Distribution of Nicotine between Water and Petroleum Oils L. B. NORTON New York State Agricultural Experiment Station, Geneva, N. Y .
Nicotine is nearly equally distributed between petroleum oils and water at low concentrations; the distribution in favor of the water increases with intermediate concentrations, reaches a maximum, and decreases at high concentration. Acids extract the nicotine completely into the water phase when present in excess and hold an equivalent amount in the water when in smaller amounts; the excess free base is thus left equally distributed between the phases. Alkalies have no appreciable effect in concentrations less than 0.1 normal. In higher amounts they drive the
nicotine into the oil phase, the extraction into the oil being complete with 5 N sodium hydroxide. It is concluded that only that portion of the nicotine which is in the free state will be shared with the oil; that the nicotine will be distributed in approximately equal concentrations in the oil and the water in the original spray mixture, regardless of the concentration of the nicotine and of the presence of fungicides and emulsifiers; and that some of the nicotine originally in the water will be transferred to the oil if the spray dries on the foliage.
T
HE distribution of nicotine between water and oils is of
at the higher concentratiohs. Cuvilier ( I ) determined the great practical importance in the preparation and use relative efficiency of several salts for reducing the solubility of insecticidal sprays. The behavior of nicotine-oil of nicotine in water, and concluded that sufficient solvent was emulsion sprays depends t o a great extent on the relative bound by hydration of the salt ions to account for the loss in amounts of nicotine in the two phases. Some of the water in solubility of the nicotine. His assumption seems justified, the spray mixture runs off immediately after application, and since the effectiveness of the salts is approximately in the the remainder evaporates in a relatively short time, while the usual order of the anions in the lyotropic series. It has long greater part of the oil remains on the foliage and serves to been known t h a t strong alkalies will reduce the solubility hold the nicotine in solution and to retard its loss by evaporaof nicotine, probably for the same reason, and thus make tion. A distribution of nicotine such that a greater proportion extraction by organic solvents practically complete. Acids, of i t is dissolved in the oil phase will therefore tend to extend on the other hand, tend to extract the basic nicotine out of the period during which it is effective. This distribution may organic solvents into the water because of the formation of be influenced by a number of factors, such as the type of oil, water-soluble but oil-insoluble salts. the concentration of nicotine, and the presence of other Experimental Procedure materials in the mixture as insecticides, fungicides, and emulsifying agents. The present work was undertaken t o determine The oils used for the measurements were a refined kerosene the quantitative effect of some of these factors. and a light lubricating oil having the following characteristics Little quantitative information is available in the literaas furnished by the manufacturers: ture on the distribution of nicotine between water and petroleum oils. DeKerosenea Lubricating Oil& Ong ( 2 ) found that one per cent of nico0.775-0.785 0.554 Specifio gravity at 600 F. 72 30-36 Saybolt Universal viscosity at 100' F., aec. tine dissolved in five different oils, rangUnsulfonated residue, % 97 or higher 94.9 ing from kerosene to light lubricating 6 . 5 % off a t 636' F. (336' C.) Distillation range 370-390' t o 480-500' F. (188-199O t o 249-260O C.) 10% off a t 641' F . (338' C.) oil, is nearly equally distributed be20% off a t 650' F. (343' C.) 30v off a t 656' F. (347' C.) tween these oils and equal volumes of off a t 664" F (351" C ) water after 5 minutes of shaking. off a t 671' F: (355' C : ) off a t 679' F. (359' C.) Ritcher and Calfee (8)state that about off a t 687' F. (364' C.) off a t 699' F. (371' C.) 99 per cent of the nicotine is extracted off a t 716' F. (380' C.) from the 50 and 95 per cent a "Deobase", produced by L. Sonneborn Sons, Ino. An experimental material furnished by t h e Shell Petroleum Corporation and refined by them from free nicotine solutions by sufficient oil mid-continent crude* to form a 1 Der cent solution of nicotine in the oil. KolosovskiI and Kulikov (4) The nicotine (a 95 per cent free alkaloid, furnished by the determined the distribution between water and a variety of Tobacco By-products and Chemical Corporation) was purified as the zinc chloride double salt according to the method of Ratz organic solvents, including petroleum ether, ligroin, and some (7). The resulting nicotine was water-white, boiled at 123cyclic hydrocarbons. They found a distribution in favor of 125' C. at 17 mm. pressure, and had an optical rotation a? of the organic solvent at low concentrations, increasing in favor -168.55". It was stored in the dark under nitrogen. It was of the water u p t o about 50 per cent, then decreasing again soluble in all proportions in the oils. 241
