May, 1917
'
T H E JOURNAL OF I X D U S T R I A L A N D ENGINEERING CHEMISTRY
481
under the boat and vapors from the oil soon ignite was found in the Cushing oil (Bartlesville sand). a t t h e mouth of tube E. The heat is carefully in- Hydrogen sulfidelis formed from the oil when it is creased under t h e boat until no more volatile matter heated above 2 2 5 ' . can be driven off. Great care must be exercised t o c5-Most of the sulfur in the Cushing oil is probably regulate the heat and the flow of carbon dioxide so t h a t in the form of alkyl sulfides. the oxides will not pass over too rapidly for complete PETROLEUM TECHNOLOGY LABORATORY OKLAHOMA A. & M. COLLEGE absorption. The flow of water from the aspirator STILLWATER, OKLAHOMA should be regulated so t h a t neither undue pressure nor suction is developed in the hot tube. After all THE THEORY OF ORE FLOTATION volatile matter has been driven over, the carbon dioxide By H. P. CORLISSAND C. L. PERKINS is shut off a t the generator and clamp I is opened Received February 5, 1917 so t h a t oxygen will pass over the residue in the boat. The physics and chemistry of ore flotation constitute Clamp 3 is shut down until a very small quantity of the subject of extensive literature, but no one contrioxygen passes through. When the residue in the boat bution presents an explanation of all the physicohas been completely oxidized the absorption tube is chemical factors involved. These articles' include disconnected and the flow of oxygen stopped. The collectively considerable information of importance, but liquid and beads are transferred from t h e absorption have failed t o elucidate this very obscure problem. tube t o a beaker and boiled with finely divided silica In the present paper there will be presented an ext o decompose any excess hydrogen peroxide. After planation of the actual factors involved in ideal flotacooling, the sulfuric acid is titrated with standard tion and also of other practical observations incident sodium hydroxide after adding a drop or two of phenol- t o the art. The theory presented herein has been phthalein indicator. I t is convenient to use alkali worked out and substantiated by actual experiment. so standardized that I cc. is equivalent to I mg. However, for the sake of brevity, only a brief resum4 sulfur. Blank analyses should be made in order t o of the experimental results is included. correct for any sulfur which may be present in the The greatest success in the a r t has been obtained reagents. The entire process may be completed in in processes in which a gas, usually air, is introduced less than two hours, b u t the length of time required into the pulp, either by chemical means from carseems t o depend upon the amount of oil weighed out. bonate and acid (Potter-Delprat process), by vacuum Four determinations by this method gave: Total (Elmore process), by agitation (Minerals Separation sulfur 0.344, 0 . 3 7 9 , 0.343 and 0.3j1 per cent, an process), or by blowing it in through a porous blanket average of 0.354. (Callow process) and with or without the use of oil. Results agreeing very closely with each other were ob- The explanation offered in this paper is for this type tained b y each method but those by the Eschka method of process especially, though the simple flotation were so much lower than the other results t h a t i t is principles involved in such processes as the Maquisten evident this method permitted some loss of sulfur. and Wood and the bulk oil process are included. The method of fusion with sodium peroxide in a Parr I n all these processes t h e material floated must not bomb and the modified combustion method appeared be wholly wet by the water or solution in the presence to give reliable results. The greater simplicity of the of this gas or the material surrounding this gas, for bomb method makes it more desirable than the com- example, an oil film on the bubble surface. If the mabustion method, which requires the greatest care and terial is completely wet by the water, it will not float, vigilance on the part of the analyst. which is the case of the ideal gangue, while the material An attempt was made t o estimate sulfur in petro- floated must go t o the interface water-air bubble or leum by burning in an ordinary combustion furnace entirely into the phase other than water, e . g., the oil and passing the evolved gas through standard iodine on the air bubble. solution but concordant results were not obtained. The relations of the forces acting t o produce this Additional investigations were made of methods in- result were first stated by Freundlich2 and enlarged volving the treatment of the boiling oil under a reflux upon by Hoffman3 and R e i n d e r ~ . ~They were first condenser with fuming nitric acid and the addition of stated for the behavior of a sol, which will be called potassium dichromate. potassium chlorate and bromine disperse phase 3 in liquid I , when shaken with an imb u t low results were obtained in each experiment. miscible liquid 2 . Let SUMMARY
I-The methylene blue test provides an extremely . delicate test for the detection of any form of sulfur in petroleum. It may also be used for determining the presence of hydrogen sulfide. 2-The Eschka method for sulfur does not appear to be applicable t o petroleum. 3-Fusion with sodium peroxide in a bomb and the modified Dammer combustion method seem t o be accurate for determining sulfur in crude oil, l q - N o sulfur as hydrogen sulfide or carbon disulfide
TI8
= Interfacial tension between phase C and liquid 1
T& = Interfacial tension between phase 3 and liquid 2
= Interfacial tension between the two liquids If T,,I> T1.i TI,*the sol will remain unchanged; If Ti,*>Ti,r TI,, the disperse phase 3 will go entirely into liquid 2; If TI,,> T2.a TI,, the disperse phase will collect,at the liquid-liquid interface and will, if posslble, separate the two llqulds from each other. Tl.3
+ + +
W. D. Bancroft. J . Phys. Chcm., 10 1 See especially the following: (1915), 275; Ralston, Minrng and Sci. Press. Oct. 23, 1911; Callow, Am. Inst. Ming. Eng. Bull.. Dec.. 1911, 2321; Anderson, Ibid., July, 1916, 1119; and Taggert and Beach, I b i d . , Bug.. 1916, 1373. For a very complete bibliography, see School of Mines & Metallurgy, Univ. of Missouri, Bull. 8, No. 1 , 1916, also Am. Insl. Min. En& Bull., 1916, 1131. 8 "Kapillarchemie," 1909, 137, 174. 8 2. phys. Chem..83 (1913). 384. 4 Kolloid 2.. 13 (1913). 235.
