J. Phys. Chem. 1980, 8 4 , 507-510
507
Crystal Modification of Freshly Precipitated Aluminum Hydroxide by Lithium Ion Intercalation Moshe Frenkel,” Abraham Glasner, and Sara Sarig Casali Instltute of Applied Chemistry, School of Applied Science and Technology, and Department of Inorganlc and Analytical Chemistry, Hebrew University of Jerusalem, Israel (Received October 17, 1979) Publication costs assisted by Casali Instltute
Aluminum hydroxide was precipitated from solutions containing lithium ions in various AkLi ratios. The chemical compositions, X-ray diffraction patterns, and DTA curves of the various precipitates were determined. While bayerite and pseudoboehmite are precipitated in the absence of foreign ions and in the presence of sodium ions, respectively, lithium ions cause a new crystal modification of the precipitate.
Introduction The specific surface area, surface properties, and the adsorption potential of precipitated alumina were extensively The surface properties of aluminum hydroxide depend on the mode of preparation: pH, initial concentration, procedure of mixing the reagents, presence of foreign ions, drying, aging, etc. All these factors can also be correlated with the crystallographic structure of aluminum hydroxide: gibbsite, bayerite, boehmite, pseudoboehmite, or a11 amorphous solid p h a ~ e . ~ ! ~ The aforementioned ~ t u d i e s l -were ~ directed toward practical purposes and, though using diverse and sophisticated research techniques, were not involved in resolving the questions about the composition of the precursors of the precipitate, i.e., the hydrolyzed aluminum species. The presence of polymeric species was inferred from numerous important theoretical studies5-10 and was even demonstrated by extraction.ll On the other hand, there was also an attempt to explain the anomaly of the aluminum ion titration curve by the penetration of the anion into the precipitate.12 The understanding of the processes which occur during the titration of an aluminum salt by sodium hydroxide was improved by recently developed new te~hniques.’~J~ Both Stol14 and Hsu7 postulated a hydrolysis scheme for aluminum. Another valuable contribution of Hsu to the understanding of the mechanism of aluminum hydroxide formation is his fundamental study of the “Effect of salts on the formation of bayerite versus p s e ~ d o b o e h m i t e ” .In ~~ general, at low sodium salt concentration and with OH/A1 ratios of 3-3.3 pure crystalline Al(OH)3 in the form of bayerite was fiormed, but, in solution with a high salt concentration, pseudoboehmite was obtained. Pseudoboehmite is an incompletely dehydrated boehmite [AlO(OH)] and is usually identified in X-ray diffractograms by a broad peak which corresponds to an interplanar distance of 6.6 A. Hsu graded the investigated cations and anions according to their effectiveness in producing pseudoboehmite. For alkali metals this effect decreases with increasing cationic radius, namely, the order of effectiveness is Na+ > K + > Cs+ for cations and C1- > C10, > NO, > Br- > I- for anions. It is interesting to note that the “directing” ions are not incorporated to any significant amount in the precipitated aluminurn h y d r 0 ~ i d e . I ~ In the present study the effect of lithium ion on the precipitation of aluminum hydroxide was investigated, thus completing the series which was investigated by Hsu. In 0022-3654/80/2084-0507$0 1.OO/O
addition, the present study may elucidate some obscure points concerning the process of lithium extraction by aluminum hydroxide from brines. It is empirically known that Li can be selectively coprecipitated with aluminum hydroxide even under conditions (no excess over OH/Al= 3 ratio, low pH) when no lithium aluminate can be formed. The existing patents for Li extraction, according to this process, took advantage of the possibility of Li coprecipitation with Al(OH)3from concentrated industrial brines.16J7 The structure of the precipitate could not be determined because the results of chemical and X-ray analyses were obscured by the accompanying salts coprecipitated from the brines. The selective coprecipitation of Li ions was vaguely explained as adsorption on alumina.16J7 It is significant that no other common adsorbant except alumina was found to be suitable for a selective Li extraction. Thus, while other alkali metal ions affect the formation of pseudoboehmite vs. bayerite without incorporation into the hydroxide structure, lithium is selectively coprecipitated by intercalation with the aluminum hydroxide formed in solution.
