4x30
HAZIME SIIIIO,TOSIIIO OGaw,i
AND
HIROSIIIYOSIIIIIASIII
Vol. 77
+
gases are present, the reaction Ag2N202 = 2Ag 2N0, which would introduce large additionala mounts of uitric oxide in the system, containing AgzO, probably does not occur.
although I is the primary change, the reaction apparently stops as soon as AgzO and AgNOz (optimum) are present and starts again when these apConclusion proach a limiting concentration. Thus, the deThe primary change in the decomposition of sil- composition oE silver hyponitrite seems to proceed by the propelling action of silver oxide formed, a t ver hyponitrite appears to be intervals, during the decomposition. Ag2Xz0, = Agio NrO (1) That I is not accompanied by AgzNzOz = 2Ag Apparently this reaction is not acconipanied by 2NO is supported by the following observations : any other change, such as ,%geNPOp = 2,lg 2N0. (a) nitric oxide destroys ,4g20 rapidly with the Reaction I is instantly followed by formation of nitrogen dioxide, nitrite and nitrate; (b) silver oxide oxidizes silver hyponitrite even at Ag2S202 + 2Ag20 = 2AgX02 + 4Ag (11) The forniation of AgNOz in this way, at 160" and 150" while the oxides of nitrogen do not do so; and above, is followed by its decomposition (above (c) when large quantities of silver nitrite are pres12s")producing, ultimately, AgNO,, 4 g and oxides ent in the system containing silver hyponitrite, of nitrogen. Nitrate is formed from nitrite and silver oxide is not found. The forniation of nitrogen is not explained. It is not directly; it may be produced, to some extent, found whenever silver oxide is present or produced by AgzO NO2 = AgN03 + Ag. As nitric oxide is formed it reacts with nitrate to regenerate nitrite in contact with nitric oxide or oxides of nitrogen. As nitric oxide and nitrogen dioxide are both present to some extent, AgN03 NO AgNOz NOz. The decomposition of silver nitrite, which fol- in the system, it is not likely that the reactionR NO = AgNO3 1/2N2occurs to any aplows reactions I and 11, retards the decomposition XgNO, of silver hyponitrite by reaction I . The formation preciable extent. Thanks are due to Dr. V. T. Oza for help in of nitrite from the hyponitrite also reduces thc amount of hyponitrite decomposing by I and ex- blowing thc systems and for gas analyses. plains low percentage of nitrous oxide in the gas; AHMEDABAD, INDIA
+
+
+
+
+
[CONTRIBUTION FROM
+
THE
+
+
CHEMICAL INSTITUTE, FACULTY OF SCIENCE, NAGOYA UNIVERSI~Y]
Measurement of the Amount of Bound Water by Ultrasonic Interferometer BY HAZIME SI-IIIO, TOSHIO OGAWAAND HIROSIIIYOSIIIHASHI RECEIVEDDECEMUER 13, 1934 A general theory which enables us to evaluate the amount of the bound water of non-electrolyte and of high polymer in solution from ultrasonic velocity measurements has been introduced, and applied for glucose, maltose and dextrin solutions T h e ultrasonic velocity was meawred by an interferometer making use of X-cut crystal of resonance frequency 1 mc. T h e measurements were made a t 20.0". The amount of bound water was found to be 0.43, 0.23, 0.40 cc./g. of glucose, maltose and dextrin, respectively.
Introduction It is a well-known fact that, in solutiolis of electrolytes, the water molecules bound to an electrolyte ion are compressed owing to the strong electric field of the ion to form a very hard structure. Passinski,' U'ada2 a i d &saki3 evaluated the degree of hydration of ions from adiabatic compressibility measuremcnts. They assumed the compressibility of bound water and of the ion itself to be both negligibly small. But the applicability of this method for the evaluation of non-electrolytes or high polymers in aqueous solutions is questionable, tht. bound water of non-electrolytes or the molecules of high polymers being reasonably supposed to be of appreciable compressibility. We have deduced a more general formula taking account of the compressibilities of bound water and of the solute particle, and have applied it to the solutions of sugars as well as high polymers and estimated the amount of bound water. Theoretical If Mgrams of solute is dissolved in Vi cc. of sol( 1 ) A . Passinski, A c l a Physicoclztri~i,U . R . S . S . . 8 , 385 ( 1 9 3 8 ) . (2) Y . IVada, ilpplied P h y s . ( J a p a u ) . 17, 237 (1048). ( 3 ) T Yasunajia a n d T . Sasaki, J . Cliern. Sor. J I I ~ I I P!cvr. II. .5 I < , 72, 3fXi (19fil),
( t i c iii
vent, somc of the solvent, say & cc., will be attached to the solute and compressed to be v2' cc., and, consequently, there will result a solution of volume V' cc. Liccordinglyit follows that V' iii
V:
=
v,:+ v: + -
(1)
L;
which =
d l / d l = volume of solute in V' ce whi dl = true denbity of 5olute i n ioln (wlv,itiuii cffeet being not taken into account)
Differentiating (1) with respect to the pressure, P d V' -
tl V; ilP- - di.
