J. inorg. nurl . Chem ., 1976, Vol . 38. pp . 1141-1147 . Pergamon Press.

Printed in Great Britain

CRYSTAL HABIT AND PHASE ATTRIBUTION OF

U(VI) OXIDES IN A GELATION PROCESS

M . H. LLOYD, K . BISCHOFF, K . PENG, H .-U. NISSENt and R . WESSICKENt

Swiss Federal Institute for Reactor Research, CH-5303 Wurenlingen, Switzerland

(First received 5 May 1975 ; in revised form 8 October 1975)

Abstract-The effects of key process variables on the product characteristics of urania microspheres prepared by an
internal gelation method are determined by studies of the crystal structure and morphology of urania in individual gel
microspheres . Three compounds in the system U0,-NH,-H 2 0 are identified, and their characteristic crystal
morphologies are determined . The results show that information obtained in this way provides a sound basis for the
attribution of phases, for evaluating process conditions and for optimizing process variables .

INTRODUCTION determined . Conditions which result in noncrystalline

THIS paper reports studies of crystallographic parameters material were also found . The results show that informa-
and crystal morphologies of a series of compounds in the tion obtained in this way provides a sound basis for
U03-NH 3-H2 0 system . The major goal of this study was understanding and evaluating process variables and
to determine the effects of process variables on the crystal optimizing process conditions . Therefore, the techniques
habit and crystallization behaviour of uranium oxide established in this study could be of general interest in
during the preparation of urania and urania-plutonia studying similar systems of various oxides ; they could
microspheres by an internal gelation method . This crystal also be of interest as methods for preparing specific
system has been the subject of numerous crystal forms.
investigations [1-6] ; however, despite the fact that de-
tailed X-ray diffraction power data have been reported, EXPERIMENTAL
the different compounds in the system are not sufficiently Microsphere preparation
known . For example, Cordfunke[2] found four com- In the procedure for preparing gel microspheres, acid-deficient
pounds in this system, while Stuart and Miller [6] reported solutions of uranyl nitrate hexahydrate (UNH) are cooled and
that the ammonium uranate system, in equilibrium, is a mixed with a cold solution of urea and hexamethylenetetramine
continuous non-stoichiometric one . (hexa) . This cold feed solution is dispensed dropwise into the
sphere-forming column concurrent with a flowing stream of hot
Microsphere preparation is based on the Kema U(VI)
silicone oil . The gelled spheres are collected, removed from the
sol-gel process, which has been shown previously to yield
oil, and washed to remove residual reaction products .
gel microspheres with a urania structure belonging to the All the procedures used in these experiments were standardized
U03-NH 3-H20 system of compounds [7] . In the overall as much as possible to ensure good reproducibility . Acid-deficient
process [8, 9], concentrated Pu(N0 3)4 solution and colloi- UNH feed was prepared by dissolving UNH in NH4OH to provide
dal carbon are mixed with the partially hydrolyzed uranyl a solution containing 1 . 5 moles of uranium per kg and having any
nitrate feed solution and formed into microspheres, which desired unneutralized NO, - /uranium mole ratio within the range
are subsequently converted to (U, Pu)C spheres by 1 . 5-2 .0 . The hexa feed solutions contained 2 .7 moles of hexa and
carbothermic reduction at elevated temperatures . These 2 . 7 moles of urea per kg . The final feed solutions, which had
hexa/uranium mole ratios from 0 .9 to 1 .6, were prepared by
products are of interest as an advanced reactor fuel
mixing appropriate amounts of these two solutions after they had
concept for the liquid metal- or gas-cooled fast breeder
been precooled to the desired temperature (-7 to 5°C) . The mixed
reactor . Experience in the production of carbide micro-
feed was further cooled for 2 min and then transferred to the
spheres suggests that the gelation process, which deter- cooled dispensing vessel . Sphere forming was started 3 min after
mines the properties of the gel microspheres, greatly mixing of the feed had been completed, and the sample to be
influences the characteristics of the final product and analyzed was collected for 3 min . The spheres were washed
possibly several aspects of the reaction sintering mechan- immediately in carbon tetrachloride to remove silicone oil and
ism as well . This is borne out by variations in product then in IM NH 4 OH to remove residual reaction products . The
density and microsphere cracking behaviour, and by washing procedure consisted of soaking the spheres in three
variations in carbide stoichiometry at constant carbon-to- consecutive carbon tetrachloride baths for 15 min in each bath and
subsequently in three consecutive NH 4OH baths for 30 min in
metal ratios in the initial spheres .
each bath . Finally, the spheres were allowed to air dry . In some
In order to determine the effects of process variables on
cases, the spheres were digested in hot silicone oil prior to
urania behavior, techniques were developed whereby washing . This was accomplished by heating the sample for 30 min
gelled microspheres from any stage in the process could at 60°C .
be ultrasonically dispersed and examined without altering Experiments were conducted to evaluate the effects of two
the chemical or physical characteristics of the primary principal chemical variables : (1) variation of the hexa/uranium
particles . Three crystalline phases were observed and mole ratio within the range 0 .9-1 .6, and (2) variation of the
identified, and their characteristic morphologies were unneutralized NO, - /U mole ratio from 1 . 5-1 .75 . Several physical
process variables which influence crystallization behavior were
evaluated as functions of these chemical variables . In one series
tGuest scientist from Oak Ridge National Laboratory Oak of experiments, microspheres were prepared at NO, -/U mole
Ridge, TN 37830, U .S .A. ratios of 1 . 5 and hexa/U mole ratios of 0 .9-1 .6 . In a second series,
#Laboratory of Electron Microscopy II, Swiss Federal hexa/U mole ratios of 1 . 2. 1 . 4 and 1 . 6 were evaluated at a NO, /U
Institute of Technology, CH-8049 Zurich-Honggerberg, Switzer- mole ratio of 1 . 75 .
land . For each feed composition, two or more temperatures were

