinely, change again a second time. The on the path of warming relations removed after 2–3 years. The fuel situation could deteriorate in the

85 POTENTIAL ADVERSE DEVELOPMENTS future, after the Yongbyon spent 15 years, would be extremely Safeguards Implementation has fuel has been removed and the radioactive, exposing an unprotect- estimated that the time required to only plutonium available to the ed individual to a lethal dose of convert plutonium in spent fuel DPRK would be that in the spent more than 1,000 rads every hour at into a weapon would be one to fuel at the KEDO site. a distance of one meter. Each fuel three months, compared to seven assembly would probably require a to ten days for metallic If safeguards at the KEDO reac- ton or more of shielding in order to plutonium.3 tors were abrogated, for instance remove it from the spent-fuel stor- following a DPRK declaration of a age pool. A shielded, remotely If the DPRK were to overtly state of emergency, foreigners such operated reprocessing facility abrogate specifically to obtain as as IAEA inspectors or Republic of would have to be built somewhere much high-quality plutonium as Korea (ROK) personnel would be and tested without being detected. possible, then one might predict forced to leave the country. Remote The testing would be hard to con- the more likely times that such an monitoring equipment would be ceal, if it used “hot” materials, event would occur. The highest disconnected or removed. After because of the release of radioac- quality plutonium exists early in such a set of events, the risk of tive material such as 85Kr into the the very first reactor cycle. If the diversion of fuel becomes a very environment, which could be DPRK were to overtly abrogate serious concern. detected by remote sensors operat- specifically to obtain as much high- ed by US intelligence. Another quality plutonium as possible, then Because of the highly radioac- problem for the DPRK would be one might predict the more likely tive nature of the fuel and high that the zirconium cladding on the times that such an event would heat-generation rate for the first fuel is extremely difficult to dis- occur. The highest quality plutoni- few years after discharge from the solve in a simple PUREX facility um exists early in the very first reactor core, the older fuel in the such as the one that they built at reactor cycle. During initial start- pool would be the most likely Yongbyon. The cladding must be up, the entire core is at very low diversion target. The 59 assemblies chopped away with a heavy, burnup. If the DPRK were to shut discharged initially from the first remotely operated machine before the reactor down early (before the core load, as was mentioned in fuel dissolution. first normal refueling), it would be Chapter 4, weigh about 0.6 ton possible to recover spent fuel with each and contain 3 kilograms of On the other hand, its experi- very good plutonium isotopics. As plutonium. The burnup will be ence at constructing and operating an example, if the reactor were about 12–15 MWd/ton, and these the Yongbyon facilities would help shut down after 4 to 6 months, the assemblies would possess the best the DPRK in building this hypo- spent fuel would have only weapon-quality plutonium of any thetical reprocessing laboratory. 5,000–7,500 MWd/MT, and could of the fuel that has cooled for a sig- Rather than build a very large contain about 100 kilograms of nificant amount of time. In theory, facility such as the one at essentially weapons-grade plutoni- the DPRK could take those 59 Yongbyon, the DPRK could rely on um (~90% 239Pu). At any other assemblies and make between 10 a simpler and lower-cost facility time in the reactor’s lifetime, the and 20 atomic bombs of the type designed to separate enough pluto- quantity of this high-quality pluto- detonated at the Trinity test. That nium for a few weapons, with little nium decreases because of addi- test, it is to be remembered, used attention paid to health or safety. tional burnup (although the total about 6 kilograms of weapon- The greatest problem would be the plutonium inventory increases). grade plutonium in an “implosion” requirement to carry the main Thus, if obtaining comparatively configuration and generated over reprocessing steps with remotely high-quality plutonium for 10 kilotons of explosive energy. operated equipment. If the facility weapons were a goal, one might While possible in principle, the is built ahead of time and the pro- expect abrogation to occur at about effort to turn this plutonium into a cedures practiced (without actually the time the first one-third core set of explosive devices would face having LWR spent-fuel assemblies fuel is actually removed from the formidable obstacles.2 available for realistic tests), in prin- reactor. We emphasize, however, ciple the time needed to separate that even high-burnup, reactor- Because the plutonium is not the requisite amount of material grade plutonium can be used for separated from the fission products might be only days or weeks if all weapons. To gain an absolute max- of the spent fuel, chemical separa- went according to plan. In practice, imum inventory of high-quality tion must be carried out with very however, the time needed is likely plutonium, this would occur dur- radioactive material. Each assem- to be much longer. The IAEA’s ing the first refueling of the second bly, even if it had cooled for Standing Advisory Group on reactor. In that case, the total

