ADSORPTION BY FILTER PAPER BY M . A . GORDON
In a paper on capillary analysis, Bayleyl says: “Sometime ago, while holding a wet patch, formed by dropping a solution of silver nitrate upon filter paper, in a stream of hydrogen mixed with arseniuretted hydrogen, I noticed that the metal was contained in the centre of the blot, and that the edges for about half a centimeter inwards were entirely free from metal. After being exposed to the gas, the blot presented the appearance of a black spot surrounded by a broad ring of water. . , . Drops of other metallic solutions were placed upon filter paper and submitted to the action of sulphuretted hydrogen. I found that in some cases the metal extends to the edge of the spot and even seems concentrated there, while in ‘others a water-ring surrounds the patch of sulphide. I’urther inquiry demonstrated that solutions of the same metal present the first or second of these phenomena, according as they are concentrated or dilute. Taking a strong solution of copper sulphate and diluting portions of it, I found that at one degree of solution the metal spreads outwards just as far as the water, and that with solutions more dilute than this, the water is separated from the metal, which remains in the centre. The more dilute the solution, the broader is the external water-ring. The exact strength of solution to give the‘ former appearance varies with the temperature, and with the kind of paper used ; the metal in a warm solution is more mobile than if the. solution were cold. “The blot formed by a drop is larger on the whole, but the mobility of the metal is increased in a greater proportion than that of the water. A close Swedish paper is more efficacious in separating the salt than a loose-textured piece of common filter paper. “ A great difference is found to exist among the salts Jour. Chem. SOC.,33, 304 (1878)
338
’
M . A. Gordovt
of various metals; the salts of silver, lead, and the persalts of mercury, when moderately concentrated, give a wide waterring, while the salts of copper, nickel, and cobalt must be much more dilute to present the same appearance. Cadmium seems especially able to pass through filter paper. “ A solution of copper sulphate containing 0.001gram of copper per cc was found to give a water-ring; at the same temperature and with the same piece of Swedish paper, a solution of cadmium sulphate of less than half that strength gave a blot in which the cadmium extended perfectly to the edge. Metals appear to act in this respect as though other metals were absent; this property of cadmium, therefore, affords an elegant means of detecting it in the presence of metals, the sulphides of which are black. The considerably dilute solution is dropped upon filter paper, and the blot allowed to extend as far as possible (this should be done in all cases) and the sulphuretted hydrogen then turned on. The black patch is found to be surrounded by a vivid yellow ring of cadmium sulphide. “ A solution rich in nickel, cobalt, or iron may in like manner be examined for these metals, in presence of smaller quantities of copper, lead, mercury, and silver. The blot spreads, is exposed to sulphuretted hydrogen, and afterwards held over a bottle of ammonium sulphide, when the waterring becomes black. This method, however, is extremely suited to the detection of cadmium. It was found that the presence of free acid much increases the mobility of copper, so that before testing for cadmium in this manner, the solution, if acid, should be made slightly alkaline by ammonia.” It seemed desirable to find out something about the conditions under which cadmium could be detected in the presence of copper by means of this diffusion method. Instead of following Bayley’s work exactly, experiments were made with strips of filter paper dipping into small volumes of the solution. A rather heavy, white, blotting paper was used and was cut into strips 2.5 cm wide. The first set of experiments was made to show the relative rates of diffusion
Adsorption by Filter Paper
339
of equimolecular solutions of cadmium sulphate and of copper sulphate. A known volume of the solution was allowed t o diffuse up the strip for a certain length of time. The height to which the solute rose was determined by placing the strip in a bottle containing a little ammonium sulphide. The data are given in Table I and are shown graphically in Fig. I .
TABLE I Volume of solution, I
1
Cadmium sulphate
I
1
j
I
1-
0.025 -
I
1
cc; time, 30 minutes
Rise in filter paper, cm
I
Molar conc.