INDUSTRIAL AND ENGINEERING CHEMISTRY
242
TABLE I. DISTRIBUTION OF NICOTINE BETWEEN WATERAND PETROLECY OILS Concn. in Oil, C
Distribu-
0.253 0,605 1.24 2.44 7.37 16.3 25.2 50.0
0.264 0.637 1.31 2.55 8.16 17.9 31.2 68.1
0.522 1.22 4.95 11.5 22.1 46.6
0.565 1.30 5.25 12.6 25.6 61.3
Between Water a n d Kerosene 0.479 1.18 53.5 1.14 1.14 94.0 4.63 1.13 164 10.4 1.22 273 18.6 1.38 409 31.8 1.93 531
Concn. r
Concn. in Water, C Grams/liter
Distribution Ratio, C/C'
Concn. in Concn. in TTater, C Oil, C 7 Grams/literBetween Water and Lubricating Oil 0.241 1.10 76.3 112.1 40.4 0.572 1.11 76.5 116.6 36.4 1.16 1.13 2.32 1.10 11 82 26 .. 18 32 10 16 .. 71 6.57 1.24 281 492 70.5 14.7 1.22 408 709 107 19.1 1.63 541 867 214 31.9 2.13 594 877 311
AV.
AV. Concn.
2;:
69.7 130 267 469 702 834
i;:: 61.4 77.0 115 227
F%?,: C/C
2.77 3.20
:::: 6.98 6.63 4.05 2.82
i::; 4.35 6.09 6.10 3.68
VOL. 32, NO. 2
oil and over six times that in the kerosene. The curves obtained by Kolosovskii and Kulikov (4) for other hydrocarbons have a similar form and show a maximum a t about the same concentration. It is interesting that solutions of nicotine and water in approximately the same concentration range also show maximum densities and viscosities and a maximum rate of change of the specific rotation (3, 9). These phenomena are explained by the formation of hydrates of nicotine, which would also account for the distribution maximum,
Effect of -4cids
Solutions of nicotine containing 1.25 grams per liter were titrated with 0.1 N hydrochloric, sulfuric, and acetic acids, in a Beckman p~ meter and glass electrode. ~h~ titration curves (Figure 2) Show t h a t nicotine a t these concentrations behaves as a monoacid base. The curves for hydrochloric and sulfuric acids are identical, \\-bile that for the weaker acetic acid shows a less pronounced break, The equivalent point with the strong acids corresponds to a P H of 5.5, and with acetic acid, to about 6.25. The primary dissociation constant calculated from the point of 50 per cent neutralization is 1 10-6,in fair agreement Tvitll the values of 4,90 10- determined by Lowry and Lloyd (6) with the hydrogen electrode, and of 7 X lop7 obtained by Kolthoff (6) with indicators, both in more concentrated solutions. The same workers found the second ionization constant to be 7,94 and 1.4 X respectively, which is almost negligible in comparison with the first. On the basis of these values, the nicotine monosulfate and monochloride will be only slightly at the point, and the acetate be about 2*5Per cent hydrolyzed. The determinations of the distribution in the presence of of the acids were made with a total nicotine 1.25 grams per liter. The equivalent amount of acid is 3.10 cc. Only the lubricating oil was used, since the general behavior of the kerosene is so nearly identical. The effect of different amounts of the three acids is shown in Table 11. The calculated values in the last column of Table I1 were obtained on the assumption that only the free portion of the 1
The distribution measurements were made with equal volumes of water and oil solutions. Twenty cubic centimeters of the oil were measured into glass-stoppered tubes of about 5 0 - ~capacity. ~. The nicotine was added either in water solution or as the free alkaloid, depending on the amount required, other substances were put in if desired, and sufficient distilled water was added to bring the volume of the water layer up to 20 cc. The sto per8 were sealed with rubber caps, and the samples shaken mecxanically in a water bath maintained at 25' =t0.1' C. Although the results of DeOng indicate that approximate equilibrium is established in a very short time, it was found convenient to leave the samples in the bath overnight. A series of tests showed that this allowed ample time for the establishment of complete equilibrium. After the period of shaking, the samples were allowed to stand in the bath until the two layers were completely separated. In a few cases the mixtures had to be centrifuged t o separate the layers. Portions for analysis were then removed from each layer with a pipet. The nicotine was determined by precipitation with silicotun stic acid and ignition of the precipitate and weighing of the resizue. The nicotine in the water solutions was precipitated directly after acidification with hydrochloric acid, The oil samples were extracted with several portions of hydrochloric acid in a separatory funnel, and the nicotine was precipitated from the combined extracts. A few of the more concentrated solutions were analyzed by titration with standard hydrochloric acid and with methyl red as indicator.
