Page 1
642
CHAPTER 22
Behavior of Radionuclides
in the Environment
Contents
22.1.
Radioactive releases and possible effects
643
22.2.
Radionuclides of environmental concern
644
22.3.
Releases from the Chernobyl accident
646
22.4.
The injection of TRU in the environment
648
22.5.
Present levels of TRU in the ecosphere
649
22.6.
Actinide chemistry in the ecosphere
651
22.6.1.
Redox properties
651
22.6.2.
Hydrolysis
653
22.6.3.
Solubilities
655
22.7.
Speciation calculations
656
22.7.1.
Calculated species in solution
659
22.8.
Natural analogues
661
22.9.
The Oklo reactor
662
22.10.
Performance assessments of waste repositories
663
22.10.1. Release scenarios
664
22.10.2. Canister dissolution
665
22.10.3. Releases from bitumen and concrete encapsulations
667
22.10.4. Migration from the repository
667
22.11.
Conclusions
672
22.12.
Exercises
672
22.13.
Literature
673
The main objection against nuclear power is the risk of spread of "radioactivity" (radioactive
elements) to the environment where it may cause health effects in humans. We have already
discussed such effects (Ch. 18). Here, we are concerned with the chemical aspects of the
sources of releases and of the migration of the radionuclides in the environment. Their chemical
properties, together with hydrology, determine how fast they will move from their point of
entry into the groundwater to water resources used by man; this is schematically illustrated in
Figure 22.1. In particular we discuss actinide behavior as these elements have the most
hazardous radionuclides which may be released in the different steps of the nuclear fuel cycle,
and, especially, from nuclear waste repositories.
Page 2
Behavior of Radionuclides in the Environment
643
FIG. 22.1. Migration path of radioactive nuclides from a waste repository to man.
22.1. Radioactive releases and possible effects
In earlier chapters, there have been brief discussions of the release of radionuclides to the
environment. Such releases occur from mining and milling operations, particularly of uranium
ores (§5.5.4), from the nuclear fuel fabrication processes, from normal operation of nuclear
reactors (§19.16),from reprocessing of spent nuclear fuel (§21.8), from nuclear weapons
production and recovery, from transportation of nuclear material, from testing of nuclear
weapons and accidents (§5.10), and from storage of nuclear wastes (which we discuss in
subsequent sections). Table 22.1 compares the estimated collective dose to the public from the
various activities of the nuclear fuel cycle, from mining and milling through nuclear waste
disposal; the significant point is the recognition that the mining operation for uranium is the
major contributor in the dose to the global public by quite a large fraction.
An estimate of the total radioactivity, by nuclide, released into the atmosphere by above-
ground nuclear tests is given in Table 22.2; the total release is 2 600 EBq. To provide
comparison, the estimated global releases from reactors and reprocessing plants are listed in
Table 22.3; globally this gives 3.9 EBq from reactors and 3.4 EBq from reprocessing plants
up to 1998. The major releases are the noble gases, H and C. The total C release over 30
3
14
14
years is less than one percent of the normal C level from cosmic ray
14
Page 3
Radiochemistry and Nuclear Chemistry
644
Source
Normalized collective effective dose (manSv/GW y)
e
___________________________________________________________________________________
Local and regional effects
Mining
0.19
Milling
0.008
Mine and mill tailings (releases over 5 years)0.04
Fuel fabrication
0.003
Reactor operation
Atmospheric
0.4
Aquatic
0.04
Reprocessing
Atmospheric
0.04
Aquatic
0.09
Transportation
<0.1
Solid waste disposal and global effects
Mine and mill tailings
Releases of radon over 10 000 y 7.5
Reactor operation
Low-level waste disposal
0.00005
Intermediate-level waste disposal0.5
Reprocessing solid waste disposal
0.05
Globally dispersed radionuclides
40
(truncated to 10 000 years)
TABLE 22.1. Collective effective dose to the public from radionuclides released in effluents from the nuclear
fuel cycle 1995-97 (UNSCEAR 2000)
production (Section 5.1.3). Thus, this is a small fraction of the radioactivity released in nuclear
weapons testing. To further place the releases in proper relation, an estimate of the total
atmospheric release in the Chernobyl accident was 2 EBq (Table 22.4), and in the Three Mile
Island Reactor (TMI) accident, 10 EBq. From such figures the United Nations Scientific
-6
Committee on the Effects of Atomic Radiation (UNSCEAR) estimates that the average annual
dose of radiation per person during year 2000 corresponded to:
Natural background
2.4 mSv
Medical diagnostics
0.4 mSv
Nuclear weapons tests
0.005 mSv
Chernobyl accident
0.002 mSv
Nuclear fuel cycle
0.0002 mSv
22.2. Radionuclides of environmental concern
Most of the radionuclides produced in nuclear tests, accidents and in the normal fuel cycle are
short lived. In Table 21.3, longer lived fission products and activation products from these
systems are listed as these represent the major concern to the general public if they are allowed
to enter the environment; we exclude nuclides with insignificant contribution
Page 4
Behavior of Radionuclides in the Environment
645
Radio-
Half-
Estimated release
Radio-
Half
Estimated release
nuclide
life
Total (EBq)
nuclide
life
Total (EBq)
___________________________________________________________________________________________________________________
H
12.33 y
186
Sb
2.73 y
0.741
3
125
C
5730. y
0.213
I
8.02 d
675
14 †
131
Mn
312.3 d
3.98
Cs
30.07 y
0.948
54
137
Fe
2.73 y
1.53
Ba
12.75 d
759
55
140
Sr
50.53 d
117
Ce
32.50 d
263
89
141
Sr
28.78 y
0.622
Ce
284.9 d
30.7
90
144
Y
58.51 d
120
Pu
24110. y
0.00652
91
239
Zr
64.02 d
148
Pu
6560. y
0.00435
95
240
Ru
39.26 d
247
Pu
14.36 y
0.142
103
Ru
1.023 y
12.2
106
___________________________________________________________________________________________________________________
For the non-gaseous fission products a total non-local fission explosion yield of 160.5 Mt, obtained from measured
Sr deposition, was assumed in deriving the total amounts released. For simplicity, all C is assumed to be due
90
†
14
to fusion.
TABLE 22.2. Radionuclides released in atmospheric nuclear testing (UNSCEAR2000)
Radionuclides
Reactor releases (PBq)
Reprocessing releases (PBq)
_______________________________________________________________________________________________
Noble gases
3631
3190
H
269
144
3
C
1.97
0.44
14
Sr
!
6.6
90
Ru
!
19
106
I
!
0.014
129
I
0.046
0.004
131
Cs
!
40
137
Particulates
0.121
!
Others
0.839
!
_______________________________________________________________________________________________
Total39003400
TABLE 22.3. Global release of radionuclides by reactors and reprocessing plants up to 1998 (UNSCEAR2000)
to the total activity after 10 years. In addition, several heavy radionuclides are formed by
neutron capture reactions such as U, Np,
Pu, Am, etc.
236
237
238-242
241
To assess the potential for these radionuclides to cause harm to humans, their geochemical and
biological behavior must be evaluated. For example, since Kr is a chemically inert gaseous
element, it would have little effect on a person who inhaled and immediately exhaled the small
amount which might be present in the air. By contrast, other nuclides with high activity,
Sr- Y and Cs-
Ba, have active geological and biological behavior and can present much
90
90
137
137m
more significant radiation concerns to humans. In normal operations, these nuclides would of
course be released in quite insignificant amounts, as represented by the value for aquatic
releases in reactor operation and reprocessing, Table
Page 5
Radiochemistry and Nuclear Chemistry
646
FIG. 22.2. Deposition of Cs in Scandinavia from the Chernobyl accident.
