ML20214R219
Text
.
- llNCu.,,._,f"%,
4#.na -
I e
\\...../
4N i
y POLICY ISSUE d
(Information)
March 11, 1986 SECY-86-87 N
For:
The Comission
$ From:
Victor Stello, Jr.
Acting Executive Director for Operations
Subject:
REVIE!! 0F G. lilLHOLLI!!'S "H~.AVY UATER I ;
IllDIA: A STUDY"
Purpose:
(U)
To provide the Comission with the staff's evaluation of the W
- tilhollin study in response to the February 3 reouest from the Chairman's office (Appendix A).
3(
tt Discussion:
I yi e
j
>1 i
el 4
]i Iea a
a j
3 3
la
('U)
Professor liflhollin's study can be summarized as follows:
5 (1) As of 1985 India had six research and power reactors
%g !! :
needing different amounts of heavy water moderator, with the j#.
total cumulative need 1239 metric tonnes (mt) through
.s f.(.
1985; (2) India produces both virgin and upgraded heavy water indigenously, with the total cumulative production
'T g
through 1985 393 at: (3) India has not indigenously produced d<j 0 enough heavy water to meet its needs, and has had to import heavy water from the U.S. (19 saf d mt the U.S.S.R.
d.;. '. m4 C0flTACT:
K. Henderson, IP 49-24660 U
h0 5 861126 j!
GARDE 86-A-27 PDR..
7 * * *
- g.,,,,.,
Jdwi e.,,....O The Commission 2
(E0 possibly unsafeguarded et and a21 safeguarded ot), and Canada (120 safeguarded nt); and (4) coroaring the total of donestic production and known imports with heavy water needs, India does not have enough heavy water to operate the reactors, for exanple, there is a shortage of about 200 nt in 1905.
(U)
Staff review, using both Professor Milhollin's references and other onen sources, shows that the infornation as collated by Professer :tilhollin, in i
particular the figuras he uses to determine heavy water needs, losses, upgrading and recovery capabilities, e
generally accord with publicly available infomation.Lstaff notes, however, that Professor Milhollin may not adequately have taken into account the fact that for extended periods of tine the RAPP-I reactor has been shut down, thus reducing predicted loss rates. While inclusion of this fact might have reduced Professor l111hollin's predicted cap between Indian heavy water requirements and production rates, the shortage, nevertheless, would continue to exist. !!oreover, heavy water production figures are not made public under India's atomic energy laws, and thus are unavailable for confirmation.9 (U) [ While it is possible, using only open sources, to derive Professor Milhollin's points sumarized above, it is the i
staff view that Professor Milhollin has cote to certain i
other conclusions which are not necessarily supported by his data. For example, Professor :tilhollin concludes that India must have a clandestine source of unsafeguarded heavy water, probably the People's Republic of China (PRC), to account for the shortage of heavy water. Given the infornation he has documented, this is an assumption representing only i
speculation, and his policy recomendations (discussed i
below) are personal prescriptions only.
Finally, it is noted that China has not signed the NPT, and has no obligation to have heavy water shipments placed under IAEA l
safeguards.1 1 (C) c A
A L-l l* *
- 4hw...
l i
h,a[,]',"I ~ ft l
The Connission 3
e 8
4 4
I I
k i
e i
(C) n 4
I I
i
\\
I'
\\
(C) l I
l a
-~
~
- h k
s- 'a j j l
ctom....,.,
~....... ;
The Conmission 4
i f
i
_~
r 1
q
- ctor Stello, Jr.
Acting Executive Director for Ooerations Apcendices:
A.
- ieno to TRehn frm NHaller dtd 2/3/86 E.
G. Milhollin's study on Heavy Water in India DISTRIBUTION:
Commissioners OGC l
OPE i
)
ELD SECY i
i l
l i
(U)
[,*SEC notes that Professor fiilhollin had access to extrenely sensitive classified data two years ago concerning Incita's nuclear progran when orenaring the report 1
"Tarapur: A Brief for the United States," for former Comissioner Gilinsky.
Professor 'tilhollin's interest in India may have been heightened by this access, and he may therefore continue to raress his interests Fy writinc additional articles. SEC plans to furnish Professor Milhollin with guidelines regarding the generation of future documents in this general subject area.]
I P
I
&g c
t
0
.)
/ s.
APPENDIX A 1
1
U:
- f,'
UNITED STATES s'!s ' '. ' -
gj "yg g
NUCLEAR REGULATORY COMMISSION WASHING TON. D. C. 20555
%, /)f Februa ry 3, 1986 CHAIRMAN MEMORANDUM FOR:
Tom Rehm, Assistant for Operations Office of the Executive Director for Operations j
'FROM:
Norm Haller, E x e c u t i v e A s s i s t a n t ])'
Il to the Chairman
SUBJECT:
MILHOLLIN STUDY OF HEAVY WATER IN INDIA The Chairman requested that the staff evaluate the above Milhellin Study.
