~ ~ t (,.; ~·t·W "J'• ~SiC~ C k•W\ • . \.\~ \
.1u~A
1
! Victor Le-nher a.nd George M. Bishop
98
~ .....
TABLE
Gm oleic acid per 50 cc fractions
Fraction
Fraction
Sodium oleate solo. I
.,
I II III IV
'. ~.:
v
~'· :··., ',.
VII
'·j;: .,-;·, VI
'
r .o68 0.008 0.038 0.123 0.241 0 .458 0.572 0.687
I
Gm oleic acid per 50 cc fractions
VIII IX X XI XII XIII XIV, XV
~
J
-"~ '\
'
"
'~~ .
~ ~~~
.&
~ -'...,. _
...
.
;~~~~·
~,..g
sa.
5 ....
w-- . ~~
OSMOTIC PRESSURE WITH MEMBRANES OF CHEMICALLYINERT MATERIALS' BY
s. L.
~ND c. s.
ROBINSON
The direct observation of osmotic pressure requires the use of a so-called semipenneable membrane, the function of which has been the cause of much discussion among investigators interested in osmotic phenomena. When two bodies of , matter are separated by a layer of a third which is of such a nature that the others pass through it at unequal rates, osmotic pressure results and the intennediate layer constitutes a semipenneable membrane. Usually the material ?n one side of the membrane is called the "solvent" and that on the other side the "solution." Generally the latter consists of a solution of some solute in the " solvent." Under suitabie conditions, \vith the proper apparatus, the solvent ·passes more readily through the membrane into the solution until fin ally a state of. · equilibrium is reached when a certain pressure is developed in • the solution which pressure is called osmotic pressure. With.;· ' out the semipermeabl~ membrane the solvent and solution merely diffuse into each other until a solution of uniform concentration is produced. No difference in pressure results. Generally speaking, the membrane "allows of a pressure being exerted on a ·solution in contact with a solvent, without exerting a pressure on the latter." 2
Previous Work and General TheoPies Involved While such is essentially the acknowledged function of the membrane the mechanism of its action is a matter of d0ubt. The theories in regard to it may be divided into two classes in which the membrane has respectively a passive and an active function in the phenomena. The former includ~;s such
"\
.
, ,
1
·: .Av. SOME EXPERIMENTS ON THE MANIFESTATION OF r.{l
III
'
'd""' \
·~
Action of Animal Charcoal on Soap Solution
~
{ ~- - · -
-~~
~.
._, "'
1
~ -Aik..~
.- ~C\7.•
tain considerable calcium carbonate and phosphate. It was purified by long digestion with moderately strong . hydrochloric acid, after which it was thoroughly washed until free from chlorides. It was dried and heated to redness for three · hours just before filling the tubes. . ·. This particular fonn of carbon shows greater .adsorptive powers than either Ceylon graphite or \villow charcoal. The first fraction showed almost complete adsorption and a total of 7so cc of soap solution was passed through the column of animal charcoal before adsorption ceased.
_\
.
'(
1 The experiments described in this article were performed' in the Chemical Laboratory of the Michiga.n Agricultural College Experiment Stat ion and the results are published with t.he perm1ssioo. of t.he Dlrector. s Findlay: "Osmotic Pri!Ssure," p. 66 (r. gt ~ I.
·-·I.· ·
j~ ...,
e
roo
5. L. Bigelow and C. S. Robinson Experiments on the !VIanijestation of Osmotic Pressure ror
theories as Traube's "atomic sieve". t)leory and Poynting's hydrate theory while the latter includes the various solution the electrostatic theories of Perrin and of Girard,
· ponents." He also clogged the pores of some of his membranes with other materials in order to reduce the size of the interstices. Such membranes varied in permeability and exhibited certain .• ~::'properties which substantiated his theory. · 1 2 ' Ostwald studied the permeability of membranes for ions in the light of Traube's theory. From his work it follows that 'I a membrane i~ permeable for a salt only when it is permeable . for both ions of that salt; and if a membrane is impermeable for · an ion it is impermeable for all salts containing that ion. 3 Tammann, however, disproved the above statements and declared himself opposed to Traube's theory. · 4 . Likewise Walden .•after an exhaustive investigation of several precipitation menl.Qran~ says (p. 718) "precipitation membranes cannot act as a~ornic sieves." Pickering 5 attempted to substantiate Traube's theory. 1 ' 2
"Arch. A nat. Physiol. tmd Wissen. Med.," 1867, p. 87. Zeit. phys. Chem ., 6, 71 (1890). ' lbid., xo, 255 (1892) .
~ Ibid., 6
xo, 699 (I892) .
Bcr. dcutsch. chem . Gcs., 24, 3629 (1891).