137
19.7 and 21.10. Similarly, the heavy elements (Np, Pu, Am) would not be released in normal
operation at levels that would be of concern. We have pointed out that some low level releases
as a result of normal operations are allowed by the health authorities, who also monitor these
levels. In the effluents from reprocessing plants (e.g. Sellafield in the United Kingdom and La
Hague in France), the relatively long-lived nuclides such as H, C, Kr, Tc and I are
3
14
85
99
129
of major concern. The liquid effluents from nuclear power plants and from reprocessing plants
are about equally responsible for the global collective dose commitment of nuclear power
generation (i.e., 0.8 man Sv per GW y of the total 2.5 man Sv).
e
22.3. Releases from the Chernobyl accident
On April 26, 1986, a low power engineering test was being conducted at one of the reactors
of the Chernobyl nuclear power station in the Ukraine (then the USSR). The reactor became
unstable, resulting in thermal explosions and fires that caused severe damage to the reactor and
its building (§§5.10.2 and 20.1.2c). Radioactivity was released over the next ten days until the
fires were extinguished and the reactor entombed in concrete. The radioactivity was released
as gas and dust particles and initially blown by winds in a northerly direction. Outside Russia,
the accident was first detected by increased
Page 6
Behavior of Radionuclides in the Environment
647
Release
Core
Contamination
Nuclide
Half-life
inventory
air (Bq m )
ground (kBq m )
!3
!2
tot EBq
%
tot EBq
Stockholm
Gävle area
†
‡
_______________________________________________________________________________________________________________________
Kr
10.73 y
0.033
.100
0.033
85
Sr
50.5 d
0.094
4.0
2.4
89
Sr
28.6 y
0.0081
4.0
0.20
90
Zr
64.0 d
0.16
3.2
5.0
0.6
5.9
95
Ru
39.4 d
0.14
2.9
4.8
6.4
14.6
103
Ru
368 d
0.059
2.9
2.0
1.8
4.0
106
I
8.04 d
0.67
20
3.3
15
179
131
Xe
5.24 d
1.7
.100
1.7
133
Cs
2.07 y
0.019
10
0.19
2.4
14.4
134
Cs
13.2 d
0.7
6.0
136
Cs
30.2 y
0.037
13
0.28
4.5
24.7
137
Ba
12.8 d
0.28
5.6
5.0
25.6
2.1
140
Ce
32.5 d
0.13
2.3
5.6
0.5
5.9
141
Ce
284 d
0.088
2.8
3.1
0.3
3.7
144
Np
2.36 d
0.97
3.2
30
2.7
239
Pu
87.7 y
3.0×10
3
0.001
238
!5
Pu
24100 y
2.6×10
3
0.0009
239
!5
Pu
6570 y
3.7×10
3
0.0012
240
!5
Pu
14.4 y
0.170
241
_______________________________________________________________________________________________________________________
First days, April 28-29, 1986. Ullbolsta, outside of Gävle, among highest depositions outside Russia;
†
‡
corrected to April 28, 1986. References: USSR State Comm. on the Utilization of Atomic Energy 1986. The
accident at the Chernobyl nuclear power plant and its consequences, IAEA expert meeting 25-29 Aug. 1986,
Vienna.
TABLE 22.4. Radionuclides released into the atmosphere in the Chernobyl accident and local contamination
Nuclide
"Hot Particle" (%) Reactor fuel
________________________________________________________
Zr
17.9
17.6
95
Nb
20.7
19.3
95
Ru
14.2
15.0
103
Ru(Rh)
3.5
2.4
106
Ba(La)
12.6
11.1
140
Ce
15.8
14.3
141
Ce(Pr)
15.8
14.4
144
TABLE 22.5 Nuclide composition of a "hot particle" from Chernobyl compared to reactor fuel after 3 years of burning
radioactivity levels at the Forsmark nuclear power plant, about 110 km north of Stockholm,
Sweden, where it caused a full alarm as the radioactivity was believed to come from the
Swedish plant. Subsequently, the radioactivity released at Chernobyl was spread more to the
west and southwest (Figure 5.8).
Page 7
Radiochemistry and Nuclear Chemistry
648
For the exposed population in the Byelorussia region near Chernobyl the estimated average
increased dose in the first year after the accident was approximately the same as the annual
background radiation. In northern and eastern Europe in general, the increased exposure during
the first year was 25-75% above background levels. The highest dose will be delivered in
southeastern Europe and is estimated to be 1.2 mSv up to year 2020, which can be compared
to 70 mSv from natural background radiation during the same period. Figure 22.2 shows the
levels of Cs deposited in Scandinavia in the first days after the accident. Table 22.4 gives
137
the fraction of the core activity released and the air and ground contamination of various
nuclides at two Swedish locations.
It was shown that the larger airborne particulates from the Chernobyl accident had a
composition which was quite similar to that of the reactor fuel. A comparison of the
composition of these "hot particles" with that of the reactor fuel is given in Table 22.5. About
a tenth of these hot particles had a high concentration of Ru and Ru while others were
103
106
depleted in the ruthenium fission products. The Ru-rich particles may have originated from a
part of the reactor where burning graphite produced CO which reduced the ruthenium to
non-volatile metallic Ru. In other sections, oxidation occurred, forming volatile RuO and/or
3
RuO which vaporized from the particles.
4
In contrast to the larger particles, the composition of the smaller ones varied considerably and
they were distributed over much greater distances. Other measurements reflect this variability
in the fall-out from Chernobyl. For example, 70% of the Cs from Chernobyl measured in
137
Great Britain was water soluble. By contrast, the Cs measured in Prague, much closer to the
137
accident site, was only 30% water soluble. Further insight into the variety of species present
in the Chernobyl dust is found in data on the deposition of some radionuclides during periods
of rain, and dry weather. Rainy periods accounted for 70-80% of the total deposition of Cs,
134
Cs,
Ru, Ru, and Te while deposition during dry weather was more important for
137
103
106
132
I.
131
These observations indicate that the speciation of radionuclides in the atmosphere is dependent
on their source, their mechanisms of production and the nature of the particular environment.
While some species are gaseous, others are associated to particles with properties and
suspension times that are strongly dependent on the particle size and density.
22.4. Injection of TRU into the environment
Of the artificial radionuclides released to the environment by nuclear activities, the
transuranium (TRU) species are a major concern. This concern arises from the very long half-
life of a number of the nuclides as well as their high radiotoxicity values. Although reactor
operation and spent fuel reprocessing activities have released small amounts of TRU's to the
environment, testing of nuclear weapons has released rather large quantities. Since the first
nuclear test detonation in New Mexico in 1945, approximately 3 500 kg of plutonium has been
released in atmospheric explosions and another 100 kg in underground tests. This corresponds
to about 11 PBq of
Pu ejected into the atmosphere. In addition 0.6 PBq of Pu were
239+240
238
released over the south Pacific in the high altitude destruction of the SNAP-9 satellite power
source in 1964. By contrast, a total of 0.58 PBq of
Pu has been released into the Irish
239+240
Sea from the Sellafield (UK) reprocessing plants between 1971 and 1999; most of this before
1985. About 37 kg of Am is present
241
Page 8
Behavior of Radionuclides in the Environment
649
Nuclide
Tests (TBq)
Sellafield, (TBq)
LaHague (TBq)
1945-1980 1971-84 1985-94 1995-99
2000
___________________________________________________________________________________
Np
-
-
0
0.33
237
Pu
6 520
239
Pu
4 350
? 559
? 15
? 0.92
240
Pu
142 000
-
-
21.8
0.039
241
†
Am
-
442
9
0.31
241
Cm
-
-
-
0.052
242
Cm-
-
-
0.024
243+244
___________________________________________________________________________________
Total "
†
TABLE 22.6. TRU released from nuclear tests and from reprocessing (UNSCEAR2000, BNFL and COGEMA)
in the environment from the decay of Pu from the nuclear testing. Table 22.6 compares the
241
TRU's released in nuclear tests in the atmosphere, in the Irish Sea from the Sellafield plants,
and into the English Channel from the LaHague plants. As the amount of spent nuclear fuel
increases, the contribution to the total plutonium in the environment could become more
significant over a longer time, especially if nuclear waste disposal sites release actinide elements
slowly to the environment. Whatever the sources of plutonium and other actinides, their
presence represents a contamination of the environment by highly toxic material. An under-
standing of the factors involved in their retention and/or migration in the ecosphere is therefore
highly desirable. Studies of the environmental behavior of releases from tests provide data
needed to understand and predict the behavior of smaller releases from the nuclear power
industry. Therefore, in the next few paragraphs we concentrate our discussion to the behavior
of the actinide elements in the environment, particulary the mobility of plutonium.
The majority of the plutonium from weapons testing was injected initially into the
stratosphere. The plutonium originally in the weapon which survived the explosion would have
been formed into high-fired oxide which would be expected to remain insoluble as it returned
to earth. Such insoluble particles would have sunk in a rather short time into the bottom
sediments of lakes, rivers, and oceans or would become incorporated in soils below the surface
layer. However, in most nuclear weapon explosions a considerable amount of plutonium is
generated in the explosion via U (n,() reactions and subsequent "-decay of the product U,
238
239
U, U, etc. In total, about two thirds of the plutonium released was generated in this way.
240
241
The nuclides from the (n,() reactions would exist as single atoms, and, hence, were never
formed into high-fired oxides. The plutonium from this formation path would have been soluble
and, as a result, more reactive and its behavior would be more similar to that of plutonium
released from nuclear reactors, reprocessing plants and from nuclear waste repository sites.
22.5. Present levels of TRU in the ecosphere
The United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR),
in 2000 reports data that corresponds to the following total average global
Page 9
Radiochemistry and Nuclear Chemistry
650
Water
Concentration of Pu (M)
_____________________________________________________________________
Lake Michigan
2.0×10
!17
Great Slave Lake, Canada
1.5×10
!17
Okefenokee River, Florida
1.5×10
!16
Hudson River, New York
1.0×10
!17
Irish Sea:
1 km from Windscale
1.6×10
!14
110 km from Windscale
1.1×10
!15
Mediterranean
2.6×10
!18
North Pacific (surface)
3.0×10
!17
South Pacific (surface)
1.0×10
!17
_____________________________________________________________________
Samples were passed through 0.45 µm filters.
