Oct., 1959
COLLECTOR-DEPRESSANT EQUILIBRIA IN FLOTATION
( 5 ) The spread in the histograms has been interpreted as being due to two effects: (a) errors in the hydrogen bond length determinations; (b) the effect on the packing of intermolecular forces other than those due to the particular hydrogen bond in question. Therefore it might be expected that the spread would be greater for larger than for smaller molecules. Tables I-XXI support this view. Most “unexpected” lengths have been observed in the structures of amino acids, di- and tripeptides, probably because the structures of other types of molecule of comparable size have seldom been determined accurately. ( 6 ) Examples have been given indicating that bifurcated bonds can exist though infrequently. The need for more data with a limit of p r o r in the hydrogen bond length of less than 0.05 A. is appar-
1717
ent from Table XXII and Fig. 1. Such data will help to separate the spread in the histograms due to errors in the structure determinations from that due to variations in length of a particular type of hydrogen bond in different structures. It will then be possible to give more reliable values of d (see diagram 1 for definition) and its standard deviation and to draw more detailed conclusions than those given. I would like to thank Dr. M. H. F. Wilkins, Mr. G. R. Wilkinson and Dr. M. Spencer for much helpful discussion and suggestions during this work. Also, I am very grateful to Professor J. Donohue who read the manuscript and made a number of valuable criticisms. During the period of this work I have been the holder of a Medical Research Council Scholarship.
COLLECTOR-DEPRESSANT EQUILIBRIA I N FLOTATION BYK. L. SUTHERLAND Division of Physical Chemistry, Commonwealth Scientific and Industrial Research Organization, Box 4331, G.P.O., Melbourne Received February 7,1969
It is shown that the competition in flotation between collector ions and depressant for a mineral surface can be expressed in the same form as an expression developed for collector-depressant competition in the neutral molecule theory. Accurate measurements over a wide pH range are made using a stable sulfur-containing collector (mercaptobensthiazole) for pyrite. Both theories fail to describe the relation between the pH value and the amount of collector required to cause flotation, but accurately describe the relation between pH value and cyanide addition a t constant collector addition.
I n the flotation process minerals are differentially floated by using depressants to prevent the adsorption of a collector on certain minerals. This collector, a polar-noli-polar substance, is adsorbed on the wanted mineral making its surface hydrophobic. The collectors usually contain an ionic group such as CibH31C00-, C2HbO.CS.S-, CizH26SO4-, cisH33(CH3)3N+ or C12H26NH3+ and depressants usually contain inorganic ions such as OH-, CN-, SH-, H f or Ca++. Depressants have the same ionic sign as the collector and compete with it for the mineral surface. Wark and Cox’ and others have considered the competition between collector ion, X-, and depressant ion, D-, such that
by Wark and Cox who mostly considered solutions where such a correction is unnecessary. In examining the mechanism of ion competition, Last and Cook suggest that the equilibria involved are
+ x+ + +
S” r’S”XS” OHSffOHS f f D- JJ S”DS’ H+ S‘H+
(1) (2) (3) (4)
where S“ is an electropositive site on the mineral and 8’ an electronegative site, the surface remaining electrically neutral after adsorption. These four relations have only three unknowns, since the concentration of H + determines that of the OH-. They then go on to show that the equations are mathematically inconsistent in this form, but this does not constitute proof of their assertion that the experimental facts are incompatible with an ionic mechanism. The hypothesis of neutral molecule adsorption is examined by Last and Cook who find the relation arising from the equilibria
[x-l = constant [D-1 This relation appears to hold approximately for a number of simple systems, e.g., Sutherland and Wark.2 Last and Cook3 have proposed an alternative mechanism in which it is postulated that the active collector is not the ion but the free acid or free base, e.g., C I ~ H ~ ~ C O O C2H,OCSSH, H, C12H26S04H1 CleH83(CH~)3NOHor C12H2,NH2; and that the S+HXSSHX (5) active depressant is HCN, H,S or presumably, S+HDZSHD (6) Ca(0H)Z. They also show that, for many colas lectors related to the xanthates or dithiophosphates, account should be taken of the extent of K”Kz [HX] - [HX]o = - [HDI (7) ionization of these collectors-a factor neglected K1 (1) I. W. Wark and A. B. Cox, Trans. Am. Inst. Minino Met. Engrn., where [HX] is the concentration of collector acid i i a , 267 (1934); i w 7 (1939). required to induce contact (flotation) at the con(2) K. L. Sutherland and I. W. Wark, “Principles of Flotation.” centration [HD] of depressant. [HXjo is the conAuetralasian Inst. Mining Met., 1955, Chapter IX. centration required in the absence of HD and K1 (3) A. W. Last and M. A . Cook, THISJOURNAL, 6 6 , 637 (1952).
