EFFECT OF pH ON THE REACTION OF SODIUM ...

EFFECT OF pH ON THE REACTION OF SODIUM FLUORIDE WITH HYDROUS OXIDES OF SILICON, ALUMINIUM, AND IRON, AND WITH POORLY ORDERED ALUMINOSILICATES K. W. PERROTT,I B. F. L. SMITH, AND B. D. MITCHELL

(The Macaulay Institute for Soil Research, Craigiebuckler, Aberdeen AB 9 z 03)

Summary The release of hydroxyl ions from silica and ferric oxide gels on treatment with sodium fluoride solution is very small above pH values of 7.6 and 9 respectively. Hydroxyl release from alumina gel and poorly ordered aluminosilicates is appreciable at pH 9 and varies little with pH, but that from crystalline forms of alumina decreases with decrease of pH. For poorly ordered aluminosilicates, the ratio (OH' released at pH 8.0) :(OH' released at pH 6.8) is directly proportional to the mole fraction Al/(Al+Si). The measurement of hydroxyl release at differing pH values may enable determination of hydkous alumina separately from the total poorly ordered inorganic gel material ; moreover, the amount of hydroxyl released at high pH values is related to phosphate sorption capacity.

Introduction INCREASE in pH of a sodium fluoride solution when in contact with soil has been recommended by Fieldes and Perrott (1956) as a suitable field or laboratory test for establishing the resence of allophanic material. Further studies (e.g. Furkert and Fie des, 1968; Bradley and Vimpany, 1969; Brydon and Day, 1970; Bonfils, 1972) have confirmed the value of this test for detecting aluminosilicate gels and hydroxy aluminium species in soils and it has been put on a semi-quantitative basis by Perrott et al. (1976) who used an automatic titrator to measure the amount of hydroxyl ions released into 0 . 8 5 ~sodium fluoride solution in z min at a constant pH of 6-8. The values obtained for a range of cottish soil types correlated well with the amount of alumina removed by 5 per cent sodium carbonate at room temperature. Like Bracewell et al. (I 70), Perrott et al. (1976) also reported a significant release of hydroxyf) ions from silica gel at pH 6.8. The effect of pH on hydroxyl release has now been examined for a range of synthetic aluminosilicate gels and hydrous oxides of silicon, aluminium, and iron. Studies on hydrous aluminium sols and gels yielding solid phases with different 0H:Al ratios have also given information on the relationship between hydroxyl release from Al-OH sites and phosphate sorption capacity.

P

2

Experimental The solids investigated were characterized by several standard instrumental techniques. Thus, X-ray powder diffraction patterns were Methods

Permanent address : Department of Agriculture, Ruakura Soil Research Station, Hamilton, New Zealand. Journal of So11 Science, Vol. 27. 348-66. 1976

EFFECT OF pH ON REACTION OF SODIUM FLUORIDE 349 obtained by a Phili s z kW diffractometer using iron-filtered CoKa

