Centrifuge Cascades and

a Final Deal with Iran

JINSA’s Gemunder Center Iran Task Force 12-14

Co-Chairs Ambassador Eric Edelman and Ambassador Dennis Ross

This report is a product of JINSA’s Gemunder Center Iran Task Force. The findings expressed

herein are those solely of the Iran Task Force. The report does not necessarily represent the

views or opinions of JINSA, its founders or its board of directors.

Gemunder Center Staff

Dr. Michael Makovsky

Chief Executive Officer

Ashton Kunkle

Research Assistant

Jonathan Ruhe

Associate Director

Co-Chairs

Ambassador Eric Edelman

Former Under Secretary of Defense for Policy

Ambassador Dennis Ross

Former special assistant to President Obama

and NSC Senior Director for the Central Region

Members

The Honorable Chris Carney

Former U.S. Representative from Pennsylvania

Professor Eliot Cohen

Director of Strategic Studies Program at Johns

Hopkins School of Advanced International

Studies

Lt. General (ret.) David Deptula

Former Deputy Chief of Staff for Intelligence,

Surveillance and Reconnaissance, U.S. Air

Force Headquarters

Larry Goldstein

Founder and Director of Energy Policy

Research Foundation, Inc.

John Hannah

Former Assistant for National Security Affairs

to the Vice President

Admiral (ret.) Gregory Johnson

Former Commander of U.S. Naval Forces,

Europe

Alan Makovsky

Former Senior Professional Staff Member,

House Foreign Affairs Committee

Steve Rademaker

Former Assistant Secretary of State for Arms

Control and Nonproliferation

Ray Takeyh

Senior Fellow for Middle Eastern Studies,

Council on Foreign Relations

General (ret.) Charles Wald

Former Deputy Commander of U.S. European

Command

Mort Zuckerman

CEO and Chairman of the Board of Directors,

Boston Properties, Inc.

Task Force and Staff

Table of Contents

Overview 5

Centrifuge Cascades 5

Issues for Implementation 7

Implications 9

Endnotes 10

Centrifuge Cascades and a Final Deal with Iran 5

Overview

Significant differences remain between Iran and the P5+1 over the parameters of a

comprehensive agreement on the Islamic Republic’s nuclear program. In particular, Tehran

has resisted agreeing to dismantling any of its existing uranium enrichment infrastructure. With

the November 24, 2014, deadline for a final deal looming, U.S. negotiators have reportedly

considered several workarounds intended to roll back Iran’s breakout timing while leaving its

existing centrifuges in place.

As this Task Force laid out in a September paper, one such possibility would be to limit the total

output of Iran’s enrichment facilities (as measured in Separative Work Units, or SWU).1 Since

then, U.S. officials reportedly have considered another route, whereby Iran would disconnect

the links between some or all of its thousands of installed centrifuges.2 Depending on the

extent of the disconnections, this could potentially increase Iran’s breakout timing anywhere

from a matter of days to months, were it ever to renege on a final deal and reconnect its

centrifuges. Unlike the SWU approach, Iran would not necessarily remain a flip of a switch

away from sprinting to a bomb. However, as with the SWU approach, Iran would maintain a

latent nuclear weapons capability, and could even expand and upgrade its existing nuclear

infrastructure without violating a final deal.

Centrifuge Cascades

Uranium enrichment is a process wherein a centrifuge increases the concentration of fissile

isotopes in uranium by separating the small quantity of fissile isotopes from the heavier, and

much more common, non-fissile ones. The uranium input for a centrifuge is referred to as its

“feed.” The higher-enriched output is called the “product” (or “heads”), and the lower-enriched

byproduct is called the “waste” (or “tails”). Because the difference in atomic mass between

these two isotopes is just over one percent, individual centrifuges can only achieve a very

minimal amount of enrichment. Therefore, large numbers of these machines are connected

in stages, via tubes, to create a “cascade” that increases the enrichment level over that of

a single centrifuge. These tubes take the product from one centrifuge, now slightly more

enriched than before it was fed into that machine, and feed it directly into another centrifuge,

which enriches it further, and so on. Separate tubes can also feed the waste from this process

back into earlier stages of the cascade for further enrichment. Finally, multiple series of

centrifuges are often connected in parallel to increase the product flow rate of a cascade.3

Under the Joint Plan of Action (JPA) interim deal on its nuclear program, Iran has slightly less

than 10,000 operating centrifuges producing 3.5 percent low-enriched uranium (LEU). These

are grouped into 60 cascades, each consisting of 164 or 174 centrifuges. Prior to the JPA, Iran

interconnected some of these cascades in pairs – often referred to as “tandem cascades” –

to enrich uranium to higher levels more efficiently than by using separate cascades. As part

of Iran’s agreement not to enrich LEU beyond five percent during the interim deal, the JPA

required Iran to remove the tubes interconnecting these paired cascades. Though the original

cascades continue producing 3.5 percent LEU, removing the interconnecting tubes between

cascades makes it more difficult, though not impossible, for Iran to return to enriching uranium

to 20 percent or higher if it chose to violate the JPA by doing so.

