Pd-Catalyzed Surface Reactions of Importance in Radiation Induced Dissolution of Spent Nuclear Fuel Involving H2

Article abstract of DOI:10.1002/cctc.201901128

To assess the influence of metallic inclusions (?-particles) on the dissolution of spent nuclear fuel under deep repository conditions, Pd-catalyzed reactions of H₂O₂, O₂ and UO₂ ²⁺ with H₂ were studied using Pd-powder suspensions. U(VI) can efficiently be reduced to less soluble U(IV) on Pd-particles in the presence of H₂. The kinetics of the reaction was found to depend on the H₂ partial pressure at p_H2≤5.1×10^?2 bar. In comparison, the H₂ pressure dependence for the reduction of H₂O₂ on Pd also becomes evident below 5.1×10^?2 bar. Surface bound hydroxyl radicals are formed as intermediate species produced during the catalytic decomposition of H₂O₂ on oxide surfaces. While a significant amount of surface bound hydroxyl radicals were scavenged during the catalytic decomposition of H₂O₂ on ZrO₂, no scavenging was observed in the same reaction on Pd. This indicates a different reaction mechanism for H₂O₂ decomposition on Pd compared to metal oxides and is in contrast to current literature. While Pd is an excellent catalyst for the synthesis of H₂O₂ from H₂ and O₂, a similar catalytic activity that was previously proposed for ZrO₂ could not be confirmed.

Full text of DOI:10.1002/cctc.201901128

Accepted Article
Title: Pd Catalyzed Surface Reactions Involving H2 of Importance in
Radiation Induced Dissolution of Spent Nuclear Fuel
Authors: Annika Carolin Maier and Mats Jonsson
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201901128
Link to VoR: http://dx.doi.org/10.1002/cctc.201901128
A Journal of
www.chemcatchem.org

ChemCatChem
10.1002/cctc.201901128
FULL PAPER
Pd Catalyzed Surface Reactions Involving H of Importance in
2
Radiation Induced Dissolution of Spent Nuclear Fuel
[
a]
[a]
Annika Carolin Maier* and Mats Jonsson
Abstract: To assess the influence of metallic inclusions (ε-particles)
the radiolytic oxidants. Under the conditions often used in
safety assessments, i.e., groundwater intrusion after more
than 1000 years, H O has been shown to be the radiolytic
on the dissolution of spent nuclear fuel under deep repository
2
+
conditions, Pd-catalyzed reactions of H
were studied using Pd-powder suspensions. U(VI) can efficiently be
reduced to less soluble U(IV) on Pd-particles in the presence of H
The kinetics of the reaction was found to depend on the H partial
pressure
on Pd also becomes evident
2
O
2
, O
2
and UO
2
with H
2
2
2
[
5]
oxidant of main importance. The reaction mechanism for
2
.
oxidative dissolution of UO
reactions 1-2.
by H
2
O
can be described by
2
2
2
-
2
pressure at pH2 ≤ 5.1 x 10 bar. In comparison, the H
2
dependence for the reduction of H
2
O
2
!!
ꢀ!
ꢀ! + ꢀꢁ
! (!)
→ ꢀꢁ
! !"#
+ 2ꢀꢁ!
(1)
(2)
-
2
below 5.1 x 10 bar. Surface bound hydroxyl radicals are formed as
intermediate species produced during the catalytic decomposition of
!
!
!"#
!!
ꢀ
ꢁ
→ ꢀꢁ
! !"##$%&'!
!
H
2
O
2
on oxide surfaces. While a significant amount of surface bound
hydroxyl radicals were scavenged during the catalytic decomposition
of H on ZrO , no scavenging was observed in the same reaction
on Pd. This indicates a different reaction mechanism for H
2
O
2
2
The competing reaction to uranium oxidation is catalytic
decomposition of H on the UO surface forming water
2
O
2
decomposition on Pd compared to metal oxides and is in contrast to
current literature. While Pd is an excellent catalyst for the synthesis
2
O
2
2
and molecular oxygen according to reactions 3-5.
of H
2
O
2
from H
2
and O
2
, a similar catalytic activity that was
could not be confirmed.
∙
previously proposed for ZrO
2
ꢀ ꢀ → 2ꢀꢁ
(3)
(4)
(5)
!
!
!"#
∙
ꢀ
ꢁ + ꢀ ꢀ → ꢀ ꢀ + ꢀꢁ∙
! ! ! !
Introduction
∙
∙
ꢀ
ꢁ + ꢀꢁ → ꢀ ꢀ + ꢀ
! ! ! ! !
Worldwide around 10 % of the electricity produced today is
[1]
Note the homolytic cleavage of H
2
Ÿ
O
2
into surface bound
based on nuclear energy.