•
He suggested that solute molecul~s possess a certain. affinity by .which they attract solvent molecules and hold them in a more or less unstable manner. These clusters of molecules .pass through the pores' of the membrane less readily than single solvent molecules. · This work was later discredited by Barlow 1 who . demonstrated · the defectiveness of. the experimental methods upon which it was based. The simple sieve theory has been more or l~ss modified in the minqs of later workers especially as regards the mechanism of its action. For instance ~piting 2 assume? a combination .. between · solute and solvent molecules which decreases. the mobility of the latter. H~s conception of osmosis is as follows: ''The solvent molecules are en~ering the .membrane from both sides, but the mobility or number set free per second from the pure solvent is greater 'than the number set free from the solution. ·. The membrane goes on absorbing the solvent from each side till it becomes saturated, i. e., holds so much that it returns as many molecules as it receives. It is . . receiving more from the pure solvent side, and therefore when · .;,!··:,~ '· sat~rated for that side i§ supersaturated for the other . . :'Consequently more molecules· are,sent into the . solution t han are received from it, and the solution grows until the growing pressure so much increases the mobility that it is eqtial on . both sides of the membrane." Later Larmor 3 while believing that " each molecule of the dissolved substance forms for itself a nidu,s in the solvent;. that is, it sensibly influences the molecules around it up to a .certain minute distance, so as to form a loosely connected complex in . the sense not of chemical union but of physical influences" still criticized Poynting' s theory as follows: 1 'that ·. Prof. Poynting's recent suggestion with a view t o evading the · neces~ity of the ionic dissociation hypothesis, cannot '
'
1
Phil. 1\'Iag., (6] xo;
(1 905); Findlay and Short: J our. Chcm. Soc., 87, ' Phil. Mag., (5] 42, 289 (1896). 'Nature, 55, 545 (L 89] ). t
8!9 (1905) . 2
\.
•
•
•
f.'· If
•L'
~~
~~f\'1 \
•
5. L. Bigelow and C. S.
102
Robinso1~
avail, as it would not lead to the desired value for the osmotic pressure ; the pressure ~epends on the number of molecular. complexes involving the dissolved substance that exists in the dilute solution, but not on their individual degrees of com·.plexity." --~~~till another variation of the sieve -theorY is that . of Sutherland 1 who states that "if we seek to picture to ourselves how a membrane allows water molecules to pass, but not sugar molecules, o r re is ~esh, amidst the threads of which the water molecules are packed in · such a manner as to give way before one .. . . ·. another almost as in ordinary water, while the sugar molecules ffl'4 . :r.~ 1~; . •. are held back by the mesh. Thus the mesh forms a solid or · quasi-solid framework · through which water can pass with high viscous resistance, while the sugar molecule is absolutely blocked. Now if the framework turns back the sugar molecules, it must take the force of their blows and shield the water· molecules from them. If then we suppose a semipermeable membrane separating water and a dilute solution of sugar in water the sugar molecules are to be regarded as replacing some water molecules, but their collisions on _the water in .the membrane are rendered inoperative by the shielding action of the framework so that the water molecules in the' membrane recdve more impacts on the side of the pure·water than on the side of the solution and therefore water flows through the membrane until in the solution there is enough excess bf_ .hydrostatic pressure established to compensate for the in. oper ative impacts of th~ sugar molecules. This inequality of pressure which can tie hydrostatically balanced is the osmotic pressure.'' ,_ M. Traube's co~ceptio~ of the actiop. of the semipermeable membrane was s'ubjected to emphatic adverse criticism by J. '1'raube 2 who cites the work of Overton. 3 Overton ~
1
~
2
Phil. Mag., [5)44, 493 (189 7). Ibid ., 16] 8, 704 (1904). ~ · zeit. phys. Chern ., 22, 189 (1897); Vicrtcljahreschr. d. Naturf. Geselsch. Zurich, 40, 1 (1895); 44, 88 (1899) .