6 Centrifuge Cascades and a Final Deal with Iran

Figure 1: Model Iranian Tandem Cascade4

Centrifuge Cascades and a Final Deal with Iran 7

Issues for Implementation

Determining the impact on Iran’s nuclear program of the P5+1’s newest proposal depends

on several unresolved issues. It can be presumed that the scale of disconnection would be

much greater than what Iran has already done under the JPA: as Figure 1 indicates, the sheer

number of tubes within a cascade is at least an order of magnitude larger than those between

cascades. Furthermore, the JPA only required Iran to disconnect the tubes between two pairs

of cascades at Fordow; under a final deal Iran possibly could have to disconnect the tubes

within several dozen cascades.

There remain several other issues that would have to be resolved before being able to

gauge the overall effect of such a proposal on Iran’s breakout timing, including: the nature of

disconnection; the number of cascades in which centrifuges would be disconnected; and the

safeguards against Iran reconnecting the tubes. Importantly, none of these factors would roll

back Iran’s latent enrichment capability. As with the idea of capping SWU output, this proposal

would not require dismantlement of any centrifuges or the facilities containing them. Iran could

thus remain in position to enrich sufficient fissile material for a nuclear weapon, should it ever

cheat on such an agreement. Moreover, it could maintain or even expand the overall capacity

of its enrichment program under such a deal, potentially leaving it well-positioned for an

industrial-sized nuclear program once a comprehensive agreement expires.

What would “disconnecting” the tubes entail?

The definition of “disconnect” would affect how far Iran’s breakout timing could be rolled back.

If it means “no connections” – paralleling the JPA’s language on “no interconnections between

cascades” – Iran would have to physically remove the tubes connecting individual centrifuges.

According to various nonproliferation experts, it could take anywhere from several days to

several months to reinstall the tubes and run the necessary tests to ensure the cascades

function properly again. The wide range of estimates stems partly from the lack of any

precedent by which to judge Iranian engineers’ proficiency at reconnecting cascade tubes.

This process may not simply be the reverse of disconnection, since the timeframe for the latter

includes both disconnection and decontamination. Thus, the 1-2 days it took Iran to remove the

interconnectors between the two pairs of tandem cascades at Fordow (as per the JPA) likely

provides no more than a rough approximation of the time required to reconnect them.5

Alternatively, the term could imply merely shutting down the tubes, without removing them from

the cascades. This would likely have minimal effect on Iran’s breakout timing, because it could

remain potentially just a flip of a switch away from reactivating the cascades in question.

Either way, even if its breakout timing is affected, Iran’s latent enrichment capability would

remain intact. Disconnecting tubes would not involve removing or in any other way dismantling

the centrifuges themselves. In fact, depending on what other constraints, if any, would be

placed on its enrichment program under such a deal, Iran could potentially expand or upgrade

its centrifuges as long as it did not connect them.

How many cascades would be disconnected?

In conjunction with the degree to which tubes are disconnected within a cascade, breakout timing

would be affected by the number of cascades in which such disconnections occur. It would do

so in two ways. First, the more tubes that would be removed, the longer it would likely take Iran

8 Centrifuge Cascades and a Final Deal with Iran

to reconnect them and run the necessary tests before reactivating the cascades. Second, the

fewer cascades Iran would have operating, the more time-consuming it would become to enrich

a bomb’s worth of weapons-grade uranium. Though much of the debate over a final deal has

concerned the number of centrifuges Iran would be allowed to keep, this is fundamentally a

negotiation over the quantity of cascades, given the miniscule amount of enrichment achieved

by unconnected centrifuges (even if thousands are operating in this manner). For example,

increasing Iran’s breakout timing to six months to a year – which would entail cutting its operating

centrifuges from 10,000 to 2,000-4,000 operating centrifuges – equates to Iran disconnecting all

but approximately 12-24 of the 60 cascades it operates under the JPA.

Consequently, the number of cascades to be disconnected could have a significant impact

on Iran’s breakout timing, assuming it would not be able to restart them. As with the issue

of defining “disconnection,” resolving this question would not by itself preclude Iran from

maintaining or expanding the number of installed centrifuges.

What safeguards would be included?

