Reactive Polymers, 1 (1983) 301-308 301

Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands

THE EXTRACTION OF URANIUM FROM SEAWATER BY

POLY(AMIDOXIME)/POLY(HYDROXAMIC ACID) RESINS
AND FIBRE
F. VERNON and T. S H A H

Department of Chemistry and Applied Chemistry, University of Salford, Salford M5 4 WT, Lanes. ( U. K.)

(Received March 11, 1983; accepted in revised form May 6, 1983)

The uranium concentration in seawater is 3 # g / l with an estimated 5 × 10 ~ tonnes of

uranium in the oceans, in solution as the triearbonato complex. An), extraction process will
encounter the problems attendant on this high dilution, the only feasible methods currently
being ion exchange on chelating resins or sorption onto inorganic materials.
Poly(amidoxime)/poly(hydroxamic acid) chelating resin has been produced with high
uranium sorption from neutral solutions containing the metal as the tricarbonato complex,
and the results of a study of the behaviour of this resin towards seawater are given. High
chemical and biochemical stability and fast sorption kinetics are properties of the resin
which can sorb 68 per cent of the uranium present using a 24 s resin to seawater contact
time. Poly(amidoxime)/poly(hydroxamic acid) fibre, prepared by the oximation of
poly(acrylonitrile) fibre, is able to sorb 12.5 per cent of uranium in seawater with a 2 s
contact time and has the advantage of being in a form capable of weaving into a chelating
cloth. Sorbing 1.8 mg uranium per gramme fibre per 10 days, the cloth can be produced as
an endless belt, for a continuous process for uranium extraction. A theoretical model
indicates that uranium production could be possible at 6 tonnes per annum.

INTRODUCTION annum. It has been suggested by the M.I.T.

Energy Laboratory (1980) that the break-even
The large-scale extraction of uranium from point in any seawater-pumping operation is
the oceans is a subject of considerable inter- through a vertical height of 10 feet', pumping
est, some 200 papers on various aspects hav- beyond this would require more energy than
ing been published over the past 15 years. would be obtained from the uranium re-
Much of the work is concerned with the sorp- covered. This has led to suggestions such as
tion of uranium by hydrated titania granules pumping through floating beds of sorbent [4],
[1-4] and, in Japan, a large-scale operation but pumping operations, unless used in con-
pumping seawater through beds of hydrated junction with other energy processes such as
titania granules is under way with a primary the Ocean Thermal Energy project (OTEC)
objective of extracting 69 kg of uranium per appear to be uneconomic. Utilization of tidal

0167-6989/83/$03.00 ,i 1983 Elsevier Science Publishers B.V.

302

flow and ocean currents to bring the seawater rous, p o l y ( a m i d o x i m e ) / p o l y ( h y d r o x a m i c

into contact with the sorbent would seem to acid) microbead resin ( P A O / P H A ) and a
be attractive alternatives and have been dis- poly(amidoxime)/poly(hydroxamic acid) fibre
cussed by Kanno [1 ]. Problems associated with produced by the oximation of poly(acryloni-
such a process are: trile) fibre. The manner in which most of the
(1) The large volumes of seawater required to problems described may be overcome by using
contact the sorbent, 1 x 10 9 litres of the fibre in a continuous seawater extraction
seawater will yield only 3 kg of uranium if process is also considered.
extracted with 100 per cent efficiency.
(2) The mechanics of supporting large area
beds of sorbent granules across ocean cur- EXPERIMENTAL
rents.
(3) The cost of transporting, eluting and re- P A O / P H A macroporous, microbead resin
placing the sorbent beds in the ocean. was produced by the oximation of 5 per cent
(4) The need for sorbents of high uranium (DVB) cross-linked poly(acrylonitrile)-
selectivity so that sorbent sites are not poly(ethylacrylate) copolymer using sodium
occupied mainly by other metals. hydroxide and hydroxylamine in methanol by
(5) Fouling of beds by suspended material in the method described in Ref. [8]. P A O / P H A
the seawater. fibres and woven cloth were produced from
(6) Stability of the sorbents to prolonged poly(acrylonitrile) materials by the same
seawater contact. oximation process.
Whilst hydrated titania exhibits high Total capacity for copper was determined
uranium selectivity, its physical strength and by batch sorption from 100 millimolar copper
resistance to attrition by seawater are low and solution in pH 4.0 acetate buffer. Capacities
the granules rapidly become deactivated by for uranium were determined by batch sorp-
calcium which is difficult to remove as the tion from a 10 millimolar uranium solution in
granules have low stability towards acid 0.5 M sodium chloride in which the uranium
eluants. The uranium is present in seawater as was converted to the tricarbonato complex by
the soluble tricarbonato complex and cannot the addition of the theoretical amount of
be extracted by Dowex A1. A chelating ion- sodium carbonate. The pH was adjusted to
exchange resin containing hydroxamic acid 7.8.
functional groups was produced as it was felt Two methods were used in the assessment
that this functionality would have the ability of metal sorption from seawater, a dynamic
to remove uranium from its complex at the method in which measured volumes of
pH of natural seawater. Our early work [5,6], seawater were pumped through columns of
with a DVB cross-linked poly(acrylonitrile) resin or fibre at known rates, and a static
granular resin converted to poly(hydroxamic method in which the material to be tested was
acid) by a synchronous oximation step using packed into a frame 2 cm thick by 20 cm
hydroxylamine in sodium hydroxide solution, square, the retaining walls being glass fibre
revealed the potential of this material. Tests cloth supported by a plastic mesh. These
on seawater demonstrated that, when passed frames were immersed in the sea, held on a
through columns of resin, up to 90 per cent of rigid structure below low water level. Elution
the uranium was sorbed. The kinetics of the of uranium and other metals was achieved by
process, however, were very slow. In this paper overnight shaking with 1 M or 2 M sulphuric
we describe the uranium-sorbing properties of acid.
that resin with a later version of a macropo- The u r a n i u m contents were usually
 303

