Reductive photo-chemical separation of the hexafluorides of uranium and molybdenum

Article abstract of DOI:10.1016/j.jfluchem.2020.109655

Two new techniques are described for the separation of molybdenum hexafluoride (MoF₆) from uranium hexafluoride (UF₆). Both separation techniques utilize the differences displayed by the hexafluorides in their ability to absorb light in the near UV region. Because UF₆ absorbs light in the near UV region and MoF₆ does not, this observation was used to selectively reduce UF₆ to uranium pentafluoride (UF₅) through irradiation with 395 nm light in the presence of a suitable reducing agent. Two reducing agents were chosen for this study: gaseous, liquid, or super-critical carbon monoxide (CO) and liquid sulfur dioxide (SO₂). Since MoF₆ is not reduced under the reaction conditions described here, it may be removed via distillation from the uranium-containing sample after complete reduction of UF₆ to solid UF₅. The molybdenum- and uranium-containing samples were measured for purity through elemental analysis using microwave plasma atomic emission spectroscopy (MP-AES). Elemental analysis showed more than 98.8 % of the Mo had been removed from the U-containing samples. Further analyses of the samples were performed by X-ray powder diffraction and IR spectroscopy.

Full text of DOI:10.1016/j.jfluchem.2020.109655

Journal of Fluorine Chemistry 240 (2020) 109655
Contents lists available at ScienceDirect
Journal of Fluorine Chemistry
journal homepage: www.elsevier.com/locate/fluor
Reductive photo-chemical separation of the hexafluorides of uranium
and molybdenum
,
,
,
,
Riane E. Stenea b, Tobias Chemnitza b, Winfried Petryb *, Florian Krausa *
a
¨
Anorganische Chemie, Phillips-Universitat Marburg, Hans-Meerwein-Straße 4, 35032 Marburg, Germany
b
¨
Neutron Research Source Heinz Maier-Leibitz (FRM II), Technische Universitat München, Lichtenbergstraße 1, 85748 Garching, Germany
A R T I C L E I N F O
A B S T R A C T
Keywords:
Two new techniques are described for the separation of molybdenum hexafluoride (MoF6) from uranium
hexafluoride (UF6). Both separation techniques utilize the differences displayed by the hexafluorides in their
ability to absorb light in the near UV region. Because UF6 absorbs light in the near UV region and MoF6 does not,
this observation was used to selectively reduce UF6 to uranium pentafluoride (UF5) through irradiation with 395
nm light in the presence of a suitable reducing agent. Two reducing agents were chosen for this study: gaseous,
liquid, or super-critical carbon monoxide (CO) and liquid sulfur dioxide (SO2). Since MoF6 is not reduced under
the reaction conditions described here, it may be removed via distillation from the uranium-containing sample
after complete reduction of UF6 to solid UF5. The molybdenum- and uranium-containing samples were measured
for purity through elemental analysis using microwave plasma atomic emission spectroscopy (MP-AES).
Elemental analysis showed more than 98.8 % of the Mo had been removed from the U-containing samples.
Further analyses of the samples were performed by X-ray powder diffraction and IR spectroscopy.
Separation
Molybdenum
Uranium
Hexafluoride
Sulfur dioxide
Carbon monoxide
UV radiation
1. Introduction
expected to decrease by at least a factor of two [7]. The decreased yield
of 99Mo per target is a direct effect of the lower 235U content. Conse-
Currently, 99mTc is used in over 80 % of all nuclear diagnostic
techniques [1–3]. Globally, over 30 million applications of 99mTc are
utilized each year [1,2]. Today, the majority of the world’s supply of
99mTc is produced by the irradiation of uranium targets in nuclear re-
actors [1–4]. This irradiation process produces the isotope 99Mo, which
decays with a half-life of 66 h to 99mTc, see Eq. (1) [3].
quently, a larger number of targets will need to be irradiated in order to
meet the worldwide demand of 99Mo.
After irradiation of a target at a nuclear facility, the target is sent to a
nearby processing facility. At these facilities, a wet-chemical process is
used to extract and purify 99Mo. During extraction, the entire target –
including the aluminum cladding – is dissolved in an alkaline solution
containing either NaOH or KOH [8]. After this initial dissolution and
extraction procedure, the dissolved molybdenum can be further purified
by passing the solution through a series of anion/cation exchange and
aluminum oxide columns, and then by sublimation [9].
235U (ntherm al,f)
⟶
99Mo 66⟶hours 99m Tc
(1)
βꢀ
For reasons of production efficiency, typical targets for the produc-
tion of 99Mo were made with highly enriched uranium (HEU, 93 %
enrichment) and composed of a uranium aluminum alloy (UAlx). How-
ever, the current political consensus requests that the usage of HEU for
civilian purposes be substituted by low enriched uranium (LEU,
enrichment 19.75 %) wherever technically and economically feasible
[5,6]. As part of this consensus, the affected irradiation facilities and
target-production facilities have spent the past several years imple-
menting changes that will allow for the production and irradiation of
LEU targets for 99Mo fabrication. However, the yield of 99Mo per irra-
diation cycle, depending on the irradiation facility and target used, is
Because the extraction and purification of 99Mo relies on a wet-
chemical process, and because a larger number of targets will need to
be irradiated to meet the worldwide demand for 99Mo, more liquid
radioactive waste will be produced during the extraction and purifica-
tion processes. In regard to radioactive waste production, Lee and co-
workers predict that the production of 10,000 6-day Ci/week using LEU
targets will lead to the production of 15,000 L of intermediate level
wastes per year (and after cementation this volume is expected to in-
crease to 375,000 L) [10]. This increase in intermediate level wastes
* Corresponding authors.
E-mail addresses: winfried.petry@frm2.tum.de (W. Petry), florian.kraus@chemie.uni-marburg.de (F. Kraus).
https://doi.org/10.1016/j.jfluchem.2020.109655
Received 11 September 2020; Received in revised form 30 September 2020; Accepted 30 September 2020
Available online 5 October 2020
0022-1139/© 2020 Elsevier B.V. All rights reserved.