Experimental Section Aqueous solutions of 0.022 M A13+(either chlorides or sulfates) in presence of various amounts of Na+ and Li+ were titrated, under stirring, with aqueous 1 N NaOH solutions and the change in pH was recorded continuously. The hydroxide was added within 10 min and the solution was stirred for another 10 min. The solids were centrifuged and dried at 110 “C. Chemical analysis of the precipitates was carried out by dissolving them in HN03, diluting the solution, and analyzing the cations with an atomic absorption spectrophotometer (Perkin-Elmer, 403), the chlorides by argentometry, and the sulfates gravimetrically (as BaS04). The amount of hydroxides in the precipitate was calculated as the difference between total equivalent of cations and that of anions analyzed. X-ray powder diffraction patterns of the precipitate were carried out on a Phillips X-ray instrument. Differential thermal analysis was carried out on a Mettler thermoanalyzer in platinum crucible, under a N2 flow and heating rate of 10 OC/min. Results and Discussion The effect of lithium ions on aluminum hydroxide precipitation was established in a series of experiments in which aqueous A1Cl3 solutions were neutralized by NaOH 0 1980 American Chemical Society
508
The Journal of Physical Chemistry, Vol. 84, No, 5, 1980
Frenkel, Glasner, and Sarig
TABLE I : Composition of Aluminum Hydroxide Precipitated in the Presence of Lithium Chloride
-
expt no.
solution Al: Li molar ratio
111 114 113 112 110
8:l 4:l 2:l 1:1
no. of moles in 100 g of precipitate Li
A1
Na
1.21 1.04 0.96 0.93 0.74
0.08 0.09 0.10 0.14 0.11
molar ratios in precipitate
c1
0Ha
Al:Li
0H:Cl
0.19 3.50 18.0 0.26 3.02 13.3 11.5 0.39 2.76 6.1 7.0 0.57 2.60 3.6 4.5 0.72 1.95 2.2 2.7 a Calculated. Analysis of Al(OH), precipitated in the presence of Li’, separated from solutions 1 0 min after the comple. tion of neutralization. 0.08 0.16 0.25 0.34
TABLE 11: Composition of Aluminum Hydroxide Precipitated in the Presence of Lithium and Sodium Salts expl no.
solution anion
solution Al:Li:Na
no. of moles in 100 g of precipitate
Li c1 1:O:l 116 c1 ’ 1:1:1 145 0.33 c1 1:1:4 146 0.29 1:O:l 144 so4 1:1:1 0.26 147 so4 1:1:4 0.24 148 so4 a Calculated. Effects of sodium and sulfate ions o n after the completion of neutralization.
A1
Na
c1
1.26 0.80 0.65 1.07 0.68 0.59
0.11 0.22 0.37 0.22 0.24 0.35
0.24 0.65 0.79
so4
OHa
0.24 0.38 0.43
3.64 2.30 1.83 3.94 1.79 1.52
A1 :Li 2.42 2.26 2.63 2.47
A1:Li ratio in Al(OH), precipitate. Separation from solutions 10 min
in the presence of varying amounts of Li+. In Table I the molar ratios of Al/Li in various precipitates are listed. The solid phase was separated from the mother liquor 10 min after the completion of neutralization, namely after the addition of the base in OH/Al = 3 molar ratio. The molar ratio of A / L i in the precipitate depends upon the initial molar ratio in solution. It reaches a value of 2.2 for Al/Li in the precipitate when the initial molar concentration ratio in solution was 1:l. The amount of chloride present in the precipitate is greatly affected by the amount of lithium in the precipitate. While the OH/C1 molar ratio in the precipitate which does not contain Li+ (Table I, expt 111) is 18.0, it is only 2.7 when Li+ was present in the solution in an Al/Li molar ratio of 1 (Table I, expt 110). The change in the composition may indicate a possible modification in the structure of the precipitate. Some irregularity in the sodium content may be noticed. The reason for this is that an appreciable amount of the sodium and chloride ions precipitate with the alumina as NaC1. Peaks for NaCl appear in the X-ray diffraction pattern but, significantly, the characteristic diffraction pattern of LiCl in the presence of lithium ions is absolutely absent. The solution in which precipitate E was formed (Figure 1, expt 110) contained Li/A1 in a 1:l molar ratio from the beginning and Na/A1 in a 3:l molar ratio at the termination of the titration due to the addition of NaOH during the titration, but the Li content in the precipitate is 0.34 mo1/100 g while the Na content there is only 0.105 mol/100 g. In order to ascertain that Li is selectively incorporated into the Al(OH)3precipitate, aluminum salts, both chloride and sulfate, were titrated while NaCl or Na2S04 were present from the beginning in solution in molar ratios Na/A1 of 1:l and 4:l; the Li/Al ratio was 1:l in all cases (Table 11). The Al/Li ratio in the precipitate ranges from 2.26 to 2.63 with a very slight increase for the lower NaCl concentration and slight increase in sulfate salts as compared to chlorides. The X-ray diffraction patterns of the precipitates listed in Table I are shown in Figure 1. Curve A (no Li+ in solution) shows the characteristic diffraction pattern of bayerite (4.72 A; 4.35 A) and only slight traces of the pseudoboehmite (6.6 A) may be distinguished. The A1(OH)3obtained in the absence of Li+ is evidently a mixture of bayerite and very slight traces of pseudo-