Although greenhouse gas
hydroxyl-radicals. Adsorption of the OH is a prerequisite for
this reaction to be spontaneous at room temperature. The
emissions from nuclear power are low, other waste streams
are of particular concern. Most fuel assemblies used in
ratio between the two competing reactions,
2 2
H O
2
nuclear reactors today are based on UO . Used nuclear fuel
decomposition and uranium dissolution, is often referred to
as the dissolution yield (expressed as the ratio between the
amount of dissolved uranium and the amount of consumed
can be reprocessed, although most countries are currently
planning to dispose the highly radioactive used fuel in deep
[
2]
repositories. In case of barrier failure, spent nuclear fuel
will be in contact with groundwater. The groundwater at
2 2
H O ). The dissolution yield has been found to depend on
[
6]
2 2
the H O concentration.
2
repository sites is expected to be reducing where UO has a
[
3]
very low solubility. However, the inherent radioactivity of
the spent nuclear fuel will induce radiolysis of the
groundwater in contact with the fuel. As a result, both
When spent nuclear fuel is discarded from a nuclear reactor
it consists of around 95 % UO , the remaining 5 % being
2
[
7]
-
Ÿ
fission products and heavier actinides. Fission products are
categorized based on their appearance in spent nuclear fuel
according to: 1) Fission gases and volatile fission products,
Ÿ
2 2 2
oxidants ( OH, H O ) and reductants (eaq , H , H ) will be
[
4]
produced in the vicinity of the fuel surface. For kinetic
reasons, the surface reactions will initially be dominated by
2
) Fission products forming metallic precipitates (ε-particles),
3
) fission products forming oxide precipitates as well as 4)
[
8]
fission products which substitute U in the fuel matrix.
ε-
[a]
A. C. Maier, Prof. Dr. M. Jonsson
Department of Chemistry
School of Engineering Sciences in Chemistry, Biotechnology and
Health
particles usually contain Mo, Ru, Tc, Rh and Pd in variable
ratios depending on the oxygen potential of the fuel and the
[
9,10]
location from the rim to the center of a fuel pin.
These
KTH – Royal Institute of Technology
SE-10044 Stockholm, Sweden
E-mail: * acmaier@kth.se
metallic particles have been shown to efficiently catalyze
several reactions of major importance in the process of
[11-15]
radiation-induced dissolution of spent nuclear fuel.
In
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201901128
FULL PAPER
several previous studies on model systems, Pd has been
fractured spent nuclear fuel, reaction rates for the reduction
[11, 12]
used to mimic the ε-particles,
or to reduce U(VI) in order
2
of uranyl at H pressures below 1.5 bar had to be estimated
[
16,17]
[23]
to determine the solubility of UO
2
(s).
Figure
1
as experimental results were not available.
summarizes the most important reactions at the interface
between spent nuclear fuel and groundwater.
To complement missing experimental results for
computational models of spent nuclear fuel corrosion in a
deep repository, we have performed experiments monitoring
the reduction of uranyl as well as H
2
O
2
consumption at H
2
2
partial pressures below 1.5 bar. The partial pressure of H
has been varied using gas mixtures containing 100, 5, 0.5,
.05 and 0 % H in N to make sure that the pressure
dependent regime can be identified. To elucidate the
mechanism for catalytic decomposition of H on Pd, we
0
2
2
2 2
O
used tris(hydroxymethyl)aminomethane (tris) as a scavenger
for hydroxyl radicals. The results are compared to results
Figure 1. Scheme of relevant processes at the interface between spent
nuclear fuel and groundwater.
from experiments performed using ZrO
2
instead of Pd. For
ZrO , the formation of surface bound hydroxyl radicals upon
2
exposure to H
2
O
2
has been confirmed and quantified in
In a deep repository after barrier failure, reactions catalyzed
by metallic inclusions are diverse. The most important
processes are the ε-particle catalyzed solid phase reduction
[
24-26]
several studies.
th
[
15]
Since the beginning of the 20 century it is known that H
2
O
2
of U(VI) to U(IV) (figure 1, reaction A),
that efficiently
can form from its elements at ambient temperatures when a
competes with the dissolution of U(VI) (figure 1, reaction B)
[
27]
suitable catalyst is used.
recently reported for ZrO
experiments the H
A similar catalytic behavior was
and the ε-particle catalyzed reaction between H
2
O
2
and H
2
2
[
28]
2
.
In complementary autoclave
(
figure 1, reaction C) that reduces the amount of H
2
O
[
11]
2
O
2
formation from H
2
and O
2
and its
available for oxidation of the fuel.
In addition to being a
successive decomposition on Pd and ZrO
2
was studied,
radiolysis product, H
2
is produced upon anaerobic corrosion
Under deep repository conditions, the hydrogen
production rate as a result of canister corrosion was roughly
Ÿ
[
18]
monitoring OH concentrations over time.
of steel.
-
1
-2
estimated to be in the order of 0.4 mol year
m
assuming
Results and Discussion
that 1 mole of hydrogen is produced per mole of consumed
[
19]
2
metal which leads to a fast buildup of H in the repository.