I
HzO ~ _ _ _ _
Cd
Copper sulphate - ~ _ _
HzO
1
-
~~~
9.1
6.0
9.2 9.2
7.4
_
Cu
8.6
5.3
8.7
6.5
I
0.05
-
1 I
i
I
0. I
-
9.2 9.5
I ~
I
8.6 8.7
9.0 9.0
1 I
7.7 7.6
I
It will be seen from these data that cadmium sulphate diffuses farther than copper sulphate, except of course for concentrations a t which copper sulphate gives no water-ring under the conditions of the experiment. At concentrations of 0 . 3 molar and higher the cadmium sulphate moves as fast as the water and copper sdlphate does the same a t concentrations of 0.4 molar and higher.. At lower concentrations the cadmium sulphate diffuses distinctly faster than the copper sulphate and should therefore be detected in theouter
M . A. Gordon
3 40
ring in case the difference in rate is not diminished in a mixed solution. Bayley’s concentration of one milligram of copper per cubic centimeter or one gram per liter corresponds to a little less than 0 . 0 2 molar. On the other hand, these experiments show that if the relative rates remain the same, one could not detect cadmium in case the molar concentration of the copper sulphate was approximately double that of the cadmium sulphate or a t higher concentrations than about 0.4 molar. Bayley’s own experiments show that the relative rates of diffusion in mixed solutions are not necessarily the
Fig. I Relative Diffusion of CuSO4 and CdSOd Volume of Solution = I cc Time of Diffusion = 30 minutes
c
same‘ as the ratio of the rates of the substances taken separately, because he found that “the presence of free acid much increases the mobility of copper.” As it was no part of the plan .to make an exhaustive study of these phenomena, the next set of experiments was made with the more concentrated solution prepared by mixing equal volumes of I molar copper sulphate and I molar cadmium sulphate. Apart from contraction or expansion the resulting solutipn was half-molar with respect to each salt and might reasonably have been expected not to show any 1
Cf. also Freundlich: Kapillarchemie, 164 (1909).
Adsorption by Filter Paper
341
separation of the salts. A strip of blotting paper was allowed to stand with one end in the solution until the liquid had risen to a height of 6 or 7 cm. It was then placed in a bottle containing ammonium sulphide. The strip was colored a dark brown by the copper sulphide as far up as the liquid had diffused. The yellow of the cadmium sulphide could not be seen. If the strip was now allowed to stand in the liquid again or in water until the liquid had diffused 0 . 2 to 0.4 cm farther and was then placed in the bottle again, the yellow cadmium sulphide became visible over the stretch through which the solution had just diffused. This presented some theoretical difficulties. If the salts had been precipitated completely by the ammonium sulphide there should have been a water-ring fa-med. If they had not been precipitated completely, why should there be only cadmium sulphide in the newly wetted zone? The most plausible explanation is that we are dealing with a case of fractional precipitation and solution. It is known that copper sulphide is precipitated before cadmium sulphide. In these experiments only the salts in the outer layer of the filter paper are precipitated. When this has taken place, the solution on the inside diffuses out and comes in contact with the mixed sulphides, and we have the reaction
+
C U S O ~ CdS = CUS
+ CdSOA.
If enough copper has been precipitated, the cadmium sulphate will diffuse alone and will be precipitated separately on treatment with hydrogen sulphide. I t seems quite likely that the diffusion method for the detection of cadmium in presence'of copper may be used when the copper salt is not present in too great excess. Let us consider the different possibilities. We may have a solution of such a concentration that both the copper and the cadmium sulphate diffuse as fast as the water. This is the case that has just been described and it is possible t o apply the method of fractional precipitation. It would also be possible t o dilute the solution until a water-ring was formed on diffusion.
Ad. A. Gordon
3 42
If the cadmium were present in sufficient relative amounts, it would show up without fractional precipitation. If we had a solution, however, so high in copper and so low in cadmium that the copper diffused as fast as the water while the cadmium did not, there would be, by hypothesis, no cadmium to be carried on as a result of fractional precipitation. Diluting the solution would not help matters because the cadmium would practically never go as far as the copper. On the other hand the solution could be concentrated by evaporation until the cadmium diffused as fast as the water and then the method of fractional precipitation could be applied. It is not likely that this test will be of much value in qualitative analysis because a negative result is not conclusive as to the absence of cadmium. It is, perhaps, desirable to say a few words about the general theory of the formation of a water-ring in filter paper because there are several points in Bayley’s work for which he offered no explanation and the text-books on colloid chemistry apparently make no reference either to Bayley or to L1oyd.l Ostwald2 mentions the work of Schonbein (which was apparently unknown to Bayley and to Lloyd) and says that the phenomenon is evidently one of adsorption, which is perfectly true as far as it goes but does not give much in the way of detail. Bayley also saw that it was a case of adsorption though he could not make any especial use of this.3 “ I therefore made a few experiments, in order to determine whether filter paper has any power of withdrawing silver salts from solution. 7.8703 grams of silver nitrate having been dissolved in water and diluted to 500 cc, a quantity of this solution was placed in a beaker, and a roll of filter paper sufficient to absorb nearly the whole plunged into it. After some minutes the filter paper was removed and as much of the liquid as possible squeezed into the beaker.