•
E xper,iments on the Manifestation of Osmotic Pressure
103
showed that permeability increased in homologous series by the substitutions of CH 3, for H, C2Hb for CHa, etc . This was contrary to M. Traube's hypothesis which postulated that an · increase in the size of the molecule should cause a decrease in the readiness with which a substance should pass through a membrane. Also Barlow 1 after demonstrating the defects in Pickering's work states that "It may justly be concluded that these experiments prove that the part played by the membrane in osmotic phenomena is not a sieve-like one. " He claimed that pure alcohol passed through a copper ferrocyanide membrane easier than pure water although when these two are sep_a rated by such a membrane the water ;fiows 'into the alcohol. Kahlenberg 2 also opposes the sieve theory. . He says: "When we think of a large molecule like that of copper oleate rapidly travelling through vulcanized caoutchouc as in No. · 35, and that under like conditions cane sugar, AgNOa, • and LiCl do not pass through the septum, it certainly must c<;mvince us that the membrane does riot act as a sieve." Finally in connection with this theory attention may be . , called to the recent work of Tinker 3 who examined , with the ultramicroscope, precipitation membranes made according to Traube.. He says (p . 368): "Since the smailest pore diameter in a copper ferrocyanide membrane is about 8 millimicrons, it follows that from 100 to 200 water molecules could be placed in a chain from one side of such a pore to 't he other, and that several thousand could be travelling across the cross section of the pore at the same time. A .selective mechanical blocking of even large hydrated crystallized molecules is ·,hence out of the question, and this hypothesis for accounting for osmotic effects-already largely rejected-is no longer tenable." , A second series of theories ascribing a mqre active lli!r..rt to the~mbrane, a~ e: eg,pWary strl!,Sture. ChronoLoc. cit. Jour. Phys. Chern., xo, 169 (19o6). · ' Proc. Royal Soc., 92A, 357 (1 916). 1
1 2
'·
•
0
-yof\
ro4 ~5. L. Bigelow a11d C. 5. Robinson
Experiments on the M anijestat1:m1- of Os·motic Pressure ros
logically tl(is was apparently the first to be advanced. From subsequent articles it s~probable that Fischer 1 must have ~on of the organization of t}le membrane. Likewise Poisson 2 attempted to show that osmosis was due to capillarity rather than electric action as suggested by . Dutrochet. The latter denied that ·c apillarity could explain the ,phenomepon although he later 3 stated distinctly that osmotic phenomena were observed when two liquids were separated by a membrane with capillary pores. Draper·1 states emphatically his belief in the capillary structure of osmotic membranes and gives details of h's conception of the action of such a membrane. . , ~he real development of this theory however dates from an article b~ri.icke. 5)He devised a cell consisting of a glass tube pressed against a glass plate and obtained osmotic effects through the crack thus formed. G He attribut~d the results to the difference in the attraction of the walls of the capillaries for the two liquids and explained in detail the mechanism of the action. Evidence tending to conf1m1 his views was contributed later by Ludwig 7 who showed that memqranes actually did take up less solution than solvent. . Fick 8 · diS,;.
I
1
'
'Gilberts Ann ., 72 1 298 (1826).; see also Pog-g. Ann., I0 1 481; II, 126 (r827). 'Pogg. Ann., II 1 134 (~827). 'Ann. Chirn. Pbys., 49, ·4;12 (1832). 'Jour. Franklin fnst., I7,' 177; IS, 27 ( 1836). 6 Pogg. Ann., 58, 77 (1843). • Sec also Fisch er (Loc. cit.) and Dutrochct (Anrr. Chim. Phys., 35 1 393 (1827); 49 1 411 ( 18.12) ) for osm~~is through cracks in glass. · • Pogg. Ann., 78, 307 (18~). 'll~ , 8 Ibid ., 94, 59 (1855). ' • Pogg. Ann., 147 1 195 (1872). ' . • 10 "Osmotische Untersuehuu~:eu," l,eipzig (1877).
•
.·
subject, expr_essed himself as favoring the capillary theqyy. Moore 1 attempted to show that osmosis ."may be produced by diffe~ence in surface tension acting along the exceedingly fine capillary openings of almost molecular dimensions whiCh place the solution in .connection with its solvent in the pores of the semipermeable wall which separates them." He says: . "It seems reasonable to suppose · that after such a process (the precipitation of a membrane) communication takes place only in the intramolecular spaces or meshes of the precipitate." The work of Tink~r.? may ~g~in be cited connection " with the capillary theory. He measured, microscopically, the sizes of the particles composing the precipitation membranes made from the same materials that Traube used and found a distinct connection between the diameters of the membrane pores, as calculated from the sizes of the particles, and the osmotiC activity of the various membranes. He _says (p. 370): "1'he relation between the extent to which the capillaries of a membrane are controlled by the surface forces and the osmotic properti~s of the .membrane is, however, an apparent. and · noteworthy one, CQ~~er ferrocyanide and Prussian .!lllle-the most perfect of membnines-h:we 'their p·ores completcly under control; the tannate and silicate membranes, ~hich are not so efficient, have a central space .within their · capillaries throughout which the surface forces are comparatively weak, while mer.n branes su~h ~s g~latin and parchment, which are permeable to all crystalloids, have pores which P,Ossess central canals outside the surface force range." The capillary theory, as Bigelow.3 has pointed out, serves as a compromise between the strictly mechanical "sieve" theory and the ~trely chemical "solution" theory. This last theery ascribes to the metnbrane a still more active part in the pr.Q_cess of osmosis. It post!flates that a membrane is permeable to such substances as are soluble in it and imper-
in
1 Phil. ?-.lag-., 38, 279 (1894). • Loc. cit.; sec also PfciTcr: "Osmotischc Untcrsuchun~en," p. 43· • Jour. Am. ChC;m . Soc., ~9. 1675 (1907). ' ·
\.