1141
J.J.N.C., Vol . 38, No . 6-D
1 14 2 M. H . LLOYD et al.

used in the sphere forming column ; in some cases, two or more suspended in a bedacryl-acetone mixture on polyester film which
feed temperatures were evaluated since these factors influence was attached to the X-ray sample plate . In the case of spheres that
gelation rate and chemical reaction time . The effect of digesting had been washed in NH,OH, powders were prepared by grinding
gelled spheres in hot silicone oil was also investigated, and efforts air-dried material. Spheres that had not been washed were
were made to determine whether the age of the feed is a significant dispersed ultrasonically, centrifuged, and decanted to remove
process variable . In several experiments, microspheres were impurities, and then suspended in the glue mixture .
analyzed both before and after the NH,OH washing step .
Electron diffraction
Microsphere dispersion techniques It was observed that the electron diffractograms (ring patterns)
Techniques were developed in this study so that gel spheres changed immediately after irradiation with the 100 kV electron
could be ultrasonically dispersed to form colloidal dispersions of beam . This indicated a phase transformation as a result of
the primary particles without significantly altering the chemical or irradiation . All diffraction patterns obtained consisted of compara-
physical characteristics of the particles. tively few lines indicating that the crystallites have been
The procedure consists of crushing ten microspheres in 40 ml converted into uranium oxide . During this process they retain,
of distilled water and dispersing the resulting suspension however, all recognizable morphological characteristics of the
ultrasonically for 20 sec at a power output of 60 W and a original compounds .
frequency of 20 kc . A B-12 Sonifier manufactured by the Branson
Sonic Power Company was used .
RESULTS AND DISCUSSION
Stable colloids were produced by this treatment when spheres
that had previously been washed in NH,OH were dispersed ; X-ray diffraction results
however, when spheres were dispersed prior to washing in The X-ray powder diagrams of three compounds in the
NH,OH, the dispersed solids coagulated shortly after the system UO,-NH 3-H20 were found . Two of the patterns
ultrasonic treatment due to the presence of impurities such as could be indexed as structures previously reported by
NH,NO,, urea, and unreacted hexa. To remove these impurities, Debets and Loopstra[3] ; the third pattern appears to be
the sample was centrifuged for 60 sec in a clinical centrifuge and identical with the structure reported by Garner[1] . These
the supernate was decanted . Stable colloids were obtained by
results can be expressed as follows :
mixing these solids with fresh distilled water .
Compound 1 : U03.2H2O ; orthorhombic : a = 14.01 A,
During initial efforts to disperse microspheres, the principal
concern was that the primary crystallites could be degraded or b=16 .76 A, c=14 . 92 A.
shattered ; however, crystal growth was found to be the most Compound 2 : U03 . 1/4NH3 .7/4H2Ot ; orthorhombic ;
significant artifact introduced by ultrasonic treatment . After about a =7-16 A, b = 12 . 24 A, c = 15 .09 A.
60 sec of treatment, the temperature of the sample began to Compound 3 : 2U0 3 •NH, •3 H2 Ot ; hexagonal ; a =
increase and changes occurred very rapidly . With a dispersion 14-03 A, c = 15 . 02 A.
time of 2 min, crystal diameters increased by as much as a factor Figure 1 shows X-ray photographs of the three
of 20, amorphous material invariably crystallized, and changes in compounds, along with the corresponding (hkl)-values for
crystal structure occurred . Aging in water at room temperature compounds 1 and 2 and the (hkil)-values for compound 3 .
also promoted crystal growth . This phenomenon was slow,
Compound 1 is well known . Its X-ray diagram has been
however ; no measurable change was observed after 2 hr, while an