3
Chem. h’ews, 51,51 (1885). Lehrbuch allgem. Chem., 2nd Ed., I , 1098 (1891). Bcyley: Jour. Chem. SOC..33, 304 (1878).
Adsorption by Filter Paper
343
It is clear that the effect of this treatment, unless the paper possesses a power of retaining the salt, would merely be t o slightly concentrate the solution by evaporation. Four experiments were made in this manner, fresh rolls of paper and fresh portions of the standard solution being used in each instance. Twenty-five cc of the solution, after treatment with the paper, were mixed with hydrochloric acid, and the precipitated silver chloride dried, -ignited, and weighed. ‘‘ The results were as follows : ~-
Solution used cc _____-.
25 25
50
1
_ ~ _ _
-;-__-- - - / 1I I
~-
Ag found gram
-~
1
0.2182
0.2306
0.2353 0.4760
Ag originally present gram 0.2500 0.2500 0.2500
0.5000
This shows that, after a roll of paper has been soaked for a few minutes in a solution of silver nitrate, the quantity of silver has materially diminished. A few experiments made by my friend, Mr. Weston, show that the same is true in the case of mercury salts.” Like other substances filter paper shows a selective adsorption for dissolved substances, taking up some very completely and others but slightly. If a few drops of different salt solutions be placed upon a piece of filter paper, the substance which is adsorbed the most will diffuse the least distance. In other words, we should expect the widest waterring with the solution containing the most readily adsorbed sa1t.l With the same salt we should get the widest waterring with the dilute solutions, which is exactly what Bayley found and what, is shown by the experiments in Table I. Bayley obtained wide water-rings with salts of silver, lead, and mercury in moderately concentrated solutions. He himself showed that silver nitrate is adsorbed by filter paper. Yorke2 found “ o n filtering a solution of oxide of lead Cf. Freundlich: Kapillarchemie, 156 (1909). Mem. Chem. SOC.,2, 399 (1845).
in lime water through a triple filter, that, whereas the original solution gave a deep black when tested by sulphuret.ted hydrogen, the filtered liquid gave but a pale brown; and it required that the unfiltered liquid should be diluted with thirty times its volume of water toproduce the same test as the filtered. I then tried the effect of mere immersion of the paper in the aqueous solutions before used. A bit of filtering paper ten inches by two inches was boiled in distilled water and then put into an ounce phial filled with the aqueous solution; after remaining six hours the liquid was poured off and tested: it gave a pale brown, and it required that the liquid which had not been in contact with the paper should be diluted with ten times its volume of water to produce the same tint.” Schwalbel says that lead salts are kept back quantitatively by cotton fibers and that one must not filter solutions of lead through paper in quantitative work. If the filter paper is washed with boiling water, the lead salt is fixed the more firmly. Skraup‘ found that lead acetate was decomposed by filter paper, the acetic acid being less adsorbed than the lead. Vignon3 has shown that cotton wool adsorbs mercury salts so much that it decomposes mercuric chloride into free hydrochloric acid and basic salt or oxide. In the case of these metals we thus have confirmatory evidence of the obvious relation between degree of adsorption and extent of water-ring. Herzog? gives some data on the adsorption of copper sulphate by cotton wool; but we cannot compare these data with those of Bayley on silver nitrate because Bayley does not give the weight of the filter paper added. It seems quite certain, however, that the silver salt is adsorbed more than the copper salt. Muller’ has shown that filtm paper possesses the property of taking up not inconsiderable quantities of Die Chemie der Cellulose, 80 (1911). Sitzungsher. Akad. Wiss. Wien, 118,IIb, j64 (1909). Comptes rendus, 116,5 1 7 , j84, 645 (1893). Zeit. Farbenindustrie, 7, 281 (1908). Jour. prakt. Chem., 83, 384 (1861).
