Aust. J. Chern.

, 1980,33, 1857-61

Diffusion Measurements in Aqueous

L-Ascorbic Acid Solutions

Mohammed Shamim and Salah M. A. Baki

School of Chemical Sciences, Universiti Sains Malaysia,
Penang, Malaysia.

Abstvact
Diffusion coefficients in aqueous 0.05-1.0 M L-ascorbic acid solutions have been measured at
298.15 K by using a modified version of the Stokes diaphragm cell. The diffusion coefficient of
un-ionized ascorbic acid molecule at zero concentration has been found to be 1.055 x m2 s-l.
The diffusion coefficient of an isolated ascorbate ion, CSH706-, is about 23 % lower than the
molecular form; this indicates, as expected, that ions strongly interact with solvent molecules.

Introduction
Physicochemical aspects of many strong electrolytes have been studied in detail.
Such data on weak electrolytes are rather meagre. L-Ascorbic acid (vitamin C)-
otherwise very well investigated-is one of them. The importance of the compound
prompted this work which reports the results of diffusion measurements in aqueous
solutions of this acid at 298.15 K.

Experimental
Matevials
Deionized water was distilled twice to obtain conductivity water. The specific conductance of
water was always less than S cm-l.
AR grade KC1 was recrystallized three times from water. The salt was dried for 36 h at 698.15 K
in a temperature-controlled oven1 and stored in a vacuum desiccator.
Analytical grade AgNO,, KOH and HC1 were used without any further purification.
Recrystallized crystals of L-ascorbic acid turned brown on drying because of oxidation in the
presence of air and light." Washing the wet crystals with alcohol and drying them in a vacuum
desiccator did not prevent oxidation. Therefore analytical grade L-ascorbic, dried over silica gel,
was used without any purification. The dried sample melted3 at 490.15 K with decomposition.
Weighed potentiometric titrations against carbonate-free KOH indicated that the acid was 99.9%
pure.

' Scatchard, G., Jones, P. R., and Prenties, S. S., J. Am.

Chem. Soc., 1932, 54, 2676.
" Oser, B. L., 'Hawk's Physiological Chemistry' 14th Edn, p. 697 (McGraw-Hill:
New York 1968).
'Handbook of Chemistry and Physics' 55th Edn, p. C-123 (Chemical Rubber Co.: Cleveland,
Ohio, 1974-75).
 Short Communications

Preparation of Solutions
Potassium chloride and silver nitrate solutions were prepared by dissolving a weighed quantity
of the salt in conductivity water. The solutions were prepared in 500-cm3calibrated flasks and were
made up at 298.15 K. The observed weights were corrected for the air buoyancy. HC1 solutions
were prepared by diluting the concentrated acid, and carbonate-free KOH solution was prepared,
stored and used in the conventional way. Exact concentrations of these solutions were determined
by potentiometric titration.
A series of ascorbic acid solutions, ranging from OS005to 2.0 M, were prepared and left on a
shelf unprotected from light. The 2 M solution became light brown within few hours while others took
12 h to a week. A similar set of solutions were prepared from degassed and nitrogen-flushed (filtered
and dried) conductivity water and were put to a similar test. No colour change was observed after
up to five days. The 2 M solution developed a light brown colour on the sixth day but the others
stayed colourless for weeks. It was thus concluded that treatment of the conductivity water did
avert (or at least retard) oxidation of the acid. Accordingly, all ascorbic acid solutions were prepared
from degassed and nitrogen-treated water in 500-cm3flasks whose outer surfaces were painted black.
Solutions were prepared by dissolving a weighed quantity of the acid and were made up at 298.15 K.
The weights, as usual, were corrected for the density of air.