aging period of 60 hr approx . doubled the crystal diameter . obtained both from synthetic products[10, 11] and from
There are no indications that significant artifacts are introduced mineral samples[12] . In this work, the interplanar
when an ultrasonic dispersion time of 30 sec or less is used or spacings and sin e 0 values determined for each reflection
when samples are prepared promptly after dispersion . Crystallite were in good agreement with those reports for /3-
degradation does not occur since only large (2000A-500014 U0 3 . 2H2O by Debets and Loopstra . Satisfactory agree-
well-formed crystals are observed even after 6 min of dispersion ment was also obtained between observed and calculated
time . Similarly, crystal growth or a change in morphology does not values .
occur until the dispersion time exceeds about 60 sec since, with Compound 2 was observed in two forms (Fig . 1) . The
shorter dispersion times, amorphous material and a wide variety
reported by Garner for U0 3 .. 7/4 H20-1/4 NH3 and is
of crystal sizes were found in various preparations .
very similar to that reported by Debets and Loopstra for
Electron microscopy their compound II (3UOVNH 3 . 5H2O) . A comparison of
Transmission electron micrographs were obtained with a lattice parameters is shown in Table 1, while comparative
Siemens Elmiskop la at 100 kV . The electron micrographs were powder data are given in Table 2 . Comparison of these
usually taken at 40,000 diameters magnification . The samples were lattice parameters clearly shows that our measurements
prepared by evaporating a small drop of the diluted colloid on a are in good agreement with those reported by Garner and
carbon foil supported by a 300-mesh copper grid . Because of are in satisfactory agreement with the strong reflections
water aging effects, the samples were prepared immediately after reported by Debets and Loopstra .
microsphere dispersion . The entire preparation required only a Compound 3 was observed in two forms (Fig . 1) . The
few minutes .
welldefined pattern (Fig . lb) was obtained from samples
in which crystal growth occurred due to heating as a result
X-ray diffraction
The X-ray analyses were performed with a Guinier de Wolff of ultrasonic dispersion and water aging . Although such
camera, using CuK a radiation at room temperature . The sine 0 samples are not representative of the true system, they
and the d-spacing were determined from each measured reflection . were helpful in indexing unaltered samples (Fig . lc) . The
The patterns obtained were complicated by the fact that many of latter samples display fewer reflections and these are all
the reflections were very diffuse . Therefore, it was rather difficult diffuse due to line broadening as a result of the very small
to determine the exact position of maximum intensity on the film, crystallite size (-100-A platelets, see next section) . The
even with a microdensitometer . Relative intensities were charac- d-values determined for this structure are indicated by
terized from visual estimations as very strong (vs), strong (s),
asterisks in Table 2 .
medium (m), weak (w), very weak (vw), and very very weak The reflections of compound 3 are broadened and
(vvw) . Samples were prepared by dispensing crushed spheres
diffuse (Fig . Id), which complicates unambiguous identifi-
cation . Nevertheless, the pattern can be indexed and
tBased on compositions prosed by Garner[1] and by Debets correlated with compound III (2U0 3 -NH3 -3 H20) re-
and Loopstra[31 . ported by Debets and Loopstra . Table 3 gives a


Crystal habit and phase attribution of U(VI) oxides in a gelation process 1143

Y- e

N a eON' ON ON eNN
O IN O(eht 1010 (0W i W I I
a a
O N
ONN
ONN
N N O O O e N
ee (OtO (D a (D

Fig. 1 . Guinier de Wolff X-ray diagrams of compounds 1-3 . The corresponding hkl values for compounds I and 2
and the hkil values for compound 3 are given. (a) Compound 1 ; (b) compound 2 ; (c) compound 2, diffuse pattern due
to line broadening ; (d) compound 3 .