K. L. SUTHERLAND
1718
and KZ are equilibrium constants for equations 5 an 6. The constant K" depends on the amount collector adsorbed and may be directly related to the contact angle (Wadsworth, Conrady and Cook4). The agreement of the experimental data to the form of the relation 7 is excellent although it is necessary t o adjust the dissociation constants of some collectors by factors of up to twenty-five.6 The Ionic Mechanism.-Mineral for flotation is ground under water and then collector or depressant is added. The following simultaneous ionic equilibria will be established S",, S"., S".,
+ Pa,+ OH- + H + SffOH- + S'H+ (8) + S'*, + D - + H + + Na+ I _ S"D- + S'H+ + Na+ (9) + Sfaq + X - + H + + Na+ S"X- + S'H+ + Na+ (10)
That is, there are three simultaneous competitions for the S" sites and the equations automatically satisfy the condition that whenever an ion of one sign is taken up electroneutrality is ensured. By subtraction of each equation from another the reS"D- + OH- 1_S"OH- + D S"XS"D-
1
+ OH- J _ S"0H- + x+ X - J_ S"X- + D -
(11)
lations are obtained which give the exchange between added ion and occupied sites. The equilibrium constants for equations 8, 9 and 10 are given by equations 12, 13 and 14 assuming the activity coefficients in solution are unity and that the activity of the solid surface is given by the mole fraction (0) of occupied sites, all sites having equal (but not necessarily unit) activity coefficients.
KH.X= (1
eaex
- h)(1- ex -
- e o ~ ) [ x - ] [ H + ] (14)
In a solution free from D- define a concentration [HXIoor [X-10 a t which flotation is just possible. Introducing this and rearranging the equations the following relation is obtained (4) M. E. Wadsworth, R. G. Conrady and M. A. Cook, THIS
JOURNAL, 55, 1219 (1951).
Dissociation constants Dissociation constant Dissociation constant by for best fit to Acid flotation data measurement Ethylxanthic 3 x 10-0 (180) 5 x 10" ( 3 9 6 n-Amylxanthic 1 X lo-' (18") 2 . 5 X lo-' (3O)E Diethylphosphoric 2 . 3 X 10-5 No measurements Diethylcarhamio 1 . 6 X 10-7 No measurements Has 5 . 7 X 10-8' 9 . 1 X 10" (18O)S HCN 7 . 2 X 10-10 4 . 7 9 X 10-10 (18O)9 (6) M . A . Cook and J . C. Nixon, ibid., 64, 455 (1950). (7) N. A. Lange, "Handbook of Chemistry," 6th Edition, Handbook Publishers, Inc., Sandusky, Ohio, 1946. ( 8 ) I. M.Kolthoff, THISJOURNAL, 85, 2715 (1931). (9) H. T.1. Britton and R. A . Robinson, Trans. Faraday SOC.,as, 536 (1932). (5)
4dl
Vol. 63
+ #)[KH,DKD[HD]+ K~.oaKwl1'/2 (15) [ I f 4 d 1 + #)KH,OHKW]*/~)
where C$ = 1/0x - 1 and [HX], is the concentration of collector required to give a collector cover of Ox in the absence of D- and K D and Kx are the equilibrium ionization constants of HD and HX, respectively. It is now required to find the condition in which relation 15 describes data which Last and Cook find to be "reasonably" fitted by equation 7. This requires that the last term in equation 15 shall be negligible or that all the terms reduce to the form of equation 7. Equation 15 is not tractable to simplification but from equations 12 and 14
and 13 and 14 form (16) and (17) and now assume that nearly all positive sites are occupied by anions. The condition for flotation is that shall exceed a critical constant ratio.1° From (16), (17) and (18) For [D-] = 0 then (20)
and by substituting for X in (19) which is precisely the form of relation given in equation 7. The single asszcmption of nearly all sites being occupied i s si@cient to enable one to describe the data. It would appear therefore that a crucial test between the two theories would be to examine a system where much less than all sites were occupied and in which the total number varied with the composition of the solution. Wadsworth, Conrady and Cook4 have also examined the free energy of adsorption. Considering only the equilibria given in equations 1 and 2, they find the energies of adsorption of xanthate and hydroxyl on galena to be improbable. However the method of separating the thermodynamic quantities is quite arbitrary. From their equation 33 and figure 4 the adsorption potential for the reaction (22) is - RT In 2.67. They now determine SX- + OHSOH- + X (22) theadsorption potential for an experiment conducted by Taggart and Hassialis" in which a clean, polished M specimen of galena is immersed in 5 X solution of potapsium ethyl xanthate. The solution is about 90% depleted and the surface coverage is about 16%. Since the solution would have a
+ +
-
ex) (10) In particular if Ox OD BOH = 1 then 1 / X = Ox/(l and ex = 1/(1 X). (11) A. F. Taggart and M. D. Hassialis, Trans. Am. Inst. Minino Met. Engrs., 169, 259 (1946).
