62
C. E. MARSHALL A N b W. E. BERQMAN
THE ELECTROCHEMICAL PROPERTIES OF MINERAL MEMBRANES. I1 MEASUREMENT OF POTASSIUM-ION ACTIVITIES IN COLLOIDAL CLAYS'S 2
C. E. MARBHALL'
AND
W. E. BERGMAN'
Department of soils, Uniuerrity of Missouri, Columbia, Missouri Received July 14, lo41 I. INTRODUCTION
The difficulties inherent in the use of amalgam electrodes for the estimation of alkali and alkaline-earth cations have greatly restricted our knowledge of the electrochemistry of many negatively charged colloidal systems. Using the recently described membrane electrode (S), which consists of a film of electrodialyzed bentonite dried at 49OoC., we are now able to make measurements over a range of potassium-gn activities from N/10 to about N/10,000 with reasonable accuracy. The application of this technique to agar and to four clays is here described. The exact significance of activity measurements in colloidal systems has frequently been discussed. Pauli and Matula (11) introduced the concept that potentiometrically determined activities represent the mean activity of the ions in the electrical double layer surrounding the particles and in the intermicellar liquids. Wiegner (13), in discussing the "suspension effect", suggested that it waa even possible for ions in the inner part of the double layer t o affect an electrode. Rabinovitch and Kargin (12) maintained that both of these v i e w were erroneous on the grounds that for any system in equilibrium the activity of all its parts must be equal; hence the intermicellar liquid should have the same activity as the interior of the electrical double layer. However, thermodynamically, it is not possible to envisage a separation of the intermicellar liquid without work being done on or by the system, nor can one expect thermodynamic arguments to throw light on the constitution of the electrical double layer. Rabinovitch and Kargin also drew attention to three practical difficulties which might lead to error: namely, the adsorption of foreign electrolytes, electrode poisoning, and the possibility of insoluble complexes being formed between the ion under investigation and the colloidal surface. It is believed that the determinations reported below are not appreciably affected by these sources of error. 11. EXPERIMENTAL
The apparatus (7) and technique (8) used in obtaining potassium-ion activities have been fully described. I n the present instance a 0.1059 molal potassium 1 Presented at the Eighteenth Colloid Symposium, which was held at Cornel1 University, Ithaca, New York, June 19-21, 1941. 3 Contribution from the Department of Soils, Missouri Agricultural Experiment Station, Journal Series No. 780. a Associate Professor of Soils. a Research Assistant.
POTASSIUM-ION ACTIVITIES IN COLLOIDAL CLAY8
53
chloride solution having an activity of 0.0810 waa used aa the reference solution. The activities reported are therefore calculated from the potentials of the systssl
K+ = 0.081 molal
I
membrane
1
K+
+ H+ in unknown
Hydrogen-ion activities were obtained from measurements with the glass membrane electrode. The potassium-ion activities were then computed using the Nernst equation : RT (0.0810) E.M.F. = - h
F
(K+)
+ (H+)
All potentirtls used represent the average of two or more determinations with different electrodes and are corrected to 25°C. The electrodes were frequently calibrated against known potassium chloride solutions and were rejected for deviations of 1 millivolt or more. Equilibrium was attained almost a t once, and final readings were made several hours after setting up. The colloidal systems were purified by prolonged electrodialysis, using the Bradfield three-compartment glass cell with platinum electrodes, until the anode and cathode chambers showed absence of soluble electrolytes. The hydrogen systems thus obtained were then converted into hydrogen-potassium or potaasium systems by titration with the hydroxide. I t was previously shown (8) that divalent ions do not affect the potwium activity appreciably, and further experiments have shown that aluminum, which might be released in active form from the clays, is also without influence on the measurements.
.