P

radiation, differentia thermal curves were recorded on an instrument similar to that described by Mitchell and Mackenzie (1959), an A.E.I. EM6 electron microscope was used for electron microscopy and electron diffraction, and infra-red absorption spectra were obtained using KBr discs with a Grubb-Parsons S4 spectrometer. Surface areas were measured by the low temperature nitrogen adsorption (77 OK) technique assuming that one adsorbed nitrogen molecule occupies an area of 0.62 nm2. Hydroxyl release in 0 . 8 5 ~NaF at 25°C was measured using a Radiometer Recording Titrigraph at constant pH as described by Perrott et al. (1976). A solid sample :solution ratio of 25 mg :5 ml NaF solution was normally employed, but for hydroxy-aluminium sols the ratio was I ml sample:^ ml solution. Titrations were carried out in a nitrogen atmosphere to prevent interference from carbon dioxide at higher pH values. Materials Hydroxy-aluminium sols and gels were prepared by mixing I M A1(N03), and O-IOSM NaOH solutions, diluting to give a final concentration of 30 mmol/l and 0H: Al ratios covering the range o Samples were aged for 9 months in the mother li uor. X-ray anr 3elec'O. tron diffraction examination revealed that the soli phase resent in the preparation with an 0 H : A l ratio of 3.0 was a mixture o bayerite and pseudoboehmite, that in the preparation with an OH :A1 ratio of 2.5 was microcrystalline gibbsite, and that in the preparation with an OH :A1 ratio of 2.0 was non-crystalline. There was no evidence of any finely particulate material in the preparations with OH :A1 ratios < 2.0 and these may have been true solutions. Two alumina samples (designated alumina-A and alumina-C) were precipitated from approximately 0 . 3 0 ~Al(NO,), solution with 9~ aqueous ammonia. Alumina-A was aged for 4 days in the mother li uor and washed sequentially with water, methanol, and acetone, w ereas alumina-C was not aged, but was rinsed immediately with acetone followed by toluene. Both were air-dried at room temperature and ground to pass Ioo-mesh. Another alumina sample, designated alumina-F, was prepared by hydrolysis of aluminium isopropoxide in distilled water, aged for z weeks in the mother liquor and then freezedried without washing. X-ray diffraction, infra-red absorption spectroscopy and differential thermal analysis showed that alumina-F contained bayerite and seudoboehmite whereas alumina-A and alumina-C were poorly or ered phases that could not be positively identified. The gibbsite sample (G) was a highly crystalline gibbsite from Gu ana ground to pass Ioo-mesh. Jerric oxide was precipitated from approximately 0 . 3 0 ~Fe(NO,), solution with 9~ aqueous ammonia, the final p H being 6.8. After washing sequentially with water, methanol, and acetone, the sample was air-dried at room temperature and ground to pass Ioo-mesh. According to X-ray diffraction the material was non-crystalline.

%

R

B

P

350

K. W. PERROTT, B. F. L. SMITH, B. D. MITCHELL

Silica gel was prepared by hydrolysis of tetraethyl silicate in distilled water with ultrasonic agitation. After ageing for 4 days in the mother liquor with intermittent boiling, the solid was dried at 1 ~ 0 ° Cand round to pass Ioo-mesh. No X-ray diffraction pattern was obtained from this preparation. Aluminosilicate gels of different composition were prepared by h drolysis of mixtures of methyl silicate and aluminium isopropoxide; a ter ageing for I week the suspensions were centrifuged to remove excess alcohol resulting from the hydrolysis and then freeze dried from a water suspension. Relevant characteristics are listed in Table I . Infra-red spectra were consistent with these products being aluminosilicates.

F

TABLE I

Characteristics of synthetic aluminosilicate gels

Al/(Al+ Si) molar ratio

Degree of order

A C

0.30 0.56

D E

0.64 0.87

Amorphous Amorphous Amorphous Amorphous trace of pseudoboehmite

Gel

*

+

Phosphate sorption capacity* pmol/ms 2-21

10.32

6.08 9.28

Sum of two Langmuir maxima (see Rajan and Perrott, 1975).

Results and Discussion The marked effect of pH on the reaction of fluoride with hydrous oxides of silicon, aluminium, and iron is illustrated in Fig. I. For urposes of comparison the OH’-release values for these gels have fee, normalized at Iooommol/Ioog at pH 6.8. The behaviour of silica gel, which does not react at pH values above 7.6, was consistent with the results of Huang and Jackson (1965), who reported a pH of 7-5 after ground quartz had reacted with I M K F for 48 h, and of Brydon and Day (1970), who noted that addition of silicic acid to IM NaF did not alter the H from 7.6. This behaviour is consistent with the observations o Judge (1971) who claimed that free F‘ does not attack protonated silica whereas the protonated HF’ ions do, and that the concentration of protonated fluoride ions increases below pH 7. The relative changes in reactivity of hydrous ferric oxide and alumina gels with increase in pH suggests that it may be possible to determine the reactivity of alumina alone in the presence of silica and ferric oxide gels and crystalline clay minerals. This hypothesis was tested for a series of mixtures of hydrous oxides of silicon, aluminium, and iron with kaolinite by comparing the amount of release of hydroxyl at pH 9 in 0 . 8 5 ~NaF solution with that of pure alumina gel. The results in Table 2 show that the calculated amount of alumina in the mixtures

P

EFFECT OF pH ON REACTION OF SODIUM FLUORIDE

351

SiO,

*0 FIG.

I.

6.0

7.0

8.0 PH O

9.0

10.0

1

Effect of pH on net amount of hydroxyl released from hydrous oxides of silicon, aluminium (preparation F), and iron.