The evaluation should place specific emphasis on matters related to NRC's responsibilities.
Please provide by February 7, 1986 an indication of'when such an evalua tion can be completed.
cc:
Commissioner Roberts Commissioner Assels tine Commissioner Bernthal Commissioner Zech OPE OGC SECY EDO --- 301387
n=
e v
/ '
i
/.
s N
f 4
y e
- l
~.
( -
w_
N i
APPENDIX B
.w b
D 4
l *
/
e s
s d
s
'u U
5 a
9 1
f a
i l4 b
1 l
l s
.-.a
___,___n_
,,,y_,,-
- \\c
- y. q THE WISCONSIN PROJECT ON NUCLEAR ARMS CONTROL i
'i
- . t
.l.
. l\\
o s ',
1 Heavy Water in India t
t 1
A Study
- t.
by Gary Milhollin Professor of Law University of Wisconsin Law School-and Scholar in Residence Natural Resources Defense Council Washington, D.C.
January 28, 1986 i
?
l k
This is a study of heavy wa te r in India.
It concludes tha t India has obtained heavy wate r on a clandestine basis, and is using it to run three new nuclear reactors in apparent violation of inte rna tional obliga tions.
The clande stine heavy wa te r will give India her first chance to build a nuclear arsenal.
Up to now India has had only small amounts of plutonium, the nuclear explosive used at Nagasaki, and those amounts we re re s tr ic te d to peace ful use by pledge s India gave he r nuclea r supplie rs.
The ne w reactor s, howe ve r, produce enough plutonium for fif tee n bombs pe r year, and the plu tonium will not be re s tr ic ted to peace f ul use.
The da ta in this study show that India must be running the new reactors -- which are not unde r inte rnational safeguards or re stricted to peace ful use -- with re stricted heavy wa te r d ive r ted from othe r reactors or with heavy wa te r imported secre tly f rom China.
A se rie s of myste rious heavy wa te r shipments in 1983 can be explained only by a secte t i
import from China, or by a transfe r of re stricted heavy water to an unre str ic ted re ac tor.
India has never produced or publicly imported enough heavy wa te r to run all of he r re ac tor s a t the same time.
She needed 1,239 tonnes to run her reactors in 1985, but had only 1,038 tonnes in her public records, which leaves a shortage of 201
__m.
tonne s.
This shor tage had to be cove red by dive rsions or secre t imports.
Because of the se facts, this study recommends that the United States (which is Ind ia 's la rge st nuclear supplie r) inquire immed ia te ly into Ind ia 's heavy wa te r source s Also, the Reagan Administra tion should make public what it knows about Ind ia 's nuclear intentions.
Ind ia 's nuclear supplie rs (the United Sta te s, Canada, and the Sovie t Union) should suspend nuclear trade with India until the large d isc repancie s in Ind ia 's pub,lic,
records are e xplained.
Also, the nuclear supplie r countrie s should suspend nuclear trade with China until China promises to stop making clandestine expor ts.
Heavy Wa te r in Ind ia Heavy wa te r (deute rium oxide) is the key to India 's nuclear program.
Heavy wa te r is essential to achieve a chain reaction in India 's natural uranium fueled re ac to r s.
He r reactors cannot ope ra te wi thou t it, and to ope ra te safely each reactor must have a full inventory.
.But India has never produced enough heavy wa te r to mee t he r needs.
She has been forced to rely upon imports, and to promise that she would use the imported heavy wate r for peace ful purposes only.
Also, she has been required to place almost all of her impor ts unde r inte rna tional inspec tion (the safeguards system administered by the Inte rnational Atomic Ene rgy Aconcy).
India has two powe r reactors at Rajasthan (RAPP-
I and II), two powe r reac tors a t Mad ras (MAPP-I and II) and two re se'a rch re ac tor s (CIRUS and Dhruva) which require heavy wate r.
Despite an installed annual capacity of over 300 tonne s, Ind ia produced only 39 tonne s of heavy wate r in fiscal year 1982.
In o the r ye a r s the prod uc t ion wa s e ve n lowe r.
When one adds up all of India's own production, and adds to that all of the amounts she has eve r impor ted, one reaches a startling conclusion:
the total -- which is India's entire supply -- is not enough to ope ra te he r re ac tor s.