Adsorption by Filter Paper
barium hydroxide from an aqueous solution of this substance and that such a solution should not be filtered through paper in quantitative determinations. I n line with this are Schiinbein’s and Skraup’s experiments’ showing that barium hydroxide solutions give a wide water-ring. Since adsorption is accompanied by evolution of heat, there will be less adsorption a t higher temperatures and consequently the water-ring will be relatively narrower. This is what Bayley found, for he says that the metal in a warm solution is more mobile than if the solution were cold.” According to Skraup’s measurements, sulphuric acid does not rise as high in filter paper as copper sulphate of equivalent concentration, and is, therefore, adsorbed to somewhat greater extent. If this is so, copper sulphate will be adsorbed relatively less from a mixture of sulphuric acid and copper sulphate; the metal will, therefore, give a narrower water-ring in an acid solution than a neutral one and this is what Bayley found. There is another way of looking a t this which does not depend on the adsorption of sulphuric acid. If we assume that cellulose has a distinct adsorbing power for cupric hydroxide or for a basic copper salt, the adsorption will be less the more acid the solution and consequently the copper will be more mobile. This seems more plausible than the other explanation. Under existing conditions it was unfortunately not possible for me to test this point experimentally. It is not enough merely to consider the adsorption of the dissolved salt or of one of its decomposition products. The taking up of water is an important factor. Vi’gnon2 has shown that one gram of cotton may take up five grams of water. That is an extreme case and does not occur in any ordinary experiments with filter paper. Suppose, however, that we let fall a drop of solution on a piece of filter paper and suppose that the nature and concentration of the solution are such that water is taken up more rapidly that the dis‘ I
’
345
Skraup: Sitzungsber. Akad. Wiss. Wien, 118,IIb, 565 (1909). Comptes rendus, 127,73 (1898).
M . A. Gordon
346
solved substance. I n that case a more concentrated solution will spread out and we shall get an outer ring of a concentrated solution. This was actually observed by Bayleyl though he was unable to account for the phenomenon. “ Experiments with the hydrates of calcium, sodium, and ammonium dropped upon turmeric paper, showed that the water-ring is formed if the solutions are dilute. When they are concentrated, the alkali extends to the edge of the blot and, indeed, seems concentrated there, for the brown patch is surrounded by a ring darker in color than the other parts.” In one place Skraup2 says that “N/50 hydrochloric acid is adsorbed by the same surface of paper more strongly than acetic acid, and consequently i t must rise less high in filter paper, which is actually the case.’’ This sounds as though Skraup were fairly clear as to the theory of the phenomenon but the following quotation3 makes one doubtful whether this is much more than an accidental phrase. It certainly never became a working hypothesis. “ I n many cases the difference in the heights to which the solutes rise is approximately as great as the difference in the deyee of electrolytic dissociation. That this is not the sole cause of the differences appears from the behavior of sulphuric acid, which rises to about the same height as the halogen acids and nitric acid at concentrations for which it is much less dissociated than these. More remarkable still is the behavior of phosphoric acid which is distinctly a weak acid and yet which rises less high than the strongest acids. Since experiments with the basic hydroxides have shown that barium hydroxide and calcium hydroxide rise less high than the oxides of the alkalies, we might account for the behavior of phosphoric acid by ascribing an important part to valency taken in its broadest sense. As against this
2
a
Jour. Chem. SOC.,33, 306 (1878). Sitzungsber. Akad. Wiss. Wien, 118,IIb, 596 (1909). Ibid., 118,IIb, 562 (1909).
