Refractory Metals & Hard Materials 11 (1992) 35-41

Pure Vanadium Pentoxide from Boiler Ash

S. Z. E1-Tawil, a I. M. Morsi, a I. S. A h m e d Farag b & F. H. A. A b d a l l a a
"CMRDI, PO Box 87, Helwan, Cairo 11421, Egypt NRC, Dokki, Cairo, Egypt (Received 27 August 1991; 3 February 1992)

Abstract: The parameters affecting the hydrolysis of sodium vanadate liquors

from boiler deposits and the purity of the produced V20 ~ were investigated. Hydrolysis of vanadate liquors in the presence of sulfuric acid gave about 95 % V20 ~ equivalent at pH 2"5 and temperature 90C. Activation energies of the first and second stage of hydrolysis are 16.7 and 27-6 k J/mole, respectively. Addition of oxidants such as HCIO 4 or HNO 3 in the presence of NH4CI leads to complete recovery of V20~, giving pure product having total surface areas of 2-91 and 2.16 m2/g, respectively, after the calcination of the produced red cakes.

INTRODUCTION Technical-grade vanadium pentoxide is essentially used as one of the raw materials for the preparation of ferrovanadium alloys either by silicon, ferrosilicon or aluminium reduction methods. Pure vanadium pentoxide is also used as catalyst for many chemical processes in various industries. A m o n g the methods used for its preparation is the oxidizing roasting technique in the presence of sodium salts such as sodium chloride, sodium carbonate or sodium sulfate at high temperatures.1 5 The formed sodium vanadates are then leached with hot water and hydrolysed by suitable dilute acid to produce the technical-grade vanadium pentoxide (known as red cake). 68 The hydrolysis process is therefore very critical and it affects the grain size and the final composition of the produced vanadium pentoxide. This process needs more studies to be clarified and to obtain a well-identified vanadium pentoxide product with specific granulation. In Egypt, the primary sources of vanadiumbearing minerals are rather limited. It is essential to use secondary sources such as the boiler deposits produced from combustion of mazot used as a fuel in thermal power stations and other allied industries 35

as a source to produce the technical-grade vanadium pentoxide and other related valuable nonferrous compounds (viz. nickel). In previous investigations carried out in our laboratory, the production of ferrovanadium ferronickel and vanadium pentoxide from boiler deposits of local power stations was performed. 3.4,9.1 The present paper is an extension of this work and it deals with a study of the influence of different hydrolysis parameters on the composition of the technical-grade vanadium pentoxide and to obtain pure vanadium pentoxide with characterised grain size. The produced vanadium pentoxide was evaluated physically and chemically.
EXPERIMENTAL Raw materials

A representative sample of boiler deposits from local power stations was collected and upgraded by a convenient and feasible technique previously carried out in our laboratory and reported elsewhere2 Chemical analysis of the main constituents of the upgraded sample and their size distribution are given in Table 1. Other materials used in this study were of chemically pure grade.

Refractory Metals & Hard Materials 0263~4368/92/$05.00 1992 Elsevier Science Publishers Ltd, England. Printed in Great Britain.

36

S. Z. El-Tawil, I. M. Morsi, L S. Ahmed Farag, F. H. A. Abdalla

Table 1. Chemical analysis a n d the particle size distribution of the u p g r a d e d boiler deposit

100
>

O---

Particle size distribution of upgraded boiler deposit Size (mesh)

- 8 0 + 120 - 1 2 0 + 160 - 160+200 -200 - for undersize. + for oversize.

Chemical analysis of upgraded boiler deposit Element %

80

Wt %

o 60
E ~ 40

45 31 15 9

V Fe Ni

31 12 7"5

> 20 0
I

O ~ p . =90C Hydrolysis time =3hrs pH=Z5

I I I I

10 20 30 40 Vanadium concentration, g/[

50

Fig. 1. Effect o f v a n a d i u m ion c o n c e n t r a t i o n on the v a n a d i u m recovery.

Formation of mother sodium vanadate liquor

1 2 5 [V a n a d i u m ~ ion c o n c . = 3 2 g / t f ab ad ef

A concentrate of boiler ash deposit was mixed with sodium carbonate in a molar ratio of 1:1 Na20/V205 and the mixture was roasted at 700C for one hour in a muffle furnace under air. These conditions were shown to be optimal from our previous work where about 99.9% of vanadium was recovered? '4 The formed sodium vanadates were leached with hot water to form a mother sodium vanadate liquor.