R.E. Stene et al.
Journal of Fluorine Chemistry 240 (2020) 109655
translates to a 200 % increase in radioactive waste production [10].
It is clear from these alarming numbers that new approaches for the
production, extraction and purification of 99Mo from uranium should be
considered in order to decrease the amount of nuclear waste produced
from this industry. Recently, our group has developed a cylindrical LEU
target consisting of a uranium foil encapsulated between two coaxial
aluminum cladding cylinders, which can be mechanically de-cladded
before target processing [11]. This design allows separate processing
of the irradiated target and cladding material during the first steps of the
99Mo extraction process [11]. Since this target may be mechanically
de-cladded, and allows for processing of the pure irradiated foil,
different approaches to processing the bare uranium foil may also be
realized, such as fluorinating the foil to obtain mixtures of uranium
hexafluoride (UF6) and molybdenum hexafluoride (MoF6) (along with
other potentially volatile fission product fluorides).
solids and MoF6 is a highly volatile liquid at room temperature (b.p.
34 ◦C), this technique could be used to separate MoF6 from UF6. In the
present work, a wavelength of 395 nm was chosen for these selective
reduction reactions. Using CO as an example, the separation process is
given by Eqs. (4) and (5).
2UF6 + CO 3⟶95 nm 2UF5 + COF2
(4)
(5)
2MoF6 + CO 3⟶95 nm No reaction
2.2. Separation of MoF6 from UF6 using gaseous carbon monoxide and
near UV light
The separation of MoF6 from UF6 is described by Eqs. (4) and (5). To
perform this separation, 200 mg of UF6 and 200 mg of MoF6 were
condensed at ꢀ 196 ◦C (using liquid nitrogen) into an approximately 30
mL quartz reaction vessel. The hexafluorides were allowed to warm to
room temperature to ensure adequate mixing. After warming, the
mixture was cooled to ꢀ 196 ◦C and the vessel was pressurized with 1 bar
of CO. The vessel was allowed to warm to room temperature and then
irradiated with near ultraviolet light having a wavelength of 395 nm for
about 24 h. After this time, a blue-green product was observed (UF5, see
Fig. 1). The reaction mixture was cooled to ꢀ 196 ◦C, the excess CO and
COF2 product were pumped off, and then 1 bar of fresh CO was added to
the vessel. The reaction mixture was irradiated once more with 395 nm
light for an additional 24 h. During these irradiation cycles, UF6 was
selectively reduced to UF5 while MoF6 remained unreacted. The
reduction of UF6 was considered to be complete after visual indication
that the UF5 formed turned from blue-green to pale green. After com-
plete reduction of UF6 to UF5, the reaction mixture was cooled once
It has previously been suggested that UF6 gas streams can be purified
from contaminants using ultraviolet light and a fluorine radical scav-
enger [12–14]. In the purification technique described by Fields, UF6
was purified from MoF6 using UV light having “sufficient energy to
photodissociate MoF6 and UF6” to MoF5 and UF5 (an exact wavelength
of 254 nm was used). Fields argues that uranium can be purified from
molybdenum using this process because UF5 is more reactive than MoF5
and will, consequently, further react with fluorine radicals/unreacted
molecules of MoF6 to regenerate UF6. However, no data concerning the
purity or the yields of either the UF6 or MoF5 samples is reported in this
work. Because the separation technique described by Fields is hard to
control and is indiscriminate towards the reduction of UF6 and MoF6, it
was suspected that this method would be unsuitable for the separation of
MoF6 from UF6, with the purpose of extracting and purifying MoF6.
Experiments in our laboratory have confirmed this opinion.
Two new techniques for the separation of MoF6 from UF6, both with
natural isotope distribution, using reducing agents and near ultraviolet
radiation having a wavelength of 395 nm, are described in this work.
Two reducing agents were studied, namely, carbon monoxide (CO,
gaseous, liquid and supercritical) and sulfur dioxide (SO2, liquid).
Future research will need to show whether or not the reactions can be
transferred to the separation of 99MoF6 from UF6 after a target has been
irradiated in a nuclear reactor.
◦
more to ꢀ 196 C and the remaining CO and COF2 were pumped off.
Once all CO and COF2 had been pumped away, the reaction vessel was
warmed to room temperature and the MoF6 was removed from the solid
UF5 sample by distillation into a new quartz vessel using liquid nitrogen
cooling. Fig. 1 depicts the laboratory set-up for this separation tech-
nique. Fig. S1 provides a step-by-step schematic of the separation.
Quantitative yields of UF5 and MoF6 were obtained after separation.
This separation technique was repeated three times to establish repro-
ducibility. The separate UF5 and MoF6 samples were then tested for
purity through elemental analysis with microwave plasma atomic
emission spectroscopy (MP-AES). For easier analysis and sample prep-
aration, the MoF6 sample obtained after this separation procedure was
further reduced to MoF5 using CO as a reducing agent and UV light
having a wavelength of 254 nm (Eq. (6) and Fig. 2).
2. Results and discussion
2.1. Selection of a suitable wavelength of UV light
It has long been known that the hexafluorides of U and Mo can be
reduced using UV light and a reducing agent, such as CO, SO2, or H2 [15,
16]. Using CO as an example, Eqs. (2) and (3) illustrate these reduction
reactions.
nm
2MoF6 + CO 2⟶52 2MoF5 + COF2
(6)
The results of the elemental analyses of the UF5 samples obtained
after each separation are given in Table 1. Each sample was measured in
triplicate (Trial 1 (U) – Trial 3 (U)); more information regarding the
analysis of these samples is given in the Material and Methods Section.
Although some amount of molybdenum was present in these samples,
over 99 % of the original molybdenum content was removed using the
separation technique described here. The percentages of uranium
observed in these samples are also given in Table 1. A pure sample of UF5
should contain about 71.48 % U, however, the samples analyzed here
contained 68.61–73.52% U. This difference in percentages is due to
hydrolysis of the UF5 samples during sample preparation and dissolution
for MP-AES measurement, see Materials and Methods Section.
2UF6 + CO UV light of sufficient energy
⟶
2UF5 + COF2
(2)
2MoF6 + CO UV light of sufficient energy
⟶
2MoF5 + COF2
(3)
For these reactions to occur under ambient conditions, careful se-
lection of the wavelength of UV light needs to be considered. If the UV
light chosen does not have sufficient energy to excite the hexafluoride,
then selective photochemical reduction will not occur. It can be seen
from the UV absorption spectrum of UF6 that weak absorption occurs
between 360–400 nm, whereas strong absorption begins around 340 nm
and continues into the far UV range [17]. In the UV absorption spectrum
of MoF6, absorption begins at about 280 nm and extends into the far UV
range [18].
To further characterize the UF5 sample obtained after the CO sepa-
ration procedure described here, an X-ray powder diffraction pattern
was obtained, of which a Rietveld refinement is shown in Fig. 3. This
powder pattern agrees nicely with previously published data for β-UF5
(Fig. S3) [20], but one peak due to an unidentified phase appeared at ca.
31.85◦ 2θ. Details of the Rietveld refinement are reported in Table S1.
Therefore, β-UF5 is essentially obtained as a pure phase, however some
Using these observations, we could demonstrate that MoF6 could be
selectively removed from UF6 by careful selection of the wavelength of
UV light. That is, if UV light having a wavelength longer than about 300
nm is used, UF6 can be selectively reduced to UF5 while MoF6 remains
unaffected. Since both known modifications of UF5 are non-volatile
2

R.E. Stene et al.
Journal of Fluorine Chemistry 240 (2020) 109655
Fig. 1. Laboratory set-up for the separation of
MoF6 from UF6 using CO and near UV light. a)
Mixture of MoF6 (colorless liquid), UF6 (white
solid) and CO (colorless gas) at room tempera-
ture. b) Reaction mixture after 24 h of irradia-
tion with 395 nm light. Blue-green UF5 has
formed, MoF6 has not been reduced. c) Reaction
mixture after the second irradiation cycle (24
h). A mixture of pale green UF5 and MoF6 is
obtained. d) The unreacted MoF6 was removed
from the solid UF5 by distilling the hexafluoride
into a new quartz vessel. Note: There is no/very
little UF6 present in the MoF6 sample because
no solid, white UF6 is visible.
Additionally, an IR spectrum in the region from 400 to 4000 cmꢀ 1
was recorded for the UF5 sample obtained after the separation (Fig. 4).
This IR spectrum agrees nicely with previously reported IR spectra for
β-UF5 [21]; slight evidence of beginning hydrolysis is present.
The elemental analyses results for the MoF5 samples, obtained by the
reduction of MoF6, are given in Table 2. Due to the oily nature of the
MoF5 sample, the entire sample was dissolved in distilled water and then
diluted for elemental analysis; Trial 1 (Mo) – Trial 3 (Mo) were obtained
from the same separation batches as those stemming from the U de-
terminations reported in Table 1. More information regarding the
analysis of these samples is given in the Material and Methods section. It
can be seen from Table 2 that no uranium was detected in these samples.
The detection limit for U on the device used to measure the samples was
approximately 0.5 mg/L, which correlates with a U impurity of less than
three percent. However, since no visual evidence of uranium as white
UF6 or green UF5 was seen in the MoF6 or MoF5 samples, the purity of
these samples is thought to be greater than 97 %. Additionally, the
percentages of Mo in the samples are given in Table 2. A sample con-
taining pure MoF5 should contain 50.26 % Mo, however, the samples
measured in this work have Mo contents ranging from 36.34 to 40.02 %.
These lower percentages are a direct consequence of the hydrolysis of
the MoF5 samples (by releasing HF to produce molybdenum oxides or
oxyfluorides) during sample preparation for MP-AES measurement.
In order to further characterize the molybdenum sample collected
after the CO separation described here, a fourth separation was per-
formed. The MoF6 collected was reduced to MoF5 using the reaction
described by Eq. (6). An oil of MoF5 was obtained that was allowed to
crystallize under inert atmosphere in a glovebox over the course of about
a week. After crystallization, an X-ray powder diffraction pattern of the
sample was obtained. The powder pattern matched that of previously
reported powder patterns of MoF5 (Fig. S4) [22]. The powder pattern
and the Rietveld refinement are shown in Fig. 5. Further details of the
Rietveld refinement are reported in Table S2, and showed MoF5 was
obtained phase-pure.
Fig. 2. Yellow MoF5 obtained after the reduction of the MoF6 sample. a) The
bulk sample inside a quartz tube. b) A section of the sample as seen under a
microscope. No visual traces of pale green UF5 can be seen in this sample, which
speaks towards the purity of the Mo sample and quality of the separation. Note:
MoF5 can readily be obtained as a supercooled liquid at room temperature [19].
Table 1
MP-AES results of the uranium sample (UF5) obtained after the near UV sepa-
ration of MoF6 from UF6 using 395 nm light and gaseous carbon monoxide as a
reducing agent. All trials were done in triplicate.
Trial 1 (U)
Trial 2 (U)
Trial 3 (U)
Mo, mg/L
% Mo
0.16 ± 0.01
0.30 ± 0.01
39.31 ±
2.27
0.16 ± 0.01
0.32 ± 0.01
35.62 ±
0.77
0.13 ± 0.00
0.26 ± 0.00
35.52 ±
0.95
U, mg/L
Additionally, an IR spectrum of the MoF5 sample was obtained. This
IR spectrum matches nicely with previously reported IR spectra of MoF5
[19,22,23], although very slight hydrolysis is obviously present, Fig. 6.
% U
73.52 ±
1.41
68.61 ±
1.70
68.89 ±
1.05
Mo/U molar ratio before
separation
1.70
1.63
1.61
Mo/U molar ratio after separation
% Mo removed
0.01 ± 0.00
99.41 ±
0.02
0.01 ± 0.00
99.30 ±
0.03
0.01 ± 0.00
99.42 ±
0.01
2.3. Separation of MoF6 from UF6 using sulfur dioxide and near UV light
Similar to the gaseous CO separation, MoF6 can be removed from UF6
using near UV light and SO2 as a reductant. During this separation, UF6
is selectively reduced to UF5 through interaction with the 395 nm light
and the SO2 solvent (Eq. (7)). MoF6 remains unaffected by the near UV
light, but does slowly react with SO2 to form MoOF4 (Eq. (8)). However,
amorphous impurities may be present as indicated by the bump in the
background, which is also due to the glass capillary used for the
measurement.
3