Table 1. Lattice parameters of compound 2 compared with values reported by Garner[1] and by Debets and
Loopstra [3]

No of lines, LatticeParameters (R)

d=9 .52to1 .62 a b c

This work 44 7 .16 12 .24 15 .09

Garner (1961) 43 7 .16 12 .21 15 .06

Debets and Loopstra (1963) 86 7 .220 12 .198 15 .072

comparison of observed and calculated d-values for our however, that these patterns can be best explained as
compound 3 with the d-values reported for their mixed crystalline phases . Electron micrographs of such
compound 3 . samples support this position since they invariably
A decrease in crystallinity and concomitant line indicate the presence of two types of crystal morphology,
broadening are characteristic of increasing ammonia as described in the next section .
content in the crystal structure[3] . This effect is
additionally complicated in our samples by the very small Electron microscopy results
crystallite size . However, despite the diffuse patterns Only two types of crystal morphology were observed in
obtained from compounds 2 and 3, it is significant that the samples in which crystal growth occurred due to
pattern are reproducible and that even small variations in excessive ultrasonic treatment or water aging . However,
the three basic patterns due to changes in process several effects were observed in unaltered samples . These
conditions are readily detectable . included a characteristic crystallite size and shape for
Variations were observed in the diffraction patterns for each of various feed compositions ; noncrystalline
several samples . Typically, such variations are manifested amorphous material ; and systems which appeared to
by the absence of some of the weaker reflections in the consist of mixed crystal phases .
principal pattern, plus one or more lines which could be Figure 2 shows electron micrographs that are charac-
assigned to a second pattern . The reason for the variations teristic of the two types of crystal morphology obtained
cannot be unambiguously determined since several when crystal growth occurred during sample preparation .
interpretations could have merit . It appears to us, These were identified as compounds 1 and 2 by X-ray


1 144 M. H . LLOYD et a!.

Table 2 . Powder data of compounds 2 compared with correspond-

ing reflections observed by Garner[1] and by Debets and
Loopstra [3] . Calculated with A= 1 . 54178 A (CuK0 )

d (A) d (A) d (A)

I obs h k 1 I Garner I Debets et al .

w 9 .54 0 1 1 vvw 9 .48 1 9 .47

m 7 .54 0 0 2* vs 7 .55 100 7 .53
m 6 .15 0 2 0* vvw 6 .13 2 6 .10
m 5 .72 1 1 1* vw 5 .72 3 5 .70
w 4 .77 1 1 2 vw 4 .77 1 4 .77
-- 0 1 3 vw 4 .63 4 4 .64
w 4 .45 1 2 1 - -- 1 4 .43
w 3 .771 0 0 4* m 3 .766 70 3 .766
vs 3 .575 2 0 0* m 3 .579 25 3 .570
vs 3 .539 1 3 0* s 3 .535 50 3 .533
-- 1 2 3 vvw 3 .407 3 3 .408
vs 3 .235 2 0 2* m 3 .233 40 3 .228
vs 3 .195 1 3 2* s 3 .200 70 3 .200
w 3 .089 2 2 0 - -- 1 3 .081
w 3 .060 0 4 0* vvw 3 .054 1 3 .049
w 3 .021 2 2 1 vvw 3 .024 3 3 .019
-- 0 1 5 vvw 2 .928 4 2 .927
m 2 .865 2 2 2 vvw 2 .854 2 2 .852
m 2 .831 0 4 2* vvw 2 .829 4 2 .825
vw 2 .642 2 3 1 - -- - --
m 2 .594 2 0 4* w 2 .593 9 2 .591
m 2 .578 1 3 4* m 2 .576 25 2 .577
vw 2 .534 2 3 2 - -- 2 2 .528
-- 0 0 6 vw 2 .509 9 2 .512
w 2 .416 0 5 1 - -- 1 2 .408
m 2 .324 2 4 0* vw 2 .322 4 2 .319
m 2 .313 1 5 0* vw 2 .310 - 2 .308
w 2 .287 1 5 1 - -- 1 2 .282
m 2 .224 2 4 2* vvw 2 .215 2 2 .216
m 2 .209 1 5 2 vvw 2 .212 3 2 .208
vw 2 .124 0 1 7 vw 2 .122 4 2 .120
3 2 .059 3 3 0* m 2 .055 - 2 .054
-- 1 3 6 mw 2 .049 15 2 .047
R 2 .038 0 6 0* vw 2 .036 4 2 .033
5 1 .986 3 3 2* w 1 .985 8 1 .982
2 1 .986 0 6 2* vw 1 .967 5 1 .963
-- 1 2 7 vvw 1 .954 2 1 .952
-- 0 0 8 vvw 1 .885 4 1 .884
w 1 .874 1 4 6 - -- 1 1 .870
1 .807 3 3 4* w 1 .806 7 1 .803
m 1 .790 0 6 4 w 1 .789 2 1 .789
a 1 .771 2 6 0* vw 1 .770 3 1 .767
m 1 .741 4 0 2 vw 1 .742 3 1 .737
m 1 .724 2 6 2* 1 .722 4 1 .720
m 1 .707 2 4 6 1 .704 3 1 .704
m 1 .697 1 7 0* 1 .693 - 1 .693
1 .676 4 2 2 -- 1 1 .671
1 .666 3 5 2 w 1 .664 . - --
1 .656 1 7 Z* vw 1 .654 3 1 .652
1 .617 4 0 4 vw 1 .617 3 1 .614
1 .603 0 4 8 vvw 1 .603 - -- Fig . 2 . Transmission electron micrographs of (a) compound 1, and
(b) compound 2 .
*Denotes reflections which are observable on diffuse patterns .