'-
106
S. L. Bigelow a11d C. S. Robinson
meable to those which are not. This theory, aeur probably tl~ most generally accepted ~e, wa2_ andcipated by Lkbig 1 · as early as I 848 when he said: "The· volume changes of two miscible liquids which are separated .from each other by a . membrane depends upon the unequal wetting or attraction . which the membrane exerts on the two liquids." In view of Liebig's own assertion later that his conception of . osmosis appeared to be identical with Graham's it is certainly permissible to interpret "Anziehung" in a broad enough way to include "attraction of solution" in which case · we have an exact statement . of the present-day conception of the solution theory. While Liebig's statement must be considered as exceedingly significant iti view of later developments, the first important experimental work performed with the object of substantiating such a theory was carried out some ye-ars later ' by L'hermite. 2 He appears to have had a very _clear idea of the possible relations between osmotic phenomena and solution and capillarity. He was the first to use the "three liquid layer" means of demonstrating that one liquid, soluble ·in a second liquid will pass into it through a layer of .a third 'liquid in which the first is more soluble than the second. He pointed out the significance of this experiment from the standpoint of osmotic phenomena. At almost the same time Graham in a series of articles dealing with the diffusion of liquids under various conditions formed practically the same idea of the process of osmosis, though his · statements were hardly as clear nor were his experiments as conclusive, ' in the light of later work, as were L'hermite's. For instance ' in his early work 3 he found that the most pronounced osmoti~ effects were obtained with acid or basic substances while neutral salts and organic compounds were inert or showed very weak effects. This appears) to 1
"Ursachcn der Siiftcbewcgung," Braunschweig (1848); also Liebig's Ann., 121, 78 (1862). • Ann. Chim. Phys., (3) 43 1 420 (1855). 3 Phil. Trans., 144 1 177 (nbs. in Phil. Mag., (4) 8, 151 (1854)).
•
Experiments on the Manifestation of Osmotic Pressure
107
have been the basis for the formation of his first theory that osmosis depended on the interaction of the membrane and the liquids bathing it. In this article he opposes the capillary explanation of osmosis since he was unable to obtain results with membranes of various materials which he knew to be porous. Later he' seems to' have approached more nearly the pure solution theory as he says :1 "The separation described (sugar from gurri arabic by dialysis) is 'somewhat analogous to that observed in a soap bubble inflated with a gaseous mixture composed of carbonic acid and hydrogen. Neither gas as such, can penetrate the water film . But the carbonic l;I.Cid, being soluble in water, is condensed and dissolved by the water film, and · so is enabled to_ pass outwards and reach the atmosphere, while hydrogen being insoluble in water, or nearly so, is retained within the vesicle." Although 'this statement was made in regard to the. process of dialysis, it is ,., certainly suggestive of the tendency of thought, especially in view of his explanation of osmosis with aqueous solutions and colloidal membranes which he gives on p. 222 of the same article. "It now appears to me that the water movement in · . osmose ~s an affair of hydration and dehydration in the substance of the membrane or other colloidal septum. " . And on · p. 223, "Placed in pure water, such. colloids a-re ~ydrated to a higher degree than they are in ·neutral saline solutions. Hen~e the equilibrium of hydration is different on the,.two sides of the membrane of an osmometer. The outer surface of the m embrane being in contact with pure water 'tends to hydrate itself to a higher degree than the inner surface does, the latter surface being supposed to be in contact with a saline solution. When the full hydration of the outer surface extends through the thickness of the membrane and reaches the inner surface it , there receives a check. , The degree of hydration is lowered, and water must be given up by the inner ayer of the membrane, and it forms the osmose." N ernst 2 describes a "three liquid layer" experiment ...... 1
1
Phil. Trans., 151 1 183 (1861) . · Zeit. phys. Chcm., ·6, 37 (1 890). '-
.'
•
108
S. '-· Bigelow and C. S. Robinson Experiments on the Manifestation of Osmotic Pressure 109
similar to those first performed by L'hermite which he used to . illustrate the action of an osmometer. He separated ether from a solution of ether in benzol lly a parchment membrane saturated with water, the water constituting the intermediate liquid layer or "membrane." Tammann, 1 after citing experimental evidence against Traube's sieve theory, calls attention to the work of L'hermite and Nernst and definitely compares precipitation membranes to their three liquid layers, saying (p. 263), "In exactly the same manner one can also explain the semipermeabilit~ j of precipitation membranes . ... Whether any substance, other I than water. can pass through the membrane depends onl);' ). upori its solubility in the membrane (the third liquid layer)." T~ idea of the £unctjon of the membrane in narcosis ~as held by Oyerton. 2 His assumptions were later contradicted by J. Traube. 2 Likewise, Barlow, 3 ~orking with gutta percha membranes and solutions of lithium chloride and campho.r, decided "that for osmotic pressure to show itself, the membrane must be able to di~solve the solvent and have a distinct 'attraction of solution' for it." 4 Wilcox concluded that experiments in osmosis are merely distribution . experiments, thereby reiterating the views expressed so~e fifty years previous by Schumacher 5 who 's ai4 (p. 338), "If we examine somewhat more closely the processes of endosmosis we can easily persuade ourselves that they arc nothing more than distribution phenomena (Mischurigserscheinungen)." Also in his statement on p. 339, that in · order to be miscible the substan<:es must possess a certain degr,ee of chemical attraction, Schumacher anticipates a somewhat ' modified form of the solution theory which has recently been proposed by Kahlenber~ 2 who has described many experiments in support of his ideas. Kahlenberg assumes not 1 Zeit. phys. Chcm., xo, 255 (1892). ' Loc. cit. 3 l'hil. Mug., (G) II, 595 (1906). j Jour. Phys. Chem., 14, 576 (1910). 6 l'ol'l'· A11lr.; uo, 337 (1.86o).