Adsorption by Filter Paper
347
we shall have rises with boric and succinic acids. Perhaps a study of other acids will clear up matters. “With the alkaline hydroxides we have in general the same results as with the acids, the solutes rising less high with increasing dilution. As previously stated, the stronger bases rise higher than the weaker ones. It is interesting to note that a t high dilutions the hydroxides, and especially those of the alkalies rise higher than do equivalent concentrations of the strong acids. It is not possible even to guess whether this has anything to do with the fact that transference numbers for hydrogen and hydroxyl vary in magnitude in the opposite direction.” “Ammonia and ethyl amine show an abnormal behavior. Both rise much higher than the strong alkali hydroxides, and the heights to which they rise decrease but little a t first with increasing dilution. At a concentration of N/IOO ammonia rises about as the alkalies. Until other amines have been studied, i t is impossible to formulate any definite statement. “The most striking thing about salts is that in general they rise much higher than equivalent solutions of the corresponding acids and bases. In many cases it is difficult to detect any lagging of the solute behind the water even a t high dilutions. Even with those salts which rise to a lesser height with increasing dilution, the rise is higher than that of an equivalent solution of the corresponding base or acid except for extreme dilution. This relatively large rise of the salts is in accord with the observations on the action of salt solutions on porous substances such as bone-black, kaolin, etc., as these substances adsorb salts of the alkalies only to a very slight extent. “Of special influence on the rise of salt solutions is hydrolysis and this can be detected even with very stable salts. With sodium chloride or potassium sulphate, for instance, the part of the filter paper outside of the solution gave a faint blue color with neutral litmus or azolitmin while the submerged end assumed a faint reddish color when tested. In-
M . A. Gordon
348
creasing dilution causes a decrease in the rise in the case of salts which we know to be strongly hydrolyzed. At concentrations greater than N / I O Opotassium carbonate rises higher than potassium hydroxide; but this difference ceases to be perceptible in N/200 solutions. This points to practically complete hydrolysis. I n fact the degree of electrolytic [ ? ] dissociation can be shown indirectly in many cases. Thus mercuric cyanide, which is scarcely dissociated a t all, rises nearly to the same height with increasing dilution, whereas the decrease is much more marked with cadmium iodide, which is more dissociated to start with and, therefore, also more hydrolyzed.” Skraup evidently did not appreciate that the filter paper causes hydrolysis when it tends to adsorb a base more than an acid or vice versa.l Shields2 found that sodium carbonate was about three percent hydrolyzed in N / I O solution. It is, therefore, very improbable that potassium carbonate is completely hydrolyzed in N/200 solution as Skraup assumes. Kohlrausch3 gives the equivalent conductivity of N/200 KzCO3as 121.6 at 18’ and the corresponding value for potassium hydroxide as 230. Lloyd’s experiments4 are very similar to Bayley’s except that he used some mixtures containing three or more dissolved substances. He found that the water-ring is widest the more dilute the solution. Owing to not working with sufficiently dilute solutions he obtained no water-ring with sodium chloride. One experiment is interesting enough to be worth quoting. “ I n carrying this series of experiments further, it is readily shown that not only can we separate liquids from each other within the paper, but we can separate them as liquids by acknowledging the fact that a liquid tends t o flow from a tube, capillary or otherwise, if the extremity is beneath the surface of the liquid in the container. Two Bancroft: Jour. Phys. Chem., 18,6 (1914). Zeit. phys. Chem., 12, 167 (1893). a Leitvermogen der Elektrolyte, 159 (1898). 4 Chem. News, 51,5 1 (1883). 1
2
Adsorptiovt by Filter Paper
349
test-tubes were placed beside each other, and into one an inch of solution of ferric sulphate (the strength before named) was poured. A strip of blotting paper was then so placed that one end reached into the liquid, while the other end rested below i t in the other vial. The paper was curved so that the height was four inches; therefore, the liquid traversed eight inches. The exposed part of the paper was covered by means of a sheet of rubber, in order to retard evaporation. I n twentyfour hours a layer of colorless liquid was carried into the empty vial, and this liquid refused to show a trace of iron by the usual reagents. ” The fundamental difference between the spreading of a drop on filter paper and the rise of a solution in filter is that the solution becomes exhausted in the first case and not necessarily in the second case. If the strips are left in the solution long enough, fresh solution can rise in them, carrying the solute higher. This is shown clearly in some experiments by PeletJolivetl with dyes (Table 11).