5
I0
20
I5
25
30
28
Flgure 1. X-ray diffraction patterns (Cu Ka radiation) of the precipitates listed in Table I: (A) expt 111; (B) expt 114; (C) expt 113; (D) expt 112; (E) expt 110; (F) expt 116 (Table 11).
boehmite. Curve B (expt 114, Al/Li = 8) shows the beginning of the formation of a new peak at d = 7.82 A. As the Al/Li ratio decreases in experiments 113,112, and 110, curves C, D, and E, respectively, the 7.8-A peak increases correspondingly and the bayerite structure disappears as evidenced by the fading of the d = 4.72 A and d = 4.32 A peaks. For comparison, curve F represents the diffraction of Al(OH), precipitated in the presence of Na+ with an initial Al/Na 1:l molar ratio in the solution. The presence of sodium in the solution favored the formation of pseudoboehmite as evidenced by this curve and is in good agreement with the observations of Hsu.16 Note the significant difference between the shape of the first peak of the pure pseudoboehmite (curve F) and that of the “Li containing Al(OH),” of curve E, as well as the difference of about 1.2 A in the d values. The presented results show unequivocally that the preferential precipitation of lithium with aluminum hydroxide involves some modifications of the pseudoboehmite structure. The particular change is manifested by the formation of a d = 7.8 A peak instead of the d 6.65 A peak.
The Journal of Physical Chemistry, Vol. 84, No.
Crystal Modification of AI(OH), with Li+
TABLE 111: Dehydration Temperatures of Aluminum Hydroxide Polymorphsa temp temp of of first second decomp decomp crystalline step, step, modification “C “C ref
9
pseudoboehmite pseudoboehmite
PH 8
140 180
460 435
5 10 T (mid
bayerite
130
300
new modification
150
345
a
Dehydration temperatures of alumina
Flgure 2. pH decrease of the solutions after the termination of the titration (0H:AI molar ratio = 3). Experiments as listed in Figure 1.
The “directing” potential of the alkali ions, as well as that of several anions, is ascribed by Hsu to their ability, when present in sufficient concentrations, to cause dehydration of some of the A1(OH)2+to A10+, which is immediately followed by p01ymerization.l~ The assumption about the dehydration of the A1(OH)2f species in solution where free water is scarce is strongly enhanced by a recent study on the injluence of the activity of water on the phase composition of aluminum hydroxides.18 Alumina precipitation was carried out in solutions of water and dioxane of varying compositions and it was found that the aging of the primarily precipitated Al(OH)3 to bayerite or pseudoboehmiite was controlled by the water content of the system. Lithium ions may be expected to promote dehydration better than other alkali ions due to their high hydration energy. The X-rays diffractograms show that the capacity of Li+ does no1 manifest itself only in the degree of effectiveness, but also in creating an additional unique effect of preferential coprecipitation and by inducing the formation of a new distinct crystallographic polymorph of A1(OH)3. Data for crystal structure of boehmitelg suggest intercalation of Li+ ions between the hexagonal aluminum layers as the most probable cause for the increased interplanar distance. This assumption is strongly enhanced by another unique phenomenon: After the aluminum salt, in the presence of lithium, has been neutralized by base (OH/Al = 3) a pH sharp drop is observed. The extent of pH decrease is proportional to Li concentration and it does not occur if the titration is (discontinuedbefore the molar ratio OH/A1 of 3 is reached. The pH drop and its dependence on the lithium content are shown in Figure 2. Before the ratio of 3:l for OH/A1 is reached, turbidity forms in the solution and a slight precipitate may be separated by use of a high-speed centrifuge. No lithium is recovered in the precipitate a t this stage. It may be inferred that lithium coprecipitates at the stage when aluminum hydroxide with the formula Al(OH), is formed. This incorporation is accompanied by the release of protons which decrease the pH in the solution. The rnechairiism of lithium incorporation in the “mature” Al(OH)3may also be demonstrated by an efficient recovery of lithium when freshly prepared aluminum hydroxide is added to a lithium containing solution. The degree of cryst,allinity of the new polymorph may be judged, besides by the sharpness of the peaks in X-ray
-
18 present study, expt 1 1 6 “Lange’s Handbook of Chemistry” present study, expt 111 present study, expt 110
360
boehmite
-7
5, 1980 509
/ /
Exo
t Endo
F
I ’ 100
500
T
C0Cl
Figure 3. DTA curve of precipitates, A, Figure 1.