The Diffusion Cell

The rotating diaphragm cell, developed by Wendt and Shamim: was used in the present work.
The volumes VA, VB and VDof the upper compartment A, lower compartment B and that of the
porous disc were respectively 2.8, 2.9 and 0.30 cm3. The effective length of the diffusion path L
was c. 0.2 cm. The pre-diffusion time, t,, required to establish steady state in the porous disc, was
estimated from the equation5
t, > 1.2LZ/D
It has been recently shown that Gordon's equation gives a t, value which is considerably longer than
that necessary to attain steady state.=
Experimental procedures for calibration and diffusion runs were the same as described in ref?
The thermostat tank was set at 298.15 5 0.02 K and controlled to 50.002 K. With ascorbic acid
diffusion runs, the cell was kept covered with aluminium foil to protect the solutions from light.
Analysis of the Solutions
The initial and final concentrations of the solution from a diffusion experiment were determined
potenti~metrically.~ A calibrated microburette and a micrometer syringe were used to read the
volumes in question. Carbonate-free KOH solution was used as a titrant for hydrochloric and
ascorbic acid solutions. Titrations with ascorbic acid solutions required special care. Accordingly,
a special titration cell was designed which was fitted with a tight Perspex lid having appropriate
openings for the combination electrode, burette tip and a glass tubing to pass a slow stream of dried
and filtered nitrogen. The titration cell, covered with aluminium foil, was placed into a constant-
temperature oil bath and the contents were continuously stirred by a magnetic stirrer. Duplicate
titrations agreed within 0.1 %.

Calibration of Diffusion Cell

Calibration diffusion experiments were carried out with KC1 solutions and the cell constant /3
was calculated from the diaphragm cell equations

Wendt, R. P., and Shamim, M., J. Phys. Chem., 1970,742770.

Gorden, A. R., Ann. N.Y. Acad. Sci., 1945, 46, 285.
Mills, R., and Woolf, L. A,, 'The Diaphragm Cell' p. 20 (Research School of Physical Sciences,
Australian National University: Canberra 1968).
Albert, A,, and Serjeant, E. P., 'The Determination of Ionization Constant' p. 13 (Chapman &Hall:
London 1971).
* Robinson, R. A., and Stokes, R. H., 'Electrolyte Solutions' 2nd Edn, pp. 7 4 8 6 (Butterworths:
London 1959).
Short Communications

in which t is the duration of the diffusion experiment in seconds, and

where COB,C o Aand C t Band C t Aare the initial and final concentrations respectively of the bottom
compartment B and upper compartment A. 4 is the calculated value of integral diffusion coefficient
given by the equation
D -
= [DO(CB)-(CA/CB)4O(L)]/[l (CA/C)I (a
where

DO(C)= ( 1 / C )
S: DdC

The values of DO(CB)and BO(CA)in equation (2) were obtained by constructing plots of Do against
C from the literature diffusion data?

Table 1. Diffusion coefficients in aqueous ascorbic acid solutions

Concentrations C in mol dm-3, diffusion coefficients D in m2 s-l

A Nernst limiting value.

From conductance work.
' Calculated from equations
K = [Caz/(l-a)] f k 2 log f* = -A J(Ca)/{l S B aJ(Ca)}

Diffusion Runs with Ascoubic Acid Solutions

From the known time t and the initial and final concentrations, the integral diffusion coefficients4
were calculated by means of equation (1). The reproducibility was 5 0 . 0 0 2 ~lo-' in 4 values.
The integral diffusion coefficients were converted into differential diffusion coefficients D by means
of a short series of approximation^'^ for solving equation (2) and the equation

The final DO(CB)values and differential diffusion coefficients D are given in Table 1.
Woolf, L. A,, and Tilley, P. J., J. Phys. Chem., 1967, 71, 1962.
Stokes, R. H., J. Am. Chem. Soc., 1950, 72, 2243.
 Short Communications