R.E. Stene et al.
Journal of Fluorine Chemistry 240 (2020) 109655
Fig. 3. Observed and calculated powder X-ray pattern of β-UF5 after Rietveld refinement. The calculated reflection positions are indicated by the vertical bars below
the pattern. The curve at the bottom represents the difference (Δ) between the observed and the calculated intensities. Rp = 2.01, wRp = 2.71 (not background
corrected R values), S = 1.71. The asterisk indicates the foreign peak.
Table 2
MP-AES results of the molybdenum sample obtained after the UV separation of
MoF6 from UF6 using 395 nm light and gaseous CO as a reducing agent. A single
sample was prepared for each trial.
Trial 1 (Mo)
Trial 2 (Mo)
Trial 3 (Mo)
Mo, mg/L
% Mo
29.14
39.99
n.d.
26.29
36.34
n.d.
28.95
40.02
n.d.
U, mg/L
n.d. = not detected.
the reaction of MoF6 with SO2 is so slow that it plays no role during the
time frame of the separation procedure.
nm
2UF6 + SO2 3⟶95 2UF5 + SO2F2
(7)
(8)
very fast
MoF6 + SO2ve⟶ry slow MoOF4 + SOF2
In this separation, 200 mg of UF6 and 200 mg of MoF6 were
◦
condensed at ꢀ 196 C into an FEP reaction vessel (containing a 1 cm
Teflon-coated stir bar). About six grams of SO2 were condensed onto the
hexafluoride mixture. The mixture was allowed to warm to room tem-
perature and the solution was stirred using a stir plate (note: the reaction
vessel must be able to withstand the vapor pressure of the liquid SO2 at
room temperature). Once the hexafluorides were completely dissolved
in the SO2 solution, the UF6/MoF6/SO2 mixture was irradiated under
constant stirring with 395 nm light for 1 h and 30 min. During irradia-
tion, UF6 was selectively reduced to UF5 while MoF6 remained unreac-
ted and soluble in the SO2 solution, whereas UF5 is insoluble in SO2 and
precipitated from solution. After irradiation, a solution of MoF6 dis-
solved in SO2 was obtained along with precipitated UF5 sitting at the
bottom of the reaction vessel. The MoF6/SO2 reaction mixture was then
distilled to a new FEP reaction vessel using liquid nitrogen cooling. The
molybdenum-containing sample was extracted from the SO2 solution
after one week as white, crystalline MoOF4 (Eq. (8)). Fig. 7 depicts the
laboratory set-up for this separation technique. Fig. S2 provides a step-
by-step schematic for the separation.
Fig. 4. IR spectrum of the UF5 sample obtained after the separation of MoF6
from UF5 using gaseous CO and near UV light. The above spectrum agrees with
previously reported spectra for β-UF5. The light-green overlays indicate very
weak bands arising from slight hydrolysis of the sample that occurred during
sample handling. The noise around 2000 cmꢀ 1 is due to the diamond ATR-IR.
After separation, quantitative yields of UF5 were obtained, however,
yields of only about 60–70 % were obtained for the MoOF4 samples.
These decreased yields can result from a variety of reasons. Most likely,
the MoF6 had not completely reacted with the SO2 solution after a week
of reaction time. Longer reaction time is probably needed for full con-
version of MoF6 to MoOF4. Another reason that the yields were rela-
tively low is due to the fact that MoF6 is soluble in the FEP plastic used
4

R.E. Stene et al.
Journal of Fluorine Chemistry 240 (2020) 109655
Fig. 5. Observed and calculated powder X-ray pattern of MoF5 after Rietveld refinement. The calculated reflection positions are indicated by the vertical bars below
the pattern. The curve at the bottom represents the difference (Δ) between the observed and the calculated intensities. Rp = 3.25, wRp = 4.73 (not background
corrected R values), S = 3.63.
should be possible to extract all the Mo by passing the MoF6/SO2 solu-
tion through an aqueous solution of NaOH or KOH, however, the in-
fluence of the formed SO23ꢀ anion then needs to be investigated. Since
the purpose of this work was to provide techniques for non-aqueous
separations, the aqueous extraction of Mo from SO2 was not tested.
Although such a dissolution process is foreseeable, the primary goal of
the 99Mo industry is making 99Mo/99mTc available to the physicians and
patients in a soluble form, preferably MoO24ꢀ . Moreover, to test the
reproducibility of this separation technique, the separation was repeated
in triplicate. Afterwards, the separate UF5 and MoOF4 samples were
tested for purity by elemental analyses with MP-AES.
The results of the elemental analyses of the UF5 samples obtained
after each separation are given in Table 3. Each sample was measured in
triplicate (Trial 4 (U) – Trail 6 (U)); more information regarding the
analyses of these samples is given in the Material and Methods section.
While some amount of molybdenum was present in these samples, over
98.8 % of the original molybdenum content was removed using the
separation technique described here. The percentage of uranium
observed in these samples is also given in Table 3. As previously stated, a
pure sample of UF5 should contain about 71.48 % U, however, the
samples analyzed here contained between 62.82–73.01% U. This dif-
ference in percentages is due to hydrolysis of the UF5 samples during
sample preparation. Additionally, it is suspected that some sulfur-
containing species are present in these samples, as discussed below.
To further characterize the UF5 samples collected after the SO2
separations, an X-ray powder diffraction pattern was obtained. This
powder pattern was very similar to the pattern observed for the UF5
samples obtained after the CO separations reported here (Fig. S5) and
agrees with previously reported patterns for β-UF5 [20], however, evi-
dence for a small amount of an unidentified crystalline impurity is
present. The results of a Rietveld refinement performed on this X-ray
powder diffraction pattern is provided in Fig. 8 and Table S3.
Additionally, IR spectra of the UF5 samples, obtained in the region
from 400 to 4000 cmꢀ 1, showed some impurities were present in these
samples (Fig. 9). In the work of Halstead and coworkers, where the
Fig. 6. Top: IR spectrum of the MoF5 obtained after the separation of MoF6
synthesis of UF5 is discussed, the authors acknowledge that UF5 can be
from UF6 using CO and near UV light. The MoF6 was further reduced to MoF5
synthesized by reduction of UF6 with SO2, but with product purities less
than those obtained with CO reductions [15]. To exclude the possibility
of these foreign IR bands arising from molybdenum-containing species,
the experiment described here was repeated, but only with UF6, that is,
no MoF6 was added. The IR spectrum obtained from the reduction of UF6
with SO2 and 395 nm light looks nearly identical to the spectra obtained
for the UF5 samples collected after the SO2 separations; Fig. 9 shows a
comparison. Therefore, it was concluded that these foreign bands arise
using gaseous CO and 254 nm light. Bottom: IR spectrum of pure MoF5. The
light-green overlays indicate bands arising from slight hydrolysis of the sample
that occurred during sample handling. The noise around 2000 cmꢀ 1 is due to
the diamond ATR-IR.
for the reaction vessels. It is suggested that glass or quartz vessels be
used for future applications of this separation technique in order to in-
crease the yield of the molybdenum-containing sample. Alternatively, it
5