The re sults are shown in the following table.
e v
r w,--
lit %W m1Ut IN IranA I85 - EdH'l (notric tems)51w;18, 1 - 3)
" PAR 1960 61 62 63 64 65 66 67 68 69 70 71 72 73 14-75 76 71 78 79 00 51 82 u) U4 05 861
_PDD M IIIMW m1TJt
~
carts
_19" IWP-I b
230 16 IG 16 16 16 16 16 16 IG 1 6
- IGJG, itAIT-I t HAIP-I 210 16 16 16 16' 16 1111tuvA 250" 12 12 12 PW P-I I
,,,90'I 18_
lura!. I;17U (OMRATIVE) 19__ 19 19 19 19 19 19 19 19 19 19 19 19 249 265 201 297 313 329 345 361 607 639 921 96512J9_1 17f 00
-imum; SIMf.Y CF lirAW My trA 19 carwg E
120 tc;n 809 55
75I 70 41 40 40 20 _ 40 _?407 35[
l'fe s t R.TI(tl k
tWRZl.
5 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 IL _10 10_
10 10 10 rwanin 1VrlO SIN le 13 6 14 14 1 20
M7rA 7"
10 15
_6 23 23 46 1UTAI, St011.Y(CituATIVE)19 19 24 34 44 54 64 74 '84. 94 104 114 124 334_ Jtt 354 419 504 504 594 652 734 n12 8
- total 5
_5 5
TCIEitVE Or straw m1 Tat (Sugply Miness tbed)
_1UrAF. Rru3tvc(OtuATIVE) 5 15 25 35 45 55 65 75 OS 55 105 OS 79 73 122_ 191 255 249 291 127 173 -G7 201 -219, SAnn wunn 55* 130 200 200 225 19P 27 15 21 31 34 intnu2 Wit tD 5 1: 25 35 45 55 65 15 OS 95 105 85 79 73 67 61 55 49 66 108__1_46 42 -232 -253 8
Note s to Table :
a)
Cirus required 19 tonnes to start up, and is assumed to have a negligible annual loss of heavy water.
The Cirus design anticipa tes a' loss ra te of.3% pe r year, which would produce a total loss of 6 tonnes during its 20 years of operation.
b)
The RAPP reactors each regaired 230 loss rate of 16 tonne s/yr is ve ry conse rva tive. tonne s to start up, and a The losses are assumed See Hart,_ infra, a t 75.
shutdown in 1982; to continue a t RAPP-I de spite the ope ra tion.
repairs are assumed to cause as many leaks as c)
MAPP-I is assumed to have for full powe r ope ra tion a t the end of 1983. required 250 tonnes of d)
Dhruva is assumed to have required only 60 tonnes of heavy wa te r for criticality.
ope ra tion in 1986.
It will require 78 tonnes for full power 1
e)
MAPP-II is assumed to require 170 tonnes of heavy water criticality. It will require for in 1986.
250 tonne s for full powe r ope ra tion f)
U.S. heavy water supplied th r'ough Ca nada..
g)
Hart, Nuclea r Powe r in India (1983) p. 45; Potte r, "Sovie t Nuclear Expor t Policy", in Snyde r and Wells, Limiting Nuclear Proli fe ra tion (1985) p. 227.
It is not clear whe ther the se 80 tonnes were supplied subject to sa fegua rds.
They are counted he re as unsafeguarded in order to give Ind ia the bene fit of the doubt.
The 80 tonnes may be subje ct to a peace ful use guarantee imposed re troactively in 1976-1978 wa te r for RAPP-II.
when.the Sovie ts supplied the h)
Hart, id. a t 4 7, a nd Po t te r, id.,a t 230.
i i
i)
Po t te r, id., a t 230.
The Sovie t impor t quantities we re confirmed by Mrs. Gandhi in March of 1983, when she 3
told the Indian parliament that total imports had, reached -546.6 tonnes.
j)
The Soviets promised in 1980 to supply another 256 tonnes.,
j With the delive ries assumed for '84 and remain to be delivered in 1986.
' 85, only 35 tonne s would 4
k)the actual amounts fluctuatedThe 10 tonnes pe r year for Nangal is an estim from year to year.
l o
1
1)
Boroda and Tuticorin are assumed to produce the same amount in 1985 as they did in 1984, and Kota is assumed to produce as much in 1985 (5 tonnes) as the total amount it produced in all the years through 1984.
m) remain the sane,Tuticorin and Paroda production is assumed to double, Kota to n)
The production statistics for Indian heavy water come se rie s of article s in The Time s of India by Praful Bidwai, from a May 6-9, 1984.
o)
It is assumed unsa fegua rded heavy wa te r through 1979 becausetha t losse s fro impor ted from 1976 throu9h 1978 was not the Sovie t heavy wa te r used until 1980.