TABLE I1 I
Time of experiment, hours
1 0.5
1
I
2
4
6
-- 1
Methylene blue, 0. I percent Rise of dye, mm Rise of water, mm Methylene blue, 0.2 percent Rise of dye, mm Rise of water, mm Crystal ponceau, 0.1 percent Rise of dye, mm Rise of water, mm Crystal ponceau, 0.2 percent Rise of dye, mm Rise of water, mm
20
88 24 83
76 91 66
79
Unless fresh solution is supplied from below, the height to which the solute rises does not vary much with the time. Blotting paper was cut into 8 cm lengths and was dipped “Die Theorie des Farbeprozesses,” 123 (1910).
350
M . A. Gordon
into solutions of such concentrations that a water zone was always formed. A known volume of the solution was taken up, usually one cubic centimeter. When all the liquid had been taken up, the strips were set aside in a moist atmosphere for different lengths of time and then the height determined to which the solute had risen. With cadmium sulphate or copper sulphate this was done by exposing the strips to hydrogen sulphide. With 0.05 molar cadmium sulphate and 8 cm strips the cadmium salt stood at a height of 7.4, 7.3, 7.7 and 7.3 cm after an hour in the moist atmosphere and a t 7.4 cm after eighteen hours. With 0.025 molar copper sulphate and 7 cm strips, the copper stood a t a height of 5.4 * 0.I cm at all periods from fifteen minutes up to eighteen hours. Theoretically, there shodd be a levelling effect and an eventual uniform distribution of the solute; but practically this effect is negligible. One of the points which excited comment is the sharpness of the dividing 1ine.l “Take an ordinary drop of porous blotting paper and drop into its centre some writing fluid drop after drop. The spot will spread, but it will not present the same appearance from the centre outward. There is usually a dark centre, and then a dark line of demarcation, after which another shade appears, which, after spreading to a certain distance, will perhaps suddenly give place to a nearly colorless liquid. Continue to add the fluid slowly to the centre of the blot, and the shades of color will expand and preserve their individuality, but the outer will usually grow more rapidly than the one immediately within. Sometimes several shades will be formed, but their individual characteristics will be maintained. If the ink be one of the purple or other colors of aniline, or a carmine, it will generally be found that the outer liquid will be colorless. The striking feature is the abrupt change from one shade to the other. It is not a gradual grading-off, for a distinct line of demarcation usually separates each shade. We have introduced this experiment Lloyd: Chem. News, 51, 51 (1885).
Adsorption by Filter Paper
351
because i t can be so readily performed, and because, upon second thought, every person must even now admit its familiarity. Mix two colors of ink, say red and blue, and t r y the experiment again, Very likely i t will be observed that, under the same conditions, one color will leave the other after both have passed together for a certain distance, and leave it completely, and by a distinct line of demarcation. Then, perhaps, this second color will cease to spread, and a colorless liquid will pass out, and form a ring encircling the ink spot.” The reason for the sharp dividing line is apparent when we consider the form of the adsorption isotherm. I n r.ost cases the amount adsorbed increases very rapidly with small increases in the concentration of the solution. Consequently the range of gradations will be correspondingly narrow and will usually escape notice. If one had a substance which gave a much flatter adsorption curve with cellulose, one would not get the apparently sharp dividing line. While i t would be interesting to have such a case, I have not been able to devote any time to searching for a suitable substance. It would, of course, be one for which the exponential factor was low in the equation for the adsorption isotherm. Pelet-Jolivet’ has studied the rise of dyes in filter paper and in other fibers. “It is not possible to explain each one of the results satisfactorily. Nevertheless we can deduce from them a certain number of general rules which are very in teresting . ‘‘ I , For a given fiber the rise of a dye is less the more readily the dye is adsorbed by the fiber. “2. If a dye is adsorbed by a fiber, the capillary rise i s generally leis for dilute concentrations and greater for high concentrations. A graphical representation of the relation between rise and concentration gives a curve which bears a strong resemblance to adsorption curves. “3. Basic dyes rise to a moderate height but usually not so high as acid dyes; rhodamine dyes, which are less basic, “Die Theorie des Farbeprozesses,” 128 (1910).