T H E J O U R N A L O F I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
482
If, however, no one interfacial tension is greater than the sum of the other two, then the disperse phase will collect a t the liquid-liquid interface, but the three phases will meet a t a certain contact angle. The application of these principles t o flotation may now be stated, for while the greater part of t h e material floated is much less disperse than t h a t which is considered colloidal, the interfacial tendencies are the same, it simply being a question if the forces holding the mineral t o t h e interface are sufficient t o overcome gravity, if the particle is t o float. WITHOUT THE USE OF OIL-This is known as the Potter-Delprat process, in which C 0 2 is generated in the acid pulp, but may be carried out successfully on some ores in a Callow cell, using air. Here, if flotation is t o result, the mineral must go t o the interface water-gas and be carried a t this interface to the top of the pulp. The word water will be used mostly t o denote the aqueous phase, whether it is pure water or a solution, and the flotable material will be called sulfide, since this is the common case. On the basis of interfacial tensions, where if
-
Ts,a Interfacial tension sul6de-air (or C o d . Ts,w = Interfacial tension sul6de-water, Tw,a = Surface tension water air (or C o d ,
+
either ( I ) T,,u> T8+ TW,=or ( 2 ) no one interfacial tension is greater than t h e sum of the other two, must be true. It is obviously ,impossible t o have Tw,a>Ts,a T8,v as the latter two are very large in comparison with the first, according t o theoretical reasoning and measurements.' Case z is t h e actual one, as can be seen if a drop of water is placed on a flat sulfide surface. Here the water does not spread over the entire surface, but comes t o equilibrium with t h e three phases, sulfide, air and water in contact a t a certain angle. Case I would require t h a t the water should not wet t h e sulfide a t all in presence of air. I n flotation then the sulfide comes t o the air-water interface and sticks through t h e bubble surface t o a certain extent, or is held in such a way t h a t the three phases are in contact. The gangue material is completely wet by water and does not float; i. e., T g , a > Tg,w T W , a . Some measurements were made t o get an idea of these interfacial tendencies, by a method explained by t h e use of Fig. I. Here a flat ground mineral surface was placed vertically in water or other solution as shown. By raising and lowering the mineral, a quite constant result was obtained for the rise of the meniscus against the mineral above the general level. Here t h e meniscus was always upward, showing a greater preference of the mineral for water than for air. I n the case of t h e sulfides, when they were raised, t h e meniscus would soon draw back t o a definite height, leaving t h e sulfide surface above quite dry. For gangue the water does not draw back quickly, but remains, wetting it for some time. The sulfides are proven interfacial in this' way and the measurements of the height of t h e point of contact above the general level are interesting. The measurements were made with a cathetometer.
+
+
1 Hulett, Z. phys.Chem., 97 (1901). 385. molten metals and fused salts are bigh.
Also the surface tensions of
MATERIAL
Water
. . . . . . 1 . 5 5 mm. . . . .. ., .. .. 23.20 .60
Chalcocite.. . . . Chalcopyrite.. , . Gangue (silicate), , ,
,
0.1 Per cent -04 2 . 1 0 mm. 2.50 3.25
Vol. 9, No. j 0.1 Per cent NaOH 3 . 0 7 mm. 2.90 3.30
The figures for the gangue are not a t t h e point of contact, for there is none, since it is thoroughly wet by water, but are at t h e point where the meniscus becomes parallel t o the face of the mineral surface. The mineral giving the smallest rise should be t h e most interfacial and the best floating. This was found t o be true, for, without oil, chalcocite is a better floating mineral than chalcopyrite, a t least for the ores t h a t were tested. The figures above also show t h a t in alkaline solution a very poor float should be made, as t h e rise is almost as much as for the gangue. This was also found t o be true. Differences even among sulfides are clearly shown, hence it is not surprising t o find all gradations in floating properties among ores. These measurements, made on large pieces of mineral with ground and partially polished surfaces, may not correspond exactly t o those for an ore surface, though in the cases mentioned above they were found t o give results agreeing with practice. Another point noticed in these measurements, which is an important one, is how quickly t h e water is displaced from a mineral surface when brought in contact with air. If an air bubble comes in contact with a sulfide particle immersed in water, it must partially displace the water from the sulfide rather quickly, if it is t o be floated in a pneumatic cell. This was tested for the same minerals, by noting $he time taken for the solution t o come back t o the final point of contact, when the mineral was raised, .with the following general results: ( I ) Water and acid solutions are removed more quickly in air from chalcocite than from chalcopyrite. ( 2 ) Little difference is noted between acid and neutral solutions. (3) Alkaline solutions are removed very slowly from all surfaces. (4) All solutions adhere strongly t o gangue. These facts also agree with t h e practical results mentioned above. The success of the Potter-Delprat process may well be due t o these facts, since the CO? is generated in contact with t h e sulfide and time is given for t h e solution t o be partially displaced by t h e gas, or, in other words, for the sulfide t o attain the interfacial condition and be floated. When a soluble frothing agent is used, without oil, the same principles apply, t h e frothing agent simply modifying the water to a certain extent. WITH on-The use of oil introduces several new factors which make t h e problem more complex, but the same principles apply. The sulfides can now be interfacial between water and air as discussed above, but in addition may be interfacial between water and oil or even go into the oil layer. This oil layer is on the bubble surface and the forces holding the sulfides t o this surface, if it has an oil film, are much greater than when no oil is used. This point will be proven a little further on. The oil layer on the bubble surface need be only of minimum thickness t o act, in contact
May,
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
1917
with water, the same as a layer of oil on water, as far as interfacial tendencies are concerned. Let TS,, = Interfacial tension sulfide-water, TS,,= Interlacial tension sulfide-oil, To,, Interfacial tension oil-water.