Uranyl Reduction by H
As can be seen in figure 2, the reduction of U(VI) to U(IV) in
the presence of H is catalyzed by Pd. A simplified reaction
2
Studies on simpler model systems have shown that the
kinetics of the reaction between H and H catalyzed by
Pd is independent of the H partial pressure between 1 bar
and 40 bar and that the reaction is virtually diffusion
2
2
O
2
2
mechanism is shown in the insert. As mentioned above, it
was previously shown that the kinetics of U(VI) reduction
2
[
11]
shows no dependence on the H
2
partial pressure between
When decreasing the H partial
pressure below 1 bar, it is evident that the reaction rate
becomes dependent on the H partial pressure. Compared to
the experiments carried out at 1.01 bar H , already at 5.1 x
the rate of U(VI)-reduction is decreased by a
factor of approximately 2. The largest apparent change in
controlled. Under these conditions, the competing catalytic
decomposition of H on the Pd particles could not be
observed. However, when lowering the H partial pressure
sufficiently, the catalytic decomposition of H on the Pd
surface will become the predominant reaction pathway. In
[
12]
1
.5 bar and 40 bar.
2
2 2
O
2
2
2 2
O
2
-2
[
20-22]
10 bar H
2
the literature,
the reaction mechanism for
2 2
H O
decomposition on Pd is claimed to be the same as for metal
oxides as described by reactions 3-5, where the surface
-
3
-4
reactivity is observed between 5.1 x 10 bar and 5.1 x 10
bar. However, this change can to a large extent be attributed
to a difference in the initial lag time of the reaction. The initial
lag phase is ascribed to the presence of oxygen in the
injected uranyl solution. Its length was found to depend on
Ÿ
bound OH is the key-intermediate.
In addition to the reactions mentioned above, it was
previously shown that in the presence of H
catalyst to reduce dissolved U(VI) back to U(IV).
rate of reduction was also found to be independent of the H
2
, Pd is a good
[
12,16,17]
2 2
the H partial pressure. While for 1.01 bar H 7 % of the
The
uranyl is reduced after 35 min, only 4 % and 2.5 %
respectively is reduced at hydrogen partial pressures of 5.1 x
2
[
12]
partial pressure between 1.5 bar and 40 bar.
When the
-3
-4
1
0
bar and 5.1 x 10 bar, during the same time interval.
same experiment was carried out in the absence of H
reduction of U(VI) was not observed. In a recent approach to
model the influence of H pressure on the corrosion of
2
,
The oxygen concentration can be calculated to be around 14
µM, almost 1.5 times as much as the uranyl present initially.
2
2
The oxygen in the system is likely to react with H in the
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201901128
FULL PAPER
2
bar N (blue diamonds).
2 2 2 2
presence of Pd to form H O . The H O in turn can react in
several ways among which, one is the re-oxidation of U(IV)
to U(VI) or it can react with H
reactions involving O / H
2
2
to from water. In any case, the
appear to efficiently compete
In table 1 the experimentally determined pseudo-first order
rate constants (k ) for different H partial pressures are
given. In contrast to what was observed in a previous
2
2
O
1
2
with the reduction of U(VI) which starts only when the
oxidants are consumed.
[
11]
study,
2 2
the reactivity of H O is fairly high also in the
absence of H
2
.
Table 1. Pseudo-first order rate constants for the consumption of
H
2
2
O in suspensions containing Pd particles.
-
1
2
H
2
partial pressure / bar
k
1
/ s
R
-
4
1
5
5
5
0
.01
(9.3 ± 0.5) x 10
(8.8 ± 0.4) x 10
0.9945
0.9963
0.9997
0.9999
0.9974
-
-
-
2
3
4
-4
.1 x 10
.1 x 10
.1 x 10
.00
-4
(4.32 ± 0.06) x 10
-4
(2.198 ± 0.006) x 10
-
4
(1.91 ± 0.07) x 10
From table 1, the H
2
pressure dependence becomes evident
below 5.1 x 10 bar. The background consumption of H
on the glass vessel as well as the dissolution of Pd species
determined by ICP-OES) are negligibly small during the
Figure 2. U(VI) reduction as a function of time on 8 mg Pd at different H
2
-2
-
2
2 2
O
partial pressures; 1.01 bar H
2
(red dots), 5.1 x 10 bar H
2
(black triangles), 5.1
-
3
-4
x 10 bar H
2
2
(green triangles), 5.1 x 10 bar H (yellow diamonds) and 1.01
bar N
2
(blue diamonds).
(
timeframe of the experiments.
H
2
O
2
Reactivity
2 2
As can be seen, the overall reactivity of H O decreases with
Figure 3 shows the concentrations of H
time in Pd-suspensions in the presence of H
Interestingly, no lag time is observed here even though the
injections also introduce approximately 15 µM O into
the system. This is attributed to the initial H concentration
being one order of magnitude higher than the expected O
concentration, i.e., O and more importantly H formed
from O reacting with H can only marginally influence the
kinetics for H consumption.
2
O
2
as a function of
decreasing H
2
partial pressure. The consumption of H
2
O
2
1
2
and/or N
2
.
displays first order kinetics for a given H
2
partial pressure (k
vs. pH2 is shown in figure 4a). Hence, the rate-determining
step is the reaction between H and Pd (reaction (6)), and
adsorption of H (reaction (7)) is fast enough to keep the
level of adsorbed H constant.
H
2
O
2
2
2 2
O
2
O
2
2
2
2
2
2 2
O
2
2
ꢀ!