86 POTENTIAL ADVERSE DEVELOPMENTS inventory of high-quality plutoni- 8.5 Scenario 4: The US or DPRK declaration had been com- um would be approximately 400 the ROK Is Unable or pleted. If verification has not been kilograms (100 kilograms of Unwilling To Meet Its completed, the DPRK might delay Beginning-of-Life fuel in the first completion of that step, bringing reactor’s spent-fuel pool, plus 300 Commitments under the the verification situation closer to kilograms in the second reactor’s Agreed Framework Even what it was before the AF was core). The time required to gain Though the DPRK Does signed, i.e., incomplete knowledge access to this material (i.e., the time of prior DPRK activities and an needed to prepare the reactor for This might be the case, for incomplete KEDO reactor. Thus, refueling and start unloading the instance, if a US administration failure on the part of the US to core) provides some (albeit limited) determines the AF is not in the US meet its commitments to the AF in response time to such an event. interest, or if Congress prevents the a timely way could, depending on completion of a nuclear coopera- the timing, itself jeopardize verifi- In either case, the challenges of tion agreement. New US legislation cation that the DPRK does not pos- rapidly removing and transporting will make gaining acceptance (non- sess nuclear materials or facilities the fuel from the reactor site would objection) from Congress more dif- to carry out a nuclear program. At be significant. Sufficient transport ficult and will probably put off the same time, if IAEA inspectors casks would have to be acquired negotiation of an Agreement for remained at the Yongbyon site, and pre-positioned near the reactor Cooperation. Lack of an activities there would continue to site. Even though the DPRK may Agreement would prevent installa- be verifiably frozen. The key factor be willing to take significant safety tion of nuclear components of US is continued presence and access compromises in shipping the fuel, origin. Again, at that point, no by the IAEA inspectors, not the the fact that it must remove the fuel nuclear fuel would have been pace of the AF itself, although the with no time for cooling means that delivered to the KEDO reactor, two are clearly linked. the casks must be capable of pro- and, of course, no plutonium viding both cooling and shielding, would have been made there. 8.6 Predicting the Future a serious engineering task. A similar issue may arise regard- The scenarios presented Presumably, for such a scenario ing nuclear liability. As noted in attempt to capture some of the to be attractive to the DPRK (given Chapter 2, Congress has prohibited key questions and decision points the potentially severe response the US from agreeing to indemnify that may occur. The DPRK’s from the US and other states), it a US manufacturer that provides efforts at normalization of rela- must also be prepared to use the nuclear components for the DPRK tions with the US and the ROK material quickly. This means the reactors. General Electric has indi- over the past ten years or so may DPRK would need to have a cated that it will not provide such make some of the more dire sce- reprocessing facility ready and components without indemnifica- narios unlikely, but they remain waiting. There would also be an tion. Negotiations have not yet possibilities. At the least, the incentive to have performed some produced an agreement to share DPRK has shown that it is capable testing, possibly “hot” testing of this liability risk. of some risky negotiating ploys. the equipment and processes. To do hot testing would require some The scenario could also be Delays in carrying out the Agreed spent fuel with which to test the brought about if KEDO ran into Framework may present benefits to systems. This spent fuel could financial difficulties, perhaps some parties as well as dangers, come from the diversion of a few owing to delays. The ROK is the such as were pointed out above. All individual fuel pins, as discussed principal financial backer of parties to the Agreed Framework in Chapter 5. If this were the case, KEDO; its contribution, in financial have, from time to time, taken then abrogation would be more and other areas, is crucial to the actions that delayed implementa- likely to occur after startup of the successful completion and safe- tion. Competing domestic, budget- second reactor, as the first reactor guarding of the KEDO reactors. ary, and security interests have reg- would have to operate for some The ROK could also decide to ularly taken precedence over the time to provide the irradiated indi- revisit its level or conditions of Agreed Framework. Some propo- vidual fuel pins for the hot-testing support if political conditions nents of the Agreed Framework program. Thus, such diversion change between the two Koreas. never intended to build the prom- could serve as an indicator of an ised reactors, but rather sought to intent to abrogate. Such a scenario could occur freeze plutonium production while before or after the verification of the DPRK ground toward an accuracy and completeness of the expected decline.