R.E. Stene et al.
Journal of Fluorine Chemistry 240 (2020) 109655
Fig. 7. Laboratory set-up for the separation of
MoF6 from UF6 using SO2 and near UV light. a)
UF6 and MoF6 dissolved in an SO2 solution.
Note: Solutions of UF6 in SO2 are slightly yel-
low. b) The solution after irradiation with 395
nm light. Pale green UF5 has formed and has
precipitated from the SO2 solution while MoF6
remains unreacted and dissolved in SO2. c) The
MoF6/SO2 solution is removed from the solid
UF5 via distillation. Note: This solution is
colorless, indicating that little or no UF6 is
present in this solution.
as intrinsic impurities in the synthesis of UF5 by reduction of UF6 in SO2
and not from molybdenum-containing species. Unfortunately, the
identity of the impurities could not be determined, although it is sus-
pected that they are sulfur-containing compounds.
Table 3
MP-AES results of the uranium sample (UF5) obtained after the SO2/UV sepa-
ration of MoF6 from UF6 using 395 nm light and SO2 as a solvent. All trials were
measured in triplicates.
The results of the elemental analyses of the MoOF4 samples are given
in Table 4. Each sample was measured in triplicate. Trial 4 (Mo) – Trial 6
(Mo) were obtained from the same separation batches as those stemming
from the U determinations reported in Table 3. More information
regarding the analyses of these samples is given in the Material and
Methods section. It can be seen from Table 4 that no uranium was
detected in these samples. As previously discussed, the detection limit
for U on the device used to measure the samples was approximately 0.5
mg/L, which correlates to a U impurity of less than 1.5 % in these
samples. However, because there was no visual evidence of uranium (as
pale green UF5) present in the white MoOF4 sample, the purities of these
samples are thought to be greater than 98.5 %. Additionally, the
Trial 4 (U)
Trial 5 (U)
Trial 6 (U)
mg Mo / L
% Mo
0.08 ± 0.00
0.15 ± 0.01
37.44 ±
0.25
0.28 ± 0.03
0.50 ± 0.02
34.78 ±
2.54
0.10 ± 0.01
0.17 ± 0.00
36.22 ±
3.77
mg U / L
% U
73.01 ±
2.35
62.82 ±
0.42
63.11 ±
0.43
Mo/U molar ratio before
separation
1.64
1.68
1.65
Mo/U molar ratio after separation
% Mo removed
0.01 ± 0.00
99.69 ±
0.02
0.02 ± 0.00
98.81 ±
0.04
0.01 ± 0.00
99.58 ±
0.01
Fig. 8. Observed and calculated powder X-ray pattern of β-UF5 after Rietveld refinement. The calculated reflection positions are indicated by the vertical bars below
the pattern. The curve at the bottom represents the difference (Δ) between the observed and the calculated intensities. Rp = 2.14, wRp = 2.83 (not background
corrected R values), S = 1.69. The asterisks indicate foreign peaks.
6

R.E. Stene et al.
Journal of Fluorine Chemistry 240 (2020) 109655
this X-ray powder diffraction pattern is reported in Fig. 10 and Table S4
and suggests that MoOF4 has been obtained phase-pure, however the
presence of amorphous impurities cannot be excluded.
Additionally, an IR spectrum of the MoOF4 obtained in this study was
collected in the region 400–4000 cmꢀ 1. A comparison with a previous IR
spectrum of pure MoOF4 synthesized in our laboratory is provided in
Fig. 11.This IR spectrum shows the presence of MoOF4 along with some
hydrolysis species or perhaps some sulfur-containing impurities [24].
2.4. Separation of MoF6 from UF6 using liquid CO or supercritical CO
and near UV light
In the work described here, a typical separation of MoF6 from MoF6/
UF6 mixtures using gaseous CO and near UV light took about two days,
whereas the same separation with SO2 took an hour and 30 min. The
shorter reaction time needed for the SO2 separation is most likely a
consequence of the higher concentration of liquid reducing agent in
comparison with the gaseous CO experiment. It was then a question of
curiosity if shorter reaction times (i.e. quicker separations) could be
achieved if the separation of MoF6 from UF6 was performed using su-
percritical CO or liquid CO and near UV light. The results from these
experiments will be briefly discussed.
Because CO has a critical pressure of 34.98 bar, all experiments using
its supercritical state were performed in a suitable high-pressure
container. This container was made from stainless steel AISI 316 L and
equipped with a 5 mm thick sapphire window as well as a liquid
nitrogen-cooled cold finger. Additionally, this vessel was connected to a
storage cylinder that contained supercritical CO. A mixture consisting of
roughly 17 mg MoF6 and 42 mg UF6 was condensed into this high-
pressure container by cooling the cold finger. The exact masses of
each hexafluoride are given in the Materials and Methods section.
Subsequently, 8.7 mmol CO having a pressure of 56 bar was added to the
reaction vessel. The high-pressure container was allowed to warm to
room temperature and then irradiated with near ultraviolet light at a
wavelength of 395 nm for 60 min, whereupon the temperature rose to
ca. 33 ◦C. During the course of the reaction, the formation of filaments of
a light-blue solid were observed, originating from the cold finger. These
became denser over time until finally covering the entire sapphire
window. Analysis of these filaments using X-ray powder diffraction
suggest this light-blue solid to be β-UF5. The course of the reaction is
shown in Fig. 12.
Fig. 9. Top) IR spectrum of a UF5 sample obtained after the separation of MoF6
from UF5 using SO2 and near UV light. Bottom) IR spectrum of a UF5 sample
obtained after the reduction of UF6 with SO2 and near UV light (in the absence
of MoF6). The light-green overlays indicate bands arising from impurities
intrinsic to the synthesis of UF5 in SO2 solution. These impurities are likely
sulfur-containing species. The noise around 2000 cmꢀ 1 is due to the diamond
ATR-IR.
The highly-volatile portion of the reaction products from the first
trial was evaluated using gas phase IR spectroscopy, Fig. 13. As ex-
pected, the IR spectrum showed the presence of unreacted CO as well as
COF2 [25]. In addition, it also showed the presence of oxalyl fluoride
(COF)2 [26], and traces of MoF6 [27,28].
Table 4
MP-AES results of the molybdenum sample (MoOF4) obtained after the SO2/UV
separation of MoF6 from UF6 using 395 nm light and SO2 as a solvent. All trials
were measured in triplicate.
For analysis of separation efficiency, the high-pressure container was
allowed to warm to room temperature and the volatile components were
distilled into a separate FEP vessel at liquid nitrogen temperature.
Because both CO and COF2 possess significant vapor pressures, even at
ꢀ 196 ◦C, they could be subsequently evaporated [29–32]. The
remaining volatile components, namely MoF6 and unreacted UF6, were
dissolved and analyzed for Mo and U content using MP-AES. The
non-volatile solid remaining in the high-pressure container, mostly UF5,
was dissolved in dilute HNO3 and also analyzed for Mo and U content by
MP-AES. Additional information on the collection of MP-AES data as
well as the IR spectrum are given in the Material and Methods section.
Table 5 shows the original composition of the UF6/MoF6 mixture as well
as the composition of the respective samples after separation. The
amount of Mo found in the high-pressure container amounts to about 1
% of the original amount, highlighting the observation that MoF6 is not
affected by the irradiation at 395 nm, even under high-pressure reducing
atmospheres. The amount of uranium found in the volatile section can
be attributed to unreacted UF6. The sapphire window is slowly covered
with a layer of UF5 during the separation procedure, which likely blocks
a considerable amount of radiation, thereby preventing a complete
Trial 4 (Mo)
Trial 5 (Mo)
Trial 6 (Mo)
Mo, mg/L
% Mo
54.24 ± 2.55
54.65 ± 1.23
n.d.
55.90 ± 3.71
50.59 ± 1.33
n.d.
55.92 ± 1.11
49.96 ± 0.83
n.d.
U, mg/L
n.d. = not detected.
percentages of Mo present in the samples are given in Table 4. A sample
containing pure MoOF4 should contain 51.05 % Mo, however, the
samples measured in this work have Mo contents between
49.96–54.65% Mo. The differences in these percentages in comparison
with the theoretical value are a direct consequence of some hydrolysis
that occurred during sample preparation.
To further characterize the MoOF4 sample obtained after the SO2
separation procedure described here, an X-ray powder diffraction
pattern was obtained. This powder pattern agrees with previously ob-
tained patterns of MoOF4 measured in our laboratory although the
crystallinity of both samples is quite low [24]. A comparison of the
powder pattern obtained here with a previously obtained pattern is
provided in Fig. S6. The results of a Rietveld refinement performed on
7