ꢀ! + ꢀꢁ ⋯ ꢀ!"# → ꢀꢁꢂꢃꢄꢅꢆꢇ
(6)
(7)
2 2
O
!
"
ꢀ!
ꢀꢁ ⋯ ꢀ!"#
The rate law for H
2 2
O consumption through Pd catalyzed
reduction by H can then be described according to equation
2
8
,
!
!!!!
−
= ꢀ [ꢀ ꢀ ]ꢀ
(8)
!
!
!
"
where Θ is the fractional surface coverage of H
2
on the Pd
particles. Equation 8 can be simplified to equation 9 if the
concentration of adsorbed
H
2
on the Pd-particles is
*
considered to be constant (k = kΘ).
!
[!!!!]
∗
=
−ꢀ [ꢀ ꢀ ]
(9)
!
!
!
"
Figure 3. H
2
O
2
consumption as a function of time on 8 mg Pd at different H
2
-
2
partial pressures; 1.01 bar H
2
(red dots), 5.1 x 10 bar H
2
(black triangles, 5.1
-
3
-4
*
x 10 bar H
2
(green triangles), 5.1 x 10 bar H
2
(yellow diamonds) and 1.01
Hence, we expect to observe first order kinetics. Note that k
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201901128
FULL PAPER
is k
1
-kcat, i.e. the observed rate constant minus the rate
on Pd. From
the experiments we have determined k at different H partial
The rate constant determined above is only valid at the solid
surface area to solution volume ratio used in the
constant for the catalytic decomposition of H
2
O
2
*
2
experiments. In a previous study, the kinetics of H
2 2
O
pressures i.e. at different Θ. Assuming that Θ can be
consumption was studied as a function of Pd surface area to
*
[11]
described by the Langmuir isotherm, we can express k
solution volume ratio at H
these conditions the rate of
independent of the pressure, i.e. ꢀ =1. The rate
expression under these conditions can be written as follows
2
pressures above 1 bar.
Under
through equation 10.
2 2
H O
consumption is
H
2
!
! [!!]
!! !!
ꢀ∗
= !
(10)
(
equation 12):
Inverting equation 10 gives a linear relationship (equation
1).
!
!!!!
!"
−
= ꢀ! ꢀ ꢀ
∙
(12)
!
!
1
!"
!
-
5
-1
!
!∗
!
!
and the rate constant k
In the present study we can estimate the corresponding H
2
was determined to 2.2 x 10 m s .
= !
+
(11)
!"[!!]
2
-5
-1
pressure independent rate constant to 2.4 x 10 m s . The
perfect agreement between the two results is mainly
coincidental, as two different batches of Pd-particles were
used and the specific surface area in the work by Nilsson
and Jonsson was estimated from the average Pd-particle
A
size. To obtain a rate expression that is valid at all H
2
pressures we can combine the two expressions to equation
1
3,
!
!!!!
!"
−
= ꢀ! ꢀ ꢀ
∙
∙ ꢀ
(13)
!
!
!
"
!
-
5
-1
where k
rate constant for a diffusion controlled process when µm-
2
= 2.4 x 10 m s . In comparison, the second order
-
3
sized particle suspensions are used is in the order of 10
m
-
1 [29]
s .
Hence, the reactions studied here are clearly not
diffusion controlled.
B
Interestingly, when the system is purged and stirred for a
longer time before the injection of H
of 30 min), the reactivity of H towards Pd is significantly
decreased. This is illustrated for N purged samples in figure
. A similar decrease in reactivity was observed in uranyl
2 2
O (i.e. for 16 h instead
2 2
O
2
5
reduction experiments, when the stirring and purging time
was increased before the uranyl injection (data not shown).
Particle agglomeration is known to be
a
common
phenomenon decreasing the catalytic activity of Pd by
[
30]
reducing the available surface area.
SEM images of the
Pd powder after different time intervals during which the
powder suspensions were stirred are shown in figure 6. From
the images, an increase in agglomerate size can be
observed, to which we qualitatively ascribe its reduced
reactivity. This observation connects the higher reactivities
from this work, with previously published results where
Figure 4. (A) First order rate constants for H
2
O
2
consumption at different pH2
2 2
and (B) a Langmuir Hinshelwood plot for the kinetics of H O consumption.
When plotting the experimental data using equation 11
[11,12]
3
significantly lower reactivities were observed.
be noted that the Pd-suspensions in the work by Nilsson and
It should
(
figure 4b) one obtains an intercept of 1.09 x 10 s and a
-2
slope of 1.34 x 10 s M. From the intercept the second order
[
11,12]
-4
-1
Jonsson
were continuously stirred in an autoclave over
respectively.
rate constant, k, can be calculated to 9.2 x 10 s . From the
night before the injection of uranyl or H
2
O
2
slope of figure 4b and the rate constant, the equilibrium
4
2
constant for adsorption of H can be calculated to 8.1 x 10
-
1
M .
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201901128
FULL PAPER
Figure 6. SE-SEM images of Pd-powder that was exposed to water without
stirring and purging (A), Pd powder that was stirred and purged for 30min
2 2
before the injection of H O (total stirring time approx. 6 h)(B) and Pd powder
that was stirred and purged for 16 h before the injection of H
time approx. 22 h)(C).