Po t te r, id., a t 230.
p)
It is assumed tha t a f te r make up losse s f rom RAPP-I and RAPP-II,1979, Sovie t heavy wa te r was to and to provide the original inventory for RAPP-II.
9 4
1
What Doe s the Ta ble show?
The table shoss' tha t India 's supply of heavy wate r kept up with her need until 1983.
Howe ve r, in tha t year he r total need
-- which consisted of all the amounts needed to star t and run her reactors -- exceeded he r supply, which was all the amounts she had ever imported or produced.
He r total need was 921 tonne s and,
he r total supply was 854, which le f t a ne t shortage ot 67 tonne s in he r re se rve.
By re se rve, one means the total amount available j
outside of reactors.
Tha t is, she needed 67 tonne s more he avy wa te r to run her reactors than she had ever produced or rece ived.
By 1985, the shortage was 201 tonne s, almost an e ntire reactor 's worth. This is a remarkable fact.
It leads to a numbe r of conclusions.
First, India did not have enough heavy water to commission MAPP-I in 1983.
In fact, she was 67 tonne s shor t.
That is, 67 tonne s short of heavy wa te r available outside of othe r reactors, regardle ss of any que stion of safeguards.
Second, the shortage of unsafeguarded heavy water in 1983 wa s e ve n g re a te r.
It totalled 82 tonne s.
MAPP-I was sta r ted without safeguards, so 47 tonne s of available safeguarded heavy wa te r we nt to RAPP-I and II, which we re safeguarded.
This inc rea sed the shortage of unsafeguarded heavy water needed to S ta r t MAPP-I by 15 tonne s, a nd me an t tha t the re was a total shortage in India of 82 tonne s (67 + 15 = 82) of unsafeguarded heavy water at the end of 1983 when MAPP-I reached full powe r.
Third, in 1985, she n Dh r u va a nd MAPP-II we n t c r i t ica l,
the a bsolu te shortage of heavy water in India reached 201 tonnes, almost a full reactor 's wor th.
That is, India had neve r produced or rece ived enough heavy wa te r to run all of her reactors at the time, and was short of that by a huge same amount.
By the end o9 1985 India was running four powe r reactors with only enough mate rial in he r public records to run three.
This is inde penden t of any que stion of sa feguards.
.I Fourth, the shor tage of unsafeguarded heavy wate r in 1985 reached 232 tonne s, mo re than a full reac tor 's wor th.
This was because Dhruva and MAPP-II we re started without safeguards.
The shortage would inc re a se in 1986, and continue inde finitely into the fu tare.
Wha t do the se shor tage s mean?
The y me a n,, qu i te simply, tha t India cannot be running her program honestly.
The data show tha t Ind ia mu s t be increasing he r total supply of heavy wate r by some secre t means, and also increasing her supply of unsafeguarded heavy wa te r, probably by the same me a n s.
The re are only three ways to do thic:
India could impor t heavy wa te r secre tly from China; she could drain it out of RAPP-I while tha t reactor was shut down; or she could do a combination of both.
It is clear tha t she didn 't produce the wa te r he r se lf.
A sudden shipment of 100 tonne s of high grade heavy wa te r f rom Bombay to MAPP-I in l
1983 is une xplained, e xcept by an import f rom China.
Also, a se ries of hurried shipments of heavy water f rom Rajasthan to Madras in 1983 is une xplained, e xcept by the transfe r of wa te r from RAPP-I to MAPP-I.
When one combines the se facts with the absence of any othe r possible source s, one conclude s tha t MAPP-I mus t ha ve bee n s ta r ted wi th he a vy wa te r imported secre tly f rom China, or with heavy wa te r taken out of RAPP-I.
The same is true of Dhruva and MAPP-II.
How did this happen?
The Evide nce In orde r to br ing MAPP-I to full power a t the end of 1983, India had to ge t 67 tonne s of heavy wa te r f rom sone whe re.
She needed tha t much more wate r to ope ra t all of he r reactors than she had on hand.
The re we re only two source s:
she could shif t wa te r to MAPP-I f rom anothe r reactor, or she could use an import.
Heavy water reactors need a full inventory of heavy wa te r to ope ra te, so India could shif t heavy water from anothe r reactor only it it we re shut down.
RAPP-II wa s ope ra ting in 1983.
RAPP-I was shut down.
RAPP-I ope ra ted until March of 1982 when it was shut down due to heavy wa te r le ak s.
It stayed shut down until January of 1985, which means that it could have been drained to start MAPP-I in 1983. When RAPP-I closed it had reactor-grade heavy wate r in its cooling and mode ration systems, and degraded heavy wate r collec ted from its leaks.