R.E. Stene et al.
Journal of Fluorine Chemistry 240 (2020) 109655
Fig. 10. Observed and calculated powder X-ray pattern of β-UF5 after Rietveld refinement. The calculated reflection positions are indicated by the vertical bars below
the pattern. The curve at the bottom represents the difference (Δ) between the observed and the calculated intensities. Rp = 3.93, wRp = 4.99 (not background
corrected R values), S = 1.46.
consisting of 18 mg MoF6 and 36 mg UF6 was condensed into an FEP
cold trap at ꢀ 196 ◦C using liquid nitrogen. Dried CO was carefully
condensed onto the hexafluorides, ensuring complete coverage, and the
cold trap subsequently taken away. Irradiation was performed for 30
min at 395 nm. The apparatus was cooled during the entire irradiation
period to ꢀ 196 ◦C. After irradiation, the volatile species were pumped
off with continuous liquid nitrogen cooling and the cold trap was then
allowed to warm to room temperature. Optical inspection did not reveal
any signs of residues such as UF5. Subsequent analysis of the remaining
volatile species was performed using gas phase IR spectroscopy
(Fig. 14).
The evaluation of the IR spectra of the mixture prior to and after
separation yields the masses listed in Table 6.
The results of the two trials are not entirely consistent. For both
trials, the major amount of UF6 could be recovered and remained
unreacted. However, the first trial showed traces of COF2 after the
separation in addition to the bands of CO in the IR spectrum, indicating
some amount of reduction of UF6. The second trial also shows bands
belonging to CO, however, the bands belonging to COF2 are absent. This
observation is in accordance with a significantly lower recovery of UF6
for the first trial, whereas the recovered amount of MoF6 is identical for
both trials. Further investigations are required to resolve this contra-
diction. Nevertheless, it may be stated that the reaction rate with liquid
CO is significantly reduced with respect to supercritical CO due to the
low temperature.
2.5. Brief comments on the reactivities of fluorides under discussion
In the separation of UF6 from MoF6 described by Fields, it is stated
that UF6 may be separated from MoF6 on the basis of the reactivities of
their corresponding pentafluorides [14]. Fields argues that UF5 is more
reactive than MoF5. Therefore, by reducing both MoF6 and UF6 to their
corresponding pentafluorides, UF5 may react further with any excess
MoF6 and fluorine radicals to reproduce UF6, thus allowing for its sep-
aration from solid MoF5. However, as is evident from the work described
Fig. 11. Top) IR spectrum of a MoOF4 sample obtained from the SO2 separa-
here, UF5 does not react with MoF6 under the conditions used in the
tion procedure. Bottom) Previously measured IR spectrum of MoOF4 obtained
present study.
in our laboratory. The light-green overlays indicate bands arising from slight
hydrolysis of the sample that occurred during sample handling. The noise
around 2000 cmꢀ 1 is due to the diamond ATR-IR.
Considering the crystal structures of UF5 and MoF5, the most reac-
tive, volatile, and soluble pentafluoride should be MoF5, because its
solid-state structure consists of isolated Mo4F20 molecules [22,33]. UF5,
on the other hand, has two polymorphs –
α-UF5 (high-temperature
reduction UF6 to UF5. This circumstance may be countered by an opti-
mized design of the high-pressure container.
modification) and β-UF5 (low-temperature modification). It is antici-
pated that -UF5 will be more reactive, volatile and soluble than β-UF5
because its solid-state structure consists of one-dimensional, infinite
α
In order to investigate the separation using liquid CO, a mixture
8

R.E. Stene et al.
Journal of Fluorine Chemistry 240 (2020) 109655
Fig. 12. View into the high-pressure container through the sapphire window. From left to right: a) Mixture of UF6 and MoF6 condensed on the cold finger. b) After
venting the chamber with supercritical CO. c) After irradiation for 7 min. d) The end of irradiation after 60 min.
Table 5
MP-AES results of the non-volatile contents remaining in the high-pressure
container (U side), and the volatile component collected in a separate FEP
vessel (Mo side) after the second supercritical CO separation trial. U and Mo
masses prior to separation were determined from the IR spectrum.
high-pressure container (U
side)
volatile section (Mo
side)
mg Mo
0.08 ± 0.06
1.06 ± 0.8
7.35 ± 0.06
98.00 ± 4.01
2.99 ± 0.60
11.16 ± 2.24
at% Mo
mg U
20.96 ± 0.63
78.20 ± 2.51
0.69 ± 0.03
at% U
Mo/U molar ratio before
separation
Mo/U ratio after separation
at% Mo removed
6.10 ± 1.22
98.93 ± 0.8
0.009 ± 0.007
ꢀ
3. Conclusions
Two separation techniques for the extraction and purification of
MoF6 from UF6 have been described in this study. Both separation
techniques utilize differences in the reactivities of the hexafluorides
under irradiation with near UV light in the presence of reducing agents.
UF6 absorbs light in the near UV region, whereas MoF6 does not. Using
this difference, UF6 was selectively reduced in the presence of either
gaseous or supercritical CO or liquid SO2 to solid UF5, while MoF6 did
not react. After complete reduction of UF6 to UF5, the volatile MoF6-
containing sample could be separated by distillation. In most cases, over
99 % of the Mo was removed from the U-containing sample.
These separation techniques provide fast and effective methods for
the extraction and purification of natural Mo from natural U. Because
these techniques do not rely on sample dissolution in aqueous solutions,
the separations could be achieved with minimum amounts of liquid
waste production. However, future work on this project must determine
if radiolysis plays any role in these separation techniques when an
irradiated U target containing 99Mo is used, and whether other fission
product fluorides participate in or interfere with the reactions described
here.
Fig. 13. Top) IR spectrum of the volatile section of the reaction products
containing CO, (COF)2 and COF2. Middle) Reference IR spectrum of (COF)2
[26]. Bottom) Reference spectrum of COF2 [25].
Since the gaseous CO separation described here lead to quantitative
yields of high-purity MoF6 samples, it is envisioned that this separation
procedure may also be suitable for transfer to industry. The chemistry
and techniques described here were performed at the laboratory scale,
but this separation procedure should easily translate to larger-scale
operations. In fact, if large enough sample containers, more than two
to three equivalents of CO, and multiple or more powerful LED lights
having a 395 nm wavelength were used, then it is anticipated that this
separation technique could, in fact, be performed within a day. How-
ever, the kinetics of the process must be investigated in the future. In
addition, before transfer to the 99Mo industry can be realized, effects due
to high radioactivity must be thoroughly studied.
strands of corner-sharing octahedra [34]. However, in the work
described here, only β-UF5 is formed, as evidenced by powder X-ray
diffraction. Consequently, β-UF5 should be less reactive than both α-UF5
and MoF5 because its solid-state structure consists of
a
three-dimensional, infinite network [35]. This supports the observation
described in this work that UF5 does not react with MoF6 to produce
MoF5 and UF6.
Because Fields does not discuss reaction yields, or give any elemental
analysis, it is hard to compare the quality of his separation reactions
with those described here. As discussed in the introduction, our labo-
ratory has failed to reproduce the separations published by Fields.
9