determined colorimetrically using arsenazo III time of 5 min. The properties of this resin
reagent. Checks on the values were made in- together with those of a later improved ver-
dependently using 8-hydroxyquinoline as a sion and of P A O / P H A fibres are given in
colorimetric reagent and also by X-ray fluo- Table 1.
rescence measurement directly on the poly- An improved PHA resin was subsequently
mer. Other metals were determined in eluates produced [7] with a macroporous, microbead
by atomic absorption spectrophotometry. structure and much faster sorption kinetics.
Resin stabilities were determined from The preparation of this resin has been de-
water regain and total capacity values for scribed in detail [8] and it was able to sorb 91
copper and uranium on samples taken from p,g of uranium (76 per cent extraction) and
the sea, as a function of immersion time. In 106 /zg of uranium (88 per cent extraction)
addition, laboratory tests involved immersion from 40 1 seawater samples using seawater-to-
of samples in a tank of aerated seawater which resin contact times of 3 rain arid 8 min,
was drip fed with glucose and biopeptone respectively.
nutrients: stability towards acids was studied The composition of this resin was de-
by shaking samples with various molarities of termined from hydrolytic breakdown studies
sulphuric acid for periods of up to 45 days. and measurement of nitrogen on the resin and
in breakdown products [9], and it was found
that the resin, as prepared, contained 40 per
R E S U L T S AND D I S C U S S I O N cent PAO and 50 per cent PHA; therefore, it
will be referred to as the P A O / P H A resin.
The poly(hydroxamic acid) resin, PHA, The stability of the amidoxime groups to-
used for initial studies of sorption of uranium wards acid eluants is not high, see Table 2:
from seawater [5] was a granular material, 5 the resin converts to a form containing 6 per
per cent cross-linked with DVB, and exhibit- cent PAO and 84 per cent PHA when in
ing poor kinetics of sorption of copper with a contact with 0.5 M sulphuric acid for a three-
time to half-saturation of the resin of 102 day period. More concentrated acids bring
rain. In spite of its slow kinetic performance, about this conversion faster but the PHA
however, this resin removed 340 btg of uranium resin resulting is then stable, loss of function-
from 300 1 of seawater (a 38 per cent uranium ality being found only in the case of resin
recovery) using a seawater-to-resin contact contact with 4 M sulphuric acid after 45 days.

TABLE 1
Properties of P A O / P H A resins and fibres tested
Early resin Optimised resin Fibre
Form granular macroporous/ 60 mm
microbead fibres
Cross-linking (~) 5 5 0
Particle size (/~m) 375-1500 150 300 3 denier
Water regain (g/g) 1.1 2.1 1.1
Total capacities (mmol/g):
Cu(pH 4.0) 2.1 2.6 1.9
U (pH 7.8) - 0.9 0.6
Kinetics. tl/2 (rain):
Cu (pH 4.0) 102 12 12
U (pH 7.8) - 13.5 12.5
304

TABLE 2
Optimised resin stability towards prolonged contact with seawater and with acid eluants

Water regain Cu capacity at U capacity at

(g/g) pH 4 (mmol/g) pH 7.8 (mmol/g)
Original P A O / P H A resin a 2.6 2.9 0.9
Resin in sea, 2 months 2.6 3.2 0.8
Resin in sea, 6 months 2.4 2.8 0.8
Resin in 0.5 M H2SO 4, 1 day 1.6 2.0
Resin in 1 M H2SO 4, 3 days 1.6 1.5
Resin in 2 M H2SO 4, 5 days b 1.6 1.5

" Nitrogen content 20.9%, corresponding to 40% PAO, 50% PHA.

b Nitrogen content 15.5%, corresponding to 6% PAO, 84% PHA.