111. POTASSIUM-ION ACTIVITIES IN AGAR SUSPENSIONS
In accordance with the findings of Hoffman and Gortner (6), the temperature of the agar suspension during electrodialysis was kept below 50°C. by the use of low current density in order to prevent autohydrolysis. However, a small amount of sulfate appeared in the anode chamber during the early stages. The electrodialyzed suspension appeared to fit their description. It wm found that the suspension could be heated in the autoclave for a short time to effect solution without appreciable autohydrolysis. The stock suspension thus prepared contained about 1 per cent by weight of the original material and had a pH of 2.89. Aliquots (250 cc.) of this were titrated with 0.1 N potassium hydroxide and diluted to 500 cc. The results obtained are summarized in table 1. In order to determine whether or not the agar salt behaves as a typical strong electrolyte, the measured potassium activities are compared with those calculated on the assumption of complete dissociation, as given in column 4. It is seen that these values agree within the limits of experimental error; hence the potassium salt of agar behaves as a typical strong electrolyte. This is in agreement with the chemical reactions of agar as summarized by Hoffman and Gortner. IV. POTASSIUM-ION ACTIVITIES I N CLAY SUSPENSIONS
According to x-ray data three types of clay have been recognized: ( 1 ) the montmorillonite type, ( 2 ) the mica type, and (3) the kaolinite type. For this
MOLARITY OF
KOH
0
KOH ADDED PERLITER
1
MILLIEQVIVALENTLI
KOHPER~WJQ. CLAY
( K+) (CALCULATED)
3.18 3.50 4.01 4.55 6.14 7.24 8.01 9.01
0.0010 0.0012 0.0015 0.0016 0.0020 0.0025 0. 0030
YOLFB
(K+)
PH
ADDED
~
1 1
pH
(YEAsUBED)
0.00110 0.00126 0.00141 0.00152 0.00173 0.00210 0.00262
I
0. oO0965 0.00116 0.00144 0.00153 0.00190 0.00237 0.00283
1
Ki ACTIVITY
1
FRAT&TvF
11.42 per cent suspensions
0.012 0.024 0.040 0.048
0.003
0.006 0.010 0.012
3.75 4.27 4.58 4.78 4.93 5.03 5.23 5.90 8.10 9.94
8.8 17.5 26.3 35.0 43.8 52.5 61.3 70.1 87.6 105.1
0.01 0.02 0.03 0.04 0.06 0.06 0.07 0.08 0.10 0.12
1
~~
26.3
i;::
105.1
26.3 52.5 87.6 105.1
~
1
0.00180 0.00249 0.00343 0.00462 0.00826 0.00753 0.00870 0.0107 0.0194 0.0216
0.124 0.114 0.115 0.125 0.126 0.124 0.134 0.194
0.180
~
E!:9.90
1I
5.51 6.08 8.34 9.78
Ij
4.98
0.00157 0.00252 0.00769 0.0128
0.131 0.105 0.192 0.266
0. 000442
0.147 0.124 0.254 0.271
0.000743 0.00254 0.00325
K+
55
POTASSIUM-ION ACTIVITIES IW COLLOIDAL CLAYS
“Grundite” was used as the mica type; kaolin ( < 2 p ) , from the McNamee Mine, South Carolina, represented the third group. As has been shown by analysis, colloidal bentonite and kaolin are essentially free of organic matter. The organic matter in the Putnam clay used was removed by treatment with 3 per cent hydrogen peroxide. The illite fraction was not treated, since it contained no appreciable amount of organic matter soluble in sodium hydroxide. TABLE 3 pH and potassium-ion activity data for bentonite sttapensions
3.277 per cent suspensions
0.040 0.050
24.4 30.5 45.8 61.0 76.3 91.6 106.8 122.1 137.3
0.008
40.0
0.010 0.012 0.016 0.020 0.022
50.0 60.0 80.0 100.0 110.0
0.008 0.010 0.015 0.020
0.025 0.030 0.035
3.28 3.55 4.11 4.79 5.40 7.03 9.55 10.49 10.78
0.00375 0.00443 0.00591 0.00678 0.00844 0.0109 0.0125 0.0156
0.375 0.295 0.295 0.281 0.312 0.313 0.312
0,00219 0.00253 0.00272 0.00349 0.W 6 7 0.00679
0.273 0.253 0.226 0.218 0.284 0.308
0.000921 0.00122 0.00146 0.00240 0.00279 0.00333
0.230 0.204
0.271
2.00 Der cent sumensions 4.29 4.68 5.15 5.74 8.45 9.69
0.004
0.006 0.008 0.010 0.011 0.0132
30’5 45.8 61.1 76.4 84.0 100.8
I , ~
~
4.44 5.26 5.81 6.56 7.12 8.68
~
I
0.183 0.240 0.254 0.252
The small amount of colloidal matter < 0.2 A, present in kaolin and illite necessitated the use of the fraction < 2 p in these cases. Stock suspensions of these electrodialyeed clays were prepared, and their colloidal content determined by drying a t 110°C. Since all suspensions prepared were diluted to volume, an error is involved in abandoning the molal basis. This error is small, since the densities of these clays are all greater than 2.2. Further, it was shown by weighing definite volumes of the concentrated suspensions that there could be little error due to,swelling.