which crystal growth resulted from excessive ultrasonic

Table 3 . Powder data for compound 3 and corresponding data of treatment.
Debets and Loopstra[3] for the compound 2UO 3 •N H 3 . 3H2O Figure 3 presents electron micrographs showing the
principal crystal morphologies observed in unaltered
d (R) d (R) d (a ) samples . Compound 1 is principally characterized by
I obs talc h k i 1 I Debets et al . *
relatively large crystals with diameters as great as 5000 A,
vw 12 .6 12 .1 1 0 1 0 -- although the crystal size varies over a wide range .
vs 7 .5 7 .5 0 0 0 2 100 7 .25 Compound 2 consists primarily of very small crystallites
w 6 .02 6 .o8 2 0 2 0 3 6 .09
(dia. = 100-200 A) ; however, some samples were ob-
w 3 .76 3 .76 0 0 0 4 22 3 .63
vs 3 .51 3 .51 2 2 4 0 6o 3 .52
served to contain small amounts of a larger crystal phase .
vs 3 .19 3 .18 2 2 4 2 70 3 .17
When this occurs, both phases appear to be compound 2
w 2 .56 2 .56 2 2 4 4 18 2 .53 because the X-ray diffraction patterns show no detectable
w 2 .29 2 .30 4 2 6 0 3 2 .31 change with variations in the ratio of the two phases .
vw 2 .20 2 .20 4 2 6 2 2 .20 Compound 3 was observed both as a very small phase and
Is 2 .03 2 .03 6 0 6 0 10 2 .03
as a large phase . These two phases are not present in the
m 1 .96 1 .96 6 0 6 2 11 1 .96
vw 1 .77 1 .78 6 0 6 4 5 1 .77
same preparation, but result, rather, from variations in
vw 1 .71 1 .71 4 4 8 2 4 1 .71 feed preparation (see the next section) .
vw 1 .64 1 .64 6 2 8 2 3 1 .65 The composition of some samples can best be explained
vw 1 .58 1 .59 4 4 8 4 2 1 .58 as mixed crystal forms . A typical electron micrograph is
shown in Fig . 4(a) . The X-ray diffraction patterns of such
All main reflections (I >5) and a part of the weak reflections samples could not be unambiguously resolved ; however,
observed by Debets and Loopstra [33 . patterns corresponding to compound 1 plus extra lines
(which correlate with compounds 2 or 3) were obtained .
diffraction . Each crystal form consists of very thin When such patterns are observed, the corresponding
platelets; and, although the crystal diameter is primarily a electron micrographs invariably show two types of crystal
function of sample preparation, the distinctive difference morphology, i .e. one with large and one with small crystal
in crystal shape is an inherent characteristic . The crystal sizes . It appears likely that the large crystallites are
form obtained was characteristic of the feed used to compound I and that the small ones belong to a phase
prepare the microspheres . At N03 /uranium mole ratios containing NH3-
of 1 . 5, compound 1 occurred at hexaluranium mole ratios A distinctive amorphous material with fibrous habit can
of 0 . 9-1 . 2 and compound 2 was obtained at ratios of also occur. An electron micrograph of mixed crystalline-
1 . 3-1 .6 . Compound 3 was not observed in samples in amorphous material is shown in Fig . 4(b) . Unlike the
 Fig. 4 . Transmission electron micrographs of mixtures of two morphologically
Fig. 3 . Transmission electron micrographs of principal types of crystal morphology observed for compounds 1-3 . (a)
different crystalline substances . (a) Mixed large and small crystal forms ; (b)
Compound 1 ; (b) compound 2 ; (c) compound 3, small phase ; (d) compound 3, large phase . mixed crystalline and amorphous phases . a
U


