Detergency of Alkaline Salt Solutions 111. Deflocculating and Emulsifying Power FOSTER DEE SNELL,130 Clinton St., Brooklyn, N. Y.
0
KE of the properties of a measure of t h e e m u l s i f y i n g For measurement of the deflocculating and detergent which must be pover. Experiments were also emulsifying power of a detergent it is recomconsidered in estimating made using oiled carbon black in mended that the size of the suspended oil droplets its value is the a b i l i t y to deplace of mineral oil, a procedure be .fixed by using an inert material such as burnt flocculate and emulsify oils or w h i c h h e c o n s i d e r e d more umber with a superjcial coating of oil. oil-covered dirt particles. By promising. such action, dirt having a superA procedure of the same type I n blends of soap and alkaline salts the deficial layer of oil, which has been as that of Stericker was used by jloccuhting and emulsifying ralue is the sum of removed from a surface by virtue B a k e r (1) in c o m p a r i n g the the aalues of the ingredients when used singly. of the w e t t i n g p o w e r of the emulsification produced by soluOf the series studied, sodium oleate had the detergent used, is dispersed and tions of s o d i u m h y d r o x i d e , greatest zdue, with sodium metasilicate a close prevented from r e d e p o s i t i n g sodium metasilicate, trisodium during the washing process and phosphate, and sodium carbonsecond, and trisodium phosphate third. is carried away at the end of ate. His method was to stir 75 Data on disodium phosphate, which is too lou the o p e r a t i o n . A detergent cc. of the solution with 75 cc. of in Cox to be a true soup builder, show that it has which has w e t t i n g power but a light motor oil for 5 minutes a t protectitTe colloidal properties but it is udmnlacks deflocculating and emulsi20" C. in a motor-driven drink tageous only in the absence of soap. fying power will m e r e l y cause mixer. The m i x t u r e s w e r e dirt to be spread over a surface ooured into bottles and allowed more evenly rather than cause it t o be removed permanently. to stand 24 hourb, and the iepths of the oil emulsion and Measurements of this property have been carried out on a aqueous layers were then measured. Comparison of the emulseries of salts which have already been compared as to neu- sion layers obtained rrith the different solutions showed sodium tralizing power as represented by available and total alka- metasilicate and trisodium phosphate to be much better emulsilinity (4),and as to wetting power as represented by inter- fying agents than the other compounds. facial tensions (6). I n these determinations the series of Measurements of deflocculating power have been made salts previously compared was shortened to include only by Fall ( 3 ) . He used manganese dioxide as his soil, and dethe sodium silicate with a KazO:SiOz ratio of 1: 1 and to termined by titration the quantity of this soil that was susomit borax and sodium chloride. pended in the detergent solution after shaking for 5 minutes under standard conditions. This method yielded quantiPREVIOUS METHODS OF MEASTJRIKG EJIULSIFYIXG tative results. He found silicates of soda to be almost as POWER effective as soap in suspending manganese dioxide, and more Various methods of measurement of emulsifying power valuable than sodium hydroxide, sodium carbonate, or trihave been used previously of which the following are repre- sodium phosphate as addition agents to soap. When a definite proportion of neutral mineral oil is agisentative: Stericker (6) compared the emulsifying power of alkaline salts and soaps by shaking 12.5 cc. of the solu- tated with an aqueous solution of emulsifying agent under tion with 0.3 cc. of a neutral brown lubricating oil. The standard conditions, it is reasonable to assume that the nature mixture was shaken 30 seconds, allowed to stand 5 minutes, of the emulsion will depend on the efficiency of the emulsiand then placed in an oven a t 80" C. for 75 minutes. The fying agent. As between an efficient and an inefficient condition of the emulsion was noted a t each step. Of the emulsifying agent, the dispersion in the case of the former salts tested, he found the most siliceous silicates to be the will be finer. It is desirable further to subdivide the effect of emulsifying agents into two effects-stabilization of oil best emulsifying agents. Vincent ( 7 ) , using highly purified petroleum oil, found droplets by preventing coalescence, and stablization by the sodium silicate to have no value for emulsifying oils and be- effect of a protective colloid in preventing concentration of lieves that the oils used by Stericker were slightly contami- the particles in response to the effect of gravity. So difnated with fatty acid so that the emulsification was due to