TABLE 22.7. Concentration of plutonium in filtered samples of natural waters
nuclide depositions from atmospheric nuclear weapon tests: Sr 11, Sr 1.2, Zr 19, Ru
89
90
95
106
12, Cs 1.8, Pu 0.013, Pu 0.009, and Pu 0.278 kBq m . Some of these radionuclides
137
239
240
241
!2
have now decayed because of their short half-lives and the end of atmospheric testing in 1980
and almost all of the Pu has decayed to Am.
241
241
Near test sites, reprocessing facilities, etc., the concentration of plutonium in the soil and
water is much higher than in more distant locations. Generally, the great majority of plutonium
is associated with sub-surface soils or sediments or with suspended particulates in water. For
example, when vegetation, animals, litter and soils are compared, $99% of the plutonium is
present in the soil. Similarly, in shallow bodies of water, more than 96% of the plutonium is
found associated with the sediments. However, it is via the species that are soluble or attached
to suspended colloids and/or particulate matter in water that plutonium is transported in the
environment. Analysis of vertical plutonium migration in soils near Chernobyl and in eastern
Europe from the Chernobyl accident has shown that most of the plutonium is still in the first
0.5 cm from the surface for soils with significant humic acid content. In these soils, the
plutonium is mostly associated with the insoluble calcium-humate fraction. In non-humic,
carbonate rich soils, the plutonium has moved several centimeters downward. Migration rates
of #0.1 cm y is associated with the humic soils and of 1-10 cm y with the carbonate rich
!1
!1
ones. Presumably, migration is retarded by the interaction with the immobilized humic material
in soils.
In subsurface oxic soil near Los Alamos National Laboratory, USA, plutonium is relatively
mobile and has been transported primarily by colloids in the 25-450 µm size range. Moreover,
the association with these colloids is strong and removal of Pu from them is very slow. By
contrast, near Sellafield in wet anoxic soil, most of the Pu is quickly immobilized in the
sediments although a small fraction remain mobile. Differences in oxidation state (Pu(V) vs.
Pu(IV)) as well as in humic content of the soils may explain these differences in mobility.
Table 22.7 lists the concentration of plutonium, after filtration (0.45 µm), in the surface layers
of some natural waters. The higher concentration in the Okefenokee River is assumed to reflect
the effect of complexing by humic materials. This agrees with the observation that
Page 10
Behavior of Radionuclides in the Environment
651
State
Pu(III+IV)
Pu(V)
Pu(VI)
__________________________________________________________________________________________________
Rain
34
66
Mediterranean
42
58
Irish Sea
23
77
0
Pacific (I)
39
52
9
Pacific (II)
40
46
14
Lake Michigan
13
87
0
TABLE 22.8. Plutonium oxidation states (in % of total Pu) in natural waters
adding humic material to seawater samples containing plutonium increases the solubility by
more than a factor of five over a period of one month.
It is difficult to obtain reliable values of plutonium concentration in natural aquatic systems
as it is very low, approximately 0.001 dpm per liter sea water. Moreover, the plutonium
associated with suspended particles may be more than an order of magnitude greater than that
in true solution. In tests of water from the Mediterranean Sea, filtration (0.45 µm) reduced the
concentration of plutonium by a factor of 25. In laboratory tests with filtered seawater to which
plutonium was added, after one month the total concentration of Pu was 1.3×10
M, but only
-11
40% (5×10 M) was in solution as ionic species and the other 60% was probably in colloidal
-12
form. The mean residence time of Pu in the water column is proportional to the concentration
of particulate matter. As a consequence, > 90% of the Pu is rapidly removed from coastal
waters whereas, in mid-ocean waters where the particulate concentrations are lower, the
residence time for Pu is much longer.
22.6. Actinide chemistry in the ecosphere
22.6.1. Redox properties
Before proceeding to more detailed discussion of the behavior of actinides in the environment,
it is useful to review some of their chemical properties. A general discussion of the actinide
solution chemistry is given in §16.3; here, the focus is on their behavior in aqueous solutions
and primarily in solutions of pH 5 - 9 which is the pH range of natural waters (e.g. the oceans
have pH = 8.2). The actinide ions have an unusually broad range of oxidation states in aqueous
solution, from II to VII; The II and VII States are not discussed further as they do not form in
ecosystems. Following the normal pattern for polyvalent cations, lower oxidation states are
stabilized by more acidic conditions while higher oxidation states are more stable in basic
solutions. Of course, this generalization can be negated by other factors, such as complexing,
which may cause a reversal of the relative stability of different oxidation states. The greater
strength of complexing of An(IV) cations relative to that of An(III) can significantly increase
the apparent redox stability of the An(IV) species compared to An(III). The greater tendency
to hydrolysis of Pu(IV) causes Pu(III), which is stable in acid solution, to be oxidized to Pu(IV)
in neutral media. The disproportionation of Pu(V) is discussed in Chapter 16 where it is pointed
out that in the
Page 11
Radiochemistry and Nuclear Chemistry
652
FIG. 22.3. Redox potential diagrams of U, Np and Pu; the reduction potentials as listed are for pH values: pH=0;
pH=8; pH=14.
higher pH and very low concentrations of Pu in natural waters, disproportionation is not a factor
in the redox behavior of plutonium.
The actinide elements in a particular oxidation state (e.g. Th(IV), U(IV), Pu(IV), Np(IV),
Am(IV)) have similar behavior. However, their redox behavior is quite different, as mentioned
in Chapter 16. The pH affects this redox behavior significantly as reflected in Figure 22.3
which compares the redox potentials of U, Np and Pu at pH = 0, 8 and 14.
Am (III) is the most stable oxidation state in aqueous solutions while Pu(III) and Np(III) are
present under reducing conditions (e.g. anoxic waters). Th(IV) is the common and stable state
for that element. U(IV) and Np(IV) do not react with water but are oxidized by O in oxic
2
systems. Pu(IV) is stable at low concentrations in acidic solution, but Pu(OH) has a very low
4
solubility product.
NpO
is stable except at high acidities and high concentrations under which conditions it
2
+
disproportionates. UO and PuO
increase in stability as the pH is increased. U, Np and Pu
2
2
+
+
form AnO
ions in solution with the stability decreasing in the order U > Pu > Np. UO
2
2
2+
2+
is the most stable uranium species in natural waters.
Redox reactions can be induced in the actinide ions by secondary effects which can be
significant in natural waters. In the presence of higher levels of radiation (e.g. "-emission from
the actinides or $,( from fission products), the radiolytic products, such as the free radicals,
peroxide, etc., would also induce redox changes.
Page 12
Behavior of Radionuclides in the Environment
653
FIG. 22.4. Eh-pH (Pourbaix) diagram showing stability areas for Pu(III), Pu(IV), Pu(V) and Pu(VI); at the division line
there is an equal concentration of the two oxidation states.
Redox properties are often described by the aid of potential-pH (or Pourbaix) diagrams,
Figure 22.4. The shaded area represents typical groundwaters in granitic rock, containing iron
minerals; such waters are usually reducing, i.e. have Eh-values below 0. Natural groundwaters
(including oceans, lakes, rivers, etc) fall within the area enclosed by the dashed curve; they can
have rather high Eh values due to atmospheric oxygen, and also be rather alkaline in contact
with carbonate rocks. The sloping lines follow the Nernst equation (defined by eqn. (9.4)). We
discuss this Figure further in §22.7.1.
22.6.2. Hydrolysis
Hydrolysis is an important factor in actinide behavior in natural waters as the pH is high
enough to result in such reactions as:
Page 13
Radiochemistry and Nuclear Chemistry
654
FIG. 22.5. Concentration of free plutonium ions at different oxidation states in solutions of different pH, showing the
effect of hydrolysis.
An
+ m H O = An(OH)
+ mH
(22.1)
+n
+(n-m)
+
2
m
The order of increasing pH for onset of hydrolysis follows the sequence:
An
> AnO
> An
> AnO
(22.2)
4+
2+
3+
+
2
2
The variation of the concentration of the free (non- hydrolyzed) cations with pH is shown for
the oxidation states of III to VI of Pu in Figure 22.5. These curves are based on estimated
values of the hydrolysis constants, but are of sufficient accuracy to indicate the pH values at
which hydrolysis becomes significant (e.g. 6-8 for Pu , #0 for Pu , 9-10 for PuO
and
3+
4+
+
2
4-5 for PuO
.