-
+
Then if ( I ) TS,, > T , , To,, , the sulfide will go into the oil layer completely. ( 2 ) N o one interfacial tension is greater than the sum of the other two, the sulfide will go t o the oilwater interface and the three phases will be in contact at a certain contact angle. The gangue is thoroughly To+,. wetted by water, i. e., T,,,>T,,, These inequalities have been stated and applied t o the flotation process by Ra1ston.l The second condition given above, where the sulfides are interfacial seems to be by far t h e most general, though the first condition may be, and probably is, realized, especially when tarry oils are used which in grinding with the ore coat the sulfides more or less with this tarry material. It is doubtful if the lighter oils or the lighter constituents of a tarry oil mixture film t h e sulfide a t all in grinding, but rather it is probable t h a t this oil is emulsified in the operation. The condition where the mineral is completely filmed by oil would be the best floating condition, and this could be realized in the flotation cell, where this film would be continuous with the oil film on t h e bubble surface. All gradations of the interfacial conditions are possible, from those t h a t show only a slight tendency to be wet by water
+
I
I
I-------
/
I
PIO.I
I
FIG. I1
in the presence of oil, to those t h a t are thoroughly wet, which is the case of the gangue material. Experimental determinations of the interfacial tendencies of various minerals were carried out in the same way as described above, except t h a t in this case t h e interface was oil-water, or aqueous solution. I n Fig. I1 is represented the case of a sulfide surface a t this interface. The flotable materials were all interfacial and the sulfides showed a decided preference for the oil. This is a very important point in showing t h a t t h e same sulfides are much more strongly held to a n oil-covered air bubble than t o one not so covered. In Fig. I the sulfide, while interfacial, shows a preference for water over air and would easily be displaced in actual flotation from the interface and go back into the water. I n Fig. I1 t h e meniscus is now pushed downward into the water, instead of upward and hence the sulfide is held much more strongly t o oil than t o air. 1
Mining and ScicntilSc Press. Oc?. 23, 191s.
483
The following measurements were made after the meniscus had come t o the true point of contact of the three phases, and this point was closely the same whether the mineral was wet with the oil or solution first. The averages of these two figures are given. Kerosene and a kerosene pine oil mixture were used mostly, as the interfaces are better defined, especially in acid and alkaline solution, than with many actual flotation oils. These other oils act in the same way. however. DEPRESSION OF MENISCUS: KEROSENE A N D CHALCOPYRITE Water 0.10Per cent HzSOd 1 Per cent HzSOi 10Per cent HpSO, 2 . 9 9 mm. 2 . 0 2 mm. 1 . 3 2 mm. 0 . i 5 mm.
Calcite in contact with neutral, acid, and alkaline solutions and kerosene showed interfacial tendencies in alkaline solution only. Malachite exhibited a small interfacial tendency, except in alkaline solutiorl in which i t was thoroughly wet by the solution. KEROSENE AND PINE OIL A N D AQUEOUS SOLUTION This was a flotation mixture of 90 per cent kerosene and 10 per cent pine oil
DEPRESSIOX OF MENISCUS SOLUTION Chalcopyrite Chalcocite Water.. . . . . . . . . . . . . . 3 . 1 0 mm. 3 . 4 2 rum. 0.10 per cent N a O H . . . 1.98 2.54 0 . 1 0 per cent HzSOc.. . 1.45 2.95
. .. .. ..
Gangue material in all cases is thoroughly wet by the solution, especially if it is wet by the solution before coming in contact with the oil, as is the case in actual flotation. The case of chalcocite in water given above is almost a condition of complete wetting by oil. These experimental results in every way justify the theoretical discussion above and also show t h a t alkali and acid lower the interfacial tension sulfidewater as the preference for oil is not as great in these solutions as in water, although the sulfide is still decidedly interfacial and hence can be easily floated from acid or alkaline pulps. These results were obtained by the use of a clean sulfide surface, but in actual flotation this may not be true for all the particles, and since t h e interfacial properties are a function of the surface only, we may expect many differences from these ideal measurements. In alkaline solution, for example, there may be some of the mineral which, like calcite, is more interfacial in this solution than in water, and hence would float although it would not do so in a neutral pulp. I n tests i t has been found with some ores and oil mixtures t h a t in an alkaline pulp a better recovery was made in the usual length of time than by prolonged floating in neutral pulp. This might also be true in an acid pulp for some minerals. NATURE O F T H E SOLID SURFACE A X D HYSTERESIS OF T H E COXTACT ANGLE
It has been noticed t h a t some surfaces have a strong tendency t o hold fast t o the liquid first wetting them and not t o allow it t o be easily displaced by another liquid. I n t h e work upon interfacial tensions, described above, such a surface would show a great difference in preferential action or angle of contact, dependent upon whether it was wet with oil or water first. It has also been observed t h a t it is principally
484
T H E J O U R N A L OF I N D U S T R I A L . A N D ENGINEERING C H E M I S T R Y