Effect of Concentration
The distribution was determined for concentrations of nicotine in the mixture as a whole, ranging from 0.01 to over 500 grams per liter. The data for the lubricating oil and the kerosene are shown in 7 Table I. The data for both oils are plotted in 6 Figure 1. The distribution ratio, C/C', is plotted against the logarithm of the concentration rather than the value itself in order to show in greater detail the behavior at low concentrations. The concentration in the water layer, C, was chosen 3 for plotting because the large amount of water in spray mixtures in comparison with the oil 2 makes this value much closer to the concentraI tion of the mixture as a whole than is the average concentration. T h e distribution is constant and practically o -I 0 I 2 3 identical for the two oils up t o a concentration of LOG C about 10 grams per liter, or 1per cent. ThroughOF C O i Y C E x T R A T I O ~O N DISTRIBCTION OF NICOTINE FIGURE 1. EFFECT out this range the concentration in the water is 1.11times that in the oil. The ratio then increases with the concentration, reaches a maximum between 50 and nicotine would be distributed between the oil and the water, and that all of the nicotine equivalent t o the acid would be 70 per cent nicotine, and decreases again at the highest held completely in the water layer. The agreement between concentrations. The greatest difference between the two oils occurs at the maximum point, where the concentration the observed and the calculated values indicates that the assumption is correct. A small excess of the acids extracts in the water is about seven times that in the lubricating
E,
INDUSTRIAL AND ENGINEERING CHEMISTRY
FEBRUARY, 1940
243
TABLE 11. EFFECTOF ACIDSON CC. 0.1 N Acid
-HCIaC
C'
C/C'
1.16 1.11 1.49 1.05 0.77 2.16 0.64 2 95 0.42 4.86 0.29 7.28 0.16 14.6 0.04 61 0.01 >zoo m 0.00 nicotine in water;
0.00 1.29 1.56 0.50 1.66 1.00 1.89 1.50 2.04 2.00 2.11 2.30 2.33 2.70 2.43 3.00 2.48 3.10 2.48 4.00 a C is concentration of
THE DISTRIBUTION OF NICOTIXE BETWEEN WATERAND LUBRICATING OIL ---HzSOtHACCalcd.--C' C C C' C/C' C C' C/Cl 1.31 1.18 1.11 1.31 1.18 1.16 1.11 1.29 1 . 5 1 0 . 99 1.33 1.41 1.06 0.96 1.42 1.48 0.80 1.70 1.96 1.65 0.84 2.20 0.75 1.65 1 . 8 9 0 . 61 2.86 1.83 0.64 3.30 0.57 1.88 0.42 2.08 1.96 4.36 0.45 0.40 2.07 5.18 2 . 1 9 0 . 3 1 6 . 8 1 0 . 3 2 2 . 1 8 7 . 5 2 0.29 2.18 2.35 0.15 14.6 2.34 0.16 0.11 2.31 21.0 2.46 0.04 49 2.44 0.05 0.03 2.40 80 0.00 2.50 0.02 > 100 2.46 >zoo 2.43 0.01 m m 2.50 0.00 2.48 0.00 0.00 2.45 C' is concentration of nicotine in oil, in grams per liter; C/C' is distribution ratio.
---
C/Cl
1.11 1.53 2.12 3.10 4 95 7 06 15.7 02 m
m
the nicotine completely into the water. The acetic acid gives slightly lower values and requires a slightly larger excess for complete extraction, because of the greater hydrolysis of the salt. OF TABLE111. EFFECTOF BASESON THE DISTRIBUTION hTICOTIx'E BETWEEN WATER AND LUBRICATING OIL
Molar Concn. of Base
0.00 0.001 0.0125 0.025 0.0375 0.05
0.125 0.15 0.25 0.375 0.5 1.0 1.53 2.5 3.06 4.00 4.59 6.12 10.0 0.00 0.0066 0,0112 0.0145 0.0167 0.0201 0.0223 (satd.)