2 2
O (total stirring
From the results presented above it is obvious that H
catalytically decomposed on the Pd surface. When
comparing the reactivity of H towards ZrO and Pd in the
presence of tris as scavenger for surface bound OH (figure
2 2
O is
2
O
2
2
Ÿ
7
a), it is clear that the reactivity of H
2
O
2
towards untreated
. It should be noted
Pd powder is higher as compared to ZrO
2
that the solid surface area to solution volume ratio is a factor
of 10 higher in the case of ZrO in the experiment and
therefore the difference in reactivity between ZrO and Pd is
even larger than it appears in figure 7a. Experiments
2
2
Figure 5. H
Purging and stirring for 30 min before the
pentagons), as well as purging and stirring for 16 h before the H
2
O
2
consumption as a function of time on 8 mg Pd and N
injection (filled blue
injection
2
purging.
H
2
O
2
Ÿ
performed on ZrO
2
in the presence of tris as OH scavenger
O (the final
product in the reaction between OH and tris) when H is
2
O
2
show that there is a significant production of CH
2
(
open orange pentagons).
Ÿ
2
O
2
consumed (figure 7b). Since the yield of formaldehyde is
Ÿ
proportional to the concentration of surface bound OH, tris
Ÿ
can be used as a quantitative probe for OH. A reaction
A
Ÿ
mechanism for the reaction between OHads and tris was
[
31]
proposed by Yang and Jonsson.
A
B
C
B
Figure 7. (A) H
2
O
2
decomposition and (B) CH
2
O production from scavenged
Ÿ
-1
OH over time with SA/V = 1620 m for Pd (green triangles down) and 16200
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201901128
FULL PAPER
-
1
m
2
for ZrO (red triangles).
concentration in air, a theoretical H
mM could be reached in the reaction vessel, assuming that
all O is dissolved and converted into H . Due to the
continuous and rapid consumption of H on Pd as well as
the time interval for sampling, the measured peak
2 2
O concentration of 190
The final yield of formaldehyde depends on the relative
surface coverage of both adsorbed H and tris (governed
2
2 2
O
2 2
O
2
O
2
by their initial concentrations and the relative affinity of the
Ÿ
adsorbent to the surface) as well as the competition for OH
concentration of approximately 0.7 mM H
the theoretical maximum.
2 2
O , is well below
[
26]
2 2
by H O and tris.
It is interesting to note that, even though H
with Pd, no CH O above background levels can be detected.
This implies that no significant amount of scavengeable OH
is formed upon catalytic decomposition of H on the Pd
surface. Consequently, the mechanism of catalytic
decomposition of H on Pd would appear to differ from
that of H on ZrO . This finding contradicts the assumption
made in the literature
decomposition on the Pd surface is the same as on
2 2
O reacts rapidly
2
When measuring CH O over time, it is evident that the
2
concentration hardly increases. The slight increase is most
Ÿ
Ÿ
likely due to scavenging of OH produced in the background
2
O
2
decomposition of H
above (addition of H
shows that hydrogen-abstracting radicals (like the OH),
2 2
O . As in the experiment described
2
O to a Pd-suspension), this result
2
2
O
2
Ÿ
2
O
2
2
which could lead to a significant production of CH
produced in the system.
2
O, are not
[
20-22]
that the reaction mechanism for
2 2
H O
metal oxides (reactions 3-5). A reaction mechanism, which
Results from autoclave experiments where ZrO
catalyze the reaction between O and H (in the absence of
Pd) are shown in figure 9. Note that the detection limit for
is approx. 0.01 mM. As can be seen, there is no
detectable H –formation (figure 9a), which indicates that
is either not produced at all, or it rapidly decomposes
on the surface of ZrO . The latter would result in the
formation of CH O. As can be seen, the CH O concentration
increases significantly in the presence of ZrO (red triangles
2
was used to
Ÿ
does not include the OH , was to the best of our knowledge
2
2
[
32-34]
only proposed in the 1960s,
in later research.
but was rarely considered
2 2
H O
2
O
2
2 2
H O
Autoclave Experiments
2
To further explore the activity of Pd as a model for ε-particles
in catalyzing reactions between molecular radiolysis products,
we studied the formation (reaction 14) and subsequent
2
2
2
pointing downwards) when a PEEK (Polyether ether ketone)
stirrer is used. However, using the same stirrer in the
decomposition of H
and O in contact with an aqueous suspension containing
Pd particles. The results are shown in figure 8.
2 2
O in an autoclave originally containing
H
2
2
absence of ZrO
2
, the background production of CH
2
O (blue
dots) overlaps the results for ZrO
the experiment.
2
within the uncertainty of
!
"
ꢀ!
+ ꢀ!
ꢀ ꢀ
(14)
!
!
In a previous study based on experiments using the same
equipment a very similar observation was made in the
[
28]
presence of ZrO
2
.