•
merely a solution of the liquids in the membrane but a , loose chemical combination. He says, "In this paper it has been shown that whether osmosis will take place in a given case or not depends. upon the specific nature of the septum and the liquids that · bathe it; and if osmosis does occur, these , 1 faCtors also determine the direction of the m,ain current and the magnitude of the pressure developed, The motive force in osmotic processes lies in the specific attractions or affinities between the liquids used, and also those between the latter and the septum employed. These attractions or affinities . have also at times been termed the potential energy of solution, etc. They are 't o 'the ·mind of the writer essentially the same as what is commonly termed chemical affinity. " In this connection Nelson 1 says: "This view (i. e., ' The Chemical or Selective Theory') of the nature of osmosis carries with it the ~ssumption of a labile compound formeq. between the membrane and the solvent on the one hand and between the solvent and the solute on the other, and ascribes the movement of liquid into the osmotic cell by assuming that the solute robs the membrane of its imbibed solvent, the · · ···· membrane in turn 'imbibing more solvent which is again "';• extracted by the solute, this process being r~peated until · ' equilibrium is established . . . , It is only by assuming a chemical attraction and loose chemical combination that these results · become at all intelligible. . The mechanism of the formation of such a labile chemical compound .bet ween the membrane and the solvent on the one hand, and the solvent and the solute on the other, cannot be conceived of simply as a 'solution' of solvent and solute in the membrane when that word merely implies an. orderly filling of the membranous interstices or intermolecular spaces by solvent or solute by virtue of capillary attraction.'' While Kahlenberg was guided in his work by L'hermite' s experiments, most of his investigations were conducted with rubber membranes and organic solvents. His methods and so~e of his conclusions were subj e~ted 1
Jour. Am. Chern. Soc., 35, 658 ~I9tJ ). \
.. ITO
5. L. Bigelow an.d C. 5.' Robinson. :
to severe criticism by Cohen and Commelin 1 who however were interested in his data more .from the standpoint of its ·r elation to van't Hoff's theories of osmosis than from his ideas of the action of the membrane. Alth,Qugh the prece~g brief review of.the tbeori~con cer._ning the fllpCtjaps gf the mewbraJ4i iR QSnaatic pressu;ce Jlhenomena is not com~te, . some details ·.and minor variations having been omitted, it nevertheless gives concise statements of these theories and affords an idea of the mai nes a w 1c ave eve o A careful study of the subject leads one to agree with Fitzgeral~ that "the theory of semipermeable diaphragms is in a very doubtful state." In recent years serious consideration of the purely mechanical sieve theory seems to have been eliminated and the field left to the various explanations based upon solutions or chemical attract~on and capillarity. A more detailed, critical discussion of these two will be taken up later. Suffice it to say here that while most of the results obtained by Kahlenberg were perhaps predicable, a priori, the assertion that they might serve as a basis for a general theory of the function of all semipermeable membranes appeared doubtful. · The exact nature of the "labile compound". between the solvent and the ,, 'membrane · seemed ·to be anything but . definite, and the assumption that a chemical combination between membrane and solvent was. always a requisite for osmosis seemed hardly to be justified by the dq.ta presented. In ·order to test this assumption, the experiments described in this paper were carried out, the intention being to obtain osmotic effects un'd.~r such conditions as to preclude the possibility of chemical_ reaction between the membrane and the solvent . . While ' it 'may be unsafe, in view of . the various new types of compounds which modern chemistry has either discovered or invented to harmonize with its theories, to claim that no chemical action can take place between ,,!, water and gold or carbon under ordinary circunl.stance!;), it must still ,be considered as a rather remote possibility.
----------. . 1
'
Zeit. phys. Cllcm ., 64, 1 ( 1908) .
Experime11ts on the M anijestation of Osmotic P ressure
I I I
Materials The materials used in the preparation of the membranes were selected . with the object of securing substances · which would be as inert as possible toward the liquids with which . the membrane W01fld come in contact, since, as has. already .: been stated, the object of the work was to obtain osmotic effects under conditions which should preclude as far as possible any chemical reaction between the membrane and cell contents. For the same reason it was necessary that the materials should be exceptionally pure. Consequently all sub,·:. stances used in preparing membranes were carefully purified 1 '' before use. · · · · The following materials were chosen from which t o make membranes: silica, carbon (graphite and amorphous), metallic copper, metallic silver and metallic gold. They were all in a I . . . state of fine powders with the exception of- one sample of graphite and the gold, which were in the .form of small flakes or plates. For use they were pressed into discs in the cells, these discs constituting the membranes.