Diffusion Coeficients of Molecular Ascorbic Acid Dm

Vitagliano and Lyonsii and Muller and Stokes12 have derived equations for the diffusion coeffi-
cients of undissociated molecules in weak electrolyte solutions. We have used the equation1'

to deduce Dmo-the diffusion coefficient of molecular species at infinite dilution. The values for
degree of dissociation sc, in the concentration range 0.05-0.30 M, were obtained from our own
conductance work and at higher concentrations these were ~alculated.'~The relative viscosities
were taken from Shamim and Khoo,13 and D: was calculated from the Nernst equation

= 30.6 S cm2 mol-l. The D,,' so obtained are given in Table 1.

with ??H+ = 349.8 and iboH,,,-

Results and Discussion

The rotating diaphragm cell seems to be the simplest technique for measuring
the diffusion coefficient. Because the cell does not have any moving parts it requires
no recalibrations and maintenance. The cell constant /3 was found to be 1 e2529 cm2
with a mean and standard deviations of 0.0015 and 0.0102 respectively. Diffusion
runs with HCI and NaCl solution^^^,^^ were reproducible and consistent with the
literature values within about t- 0.3 %.
A plot of Dmragainst C, on extrapolation to zero concentration, led to the limiting
diffusion coefficient for the molecular form

This value is considerably larger than the diffusion coefficient Do of an isolated

ascorbate ion (C,H,O,-) calculated at the same temperature from the equation

This indicates, as expected, that the negatively charged ascorbate ion strongly interacts
with water molecules and thus suffers a decrease of about 23 % in its mobility as
compared to the neutral molecule. The observation is consistent with the viscosity
measurements13 carried out at the same temperature. Acetate ions and undissociated
acetic acid molecules show a similar behaviour.ll
The limiting value Dm0 of the ascorbic acid molecules given above is based on
equation (7) which does not take into consideration the thermodynamic non-ideality
of the molecules. Evidently activity data, which are not available as yet, are needed
to obtain a better value of Dm0. It is to be noted, however, that such a step of refine-
ment made no change in the Dm0 value of citric acid m ~ l e c u l e s . ~ ~
Recently, Schneeberger, Stahl and Loncin" have reported diffusion coefficients in
0.07 and 0.10 M L-ascorbic acid solutions using the diaphragm cell method. The D
value at 298.15 K does not show, contrary to expectation, any concentration
dependence and is 0.84 x lo-' m2 s-l at both concentrations. Their value is c. 25 %
lower than ours at the corresponding concentrations.

l1 Vitagliano, V., and Lyons, P. A,, J. Am. Chem. Soc., 1956, 78, 1549.
Muller, G. T. A., and Stokes, R. H., Trans. Faraday Soc., 1957, 58, 642.
l 3 Shamim, M., and Khoo, S. B., Ausf. J. Chem., 1979, 32, 2293.
l 4 Robinson, R. A., and Stokes, R. H., 'Electrolyte Solutions' 2nd Edn, Appendix 11.2, p. 515
(Butterworths: London 1959).
l 5 Schneeberger, R., Stahl, R., and Loncin, M., Lebensm-Wiss, Technol., 1975, 8, 222.
Short Communications

There are two reasons for this discrepancy. Firstly, the authors did not have
sufficient experimental data to obtain accurate BO(C,)values for converting1' integral
diffusion coefficients D into D. Secondly, they did not exclude air from the solutions
and did not protect them from light. In the presence of air and light, aqueous solutions
of L-ascorbic acid readily oxidized into dehydroascorbic acid, and the aged solutions
undergo further oxidation into oxalic and theronic a ~ i d s . ~ ,From
' ~ the dimensions
of their cell (small p value) we gather that it would have taken them 5-6 days to
complete a run. Thus the solutions inside the cell well might have contained both
L-ascorbic acid and the product(s) of oxidation. Consequently the theory and equations
of a binary system cannot be used to interpret the experimental data of a ternary or
a multicomponent system. Under the circumstances it is difficult to assign any
physical meaning to the reported values.

Manuscript received 23 January 1980

Kiss, L., and Scholz, T., Acta Pharm. Hung., 1967, 37(6), 254.