All of that heavy wate r was sitting
idle, not be ing used.
This wa te r could have been shif ted to MAPP-I in only two ways.
It could have been " upgraded" af te r leaking out of RAPP-I and becoming "de g raded, " or it simply could have bee n transfe rred directly.
First, le t us conside r upgrading.
Wha t doe s tha t mean?
When heavy wa te r e scape s f rom the reactor cooling system (a round pumps, seals, e tc.) it becomes vaporized. This is becausa the wa te r in the re ac tor ir at high te mpe ra tu re and pre ssure.
The e scaped heavy water mixe s with ordinary wate r in the air and is e ithe r lost 1/ or is condensed and re cove red.
The recove red wa te r (called "de g raded ") must be " upgraded" ( reconce n t ra ted )
be fore i t can be pu t back into a reac tor.
According to David Ha r t, Nucle a r Powe r in Ind ia (1983) at
- p. 75 ', 60-70 tonne s of degraded heavy wa te r we re collec ted and upgraded each year at RAPP-I. and 20 tonnes per year e scaped
. comple te ly.
A more conse rva tive loss figure of 16 tonnes per year was used in the table above to g ive India the bene fit of the doubt.
1/ The re is the theore tical possibility tha t wa te r actually escaping from RAPP-I was smalle rthe amount of heavy than 16 tonnes pe r yea r, and tha t be tween 16 tonnes andInd ia could have accumula ted the difference the actual loss.
Howe ve r,
same annual loss that Canada first experienced at the Douglas 16 tonnes i Point reactor, the prototype for both MAPP and RAPP, all the experts on theand persistent leaks at RAPP-I have been publicly d The large Indian nuclear program believe losse s are real.
In light of all this, that the se faking losse s. India could have accumulated anything significantit seems that In fact, from large r than 16 tonne s pe r year, andthe available evidenc losse s we re that the Ind ian re serve was probably smaller than the table shows.
-,,, _ _ _ ~ _ _.
According to an article in the _ Time s of Ind ia, July 27, 1983, by Pra ful Bidwai, about 100 tonne s of degraded heavy wa te r we re trans fe r red from Rajasthan ( RAPP-I) to Mad ras (MAPP-I) during the two year s be fore MAPP-I was s tar ted.
In April and May of 1983, for example, a convoy of trucks rushed heavy water to MAPP which had been flown in from RAPP.
The de u te r ium concentra tion of' the degraded wa ter varied from 3 to 30%.
In July of 1982, Mrs. Gandhi told members of the Ind ian pa rliame n t tha t MAPP-I would be started by upgrading a large quantity of " low g rade " he avy wa te r.
The se s ta te me n t s by Ha r t, Bidwai, and Mrs. Gandhi show tha t RAPP-I wa s the source of at le a st par t of MAPP-I's supply.
Mrs.
Gandhi said tha t " low grade " heavy wa te r would be upgraded to start MAPP-I; RAPP-I was the only significant source of degraded heavy wa te r in India; Hart confirms that large amounts of RAPP-I wa te r we re upgraded ; and Bidwai shows tha t the wa te r was shipped to MAPP-I.
How much could be trans fe r red in this way?
It depends upon how degraded the wa te r is wh ich one is upgrading, and how much upgrading one can do.
He a vy wa te r is upgraded by beginning with degraded water containing some percentage of re tained m
7
,..----a
,-.4
de u te r ium. 2/
In o f fect, upgrading consists of ge tting rid of the :
ordinary water with which the deuterium has become mixed.
The wa te r taken to MAPP-I, according to Bidwai, va ried f rom 3 to 30%
de u te r ium (an ave rage of 17%).
Upgrading 100 tonne s of this would produce about 16 tonne s of reactor grade heavy wa te r if the ave rage concentra tion we re 17% and if one subtracts a 10%
proce ssing loss.
This is not enough to cove r a shortage of 67 tonne s.
To cove r the whole shortage, India would have to upgrade enough heavy wa te r to produce 67 tonne s of reactor grade he avy wa te r.
This would mean upgrading about 394 tonnes of degraded wa te r a t a concentra tion of 17%.
To cover the 82 tonne shortage of unsa feguarded heavy wa te r, which she needed for MAPP-I, India would have to upgrade 4 82 tonne s of degraded wa te r at 17%.
Nobody knows whe the r th is ac tually happe ned..
The only hint is Mrs. Gandhi's sta tement in July of 1982, in which she said tha t India got 140 tonnes of reactor grade heavy wa te r for MAPP-I by upgrading " low g rade " wa te r.
RAPP-I was the only source of degraded unsa feguarded heavy wate r in India.