ferentiated, it is possible that two agents will give equal stability of emulsions, but the nature of the emulsions may soap formed. He found that other alkaline compoundsin particular, trisodium phosphate, sodium hydroxide, and be radically different. I n laundry operations the dispersion of large amounts ammonium hydroxide-aided emulsification. His method was to shake 25 cc. of soap or salt solution and 25 cc. of of oil alone is unimportant. The procedure and discussion kerosene oil twenty consecutive times in a 100-cc. oil sample assume that a good wetting agent does not remove the superbottle a t a temperature of 40" C. Comparative emulsify- ficial film of oil from a small particle but rather deflocculates ing powers were judged according to the appearance of the the oil-coated particle as a whole, so that the interface is one between oil and detergent solution. KO positive data emulsions after standing 16.5 hours. Chapin ( 2 ) used a somewhat different method to determine as t o the correctness of this assumption have been obtained the emulsifying power of stearate soap solutions. His pro- nor does an answer as to the correctness seem possible a t cedure was to add one cc. of mineral oil free from acid to this time. The data are presented because of the impor20 cc. of soap solution in a tube which vias rotated a t 20 to tance of the property measured and the correlation obtained 22 r. p. m. for 40 minutes. The contents were then filtered between such experiments and experimental washing operaand the quantity of oil found in the filtrate was taken as a tions to be reported. Denial of the correctness of the hy162
Februar?, 1933
I N D U S T R I A L A N D E N G I N E E €11 N G C H E $1 I S T R Y
pothesis would invnlidate much of the discussion of the mechanism of the operation of suspending soil as discussed here, but would affect the conclusions reached only to the extent of altering the terminology used. Emulsions of pure neutral oil are of theoretical interest, but, in dealing with soiled fabrics, the dirt contains materials of the nature of fatty acids. Soap formed by the soap builder a t the surface of the dirt particle is a n important factor in stabilizing the suspension. It is also desirable to use a method such that the results may be expressed in quantitative terms rather than by the appearance of the emulsions. Chapin's method with oiled carbon black is designed to meet the latter specification, but involves difficulties in technic which he recognized, and which introduce error into the results. A method has been developed for measuring a property of alkaline salt solutions which is a combined effect of deflocculation followed by emulsification-subject to one assumption. It is believed to be fairly representative of actual conditions and givw results which may be expressed yuantitatively.
163
residue to give an approximation of the weight of umber suspended in 25 cc. of solution. Such values were multiplied by four to give the weight of umber per 100 cc. of suspension, which was taken as a tentative expression of the emulsifying and deflocculating power of the detergent. This method possessed inherent defects. The builders would not be completely dehydrated a t 110" C. in all cases. Sodium hydroxide, and to a lesser extent others, may be expected to become carbonated. Kegative values were obtained for sodium oleate and for a mixture of modified soda and sodium oleate, n-hich showed that the calculated weight of dissolved substance did not represent the actual amount present in the residue. Presumably some of the soap, and possibly some of the alkaline salts, were removed from solution by adsorption on the oil-coated umber, a considerable portion of which had settled to the bottom. The actual concentration of solution when sampled would therefore be lower than that on which the calculations were based. These factors would cause the calculated values to be a t variance with the true values with all of the solutions in the series, and such variance would not be the same in all cases. As a more accurate method, the weight of umber susPROCEDURE FOR DETERMINISG EMULSIFI ISG ASD pended in 26 cc. of the detergent solution was determined DEFLOCCULATIKG POWER directly by a procedure similar to that used by Fall ( 3 ) in To ohtain droplets of uniformly distributed sizes, oil- determining quantities of manganese dioxide in suspension. coated burnt umber was used instead of oil alone. The In order to reduce the amount of soap or alkaline salt reumber was a finely powdered grade giving a mesh analysis moved by adsorption on the umber which settled to the of 0.37 per cent retained on 200 mesh and 2.90 per cent re- bottom, only one gram of umber per 100 cc. of solution was tained on 325 mesh. Since the size of the droplets depends used. The procedure was