R.E. Stene et al.
Journal of Fluorine Chemistry 240 (2020) 109655
4.2. Techniques and procedures for the separation of MoF6 from UF6
using carbon monoxide and near UV light
Because all trials performed to demonstrate this separation tech-
nique are similar to each other, only Trial 1 will be discussed in detail
here. For Trial 2 – Trial 3 and Trial 7 (performed to obtain analyses on
the MoF5 samples), only the amounts of starting material, yields, and
sample used for MP-AES analysis will be discussed.
Trial 1: A 30 mL, homemade quartz reaction vessel was attached to a
glass-to-metal adaptor using pizein plastic to provide a vacuum-tight
seal. Using this adaptor, the quartz vessel was attached to a stainless-
steel valve, which was then attached to a stainless-steel vacuum line.
Afterwards, the quartz reaction vessel was flame dried 5 times using a
Bunsen burner (propane gas). After adequate drying of the quartz vessel,
197.12 mg (0.56 mmol) UF6 and 200.28 mg (0.95 mmol) MoF6 (about a
1:1 wt mixture) were distilled into the quartz reaction vessel using liquid
nitrogen cooling. The hexafluoride mixture was allowed to warm to
room temperature to ensure adequate mixing. Afterwards, the hexaflu-
oride mixture was cooled with liquid nitrogen and then 849 mbar (1.05
mmol) CO was added to the reaction vessel. The UF6/MoF6/CO mixture
was allowed to warm to room temperature. Afterwards, the mixture was
irradiated with a 395 nm LED light for 24 h. After the first irradiation
cycle, the mixture was cooled with liquid nitrogen and the CO/COF2
mixture was pumped off under dynamic vacuum conditions. Another
853 mbar (1.05 mmol) CO was added to the reaction vessel. The mixture
was then irradiated with a 395 nm LED light for another 24 h. After
irradiation, the reaction mixture was cooled using liquid nitrogen and
the CO/CO2 was pumped off under dynamic vacuum conditions.
Once all CO/CO2 had been pumped off, the UF5/MoF6 reaction
mixture was allowed to warm and the MoF6 was removed from the UF5
sample via distillation to a 20 mL quartz vessel. This 20 mL quartz vessel
was previously flame dried 5 times with a Bunsen burner. The MoF6
sample was then cooled with liquid nitrogen and 913 mbar (7.96 mmol)
CO were added to the vessel. The MoF6/CO mixture was irradiated with
395 nm light for 2 h. During this time, no pale green UF5 had formed.
Afterwards, the mixture was moved to a homemade UV reactor and
irradiated with 254 nm light for 18 h in order to reduce the MoF6 to
MoF5. The authors would also like to note that, if a larger quartz vessel
was used, then this separation technique could be performed in a one-
step process, i.e., “fresh” CO would not have to be added after a day
of irradiation. In addition, with the use of a larger reaction vessel, and
more than one 395 nm LED light, the time needed for the separation
should significantly decrease.
Fig. 14. Top) IR spectrum of the mixture in trial 1 prior to separation. Bottom)
IR spectrum of the mixture after separation. The light-green overlay indicates
the bands belonging to COF2, whereas the light-blue overlay indicates the bands
belonging to CO.
Table 6
Masses of MoF6 and UF6 calculated from IR spectra prior to and after irradiation.
Trial 1
Trial 2
MoF6 initial, mg
18.9 ± 0.1
16.0 ± 0.1
84.7 ± 0.7
37.0 ± 0.1
31.1 ± 0.1
84.1 ± 0.9
17.7 ± 0.1
15.0 ± 0.1
84.7 ± 0.7
35.3 ± 0.1
32.9 ± 0.1
93.2 ± 1.1
Quantitative amounts of UF5 and MoF6 were obtained. The UF5 was
collected and stored in an PTFE vessel in an inert atmosphere glovebox.
All further manipulation of this UF5 sample was done in the same inert
atmosphere glovebox. In order to perform elemental analyses on the UF5
sample, the entire sample was ground using an agate mortar and pestle
to ensure homogeneity (it was noticed that some hydrolysis occurred
during this process, as the sample turned from pale green to grey green).
Some of the UF5 sample was subsequently transferred to three gelatin
capsules. Capsule 1 contained 5.02 mg UF5; Capsule 2 contained 5.42
mg UF5; Capsule 3 contained 5.60 mg UF5. The capsules were then
dissolved in dilute nitric acid and microwaved to ensure complete
dissolution. Finally, the dissolved samples were diluted with dilute nitric
acid to 100 mL and analyzed by MP-AES. The remaining UF5 sample was
used for powder X-ray diffraction and IR spectroscopy.
MoF6 final, mg
Mo ratio (after / before) [%]
UF6 initial, mg
UF6 final, mg
U molar (after / before) [%]
4. Material and methods
4.1. General procedures and materials
All operations were performed in either stainless steel (316 L) or
Monel metal vacuum lines, which were passivated with undiluted
fluorine at various pressures before use. Preparations were carried out in
an atmosphere of dry and purified Argon (5.0, Praxair). Molybdenum
hexafluoride (99 %, abcr) was distilled once prior to use. Uranium
hexafluoride (99 %, homemade) was distilled once prior to use.
Homemade quartz reaction vessels were used for the CO separations.
Fluorinated ethylene propylene, FEP, vessels were used for the SO2
separations.
In order to perform elemental analyses on the MoF5 sample (pro-
duced by the reduction of the MoF6 sample), the entire sample was
dissolved in ca. 50 mL of distilled water. Cation: This dissolution pro-
cess produces significant amounts of hydrogen fluoride. This procedure
must be performed in a fume hood. Afterwards, 8 mL of concentrated
(65 %) nitric acid was added to the sample and then the sample was
diluted with distilled water to 100 mL. For analysis of this sample, 1 mL
of the Mo solution was taken and further diluted to 25 mL using dilute
nitric acid.
10