66
C. E. MARSHALL AND W. E. BERQMAN
0.005 0.008 0.012 0.016
0.020 0.024 0.028 0.032
0.004 0.008 0.008 0.010 0.012 0.014 0.016
5 8 12 16 20 24 28 32
3.81 4.25 4.85 5.24 5.56 6.16 7.42 8.75
0.00210 0.00254 0.00301 0.00434 0.00559
8 12 16 20 24 28 32
4.88 5.15 5.51 5.93 6.39 7.48 8.93
o.Ooo849 0.000947 0.00109 0.00144 0.00189 0.00262
0.00152
0.304
0.00166
0.208
0.00189
0.158 0.131 0.127 0.125 0.155 0.174
0.196 0.141 0.118 0.120 0.135
TABLE 5 -
p H and potcllrsium-ion activity data for 10.00 per cent kaolin suapensiona
Yobm KOH ADDliD PER UTMB
O.OOO4 0 . W 0.0010 0.0020
0.0030 O.Oo40 0.0050
O.OO60 0.0070
M W E 9 0 1 V A L ~ 100 a. CLAY
PH
KOH m n
0.40 0.70 1.00 2.00 3.00 4.00 5.00 6.00 7.00
1
4.58 5.34 5.67 5.90 6.49 9.18 10.04 10.38 10.59 10.79
K+ *CRYITY
0.0000944 0.000113 o.Ooo183 0.000414 0.000977 0.00167 0.00255 0.00319 0.00432
FRACTION OB ACTIVE
0.236 0.161 0.183 0.207 0.326 0.418 0.511 0.532 0.617
K '
POTASSIUM-ION ACTIVITIES I N COLLOIDAL CLAYS
57
FIG.1 FIG.2 FIG.1. pH and pK curves for suspensions of Putnam clay. The dotted line represent8 the pK curve which would be obtained for the 11.42per cent suspension if the clay “8alt” acted aa a typical strong electrolyte. FIG.2. pH and pK curves for bentonite
FIG.3. pH and pK curves for illite
FIQ.4. pH and pK curves for kaolin
The potassium-ion activity data show that the clay “salts” act as moderately weak electrolytes, since the active fraction remains small, although greater than for hydrogen. In the case of montmorillonite, beidellite, and illite, potassium
58
0
C. E. U R S H A L L AND W. E . BERGMAN
added beyond the point of equivalence also has a low activity, whereas with kaolinite 60 to 100 per cent of the potassium added above p H 9 contributes to the total activity. The changes in the status of the added potassium as neutralization proceeds can best be appreciated by considering the changes in potassium-ion activity for unit increments in base. In all cases the first additions of base up to 15 per cent of equivalence contribute some 18 to 37 per cent of the potassium in the active form. The next additions are relatively inactive, the mean percentages over the pH range from 4 to 6 being as follows: bentonite 21 per cent, Putnam clay 13 per cent, illite 8 per cent, kaolin (5.34.5) 17 per cent. From pH 6 onwards the percentage activity of successive increments of base increases to values of from 20 to 50 per cent. Only in the case of kaolinite, as noted above, does it approach 100 per cent. The relatively high activity of the first additions of potassium may conceivably be ascribed to the solubility of the acid clays and of their salts. Nutting (10) has studied the solubilities of a number of reactive clays in acid, neutral, and alkaline solutions and has concluded that the clays of the montmorillonite group have a definite true solubility. However, the proportions of silica and aluminum found in the ultrafiltrates vary according to the reaction of the equilibrium mixture. In distilled water the solubilities of various montmorillonite clays were found to range from 30 to 100 parts per million. The latter figures could account for 1.5 milliequivalents of potassium per liter for neutralization of soluble clay, which, for montmorillonite, beidellite, and illite, would be an appreciable fraction of the active potassium in the first addition of base. For kaolinite, however, the solubility, if present, must be much smaller, since the first addition of base was only 0.40 milliequivalent, of which only 24 per cent was active. It will be seen from the pK curves for montmorillonite, beidellite, and illite (figures 1 , 2 , and 3) that a change in curvature occurs in each caBe with the most concentrated suspensions and that its position corresponds to 75 to 85 per cent saturation of these clays with potassium. Freundlich, Schmidt, and Lindau (9, working with bentonite suspensions, found that their regions of maximum swelling, maximum sedimentation volume, and maximum thixotropy occur also at about 80 per cent saturation with potassium, although the tihxotropy maximum extends over a considerable range from pH 6 to 8. Baver (l), using Putnam clay, has also shown that maximum viscosity is reached just before complete neutralization with potassium.