Concn. in Conon. in Water, C Oil, C' --Grama/ZiterEffect of Sodium Hydroxide 1.81 9,33 1.18 1.14 1.32 10.99 1.06 1.22 11.97 1.30 1.44 12.30 1.19 12.43 1.37 1.24 12.45 1.37 1.23 13.05 1.27 13.13 1.25 1.24 1.20 13.33 1.09 1.43 13.49 1.13 0.98 1.38 13.60 13.86 0.651 1.72 14.00 0.524 1.99 14.14 2.12 0.172 2.22 14.18 0.056 14.25 2.18 0.023 14.27 2.40 0.019 14.28 2.32 0.020 13.97 2.55 0.038 Effect of Calcium Hydroxide 9.33 1.32 1.16 11.96 1.44 1.27 12.23 1.75 1.60 12.32 1.31 1.15 12.38 1.50 1.35 12.45 1.33 1.19 12.50 1.18 1.00 PH
DiRtribution Ratio, C/C'
1.11 1.16 1.15 1.11 1.15 1.10 1.03 0.99 0.91 0.79 0.71 0.38 0.26 0.08 0.03 0.01 0.01 0.01 0.01
I
21 0
1
1
' I 3 4 5 CC. 0.1 N ACID TITRATION OF NICOTINE WITH
2
6
7
ACIDS
Discussion
1.14 1.13 1.09 1.14 1.11 1.14 1.18
The distribution of nicotine in the presence of sodium hydroxide and calcium hydroxide mas determined in solutions having a total concentration of approximately 1.25 grams per liter, the same as that used with the acids. Again only the lubricating oil was used. The concentration of sodium hydroxide was varied from 0.001 molar, or 0.004 per cent, to 10 molar, or about 40 per cent; and t h a t of calcium hydroxide from 0.006 molar to saturation, or 0.0223 molar. The results are shown in Table 111. The pH values were measured with the glass ;
5
'
'
I
FIGURE 2.
Effect of Bases
concentrated solutions. The data are plotted in Figure 3, where the logarithm of the concentration is again used in order to shon more detail in the lower concentrations. The effect of the bases on the distribution of nicotine is negligible in concentrations less than 0.1 molar, which includes all of the calcium hydroxide solutions. Increasing the concentration of sodium hydroxide beyond this point pushes the
1
I n applying distribution studies to the behavior of actual spray mixtures, it should be emphasized that the distribution ratio is a ratio of concentrations and not of total quantities. For example, if the distribution ratio is 1, indicating equal concentrations in the two phases, an emulsion containing 1 per cent of oil will have only 1 per cent of its total nicotine content in the oil. The ratio must be reduced to 0.01 to bring about half of the nicotine into the oil. The distribution ratio must therefore be made very small to concentrate most of the nicotine in the oil in a dilute emulsion. It is to be expected from the basic nature of gicotine t h a t acids would have a more pronounced effect on it than any other class of compounds. The presence of even a slight excess of a moderately strong water-soluble acid will prevent any of the nicotine from dissolving in the oil (Table 11). lu'icotine sulfate solutions, in the absence of any alkaline material, mill therefore remain completely in the water phase. Only t h a t part of the nicotine in a spray mixture which is added as free nicotine, or which has been set free from the salt by the addition of some basic substance, will be shared with the oil. Free nicotine and nicotine sulfate can be used interchangeably t o give exactly the same effect with oil emulsions I
.'
7.
0
I +
__.-
I
I I
I
o,5.
7
1
I
I
d
?