100[-

sol75 50

75t-

d a cf ~ ,
f f Io,;bb I b&

If

la

,~"=
"~

Vanadium

+~ C

> 250 ~ --~

n.," 75-

ad

~ a

ion c o n c . = 2 2 . 5 g / t e_ f d c fe c f b af C e~ b e c a a d b b c C bba a C ebf - -

__
Vanadium

dL
ion c o n c . = 15-8 g/L
b

~
d a a

~f

fc

fe

f
a

50-

Hydrolysis process
This process was carried out on solutions having different vanadium concentrations from 7.5 to 50 g/litre, at different pH values from 1 to 3, for various periods from 1 to 5 h and at different temperatures from 50 to 90C. Dilute sulfuric acid solution was used as acidifying agent to control the pH to precipitate the commercial grade V205 (red cake).

251,2

~@ e a
2

"

I, b a I o10 d

b bb @
3

d,~
Fig. 2. X-ray p a t t e r n s of the dried cakes o b t a i n e d from various v a n a d i u m ion c o n c e n t r a t i o n s at 90C for 3-h hydrolysis time a n d at c o n s t a n t p H = 2.5. (a = V204, b = V407, c = VaO 5, d = V~O~, e = Na2SO 4 a n d f = V~Og. )

Calcination process
The red cakes were dried and then calcined at different temperatures from 300 to 800C for 3 h in a muffle furnace. The dried and calcined cakes were chemically analysed and identified by X-ray diffraction. The surface area of some selected samples was also measured. R E S U L T S AND D I S C U S S I O N

Effect of vanadium concentration

Hydrolysis was carried out on solutions having different vanadium ion concentrations (7-5-

50 g/litre) at 90C for 3 h and at constant pH of 2-5. The results, given in Fig. 1, show that the precipitated vanadium pentoxide increases from 21 to about 97% by increase of the vanadium concentration from 7"5 to 32 g/litre after which a steady state is attained. The chemical analysis of the dried cake at these optimal conditions shows that it consists of about 87 % V~O 5 equivalent. Recovery of V~O5 equivalent of calcined sample at 550C reaches about 89 %. Figure 2 shows the X-ray patterns of the dried cake obtained from various vanadium concentrations. These patterns reveal that the cakes consist mainly of a mixture of vanadium oxides V204, V407, V305, VsO 9 and V20 ~ having a general formula V~O~. nVO2, where n varies from zero to three. It is apparent that the vanadium ions under the above conditions are precipitated in tetra- and trivalent valences with different proportions. The

Pure vanadium pentoxide from boiler ash

lOO Vanadium conc.=32g/( ~ / ~ : ] 1 - - - - ~ 0 90oc lOO 75 50 pH =3

37

~" L 80 -- Hydrolysis t i m e ~ . / ~ ~

IllL ,I

L I

L,

8 E

6o

~. 7 5 -

pH=2.5

c_

li
pH =2

_~ 5o
o
75

m 75~ >

ILh, ,I
aa @e c I

I I
cd e ed ab ceab

,,
c

pH Fig. 3. Effect of pH on the vanadium recovery at different hydrolysis temperatures.

c f f'f f fe ~el i

0 pH=l !

ii

L
i

25

''ta~

11~ 2ILl ?

iT
d,~,
a

cb

, 11

pentavalent vanadium did not appear under these conditions. X-ray analysis of calcined samples at 550C reveals the presence of the phases V80 u, VTOt.~ and VsOts.