TABLE 2 Determination of alumina, in the presence of silica and iron oxide, from fluoride reactivity Composition of mixtures Fe,Os Kuolinite Also, SiO,

%

%

%

%

0 0

90

5

5

I0

I0

15

I5

15

Alsos determined f r m fluoride reactivity at PH 9'

%

5'6

80

11'0

55

16.5

Calibration curve constructed from mixtures of alumina-C with kaolinite in proportions I , 5 , 10, 15, and 20 per cent of total weight of mixture.


352

was about 10 to 12 per cent higher than the amount present, resumably because of some reactivity from the other components, a though the release of hydroxyl ions was associated mainly with the aluminous com onent. Tge hypothesis that the rate of release of hydroxyl ions from aluminous species is dependent on crystallinity (Bracewell et al., 1970;

P

loo.

-2

h

E E

Y

80. .

.-C

E

% .-C

60,-

W

-mW e

-

I 40 0 c

2

pH6.8 -

pH8.0 pH9.0

20.

O4

1.o 2.0 OH/AI molar ratio

2

3.0

FIG. 2. Influence of crystallinity on the effect of pH on the amount of hydroxyl released from hydrous aluminium oxides. No crystalline phase was observed in preparations with OH :A1 molar ratio < 2.0.

Perrott et al., 1976) was checked using a series of sols, gels, and crystalline modifications of alumina. For the three pH values tested (Fig. 2) the amount of hydroxyl ions released in 25 min is similar for all sols with 0 H : A l ratios less than 2 and aged for 9 months. The value for these, approximately 90 mmol/l, indicates that maximum replacement of OH' by F' has occurred. In contrast, hydroxyl release decreases sharp1 for gibbsitic gels (OH :A1 2.5 and bayerite or pseudoboehmite els ( H :A1 3.0). Moreover, at pH the amount of hydroxyl released from gibbsite (25 mmol/roog, air dried) and alumina-F (22 mmol/ Ioog) is much less than that from non-crystalline species (alumina-A 2300 mmol/Ioog and alumina-C 1400mmol/Ioog). The effect of pH on the reaction of fluoride with hydroxy-aluminium sols and crystalline hydrous alumina species is illustrated in Figs. 3 and 4. The hydroxyl released from crystalline materials decreases with

B

8'

EFFECT O F p H ON REACTION OF SODIUM FLUORIDE

353

increasing pH, whereas that from poorly ordered materials remains constant or even increases. This increase can probably be accounted for by conversion of aquo-groups, located in soluble complexes or on the surfaces of solids, to hydroxy groups (Tomesanyi and Lanyi, 1972). OH/AI

g 0.6

z

1

0.5

8

7.0

8.0

9.0

C

PH

FIG.3. Influence of crystallinity on the effect of pH on the net amount of hydroxyl released from hydroxy-aluminium sols. No crystalline phase was observed in preparations with OH :A1 molar ratio < 2.0.

FIG.4. Influence of crystallinity on the effect of p H on the amount of hydroxyl released from aluminium oxides. A and C s y n t h e t i c alumina gels containing no crystalline phase; F-synthetic alumina containing bayerite andpseudoboehmite; G-gibbsite, Guyana.

The influence of crystallinity on variation of hydroxyl release with pH renders it impossible to assess with any accuracy the contribution of A1-OH sites to the total reactivity at pH 6.8 for highly crystalline oxides in the presence of Si-OH sites. For non-crystalline hydroxy aluminium sols and gels 0 H : A l = 0-5--2-0, Fig. 3; A and C, Fig. 4 variation in amount of hy roxyl released with H is not so marked as or crystalline material (0H:Al = 2-5 and 3.0, &g. 3; F and G, Fig. 4). At about pH 8 the amount of hydroxyl released from soil clays containing large

6

2

K. W. PERRO'M', B. F. L. SMITH, B. D. MITCHELL amounts of poorly ordered inorganic gel material should be almost entirely due to Al-OH. The contribution from iron oxides should be minimal, since these do not react with fluoride at pH 8 to the same extent as alumina. An approximatel linear relationshi was found between the composition of the synt etic gels listed in 'fable I, as expressed by the mole 354

l

FIG.5. Effect of composition of aluminosilicate gels on the ratio of the net hydroxyl release above pH 7.6 to the net hydroxyl release at pH 6-8.