R.E. Stene et al.
Journal of Fluorine Chemistry 240 (2020) 109655
Trial 2: Original amounts/first irradiation cycle: 204.42 mg (0.58
Capsule 4 contained 2.58 mg of MoOF4; Capsule 5 contained 2.65 mg of
MoOF4; Capsule 6 contained 3.08 mg of MoOF4. All capsules were dis-
solved in dilute nitric acid and then microwaved to ensure dissolution.
All samples were then diluted to 50 mL with dilute nitric acid for
analysis by MP-AES. The remaining samples of UF5 and MoOF4 were
used to obtain powder X-ray diffraction patterns and IR spectra.
Trial 5: Starting amounts: 197.12 mg (0.56 mmol) of UF6, 197.38 mg
(0.94 mmol) of MoF6, and 6.61 mg (0.10 mol) of SO2. The yield of UF5
was quantitative, whereas the yield of MoOF4 was 60.42 %. Amount UF5
used for elemental analyses: Capsule 1 contained 2.58 mg of UF5;
Capsule 2 contained 2.65 mg of UF5; Capsule 3 contained 3.08 mg of
UF5. Amount MoOF4 used for elemental analyses: Capsule 4 contained
5.15 mg of MoOF4; Capsule 5 contained 5.22 mg of MoOF4; Capsule 6
contained 6.24 mg of MoOF4.
mmol) UF6, 198.83 mg (0.95 mmol) MoF6, and 847 mbar (1.04 mmol)
CO. Second irradiation cycle: 845 mbar (1.04 mmol) CO. Quantitative
yields of UF5 and MoF6 obtained. Amount of CO added to vessel con-
taining MoF6 (to reduce MoF6 to MoF5): 924 mbar (0.81 mmol).
Amounts of UF5 used for MP-AES analysis: Capsule 1 contained 5.49 mg
of UF5; Capsule 2 contained 4.94 mg of UF5; Capsule 3 contained 5.16
mg of UF5.
Trial 3: Original amounts/first irradiation cycle: 206.86 mg (0.59
mmol) of UF6, 198.83 mg (0.95 mmol) of MoF6, and 845 mbar (1.04
mmol) of CO. Second irradiation cycle: 880 mbar (1.09 mmol) of CO.
Quantitative yields of UF5 and MoF6 were obtained. Amount of CO
added to the vessel containing MoF6 (to reduce MoF6 to MoF5): 929
mbar (0.81 mmol). Amounts UF5 used for MP-AES analysis: Capsule 1
contained 5.31 mg of UF5; Capsule 2 contained 5.11 mg of UF5; Capsule
3 contained 5.05 mg of UF5.
Trial 6: Starting amounts: 201.99 mg (0.57 mmol) of UF6, 198.83 mg
(0.95 mmol) of MoF6, and 6.75 mg (0.11 mol) of SO2. The yield of UF5
was quantitative, whereas the yield of MoOF4 was 71.66 %. Amount UF5
used for elemental analyses: Capsule 1 contained 3.28 mg of UF5;
Capsule 2 contained 2.77 mg of UF5; Capsule 3 contained 2.56 mg of
UF5. Amount MoOF4 used for elemental analyses: Capsule 4 contained
5.37 mg of MoOF4; Capsule 5 contained 5.85 mg of MoOF4; Capsule 6
contained 5.37 mg of MoOF4.
Trial 7: This trial was performed only to collect MoF5 in order to
obtain a powder X-ray diffraction pattern and IR spectrum. Instead of a
quartz vessel, a 37 mL FEP reaction vessel was used for the separation
reaction. A 36 mL FEP vessel was used for the reduction of MoF6 to
MoF5. Note: FEP vessels are not recommended for these separation
reactions because the hexafluorides are somewhat soluble in the
plastic. Original amounts/first irradiation cycle: 172.68 mg (0.49
mmol) of UF6, 161.04 mg (0.77 mmol) of MoF6, and 1.03 bar (1.56
mmol) of CO. Second irradiation cycle: 1.03 bar (1.56 mmol) of CO.
Quantitative yields of UF5 and MoF6 were obtained. Amount of CO
added to vessel containing MoF6 (to reduce MoF6 to MoF5): 1.03 bar
(1.49 mmol).
4.4. Techniques and procedures for the separation of MoF6 from UF6
using supercritical carbon monoxide and near UV light
The preparation and irradiation procedure was identical for both
trials and was performed according to the previously given description.
Both trials differed in the analytical treatment of the reaction products.
4.3. Techniques and procedures for the separation of MoF6 from UF6
using sulfur dioxide and near UV light
Trial 1: Starting amounts: 44.58 (127 μmol) mg of UF6, 17.99 mg (86
μ
mol) of MoF6, and 42 bar (6.5 mmol) of CO. At end of irradiation, the
cold finger of high-pressure container was cooled to ꢀ 69 ◦C in order to
maintain MoF6 and unreacted UF6 in the solid phase. The volatile species
were removed at this temperature from the high-pressure container and
analyzed using IR spectroscopy. The pressure in the measuring cell was
1.7 bar.
Because all trials that were performed to demonstrate the separation
techniques are similar to each other, only Trial 4 will be discussed in
detail here. For Trial 5 and Trial 6, only the amounts of starting mate-
rials, yields, and samples used for MP-AES analyses will be discussed.
Trial 4: A 1-cm Teflon-coated stir bar was added to an approximately
30 mL homemade FEP reaction vessel. The vessel was then connected to
a stainless-steel valve, and then to a metal vacuum line. The reaction
vessel was placed under dynamic vacuum and then heated with a heat
gun set to 120 ◦C for 2 h to ensure adequate dryness. Afterward, 201.99
mg (0.57 mmol) UF6 and 197.38 (0.94 mmol) MoF6 were distilled into
the vessel using liquid nitrogen cooling. Onto the hexafluoride mixture,
7.53 mg (0.12 mol) SO2 were condensed at ꢀ 196 ◦C. The reaction
mixture was warmed to room temperature and stirred using the stir bar
and a stir plate. Once the hexafluorides were dissolved, the reaction
mixture was irradiated with 395 nm light for 1 h and 30 min. Note:
Irradiation times seem to be dependent on the amount of UF6 present.
Larger samples of UF6 may call for longer irradiation times. Afterwards,
the MoF6/SO2/SO2F2 mixture was distilled into another 30 mL FEP
vessel, which was previously dried through the procedure described
above. This reaction mixture was warmed to room temperature and left
for one week. After one week, the remaining SO2/SO2F2 was pumped off
under static vacuum and a white, crystalline sample of MoOF4 remained
in the FEP vessel.
Trial 2: Starting amounts: 39.70 mg (112
μmol) of UF6, 16.43 mg (78
μ
mol) of MoF6, and 56 bar (8.7 mmol) of CO. At end of irradiation, the
high-pressure container was not cooled, but instead all components that
were volatile at room temperature were condensed into a passivated
stainless steel AISI 316 L storage container using liquid nitrogen. The
volatiles were pumped off at this temperature. The storage container
was allowed to warm to room temperature and its contents were again
sublimed into a dried FEP vessel using liquid nitrogen. The vessel was
vented under constant cooling, water was added and finally the liquid
nitrogen was removed, allowing the vessel to warm to room tempera-
ture. MP-AES was performed on the resulting solution. The double
transfer was performed in order to avoid a potential overloading of the
FEP vessel by a rapid decompression of the high-pressure container. The
high-pressure container was emptied at room temperature under an
argon atmosphere and the extracted residues were dissolved in 50/50
(V/V) water/conc. HNO3 and the solution also analyzed by MP-AES.
Small amounts of the residues were lost, because they adhered to the
FFKM O-ring and remained in difficult to reach regions of the high-
pressure container.
The UF5 and MoOF4 samples were collected in an inert atmosphere
glovebox and stored in PTFE containers. The yield of UF5 was quanti-
tative, whereas the yield of MoOF4 was 72.00 %. To perform elemental
analysis, the entire UF5 sample or the entire MoOF4 sample were ground
using a agate mortar and pestle to ensure homogeneity. During this
process, slight hydrolysis of both samples occurred, as noted by the color
change (both samples turned somewhat greyish). For elemental analyses
of UF5, some of the sample was loaded into three gelatin capsules.
Capsule 1 contained 2.64 mg of UF5; Capsule 2 contained 2.47 mg of
UF5; Capsule 3 contained 2.59 mg of UF5. For elemental analyses of the
MoOF4 sample, some of the sample was loaded into 3 gelatin capsules.
4.5. Powder X-ray diffraction
Powder X-ray diffraction patterns were obtained with a Stadi-MP-
Diffractometer (STOE) using Cu-K
α radiation (λ =1.54051 Å), a
germanium monochromator, and a Mythen1K detector. The data were
handled using the WINXPOW software [36]. The compounds were filled
into borosilicate capillaries, which were previously flamed dried under
vacuum, and sealed using a hot tungsten wire under inert atmosphere in
a glovebox.
11