oop
o 75F
~

50 f 25
0 1
I

, X@ o 2
:

pH of the solution
Hydrolysis experiments were conducted at various pH values using a solution containing vanadium at 32 g/litre and at different temperatures 50-90C and for a constant period of 3 h. It is obvious from Fig. 3 that at each temperature, the vanadium recovery increases up to pH = 2"5 after which the recovery decreases. Also the recovery increases from 32 to 97"5 % for a temperature increase from 50 to 90C, which is the optimum value. The chemical analysis of the dried cake at this optimum condition showed that it consists of 86"3 % V205 equivalent. However, the calcined sample at 550C consists of about 89 % equivalent V205. X-ray patterns of the dried cakes (Fig. 4(a)) obtained at 90C after hydrolysis at different pH values reveal the presence of V204, V407, V30 ~, V~O3 and V,-,O:,, indicating the predominance of the tetra- and trivalent states of vanadium. Comparing the relative intensities of lines corresponding to these phases for the different cakes, Fig. 4(b) reveals: (i) no dependence of VaO ~ and V203 lines on the pH; (ii) the relative intensities of lines corresponding to the phase V20 ~ increase with increasing the pH until they attain a constant value after pH = 2; and (iii) the relative intensities of the lines corresponding to the phases V~O9 and V407 decrease with increase of pH to a maximum at pH of 2-5 after which they increase again. It is obvious, therefore, that the increase in pH in the hydrolysis process results in a partial conversion of the trivalent vanadium to the tetravalent state. This

(b)

pH

Fig. 4. (a) X-ray patterns of the dried cakes obtained at different pH values; (b) variation of relative intensities of the phases with these pH values. (a = V20 ~, b = V~O 7, c = V:~O~, d = V,.,O3, e = Na2SO ~ and f = V~O:).)

conversion is accompanied by an increase in the vanadium recovery at pH values 2-2"5. The formed tetravalent vanadium is again converted to V 3+ at pH > 2"5 by forming polyvanadates. This behaviour leads to an decrease in vanadium recovery at pH > 2.5.

Hydrolysis isotherms
It is obvious from the isotherm results in Fig. 5 that the vanadium recovery increases with time and temperature of hydrolysis to approach its maximum
tOO

90oc

~ 80
>

~ 6o E ~5 40
r0 c m

2O

IO0

Vanadium conc.= 32 g/( pH = 2"5 1 i 200 300 Time, rain

Fig. 5. Effect of time and temperature of hydrolysis on vanadium recovery at constant pH = 2"5.

38

S. Z. El-Tawil, L M. Morsi, L S. Ahmed Farag, F. H. A. Abdalla

10o 75 50
~ 25 0-247 0'60 0-71 0"75

Table 2. The rates of precipitation of the first and second stages

HydroLysis time, rain.

Temperature (dR/dt) first stage (c)

50 70 80 90 0" 128 0" 156 0"185 0"25

(dR/dt) second stage

~

180

III,;,I
-

IIII ,I,
I d

I~

11
120

c~ 75 o ~ 5o 25 >

,, lib, ,1
?a~)(E)e
(~b

] Ill ,11
ab

,,

II
6O
b

a: 75
0 5o 25 (a)

e~e ~ , .

o First stage Second stage

;oflII~p,1 I rTII ,~,1 I3 , 2

/
~ 60 120 180 Time, min

I fC a d el e C(~)IT e b cfd a beb b , *

i 4

rv

-0.5
-

100 r 751

d spacing, A d,.& o @
e

_J

~5o f 25 0
(b) -1.0

e.@ I I 240 300

Fig. 7. (a) X-ray patterns of the dried cakes obtained at 90C and p H = 2-5 for different periods; (b) variation of relative

I 27

I 28

I 29

I I 30 31 l i t , xlO -3

I 32

I 33

I 34

intensities of the phases obtained with hydrolysis time. ( a = V ~ O 4, b = V 4 0 7 , c = V 3 0 5 , d = V ~ O 3, e = N a 2 S O 4 and

f = V509. )

Fig. 6. Effect of temperature on the rate of hydrolysis at first

and second stages.