2
In this case the CH O concentration
reached approximately 225 µM after 300 h. This
concentration is only slightly higher compared to the results
presented above. From the previous results (where also
[
28]
H
2
O
2
was detected as a transient)
catalyzes the reaction between H
it was concluded, that
and O to form H
ZrO
2
2
2
2 2
O .
However, in that study background measurements were
never made as PEEK was assumed to be an inert material.
When removing the stirrer and in the absence of any powder,
neither H
Figure 9b orange triangles). The same was found in the
presence of ZrO when the PEEK stirrer was replaced with a
glass stirrer (Figure 9b green diamonds). As neither, H
nor CH are detectable in concentrations above
background, we must ascribe the catalytic activity that was
2 2 2
O nor CH O are produced in detectable quantities
(
2
2
O
2
2
O
[
28]
Figure 8. H
2
O
2
production and consecutive consumption on Pd at pH2 = 40 bar
reported in
with Pd. We found that 2 mg of Pd is enough to produce a
significant amount of H (figure 8). As Pd was previously
used in the same autoclave,
to aging of the PEEK stirrer or a contamination
and pO2 = 0.2 bar. The displayed data points are the result of two identical
experiments to assure reproducibility.
2
O
2
[11,12]
small quantities could have
contaminated the stirrer still being enough to see the
enhanced background in the presence of the PEEK stirrer.
Based on the volume of the autoclave and the oxygen
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201901128
FULL PAPER
2 2 2 2
catalyze the reaction between H and O to form H O , we
were not able observe the same catalytic behavior ascribing
the catalytic behavior reported earlier to contamination or
aging of the experimental setup.
A
B
Experimental Section
ZrO
Table 1 and were used as received. The BET surface area of ZrO
was determined by isothermal adsorption and desorption of a 30 %
in He gas mixture on a Micrometrics Flowsorb II 2300 device. For
2
as well as Pd powder used during the experiments are listed in
2
N
2
Pd, the specific surface area was calculated from the particle sizes
given by the supplier. Throughout all experiments purified water
(
18.2 MΩ cm, Merck MilliQ) was used.
Table 1. Powders used during the experiments
Metal / Metal Oxide
Supplier
Specific Surface
2 -1
Area / m g
ZrO
Pd
2
Aldrich 99 %
6.44 ± 0.05
Aldrich ≥ 99.9 %
1.35 calculated
Uranium concentrations were measured spectrophotometrically at
53 nm using the Arsenazo III method [35], whereas
concentrations were measured using the Ghormley triiodide method
36]. U(VI) in solution absorbs light at the low wavelength side of the
6
2 2
H O
[
Figure 9. (A) H
2
O
2
and (B) CH
2
O concentrations measured over time in the
triiodide peak at 350 nm. To avoid an overlap of the two peaks,
samples containing uranium were therefore measured at 360 nm
instead. Individual calibrations were made for both wavelengths.
absence of Pd at pH2 = 40 bar and pO2 = 0.3 bar.
Hydroxyl radicals were scavenged using tris (BDH chemicals ≥
Conclusions
9
9 %). Tris reacts with hydrogen abstracting radicals to form
formaldehyde, which can be measured indirectly by
The experimental results presented above give new insights
spectrophotometry using the modified Hantzsch method [37].
Samples containing formaldehyde were left to react with 2 M
on the effects of ε-particle (mimicked by Pd) catalyzed
reactions in
a deep geological repository. In case of
ammonium acetate (Sigma Aldrich
≥ 99 %) and 0.04 M
groundwater intrusion into a repository, ε-particles can
◦
acetoacetanilide (SAFC, Sigma Aldrich ≥ 98 %) for 15 min at 40 C
forming a dihydropyridine derivative, which can be measured at 368
nm. Before each experiment with tris, the pH was adjusted to 7.5
with HCl.
efficiently reduce dissolved U(VI) to the considerably less
soluble U(IV), even in the presence of small quantities of H
2
.
The reduction of U(VI) was found to become depended on
-2
the H
of H
dissolution also decreases with decreasing
pressure, where the dependence also becomes evident
2
partial pressure at pH2 ≤ 5.1 x 10 bar. The reactivity
, the key radiolysis product to induce oxidative
partial
2
O
2
Experiments on aqueous powder suspensions where the partial
H
2
2
pressure of H was varied, were carried out in a three neck round
bottom flask. One of the necks was permanently closed with a glass
stopper, the second one with a septum and a custom made in-/outlet
for purging and sampling was inserted into the third neck. The in-
and outlet is equipped with two valves and a glass frit for purging.
When closing the valve that is used to release excess gas, the
vessel can temporarily be slightly pressurized. The overpressure in
the vessel is then used to drive a liquid sample up through the glass
frit from where it can be released through the second valve. After
sampling, the overpressure was quickly released back to 1 atm. By
using the custom made in-/outlet the aqueous powder suspensions
could be continuously purged with pre-mixed gases (Table 2) without
-
2
below 5.1 x 10 bar. We found that formaldehyde is not
produced in the reaction between H and Pd in the
presence of tris, which indicates that OH is not formed
during catalytic decomposition of H on Pd. In contrast, a
significant amount of hydroxyl radicals are scavenged in the
2
O
2
Ÿ
2 2
O
Ÿ
2 2 2
same reaction on ZrO . The absence of OH during H O
decomposition on Pd indicates that the reaction mechanism,
which is proposed in current literature, needs to be revised.