--
I ,
--_~ I
I ___-
-~
_-I
__ -
__ __
:I,,__-
, 1
1
L L -2 0
-3
-1
INDUSTRIAL AND ENGINEERING CHEMISTRY
244
only when the mixture is sufficiently alkaline to react completely with the nicotine salt. A number of oil-soluble nicotine compounds have been proposed, particularly salts of nicotine with high-molecularweight organic acids, in the hope that the nicotine would be kept in the oil layer. Some preliminary experiments have indicated t h a t although most of these acids are insoluble in water, they form salts which are somewhat soluble in water and which, in addition yield some free nicotine to the water by hydrolysis. They have much less effect than might be expected in keeping the nicotine in the oil, and some of them actually favor the water more than free nicotine. T o ensure holding the nicotine in the oil, it is not sufficient to find an oil-soluble compound; the compound must also be insoluble in water and relatively slightly hydrolyzed. The influence of concentration and of constituents of the spray mixture other than acids on the distribution of nicotine is not so great as has generally been assumed. Under practically all ordinary conditions the spray as applied will contain nearly equal concentrations of nicotine in the oil and the water. The ratio will vary little with the type of oil and will be practically independent of the concentration, since the largest amount used in practice is well within the range of constant distribution (Table I). It will not be seriously affected by alkaline fungicides, since saturated lime and sodium hydroxide up to 0.1 molar have practically no effect (Table 111). Emulsifying agents should not produce much change. I n the concentrations which are used, neither the substances themselves nor their possible hydrolysis products are present in sufficient quantity to affect the solubility of the nicotine appreciably. I n view of the rapid transfer of nicotine between the oil and the water, no advantage will be gained by dissolving the nicotine in the oil first. I n general, then, practically any nicotine-petroleum oil spray will have approximately equal concentrations of nicotine in the two phases a t the time of application. The situation may be somewhat changed after the spray has been on the foliage for a time. Any material running off
VOL. 32, NO. 2
will cause the same loss of nicotine whether it consists of the unbroken emulsion or only of the water phase, because of the equal distribution. I n the remainder of the spray which remains on the foliage, the water will tend to evaporate more rapidly than the dissolved nicotine, so that the solution will become more concentrated. This increase in concentration will cause some of the nicotine to be transferred from the water into the oil with which it is in contact. After the concentration has exceeded one per cent, the distribution in favor of the water increases, so that further increases in the concentration of nicotine in the water phase will cause proportionally smaller increases in the oil phase. Eventually, however, all of the nicotine which was originally in the water and which has not been lost because of run-off, evaporation, or other agency, will be dissolved in the oil. The effect of rain on the residues will be a reversal of these changes. The extraction of nicotine will be rapid a t first, because of the distribution in favor of the water, as well as the relatively high concentration, and will become progressively less, until from solutions containing less than one per cent, half of the remaining nicotine can theoretically be extracted by each equal volume of water. I n actual practice the extraction will be much less, because most of the water will run off before it has had time to dissolve the maximum possible amount of nicotine from the oil.
Literature Cited Cuvilier, B. V. J., 2. anal. Chem., 105, 325 (1936). DeOng, E. R., IND. EX+. CHEM.,20, 826 (1928). Jephcott, H . , J. Chem. Soc., 115, 104 (1919). Kolosovskii, N. A., and Kulikov, F. S.,Acta Univ. Asiae Mediae (Tashkent), VI, No. 8 (1935). Kolthoff, I. M., Biochem. Z., 162, 289 (1925). Lowry, T . M., and Lloyd, W. V., J . Chem. SOC.,1932, 1626. Ratz, F.,Monatsh, 26, 1241 (1905)., Ritcher, P. O., and Calfee, R . K., Ky. 4gr. Expt. Sta., Bull. 370, 47 (1937).
Tsakalotos, D. E., Compt. rend., 148, 1324 (1909). PBEBENTED before the Division of Agrioultural and Food Chemistry at the 97th Meeting of the American Chemical Society, Baltimore, Md. Approved by the Director of the New York State Agricultural Experiment Station for publication an Journal Paper No. 316.
POTASSIUM METAPHOSPHATE A Potential High-Analysis Fertilizer Material S . L. MADORSKY
AND K.
G . CLARK
OLYMERIC modifications of the alkali metaphosphates ranging from mono- to decameric forms have been described in the literature (4, 5, 15-18, 26). The sodium salts-Maddrell’s salt (an insoluble crystalline trimer), Knorre’s salt (a soluble crystalline trimer), Graham’s salt (a very soluble amorphous hexamer), and Kurrol’s salt (an insoluble octamer)-appear to have received more attention than potassium analogs. Graham’s salt in particular has been publicized in connection with recent developments in the treatment of water for industrial purposes (7-10,19). Soluble, insoluble, crystalline, and amorphous forms of the potassium salts have been reported (15, 16). Water-insoluble potassium metaphosphate may be pre-
P
Bureau of Agricultural Chemistry and Engineering, U. S. Department of Agriculture, Washington, D. C.
pared by a number of reactions, including ( a ) dehydration of monopotassium phosphate or dipotassium pyrophosphate, (6) neutralization of metaphosphoric acid with potassium hydroxide or potassium carbonate followed by dehydration of the product, and ( c ) reaction of potassium chloride with phosphoric acids. Such an insoluble product, composed entirely of the plant foods, potassium oxide (K20) and phosphoric if capable of releasing the potassium and phosoxide (P20s), phorus to growing plants under soil conditions, should possess many of the properties desirable in a high-analysis fertilizer material. Bartholomew and Jacob (3) in a series of pot tests found the fertilizer efficiency of the phosphorus content of water-insoluble potassium metaphosphate, prepared