(tv:
(11;
d P
tlP
+
+
- tlv:
dP
Let 8, PO, and B2 represent the adiabatic COIIIpressibility of solution, solvent, solute and bound water, respectively. From definition these 6 ' s are given by
AMOUNT OF ~ V A T E R BOUNDBY NUN-ELECTROLYTES
Oct. 5 , 19-55
TABLE I
where
v, = v:/vt, vo = v;/vr,v 2
= 'd:/vf. vo = vA/V'
Vo
(d
- c)/do
(3)
in which d = density of solution c = concn. of solute in g./cc. s o h . d o = normal solvent density
From (2), we obtain the formula A
= ?/Pa - vo
= V,P,/?O
- (V"
-
V2??/?0)
(4)
And per 1 g. solute
= (?/Po
- V")/L = !%/?a
x
l/dl ('do
- v,.?2/?o)/c
(5)
P, Po can easily be calculated from the results of measurements of the sonic velocities in solutions and solvents, and VOfrom ( 3 ) through measurements of the concentration c together with densities d and do; therefore, A/c may be determined experimentally. The first term of the right side of ( 5 ) is due to the compressibility of solute PI, and the second represents the term due to hydration. Now, as ethanol, which acts as a precipitating agent for the solutes described above, is added to the solution, the second term, which is due to hydration, decreases g r a d ~ a l l y . ~ The ? ~ compressibility of the solvent PO, which is, in the present case, a binary mixture of ethanol and water in various proportions, varies with ethanol In order to correct this effect, both sides of ( 5 ) are multiplied by Po/&; then, it follows that K
=
A/c X Po/?,
=
P1/Pw
EXAMPLE OF
THE
MEASUREMENTS (GLUCOSESOLUTIONS) K
Alcohol,
and there exists a relation for VO
A/C
498 1
X l/di ( v Q ? Q / ? ~- ~ 3 2 / ! 3 W ) / c
(6)
(4/c
x
d do vo @/Bo 5% c ( g . / c c . ) Bo/Bw) 0 . 0 0.0588 0.9982 1.0206 0.9635 0.9486 -0.253 ,9582 - ,128 ,9663 10.2 .0552 ,9814 1.0035 ,9567 - ,117 .9791 1.0005 ,9639 12.0 ,0568 ,9657 ,9660 ,005 .9694 0.9922 1 9 . 3 ,0560 ,9636 9649 ,019 ,9602 ,9837 2 7 . 0 ,0584 ,9641 ,9752 ,017 .9432 ,9670 3 4 . 2 ,0576 ,9642 ,9654 ,020 ,9324 ,9564 4 2 . 3 ,0574 ,9656 ,9656 ,000 .9222 ,9456 47.2 ,0551 ,9685 ,9682 ,006 .8987 ,9200 5 6 . 1 ,0496
The plots of K against ethanol Figs. 1, 2, 3 .
1
I
0 Fig. 1.-Effect
where Pwis the compressibility of pure water. If we assume that the hydration has been brought to naught on the point of precipitation, K,, which represents the K a t the precipitation point, is given only by the first term of ( G ) , and consequently the compressibility of solute particle PI can be found directly from this value. And if is assumed to remain constant irrespective of variation of ethanol concentration, we obtain the magnitude of the hydration term from the difference of K," and K,, which are the value of K for aqueous solution and a t the precipitation point in the alcohol-containing aqueous solution, respectively.
70are shown in
1
I -
10 20 30 40 50 Concn. of ethanol (wt. %). of ethanol concentration on K's value of glucose. 0
/
Experimental T h e ultrasonic velocity was measured by an interferometer making use of X-cut crystal of resonance frequency 1 mc. The methods of measurement were described in previous paper^.^^^ The materials used were purified by reprecipitation from G.P. or E.P. grade chemicals.
Results and Discussion Measurements were carried out with solutions of Table I glucose, maltose and dextrin a t 20.0'. shows the results obtained for glucose solutions as an illustrative example, in which Vo, Po and do are values of water-ethanol mixed solvents. (4) H. R. K r u y t and €1. G. Bungenberg de Jong, Kolloidchem. Beihcfle, 28, 1 (1920). (5) H. Shiio, J. Chem. S O L .J a p a n , Pure Chem. Sec., 74, 203 (1953). (6) Y .Miyahara, Bull. Cheln. SOC.J a p a n , 25, 326 (1952). ( 7 ) Y .Miyahara, i b i d . . 26, 380 (1953).
0 Fig. 2.-Effect
10 20 30 40 50 Concn. of ethanol (wt. yo). of ethanol concentration on K's value of maltose.