50 0
700
E, and F designations as in
diffractograms, by the temperature of dehydration and the shape of DTA curves (Table I11 and Figure 3). The dehydration of all species of aluminum hydroxide is a distinct two-step reaction. First the adsorbed water molecules are evolved in the temperature range 130-180 “C. In Table 111the temperatures of the second decomposition step of several crystalline forms of aluminum hydroxide are listed. They range from 300 to 460 “C and involve a dehydroxylation step. The data for the temperature of pseudoboehmite dehydration obtained in the present study and in recent researchls are in good agreement. However, the dehydroxylation temperature of “Li containing Al(OH),” differs significantly from that of pseudoboehmite. This would classify the former as a new crystal modification which is in excellent agreement with the X-ray diffraction pattern. The shapes of DTA peaks of the various dehydroxylation experiments (Figure 3) enhance the conclusion that the “Li containing Al(OH),” has a degree of crystallinity lower than bayerite but much higher than pseudoboehmite. The DTA peak of bayerite is sharp and narrow, indicating a well-defined reaction as would be expected for the decomposition of a well-crystallized compound. The pseudoboehmite has very wide peak which is anticipated for
510
J. Phys. Chem. 1980, 84, 510-512
the decomposition of a nearly amorphous compound. The chemical composition, X-ray diffraction patterns, the dehydroxylation temperature of the “lithium containing A1(OH)3” precipitates as well as the p H decrease in the mother liquor, and the shape of the DTA curves indicate that lithium ions cause a crystallographic modification in precipitating pseudoboehmite. Such a definite effect may be well explained by intercalation of lithium ions. The is both and by taking into account the structural features of boehmite and pseud~boehmite’~ and the remarkable difficulty encountered in Li+ extraction from A1(0H)3 precipitates‘ The exact way of lithium incorporation should be further investigated. Acknowledgment. The authors are indebted to Mr. Aryeh Raz of the Industrial Research Administration of the Ministry of Commerce and Industry, and to Dr. J. A. Epstein of the Dead Sea Works for their initiative and interest. The financial support of the Ministry and of Israeli Chemicals Ltd. is appreciated.