to withdran- 25 cc. from the layer on the size of the umber particles, the droplets in different of suspension as before. This was evaporated to dryness emulsions would be uniformly distributed as to size. The a t 110" C. The residue after ignition a t a dull red heat oil mixture used to impregnate the umber consisted of 49.5 was completely dissolved in sulfuric acid Metallic zinc per cent by volume of commercial cottonseed oil, 49.5 per was used as a reducing agent, and a titration was then carcent of neutral mineral oil, and one per cent c;f commercial ried out with standard potassium permanganate solution oleic acid. The cottonseed oil was a grade containing 0.38 for determination of the iron content. The ferric oxide conper cent of free fatty acid, and the oleic acid rvas a refined tent of the umber used r a s 46.77 per cent. K i t h this factor grade containing 95 t o 96 per cent of oleic acid. This mix- the umber present in the residue on evaporation could be ture corresponds roughly to the oils in dirt, although the calculated. acidity is probably low in comparison with 1hat actually Determinations according to this procedure were made present in average dirt. The oil mixture was applied to the after the umber had been allowed to settle for one hour and umber as a solution in benzene, the proportions used being for 24 hours a t 20" C. The results after standing for only one gram of oil mixture dissolved in 25 cc. of llenzene to 25 one hour were discarded because of inability to obtain satisgrams of the ground umber. The oil and umber were thor- factory checks. Presumably a uniform rate of sedimentaoughIy mixed to give uniform distribution of 1 he oil. The tion had not yet been reached a t that time under the experimixture was then dried until no odor of benzene could be mental conditions adopted. The results for the 24-hour detected. The resulting oiled umber contaiiied 1.98 per determinations are shown in Table I. cent of mineral oil, 1.98 per cent of qaponifiablc oil, and 0.04 I n order to determine the variance which existed between per cent of oleic acid. the calculated and actual residues from substances in soluThe fint procedure tried was t o add 5 gram> of the oiled tion and which Tvas assumed to be due to incomplete dehyumber to 100 cc. of 0.1 per cent solutions of the Poap builders dration and to carbonation of the more alkaline materials, to Le tested at 20" C:. Uniform 110-cc. oil bottles were used. residues on evaporation were determined for the various The mixtures were shaken vigorously by hand about twenty- solutions alone at the same concentrations as those used in five times and allowed to stand for 24 hours. At the end the tests. These are included in Table I. Wide variations of this period, comparisons were made of the depth of are shown to occur between the calculated and determined sediment in the bottom of the bottles, the height of the values for the residues obtained in evaporating solutions of layer of suspension above the sediment, and the weight of this series of substances. It was anticipated that some residue obtained a t 110" C. from 25 cc. of the solution carbonation would occur in the case of sodium hydroxide. pipetted from the center of the layer of suspension. Stability of the monohydrate v a s also anticipated. The Comparisons of the depth of the layers of sediment and of difference of an excess residue of 48.4 for sodium hydroxide suspension were unreliable as a basis for concluFions because as compared with a calculated value of 40 for the monothe differences were not large and checks were p' lor. -4com- hydrate indicates snme carbonation. Sodium metasilicate parison of this type would probably prove 1131x8 valuable n-ill carbonate slightly, but the excess is probably largely if long narrow tubes were used in place of 110-cc. oil bottles. due to incomplete dehydration. The value obtained is The residue obtained on evaporating 25 cc. of the sus- near that for a tetrahydrate. Trisodium phosphate conpension consisted of the wspended umber and of substances tains free sodium hydroxide in excess over the amount rewhich were present in solution. Calculations were made quired to completely neutralize the phosphoric acid, so that of the neight of anhydrous diesolved substance in 25 cc. of both carbonation and incomplete dehydration are factors the various solutions on the basis of the concentrations used. entering there The comparisons given in Table I show that, Thew n-eights were then deducted from the total n-eight of in the case of the solutions containing no soap, the portion
INDUSTRIAL AND ENGINEERING CHEMISTRY
164
Vol. 25, No 2
TABLEr. UMBERSUSPENDED AND RESIDUES ON EVAPORATIOX AFTER 24 HOURS (From shaking 1 gram of oiled umber a t 20' C. with 100 cc. of emulsifying solution)
TOTAL RESIDUE
O.1y0 SOLNS.