those substances having smooth or shiny surfaces which float, while those having dull or rough surfaces do not float. These observations and others point t o the following explanation of t h e mechanism of this action: there is first the inherent property of each substance t o adhere t o oil or to water t o a certain degree. When the substance is brought t o t h e interface between water and oil, these forces tend t o come to equilibrium with the third force, the interfacial tension between oil and water a t some definite contact angle. Here is where the physical nature of t h e solid surface comes into play. If the surface is smooth and shiny, such as t h a t of a polished metal or a freshly fractured sulfide crystal, then t h e liquid first touching i t is easily pushed back t o the equilibrium position when brought to the interface with another liquid. If, however, the substance has a dull, i. e . , a capillary surface, so t h a t t h e liquid first wetting it is strongly held in its pores, then, when it is brought t o the interface i t may exhibit no interfacial properties a t all although, if it were smooth, i t might even show a preference for the other liquid. This shows the reason for the difference, “or hysteresis,” of the contact angle noted for some surfaces. * It also explains why a particle having such a surface, if first wet with water, as is the case in flotation, will be very difficult t o float, since it will not easily be brought into contact with oil. F U N C T I O N OF T H E BUBBLE-The function of the bubble is t o give a large surface to which the sulfide may go and be floated. As already stated, the air bubble in oil flotation is covered wholly or in part by a n oil film. For the action of oil on water, see Devauxl and Langmuir.2 It is not necessary t h a t the oil completely cover the bubble, and i t probably does not in the greater proportion of the bubbles. The supply of oil for the bubbles will be discussed under the action of emulsions. If an oil droplet is placed on water or aqueous solution, i t will spread out over the surface provided the surface tension of the water is greater than the sum of the surface tension of the oil plus the interfacial tension oil-water, i. e., this inequality must be true: Tv,o>To,a To,w* For oil flotation this must be true for all solutions used, as the air in the bubble, surrounded by t h e pulp, presents this same condition. If t o water is added some material which lowers its surface tension (TW+), without lowering Tw,o To+t o an equal amount, the inequality is reduced and finally a point is reached where the oil will not spread on the solution. This is easily realized in case of soap solutions, and with many other substances which lower the surface tension greatly, I n this condition a poor float would result. I n flotation, in order t o produce a froth, material such as the soluble portion of pine oil is added which lowers the surface tension of a-ater. Unless this helps in other ways than producing a froth, i t should be used in as small a quantity as possible, and this agrees with many practical observations The froth-
+
+
1
a
Ann. Report Smithsonion Insl.. 1915, 261. Met. & Chem. Eng., 16 (1916). 469.
Vol. 9, No, 5
ing agent added also lowers the interfacial tension oil-water, but here it must be remembered t h a t even if the interfacial tension is lowered in the same proportion as the surface tension, the inequality is less than before, since the interfacial tension is much smalIer than the surface tension of water. The other factor, the surface tension of oil (Toso),is not changed much, for inorganic salts, etc., do not dissolve in it. If, however, some substance is added which will not lower the surface tension of water but will lower the interfacial tension oil-water, then this should produce better oiling of the bubble. This can be done with alkalies and in the case of some oils by acids. An important point in connection with the use of the pneumatic cell is the time during which the bubble is in contact with the pulp as it passes through, as here i t must be attached t o the sulfide particles. Any reagent t h a t will give a quicker filming of the bubble surface by oil, after it comes through the blanket, will be of benefit in the rapidity with which the mineral is attached and raised. Alkalies, as explained, produce a greater inequality between Tu,o and Tu,o To,, and hence the oil will be spread out thicker over the surface than without their use. A large number of surface and interfacial tension measurements were made, a few of which are as follows :
+
SURFACE TENSIONS Dvnes Der cm. 71.8 Water 25’ C . . . . . . . . . . . . . . . . . . ....... 25.2 Kerosene. ........... ................ ....... 28.0 Coke-oven Oil., ..... ................ ....... .... 30.0 Pine Oil.. ................ ... .... 6 8 . 6 0.01 per cent solution Terpineol. 49.2 0.10 per cent solution Terpineol.
.......
.... .. .... .... ... .... ........... ....... ....... ....... .... ....... INTERFACIAL TENSIONS 32.8 Kerosene-Water.. .................................. Kerosene and Pine Oil-Water.. ......:................ 11.6 Kerosene and Pine Oil-0.05 per cent solution NaOH.. , . 7.3 Kerosene and Pine Oil-0.20 per cent solution NaOH.. .. 4 . 5
.... 13.2
Kerosene and Pine Oil-0.20 per cent solution HIS04 Coke-oven Oil- Water.. Coke-oven Oil-0.05 per cent solution NaOH.. Coke-oven Oil-0.20 per cent solution NaOH.. Coke-oven Oil-0.10 per cent solution NarCOi.. Coke-oven 011-0.20 per cent solution NazCOa.. Coke-oven Oil-0.20 per cent solution NaaBdOi.10 Aq.. Coke-oven Oil-0.10 per cent solution NarP107.10 A s . . Coke-oven Oil-0.20 per cent solution NarPaO7.10 Aq.. Coke-oven Oil-0.40 per cent Solution HzSO4.. Coke-oven Oil-0.01 per cent solution Saponin.. Coke-oven Oil-0.01 per cent solution Tannic acid Coke-oven Oil-0.01 per cent solution Hemoglobin..
............................. ......... 14.1 5.8
......... 2.6 ......... 6.6 ......... 4.4 .. .... 978 ... 640 .......... 14.4 ......... 9.3 ........ ..... 12.7 8.9
Considerable data of this kind are given by Lewis’ and Shorter and Ellingsworth2 on the action of dyes, salts and soap. The drop number apparatus used was the same as described by Shorter and Ellingsworth. Their work also shows t h a t soap and alkali together are very active in lowering the interfacial tension oil-water. This would be the condition in an alkaline pulp, as there would then be free alkali and some saponified material with many of the oils used. The results when colloidal material is present are subject to great variation, due to different speed of formation of drops. The figures given above for these materials approach the dynamic value, as the rate of dropping was fairly rapid. The static values are very much smaller and are interesting in connection with the emulsifying power of these substances. As an example of this the following result on coke12.p h y s . 2
Chem.. 74 (1910). 619. PYOC. Roy. Soc., 99 (1916), 231.