R.E. Stene et al.
Journal of Fluorine Chemistry 240 (2020) 109655
[10] S.-K. Lee, G.J. Beyer, J.S. Lee, Development of industrial-scale fission 99Mo
production process using low enriched uranium target, Nucl. Eng. Technol. 48
(2016) 613–623, https://doi.org/10.1016/j.net.2016.04.006.
4.6. Infrared spectroscopy
The IR spectra were measured on an alpha FTIR spectrometer
(Bruker) using a diamond ATR unit under an Ar atmosphere. The
spectrum was processed with the OPUS software package [37]. Gas
phase IR was measured in a specially designed and passivated measuring
cell manufactured from stainless steel AISI 316 L, equipped with BaF2
single crystal windows.
[11] T. Hollmer, W. Petry, A novel monolithic LEU foil target based on a PVD
manufacturing process for 99Mo production via fission, Appl. Radiat. Isot. 118
(2016) 290–296, https://doi.org/10.1016/j.apradiso.2016.10.003.
[12] J.V. Beitz, C.W. Williams, Photochemical Removal of NpF6 and PuF6 from UF6 Gas
Streams, 1990.
[13] J.V. Beitz, Removal of Fluoride Impurities from UF6 Gas, US4555318, 1985.
[14] Mark Fields, Metal Fluorides Separation, WO 97/38785, 1997.
[15] G.W. Halstead, P.G. Eller, L.B. Asprey, K.V. Salazar, Convenient multigram
syntheses of uranium pentafluoride and uranium pentaethoxide, Inorg. Chem. 17
(1978) 2967–2969, https://doi.org/10.1021/ic50188a059.
4.7. Microwave plasma atomic emission spectroscopy
[16] J.R. Geichman, E.A. Smith, S.S. Trond, P.R. Ogle, Hexafluorides of molybdenum,
tungsten, and uranium. I. Reactions with nitrous and nitric oxides, Inorg. Chem. 1
(1962) 661–665, https://doi.org/10.1021/ic50003a042.
The quantification of the elements was done by microwave plasma
atomic emission spectroscopy using a 4200 MP-AES spectrometer
(Agilent). Nitrogen was used for plasma generation. Samples were sol-
ubilized using 8 mL concentrated HNO3 within 100 mL of solution, or
with 4 mL concentrated HNO3 within 50 mL of solution. Commercial
uranium and molybdenum standard solutions were used for calibration
of the instrument.
[17] H.J. Hurst, P.W. Wilson, The electronic absorption Spectrum of gaseous uranium
hexachloride and hexafluoride, Spectrosc. Lett. 5 (1972) 275–279, https://doi.org/
10.1080/00387017208064964.
[18] K.N. Tanner, A.B.F. Duncan, Raman effect and ultraviolet absorption spectra of
molybdenum and tungsten hexafluorides1, J. Am. Chem. Soc. 73 (1951)
1164–1167, https://doi.org/10.1021/ja01147a085.
[19] T.J. Ouellette, C.T. Ratcliffe, D.W.A. Sharp, Vibrational spectra of molybdenum
and tungsten pentafluorides, J. Chem. Soc. A (1969) 2351, https://doi.org/
10.1039/j19690002351.
[20] W.H. Zachariasen, Crystal chemical studies of the 5 f -series of elements. XI. The
Declaration of Competing Interest
The authors report no declarations of interest.
Acknowledgments
crystal structure of α-UF 5 and of β-UF5, Acta Cryst. 2 (1949) 296–298, https://doi.
org/10.1107/S0365110X4900076X.
¨
[21] E. Jacob, Beitrage zur Chemie der Uranfluoride und -oxidfluoride. II. Darstellung
und Schwingungsspektren von? - und? -Uranpentafluorid, Z, Anorg. Allg. Chem.
400 (1973) 45–50, https://doi.org/10.1002/zaac.19734000107.
[22] R.E. Stene, B. Scheibe, C. Pietzonka, A.J. Karttunen, W. Petry, F. Kraus, MoF 5
revisited. A comprehensive study of MoF 5, J. Fluorine Chem. 211 (2018) 171–179,
https://doi.org/10.1016/j.jfluchem.2018.05.002.
We thank Dr. Uwe Linne, Marburg, and his team for aid with the U
and Mo analyses.
[23] N. Acquista, S. Abramowitz, Vibrational spectrum of MoF5, J. Chem. Phys. 58
(1973) 5484–5488, https://doi.org/10.1063/1.1679170.
[24] A.J. Edwards, B.R. Steventon, Fluoride crystal structures. Part II. Molybdenum
oxide tetrafluoride, J. Chem. Soc. A (1968) 2503, https://doi.org/10.1039/
j19680002503.
Appendix A. Supplementary data
[25] A. Lee Smith, The Coblentz Society Desk Book of Infrared Spectra, 2nd ed., The
Coblentz Society, 1977.
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.jfluchem.2020.10
9655.
[26] J.L. Hencher, G.W. King, Oxalyl halides, J. Mol. Spectrosc. 16 (1965) 168–178,
https://doi.org/10.1016/0022-2852(65)90095-0.
[27] T.G. Burke, D.F. Smith, A.H. Nielsen, The molecular structure of MoF 6, WF 6, and
UF 6 from infrared and Raman spectra, J. Chem. Phys. 20 (1952) 447–454, https://
doi.org/10.1063/1.1700440.
References
[28] R.S. McDowell, R.J. Sherman, L.B. Asprey, R.C. Kennedy, Vibrational spectrum and
force field of molybdenum hexafluoride, J. Chem. Phys. 62 (1975) 3974–3978,
https://doi.org/10.1063/1.430320.
[1] Nuclear and Radiation Studies Board, Division on Earth and Life Studies, National
Academies of Sciences, Engineering, and Medicine, in: Opportunities and
Approaches for Supplying Molybdenum-99 and Associated Medical Isotopes to
Global Markets: Proceedings of a Symposium, National Academies Press,
Washington, D.C., 2018, https://doi.org/10.17226/24909.
[29] A. Michels, T. Wassenaar, Th.N. Zwietering, The vapour pressure of carbon
monoxide, Physica 18 (1952) 160–162, https://doi.org/10.1016/S0031-8914(52)
80018-7.
[30] J. Stoll, J. Vrabec, H. Hasse, A set of molecular models for carbon monoxide and
halogenated hydrocarbons, J. Chem. Phys. 119 (2003) 11396–11407, https://doi.
org/10.1063/1.1623475.
[2] OECD, Nuclear Energy Agency, The Supply of Medical Isotopes: An Economic
Diagnosis and Possible Solutions, OECD, 2019, https://doi.org/10.1787/
9b326195-en.
[31] T. Shinoda, Vapor pressure of carbon monoxide in condensed phases, Bull. Chem.
Soc. Jpn. 42 (1969) 2815–2820, https://doi.org/10.1246/bcsj.42.2815.
[32] R.P. Pohanish, Sittigs Handbook of Toxic and Hazardous Chemicals and
Carcinogens, 7th ed., Elsevier, 2017.
[3] Committee on State of Molybdenum-99 Production and Utilization and Progress
Toward Eliminating Use of Highly Enriched Uranium, Nuclear and Radiation
Studies Board, Division on Earth and Life Studies, National Academies of Sciences,
Engineering, and Medicine, Molybdenum-99 for Medical Imaging, National
Academies Press, Washington, D.C, 2016, https://doi.org/10.17226/23563.
[4] C. Green, Technetium-99m production issues in the United Kingdom, J. Med. Phys.
37 (2012) 66, https://doi.org/10.4103/0971-6203.94740.
[33] A.J. Edwards, R.D. Peacock, R.W.H. Small, 868. The preparation and structure of
molybdenum pentafluoride, J. Chem. Soc. (1962) 4486–4491, https://doi.org/
10.1039/JR9620004486.
[34] P.G. Eller, A.C. Larson, Crystal structures of alpha-UF5 and U2F9 and spectral
characterization of U2F9, Inorg. Chim. Acta 37 (1979) 129–133.
[35] R.R. Ryan, R.A. Penneman, L.B. Asprey, R.T. Paine, Single-crystal X-ray study of
β-uranium pentafluoride. The eight coordination of U(V), Acta Crystallogr. B 32
(1976) 3311–3313, https://doi.org/10.1107/S0567740876010182.
[36] STOE WinXPOW, STOE & Cie GmbH, Darmstadt, Germany, 2015.
[37] OPUS, Bruker Optik GmbH, Ettlingen, Germany, 2009.
[5] The Hague Nuclear Security Summit Communique, 2014.
[6] The Washington Nuclear Security Summit Communique, 2016.
[7] NEA/SEN/HLGMR, Market Impacts of Converting to Low-enriched Uranium
Targets for Medical Isotope Production, 2012.
[8] S. Dittrich, History and actual state of non-HEU fission-based Mo-99 production
with low-performance research reactors, Sci. Technol. Nucl. Install. 2013 (2013)
1–9, https://doi.org/10.1155/2013/514894.
[9] S.A. Ali, H.J. Ache, Production techniques of fission Molybdenum-99, Radiochim.
Acta 41 (1987), https://doi.org/10.1524/ract.1987.41.23.65.
12