87 POTENTIAL ADVERSE DEVELOPMENTS

Despite the many hardships provision of electricity at conces- whether it will be verified is up to imposed on the North Korean peo- sionary rates, but also the opportu- the parties. With these conditions ple, however, few experts predict a nity to become fully cooperative met, verification is robust under DPRK collapse. Indeed, the coun- and open in an important area of most scenarios. tries with the most at stake—South international concern. Obviously, Korea, China, the United States, these opportunities are also chal- Notes to Chapter 8 and Japan—have gone to some lenges, both for the DPRK and for length to prevent a collapse. the US, ROK, and other KEDO 1. National Academy of Sciences, Beyond nonproliferation, the members. The challenges include Management and Disposition of Agreed Framework is part of a the challenge of verification that Excess Weapons Plutonium, “soft landing” strategy that seeks has been taken up in this report. Report of the Committee on to contain the DPRK’s problems International Security and through an extended period of rec- We believe, based on the consid- Arms Control (National onciliation with South Korea. erations of this report, that the chal- Academy Press, Washington, Continued delay could endanger lenges of verifying the Agreed D.C., 1994), p. 151. this strategy. The development of Framework can be met, under the contingency plans in cooperation conditions outlined. In essence, 2. David Albright and Kevin with the ROK and Japan must be these conditions account for the O’Neill, eds., Solving the North part of an execution strategy for IAEA to be ready and capable of Koran Nuclear Puzzle, Institute the Agreed Framework. special efforts in this case, efforts for Science and International that will probably require enhanced Security Press, Washington, The Agreed Framework faces its US and ROK support, and for the D.C., 2000), Chapter 3, p. 9. parties with both opportunities DPRK to fully cooperate with the and challenges. The opportunities IAEA, and open as to its past activ- 3. Albright and O’Neill, Chapter 8, for the DPRK include not only the ities. In other words, the Agreed p. 4. Framework is verifiable, but

88 Verifying the Agreed Framework: Conclusions and Recommendations

Safeguarding the • The IAEA is under severe budget • With adequate preparation and KEDO Reactors constraints and the member unhindered measurements and states (in particular, the US, ROK, inspections (including access to • The applicable International and Japan) can usefully assist the appropriate records), the IAEA Atomic Energy Agency (IAEA) IAEA by making available the can assess past DPRK nuclear safeguards are adequate for the technical and financial resources activities to reasonable accuracy timely detection of the diversion necessary to implement the and confidence. With DPRK of nuclear material. Timeliness is DPRK’s Safeguards Agreement. cooperation, the process is esti- taken as the estimated time to mated to take 2–4 years. convert diverted nuclear materi- • Removal of Beginning-of-Life fuel al into weapons-usable form. from the DPRK as soon as practi- • The exact quantity of plutonium cal is desirable due to both the separated can only be approxi- • Advanced safeguards technolo- quality of the contained, weapons- mately determined. Depending gies, such as real-time remote usable plutonium in such fuel and on the reactor operating history monitoring, are being tested on the fact that it will be the first to available, there may not be high the type of reactors being pro- cool (i.e., lose radioactivity), which confidence in the exact number vided by Korean Peninsula facilitates ease of handling. Nego- of kilograms separated. Energy Development tiating an agreement for such Organization (KEDO). Upon removal as early as possible • The IAEA is likely to need completion of this testing, it will would be usefull. added resources to prepare for be desirable to implement these and deal with the problem of measures in the Democratic Past DPRK Nuclear Activities verifying the DPRK declaration, People’s Republic of Korea and beyond this, to verify that (DPRK) as they will enhance the • The DPRK’s initial declaration to the DPRK is complying with the effectiveness of safeguards. the IAEA identifying facilities Non-Proliferation Treaty as Implementation by the Republic and quantities of nuclear materi- required by the Agreed of Korea (ROK) is probably a al subject to safeguards appears Framework. prerequisite to implementation to be incomplete. At least one by the DPRK. undeclared waste site has been • Access to third-party informa- identified, probably containing tion provided by member states • All safeguards-relevant monitor- additional plutonium. There is will continue to be an important ing and data-transmission evidence indicating more fuel part of IAEA verification and equipment must properly and removal and more plutonium- safeguards activities. securely installed and adequate- separation activity than the ly maintained, and data trans- DPRK has declared. • The US and other states support- mission must be secure and ing nuclear non-proliferation uninterrupted. The IAEA must • An amended DPRK declaration, objectives, especially the ROK also have adequate resources to confirmed by IAEA inspections and Japan, need to support the ensure that inspectors are prop- and measurements, will very IAEA in maintaining a “standard erly trained, and that safe- likely be required. With such an of verification” for the DPRK, as guards-relevant information is amended declaration, the US expected for other member states. reviewed in a timely manner. and other interested parties would have more complete and • The Agreed Framework requires • Full cooperation and openness by reliable knowledge of the dismantlement and disposal of the DPRK is essential to the suc- DPRK’s nuclear materials and the identified nuclear facilities at cessful implementation of its facilities, and more complete Yongbyon after the first KEDO IAEA Safeguards Agreement. and ongoing safeguards over reactor has been completed. The such material and facilities cost to dismantle these facilities, would be possible. based on past experience, is likely