The values of K are found to increase with ethanol concentration (Figs. 1, 2 ) ; this might be due to the decrease of the hydration term in ( 6 ) . At ethanol concentrations of about 20% and more, the curve becomes horizontal, this being possibly attributed to the effect that the bound water has been removed from the solute molecules by ethanol, and if this is the case, K values would be given only by the first term of (6) having a bearing on the com-
HAZIMESHIIO,TOSIIIO OGAWA AND HIROSIIIYOSHIHASIII
4982
VOl. 77
the precipitation point. Namely our assumption is regarded as a definition of bound water. From the difference of K, and K , in Figs. 1, 2, we can obtain the value of hydration volt. These values are given in Table 11. TABLE I1 THEAMOUNTOF BOUND WATER (vo/c) cc
Solute
Glucose Maltose Dextrin
/g
0 43 23
hIole/mole
4 3 3 3
40
Now in the case of glucose, we have obtained a reasonable result that the sugar molecule possesses -0.1 1 1 1 1I for each of its free OH-radicals nearly one water molecule, but with maltose, the value is fairly 1(1 LO small, and this might be considered to be due to the Coricn of ethmol ( i t t % ) Fig 3 -Curve 1, effect of ethanol concentration on K’s decrease of free OH-radicals owing to intramolecuvalue of dextrin; curve 11, effect of ethanol concentration lar hydrogen bonding. In dextrin solution, K initially takes a negative on Kb value (only hydration part of K’s value) value, but with the addition of alcohol, it gradually pressibility of solute molecule P1. From this limiting increases toward a maximum, and finally drops off we can calculate PI using 0, and d, (curve I). In the initial increasing part the effect of value (K,), which can also be determined experimentally; the hydration term may be predominant as with gluhowever, in the case of glucose and of maltose, the cose; in the latter part in which the curve is devalues of K, thus obtained are found to be so small scending, the effect of decreasing 01 may be prevailthat they are beyond the limits of experimental ac- ing and the dextrin molecule would be considered curacy and, therefore, it would be meaningless to to be of a compressed structure as the result of ethanol addition. discuss the values of PI further. Beyond the maximum point, the curve runs Then it is clear, from this point of view, that the downward almost linearly; then we might well astreatments by Passinski,l Sasaki3 and our previous reports,8in which the compressibility of solute mole- sume that the above mentioned effects of & are Iincule is put to zero, are only of the first approxima- ear with respect to the ethanol concentration. By tion. Now, the difference of K , and K , give us the subtracting this effect of PI froin K , we obtain Kl, value of (VO - v & / P , ~ ) / c which is proportional to (curve 11) which represents the hydration effect the hydration of one gram of solute. However, it alone in K. From this curve we calculated the is more easily understood, if the numerical values of degree of hydration of dextrin as in the case of gluvO/c are given concretely and, hence, we may evalu- cose. The results are shown in Table 11. From barye-’ as ate the value vO/c on the following assumptions: curve I1 in Fig. 3 we obtain 9.2 X “IVith non-electrolytes, a strong compression as the value of the compressibility p1 in aqueous soluobserved with ions would not result, then vO = v 2 ; tion. Using the several methods of nieasurernent, variin that case, hydration may be caused by hydrogen bonds principally, then we can use without much ous authorsglo have estimated the degree of hydration of starch to be 0.8-0.4 cc./g. Our results are error the compressibility of ice (1s X lo-’? barye-]) for Pz, since their bound water is not so strongly completely in accord with their values compressed as ion (,& = 0), but not so soft as normal Conclusion water (ply = 43 X lo-’* barye-’).” (1) -4 general theory is introduced which enables “Bound water” should be defined according to the methods of measurement, and it has only a rela- us to evaluate the amount of bound water of nontive significance characteristic to each method. For electrolytes and of high polymers in solutions froin example, in the determination of the bound water ultrasonic velocity measurements. (2) The compressibility of the solute molecule from measurements of viscosity or cliffusion conof lower molecular weight may, in the first approwstant, bound water is taken as that moving together with the solute particles under ordinary conditions, mation, be regarded as being negligibly small. (3) I n high polymers, however, the compressiand in the cases from measurements of negative adsorption of low molecular weight solute (e.g., sugar, bility of the molecule itself must be taken into conCuS04, etc.), the portion of water adsorbed by the sideration. (4) The amounts of bound water are 0.43, 0.28, solute particle not accessible to the molecules of low molecular weight is defined the bound watersg So 0.40 cc /g. for glucose, maltose and dextrin, respecwe think that bound water due to hydrogen bond tively. ( 5 ) The compressibility value obtained for the principally has the compressibility of ice on averbarye-] in aqueous devtrin molecule is 9.2 X age, and t h a t amount of bound water is that of water dehydrated while precipitating agent is added to solution.
t
( 8 ) Y .Mihahara a n d H . Shiio. . I . Clicii?..cor. J n p n r t ,
73, 876 (1951). (9) R. Yo.;hinka, i h i d . , 71, 4 5 0 (1950).
pur^
C h e w . .Ye(..,
NAGOYA,JAPAN (10) 2 411 (IO$?)
V Diiman?kii a n d R V Veil ?hh