1 146 M. H . LLOYD et alL

crystalline phases, the fibrous material does not give an Table 4. Variations in crystal structure and morphology with
electron diffractogram of a crystalline substance. chemical process variables
N03 /U Hexa/Uranium Crystal
Correlation of crystal form and morphology with process Mole Ratio Mole Ratio Diameter (a)* Compound
variables
As previously indicated, this study was designed to 1 .5 0 .9 2500-5000 1

better understand process variables during the prepara- 1 .5 1 .0 1

tion of dense microspheres of urania and urania-plutonia 1 .5 1 .1 1
oxides or carbides . Two of the principal chemical
1 .5 1 .2 1000-2500 1
variables in this system, the hexa/uranium and the
<100-200 3
N03 /uranium mole ratios, were studied in conjunction
with several physical process variables which potentially 1 .5 1 .3 100-250** 2
1000-2500 2
influence crystallization behavior . Microspheres were
1 .5 1 .4 100-250** 2
dispersed and examined at several stages in the process to
1000-2500 2
aid in the evaluation .
When the cold feed solution is heated, uranium 1 .5 1 .5 <100-250 2

hydrolysis occurs because NH 3 is liberated by decomposi- 1 .5 1 .6 <100-250 3

tion of the hexa . Therefore, increasing the hexa/U mole 1 .75 1 .2 2000-5000 1
ratio in the feed increases the concentration of NH,OH 2000-5000 3
which is present during hydrolysis and crystallization . 1 .75 1 .4 2000-5000 3
The effect of this variable was studied primarily with feed
1 .75 1 .6 2000-5000 3
solutions in which the uranyl nitrate was prehydrolyzed
with NH,OH to provide an effective N03 - /U mole ratio of
1 .5 because microspheres can be prepared using a wide *Although the size range indicated is characteristic of the
range of hexa/U mole ratios under these conditions . In preparation, both smaller and larger sizes are generally observed .
this study, the ratio was varied from 0 .9-1 .6. The physical **The small size is predominant, and relatively few large
appearance of air-dried microspheres changes considera- crystals are observed .
bly as this ratio changes ; for example, spheres prepared at
low ratios are opaque and light yellow while those were not prepared at ratios below 1 .2 because excessively
prepared at high ratios are transparent, orange-brown, long gelation times are required . All the microspheres
and much smaller . At intermediate ratios the spheres prepared at this N03 /U ratio were opaque and light
become progressively smaller, less opaque, and darker in yellow in color, indicative of large crystallites ; this was
color as the ratio increases, which suggests that the again verified by electron micrographs (Table 4) . Com-
crystallite or particle size of the precipitated urania pound 3 was obtained for hexa/U ratios above 1 .2, while a
decreases as the hexa/U mole ratio increases . This general ratio of 1 .2 produced a mixed phase which appeared to
trend was verified by the electron micrographs . consist of compounds 1 and 3 . Crystallite diameters in all
At a N03 /U mole ratio of 1 . 5, three types of crystal cases were predominantly in the range 2000-5000 A .
morphology can be obtained when the hexa/U mole ratio There is probably a decrease in average crystal size as the
is varied from 0 .9-1 .6. This change occurs because some hexa ratio increases, but the change is not sufficiently
of the water of hydration in the compound is replaced by great to determine by qualitative observations . The large
NH3 . effect of variations in the N03-/U mole ratio on crystal
The data obtained for eight hexa/U mole ratios are size was not expected ; nevertheless, it explains several
tabulated in Table 4. As indicated, only compound 1 is aspects of sphere-forming behavior and affords a mechan-
obtained at ratios 0 .9-1 . 1 and is the principal form ism by which, crystal diameters can be controlled within
obtained at 1 .2 . Multiple crystal diameters can be certain limits .
observed in most preparations ; on the other hand, there is The results described above were obtained when
a characteristic size range for each preparation which conditions were such that a high degree of crystallization
decreases with increasing hexa/U ratio . This decrease is occurred . When gelation takes place too rapidly, e .g.
apparent from line broadening in the X-ray diffraction when the oil temperature in the sphere-forming column is
patterns as well as from the electron micrographs . too high, complete crystallization is not obtained . This
The intermediate ratio of 1 .2 appears to consists of effect was demonstrated in a series of experiments with
mixtures of compounds 1 and 2, the relative amounts of various feed compositions . Microspheres were prepared
each phase depending on several physical variables . at different gelling rates by varying column and feed
These effects are described below . Only compound 2 was temperatures . A minimum gelation time for complete
observed at ratios of 1 .3-1 .5 . In all cases, a very small crystallization was required to each composition, and the
crystal size was predominant for these ratios ; however, a time required varied with both the hexa/uranium and the
significant but small concentration of large crystals was NO3/uranium mole ratios . If the gelation time was too
noted at 1 .3, while less of these were obtained at 1 .4 and short, catastrophic cracking of the microspheres resulted
none were observed at 1 . 5 . Since the X-ray patterns for during air drying or calcination . Transmission electron
the three compositions are identical, it appears that the micrographs of such material show high concentrations of
two crystal sizes have the same chemical compodition . amorphous material as well as urania crystallites (Fig . 4).
At a ratio of 1 .6, compound 3 was always obtained and A comparison of these micrographs with micrographs of
only a small crystal size was observed . highly crystalline products allows the conclusion that the
The effect of varying the N03/U mole ratio was crystal size is essentially the same in each case . These
evaluated by comparing the above results with experi- factors suggest that crystal size and the attendant gelation
ments in which the N03/U mole ratio was 1 . 75 and the time requirement are determined by the chemical
hexa/U mole ratio was varied from 1 .2-1 . 6 . Microspheres composition of the feed solution, while the physical