Apparatu s .
Owing to
th~· ·p~~uliar
physical nature of the materials,
i. e., the fact' that they were in the form of powders, special apparatus and technique had to be developed for the construction of the membranes. The type of cell finally adopted was patterned after that used by Bigelow and Bartelll in working with unglazed porcelain. It consisted of the following parts: (Fig. 1 .) A base (A) for holding the membrane; a plate (B) covering the open side of the base to hold the membrane in place; a second plate (C) for holding the glass parts of the cell against the base; an attachment (D) . for U!;)e in connecting the cell t o the pressure tank and gauge ,..,i _., .,for measuring pore diameters and saturating the membranes, threaded rods G, G, G, and nuts for holdirig the various parts together. The base and plate ' (B) were made of manganese bronze and all other metal parts of ·yellow brass. A ~-cell 1
J our. Am. Chem. Soc., 31, 1194 .(1909). '-
•..:·+-' ,...1 .~ ' ··- ,,..,
•
~-~ .. ~ S>tx--
• 1576
~
S. L.\W R I!NCE B IG ELO W .\ND ADELAIDE GEMBERLING
\5'1~- \~ \q-o')
_,., '"
~~ .1-\'S.·~· f'~· -r.
1577
COLLODION MEMBRANES
drogen ion concen tration and the rate of mutarotation in sol uti ons of such acid strength that the influence of th e hydrox y l ions upon the rate can be neglected .
and abandoned their use. \V. Schumacher• had more success with them and devised a method for making sma ll closed tubes or sacs of collodi on with which he carried out a ,·a ri e t y of experi men ts. But these were not numerou s enough nor methodical eno ugh to serve as the basis for any important conclusions n:garding osmosis o r th e actio n of membranes. An article byLJ]1 rattt:f7k"Vcnnta in s m or e infor mation about th e cl~ c~ and pl!):..Siod p roperties of collodion mem bran es th an any other which ~e haye os r et fo~l. IIe made collodion membranes, 5 em. in diameter, in cells huilt up with filter paper ou a g lass plate. He obser\'ed, that, in order to make th ese membran es permeable to wa t er, they must be put in water before the solvents have completely evaporated. Tl~ ether of the ~oh·e nt e \·aporates quickly. and soon the membrane contaius a lmost exclusi ,·ely alcohol, this alcohol is replaced by water , attd th e res ultin g membrane is permeable to th e latter. Bu t if the m e mbrane dries, it becomes hard and brittle, and when pu t in w:-tter ab,.,o rbs on ly an in significant quantity and re main s imper\'ions to th at soh·ent. \ Ve haYe thu s, with collod ion, a method for obtaining membran es, otherwise alike, but differing in pe rmeability. The \\' a y in \\·hich the g r adual diminution in the size of the inter,tices in collodion may be stopptd by plunging- it into water, may be likened to the arresting of the deyeJopment of a photographic plate by washing and dipping into a fixing bath. A collodion membrane, under n·ater, retains for a good \\·hile the parttcular den~ i tv or structure \Yhi c h it happ .:! ns to IHt\·e reach ed through evnpora tion of the soh·e nts befort: immer!:-i un; after se,·eml \Yeek~. how e n~ r. it
2. The conclu~ion of Osaka that the r ate of mutarotatio n is proportional to the square root of the hydrogen ion concentration holds fairly n·ell between the concentrations o.or too . ro molal but does not agree at all well with measurements outside this region of concentration. 3· Using Osaka's val ues fo\ the rate in alkaline so lution and new values for it in acid solutions, it 1s found that the following formula expresses accurately the rate of mutarotation of g lu cose at 25° ~n pure water and in acid and alkali ne solutions. Rate - 0.0096 + o.258(H·) -t 975o(OH'). {. Hydro,yl IOns are nearly forty thou;.:and ti mes :,tronger catal) zing agents of the mutarotation of glucose than hydrogen ions.
5· This stronger catalyzing action of the hydroxyl ions ca uses a lower rate of mutarotation in weakly acid solutions than is obsen-ed for pure water. Thi~ depression of the rate, or "uegative cata lysis'', has been measured in a o.oo r molal hydrochloric acid solution and found to be in c lose agreement \\'ith the predic tions of the aboye formula. The measurements recorded in this article n·ere made po"!>ible by the kindne"s of Professor Geo. A . Hulett, who allo\\'ed th e author the use of of his \\'ell equipped laboratory at Princeton Uui ,·ersity , ·,·wpiJJt
x, \\''-;
\·irgi11L1
-----J:y
s
COLLODION MEMBRANES. L -"'''-KJ:.:-;cE ntcEr...ow .-\XJ> .i,D LL\ rnr: Cg!\rn•
H.eceived Sept. _;,
Bibliographical.-Comp.tri~on
contract-; somen·hat. Like S chum acher, Baranetzky n·orked exclusi,·ely \\' :t h solu ti ons and was particularly interested in sho\Ying that Jolly's "endosmotic equiYalent"" \\'aS of no great significance, and in substantiat ing Briicke's theory' regardinrr th e assage of substances throu His resu ts show that pyroxylin membranes are distinguished from others , such as animal membranes , parchment pape r, and cdlu lo,-e. hy a less permeability for dissolYed ~alts. Neither Graham" nor Pfeffetn made u se of collodion membranes; they haye been overlooked or ueg lected by chemists and phYsicists since 1872 . R efere nces to them are more numerous in medical lit~:rature. Intere:,t-
f{ r.t:"\ 1;
Jqo -
Contents.