of about 97 % at 90C for 180 min. It is apparent that the hydrolysis process proceeds through three stages: (i) the first stage at the beginning of precipitation at 90C lasts for 80 rain and becomes longer at lower temperatures, up to 120 mins at 50C; (ii) the second stage is characterised by an abrupt increase in the recovery at about 80 min; (iii) the third stage is for long periods where a steady state is attained. The gradients of the curves at the first and second stages, which indicate the rate of precipitation, are given in Table 2. It is clear that the rate of precipitation of vanadium oxides is very low at the beginning of the hydrolysis process and then increases abruptly. This can be explained thus: at the first stage, the nucleation of vanadium oxides takes place, followed by nuclei growth during the second stage. At the end of the second stage, most of the vanadium ions in the solution are deposited and a steady state is attained. The activation energies of the first and second stages were calculated by plotting log (dR~dO of the first stage and (dR/dt) of the second stage against 1/T and applying the Arrhenius equation (Fig. 6). The values obtained are 16-7 and 27.62 k J/mole for the first and second stages, respectively. This indicates that the second stage is highly dependent on the hydrolysis temperature. Figure 7(a) shows the X-ray diffractograms of the

dried cakes obtained from the hydrolysis process at 90C for different periods 60, 120 and 180 rain. The phases identified are V~O 4, V4Q, V305, V~O 3 and V509, which indicates the existence of the tetra- and trivalent vanadium. Comparing the relative intensities of the phases present (Fig. 7b) shows that the lines corresponding to the phases V204, VsO 9 and V20 ~ do not change with hydrolysis time. On the other hand, the intensities of the lines corresponding to the phase V407 and V~O s decrease slightly with increase in time of hydrolysis. Calcination of red cake In order to get an identified V~O s product, the red cake obtained at optimum conditions is subjected to
4~3-

o 2

/
~ 300
I

o o

400

500 600 700 800 Roasting temperature, C

900

Fig. 8. Effect of calcination temperature on the weight loss of red cake obtained at optimum conditions; p H = 2.5, hydrolysis temperature 90C and hydrolysis time 3 h.

Pure vanadium pentoxide from boiler ash

100 F mm t ~
100

39

o 75[-

kltkr~

L 2

3 d spacing, ~,

1 4

80

Fig. 9. X-ray pattern of calcined sample at 550C for 3 h. The sample was prepared at optimal conditions; pH 2'5, hydrolysis temperature 90C and hydrolysis time 3 h. (1 = V,O u, k = VsOl~ and m = VvOta.)

$
> o L

60

calcination at various temperatures for a constant period of 3 h. The results in Fig. 8 show the variation of the percentage of weight loss as a function of calcination temperature up to 800C for 3 h. It is obvious that the weight loss increases gradually from 0"3 to 0.8% by increase in temperature from 300 to 500C, respectively. This is due to the volatilisation of chemically bonded water. The weight loss is abruptly increased from 0.8 to 3.2% by rise in temperature to 550C. No variation in loss in weight is observed at higher temperatures ( > 550C) and melting of the product is noticed. Analysis of the calcined product at > 550C revealed that it contains 88"89% V205 equivalent. The abrupt loss in weight is probably due to the dissociation of some lower vanadium oxides with removal of some oxygen (about 3%) as in the following equations : 3 V.,O4------~V6O11+02 2 V40<---~VsO15 + 0., (1) (2)

E
E >
40

V a n a d i u m conc.= 32g/I
Hydrolysis temp.=90C pH = 2.5 Type of oxidant

(1Oral, 1M sol.)
o Without oxidant

20

A X

H202 HNO 3 KMn04 HCIO 4

I 100

i 200

l 300

H y d r o l y s i s t i m e , rain

Fig. 10. Effect of the presence of different oxidants on the hydrolysis rate of vanadium liquor.

~ooF
75 f
50

HCI04

25 O
75t
/

,,,,,

IL,liT

t
KMn04

o5O 25,~ ,,,

~_ 5op

in , 1 ;,,Ill
I I

, I

The existence of these phases in the calcined samples at 550C was confirmed by X-ray investigations (Fig. 9). The main phases identified are V601~, VsOt.~ and VvOt3 which are the same phases present in X-ray patterns of samples calcined at temperatures > 550C and up to 800C. Oxidant additions The former results indicate that it is not feasible to produce pure V.,O 5 only by calcination. Addition of an oxidant to the liquor before the hydrolysis process to convert the tetra- and trivalent vanadium ions to pentavalent ions, was tried. Different oxidants, H20 ~, NHO 3, HC10 4 and KMnO 4, were tested. It is clear from the results in Fig. 10 that the period for vanadium recovery of about 95% decreased from 178 to 116, 150, 170 and 174 minutes for solutions containing HC104, KMnO 4, HNO3 and H.,O 2 at 90C and pH 2"5, respectively.
4

_~ o >~25~/
cv 75[50 F

II~ 9'
o I
d b

iLl fill
, i ?