PER
100 CC.
SUBPENSION Me
.
158.4
111.4
Sodium oleate Sodium hydroxide 0.1% sodium oleate Sodium metasilicate 0.1% sodium oleate Trisodium hosphate 0.1% sodium oleate Sodium c a r i o n a t e 0.1% sodium oleate Modified soda 0.1% sodium oleate
++ + + +
87.4 126.8 106.4 93.0 281.2 201.4
178.0 226.2 203.2
of the residue which is material other than umber is approximately the same as the residue which would be obtained if umber were not present. The small differences shown may be assumed to be due chiefly to experimental error. When soap is present in the solution, however, either alone or in a mixture, the residue due to dissolved substances is considerably smaller when umber has been suspended in the solution than when it is absent. This bears out the hypothesis that adsorption of the soap by the oil-coated umber occurs. Adsorption of the alkaline materials may occur also but evidently to a much smaller extent.
DISCUSSION OF RESULTS The latter part of Table I shows that the adsorption of 45 mg. of soap per 100 cc. corresponds to the three builders with a high sodium-ion content-sodium hydroxide, sodium carbonate, and modified soda with equivalent weights of 40, 53, and 63.3. The adsorption of 60 mg. of soap per 100 cc. corresponds to the builders with lower total sodium-ion concentration-sodium metasilicate and trisodium phosphate with equivalent weights of 115 and 127. Although these are only approximations, the differences are great enough to indicate that the degree to which soap is adsorbed from the solution is directly related to repression of hydrolysis of the soap by sodium ion from the builder. Such an explanation is far too simple to be the whole explanation. The values for deflocculating and emulsifying power, as represented by the weight of umber suspended per 100 cc., show a distinction between colloidal and noncolloidal materials. The materials possessing known colloidal properties-sodium oleate and sodium metasilicate-are better agents than the noncolloidal materials. The properties of trisodium phosphate correspond with those of a colloidal builder in this case and have been found to do so in other cases. This is true, irrespective of the degree of alkalinity of the solutions of these various substances. I n order to illustrate this, the materials have been listed in all tables in decreasing order according t o their degree of alkalinity as previously determined (4). Of the series, sodium hydroxide, which gives the highest pH, has comparatively low deflocculating power and therefore shows a low value for the combined deflocculating and emulsifying power. Sodium carbonate has a considerably higher pH in solution than modified soda but both give the same low values. Sodium oleate has a low pH and the highest deflocculating and emulsifying power. Another distinguishing feature of these determinations is the additive effect shown by mixtures of soap and alkali. Where the pH of the solution is a determining factor in tests of a detergent property of an alkali or alkaline salt, as in interfacial tension determinations, substances m-hich give a p H lower than that of a soap solution increase the value of the soap very little when added to it and may decrease it. I n suspension or emulsification, however, the effect of the builder is apparently added t o that of soap. This is
UMBER PER 100 cc. Due t o materials SUBPENBION in soln. b y difference I
Me.