+
COLLECTOR-DEPRESSANT EQUILIBRIA IN FLOTATION
Oct., 1959
1719
pH value of about 6, then the adsorption potential for equation 22 is given by which is in good agreement with the preceding calculation and not, as they state, in contradiction. The free energies of exchange between adsorbed depressants and collectors are from about one to several kilocalories per mole. These seem entirely reasonable as judged, for example, from free energies of exchange on clays. Thus, Merriam and Thomas,12 who studied the exchange of Li+, Na+, K+, Cs+ on attapulgite, find values ranging from 0.2 to 2 kcal. per mole. Mercaptobenzthiazole as Collector.-One of the difficulties presented by the data of Wark and Cox lies in the instability of the free acids produced from their collectors at low pH values. Sutherland and Wark13 have shown that galena, pyrite, chalcopyrite and copper-activated sphalerite respond to mercaptobenzthiazole in the same way as the other thio-collectors such as the xanthates which are unstable. The dissociation constant of mercaptoa t 35" and benzthiazole was found to be 3.2 X the response of pyrite over a wide pH range is shown in Fig. 1. The broken curves are lines of constant free acid concentration or of constant ratio of collector ion to hydroxyl ion based on ionization of the SH group alone. Clearly neither the theory of constant free acid nor the ionic theory, when subjected to the restriction of nearly all sites occupied, is able to describe the results. Nor is it possible to obtain agreement between theory and experiment by "adjusting" the dissociation constant of the collector acid. The increased solubility (see Experimental section) of mercaptobenzthiazole in acid solution indicated that an amine ion (IV) is also formed
2
4
8
6
10
PH. Fig. 1.-Flotation of pyrite at 35' by mercaptobenzthiarole. The full curve is a line of 30% flotation. The broken curves me for constant free acid. Concentration (in mg./l.) for lines of constant ratio of collector ion to hydroxyl ion.
TABLE I RELATION BETWEEN pH VALUEAND ADDITION OF MERCAPTOBENZTHIAZOLE [XI TO CAUSE 30% FLOTATION OF PYRITE AT 35" (pKw = 13.6,pK. = 6.5) 1x1 x IO&, [HX] X 106, PH
6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5
mole/l.
6.0
7.8 10.2 12
18 30 45 132
mole/l.
5.7 4.9 3.1
1.37 0.69 -37 .18 .17
4.6 3.9 2.5 1.1 0.55 .30 -14 .13
It can be readily shown from equations 12 and 14 that, when the adsorption sites are not entirely covered by one of the species OH- or X- that I
I1
[X-I =
IOH-1
(23)
IV Assuming no salt effect and that the solubility (170 mg./l.) of the free acid I1 is the same as for the tautomer 111, the acid dissociation constant for the amine is calculated as 5 i= 3 X low3but is probably lower than this. If the free acid was the collector, then at a pH value of 2 a large increase in added amount of collector would be required: this is not found. However even if the only data considered are in the p H range 6 to 9, where the influence of amine ion must be vanishingly small, the quantity [X-]/(OH-) or [HX] decreases rapidly as shown in Table I. Neither theory is satisfactory. I11
(12) C. N. Merriam and H. C. Thomas, J . Chsm. Phys., 84, 99 (1956). (13) K.L.Sutherland and I. W. Wark, Trans. A m . I n s t . Minine Met. Engra., 184, 53 (1939).
where 4 = 1/Ox - 1. The term on the right-hand side is constant so that partial constant occupation of sites provides no explanation of the inconstancy of the ratio. I n this and other systems, e.g., dodecylamine as collector, it appears that both forms of the collector-ion and free acid or free base-may act as collectors. It is entirely logical to suppose that a t the pH value concerned it is the predominant species which is the collector. This behavior is apparent for dodecylamine or octylamine adsorbed on platinum where a constant contact angle, presumably due to the amine ion, arises in solutions more acid than a pH value of 8 and then increases considerably in the neighborhood of the pK, (= 10.4) of the amine (see Zisman and ShafrinI4: presumably due now to adsorption of free base.