May, 1917
T H E J O U R N A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
oven oil against 0 . 0 0 5 per cent hemoglobin solution is given. The time is for t h e total number of drops formed. TIME 2 min. 40 sec. 1 hr. 4 min.
DROPNo. 22.5 84
INTERFACIAL T~NSION 13.2 dynes per cm. 3 . 5 dynes per cm.
It is seen from t h e table above, t h a t besides NaOH itself, any salt t h a t hydrolyzes t o give an alkaline solution lowers the interfacial tension, and all these salts are beneficial to flotation. EB6uLsIoNs-The behavior of the oil a t t h e bubble and sulfide surfaces has been given. I n the pneumatic cell this oil is supplied by an emulsion or a coarser suspension of oil in water. I n the agitator type machine, the oil may be beaten in a t the cell, though i t is also customary t o grind the oil with the ore. I n either case the problem of emulsions comes in. I n the pneumatic process this emulsion is formed in the grinding and must be good enough to last throughout the float, yet not so good as t o fail t o break down with sufficient rapidity to give free oil for the bubble surface. The subdivision of the oil is such t h a t no doubt almost all degrees of dispersion exist; the larger droplets may be of sufficient size for one t o coat a fair area of a bubble surface, but the better emulsified portion is of such size t h a t many particles have to unite t o give oil enough for the minimum thickness of an oil film, to spread over even a square centimeter. This can be calculated from the minimum thickness of an oil film' and the size of t h e particles in an ordinary oil emulsion.* Experimental evidence on these points is very conclusive. If a coarse suspension of oil is made simply by shaking the ore, oil and water together in a bottle by hand, and then put in a small Callow cell, only a partial float results and the operation must be repeated several times, adding more oil each time, in order t o get a good recovery. If, however, too good an emulsion is had, a poor recovery also results. For this purpose a kerosene pine oil mixture was emulsified with water in a De Lava1 emulser and allowed t o stand over night and a middle portion of this emulsion was removed for the tests. This emulsion added a t the cell gave a small float a t first and then stopped. On adding a little acid no further float resulted, but by allowing t h e pulp t o stand for a few minutes an additional amount of sulfide was raised and finally a good recovery was made, though considerable time had t o be given for the emulsion to give up its oil. This was also found to be true for another oil which gave an excellent emulsion on simply adding i t t o water. It is interesting t o note t h a t in these cases it was proven t h a t it was not necessary t o grind the oil with the ore,-but by adding it as an emulsion prepared by itself as good a recovery results. This probably is not true for oils containing tarry matter as explained above. It was also noticed in using t h e second emulsion, named above, t h a t flocculation of the slimes took place in neutral solution and t h a t these then floated 1 1
Devaux, LOG.cit. Ellis, Z . physik. Chem., 80 (1912), 597.
485
t o a large extent, giving a non-preferential float; when, on t h e other hand, the emulsion was broken b y acid and alum, a good preferential float resulted. It was found t h a t these slimes in neutral pulp had flocculated with the oil emulsion so t h a t o n standing all the oil was carried down, though the emulsion was not appreciably broken. The value of acid and salts having a polyvalent cation has been demonstrated in some cases, usually in connection with the M. S. type process. I n this process there is greater danger of getting too good an emulsion than in the Callow process, and the value of acids and salts of this type consists in their power of breaking down an emulsion, or preventing too good a one being formed. These salts should be used in acid solution, or otherwise, due t o hydrolysis, the insoluble hydroxides formed, e . g., Fe(OH)3, and Al(OH)3 have the opposite effect, i. e . , of preventing the breaking down of the emulsion or promoting its formation.' Oil emulsions in FeC13 solution on standing give a yellow flocculent precipitate, but the emulsion is not broken. The mechanism of this is discussed by E1lis.l I n a neutral, pneumatic, Callow float, such salts have been found to be harmful. If salts of iron or aluminum are present in the feed water then acid may be necessary to prevent this action between them and the oil emulsions. The value of alkalies has been discussed as giving a better oiling of the bubble surface. I n connection with emulsions, however, a greater effect can be ascribed t o the action of alkalies or salts which hydrolyze t o give an alkaline reaction and t o those which have a polyvalent anion. If a neutral ore pulp is shaken with a small quantity of an oil emulsion i t is found t h a t the slimes are coagulated with the emulsion and settle out, often leaving the liquid quite free from oil emulsion. The emulsion is not broken, but simply carried down with the flocculated slimes. If alkalies are used, or salts such as last mentioned, then the slimes are deflocculated in sthe great majority of cases. They then settle more slowly, and when they have settled the emulsion is left free and still standing. This is very important, for now the emulsion is free t o function as it should, i. e., t o give oil t o the bubble surface. The ore particles, both sulfide and gangue, are also free t o show their own behavior toward the water and the oil. This deflocculation should, and does, result in a higher grade concentrate and a greater and quicker recovery, since now no sulfide particles are coagulated with, or surrounded by, gangue particles t h a t prevent their flotation. The use of lime has not been found t o be as beneficial as t h a t of NaOH. This is explained by the fact t h a t this substance, due t o t h e predominating effect of t h e calcium ion, coagulates instead of deflocculating t h e slimes, and hence part of the emulsion is removed and the individual particles are not free t o float as they should. This coagulating action may be more noticeable in a Callow cell than in a cell of the Minerals 1 1
Briggs and Schmidt, J . P h y s . Chem.. 19 (1915), 478. Z . physik. Chem., 89 (1914), 149.