18.9
139.5 79.2
32.2
Detd. on soln. alone Milligrams per 100 cc. 148.4
83.8
62.1
25.3 12.2
66.0 108.8 93.8 105.8 257.2 189.6 169.6 205.2 192.4
114.6 94.2 44.4 214.8 124.4 104.6
12.2
48.6 66.4 77.0
73.4
159.8 145.6
66.4 57.6
-
RESIDUE
Difference
-8.9
-4.6 +5.8 -3.9
+0.4
-61.4 -42.4 -65.2
-65.0
-45.4 -46.8
demonstrated in Table 11. It follows that in these tests the importance of wetting power is small as compared with suspending or emulsifying power. Even those salts which do not raise the pH of a soap solution, as represented by modified soda, are capable of raising the deflocculating and emulsifying power of a soap solution by the extent of the effect of the salt solution alone. This is partially due to the soap builder reacting with the fatty acid of the coating on the umber particles. That this is not the entire explanation is indicated by the fact that the deflocculating and emulsifying values found for the alkaline salts do not parallel their pH or CORvalues as previously determined. Soap serves as a deflocculating agent and protective colloid, and therefore stabilizes the suspension of oiled umber particles. 11. COMPARISON O F DEFLOCCULATIXQ ASD EMULSIFYINQ POWER OF ALKALINESALTAND SOAPSOLUTIONS AT 20' C.
T.4BLE
UMBER SC6PENDED Sum of values for Values d e t d . alkalies alone for alkalies and 4 8 . 6 , value for sodium oleate sodium oleate alone together Mzliigrams per 100 cc. r
+
0.1%
SOLNB.
Sodium hydroxide Sodium metasilicate Trisodium phosphate Sodium carbonate Modified soda
67.5 80.8
73.9 60.8 60.8
66.4
77.0 73.4 66.4 57.6
Difference -1.1
-3.8
-0.5 +5.6
-3.2
Considering the deflocculating and emulsifying values of these individual materials, the highest is sodium oleate. Sodium metasilicate follows next, with trisodium phosphate third. The effects of blended materials need not be classified other than to say that they are the sums of the effects of the individual materials used. The effect of soap alone is not far greater than that of metasilicate alone.
DISODIUM PHOSPHATE Disodium phosphate is not included in the tables and general discussion because it is not an alkaline salt of the type being considered. Under the conditions of Table I, in the absence of soap, where trisodium phosphate gives a value of 25.3, disodium phosphate gives a value of 28.8 as compared with 12.2 given by sodium carbonate and by modified soda. This is further confirmation of the apparent colloidal behavior of alkaline phosphates and is entirely out of line with the COHvalues. I n the presence of soap, effect of the relatively low CORof the disodium phosphate shows in the value of 57.5, corresponding with 57.6 for modified soda rather than with i3.4 for its more alkaline homolog, trisodium phosphate, The pH value is so markedly below the neutral point for soap that the deflocculating and emulsifying power does not correspond to the rule stated for true builders as brought out in Table 11. The sum of the individual values for disodium phosphate and soap is 77.4, but the mixture of disodium phosphate and soap gives a value 19.9 units lower, because of the detrimental effect of the low COB of disodium phosphate on the soap. Therefore,
February, 1933
IXDUSTRIAL AND ENGINEERING
while disodium phosphate has merit as an emulsifying agent below p H 10.2, in the presence of soap, trisodium phosphate rather than disodium phosphate should be used.
( 3 Snek Ib&
t:;
~
165
CHEMISTRY
24, 1051-7 (1932).
~
~
~
~ (1927), ~
~
RECEIVED August 2, 1932. Presented before the Division o€ Colloid
LITERATURE CITED
Chemistry a t the 84th Meeting of the American Chemical Society, Denver, Colo., August 22 t o 26. 1932. Publirrhed b y permission of The Provident Chemical Works, a subsidiary of The Swann Corporation. The experimental work was carried out b y Elbert L. Jung a n d t h e results prepared for publication b y Beatrice F. Grey
(1) Baker, IND.EBG.CHEV.,23, 1026-32 (1931). (2) Chapin, 0 2 1 &. Fat I n d . , 4, 15-21 (1927). (3) Fall, J. Phys. Chem., 31, 801-49 (1927). (4) Snell, ISD.EXG.C H E V, 24, 76-80 (1932).