ARTHURVEIS AND JOANANESEY
1720
Vol. 63
Sutherland and Wark13 studied the depressant action of cyanide on pyrite with mercaptobenzthiazole as collector. Table I1 shows the data (columns 1 and 2) determined from Fig. 2 in their paper.
varying greatly with the extent of occupation of the surface or that OH- and CN-, say, can compete simply with one another and provided the sum of these is large enough, flotation is inhibited. Experimental
TABLE I1
The mercaptobenzthiazole was purified by partial precipitation from a solution of the sodium salt by addition of acid followed by two crystallizations from hot alcohol. The melting point was 177'. The dissociation constant of the mercaptobenzthiaeole was determined from the e.m.f. of a hydrogen-calomel cell and solutions of the free acid and sodium salt of the thiazole. The solutions were prepared from conductivity water free from carbon dioxide. The method was checked by determining the pK, of pure benzoic acid a t 25' which gave 4.3 f 0.1. The pK. of mercaptobenzthiazole was found to be 7.0 at 25' and 6.5 a t 35' both & 0.2 unit. The solubility of the mercaptobenzthiazole was determined by shaking with water or lo-* N HCl solutions for periods up to a month. The total solubility in water was 180 f 5 mg. per liter and the final p H value of the solution was 5.45. The calculated solubility of the free acid is 170 i 5 mg. er liter at 35'. In solutions containing 10-8 N HC1 the totarsolubility was 200 f 10 mg. per liter.
&fERCAPTOBENZTHIAZOLE(20 MG./L.) AS A COLLECTOR FOR PYRITE Temp. 35"; cyanide as depressant: lHXl = free acid
9.3 9.0 8.5 8.0 7.5 7.0 6.5 6.0
0 1 3 8 20 47 102 184
0 0.25 2.8 7.8 20 47 102 184
0.019 .038 * 121 .37 1.11 2.9 6.1 9.2
0 0.019
.lo2 .35 1.09 2.9 6.1 9.2
.. 7.6 3.6 4.5 5.4 6.2 6.0 5.0
-
Mean 5 . 5
The ratio [OH-]o/[X-]o - [OH-]/[X-l X [D-]/[X-] is constant as required by the ionic theory: this ratio is more succinctly written as in the last column of Table 11. Thus we have the apparent contradiction that, in terms of the model, competition between OH- and mercaptobenzthiazole ions is quite inadequately described but that the competition between OH- and CN- ions with this collector can be accurately described! This feature is common to many flotation systems including a wide range of types of collectors and depressants. It suggests either that the adsorption of collector ions (or molecules) is quite different from that of the competing depressant ions, e.g., the activity coefficient of the surface species (14) W. A. Ziaman and E. G. Shafrin, J . Colloid Sei., 4,571 (1949).
I
The flotation tests were performed by shaking mineral with solution in a closed cylinder (Edwards and EwersI5). The pyrit,e was ground and elutriated to give a sized fraction between 100 and 200 mesh. The solution was adjusted to the pH value required using either NaOH, K2C03 or HC1. Terpineol (40 mg. per liter) was used as frother: it produced no flotation without addition of mercaptobenzthiazole. The curve recorded in Fig. 1 was the result of 114 tests and is the line representing 30% flotation. The p H value of the solution was measured with a glass electrode. Acknowledgments.-I wish to thank Mr. L. F. Evans and Mr. F. Meadows for the experimental determinations of flotation, dissociation constant and solubility. (19)G. R. Edwards and W. E. Ewers. Australian J . Sci. Research* 8 4 , 637 (1951).
CONFIGURATIONAL TRANSITIONS IN GELATINS IN NON-AQUEOUS SOLUTIONS BY ARTHURVEIS AND JOANANESEY Central Research Laboratories, Armour and Company, Chicago, Illinois Received Februara, 98, 1969
In aqueous solutions disordered gelatin molecules cannot be returned t o the ordered helical structure of native collagen. However, since synthet)ic polypeptides and some proteins can be made to go through the helix-coil transformation reversibly by the appropriate choice of solvent systems, a similar attempt has been made with gelatins. The behavior of gelatin in formic acid (FA)-dimethylformamide (DMF) and in ethylene dichloride-dichloroacetic acid has been examined. The studies cover ( 1 ) an analysis of the optical rotatory power and rotatory dispersion as a function of solvent composition; (2) the configurational changes as described by viscosity measurements; and (3) the roles of aggregation and degradation as determined by light scattering. [ c Y ] * ~ . ~isD -116' for gelatin in FA and becomes more positive as the D M F content is increased, reaching a value of -58' a t 5y0 FA in DMF. There is a discontinuity in a plot of [a] D us. FA concentration a t the molar mixing ratio FA/DMF = 1. Similarly, there is a discontinuit in a plot of [v] us. F A concentration at FA/DMF = 1, with a sharp rise in [ v ] when FA/DMF