For India to have taken 140 tonne s out of RAPP-I, howe ve r, would have removed 60% of RAPP-I's inventory, and all of its unsafeguarded 2/
The relevant accounting principle for heavy water safeguards is "re tained de ute r ium".
This means that the re ac tor ope ra to.r must put back into safeguards the sane amount of deute rium that he takes out for upgrrding.
If he re move s, for e xample, one hundred tonnes of degraded wa te r which is 49% deute rium and 51%
ordinary water, he must put 49 tonnes of deuterium back into the sa fe gua rded inventory, which was the amount of "re tained de u te r ium" in th.100 tonnes of degraded wate r he took out.
He may put back upgraded heavy water containing 98% deuterium,'which is near reactor grade.
But he will only put back 50 tonnes of it, because the 50 tonnes at 98% will contain the required 49 tonnes of deuterium he re moved for upgrading.
hea vy wa te r, which totalled 141 tonne s a t the e nd o f 1982. 3/
This would amount to cannibalizing all of RAPP-I's unsafeguarded inventory just to start MAPP-I f ree of safegua rds.
The alte rna tive to upgrading was simply to shif t heavy wate r from RAPP-I to MAPP-I d irec tly a t reactor grade, kithout upg rad ing.
India needed a total of 82 tonnes of unsafeguarded heavy wa te r.
The re we re 141 tonne s of unsafeguarded wa te r in RAPP-I, so such a transfe r would have be e n le ga l.
Regardle ss of which of the se alte rna tive s India chose whe the r she shif ted heavy wate r from RAPP to MAPP af te r upg rad ing, or shif ted i,t direc tly by transfers a t reactor grade
-- the re would have been the same prac tical e f fect on RAPP-I.
Each alte rnative would have reduced drastically the amount of heavy wate r in its inventory.
If India had shif ted only the 82 tonne s of unsafeguarded wa te r, this would have le f t RAPP-I 35%
e mp ty, a nd required an 82 tonne shipment back to RAPP-I some time in 1984 in orde r to have RAPP-I ready to start up again in 3/
The amount of unsafeguarded heavy wate r in RAPP-I at any particular time depends upon how losses are re ple n ishe d.
It is logical to assume tha t since 55% of the original inventory was sa fe gua rde d, 55% of the yearly loss would be of safeguarded heavy wa te r.
RAPP-I lost 16 tc.ne s pe r year during the nine years from 1974 through 1982, a total of 144 tonnes.
Fif ty-five pe rcent of that is 79 tonnes.
This reduced the 120 tonne s unde r safeguards (from Canada) in the original inventory to 41 tonnes (120 - 79 =
- 41).
Howe ve r, dur ing the three years from 1980 through 1982,'16 tonne s per year of safeguarded Sovie t heavy wate r we re added to RAPP-I to make up for annual locce s.
This l'
- e a sed the sa fe gua rded inventory in RAPP-I 4 8 tonne. g 3 x 16 = 4 8), to a total of 89 tonnes (41 + 48 = 89).
Tnus, the re we re 89 tonne s of safeguarde; heavy wate r, and 141 tonnes of unsafeguarded heavy wa te r, in RAPP-I a t the end of 1982.
It is assumed that the rest of the safeguarded Sovie t imports rece ived through 1982 we re used in RAPP-II.
Thus, all the heavy wate r in RAPP-II is under safegua rds.
January of 1985.
If India had shif ted 140 tonne s, a s Mr s. Ga nd h i said, the e f fect would have been to leave RAPP-I 60% empty and require a 140 tonne shipment back.
Could this ha ve happe ned, in light of the fact that RAPP-I s tar ted up. a t full powe r in January of 1985?
At the end of 1984 (i.e., Janua ry of 1985), India needed 965 tonnes to run all of her re ac tor s.
But she only had 946 tonne s on hand.
This produced a re se rve shor tage of 19 tonne s, regardle ss of safeguards.
Thus, whe n RAPP-I s ta r ted back up again, with a full inventory in January of 1985, Ind ia was 19 tonne s shor t of having enough heavy wa te r.
So even if India had tr ied to ship wa te r back to RAPP-I some time in 1984 to pre pa re for s ta r t-up, she would not have had enough heavy water to do so.
And, because MAPP-I had to be run on unsafeguarded heavy wa te r, India was short 42 tonne s instead of me rely 19.
When one combine s this shor tage with the transportation and planning problems which shipments back to RAPP-I would have caused, it become s clea r tha t RAPP-I alone could not have provided enough heavy wa te r for MAPP-I.
One could not have run MAPP-I at full powe r through 1984 and at the same time had enough heavy water in 1984 to ship back to RAPP-I to ge t RAPP-I ready for full powe r in i
January of 1985.