Crystal habit and phase attribution of U(VI) oxides in a gelation process 1 14 7

factors that affect the gelation rate must be selected to metal ratios are used ; this appears to be due, at least in
provide adequate reaction time for complete crystalliza- part, to variations in the crystal form of the urania .
tion . The very large effect on crystal size with variations in
The precise gelation time requirements for various feed the NO3 /U mole ratio is also a highly significant process
compositions depend on several parameters, including the variable . It affords a mechanism for altering the crystallite
mechanical operations of the sphere-forming column, and size without changing the chemical composition . In
therefore will not be described here ; however, several addition, it provides an explanation for an important
generalizations can be made . aspect of microsphere preparation . For example, during
The gelation or reaction times involved are of the order the preparation of (U, Pu)C microspheres, the best
of 2-6 sec . The time required for complete crystallization products were generally prepared at high hexa/U mole
decreases with increasing hexa/U mole ratio at a given ratios . This behavior was difficult to correlate with the
N0,-/U mole ratio, and is much more critical at low hexa concept of the very small crystallite size, which was
ratios where the crystallite size is large . Similarly, the thought to be produced at high ratios because such
gelation time is very critical for a NO3 /U mole ratio of material has a much greater tendency to crack during
1 .75 because of the large crystal size . For example, drying and sintering (this tendency is due to the very high
changing the column temperature by 5°C, which alters the surface area of the small crystallites and the smaller pore
gelation time by less than I sec, significantly affects size in the gelled spheres) . Such behavior can now be
formation of the larger crystals (>1000A) and the understood ; when plutonium nitrate is introduced into the
cracking behaviour of the microspheres . system, it is accompanied by a small amount of HNO,,
A useful process option involves curing gelled spheres which increases the effective N0 ; /U mole ratio of the
in hot oil after their formation . This technique improves uranyl nitrate and therefore increases the crystallite size
the physical characteristics of microspheres prepared obtained in the gelled sphere . Because of this effect, high
from most feed compositions . Transmission electron hexa/uranium mole ratios can be tolerated in mixed
micrographs of crystallites from cured and uncured plutonium-uranium feeds ; and high ratios are desirable
spheres indicate that the principal effect of curing is an because of their fast gelation rates, which counteract the
increase in crystallite size and, to a limited degree, decrease in gelation rate caused by increasing the NO, /U
continued crystallization of amorphous material . How- mole ratio . Such factors must be balanced in an operating
ever, when preparations contain large amounts of system because an excessively slow gelation rate will lead
amorphous material, the crystallization will not be to distortion of the microspheres and, in an extreme case,
sufficient to prevent subsequent microsphere cracking . will prevent gelation .
Curing has a very pronounced effect on product prepared A third highly significant factor is the interrelationship
at a hexa/U ratio of 1 .2, which contains two crystalline of feed composition and gelation rate . The chemical form