352
M . A . Gordon
come in between these two classes. Direct cotton colors usually rise less high than the acid colors. Among the basic colors those containing NH2 groups usually rise less high in cellulose (paper and linen) than those containing N (CH& or N(C2H&groups. This difference does not occur with animal fibers. “4. Dyes having a distinctly marked colloidal character usually rise less high than others (fuchsine, alkali blue, benzopurpurine) of similar constitution. “According to a still unpublished paper1 of which my colleague, Professor Fichter, of Basel, very kindly informed me, it seems certain that changes in the capillary rise depend to some extent on the colloidal state. He has studied the rise of positive and negative colloids in strips of paper and has found that positive colloids rise but slightly while negative colloids rise freely. These interesting facts bring out a new analogy between dyes and colloids although it has been shown that dyes are electrolytes. “Has the basic or the acid character of the molecule alone an effect on the rise? This seems improbable. There must be other properties which affect the capillarity. With dyes of similar constitution the, molecular weight appears to be an important factor, as for instance ___
_ ___
I
I Fluoresceine Eosine (tetrabromfluoresceine) Rose Bengale (dichlortetrabromfluoresceine)
Mean rise Mm 1.51 121
114
“If we sum up these general rules we may say that, subject to the limited accuracy of the method, the capillary rise and the adsorption are closely connected; the capillary rise i s less the greater the adsorption and vice versa.” 1 Fichter: Verh. naturforsch. ges. Basel, chernie, 8, I (1911).
21, I
(1910); Zeit. Kolloid-
Adsorption by Filter Paper
353
This final conclusion is entirely right; but it can be interpreted in several ways and the context shows, I think, that Pelet-Jolivet intended merely to say that adsorption was one of the factors and not that it was the factor. He discusses the questions of the molecular weight, of the constitution, of the basicity and acidity, with reference to the capillary rise and not with reference to the adsorption, which is where they belong. He apparently accepts Fichter’s views that the precipitation of a positive colloid is fundamentally different from adsorption. Working with dilute solutions, Pelet-Jolivet did not observe the formation of a darker outer zone and he does not refer to the work of others who have observed this. It would have been impossible for him to> have accounted for this phenomenon so long as he considered only the adsorption of the solute. The question is one of the relative adsorption of solvent and solute. There is still a great deal of work to be done before all the details of the rise of solutions in filter paperehave been cleared up. We need more facts in regard to specific adsorption and there is also the problem of the movement of water or of solution through the fiber instead of over the surface. It all comes back, however, to a question of adsorption of water and of the constituents of the solution by the paper. The general results of this paper are: I . Filter paper shows selective adsorption for water and for each constituent of the solution. 2 . If the solute is adsorbed relatively faster than t h e water there will be formed a water-ring. 3. If the water is adsorbed relatively faster than the solute,, the latter will concentrate in the outer zone. 4. Whether one gets a dilute solution (water-ring) or a more concentrated one in the outer zone depends on the concentration as well as on the nature of the solute. 5 . The more readily a salt is adsorbed relatively to the water, the. higher will be the limiting concentration giving a water-ring. 6 . The amount of base or of acid taken out of a salt
M . A. Gordon
354
solution by filter .paper is not a measure of the degree of hydrolysis before the paper was dipped into the solution. 7. >Hydrolysis is not complete in a N/2oo potassium carbonate solution. 8. Since adsorption decreases with rising temperature, the waler-ring will be relatively narrower at higher temperatures; the salt will appear more mobile. 9. Addition of anything which cuts down the adsorption -sulphuric acid to copper sulphate solution, for instanceincreases the mobility of the salt. IO. With a strip of filter paper dipping into an excess of solution there will be a continual supply of solution to the filter paper, which will give rise to phenomena which we do not get when the filter paper is present in excess, as in the case of a drop of solution upon a piece of filter paper. 1 1 . Theoretically, there should be a tendency for the dissolved substance to distribute itself uniformly over a short strip of filter paper in time. Practically this tendency is negligible. 1 2 . The apparently sharp dividing line in the case of a water-ring is a necessary consequence of the usual form of the adsorption isotherm. With a flatter adsorption isotherm the contrast would not be so marked. This work was suggested by Professor Bancroft and has been carried on under his supervision. 1
Cornell University