of rlialysi' through collodion, parcltmL·ttt pap,· t· an d ~old heater's ~ktn.-The quantity of water which passe, through collo,Jion a t a deliuitL' temperature and pre"ure.-Pressure nnd te mpemture coeti'ciutt~.-The l'ITcet of thiekne~~ on the permeabi lity. The pcrmeabilitte~ ui difTercnt samples of collodion.-The effec t o f ag-e on the pennea hility. Sumutary.
Th e cot t\·cnience and u~efulne~s o f collodion ntembranes are uot, at the preseut time, fully apprecinted by chemists and physicists. This article ·outains n brief bibliography, methods for making the~e membr a ne:, ami n accouu t of on r ex peri men t.;..
.-,(e,u '1)
1 Uebe r ::\Iembra udiiTusiou. Pogg. :\n ual., uo, 337·3iU ( t86o ). ' Iliosmotische Culer:-.uchungen . L~Cp ~-' ~ 3 Jolly. Ph. Experimeutal-Culersuclntngen l'eber End osuwse.
.::m·' . .
l'ogg. Annal.,
78, 261-271 ( r8~ 9).
1 E. Briicke. Beitriige zur Lc ltre \'Oil der Diffusion tr• ph~ · r 'is!;igcr Kqrper durch poriiFe Schncidewall(le. Pogg. Anual .. 77·~4 (11'~3 ."Z, _:. o On Osmotic Force. Phil. Trans., 144, 177-228 ( tSq ). Liqt < Diffusion appli e
@
•
e
>I
1
' \"\(~
1-/ ,'1/" '•
17+
5. L. Bigelow and C. S. Robinson
Bearing of the Results on the Cause of Osmosis an.d the .. Function of the Semipermeable Membrane
'1f:.--_
· ~ In
the review of the literature given in the first pages of . this article the theories concerning the cause of osmosis may be divided into comparatively few classes, namely those attributing osmosis to. I. A difference in the surface tension of solvent and solution. , 2. A difference in the solubility of the solvent and the solution in the membrane. · _3 : An attraction of the solute molecules for the solvent molecules: .,I 4· The so-called kinetic theory. 5· A difference in the vapor pressures of the solvent and solution. 6. A chemical reaction between the membrane, solvent and solution. 7·. · A difference in electrical 'potential between the op- posite faces of the membrane. The advocates of these theorie.s and their various modi. fications advanced them as applying to osmotic phenomena ·,= in general and, if later, certain exceptions were found for - any ~-·, 1 ' ~me theory, that theory was thrown into more or less disrepute. - ·J· : '"'1i;'"'' The' main endeavor of investigators has been ' to find one I theory to explain omosis, by osmosis being meant the flow of one liquid through a membrane permeable for it, into another liquid, for which the membrane is less permeable, until a certain difference in hydrostatic pressure is produced on opposite sides of the membrane. yet the conditions under which "osmosis" takes place differ greatly in different experiments. · Thus, membranes range from a layer of liquid to a layer · of solid such as metallic gold or unglazed porcelain and include ~·. 1·ubber, parchment and various other 'animal, vegetable and ·inorganic membranes. The liquids used in producing osmosis vary . as widely in character. When these things are considered. it seems improbable that the phenomena in question should all be traceable to a single cause. It is more probable
•
£.~
\7
.
I
Membranes of Chemically Inert Maten:als· IJ
175
that we are dealing with several which manifest themselves in the same way, i.e., in causing a transference of liquid through the membrane. Nor is this contrary to experience, In a "three-liquid layer" experiment for instance, the fo! ce is un-. doubtedly the difference in the "attraction of solution" between the liquids involved. The corresponding result may be brought abot1t with a porous wall and two solutions under . the influence of an electric current, · i. e., in electric osmose. ·. · . Yet in the latter case we have no "attraction of solution" between the liquids and the membrane. In the case of sugar solution and water separated by a membrane of .tmglazed porcelain we have but little evidence of either of these forces being operative. The difference between ordinary osmosis and electrical osmose is that in the former the energy is derived from the system itself while in the latter it comes from an external source. That this same kind of energy may aJso be developed internally for the production of strictly analogous phonomena is fairly evident ' ~rom Bartell's work on "negative osriwse." 1 ' If we grant then that " attraction of solut ion" may produce osmosis in a three-liquid layer experiment and difference in electrical potential may produce it in either electrical osmose or "negative osmose," we have differentiated "osmose" into at .· least two phenomena and, unless we conceive of. " solution" as the filling of membrane pores of capillary. dimensions with the liquid, we must a-pparently admit the existence of still a third class of osmotic phenomena obtainable with aqueous solutions of sugar, water and membranes of the inert materials used in the present work. Many of the various theories advanced appear plausible under certain circumstances and the fact that they are not always so should .not cause their complete rejection. For instance, the relationships between sllrface tension and osmotic effect pointed out .bY Traube 2 seem too consistent to be attributed to coincidence and their significance should not be entirely denied because Barlow 3 Jour. Am. Chem. Soc., 36, 646 (xgx4). Mag., (6) 8, 704 (x 904). 3 Ibid., (6) I0 1 I (1905).