phases . Typically, such material converts primarily to and crystal morphology are determined by the feed
compound 1, thereby significantly improving the physical composition, although an appropriate gelation rate must
characteristics of the microspheres . be determined for each composition . If gelation occurs
Examination of microspheres both before and after too rapidly, crystallization will not be complete and the
washing showed that crystallization occurs during the resulting high concentration of amorphous material will
initial gelation ; however, interesting changes in the X-ray invariably lead to catastrophic cracking of the micro-
diffraction pattern were frequently noted during the sphere during air drying .
sphere-washing step . This was particularly true for It is apparent from the above considerations that
NH,-containing crystal material, in which diffraction lines correlation of the process variables with crystallization
found in the NH,OH-washed sample were frequently behavior by the techniques used in this study affords a
missing ; conversely, an additional line was sometimes useful method to better understand processes of this type
observed in samples prior to washing . A detailed analysis and provides a means for optimizing process variables .
of these features is beyond the scope of this paper .
Examination of microspheres both before and after air
drying indicates that no changes occur during this step .
REFERENCES
CONCLUSIONS 1 . E . V . Garner, J. Inorg. Nucl. Chem . 21, 380 (1961) .
The crystal morphologies obtained in the experiments 2 . E . H . P . Cordfunke, J. Inorg . Nucl. Chem . 24, 303 (1962) .
reported here are not highly suited for study of the 3 . P . C . Debets and B . 0 . Loopstra, J. Inorg. Nucl. Chem . 25,945
crystalline ammonium uranate system . However, our (1963) .
interpretation of the data supports the idea of chemically 4 . H . R . Hoekstra and S . Siegel, J. Inorg. Nuc!. Chem . 35, 761

defined compounds rather than a continuous solid (1973) .

5 . R . Sobry, J . Inorg . Nucl. Chem . 35, 1515 (1973) .
solution system . It seems possible that mixtures of two
6 . W . I . Stuart and D . J . Miller, J. Inorg. Nucl. Chem . 35, 2109
phases may give the appearance of a continuous system, (1973) .
since such mixtures should alter the X-ray diffraction 7 . F . W . v . d . Brugghen et al., A U-Process for Microsphere
pattern observed for a single phase in a complicated way . Production, presented at the symposium on so]-gel processes
With regard to internal gelation processes, several and reactor fuel cycles, Gatlinburg, Tennessee (May 1970) .
observations may be emphasized . The knowledge gained 8 . Plutonium-Brennstoffprogramm, Jahresbericht 1971, EIR-
concerning the composition of the phases that can be Bericht Nr . 207 (April 1972) .
9. K . Bischoff, M . H . Lloyd and H . Schumacher, Sol-Gel
obtained, and more particularly their NH3 content, is
Processes for Carbide Preparation, EIR-Bericht Nr . 236 (May
important for carbide preparation because NH3 is a strong
1973) .
reducing agent at elevated temperatures . This factor is
10 . J . K . Dawson et al., J. Chem . Soc. 35, 31 (1956) .
important in determining the carbon content required for
11 . H . W . Dunn, USAEC Rep. ORNL-2092, Oak Ridge National
reaction sintering . Variations in carbide stoichiometry Laboratory, Oak Ridge, Tennessee (1956) .
have frequently been observed when constant carbon-to- 12 . C . L . Christ and J . R . Clark, Am . Mineral 45, 1026 (1960) .