,
h

I
H202

, I
]

2q o/
75

I ,I

vg,o01v;I h r L,, I ,,I , ,

f e c e Ic f ere

Without oxidant
bf

2
0

5
1 "

ac~ C ~ d
2 2

a
3

4 d spacing,

Fig. 11. X-ray patterns of the red cakes obtained at optimal conditions and in the presence of different oxidants. ( a = V 2 0 4 , b=V4Ov, c = V a O s , d = V ~ O 3, e=Na2SO4, f = VsO~, g = Na20. X(V204). Y(VsOs), h = V409 and i = X(V204). Y(V~Os). n(HsO).)

However, complete vanadium recovery is achieved after 240 min for the addition of H,,O2, HNO 3 and KMnOa and after 180 min for the oxidant HC10 4.
ERM 11

40

S. Z. El-Tawil, I. M. Morsi, L S. Ahmed Farag, F. H. A. Abdalla

100 A

Table 3. The surface areas of the dried red cakes obtained in presence of oxidising agents
Type of oxidant (Quantity added = 0"01 M) Surface area (m2/g)

80

Without oxidizing agent H202 KMnO4 HNO 3 HC104

3-47 1-79 7-89 10-65 16"02

>:,
60
>

It is also shown that the rate of hydrolysis in presence of oxidants increases in the order H202 < H N O 3 < KMnO~ < HC1Q. Chemical analysis shows that the dried cake contains about 88% V205 equivalent. X-ray diffractograms of the dried cakes obtained in the presence of different oxidants are shown in Fig. 11. The presence of lines corresponding to the phases N a 2 0 . X(V20,). Y(V2Q ), V~O 9 and X(V204). Y(V2Os). n(H20 ) proves that vanadium is precipitated mainly in the form of a mixture of penta- and tetravalent vanadium. It also shows that the addition of the aforementioned oxidants did not completely convert the vanadium ions to the pentavalent state under the applied conditions. The results of the surface area measurements of the dried cakes (Table 3) show that they increase as a result of oxidant presence. They approach 16.02, 10.65, 7.89 and 1.79mZ/g for HC1Q, H N O 3, KMnO~ and H~O 2, respectively. It is concluded that the more powerful the oxidising agent used, the finer the precipitate obtained. The precipitated V~O5 is granular in nature and its particle size is governed by the type of oxidant added.
Ammonium chloride addition

E
::3 t-

40

>

Vanadium c o n e . =32g/I Hydrolysis temp. =90C pH =2.5 A HN03 (lOml 1M sol.) + 5 g NH4CI HC104(10 ml 1M sol.) + 5 g NH4CI

20

100 200 Hydrolysis time, min

Fig. 12. Effect of presence of ammonium chloride on the rate of hydrolysis of vanadium liquor containing oxidants.
loo F
o 75 b

V205

5of
25
d| ,

ll,ll,llt,
2

II

11

3 4 d spacing,

Fig. 13. X-ray pattern of the calcined red cakes obtained from the hydrolysis of boiler ash vanadium liquor (32 g/litre) in the presence of ammonium chloride and one of the oxidants nitric acid or perchloric acid. is mainly V 2 Q as proved by X-ray analysis of the products (Fig. 13). Measurement of the surface area of the calcined products gives 2.16 and 2-19 m2/g for the hydrolysed products in the presence of nitric and perchloric acids, respectively. The lower surface area in this case as compared to that obtained in the dried red cakes may be due to coarsening of the very fine particles by sintering during the calcination process. It is recognised, therefore, that hydrolysis of vanadate liquors produced by leaching of soda ash roasted boiler deposits suffer condensation and polymerisation of vanadic acid of deca or of sesqui forms (H2VIoO28 or H2V6016) n'12 when they are neutralised to pH 2-3. The produced red cakes in this case are mixtures of tri- and tetravalent vanadium oxides. When the hydrolysis process is