2
2+
The study of plutonium hydrolysis is complicated by the formation of oligomers and polymers
once the simple mononuclear hydrolytic species start forming. The relative mono/oligomer
concentrations are dependent on the plutonium concentration: e.g. the ratio of Pu present as
(PuO ) (OH)
to that as PuO (OH) is 200 for total [Pu] 0.1 M, decreases to 5.6 for 10
2 2
2
2
T
2+
+
-4
M and is only 0.05 for 10 M. The hydrolysis of Pu
can result in the formation of polymers
-8
4+
which are very difficult to convert back to simpler species. Generally, such polymerization
requires [Pu] > 10 M. However, the irreversibility of polymer formation prevents the
T
-6
destruction of the polymers by dilution of more con-
Page 14
Behavior of Radionuclides in the Environment
655
centrated hydrolysed solutions to concentrations below 10 M. Soon after formation, such poly-
-6
mers in solution can be decomposed to simpler species by acidification or by oxidation to
Pu(VI). However, as the polymers age, the depolymerization process requires an increasingly
vigorous treatment. A reasonable model of the aging involves initial formation of aggregates
with hydroxo bridging with conversion over time to structures with oxygen bridging:
The relative percentage of oxygen bridges presumably determines the relative inertness of the
polymers. The polymers apparently increase in aggregate size as the pH increases. At pH 4,
the polymers are small enough that essentially all of the Pu remains suspended in solution after
a week while at pH 5 less than 10% remains in solution and at pH 6, only 0.1% remains.
22.6.3. Solubilities
In marine and natural waters, the limiting solubility is usually associated with either the
carbonate or the hydroxide depending on the oxidation state, pH and carbonate concentration.
For example for Am , the reported value of the solubility product (logK , eqn. 9.21) is !26.6
3+
s0
for crystalline Am(OH) (c) at very low ionic strength and -22.6 for Am(OH)(CO )(c). At pH
3
3
6, if [CO ]
> 10 M and at pH 8, if [CO ]
> 10 M , the solubility of Am
would
3 free
3
free
2-
-12
2-
-8
+3
be expected to be limited by the formation of Am(OH)(CO ).
3
Plutonium solubility in marine and natural waters is limited by the formation of Pu(OH) (am)
4
(for amorphous) or PuO (c) (for crystalline). The K of these species is difficult to measure,
2
s0
in part due to the problems of the polymer formation. A measured value for Pu(OH) (am) is
4
log K = -56. This value puts a limit on the amount of plutonium present, even if Pu(V) or
s0
Pu(IV) are the more stable states in the solution phase. Moreover, hydrolyzed Pu(IV) sorbs on
colloidal and suspended material, both inorganic and biological.
The strong preference for neptunium to form the NpO , which has relatively weak
2
+
complexing and hydrolysis tendency, lead to solubilities as large as 10 M under many geo-
-4
chemical conditions. However, the reducing environment found at large depths in some granites
would make Np(IV) the dominant oxidation state. As with plutonium, the solubility of
neptunium in all oxidation states seems to be limited by the low solubility of Np(OH) (am) or
4
NpO (c).
2
Page 15
Radiochemistry and Nuclear Chemistry
656
Species
log$
Species
log$
n
n
_______________________________________________________________________________________
Am(OH)
6.44
Am(CO )
5.08
2+
+
3
Am(OH)
13.80
Am(CO )
9.27
2
3 2
+
-
Am(OH)
17.86
Am(CO )
12.12
3
3 3
3-
TABLE 22.9. Equilibrium Constants for Am(III)
In oxic waters, uranium is present as U(VI) and strongly complexes with carbonate; e.g. in
sea water, the uranium is present at 10 M concentration as UO (CO ) . The solubility of
-8
4-
2
3 3
uranium in some waters may be limited by the formation of an uranyl silicate species.
22.7. Speciation calculations
An essential step in the safety analysis of potential waste repositories is the prediction of what
chemical species are formed in the actual water. For example, the relatively high solubility of
uranium in sea water is due to this strong carbonate complexation which forms UO (CO ) .
2
3 3
4+
Figure 22.6 shows the variation of uranyl species in a surface water under normal atmospheric
pressure of CO (p3.2×10 ; log[CO ]=2pH!18.1+logp
). These speciation
2
CO2
3
CO
!4
2!
2
diagrams are calculated from the equilibrium constant for formation of each species plus mass
balance equations. In this section we describe the use of equilibrium constants in modeling the
speciation in a natural water.
Assume the reactions M + nX = MX where X is OH , and CO and n = 1 to 3. The
n
3
!
2!
equilibrium constants, expressed as $ are given by (cf. eqn. 9.22):
$ = [MX ]/[M] [X]
(22.3)
n
n
n
The equilibrium constants for Am(III) are listed in Table 22.9 for hydrolysis and CO
3
2-
complexation. Let us consider the case of Am(III) should it be released into the environment.
The first step in modeling the speciation of Am(III) is to rearrange the above equation to
express the ratio of complexed to free metal, e.g. (omitting ionic charges)
[AmX]/[Am] = $ [X]
(22.4)
1
[AmX ]/[Am] = $ [X]
(22.5)
2
2
2
and so forth. The mass balance equation is:
[Am] = [Am] + [AmX] + [AmX ] + …
(22.6)
T
2
where [Am] is the total analytical concentration of americium. Dividing by [Am] gives:
T
[Am] /[Am] = [Am]/[Am] +[AmX]/[Am] + [AmX ]/[Am] + …
(22.7)
T
2
Page 16
Behavior of Radionuclides in the Environment
657
FIG. 22.6. Fraction of uranium species in water with natural carbon dioxide content and at different pH, showing
hydrolysis and CO
complexation.
3
2!
FIG. 22.7. Results of speciation calculations for Am(III) in natural water.
Page 17
Radiochemistry and Nuclear Chemistry
658
FIG. 22.8. Results of speciation calculations for Am(III) in natural water, taking into account the low solubility of
Am(OH) .
3
[Am] /[Am] = 1 + $ [X] + $ [X] + …
(22.8)
T
1
2
2
and
" = $ [X] ([Am]/[Am] )
(22.9)
n
n
T
n
where " is the fraction of all americium in the form of AmX . With these equations, $ 's, and
n
n
n
a given value of [Am] , we can calculate the concentration of each species for any value of [X].
T
In Figure 22.7 the speciation of Am(III) as hydroxide and carbonate complexes is shown as a
function of pH. The concentration of CO
is based on the atmospheric partial pressure
3
2-
(3.2×10 atm) and the resultant concentration of CO
at different pH values in a natural
-4
2-
3
water.
If the solubility is limited by a solid phase for which the K is known, the actual value to be
s0
expected for each species can be calculated by assuming no over-saturation. These can be
included in the mass balance equation to predict the total minimum solubility. If log K = -
s0
28.9 is assumed for Am(OH) , Figure 22.8 shows the results of such a calculation for the
3
solubility of americium in the same system as in Figure 22.7, but now including the effect of
the limited solubility of Am(OH) .
3
We have stated that Pu(V), as PuO , is the dominant dissolved species while Pu(OH) is the
2
4
+
solubility limiting precipitate. The redox reaction can be written as:
PuO
+ 4 H + e = Pu
+ 2 H O
(22.10)
2
2
+
+
-
4+
Page 18
Behavior of Radionuclides in the Environment
659
The EE value for the IV/V pair is 1.17 V. Using the equilibrium expression with the Nernst
equation, assuming the redox potential in a fresh water lake to be 0.4 V, we obtain at pH 7:
[PuO ]/[Pu ] . 10
2
+
4+
15
From the value of log K = -56 for Pu(OH) (am) we can calculate that [Pu ] in a solution
s0
4
4+
of pH = 7 in contact with solid Pu(OH) is 10
M. This gives us a value of 10 ×10
.
4
-28
15
-28
10
M for the expected concentration of PuO
in this solution.
-13
+
2
A number of geochemical modeling codes have been developed which use such speciation and
solubility equilibrium equations to calculate the concentration of different species of a metal ion
as well as its net solubility in various waters (common codes are PHREEQE and EQ 3/6). The
results from such calculations are as good as the equilibrium constants or the thermodynamic
values used in the calculations. Also, most important, the calculations must include the
equilibrium equations for all species in solution which contribute significantly to the solution
phase concentrations and all solids which can provide the limiting solubility to the solution
species. Furthermore, the degree of oversaturation with regard to each solid phase must be
prescribed in order to calculate a realistic speciation. These modeling codes are presently based
on the assumption that the natural systems are all at equilibrium whereas in nature this may not
be true. Many systems are kinetically controlled and are often in a steady state, but not in true
equilibrium. In these cases, perhaps the majority of the systems, the equilibrium modelling
codes cannot accurately describe the actual conditions, but may provide a set of limiting
species, approximate relative concentrations and baseline net solubilities. A further complication
arises in assessing the role of colloids, and of sorption which may reduce the concentration of
soluble species below that estimated for the least soluble solid phase. On the other hand,
sorption on suspended colloids may also increase the total concentration (dissolved plus amount
in colloids, e.g. Pu in sea water §22.5). In general, the equilibrium code calculations can easily
give lower limit values of maximum solubilities by assuming no oversaturation. Such
calculations are valuable in waste management risk assessment since, if the lower limit
solubilities from the equilibrium calculations fall well below the accepted safety limits even
when assuming reasonable degrees of oversaturation, it is very likely that the actual total
concentrations will also be below the acceptable limits, cf. Fig.22.10, Table 22.10 and §22.10.