2
Compared to previous results, where ZrO was shown to
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201901128
FULL PAPER
air intrusion during sampling. In the absence of H
2
, the samples
Autoclave experiments
2
For the autoclave experiments 2 mg of Pd, or 2.33 g of ZrO were
were purged throughout the experiment using N
2
.
suspended in 100 ml 20 mM Tris. As for the previous experiments
the pH was adjusted to 7.5 with HCl before the reaction vessel was
transferred into an aerated autoclave. The autoclave was then
Table 2. Overview of gases used during the experiments
pressurized with H
0 bar H . Throughout the experiments the powder suspensions
were stirred with a paddle stirrer to avoid sedimentation. Samples
2 2
to give final partial pressures of 0.2 bar O and
H
2
partial pressure / %
H
2
partial pressure / bar
Purity / %
4
2
1
5
0
0
0
00
1.01
≥ 99.995
were taken during two weeks and the concentrations of H
2
O
2
as well
-
-
-
2
3
4
Ÿ
5.1 x 10
5.1 x 10
5.1 x 10
0.00
as OH were monitored.
.5
.05
(100 % N
2
)
≥ 99.999
Acknowledgements
The Swedish Nuclear and Fuel Waste Management Company
SKB) is gratefully acknowledged for financial support.
(
Uranyl reduction by H
To measure the kinetics of UO
partial pressure below 1 bar, different gas mixtures were used (table
2
2
+
2
reduction depending on the H
2
Keywords: ε-particles • heterogeneous catalysis • hydrogen •
-
2
3
). 8 mg of Pd were suspended in HCO solution and the mixture
palladium • spent nuclear fuel
was deoxygenized through purging for 30 min in the custom made
setup as described above. The gas flow was kept constant at 0.9 l
[
1]
IAEA, Energy, electricity and nuclear power estimates for the period
-
1
min and the sample was stirred using a magnetic stirrer throughout
up to 2050, IAEA-RDS-1/38, 2018.
the experiment.
After the 30 min equilibration time, 0.56 ml 5 mM uranyl nitrate were
[
2]
B. Faybishenko, J. Birkholzer, D. Sassani, P. Swift, Eds.,
International Approaches for Deep Geologic Disposal of Nuclear
Waste: Geological Challenges in Radioactive Waste Isolation, Fifth
World Wide Review, LBNL Report 1006984, Lawrence Berkeley
National Laboratory, Sandia National Laboratories, University of
California, Berkeley, USA, 2016, https://doi.org/10.2172/1353043.
injected into the reaction vessel through the septum using
hypodermic needle to reach a total volume of 280 ml, a HCO
a
-
3
concentration of 2 mM and a uranyl concentration of 0.01 mM. U(VI)
concentrations were measured over time.
[
[
[
3]
4]
5]
A.S. Kertes, R. Guillaumont, Nuclear and Chemical Waste
Menagement 1985, 5, 215-219.
H
2
O
2
reactivity as a function of the H
To determine the reactivity of H
concentrations were measured over time as a function of H
2
partial pressure
towards H on Pd, H
partial
2
O
2
2
2 2
O
2
J. W. T. Spinks, R. J. Woods, An Introduction to Radiation Chemistry,
pressure below 1 bar. 8 mg of Pd powder were suspended in water
and the mixture was left to deoxygenize through purging for 30 min.
third ed., John Wiley and Sons Inc., New York. 1990.
After this equilibration time H
into the reaction vessel using a hypodermic needle to a total volume
of 280 ml and a H concentration of 0.22 mM. The gas flow was
kept constant at 0.9 l min and the sample was stirred using a
2 2
O was injected through the septum
E. Ekeroth, O. Roth, M. Jonsson, J. Nucl. Mater. 2006, 355, 38-46.
2
O
2
[6]
A. Barreiro Fidalgo, Y. Kumagai, M. Jonsson, J. Coord. Chem. 2018,
71, 1799-1807.
-
1
magnetic stirrer throughout the experiment.
were measured over time.
H
2
O
2
concentrations
[
[
[
7]
8]
9]
J. Bruno, R.C. Ewing, Elements 2006, 2, 343-349.
H. Kleykamp, J. Nucl. Mater. 1985, 131, 221-246.
2 2
Catalytic decomposition of H O
H. Kleykamp, J.O. Paschoal, R. Pejsa, F. Thümmler, J. Nucl. Mater.
1985, 130, 426-433.
Ÿ
For experiments where the production of surface bound OH was
measured simultaneously to H
powder were added to a 20 mM tris solution and the pH was
adjusted to 7.5 using HCl. The mixture was purged with N to
deoxygenize for 30 min and H was added to a concentration of 5
mM and a total volume of 25 ml. As a comparison to the Pd case,
the same experiment was performed using 63 mg of ZrO powder
instead of Pd. The background consumption of was
determined in the absence of any powder. All samples were stirred
using a magnetic stirrer, purged with N and protected from light
as well as CH O concentrations
2 2
O decomposition, 30 mg of Pd
[10]
D. Cui, J. Low, C.J. Sjöstedt, K. Spahiu, Radiochim. Acta 2004, 92,
5
51-555.