1
~Phil.
\
.
.. I
e
176
5. L. Bigelow and C. 5 . .Robinson
· M em.branes of Chemically Inert 1\.faterials II
later found them incapable of universal application. It is quite · conceivable that difference in surface tension should under some circumstances produce liquid movements across · .diaphragm but it is ·quite unnecessary to conclude therefrom . that all such movements are due to this cause. It is equally ,· ·. absurd to postulate that since this explanation is not always tenable it is never so. It seems fairly obvious therefore that osmosis may be due not to one but to a variety of causes. Nevertheless, several of these groups of theories which .:~· ascribe to the membrane an active role in the phenomena ,.,.,.,; of osmosis and which are apparently ~t first ;:;ight irreconcilable, !.r;:· prove to be, upon closer scrutiny, very closely related to each · ...::;ther.
a
.' tfK.
According to these theories osmosis takes place : I. Through capillary spaces in the membrane, 2. By the solution of the solvent in the membrane, and · 3. By the formation of a labile chemical compound between the membrane and the solvent. ·That osmosis can take place through capillaries, as distin. guished from molecular interstices, seems probable from the fact : •· ·· that it has been observed to occur when two liquids ?-re in .con- ·:·., tact only through fine cracks in glass .tubes, 1 through the pores in unglazed porcelain, through the various membranes described by 2 Tinker and through those used in the experiments reported in the present article. The limits of these pore diameters between . which osmosis can take place are not definitely known but Bi.gelow and BartelP and Bartel1 4 have shown that wiili un' various materials the upper glazed porcelain clogged wit!J. limit seems to. be about . . o.9 mkron,. a figure much too large to represent the . dimensiops of molecular interstices. This work has been corroborated by Tinker 5 and is in agreement 1
See in this connection, Fischer: Pogg. Ann., 10 481 .(182 7); Dutrochet: Ann. Chim. Phys., 35, 383 (1827); 49, 411 (1832); Brucke: Pogg. Ann., 58 77 0
(1843).
'
1
2
Loc. Ctt. 3 Jour. Am. Chern. Soc., 31, 1194 (1909). 4 Jour. Phys. Chern., 15, 318 (19n); Dissertation, Uuiv. of Mich., 6 Loc. cit. '
-
1910,
177
with that of Draper 1 on osmosis as applied to gases. That the lower liinit may be of the dimensions of molecular inter., . spaces is equally certain from the results of various investigators who have used th~ three-liquid layer arrangement, and in these ... . cases t)J.ere is certainly a solution of the liquids in the membrane. The validity of the last of the three suggestions, namely the formation of a labile chemical compound between . the ,~ ,,. . · membrane and the solvent, has not been so certainly demon( •····M~· .·strated. It is however a possibility as will appear later. : , . ' · A~ has been pointed out by Bigelow, 2 these views have always been held to be "mutually exclusive, " although this was not at all necessary and in reality it is possible to bring them all into harmony. He showed that "the rate of passage ; of liquids ·. through molecular interstices is expressible by the same laws which formulate the . rate of passage of liquids through capillary· tubes," thus affording experimental proof that there is pp essential difference between ·capillaries and molecular interstices in . this respect. Theoretically, interspaces of all dhriensions from macroscopic capillaries down to ., infinitely _small .. interstices are capable of existence. All , \:· phenomena which · result from forces inherent · in such interspaces should also manifest themselves and furthermore in a continuous 'fashion except as they may be influenced by certain secondary factors. There is really no basis for assuming that between the pores of solid membranes of capillary structure on the one hand and the molecular interspaces of liquid membranes on the other there exists a region of osmotically inactive spaces. It is far more logical to suppose that osmotic phenomena with the membranes are but extremes of one and · ·the same process. From this it follows that there is no s_harp l.1 •;;: line . between the . changes occurring in ordinary capillaries. · and in intermolecular·interstices. . Furthermore, it is· entirely conceivable that in the latter category may be included the formation of certain chemical compounds such, for instance, as certain so-called molecular 1
'
''
•
I Jour. Franklin Inst., 17, 177 (1827). z Jour. Am. Chem. Soc., 29, 1690 ( 190i) -
• •
.,