B . The effect of dilution From tables 2, 3, and 4 one can deduce the effect of dilution upon the fraction of potassium active a t various degrees of saturation. For bentonite and illite, dilution of the clay system decreases the active fraction throughout, whereas the Putnam clay is variable in this respect. Earlier work on silver-ion activities (9) showed that with bentonite and Putnam clay there existed a minimum fraction dissociated and that both higher and lower concentrations gave greater values. In the present csse the Putnam clay shows an increase in the active
59
POTASSIUM-ION ACTIVITIES IN COLLOIDAL CLAYS
0.012 0.012 0.012 0.012 0.012 0.012
0.006 0.012 0.018 0.030 0.048
0.024
0.012 0.036
0.024
0.096
0.012 0.012 0.020
0.024 0.036
0.020 0.020
0.010
j i 1 i
~
I 1
5.15 4.79 4.67 4.60 4.42 4.34
5.12 5.06 4.59
~
~
i ~
1
~
0.00272 0.00859 0.0152 0.0182 0.0287 0.0397
0.0131 0.0226 0.0303 0.0397 0.0743
~
I I
1
I
0.00587 0.0125 0.0155 0.0260 0.0370
0.00553 0.0107 0.0157 0.0254 0,0393
0.0106 0.0201 0.0372 0.0718
0.0107 0.0204 0.03OO 0.0393 0.0742
0.0195 0.0264
0.0200 0.0295
0.0098
0.0088 0.0169 0.0322
0.0278
Bentonite and potassium phthalate 5.20 5.36 8.45 8.36 8.13 8.12
0.020 0.070
0,020
0.0222 0.0291 0.00567 0.0155 0.0237 0.0352
0.0180 0.0295
Bentonite and potassium acid phthalate 0.012 0.012
j 1
0.012 0.024
1
⋘
1
0.0147 0.0214
1
0.0120 0.0187
0.011 0.020
potassium of a clay and potassium added in true solution. In the last two columns are compared the measured increases in potassium-ion activity due to salt additions and the increases which would be found if the salt were added to pure water. It is seen that these columns are in good agreement. Hence one may conclude ( a ) that potassium montmorillonite and beidellite act as strong electrolytes, since addition of the common ion does not repress the dissociation, and ( b ) that the highly polyvalent negatively charged clay anion has little in-
60
C. E. MARSHALL AND W. E. BERGMAN
fluence on the activity of the added salt. This would seem to bear out Hartley’s (4) theoretical examination of the interionic effects in colloidal system. The experiments are not sufficiently accurate, however, to preclude the presence of small “second-order” effects.
D. Other factors Space does not permit of a detailed account of experiments dealing with the aging of potassium clays. It was found that, whereas considerable changes of pH with time followed neutralization of the hydrogen clays, yet the potassiumion activities showed little change. The activity of exchangeable potassium in the presence of varying amounts of exchangeable calcium was also studied. It was found that calcium increased the relative activity of the potassium only when the clay was over 30 per cent saturated with potassium.