T H E J O U R N A L OF I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
486
separation type, as in the latter the coagulated slimes may be broken u p considerably, but the tendency is the same in either case. QUANTITY OF oIL-The principles involved when varying quantities of oil are used, is a question on which there is a great difference of opinion. From the theory there should be no difference whether a large or small amount of oil is used, provided this oil is properly emulsified. If a large amount, 2 per cent or 3 per cent, is used and is not emulsified sufficiently, the excess may float and be of disadvantage in several ways. T o test this point a float was made with an amount of oil equivalent t o 2 per cent of the weight of the ore, emulsified in a De Lava1 emulser, and added a t the cell (Callow) and a float made. It behaved in every way the same as when 0 .2 per cent or less of oil was used and the recovery was better, with as high a grade of concentrate. Of course, economy would settle the minimum amount of oil t o use. This was repeated with other oils and ores. The extra amount of oil used gave a greater oiling of the bubble surface, and in fact these floats were better than when alkali was used to make the smaller amount of oil more efficient. COLLOIDS A N D (LPoIsoNs”-In the light of the above work the question of flotation “poisons” was taken up with the idea t h a t any substance which will prevent the breaking down of a n emulsion or coalescence of oil droplets, or which gives adsorption of colloidal particles a t t h e oil-water interface is harmful t o flotation. I n the first two cases the proper amount of oil will not be freed, and in the other case the oil surface, if formed, would be covered by an adsorbed layer, so t h a t no oil surface would be presented for attachment of the mineral. Experimental work, by actual flotation, had shown what substances, including many dyes, were harmful. Solutions of these substances of 0.01per cent strength were shaken in test tubes, with about 2 cc. of oil, for a few minutes t o t h e same extent and a t the same time. The tubes were then placed upright and the amount of emulsification and rapidity of coalescence of the oil droplets rising t o the top noted, with the following results : ( I ) SLIGHT
OR
NO
EMULSIFICATION
AND
RAPID
DROPLETS-Methylene blue, saffranine and Bismarck-brown. These dyes really act like salts and are not colloidal, nor are they harmful t o flotation. I n fact, these dyes assist slightly in breaking an emulsion. COALESCENCE OF
(2)
E X T R E M E L Y S L O W C O A L E S C E N C E O F DROPLETS-
The finely divided oil layer lasting for several hours t o days : Congo-red, bengoazurin, azo-blue, saponin, tannic acid, waste sulfite liquor, hemoglobin and eosin. These substances are all very injurious to flotation. Most of these are negative colloids. Hemoglobin is highly colloidal and positive and its adsorption is probably enhanced because it is oppositely charged t o the oil emulsion. Several of this last class of substances, especially saponin, gave marked emulsification, even with the small amount of shaking received. Some of these substances also form quite stable and viscous skins at oil surfaces.
Vol. 9, No. 5
Another experiment consisted in dividing an oil emulsion in two parts t o one of which tannic acid was added, and then frothing over equal volumes of each in a small cell. The one t o which tannic acid had been added contained 3 . j times as much oil in the residue or tail water as the other. This shows t h a t the oil emulsion had been kept from breaking down, and the oil from being frothed over. Besides the substances given above, the injurious effect of insoluble hydroxides of the heavy metals has been explained under emulsions. Other inorganic colloids have been found to be injurious, e. g., when floating with K4Fe(CN)e, the Cu2Fe(CN)Gformed from the oxidized and soluble copper hurts the float very noticeably. The experimental evidence proves that the action of these colloids is, without doubt, as stated, though they may also adsorb a t the solid surfaces and in t h a t way cause a poorer result t o be obtained. It is easily seen how the water used in flotation and the slimes coming from certain ores have a great effect in flotation. This has caused some t o say t h a t it is t h e gangue t h a t determines the success of the process, and if the water supply be included in this, they are t o a certain extent correct. FROTHS-The froths produced in flotation are useful as a mechanical means of removing the mineral brought up by the bubble. The formation of a froth and its stability are due principally t o dissolved materials in the water which give to the solution a variable surface tension. The static surface of a solution has a lower tension than a fresh surface, whether the substance added lowers or raises the surface tension of t h e solvent. Since a large lowering may be caused by a small amount of solute and only a small rise may be obtained, the best frothing agents are those t h a t lower the surface tension. Pine oil is used t o a large extent for this purpose in practice, the soluble portion causing a considerable lowering of the surface tension of water. I n many articles t h a t have appeared on the theory of flotation, it has been stated t h a t oils lower the surface tension of water. This is not very clearly stated, since, as ordinarily understood, oil is insoluble in water and only soluble material can affect the surface tension of water. Besides the soluble portion of pine oil, a part of many other flotation oil mixtures is soluble and gives a froth. Terpineol, menthol and many such substances are very powerful frothing agents. The lasting qualities of a froth, as stated above, are due to its variable surface tension, for if a bubble starts t o thin out or break’at a certain point this fresh surface has a greater surface tension than before and hence is automatically strengthened a t this point and resists rupture. I n using alkalies it is observed t h a t a more quickly breaking froth results in a pneumatic cell. This can be explained by the fact, as stated before, t h a t a greater extent of bubble surface is covered with oil and hence there is less surface which contains only the adsorbed frothing agent, and since oils themselves do not produce good froths, the froth breaks more quickly than when alkalies are not used. Or, this observation may be used