India must have used an import.
There is no othe r e xpla na tion.
Whe re did the import come from?
China is the only source of unsafeguarded heavy water in the world.
And China is the only country with a record, and a policy, of making sensitive l
e
,,ic-..,
exports with the expre ss purpose of foste ring nuclear arms proli fe ra tion.
In the same ar ticle in the Time s of Ind ia ci ted above,
Bid wa i s ta te s tha t 100 tonnes of reac tor-grade heavy wate r we re sh ipped f rom Bombay to MAPP-I just be fore MAPP-I started 'up.
The se 100 tonnes are not on Mrs. Gandhi's list, nor on B id wa i 's.
Bidwai repor ts tha t the re we re only 70 tonne s of reactor grade wate r available at MAPP-I in Se p te mbe r o f 19 82, and-only 70 additional tonne s in accumula ted dome stic production i'n the spring of 1983.
MAPP-I s ta r ted in the summe r of 1983.
The se e stima te s are quite close to the amounts reported in the ta ble a bove, and are not too far from Mrs. Gandhi's own figure s.
- Thus, one see s 70 tonne s waiting at MAPP-I (ob ta ine d in part by "upg rad ing " wa te r f rom RAPP), and 70 more accumulated at the various production plants.
A total of 140.
And tha t is all the
-he avy wa te r in Ind ia.
Everything is accounted for.
Nothing more is le f t in the sys te m.
Then suddenly, 100 tonne s of absolutely pure reac tor-grade wa te r appe ars in Bombay.
I t d id no t come from upgrading, or from domestic production, or from the previous impor ts, which we re all keeping reac tors running.
The 100 tonnes simply ente rs, all a t once, from outside the sy s te m.
And in the nick of time to start MAPP-I.
Did it come from China?
No other supplier would have sold it without safeguards, or without reporting it to the Inte rnational Atomic Energy Agency.
The only othe r e xpor te rs of heavy wa te r a re the Uni ted S ta te s, Canada, and the Sovie t Union.
Canada cut of f all nuclear trade with India in re sponse to Ind ia 's " peace ful nuclea r de vice "
l
in 1974.
The Sovie ts are bound by the Non-Proli fe ra tion Trea ty and the ir membe rship in the Nuclear Supplie rs' Group not to e xport heavy wa te r without safeguards, and have adhe red strictly
- to those obliga tions.
The Sovie t re qu i re me n ts on the he avy wa te r for RAPP-II we re the str ic te st e ve r imposed on heavy wate r and we re far more strict than Canada's requirements had been on the wa te r for RAPP-I.
It is inconce iva ble tha t the - U. S., the U.S.S.R., or Canada would have supplied heavy water to India unde r the ta ble.
China is the only source remaining.
Whe n one conside rs this fact in light of China 's record of foste ring nuclea r arms prolife ra tion with sec te t e xper ts of heavy wa te r,A/
the evidence is ve ry strong, strong enough to create an obligation on India eithe r to admit she got a secre t import from China, or explain the source of the myste r ious 100 tonne s.
In 1985, India started Dhruva and MAPP-II without
'sa fe g ua rd s.
Her public shortage was vast:
201 tonne s, almost an e n tire reactor 's wor th of heavy wa te r.
Her unsafeguarded shortage was more vast, 232 tonne s.
If China had supplied the 100 tonne s in 1983, tha t would help in 1985.
But the re would still be 101 tonne s le ss in hand than the amount needed to run all the re ac tor s.
And 132 tonne s le ss in hand than needed to run 3/
Press reports, based on U.S. in te llige nce informa tion, have s ta ted that China has done the following:
discussed with Iran the possibility of providing sensitive nuclear technology; assisted Pakistan in ope ra ting that country's unsafeguarded uranium enrichmen t plant; conducted a nuclear te s t in the presence of a high-level Pakistani of ficial; provided nuclear weapons de sign da ta to Pakistan; provided enriched uranium without cafcgurds LO South Atrica; provided be tween 1980 and 1982 heavy water to Argentina without safeguards; and more heavy wate r to Argentina without safeguards. in 1985 provided
the ne w reactor s f ree of ca fegua rd s.
At this point, Ind ia e i the r got another secre t import from China, or d rs ined wa te r out of RAPP-I.
The re a re no othe r source s.
RAPP-I started up again in January of 1985 with a full inventory.
But it only ope rated until July, when it was shut down again.
It may now be shut down pe rma ne n tly.
Thus, i ts wa te r theore tically would have bee n a va ila ble -- tha t is, it could have bee n dra ined out -- be fore Dhruva and MAPP-II went critical in August of 1985.