References a n d Notes (1) M. E. Harris and K. S. W. Sing, J . Appl. Chem., 223 (1955). (2) J. W. Lucas, G. W. Newton, and K. S. W. Sing, J. Appl. Chem., 265 (1963). (3) E. Calvet, P. Boivinet, M. Noel, H. Thlbon, A. Malllard, and R. Tertian Bull. SOC. Chim. Fr., 99 (1953). (4) D. Pap&, R. Tertian, and R. Bais, Bull. Soc. Chim. Fr., 1301 (1958). (5) E. Matilevie, K. G. Mathai, R. H. Ottewill, and M. KerkeF, J . Phys. Chem., 65,826 (1961). (6) S. S. Singh, Can. J. Chem., 47,663 (1969). (7) P. H. Hsu and R. F. Bates, Miner. Mag., 33, 749 (1964). (8) J. Aveston, J . Chem. Soc., 4438 (1965). (9) C. Brosset, G. Blederman, and L. 0. Sillen, Acta Chim. Scad., 1917 (1954). (10) J. J. Fripiat, F. Van Cauwelaert, and H. Bosmans, J . fhys. Chem., 69, 2458 (1965). (11) T. Okura, K. Goto, and T. Yotuajanagi, Anal. Chem., 34,581(1962). (12) C. R. Frink and B. L. Sawhney, Soil Sci., 103, 144 (1967). (13) A. C. Vermeulen, J. W. Geus, R. J. Stol, and P. L. de Bruyn, J. ColW Interface Sci., 51, 449 (1975). (14) R. J. Stol, A. K. van Helden, and P. L. de Bruyn, J . ColloidInferface Sci., 57, 115 (1976). , sei., 103, 101 (1967). (15) p. H. H ~ U soil (16) R. D. Goodenough, U S . Patent 2964381 (1960). (17) N. P. Neipert and C. L. Bon, U.S. Patent 3306700 (1967). (18) 0.Lahodny-Sarc, L. Dragcevic, and D. Dosen-Sver, Ckys C/ayMiw., 26 153 (1978). (19) R. W. G. Wyckoff, “Crystal Structures”, Interscience,New York, 1964.
Surface Effects of Anisotropic London Dispersion Forces in n-Alkanest Frederlck M. Fowkes Depattment of Chemistry, Lehigh University, Bethlehem, Pennsylvania 18015 (Received Ju/y 30, 1979) Publication costs assisted by Lehigh University
Light-scattering and heats of mixing measurements by Bothorel, Patterson, and Tancrede have shown that in the higher n-alkanes adjacent chains are oriented parallel to one another, giving rise to an appreciable enhancement in adhesion; the effect has been termed correlated molecular orientation (CMO). Surface and interfacial tensions of alkanes vs. water show a specific CMO interaction between adjacent molecules of the higher n-alkanes which results in an anisotropic dispersion force component of the surface energy (ya) which is in excess of the normal isotropic dispersion force component of surface energy (yd). These findings suggest that the CMO contribution to cohesive energy and surface tension arises from an enhancement in the London dispersion forces between pairs of parallel molecules because of the much greater polarizability parallel to the chains. This interaction gives rise to local regions of optical anisotropy, to enhanced cohesive energy, and to enhanced surface tension, but does not contribute to the isotropic London dispersion farce field which governs the intermolecular interaction of alkanes with water. At 25 “C the CMO contribution to the surface energy of n-alkanes increases from zero for n-hexane to 2.9 mJ/m2 for n-hexadecane, but it is negligibly small for cyclic and highly branched alkanes. The dispersion force component of the surface energy of water is found to be 22.0 mJ/m2 at 25 “C, independent of what alkane is used as reference liquid.
Correlated Molecular Orientation In 1967 and 1968 Bothorel and co-workers published light-scattering studies1*2of n-alkanes which showed optically anisotropic regions attributed to parallel orientation of parts of adjacent molecules. The “correlated molecular orientation” (CMO) effect was strong in n-hexadecane and n-dodecane and negligible in n-hexane and in highly branched n-alkanes such as isohexadecane (2,2,4,4,6,8,8heptamethylnonane). In the following decade Patterson, Delmas, Tancrede, and others studied the calorimetry of mixing the higher Presented at the Centennial Meeting of the American Chemical Society (New York, April, 1976) in the Kendall Award Symposium honoring R. J. Good. 0022-3654/80/2084-0510$0 1 .OO/O
n-alkanes with lower n-alkanes or with highly branched or cyclic alkane^.^-^ It was found that a positive term in
the heats of mixing always occurred when diluting the higher n-alkanes with branched or cyclic alkanes and this term was attributed to an enhanced interaction energy of the higher n-alkanes resulting from the parallel orientation of adjacent chains indicated by the light-scattering studies. This CMO contribution to the cohesive energy was observed at 25 “C only with n-alkanes above n-hexane and it increased with chain length. It was absent in highly branched or cyclic alkanes. The magnitude of this effect was determined by heats of mixing of many pairs of alkanes and over a wide range of compositions. An example is the heat of mixing of n-hexadecane with cy~lohexane,~ in which the CMO contribution to the partial molal heat 0 1980 American Chemical Society