22.7.1. Calculated species in solution
The diversity of reactions which actinides can undergo in natural waters is presented
schematically in Figure 22.9. Complexation by anions such as hydroxide, carbonate, phosphate,
humates, etc. determine the species in solution. Sorption to colloids and suspended material
increases the actinide concentration in the water while precipitation of hydroxides, phosphates,
carbonates, and/or sorption to mineral and biological material limit the amount in the solution
phase.
In natural oxic waters, americium is present in the trivalent and thorium in the tetravalent state
while uranium is hexavalent, UO
. The total concentrations of uranium and thorium in
2
2+
surface sea water are 1.1-1.5×10 and 2.5×10
M, respectively in both the Atlantic
-8
-12
Page 19
Radiochemistry and Nuclear Chemistry
660
FIG. 22.9. Speciation diagram for the range of reactions to be considered in studying the environmental behavior of
plutonium.
and Pacific oceans. The amount of uranium associated with particulate matter in the water is
small. By contrast, for tetravalent Th (originating from thorium minerals), about 50% is
232
bound in aluminosilicate particles and 50% is dissolved (e.g. passes through a 1 µm filter). For
Th and Th (radiogenic decay products), about 90% is found in solution. Sorption of Th
228
230
234
tracer is a reversible process, possibly due to an organic material coating on the surface with
increased sorption as the particles age. In such studies, solubility is defined as passage through
a 0.45 µm filter. Speciation calculations show typically that thorium occurs as a hydrolyzed
species whereas uranium is present in surface waters normally as the UO (CO )
and
2
3 2
2-
UO (CO )
species. In neutral waters of very low carbonate content, speciation calculations
2
3 3
4-
indicate that uranyl cations hydrolyze to form oligomers for total uranyl concentrations as low
as 10 M.
-7
Neptunium, in oxic waters, is present in the pentavalent state. The hydrated cation is
calculated to be the dominant species for pH#8 unless the free carbonate concentration exceeds
approximately 10 M in which case NpO (CO )
is more common.
-4
1-2n
2
3 n
The rather similar values for the plutonium reduction half reactions at pH 8 indicate that
plutonium may exist in oxic waters in more than one oxidation state. The reduction potential
at pH 8 of the Pu(III)/Pu(IV) couple indicates that Pu(III) is unlikely to exist in oxic waters in
the absence of a reductant, but may be present in anoxic waters. Each oxidation state of
plutonium differs in chemical behavior from that of the other states so modelling the geoche-
mical behavior of plutonium must include the correct oxidation state, or states, of plutonium
which are likely to be present in a particular system.
In principle, the calculation of the oxidation states of plutonium requires knowledge of the
redox potential, Eh, of the aqueous phase. However, the Eh measured with a certain type of
electrode may not be the potential for the particular redox couple with which the plutonium
reacts. One of the reasons for this is that the Eh-electrode usually catalyses the reaction rate for
its specific redox couple. For example, in surface sea water, the measured Eh is about 0.8 V
and is due to the O /H O couple. In the log Eh versus pH diagram,
2
2
Page 20
Behavior of Radionuclides in the Environment
661
Figure 22.4, the area of existence of plutonium in different oxidation states, including the
effects of hydrolysis and carbonate complexation, is marked by roman numerals. From that
diagram, it can be seen that Pu(VI) would be the predominant state in solution in the ocean. In
fact, the predominant species is Pu(V). Table 22.8 summarizes the reported oxidation state
distribution in several natural systems.
Like NpO , PuO
has a low tendency to hydrolysis and complexation and is much less
2
2
+
+
likely to be sorbed to solid surfaces and on colloidal particles than the Pu species in other
oxidation states. As a consequence, plutonium can be expected to migrate most rapidly if it is
in the pentavalent oxidation state. The total solubility is limited by the formation of the highly
insoluble Pu(OH) . The sorption of hydrolyzed Pu(IV) in neutral water on mineral and
4
organic-coated surfaces is accountable for the very low concentrations of dissolved Pu even in
the absence of Pu(OH) (am) or PuO (c). Desorption is accomplished only by strong complexing
4
2
and/or redox reagents. For example, citrate extracts little plutonium from soil, but a
combination of citrate and the redox agent dithionite provides good extraction. The intractable
nature of Pu(OH) and its strong tendency to sorb on surfaces is a dominant and often
4
controlling feature of plutonium geochemistry.
Silicates and humic substances present in natural waters form colloids and pseudocolloids with
which actinides can react. The pseudocolloids formed by humic substances in ground waters
have been shown to be efficient scavengers of americium and plutonium.
22.8. Natural analogues
The geochemical modeling calculations based on measured equilibrium data are of primary
importance in the safety assessments of proposed nuclear waste repositories (and should be
important for other waste repositories as well). Another useful tool in such assessments is the
data from studies of natural analogue sites.
The modeling calculations use data from laboratory and field studies which have been obtained
over the last few decades. These data are used in the codes to predict the solubilities, and
nuclide migration of material which might be released from nuclear repositories over thousands
and, even, hundreds of thousands of years. It is not possible to demonstrate rigorously that the
models used are accurate as they may simplify the natural system, use incorrect data or
misrepresent (possibly, ignore) important processes which occur over very long time periods.
However, some validation of the models and data used in the modeling calculations can be
obtained from careful comparison of calculated values with those measured at appropriate
geologic sites, known as a natural analogues.
The natural analogue sites are areas in which uranium ores have been present for geologic
time periods. In most cases, these sites have not been affected by human activities, so the
record of geologic, long term effects are well preserved. A number of such sites are being
studied around the world; studies from a few sites of different characteristics are reviewed here.
From studies of granitic sites, the dominant role of fractures and fissures in the transport of
fluids has been convincingly demonstrated. Thus any model of water in a granitic site must
include both advection which is dominant in the fractures and fissures, and diffusion which is
important in regions of highly altered rocks ("alteration rims"). In clays, mass
Page 21
Radiochemistry and Nuclear Chemistry
662
transport seems to proceed primarily by ionic and molecular diffusion although some fluid mass
transport occurs at discontinuities in the formation.
Often, the mobilization of many elements (e.g. U, but not Th, in crystalline rocks) is
correlated with the flow of oxidizing water. The mobilization and fixation of uranium involves
complexation, redox and retention on minerals via adsorption, and ion exchange. In clay media,
the redox potential is strongly buffered if significant amounts of organic substances are present.
In a section of the Pocos de Caldas (Brazil) formation, most of the thorium and the rare
earths, and, to a lesser extent, the uranium, is associated with goethite (FeOOH) particles and
transport by organic colloids is much less important. This region is reducing, as shown by the
presence of Fe(II). However, in another region of this formation, the thorium and rare earths
have a much higher mobility and are associated with organic (humic) colloids. At an analogue
site in Scotland, the flow from the ore has passed into a peat bog in which the uranium is
associated predominantly with humic material while the Th is found on Fe/Al oxyhydroxide
colloids and particles. In many clay deposits, the organic material is most significant in main-
taining a reducing potential which restricts actinide migration and provides a sorption source
of the mobilized fraction.
22.9. The Oklo reactor
Analysis of the Oklo natural reactors (§19.10) indicates that they must have lasted for 100 000
to 500000 years with criticality occurring periodically. They probably would have consumed
about twelve metric tons of U, releasing a total energy of 2 ! 3 GWy at a probable power
235
level of 10 kW. About 1.0 ! 1.5 t of Pu were formed by neutron capture in U, but the
239
238
half-life of Pu has resulted in its total decay to U. Since a few samples enriched in U
239
235
235
(presumably due to this decay) have been found, it is believed that in some places in Oklo
breeding conditions may have temporarily existed. The decaying fission products are estimated
to have had 10 $,( disintegrations over the operating time of the reactors. The average energy
28
release in the reactor zone was 50 W m which is several times greater than that planned
!2
for geological nuclear waste repositories. As a consequence, it is estimated that the fluids in the
inclusions of the mineral grains in the Oklo reactors had temperatures of 450 to 600EC, well
above those anticipated in deep waste repositories. There is evidence of connectively driven
circulation of the fluids for distances of 30 m from the main reactor zones as well as significant
dissolution and modification of minerals due to the thermal and radiation conditions.
Redistribution of some elements resulted from the convective flow of the hot liquids.
The uranium minerals appear to have remained relatively stable despite this heating process.
The uranium and the lanthanide elements show evidence of some small degree of localized
redistribution, but were mostly retained within the reactor zone. By contrast, fission product
rare gases, halogens, molybdenum and the alkali and alkaline earth elements migrated
significant distances from the reactor zones. In general, it seems that elemental redistribution
took place over a period of 0.5 to million years while the area was thermally hot (during and
after nuclear criticality). It has been estimated that as much as a total of 10 liter of hot water
12
flowed through the reactor zones. The water leaving the reactor zones had 5×10 g U m and
-3
!3
10
M concentrations of Tc, Ru, and Nd. The Tc and
!10
Page 22
Behavior of Radionuclides in the Environment
663
Ru were oxidized to TcO and RuO , and these soluble oxyanions moved with the water as
4
4
!