fraction Al/(Al+Si , and the ratio of net OH' release at pH 8 to net OH' release at pH -8(Fi 5 ) . This suggests that it may also be ossible to use the reaction with uoride to assess the roximate mo ar fraction of alumina in X-ray amorphous inorganic material in soils. The phosphate sorption characteristics of synthetic aluminosilicate gels have also been examined (Table I), because there is evidence (Rajan and Perrott, I 75) that ghosphate so tion by noncrystalline aluminosilicates, sucB as allop me, occurs 'gy reaction at hydroxy-aluminium sites. OH' release and phosphate so tion have been expressed in terms of unit area in order to eliminate t e effect of surface area on the relationship (Fig. 6). The results fitted different Langmuir equations at low and medium phosphate concentrations (Rajan and Perrott, 1975) but not at high concentration where disruption of gel structure occurred. The total phosphate sorption capacity,

2 !it

P

'R

EFFECT OF pH O N REACTION OF SODIUM FLUORIDE

355

which is the sum of the two Langmuir sorption maxima, is not linearly related to the amount of hydrox 1 released at pH 6-8, presumably because fluoride reacts with both &-OH and Al-OH sites, but at pH 8 fluoride reaction is directly proportional to phosphate sorption. These results suggest that an a propriate choice of pH for examining the reaction of fluoride with sois should enable simultaneous assess-

P

d

0

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

40 80 120 160 NetOH- release after 25 rnin (pmol OH-/ma)

FIG.6. Relationship between total phosphate sorption of aluminosilicate gels and the net hydroxyl release at pH values of 6.8, 8.0, and 9.0. A, C, D, and E-synthetic aluminosilicate gels with Al/(Al Si) molar fractions 0~30,0~56,0~64,0~87, respectively.

+

ment to be made of phosphate sorption capacity associated with Al-OH sites. However, the relative reactivities of Al-OH and Fe-OH sites with respect to phosphate and fluoride remain to be tested.

Acknowledgements One of the authors (K. W. P.) wishes to thank the National Research Council of New Zealand for the award of a Postgraduate Fellows ip.

AdvisoX

356


REFERENCES P. 1972. Characterization of the Escandorgue andosols and of the Lodevois BONFILS, andic brown soils. Bull. Ass. fr. Stude Sol No. 3, 113-27. BRACEWELL, J. M., CAMPBELL, A. S., and MITCHELL, B. D. 1970. An assessment of some thermal and chemical techniques used in the study of the poorly ordered aluminosilicates in soil clays. Clay Miner. 8, 325-35. BRADLEY, J., and VIMPANY, I. 1969. Allophanic soils in New South Wales. J. Aust. Inst. agric. Sci. 35, 26870. BRYDON, J. E., and DAY,J. H. 1970. Use of the Fieldes and Perrott sodium fluoride test to distinguish the B-horizons of podzols in the field. Can. J. Soil Sci. 50, 35-41. FIELDES,M., and PERROTT,K. W. 1966. Nature of allophane in soils. 111. Rapid field and laboratory test for allophane. N.Z. J. Sci. 9, 623-9. FURKERT, R. J., and FIELDES,M. 1968. Allophane in New Zealand soils. Trans. 9th int. Congr. Soil Sci. Adelaide, 3, 133-41. HUANG,P. M., and JACKSON, M. L. 1965. Mechanism of reaction of neutral fluoride solution with layer silicates and oxides of soils. Proc. Soil Sci. SOC.Am. 29, 661-5. JUDGE, J. S. 1971.A study of the dissolution of SiO, in acidic fluoride solutions. J. electrochem. SOC.118, 1772-5. MITCHELL, B. D., and MACKENZIE, R. C. 1959. An apparatus for differential thermal analysis under controlled atmosphere conditions. Clay Miner. Bull. 4, 3 1-43. PERROTT,K. W., SMITH,B. F. L., and INKSON,R. H. E. 1976. The reaction of fluoride with soils and soil minerals. J. Soil Sci. 27, 58-67. RAJAN,S. S. S., and PERROTT,K. W. 1975. Phosphate adsorption by synthetic amorphous aluminosilicates. Ibid. 26, 257-66. TOMESANYI, L., and LANYI, G. 1972. Acid-base determination of aluminium in bauxite and red mud with potentiometric end-point detection. Anal. chim. Acta 62, 377-84.

(Received 16 December 1975)