A m m o n i u m chloride in a solid/liquid ratio of 5 g NH,C1/250 ml solution was added to the solutions containing either 0.01 M oxidizing agents HC10 4 or H N O 3. It is shown from Fig. 12 that the presence of a m m o n i u m chloride in the solution increases the hydrolysis rate and the optimum recovery of vanadium of about 99.8 % is achieved after 120 min and 90 min hydrolysis with respect to nitric and perchloric acids, respectively. Analysis of the dried red cake reveals that it consists of 92 % V205 equivalent. Calcination of the dried cake at 450C causes a loss in weight of about 8.7 %. Chemical analysis of the calcined product shows that it contains about 99 % V20 ~ equivalent. This means that the precipitated vanadium product

Pure vanadium pentoxide from boiler ash

41

performed in the presence of a m m o n i u m chloride and a suitable oxidant, the lower-valency vanadium ions are oxidized to the pentavalent state. Probably the c o m p o u n d (NH4) 2 V6010 is formed. Such a compound manifests a loss in weight of 8-7 % on thermal decomposition in oxygen or in dry air at 450C, according to the following equation:
(NH4),~ V6016

4. The hydrolysis process of the vanadate liquors comprises two stages; the first one is the nucleation stage with an activation energy of 16"7 kJ/mole and the second is the growth stage with an activation energy of 27"62 k J/ mole.
REFERENCES

,3 V,_,O.~+ 2 N H a + H,,O
1. F. Gong, Y.L., Nonferrous Met. (China), 3 (1) (1989) 49-53. 2. Toda, S., J. Min. Mater. Process Inst. Japan, 105 (3) (1989) 2614. 3. E1-Tawil, S.Z. & Abdel-Salam, M.W., Production of ferrinickel from ash boiler deposites. Egyptian Patent No. 13598, 1978. 4. El-Tawil, S.Z. & Abdel-Salam, M.W., Production of ferrovanadium from boiler deposites. Egyptian Patent No. 12002, 1975. 5. Anon., Thermal treatment of minerals as an aid to metal extraction--A survey. Mineral Processing Information Note No. 4, Warren Spring Laboratory, 1970, pp. 68-9. 6. Gomez-Bueno, C.O., Spink, D.R. & Rempel, G.L., Metallurgical Trans. B., 12B (June) (1981) 341-52. 7. Kirk-Othmer, Encyclopedia of Chemical Technology, 2rid edn, Vol. 21. John Wiley & Sons., London, 1970, p. 160. 8. Clark, R. T. H., The Chemistry of Titanium and Vanadium. Elsevier Publishing Co., London, 1968 p. 203. 9. Kolta, G.A., El-Tawil, S.Z., Abdel-Salam, M.W. & Elzewel, K. A. Chemical composition and upgrading of ash boiler deposits of Egypt. Paper presented at 2nd Metal Conf. in A.R.E., 9-12 Dec., 1978. 10. E1-Tawil, S. Z., Rabah, M., E1-Barawy, K., Morsi, M. B. & Abd-Ghaffar, M. A., Evaluation of vanadium pentoxide produced from boiler deposits. Paper presented at 1st Chem. Conf., Faculty of Science, Mansoura University, 24-26 Sept. 1986. 11. Aldebert, P., Baffler, N., Gharbi, N. & Livage, J., Mat. Res. Bull. 16 (1981) 669 76. 12. Lemerle, J., Nejem, L. & LeFebvre, J., Chem. Res. (M), 8 (1978) 5301.

This value of loss in weight (about 8'7 %) is nearly equal to that obtained after calcination of the dried red cakes at 450C for half an hour. CONCLUSIONS 1. Hydrolysis of sodium vanadate liquors by dilute sulfuric acid produces red cakes containing a maximum V.,Q equivalent of about 100% at pH 2.5 and 90C. It is composed of mixtures of tetra- and trivalent vanadium oxides. Calcination does not produce V205 but gives mixtures of V60 n, V8015 and VTOla. 2. The presence of oxidising agents such as perchloric acid promotes the hydrolysis process as well as an increase in surface area of the produced cake (16.02 m"/g). 3. Addition of a m m o n i u m chloride to the vanadium liquor in the presence of an oxidising agent such as perchloric acid, accelerates the hydrolysis process, and produces highest vanadium recovery (about 97 %). Calcination of the produced red cake gives very high purity V~,O~ (99 %), with a surface area of 2.91 m2/g.

4-2