E . Discussion In their electrochemical behavior montmorillonite, beidellite, and illite have shown themselves to be closely similar. Kaolinite is quantitatively different. All four clays show some evidence of passing into true solution in water; but if one uses the activity of the first increment of potassium as a guide, then the solubility of kaolinite is very much less than that of the other clays. Its low exchange capacity and the high activity of the potassium added beyond pH 9 are also distinctive. The general behavior of kaolinite suggests that its capacity for rendering potassium inactive is less than that of the other clays, as would be expected from its relatively compact structure. The close similarity of the micaceous clay illite to the expanding lattice types montmorillonite and beidellite is somewhat surprising. The base-exchange capacity of the sample of illite used is much greater than would be expected from the particle size if the exchangeable cations were all held externally, but its capacity is less than half that of the beidellite and montmorillonite. It seems likely, therefore, that some of the exchange is internal. This is only possible in places where the structure has already been loosened by the interposition of water molecules. Thus some molecular layers would be hydrated and would allow of base exchange, while others retained the micaceous structure. The lattice would therefore be miued,-partly hydrated and partly compact. Hendricks (5) has already suggested that such mixed lattices may be common in the clay group. V. SUMMARY
1. Complete titration curves involving determinations ot hydrogen-ion and potassium-ion activities have been obtained for agar and for four clays: namely, montmorillonite, beidellite, illite, and kaolinite. 2. Potassium agar behaves throughout as a completely dissociated colloidal electrolyte. 3. The clays show varying degrees of dissociation according to the extent of neutralization.
ORIGIN OF UNDERGROUND CARBON DIOXIDE
61
4. Within the accuracy of the experiments, no interionic effects were found between potassium ions associated with the clays and those added in true solution as potassium salts. REFERENCES (1) BAVER,L.D.:Missouri Agr. Expt. Sta. Research Bull. 129 (1929). (2) BRADFIELD, R.: J. Am. Chem. SOC.U,2669 (1923). H.,SCHMIDT, O., AND LINDAU,G.: Kolloid-Beihefte 38, 43 (1932). (3) FREUNDLICH, (4) HARTLEY, G.S.: Trans. Faraday SOC.S1, 31 (1935). (5) HENDRICKS, S.B., AND Ross, C. S.: Z. Krist. mineral. Petrog. lWA, 2.51 (1938). (6) HOFFMAN, W.F., AND GORTNER, R. A,: J. Biol. Chem. 66,371 (192.5). C. E.: J. Phys. Chem. 49,1155 (1939). (7) MARSHALL, (8) MARSHALL, C. E., AND BERGMAN, W. E.: J . Am. Chem. 900.89, 1911 (1941). (9) MARSEALL, C. E., AND GUPTA,R. S.: J. Soo. Chem. Ind. 62,433T (1933). (10) NUT TIN^, P. G.: J. Franklin Inst. 224, 339 (1937). (11) PAULI,W.,AND MATULA, J.: Kolloid-Z. 31, 49 (1917). A. F., AND KARGIN, V. A.: Trans. Faraday SOC. S1,M (1936). (12) RABINOVITCH, (13) WIEGNER, G.:Kolloid-Z. 61, 49 (1930).
THE ORIGIN OF UNDERGROUND CARBON DIOXIDE' FRANK E. E . GERMANN
AND
HERBERT W. AYRES
Department of Chemistry, University of Colorado, Boulder, Colorado Received August 18, 1941 I. INTRODUCTION
Many theories of the origin of subterranean carbon dioxide have been advanced (4), and it is a reasonable assumption that there is truth in a number of them. It is the purpose of the present paper to study only one of these, Le., thermal dissociation, and to show that actual underground conditions may lead to quite different results from those obtained under idealized conditions realized in a modern laboratory. The first careful study of the equilibrium established between calcium carbonate, calcium oxide, and carbon dioxide was reported by Debray (2), whose results are incorrect principally because of incorrect values assigned to the temperatures of boiling cadmium and zinc. Since that time numerous studies of the thermal dissociation of pure calcium carbonate have been reported and the dissociation pressures over a wide range of temperatures are accurately known. Carbon dioxide has been encountered in various deep oil wells in the United States and Mexico. Volumes estimated as high as fifty million cubic feet per day have been reported from wells the measured closed-in pressures of which Presented a t the Eighteenth Colloid Symposium, which waa held a t Cornel1 University, Ithaoa, New York, June 1!3-21,1941.