89 CONCLUSIONS AND RECOMMENDATIONS

to be at least a few hundred mil- Possible Adverse • Lack of agreement over or inter- lion dollars. Developments ference with the extent and nature of advanced safeguards • Spent fuel of the type at the • Delays at this point of imple- technologies to be applied to the Yongbyon site cannot be stored mentation of the AF have little KEDO reactors could limit the indefinitely. It will have to be direct effect on verification abili- effectiveness of the IAEA’s transported to a facility that can ty. Delays later, after one or both safeguards. safely handle and treat this type KEDO reactors are completed, of spent fuel within a few years. for instance in allowing special • Abrogation and other overt viola- Such an operation raises signifi- inspections (in addition to those tion of the Agreed Framework or cant technical and political before completion and already Non-Proliferation Treaty cannot challenges. Negotiating and requested as part of the Supply be prevented by verification identifying a site to which this Agreement), could have more measures. Early removal of spent material can be removed should serious consequences. KEDO fuel would minimize the be done as early as possible. amount of plutonium-containing • Disagreement over the need for material in the DPRK in the event • Some of the crucial pipes and an amended DPRK declaration or of overt abrogation. Application the special equipment at the over the means used by IAEA to of additional methods beyond identified Yongbyon facilities verify the declaration could pre- INFCIRC 153, notably broader must be removed or destroyed vent the US and other interested environmental sampling and eas- to make the dismantlement veri- parties from having a full knowl- ier inspections at undeclared fiably irreversible. It is desirable edge of past DPRK activities. facilities, would diminish the to do this as soon as possible. uncertainty associated with the • Failure to remove the Yongbyon detection of undeclared activities spent fuel would leave weapons- and increase the warning time of usable material in the DPRK. possible overt violations.

• Failure to dismantle and dispose of identified nuclear facilities in the DPRK would leave them available for later use.

90 DISCLAIMER

This document was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor the University of California nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, complete- ness, or usefulness of any information, apparatus, prod- uct, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or oth- erwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or the University of California. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or the University of California, and shall not be used for advertising or prod- uct endorsement purposes.

Work performed under the auspices of the U. S. Department of Energy by the University of California Lawrence Livermore National Laboratory under Contract W-7405-Eng-48.

This report has been reproduced directly from the best available copy.

Available to DOE and DOE contractors from the Office of Scientific and Technical Information P.O. Box 62, Oak Ridge, TN 37831 Prices available from 423.576.8401 http://apollo.osti.gov/bridge/

Available to the public from the National Technical Information Service U.S. Department of Commerce 5285 Port Royal Rd., Springfield, VA 22161 http://www.ntis.gov/

or

Lawrence Livermore National Laboratory Technical Information Department’s Digital Library http://www.llnl.gov/tid/Library.html

G1210