Table 1 shows that the total capacity of the Seawater tests

resin for uranium, when present initially as
the tricarbonato complex in a simulated The uptake of uranium by the resin when
seawater at pH 7.8, is 0.9 m m o l / g dry resin. used in column form and seawater pumped
Elution studies using 1 M sulphuric acid or 1 through is shown in Table 3 where it is seen
M sodium carbonate, revealed that either of that for contact times of 24-120 s uranium
these eluants can effect complete recovery of recovery is high, very much better than when
uranium from the resin. using the earlier PHA resin, as shown in Fig.
The resin stability towards the marine en- 1. Resin samples in a sandwich construction,
vironment is very good. Table 2 shows the 2 cm thick by 20 cm square and retained by
stability, in terms of uranium and copper glass cloth, were immersed in the sea for
capacities of the resin before and after pro- various periods. On recovery, the uranium
longed immersion in the sea. Very little change contents of the samples were determined, pro-
in properties occurred after 6 months immer- ducing the results given in Table 4. It was
sion. Additionally~ laboratory tests where the found that the uranium loading increased as a
resin was kept in aerated seawater with addi- function of immersion time over periods of up
tions of glucose and biopeptone as nutrients to 4 months, but the values are disappoint-
showed that, although heavy algal growth oc- ingly low. The problem is one of obtaining
curre& the resin was not affected by this sufficient current flow through the beds. It
environment. Table 2 also shows the effects was found that resistance to flow increased
mentioned of prolonged contact with acids,
the copper capacity of 1.5 m m o l / g corre-
TABLE 3
sponding to the resin containing almost 90
per cent PHA functionality. Uptake of uranium by P A O / P H A resin (2 g) on pump-
ing seawater through it at various rates
Sorption elution of uranium cycles were
carried out with 24 hour periods of equilibra- Seawater Contact U sorbed U recovered from
tion with uranium in simulated seawater fol- volume time (/~ g) seawater
(1) (s) (%)
lowed by 24 hour elution periods with 1 M
sulphuric acid. After ten such cycles the resin 4.2 120 11 87
8.2 48 19.5 79
was examined and found to have undergone
12.0 24 24.5 68
no reduction in its uranium capacity.
 305

ever, it has been shown [8] that the P A O / P H A

i
~0 2¸
resin is able to quantitatively sorb uranium
from neutral solutions of high magnesium
concentration so the alkaline earth elements
I
are not regarded as a problem.
f

~ / Iron, however, is a serious problem as it is

sorbed in a five-fold excess over uranium and
is difficult to desorb. Sugasaka and coworkers
[10] have shown that P A O / P H A resins ex-
r i
hibit higher selectivity for uranium sorption
L
60 120 ] 80
Cantoct t i m e (s~ from seawater than for other metals with the
exceptions of calcium and chromium. Their
Fig. 1. Efficiency in extraction of uranium from seawater
of resins and fibre as a function of contact time.
selectivities are based on concentration fac-
tors of the metals on the resins. Laboratory
tests have shown [11] that P A O / P H A resin
TABLE 4 exhibits a greater affinity for iron(III) than
P A O / P H A resin (as sandwich structure) showing for uranium, the former being the more dif-
uranium loading as a function of time of immersion in ficult to desorb. Successful desorption of iron
the sea requires the use of sulphuric acid of at least 2
Immersion time (days) 4 30 60 120 M concentration and any uranium extraction
Uranium loading ( # g U / g resin) 5 27 45 72 process would require either a uranium-from-
iron separation stage after desorption, or the
use of eluants of carefully controlled acidities
with immersion time, the resin bed becoming and elution times to bring about the selective
gradually blocked by mud and sand sus- desorption of uranium. A possibility is the
pended in the water. desorption of uranium by sodium carbonate
followed by regeneration of the resin by wash-
Resin selectivity for trace metals from seawater ing with sulphuric acid. X-ray fluorescence
techniques on the resins were used to de-
Two of the samples, immersed for pro- termine the extent of sorption of uranium and
longed periods, were analyzed for full metals molybdenum: whilst conditions were used en-
content. The results, given in Table 5, reveal a abling quantitative uranium sorption, it was
high uptake of calcium and magnesium. How- found that the resin sorbed approximately 50
per cent of the molybdenum present in
seawater. This metal has to be removed by
alkaline elution with sodium hydroxide solu-
TABLE 5
tion. The resin, however, shows high stability
P A O / P H A resin loading for various metals sorbed as a
towards alkaline solutions; the sodium form
function of resin immersion time in the sea
of the resin exhibits a faster uptake of uranium
Immersion Metal content of resin ~ ( f i g / g ) from seawater than does the resin in the hy-
time Ca Mg Fe Mn Zn U Cu Ni drogen form [8].
(days)
60 570 79 195 99 83 45 18 0 PA 0 / PHA fibres
120 2,400 1,200 471 96 49 72 20 3