May, 1917
T H E J O C R - V A L OF I N D U S T R I A L A N D E N G I N E E R I N G C H E M I S T R Y
t o support the view t h a t the bubbles are better oiled . froth is also stabilized by in an alkaline pulp. 4 the slimes present in a pulp or by other colloidal matter. Colloidal material dissolved in the oils will make an oil froth more lasting. A mixture of oils, the same as an aqueous solution, gives a better froth than a pure oil. E L E C T R I C A L EFFECTS-Considerable weight has been placed by many upon the electrostatic forces that might be present in the flotation process. Some have even considered the attraction t h a t holds the sulfide t o the bubble surface t o be of this origin. Air bubbled through water has been found t o carry ions,' and from this and t h e fact t h a t most substances have a contact difference of potential when in contact with water or solutions, an electrical theory has been built up, though in many cases serious errors have been made regarding the action of these forces. Measurements were made t o determine these forces. The small metal Callow cell used was grounded, as this condition prevails in actual practice. The charge carried by t h e air issuing from the flotation pulp was discharged on a metal screen placed above the cell, and the effect measured b y means of a Dolezalek electrometer. The readings in this case are measured in 1-olts per minute. The charge upon the air from several pulps was measured and in no case did it exceed 0 . 0 1 1 volt per minute, and was usually only about half the value. The air was negative in neutral pulps, but slightly positive in one of the alkaline pulps. The charge on the froth was also measured and this varied from zero t o 0.011 volt as the maximum. This was sometimes positive and under other conditions negative. I n two good floating pulps the froth was a t almost zero potential, though 0 . 0 0 2 volt could easily be determined. It seems, then, t h a t these electrostatic effects are far too small t o have any important part in flotation, and cannot possibly he the force t h a t holds the sulfide t o the bubble. This, too, would require a dielectric film, e. g., oil, between the two oppositely charged bodies, the sulfide and the gaseous ions in the bubble; but since flotation results without the use of oil in many cases, and without doubt the bubble surfaces are often not completely covered b y oil even when oil is used, it seems t h a t this theory cannot hold. The contact difference of potential of various minerals has been used in some theories. These were also measured by an electro-endosmose method as described b y Perrin.* T o this apparatus a small calibrated tube was sealed a t t h e top of the diaphragm side, so t h a t when dilute electrolytes are used t h e gas generated can be forced over into this tube, after the experiment is over, and this correction applied t o the amount of liquid apparently transferred through the powdered material. The distance between the electrodes was I Z cm. and the potential I I O volts. The results obtained give the sign of t h e charge on the solid in contact with the water or solution, but Lord Kelvin, McLean and Galt, Proc. Roy. SOC.,1894, 5 7 ; Coehn and Mozer. A n n . Physik, 48 (1914), 1048. 2 J . chim. phys., 2 (1904), 601.
487
quantitative results as t o the actual potential differences are very difficult t o obtain in this way. However, some idea can be had by comparing t h e amount of liquid transferred for the minerals, t o t h a t transferred in the case of silica, whose potential difference against water has been found by cataphoresis measurements. This is found t o be approximately -0.042 volt. For quartz and ferric hydroxide, see Whitney and Blake.' The results obtained are as follows: MINERAL Silica Alumina Chalcopyrite Galena Sphalerite Molybdenite Malachite Malachite Galena
LIQUID Water X/lOO HC1 Water Water Water Water Water N/500 HC1 F. W./500 FeCla
Sign of Solid Negative Positive ? Negative Negative Negative Positive Positive Positive
Liquid Transferred Cu. mm. per min. 30.7 40.0 Approx. 0 3.6 6.1 3.7 4.0
17.8 44.3
Here the sulfides tested are seen t o be slightly negative against water or practically zero in case of chalcopyrite. This agrees with our ideas concerning the contact difference of potential of these substances and with cataphoresis experiments on colloidal sulfides, etc. Malachite is positive, as would be expected from its basic character. The last result given in the table is probably due t o the formation, by hydrolysis, of ferric hydroxide, and its adsorption on the surface of the mineral, so that the action is exactly the same as for ferric hydroxide itself. I n this case again, we see t h a t no attraction can exist on the basis of electrical charges between sulfides and oil in emulsions, since they are of the same sign. The charges on oil in emulsions in dilute salt solutions, etc., are given by Ellis2 and Powiss and others. This, however, would not determine the charges on a mineral and oil, if the two were in actual contact, as is necessary for flotation. The charges carried by the oil in emulsions are important probably in connection with positively charged colloids which act as poisons, and, of course, the coagulation of slimes and the breaking of an emulsion by electrolytes is a function of the charge carried by them; but it is not possible to use these charges as an explanation of the primary principles involved in flotation. CONCLUSIONS