2
2
O
2
[
[
[
11]
12]
13]
S. Nilsson, M. Jonsson, J. Nucl. Mater. 2008, 372, 160-163.
S. Nilsson, M. Jonsson, J. Nucl. Mater. 2008, 374, 290-292.
2
2 2
H O
M. Jonsson, F. Nielsen, O. Roth, E. Ekeroth, S. Nilsson, M.M.
2
Hossain, Environ. Sci. Technol. 2007, 41, 7087-7093.
throughout the experiment.
were measured over time.
H
2
O
2
2
[14]
M. Trummer, M. Jonsson, J. Nucl. Mater. 2010, 396, 163-169.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201901128
FULL PAPER
[
[
[
[
15]
16]
17]
18]
M.E. Broczkowski, J.J. Noël, D.W. Shoesmith, J. Nucl. Mater. 2005,
46, 16-23.
[27]
H. Henkel, W. Weber, US Patent No. 1108752A, 1914.
3
[
28]
A. Barreiro Fidalgo, B. Dahlgren, T. Brinck, M. Jonsson, J. Phys.
Chem. C 2016, 120, 1609-1614.
J. Bruno, I. Grenthe, B. Lagerman, Mat. Res. Soc. Symp. Proc. 1985,
0, 299-308.
5
J. Bruno, I. Casas, B. Lagerman, M. Munoz, Mat. Res. Soc. Symp.
Proc. 1986, 84, 152-160.
[29]
M. Jonsson in Recent Trends in Radiation Chemistry (Eds.: J. F.
Wishart, B. S. M. Rao), World Scientific Publishing Co. Pte. Ltd.,
Singapore, 2010, pp. 301-323.
J.P. Simpson, R. Schenk, B. Knecht, Mat. Res. Soc. Symp. Proc.
1
985, 50, 429-436.
[
30]
P. Albers, J. Pietsch, S.F. Parker, J. Mol. Catal. Chem. 2001, 173,
2
75-286.
[
[
19]
20]
B. Bonin, M. Colin, A. Dutfoy, J. Nucl. Mater. 2000, 281, 1-14.
[
[
31]
32]
M. Yang, M. Jonsson, J. Phys. Chem. C 2014, 118, 7971-7979.
J. Li, A. Staykov, T. Ishihara, K. Yoshizawa, J. Phys. Chem. C. 2011,
1
15, 7392-7398.
T. A. Pospelova, N. I. Kobozev, E. N. Eremin, Rus. J. Phys. Chem.
Trans. 1961, 35, 299-305.
[
21]
A. Plauck, E. E. Stangland, J. A. Dumesic, M. Mavrikakis, Proc. Natl.
Acad. Sci. U.S.A. 2016, 113, E1973-E1982.
[
[
33]
34]
T. A. Pospelova, N. I. Kobozev, Rus. J. Phys. Chem. Trans. 1961,
3
5, 535-542.
[
[
22]
23]
N.M. Wilson, D.W. Flahety, J. Am. Chem. Soc. 2016, 138, 574-586.
N. Liu, L. Wu, Z. Qin, D.W. Shoesmith, Environ. Sci. Technol. 2016,
T. A. Pospelova, N. I. Kobozev, Rus. J. Phys. Chem. Trans. 1961,
35, 1192-1197.
5
0, 12348-12355.
[
[
35]
36]
S. B. Savvin, Talanta 1961, 8, 637-685.
[
[
[
24]
25]
26]
C.M. Lousada, M. Jonsson, J. Phys. Chem. C 2010, 114, 11202-
1208.
1
A. O. Allen, T. W. Davis, G.V. Elmore, J. A. Ghormley B. M. Haines,
C. J. Hochanadel, ORNL Publications 130, Oak Ridge National
Laboratory, Tennessee, USA, 1949.
C.M. Lousada, A.J. Johansson, T. Brink, M. Jonsson, J. Phys. Chem.
C 2012, 116, 9533-9543.
[
37]
Q. Li, P. Sritharathikhun, S. Motomizu, Anal. Sci. 2007, 23, 413-417.
M. Yang, M. Jonsson, J. Mol. Catal. Chem. 2015, 400, 49-55.
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.201901128
FULL PAPER
Entry for the Table of Contents
FULL PAPER
Annika Carolin Maier*, Mats Jonsson
Page No. – Page No.
Pd Catalyzed Surface Reactions
involving
H
2
of Importance in
2
+
Pd catalyzes UO
decomposition of H
While Pd catalyzes H
that was proposed for ZrO
2
and H
on Pd has a different mechanism than on metal oxides.
formation from its elements, the same catalytic activity
could not be confirmed.
2 2 2
O reduction even at low H partial pressures. Catalytic
Radiation Induced Dissolution of
Spent Nuclear Fuel
2
O
2
2
O
2
2
This article is protected by copyright. All rights reserved.