Purification of Brine by Ammoniation w. c. HSIEH,E. 0. WILSON, Yenching University, Peiping, China, , ~ N DT. P. Hou, Pacific, Alkali Company, Tangku, China T T H E Pacific Alkali Company a t Tangku, near
A
Tientsin, China, sodium carbonate is manufactured by the Solvay process from crude salt obtained by the solar evaporation of sea brine. The salt used contains a higher percentage of magnesium than the rock salt or natural brines that are largely used as raw materials for the manufacture of soda ash in other parts of the world. I n order to produce a product of sufficient purity and to avoid clogging the carbonating towers, it is essential that the magnesium be largely removed early in the process. This paper presents the results of a study of the factors influencing the precipitation of magnesium by ammoniation. Recovered gas containing ammonia and some carbon dioxide from the filters, carbonating towers, ammonia stills, etc., meets the crude brine and serves to introduce the required amount of ammonia into the brine, a t the same time precipitating the calcium, magnesium, and oxides of iron and aluminum. TElis process avoids the addition of any special precipitating agents, such as lime or soda ash, and under proper conditions efficiently removes the calcium and magnesium. The chief defect of the method is the coprecipitation of sodium carbonate and sodium chloride which results in a considerable loss of alkali. Further work on this phase of the problem is now in progress.
gas through the solution. To obtain the higher concentrations of ammonia, it was found necessary to cool the brine after a preliminary heating to 60-65" C. to cause coagulation. Samples were withdrawn through a rubber-tipped pipet, the rubber tip being covered with filter paper. Total ammonia was determined in the filtered samples by the Kjeldahl method, and magnesium by the gravimetric method, weighing as magnesium pyrophosphate. The results are shown in Figure 1 and the following table: EXPT. T O T A L37% Grams/Ziter 1 10.49 2 14.49 3 28.59 4 41.02
JIg Mg./lite? 511.9 328.9 160.7 128.8
EXPT. TOTALNH3 M g Grams/laler M g . / h t e r 5 65.30 114.0 6 81.36 111.5 7 88.78 109.9
The magnesium remaining in solution in the brine decreases as the concentration of total ammonia increases, the optimum concentration being between 20 and 40 grams per
EFFECTOF COKCENTRATJON OF ANMOKIA ON PRECIPITATION
OF
MAGNESIUM
The sea salt, obtained by solar evaporation, is collected in large piles. On standing through the rainy season, a large amount of the more soluble magnesium compounds is leached out. For this experiment a sample of freshly harvested salt was used as representing the worst possible conditions. Analyses of samples of new and one-year-old salt are as follows: OLD SALT KE% SALT
% Cas04 MgSOi MgClz NaCl
1.17 0.12
g;:$;
OLD
72 1.I02
s!:g0; . 8 1
Insol. matter Moisture and combined HzO
SILT. % 0 58
s 4 L T ' NET5
% 0.64 5.34
10 43
A solution of the salt of specific gravity 1.2 was made, using distilled water. Insoluble impurities wire removed by filtration. Sufficlent ammonium carbonate was added to precipitate all of the calcium in the brine and leave an excess; then the magnesium was precipitated by bubbling dry ammonia gas through the solution. The temperature of the brine during ammoniation was kept between 60" and 65" C. by means of a water bath. The temperature and pressure of the gas were keot constant. No agitation of the brine was provided except that furnished b y t h e bubbling of the
Total NH3 in AmmoniatedBrinc,gram//iter
FIGURE 1. AMMONIA AND MAGNESIUM IN AMMONIATED BRINE
liter. It is probable that the precipitate formed is a basic carbonate, and the solubility is certainly influenced by the presence of sodium chloride as well as the ammonia.
EFFECT O F TEMPERATURE O S SETTLIKG RliTE O F CALCIUMASD MAGKESIUM PRECIPITATES Ammonia gas was bubbled through a solution of the brine, made as above and containing an excess of ammonium carbonate, until the concentration- of total ammonia was
~
~