!
it flowed from the area, creating significant (25 ! 35%) deficiencies of these elements in the
reactor zones. The Nd was less soluble and, apparently, migrated much less during this hot
period. For Tc and Ru, the migration rate seems to have been 10 m y in water moving
!5
!1
at a flow of 5 m y .
!1
A very important observation is the evidence in Oklo that the plutonium produced by the
reactors did not move during its lifetime from the site of its formation.
In summary, essentially 100% of the Pu, 85 ! 100% of the Nd, 75 ! 90% of the Ru and 60
! 85% of the Tc were retained within the reactor zones. The migrating fission products were
held within a few tens of meters of these zones. Thermodynamic calculations of the temperature
dependent solubilities indicate that the loss of fissiogenic elements is diffusion controlled,
whereas, retention in the surrounding rocks is due to temperature dependent deposition from
an aqueous solution.
While the conditions at Oklo differ in a number of aspects from those expected in nuclear
repository sites, they frequently were much less favorable to retention of the radionuclides. The
lack of migration of the actinides and the much slower release of Tc agree with the predictions
of laboratory studies and indicate their value in validating the safety of nuclear repositories.
22.10. Performance assessments of waste repositories
In this Chapter we have presented some information on studies of the environmental behavior
of some fission products and actinides. We have shown how laboratory data can be used in
speciation calculations to predict solubilities, etc. The knowledge gained from studies of natural
analog sites and, very importantly, from the Oklo natural reactor site, has been reviewed. All
of these types of studies and data are of use in assessing the probable ability of nuclear waste
repositories to protect the public over the millennia required for the radioactivity to decay. The
most common design goal for these repositories is that no radioactivity shall be released within
the next 1 000 years, and that the leakage after that time would be so small that it presents no
danger to living species. The Environmental Protection Agency (EPA), USA, standard requires
that no more than 1000 cancers shall be caused in 10 000 years by radionuclides released from
the waste, using the ICRP predictions of radiation dose effects (Ch. 18).
The initiated radioactive inventory for spent reactor fuel consists of actinides, fission products
and activation products. As noted previously, (Ch. 21) the shorter lived fission products, such
as
Sr and
Cs, and transuranic elements, such as
Pu,
Pu,
Cm, are the main
90
137
238
241
244
contributors to the radioactivity. However, performance assessments strongly indicate that the
waste form matrix and the near field engineered barriers (e.g. clay backfill, etc.), can
successfully retain and prevent any migration to the far field environment for one thousand
years and probably much longer (>10 years). After the first thousand years the long lived
4
nuclides such as I, Cs, Sn, Tc and Se among the fission products and the actinides
129
135
126
99
79
U, U, Np, Pu, Pu, and Am become the major concern.
234
236
237
239
240
241
Page 23
Radiochemistry and Nuclear Chemistry
664
FIG. 22.10. Coupling of models in model chain for safety assessment of leakage of radionuclides from fractured fuel
rod and canister in a bentonite filled rock hole. (SKB91).
22.10.1. Release scenarios
Two major types of scenarios are considered in the performance assessments for release of
radionuclides from the repository. One scenario evaluates all the processes which are expected
to occur normally in the region of the repository which could effect the rate of release and
migration. In this scenario, ground water penetrates the waste packages and leaches the
radionuclides which, then, can migrate through and out of the repository, Figure 22.10. R1,
R2, etc represent mathematical models describing the migration of one particular radionuclide
in that specific region. The Ri's are then added into the overall equation.
Page 24
Behavior of Radionuclides in the Environment
665
Reducing conditions
Oxidizing conditions
________________________________________________________________________________________________________________
solubility
limiting
solubility
limiting
(M)
phase
(M)
phase
________________________________________________________________________________________________________________
Se
very low
M Se
high
-
x
y
Sr
1×10
Strontianite
1×10
Strontianite
!3
!3
Zr
2×10
ZrO
2×10
ZrO
!11
!11
2
2
Tc
2×10
TcO
high
-
!8
2
Pd
2×10
Pd(OH)
2×10
Pd(OH)
!6
!6
2
2
Sn
3×10
SnO
3×10
SnO
!8
!8
2
2
I
high
-
high
-
Cs
high
-
high
-
Sm
2×10
Sm (CO )
2×10
Sm (CO )
!4
!4
2
3 3
2
3 3
Am
2×10
AmOHCO
2×10
AmOHCO
!8
!8
3
3
Pu
2×10
Pu(OH)
3×10
Pu(OH)
!8
!9
4
4
Pa
4×10
Pa O
4×10
Pa O
!7
!7
2 5
2 5
TABLE 22.10. Radionuclide solubilities and limiting phase in reducing and oxidizing Finnsjö fresh water (SKB91)
To assess the release by this scenario, it is necessary to evaluate the rate of release from the
package, the flow rate of the underground fluids, the speciation and solubility of the different
radionuclides and their diffusion (migration) rates.
The second scenario includes "disturbed" conditions which could be the result of geologic
events such as earthquakes, volcanic activity and changes in hydrological conditions. This
scenario also includes the effect of "human intrusion," in which people in future generations
unknowingly penetrate a repository and release a portion of its radioactive contents to the
earth's surface via the groundwater system. For example, this release could be the result of a
drilling operation. Such events are assumed to occur over the history of the repository including
some at very early times (less than 1000 years). In this case, sorption and half-life and, for the
most part, solubility are not major factors in determining the potential release. The dominant
contribution would be from transuranic elements (mainly Pu, Pu, Am) with some
239
240
241
contributions from fission products.
22.10.2. Canister dissolution
The most common canister materials are copper and iron. In rocks, where the ground water
is reducing (as e.g. in the Canadian and Scandinavian shields), copper is practically insoluble
as shown by the existence of native copper, million years old, found in this environment.
Detailed studies have shown that the radiation from the waste nuclides have a very small effect
on the dissolution. It is therefore predicted that a 50 mm thick copper canister will be intact for
at least one million years. Iron, or steel, can be expected to dissolve more rapidly, especially
in oxidizing groundwater. However, the canister surrounding would become saturated by
Fe(II), resulting in a reducing media, which would be important in limiting the waste nuclide
migration.
Page 25
Radiochemistry and Nuclear Chemistry
666
FIG. 22.11. Fraction of released cesium from BWR and PWR fuels in different waters as a function of contact time
(SKB91).
As the metal encapsulation is dissolved or cracks (Figure 22.10.B), the radionuclides in the
waste matrix (UO or glass) will be released mainly with the dissolution rate of the matrix
2
(congruent dissolution). A reasonable figure for the glass dissolution rate is 2×10 g m d
!3
!2 !1
exposed glass surface, assuming no limit with regard to the solubility product (i.e. unlimited
amount of water); this corresponds to a corrosion rate of the glass surface of 2.7×10 mm
!4
y . Experience shows that this rate rapidly decreases with time. Thus strontium release rates
!1
were reduced by a factor of 10 in 15 y. Figure 22.11 shows measurements of the fraction of
6
Cs released from spent fuel as a function of contact time with simulated ground water. By the
time, the fuel matrix will be converted into hydrous oxide. The fraction altered has been
measured under various conditions and are extrapolated to very long times in Figure 22.12.
These results indicate that even if the barrier surrounding the canisters break down, the leakage
of waste products into the near field would be quite small.
The dissolution rate of canister encapsulation and waste matrix is limited by water solubility.
As mentioned above, of the actinides the most hazardous, plutonium, would most likely be in
the +4 state and form a very insoluble hydroxide. This is particulary true in an Fe(II)
environment caused by dissolution of an iron canister. Table 22.10 gives limiting solubilities
and phases for the more important waste elements in reducing and oxidizing groundwater.
Figure 22.13 shows the leakage rate from the near field of an initially defective canister,
extrapolated to long times; all nuclides that are expected to leak out at a rate of more than 1 Bq
y are included in the Figure. In this prediction, the stipulations mentioned in the beginning
!1
of §22.10 are met.
Page 26
Behavior of Radionuclides in the Environment
667
FIG. 22.12. Fraction of fuel altered in ground water as a function of time; the reaction is assumed to start 40 years after
fuel discharge from the reactor (SKB91).
22.10.3. Releases from bitumen and concrete encapsulations
The low-level and intermediate-level wastes are to be encapsulated in bitumen, concrete or
glass. The bitumen would be highly water resistant, but it ages with time and begins to lose
strength in 10 ! 20 years. Special bitumen materials have to be developed for wastes which
must be contained in >50 years. The bitumen drums are normally stored in a containment
building, e.g. of concrete.