Sn, Al, Sr and Ba were not found in measurable The conversion of linear poly(acrylonitrile)
quantities in either sample. fibre to P A O / P H A fibre may be carried out
306

TABLE 6
Properties of P A O / P H A woven cloth
Sample Water regain Cu capacity at U capacity at Strength
(g/g) pH 4 (mmol/g) pH 7.8 (mmol/g)
1 0.5 0.4 0.1 Excellent
2 0.8 1.0 0.3 Very good
3 1.0 1.9 0.6 Good
4 1.8 2.5 0.7 Medium
5 2.5 3.1 1.4 Very poor

in a manner similar to microbead resin con- Further improvement of fibre led to the
version, the fibre being heated with a sodium highest rate yet reported of 4 mg U per
hydroxide-hydroxylamine solution in metha- gramme fibre over ten days [10]. This is very
nol. As this fibre is not cross-linked, the intro- impressive when compared with the kinetics
duction of excessive amounts of hydrophilic of sorption by hydrated titania, values being
groups such as amidoxime and hydroxamic of the order of 0.2 mg U per gramme titania
acid leads to increased swelling properties and over 10 days.
a loss of fibre strength. An optimised conver- Figure 1 shows that providing the seawater
sion is therefore necessary, balancing uranium is in contact with the fibre for a period longer
capacity against fibre strength. Table 6 shows t h a n 20 s, more than 50 per cent of the
how oximation of P A O / P H A woven cloth uranium present in seawater is sorbed. This
can lead to high copper and uranium capaci- work on sorption as a function of flow rate
ties of 3.1 and 1.4 m m o l / g , respectively, but through fibre-packed columns, was followed
the resulting cloth had excessive swelling by a study of the sorption characteristics of
properties and very low mechanical strength. the fibre using constant flow rates and much
The optimum cloth would appear to have smaller contact times. A nominal flow rate of
copper and uranium capacities around 2.0 40 1/min gave seawater-to-fibre contact time
and 0.6 m m o l / g , respectively. It should be of 2 s and, from total volume measurements,
noted that such a cloth, in the dry state and it was shown that approximately 13 per cent
fully loaded with uranium, would contain 14 of the uranium in the seawater was sorbed
per cent by weight of the metal. As shown by under these conditions. Table 7 gives the re-
Fig. 1, the kinetics of uranium uptake by suits of this study from which it can be seen
P A O / P H A fibres are similar to those of the that the uranium uptake by the fibre is of the
optimised resin, whereas Sugasaka et al. [10] order of 8 ~g U per gramme of fibre per hour.
obtained much faster kinetics for such fibres This figure corresponds to 1.8 mg U per
than for either their resin or hydrated titania gramme per 10 days, the value reported in
granules. Our results indicate that the fibres [12]. Hence our fibre and the Japanese fibre
are comparable whereas our resin has much appear to show very similar kinetics with re-
faster kinetics than the resins described in spect to uranium sorption. However, the
[10]. Japanese workers have reported more recently
The Government Research Institute, Shiko- that the maximum uranium sorption from
ku, have described [12] the use of P A O / P H A seawater by the fibre is 100 #g per gramme
fibres for sorption of uranium from seawater fibre per week [13] and it is not clear how the
and have obtained a uranium absorption rate apparent discrepancy between this low value
of 1.8 mg U per gramme fibre over ten days. and the value given in earlier work, and by
 307

TABLE 7
Uptake of uranium by P A O / P H A fibre (2 g) on pumping seawater through it at a constant rate (nominally 40 l / h )