The following is a summary of the conclusions arrived a t as a result of the work reported in this communication: I-For an ore particle to float, it must be interfacial between oil and water or i t must go completely into the oil phase. If no oil is used, the particle must be interfacial between water and air. The force holding t h e particle t o t h e bubble is much greater when oil is used. 11-In addition t o its value as a lifting agent, the bubble serves t o produce a large air surface, in contact with the pulp. This surface is covered t o a greater or less extent by an oil film, t o which the mineral may go, so t h a t a small amount of oil is very efficient. 111-The oil should not be so well emulsified t h a t it will not be given up to the bubble surface; and yet J . A m . Chem. Soc.. a6 (1904). 1339. 2. phys. Chem., 18 (1911). 325. a I b i d . , 89 (1914). 91. 1 2
488
T H E J O U R N A L O F I N D U S T R I A L A N D ENGINEERING C H E M I S T R Y
should be sufficiently emulsified, in a pneumatic process, to last during the time of floating. IV-Colloids in general are harmful, owing either t o their causing too stable an emulsion, or t o their adsorption on the oil film at the bubble surface preventing mineral attachment. This is t h e action of the so-called “flotation poisons.” V-The froth formed is attributable either t o the soluble portion of the flotation mixture, which produces a variable surface tension, or to finely divided or colloidal materials. VI-Acids, alkalies and salts affect all these factors, as discussed under the several headings in the paper. VII-The electrical effects, other than the colloidal charges, are not important in flotation. VIII-The nature of the solid surface in relation t o its wetting properties has been discussed and an explanation of the “hysteresis” of the contact angle advanced. I n the light of present knowledge, it is impossible to measure many of the forces operative in flotation, such as, for example, the interfacial tensions between solids and liquids, or t o explain the mechanism of adhesion. Such problems are, however, nearer solution, due t o the material advances made recently by Lauel and by Bragg and Bragg,z by which the actual arrangement of the atoms in a crystal may be determined, and also by Langmuir,a whose work on the constitution of solids and liquids, the structure of solid surfaces, and the mechanism of adsorption leads toward the solution of this problem. While the flotation of each ore still remains more or less of a problem in itself, yet a clear understanding and the proper application of the principles involved will lead t o an earlier solution of the problem. I n conclusion, the authors wish to express their thanks to Dr. R. F. Bacon and t o Mr. E. R. Weidlein, under whose direction this research has been carried out. MELLONINSTITUTE
OF INDUSTRIAL
RESEARCH
PITTSBURGH
-~ NOTES ON THE ANALYSIS OF CAST NICHROME B y E. W. REID Received March 6, 1917
It is no new experience for chemists t o find that each alloy has its own personal equation in yielding itself t o sharp analytical results, and the remarkable group of mixtures which are coming into such general use under the popular name “Nichrome” form no exception. The following notes are submitted, not with the supposition t h a t they are the best ideally possible, but with t h e hope that they may help t o call the attention of those specially interested in related lines of research to the need of reliable methods for the analysis of this group of alloys. 1
1
Site. Akad. Wh., Wien, June, 1911. Proc. Comb. Phil. Soc.. 17 (1912). 43; and treatise on “X-Rays and
Crystal Structure.’’ I J . Am. Chcm. Soc., 88 (1916). 2221.
Vol. 9, No. 5
Some difficulty will usually be experienced in getting the alloy into solution, and after several solubility determinations, the following method was adopted. There is always a slight residue left in the bottom of the casserole after treatment with hydrochloric and nitric acids, which appears t o be small particles of metal enclosed by gelatinous silica; hence the usual necessity of first removing the silica, and then dissolving the residue in acid, with subsequent fusion of any undissolved chromium with sodium peroxide. Cast nichrome contains approximately 58 t o 6 2 per cent of nickel, 2 3 to 2 6 per cent of iron, 8 to 14 per cent of chromium, 0.5 to 2 . o per cent of manganese, zinc and silica, 0 . 2 t o I . o per cent of carbon and sometimes a bare trace of copper. The ingredients were determined in the order given.
.
SOLUTIONS
AMMONIUM
CHLORIDE-saturated
solution.
HYDROCHLORIC ACID (SP. GR. I . 12)--8 12
cc. hydrochloric acid (sp. gr.
N/IO
POTASSIUM
Water t o
cc. water t o 40 cc.
SULFURIC ACID (SP. G R . 1 . 4 0 ) - 4 3
sulfuric acid (sp. gr.
CC.
I . 20).
1.83). PERMANGANATE-3.161
salt dissolved in water, diluted to
I
g. pure
liter.
F E R R O U S A M M O N I U M SULFATE-39. 2 g. pure salt dissolved in 500 cc. water and 50 cc. concentrated sulfuric acid added; diluted t o I liter. POTASSIUM IODIDE--:! per cent solution. POTASSIUM CYANIDE-13. j g. pure Salt and 1 5 g. Of potassium hydroxide dissolved in water, diluted t o I liter. N / I O SILVER NITRATE-8. 49 5 g. of the salt dissolved in water, diluted to I liter. TARTARIC OR CITRIC ACID---:! 5 per cent solution. BROMINE WATER-satUrated solution. POTASSIUM FERROCYANIDE-2 I . j 5 g. pure Crystallized salt dissolved in water; diluted to I liter, DIMETHYLGLYOXIME-I per cent solution. URANIUM ACETATE O R NITRATE-I5 per cent solution. “STOCK” soLuTIoN-(Described under Silicon). SILICON
Dissolve 2 . j g. of nichrome turnings in a 2 50 cc. casserole, with 20-30 cc. concentrated hydrochloric acid and 3-5 cc. concentrated nitric acid. Evaporate the solution t o dryness, take up with the above amounts of acids, again evaporate’ to dryness, and ignite t o redness for a few minutes. Take up with hydrochloric acid (sp. gr. 1 . 1 2 ) and a few drops of nitric acid; bring t o boiling; dilute with cold water and filter. Wash the residue on the filter paper thoroughly with dilute hydrochloric acid ( I : 3 ) , and finally with hot water. Carefully ignite the filter paper with its contents by means of a platinum wire over a platinum crucible, and after the ash and residue are allowed t o fall into the crucible, ignite to a high tempe;ature for ~j to 2 0 minutes, or until all the carbon of the paper is burned: cool in a desiccator and weigh. Then moisten the residue with a few drops of sulfuric acid I The evaporations may be accomplished in a short time by manipulntion of the casserole over a free flame. observing the usual precautions.