The groundwater surrounding concrete rapidly becomes very basic, pH>13, as a consequence
of leakage of Na and K ions. At a later stage, Ca
ions begin to leak, reducing the pH to
+
+
2+
about 10.5. Thus nuclides migrating through the concrete encounter very basic media, leading
to the formation of insoluble hydroxides for most polyvalent ions, including all actinides. The
leaching of Ca
from the concrete causes it to lose its mechanical strength and to begin to
2+
deteriorate. By suitable additives, the onset of deterioration can be delayed, and at pH<10 the
concrete is stable for very long times. However, concrete containers for LLW and MLW can
not be considered to have an infinite lifetime, and therefore further containment is necessary.
22.10.4. Migration from the repository
After release into the near field, the radionuclides can only migrate via a water transport path
(Figure 22.10). Migration in the far field may occur for radionuclides with long lifetimes, high
solubility in ground water, and low sorption along the transport pathway. In repositories where
the water is confined to interstitial fracture and pore areas, the
Page 27
Radiochemistry and Nuclear Chemistry
668
Element
k value
Element
k value
d
d
(m kg )
(m kg )
3
!1
3
!1
______________________________________________________________________________________________
Zirconium
1
Carbon
0.001
Radium
0.15
Chlorine
0
Protactinium
1
Palladium
0.001
Thorium
2
Selenium
0.001
Uranium
2
Strontium
0.015
Neptunium
2
Cesium
0.15
Plutonium
0.2
Iodine
0
Technetium
1
______________________________________________________________________________________________
Conditions: specific surface area 0.1 m per m of rock, (equal to) 1 000 m per m of water;
2
3
2
3
matrix diffusion coefficient 3.2×10 m y ; diffusion porosity in the rock matrix 0.005.
!6 2 !1
TABLE 22.11. Far field radionuclide distribution values for granitic rock (SKB91)
Radionuclide
100 years
100 000 years
_______________________________________________________________________________________________________________
Dose Rate Risk
Dose Rate Risk
†
‡
(Sv y )
(y )
(Sv y )
(y )
!1
!1
!1
!1
_______________________________________________________________________________________________________________
Tc
5.0×10
1.5×10
1.9×10
2.5×10
99
!6
!14
!9
!14
I
1.8×10
5.4×10
3.2×10
4.2×10
129
!6
!16
!10
!15
Cs
2.4×10
7.0×10
1.3×10
1.7×10
135
!6
!15
!7
!12
Cs
1.5
1.8×10
- - -
- - -
137
!7
U
7.3×10
2.1×10
5.0×10
6.5×10
234
!4
!12
!7
!11
U
1.8×10
5.1×10
1.4×10
1.9×10
235
!5
!14
!6
!11
U
1.9×10
5.4×10
2.8×10
3.7×10
236
!4
!13
!7
!12
U
2.0×10
5.9×10
5.9×10
7.7×10
238
!4
!13
!7
!12
Np
1.1×10
3.1×10
3.6×10
4.7×10
237
!3
!12
!6
!11
Pu
3.6
1.8×10
5.1×10
6.7×10
238
!7
!6
!11
Pu
1.0
1.8×10
1.2×10
1.5×10
239
!7
!4
!9
Pu
1.5
1.8×10
3.1×10
4.0×10
240
!7
!7
!12
Pu
1.0×10
1.8×10
9.2×10
1.2×10
241
1
!7
!6
!9
Pu
5.4×10
1.6×10
8.9×10
1.2×10
242
!3
!11
!6
!10
Am
2.0
1.6×10
1.8×10
2.4×10
241
!7
!6
!11
Am
6.4×10
1.9×10
3.4×10
4.4×10
243
!2
!10
!6
!11
Cm
8.7×10
2.5×10
4.7×10
6.2×10
244
!2
!10
!9
!14
Cm
4.8×10
1.4×10
7.3×10
9.6×10
245
!4
!12
!9
!14
_______________________________________________________________________________________________________________
Total
2.1×10
1.8×10
1.6×10
2.1×10
1
!7
!4
!9
_______________________________________________________________________________________________________________
From S. F. Mobbs et al. Dose-risk conversion factor = 1.65×10 Sv recommended by ICRP in 1977
†
‡
!2
!1
[ICRP, 1977]. However, ICRP recommended a higher value, 5.0×10 Sv , in 1990 [ICRP #60, 1990].
!2
!1
TABLE 22.12. Individual doses and risks from intrusion scenarios in a granite repository of unreprocessed LWR spent
fuel (Mobbs et al.)
actinides would have low solubility and high sorption compared with the fission products. All
nuclides which adsorb would move slower than the free groundwater due to the
Page 28
Behavior of Radionuclides in the Environment
669
FIG. 22.13. Leakage from the near field of radionuclides from an initially defective canister; all nuclides that leak out
at a rate of > 1 Bq/y are included (SKB91).
sorption-desorption equilibria. As a consequence I and Tc would be expected to be the
129
99
principle contributors to a radioactivity release to the environment, as they are poorly adsorbed.
This is also true for C, which contributes because it is transported as dissolved CO (Fig.
14
2
22.13).
The effect of elevated temperatures resulting from the energy released by the radioactive
decay must be included in the evaluation of release and near field migration (i.e. within the
repository volume). Elevated temperatures could alter the geology as well as the chemical
speciation and solubility of the released nuclides. If the temperature exceeds the boiling point
of water in the fluids, it would result in a drier repository with reduced or no release and
migration.
When granite or clay is contacted with water containing dissolved cations, sorption or
exchange of these ions with ions of the solid phase are observed. For example, montmorillonite
has such a high exchange capacity that it is used as a natural ion exchanger, e.g. for water
purification. The ion exchange elution curves in Figure 16.7, however, do not depend on
sorption capacity, but on sorption strength and on aqueous complexation. The sorption strength
depends on ionic charge (higher charged species sorb more strongly), ionic size (smaller ions
sorb more strongly), etc., while aqueous complexation depends on the nature of the complexant
(ligand), as well as on the cation properties.
All ground/soil/rock materials sorb ions,but the sorption distribution value (defined by k , eqn.
d
(22.10)) depends on so many factors that they may be considered site specific; i.e. dependent
on the ion released, the near field matrix composition (buffer material, dissolved canister, etc),
the ground water composition and the rock mineral composition. However, clay and granite
have about the same k values in similar groundwaters.
d
The distribution coefficient, k , is defined by
d
Page 29
Radiochemistry and Nuclear Chemistry
670
FIG. 22.14. Predicted dose rates to individuals from the release of a defective canister, assuming the release occurs
directly to the biosphere, i.e. without retention in the ground (SKB91).
conc. of radionuclide per kg soil/rock/etc
k =
(22.11)
d
________________________________________________
conc. of radionuclide per m water
3
Typical k values are listed in Table 22.11. From the k values and the soil and groundwater
d
d
properties, the retention time for a radionuclide may be calculated. The radionuclide retention
factor (RF) is defined as
RF = v /v = 1 + k * (1 - ,) / ,
(22.12)
w n
d
where v is the groundwater velocity (in granite typically 0.1 m y ), v the nuclide transport
w
n
!1
velocity (which must be <v ), * the soil density (typically 1 500 ! 2500 kg m ), , the soil
w
!3
porosity (or void fraction, typically 0.01 ! 0.05), The groundwater velocity can be obtained
from eqn. (20.11), if geologic parameters are known. Observed k values for Swedish granite
d
are given in Table 22.11. Typical retention times in granite are about 400 y for Cs, 1 500 y for
Sr, 200000 y for Ln(III) and 40000 y for An(IV). Rather similar (within a factor of ten) values
have been found for other geologic formations such as tuff from New Mexico, basalt from
Idaho, and limestone from Illinois. The high retention values combined with a water transport
time of a few years lead to retention of Sr, Cs, and most Pu, Am, and Cm isotopes over
90
137
their lifetimes. With the negligible water flow rate through clay, diffusion is the dominating
transport process. It is estimated that it will take 700 y for Cs, 1 800 y for Sr, 22 000 y for Am
and 8 000 y for Pu to penetrate 0.4 m bentonite. During those times most of the radioactivity
of these nuclides have decreased to negligible values. Studies for a salt repository (WIPP) in
New Mexico indicate that, of all waste nuclides released in the repository, only C, Tc, and
14
99
I may
129
Page 30
Behavior of Radionuclides in the Environment
671
reach the surface before they have decayed. Thus, ground retention of the radionuclides plays
an essential role in the risk evaluation.
Both dissolution rate and transport rate of nuclides depends on the existence of complex
formers in the groundwater. Such complex formers are Cl , F , SO , HPO , CO , and
!
!
2!
2!
2!
4
4
3
organic anions (e.g. humic acid). Complexes with these anions, in most cases, increase
solubility, and ! through formation of less positively charged metal species ! reduce the
retention factors. The groundwater conditions, therefore, play a central role in evaluation of the
risks of a waste repository. Since these conditions vary, they have to be evaluated for each site.