As the preceding sections indicate, simply disconnecting the tubes within cascades would

not be certain to roll back Iran’s breakout timing significantly, and would not limit the size of its

overall enrichment program. Thus, even if Iran were required to physically remove every tube

from every cascade, safeguards would still be crucial to determining the effect on breakout

timing. Physically removing tubes from the cascades would simplify the verification of Iranian

compliance, as long as the equipment was mothballed off-site in locations under constant

supervision of International Atomic Energy Agency (IAEA) inspectors.

Merely turning off the tubes without removing them would complicate verification. IAEA

cameras at Iran’s enrichment facilities monitor the feed and product levels at the beginning

and end of a given cascade, respectively, but not the tubes within them, which must be

accomplished during on-site inspections by IAEA personnel. If it chose to renege on the final

deal, Iran could thus change the configuration of its dormant cascades to increase their latent

enrichment capability without automatically being detected.

The IAEA Additional Protocol, to which Iran has agreed to adhere under a final deal, would

not necessarily spell out airtight safeguards on either count. Access to storage facilities could

presumably be the product of Iranian negotiations with the P5+1 and IAEA – as it is under the

JPA – and therefore unlikely to include complete removal of the equipment from Iranian control.

Moreover, while the Additional Protocol does allow unannounced inspections of cascade halls,

these are regulated by the Low Frequency Unannounced Access (LFUA) regime, which limits

such visits to 4-12 times per year. Even if the IAEA could carry out the maximum number of

these visits annually, Iran might conceivably reconfigure its cascades before being detected,

should it ever choose to do so.6

These IAEA safeguards would need to be understood in the context of any other constraints

Iran negotiates with the P5+1 on its enrichment program. The most stringent limit on cascade

tubing would do nothing to roll back Iran’s nuclear program for the long term if it is not

accompanied by dismantlement of key elements of its existing enrichment infrastructure

– specifically centrifuges – and verifiable limits on centrifuge output, number and types

of operating and installed centrifuges, research and development (R&D) activities, and

enrichment levels and facilities, among others. Without these additional restrictions, Iran could

expand its latent enrichment capability while adhering to a final deal.

Centrifuge Cascades and a Final Deal with Iran 9

Implications

Beyond technical concerns over the viability of disconnecting cascade tubes, such a proposal

would represent a rollback of U.S. redlines – and a reinforcement of Iran’s – regarding the

latter’s nuclear program. Specifically, it would contradict statements by Administration

officials since the JPA was agreed that Iran must dismantle significant amounts of its nuclear

infrastructure, and that it must close its Fordow enrichment facility. Simultaneously, the proposal

would underscore declarations by Iran’s Supreme Leader Ali Khamenei, President Hassan

Rouhani and leading Iranian negotiators that they would never agree to dismantle a single

centrifuge or to close Fordow.

This would limit U.S. credibility when it comes to enforcing adherence to a comprehensive

agreement. Promises to punish violations – whether by Iran, other countries or companies

eager for the lifting of sanctions – would likely gain less traction if the United States was

attempting to uphold a deal whose terms it had previously said were unacceptable.

Furthermore, were Iran ever to decide to reconnect the tubes, the potential difficulties for the

United States and its diplomatic partners of detecting such activities, discerning whether

they constitute a clear violation and agreeing to an appropriate punishment before Iran had

completed the process, could all compound the challenges stemming from limited credibility at

the outset of the final deal.

10 Centrifuge Cascades and a Final Deal with Iran

Endnotes

1. JINSA Gemunder Center Iran Task Force, “Separative Work Units (SWU) and a Final Deal with Iran,” September

22, 2014.

2. David E. Sanger, “U.S. Hopes Face-Saving Plan Offers a Path to a Nuclear Pact with Iran,” New York Times,

September 19, 2014.

3. For an overview of uranium enrichment technology, see: Ivanka Barzashka and Ivan Oelrich,”Enrichment

Cascades,” Federation of American Scientists (accessed October 2014).

4. This figure is derived from models suggested by staff at the Institute for Strategic and International Studies. See,

for example: Wiliam C. Witt et al., “Modeling Iran’s Tandem Cascade Configuration for Uranium Enrichment by

Gas Centrifuge,” paper presented at INMM 54th Annual Meeting (Palm Desert, CA), July 14-18, 2013.

5. Time estimates for Iran to disconnect and reconnect centrifuge tubes were provided to Gemunder Center staff

in not-for-attribution discussions with nonproliferation experts.

6. Nuclear Safeguards and the International Atomic Energy Agency (Washington, D.C.: Office of Technology

Assessment, 1995), 72.

1307 New York Ave., NW • Ste. 200 • Washington, DC 20005 • (202) 667-3900

www.jinsa.org