Howe ve r,
tha t would mean tha t India would have been planning to bring t'o criticality two reactors for which she had no wate r until just a mon th be fore they we re due to go critical, and was planning to produce tha t wa te r by taking off the grid another reac tor which had just started up in another part of India.
Also, large amounts of heavy wate r would have to be transpor ted almost ove rnigh t.
Th is d oe sn ' t seem likely.
An import seems far more proba ble.
So it appears tha t India got a second import from China in 1985, or the import she rece ived in 1983 was greate r than 100 tonne s.
The total amount of impor ts would have to be 201 tonne s in order to run all of the reactors a t the same time, and would have to be 232 tonne s to start the new reactors without sa fe gua rd s.
One can re ject this as purely circumstantial, and re fuse 'to be lie ve that the Chine se imports happened.
Howe ve r, the price of re fusing to be lie ve in Chine se imports is to belie ve tha t Ind ia got the 201 tonnes from RAPP-I, the only other source.
This would mean draining 90% of RAPP-I's inventory as soon as it shut
down in 1985, transfe rring the wa te r immed ia te ly to MAPP-II and Dhruva (which would have been waiting for it in orde r to achieve the ir planned cr i ticali ty) and breaching safeguards.
By 1985, only 109 tonne s of RAPP-I's heavy wa te r was unsa fe gua rded. 5/
Thus, if India took 201 tonne s out of RAPP-I, she would have been forced to d ive r t illegally 92 tonnes of safeguarded heavy wate r (201-109 = 92).
This would have bee n nece ssary to physically bring MAPP-II and Dhruva to criticality.
To bring them to criticality outside 'sa fegua rds, India needed 232 tonne s, which would have forced her to d ive r t about 123 tonne s of RAPP-I's sa fe guarded he avy wa.te r (232 - 109 =
123).
Which is most probable, secre t imports or illegal d ive r s ion s?
It had to be one or the othe r.
India will st'.ill be publicly short of both safeguarded and unsa feguarded heavy wa te r in 1986 and into the inde fini te future.
Even assuming dome: tic production in 1986 at double the ra te in 1984, the unsafeguarded shortage will be 253 conne s.
India must cover it some h ow.,6/
Will she circumvent sa fe gua rd s,
5/
By the beginning of 1985, RAPP-I contained 121 tonne s of safeguarded heavy water.
This is calculated by starting with the 89 tonnes of safeguarded heavy water it con.tained a t the beginning of 1983, and adding 16 tonnes of safeguarded make up wa te r for each of the years 1984 and 1985 ( 8 9 + 16 + 16 = 121).
The balance of the inven tory, 109 tonne s, we re unsa fe guarded.
5/ It is not possible for a significant quantity of heavy water to have been dive rted slowly over the span of previous years.
When RAPP-I started up in 1973, 55% of its inventory was safeguarded and 45% was unsafeguarded.
Thus, 55% of all the heavy water which leaked out and was collected as " degraded" would be under safeguards, and 45% would not.
The de u te rium content of the safeguarded part of the degraded heavy water was required to be put back into the. safeguarded inventory af te r upgrading.
But the deuterium content of the unsafeguarded part (continued next page)
e 4
or make se c re t imports?
She must continue to do one or the othe r inde finite ly.
Wha t doe s one conclude ?
Ind ia is eithe r importing heavy wate r - secre tly f rom Chine -- and will continue to do so --to keep Dhruva and the MAPP reac tor s f ree of sa feguard s, or India has found a way to circumvent sa feguards, and is using deuterium from sa fe gua rded impor ts in he r unsa feguarded re ac to r s.
To stop this, India 's nuclear supplie rs (the United Sta te s, Canada, and - the Sovie t Union) should suspend furthe r nuclear trade until Ind ia explainn he r public shortages.
They should also suspend f u r the r nuclear trade with China until China stops the expor ts or o ive s convincing assurances that they we re not made.
If the supplie r s take no action, it will mean tha t India can officially thumb he r nose at the world, and tha t China, through secre t expor ts, can con tinue to foste r nuclea r arms prolife ra tion.
did not have to be accounted for.
Howe ve r,
deuterium we re not put back into the reactor, or replacedif the unsafeguarded in the reactor by othe r deute rium f rom the heavy water re serve, the reactor would not be safe to ope ra te.
The wa te r le ve l would be too low.
RAPP-I operated continuously until early 1982.
The re fore, one can assume that the unsafeguarded deute rium which leaked out and was collected as degraded was put back into the reactor or replaced from the unsafeguarded have deple ted tha t rese rve by an amount equivalent to the amount re se rve, which would of unsafeguarded water collected.
Thus, India could not have accumulated any additional unsafeguarded heavy water over the years by upgrading.
.