Seawater volume Average Uranium Uranium Uranium

pumped fibre-seawater sorbed uptake rate recovered from
(1) contact time (t~g) ( f f g / g h) seawater
(s) (~)
21.3 1.8 7 7 It
39.5 1.8 17 8.5 14.3
74 1.9 28 7 12.6
151 2.3 66 8.2 14.5

our studies, has come about; particularly, in process utilizing the woven P A O / P H A cloth
view of the statement in Ref. [12] that alkaline in the form of an endless belt which Would
pretreatment of the P A O / P H A fibre leads to pass over the sea bed. The belt would be
enhanced sorption kinetics, although a figure located so that two-thirds of it was below the
of 2 mg per gramme fibre per week has re- sea, the remainder passing through uranium
cently been reported for this material [14]. desorption tanks located on a pontoon. Fibre-
Our value of 1.8 mg U sorbed per gramme to-seawater contact time would not be a func-
of fibre in 10 days may be extrapolated to 5.4 tion of belt speed but of the flow rate of the
kilo of uranium per tonne of fibre per month, ocean or tidal current in which the plant was
a figure revealing the magnitude of the prob- situated. From desorption rates, calculations
lem of extracting tonnage quantities of urani- show that the belt speed need not be more
um from seawater. The fibrous nature of the than a few metres per hour so the energy for
adsorbent means that, woven into cloth, the belt drive could possibly be derived from
structures suitable for immersion in the sea tidal energy. Figure 2 shows, in schematic
may be produced. Of the problems outlined layout, how such a continuous exchange pro-
in the introductory section, the problem of cess could operate and, by projection of the
cost-operating a sorption-desorption batch results to date, our calculations show that
process must be considered. The alternative to such a process, employing a belt 400 metres
such a batch process could be a continuous by 10 metres by 1 cm thick, operating in a 4
knot current, could yield uranium at a rate of
up to 6 tonnes per annum. This figure is
Belt dr,re
dependent upon free access of seawater to
Prccessmg
fibres within the belt, however, and current
Seo
work is concentrated on this aspect, determin-
Pontoo~ with
ing the efficiency of uranium sorption as a
extraction bath5
function of P A O / P H A fibre cloth thickness.
Consideration of the problems of supporting
large areas of titania or resin bead beds across
ocean currents, leads to the conclusion that
the type of continuous process outlined here
Fig. 2. Schematic diagram of continuous extraction of
is the most promising approach to the prob-
uranium from seawater with the P A O / P H A - f i b r e cloth lem of large-scale extraction of uranium from
used as an endless belt. seawater.
308

REFERENCES and uses of polyhydroxamic acid resins, Pure Appl.

Chem., 54 (1982) 2151.
1 M. Kanno, Design and cost studies on the extraction 9 E. Ichsan, Stability and chelating properties of an
of uranium from seawater, Sep. Sci. Technol., 16 amidoxime resin, M.Sc. Thesis, Salford, 1982.
(1981) 999. 10 K. Sugasaka, S. Katoh, N. Takai, H. Takahashi and
2 M. Kanno, Present status of study on extraction of Y. Umezawa, Recovery of uranium from seawater,
uranium from seawater in Japan, N i p p o n Sep. Sci. Technol., 16 (1981) 971.
Genshiryoku Gakkaishi, 23 (1981) 36. 11 F. Vernon and T.W. Kyffin, Iron, copper and
3 A.D. Kelmers, Status of technology for the recovery uranium separations on Duolite CS-346, Anal. Chim.
of uranium from seawater, Sep. Sci. Technol., 16 Acta, 94 (1977) 317.
(1981) 1019. 12 Extraction of uranium and lithium from seawater,
4 P. Koske, Recovery of uranium from seawater, Ger- The Government Research Institute, Shikoku,
man Patent No. 2936399 (1979). JETRO, Sept. 1981, p. 5.
5 T.W. Kyffin, Chelating ion exchange on amidoxime 13 S. Katoh, K. Sugasaka, K. Sakane, N. Takai, H.
and hydroxamic acid resins, Ph.D. Thesis, Salford, Takahashi, Y. Umezawe and K. Itagiri, Preparation
1976. of the fibrous adsorbent containing amidoxime group
6 H.A.A. Abdalla, Extraction, adsorption and precipi- and adsorption property of uranium, Nippon Kagaku
tation of seven metals from seawater matrix, Ph.D. Kaishi, 9 (1982) 1454.
Thesis, Salford, 1980. 14 S. Katoh, K. Sugasaka, K. Sakane, N. Takai, H.
7 F. Vernon and W.M. Zin, Sorption and separation Takahashi, Y. Umezawe and K. Itagiri, Enhance-
of gold and silver on a polyhydroxamic acid resin, ment of the adsorptive properties of the amidoxime
Anal. Chim. Acta, 123 (1981) 309. group containing fibre by alkaline treatment, Nip-
8 F. Vernon, Chelating ion exchangers--the synthesis pon Kagaku Kaishi, 9 (1982) 1458.