This point is illustrated by the Maxey Flats facility in the USA, where radionuclides were found
to move from a storage basin much more rapidly through soil than would be expected from the
equations above. This was found to be caused by the addition of strong metal chelating complex
formers like diethylenetriaminopentaacetic acid (DTPA) to the basin liquid. The DTPA
complexes have negative charge, resulting in lower k values (usually 1). This reflects the
d
necessity to know in adequate detail the chemistry of the waste and of the storage site for proper
evaluation of the safety.
Figure 22.14 shows the result of a safety analysis, indicating the estimated amounts of
released radioactivities carried by ground water from a spent fuel repository to a recipient, and
the expected doses delivered, without consideration of retention. Retention will cause the Ra
226
and Pa to reach zero value before 10 000 years.
231
Calculations for individual doses and the risks for intrusion into a granite repository of LWR
spent fuel at 100 and at 100000 years after burial are listed in Table 22.12. These calculations
are based on very extensive modelling (cf. Fig. 22.10) of the releases and uptakes from wastes
produced from a 20 GW reactor park operating in Europe for 30 years, including reprocessing
e
and storage of LLW and ILW in surface facilities, and HLW, either in the form of spent fuel
elements or vitrified HLLW in an underground granite repository. To model the dissolution of
radionuclides and their migration in the ground existing facilities in their actual environments
have been chosen. Similarly, for the uptake actual food chains (crops, fish , etc and eating
habits) have been used. The quantity of spent fuel is assumed to be 18 000 t IHM. The LLW
and ILW repositories are assumed to begin to leak at time 0, the dissolved radionuclides
migrating into the actual soils, reaching streams and wells, etc, while it is assumed that the
spent fuel elements or high level waste glass begins to leak after 1 000 years. The radionuclides
migrate by groundwater under various retention conditions, depending on hydrogeologic
conditions and chemical properties. The exposed population is divided into four groups: the
critical group around the installations, the population of the country, the population in Europe
and in the world. As expected, the population dose is the highest for the critical and national
groups for aqueous releases, but about the same for all groups for gaseous releases. The model
also includes an intrusion scenario using a well drilling probability, beginning after 300 years
(in the main case), as it is expected that control will last that long. It is assumed that as soon
as a drill core from a waste repository is taken, it will immediately be recognized; this means
that high doses from the core will only be received by drilling and laboratory personal. The risk
from intrusion after 100 years is calculated to be 8.1×10 y and after 100 000 years 2.1×10
-7 -1
-9
y . By comparison, the calculated risk from the "normal" scenario of releases from spent fuel
-1
elements and migration in the environment over a long period is 5×10 y . With this exception
-5 -1
all the maximum calculated individual risks from disposal of solid wastes are below the limit
of 10 y recommended by the ICRP. For
-5 -1
Page 31
Radiochemistry and Nuclear Chemistry
672
vitrified high level waste and spent fuel disposal, the doses calculated for intrusion scenarios
are very high but the probability of occurrence are low, so the risks from intrusion are lower
than those from radionuclide migration with groundwater. The maximum individual doses from
the migration scenarios for solid waste disposal are predicted to arise beyond 10 000 years after
closure of the repository.
These values indicate that release and migration is the more significant concern, although even
for this scenario, geologic disposal is calculated to present very small risks to future
generations.
22.11. Conclusions
A rather large amount of nuclear fission products and actinide elements have been released
to the environment from nuclear weapons testing and from accidental and intentional discharges
from nuclear reactor operations and fuel reprocessing. The research on the fate of these
released radionuclides suggest that the long lived actinides form quite insoluble or strongly
sorbed species while I and Tc have relatively high dissemination in natural systems. The
129
99
most active shorter-lived species ( Sr, Cs) also have more mobility in ecosystems.
90
137
These conclusions are confirmed by the results of the investigations of radionuclide behavior
in natural analog sites. Even more relevant are the results of the natural reactor region at Oklo.
The data from field studies is largely confirmed by the performance assessments of proposed
nuclear waste repositories. The "intrusion" scenario is calculated to have a lower risk than the
undisturbed natural leach and migration scenario. The latter qualitatively agrees with the Oklo
data and indicates no unacceptable risks result from a carefully chosen and designed geologic
repository in which the nuclear wastes are emplaced with appropriate packaging.
22.12. Exercises
22.1. What is the purpose of the clay buffer of a waste repository in granitic rock?
22.2. Which of the following ions move slower than the groundwater: K , Cs , La , TcO , HCO ?
+
+
3+
!
!
4
3
22.3. Why is Np assumed to move faster than Pu in most groundwaters?
22.4. (a) What types of geologic formations are considered for waste repositories? (b) Which types are used/planned for
Asse and WIPP?
22.5. What will be the concentration of Am(CO ) at pH 7 under the conditions given in §22.7.
3 2
!
22.6. What will be the ratio of [Pu(III)]/Pu[(IV)] in a groundwater containing small concentrations of iron in the relation
Fe(II) 99% and Fe(III) 1%? EE
= 0.743 V. Neglect hydrolysis.
Fe(II)/Fe(III)
22.7. (a) Calculate the Pourbaix-line for Np(IV)/Np(V) in Figure 22.4. (b) What value must the Eh value of water exceed
at pH 6 for Np(V) to dominate? Neglect hydrolysis.
22.8. A miner has a deep well 160 km away from a waste repository. During an earthquake, the rock fractures and a
groundwater stream opens between the repository and the well so that 10 Ci Sr momentarily dissolves and moves toward
90
the well at v 160 km y . (a) What will be the amount of Sr (in Bq) reaching the well (assume plug flow)? (b) Will
w
!1
90
that water be harmful to the miner?
Page 32
Behavior of Radionuclides in the Environment
673
22.13. Literature
B. A
LLARD
, H. K
IPATSI
and J. O. L
ILJENZIN
, Expected species of U, Np and Pu in neutral aqueous solutions, J. Inorg.
Nucl. Chem. 42 (1980) 1015.
D. L. P
ARKHURST
, D. C. T
HORSTENSON
and L. N. P
LUMMER
, PHREEQE ! A Computer Program for Geochemical
Calculations, U. S. Geologic Survey, Water Resources Investigations, Report 90-86, Denver, 1980.
E. R. S
HOLKOVITZ
, The Geochemistry of Plutonium in Fresh and Marine Environments, Earth-Science Rev. 19 (1983)
95.
P. J.C
OUGHTREY
, D. J
ACKSON
, C. H. J
ONES
and M. C. T
HORNE
, Radionuclide Distribution and Transport in Terrestrial
and Aquatic Ecosystems: A Critical Review of Data, CEC and U.K. Ministry Agriculture, Fish and Food, A. A.
Balkema, Rotterdam, 1984.
B. A
LLARD
, U. O
LOFSSON
and B. T
ORSTENFELT
, Environmental actinide chemistry, Inorg. Chimica Acta 94 (1984) 205.
D. R
AI
, Solubility Product of Pu(IV) Hydrous Oxide and Equilibrium Constants of Pu(IV)/Pu(V), Pu(IV)/Pu(VI), and
Pu(V)/Pu(VI) Couples, Radiochim. Acta, 35 (1984) 97.
D. C. H
OFFMAN
and G. R. C
HOPPIN
, Chemistry Related to Isolation of High-Level Waste, J. Chem. Educ. 63 (1986)
1059.
J. W. M
ORSE
and G. R. C
HOPPIN
, The Chemistry of Transuranic Elements in Natural Waters, Rev. in Aquatic Sci. 4
(1991) 1.
S. F. M
OBBS
, M. P. H
ARVEY
, J.S. M
ARTIN
, A. M
AYALL
and M. E. J
ONES
, Comparison of the Waste Management Aspects
of Spent Fuel Disposal and Reprocessing; Post Disposal Radiological Impact, NRPB report EUR 13561 EN, UK, 1991.
Swedish Nuclear Fuel and Waste Management Co., SKB 91: Final Disposal of Spent Nuclear Fuel, Importance of the
Bedrock for Safety, SKB Tech. Report 92-20, Stockholm, 1992.
J.-C. P
ETIT
, Migration of Radionuclides in the Geosphere: What Can We Learn from Natural Analogues?, in Chemistry
and Migration Behavior of Actinides and Fission Products in the Geosphere, Radiochim. Acta, special issue 58/59 (1992).
F. B
RANBERG
, B. G
RUNDFELT
, L. O. H
OGBUND
, F. K
ARLSON
, K. S
HAQUIS
and J. S
MELLIE
, Studies of Natural Analogues
and Geologic Systems, SKB Tech. Rept., 93-05, Stockholm, 1993.
H. R.
VON
G
UNTEN
and P. B
ENES
, Speciation of Radionuclides in the Environment, Paul Scherrer Institute Report, No.
94-03, Würenlingen, 1994.
U
NITED
N
ATIONS
S
CIENTIFIC
C
OMMITTEE ON THE
E
FFECTS OF
A
TOMIC
R
ADIATION
, Sources and Effects of Ionizing
Radiation, UNSCEAR 2000 Report to the General Assembly, ISBN 92-1-142238-8, United Nations, 2000.