JOURNAL OF MATERIALS SCIENCE LETTERS 16 (1997) 1945±1947
X-ray study of the in¯uence of water on the cellulose membrane
structure
H. GRIGORIEW, A. G. CHMIELEWSKI
Institute of Nuclear Chemistry and Technology, Dorodna 16, 03±195 Warsaw, Poland
In spite of its application in technology the process
of water permeation through a cellulose membrane
has not been explained fully. Hydrophilic membranes
play an important role in membrane processes.
Recently, a swollen dry-layer model has been
developed to describe permeation process through
such membranes [1]. In the process, the permeation
has been described generally as the sorption of
liquid, its diffusion, and the evaporation and
desorption of vapour. According to this model the
layer containing liquid water in the swollen mem-
brane exists on the upstream side of the membrane
and another layer with vapour water on the down-
stream side. Also an isotope effect of H±D and
18
O±
16
O in the permeation process was observed, for
the cellulose membranes Tomophan and Cuprophan
[2, 3].
Structural research on the water permeation
through membranes was carried out mainly in the
static stage concerning the structure of water during
the water±membrane contact (sorption). The research
included infrared and nuclear magnetic resonance
studies and molecular dynamics approaches [4±8].
Also a study of some parameters of the permeation
process was performed [2]. It was found that water
molecules form two kinds of bond of different
strengths named ``free water'' and ``bound water''.
The studies give evidence mainly about structural
changes at a short distance (interatomic).
In the present letter we have used X-ray methods
to study the structural changes in cellulose caused by
the sorption process. The expected results should
inform us about the changes at larger distances than
the interatomic distance. We studied the sorption of
light and heavy water, in both the liquid and the
vapour phase.
Samples were prepared from cellulose membrane
Tomophan I, 0.02 mm thick. First, we studied
swelling of the membrane in contact with water.
After soaking the membrane in water for a short
time (about 3 s) its weight increased about two fold.
A longer time of soaking did not effect the real
increase. Next we tried to expose the membrane to
the vapour. Independently of the exposure conditions
(20±80 8C; 1±6 h), the weight of the membrane
increased by 20%. For this weight increase the stage
of saturation of the water-backbone bonds was
reached. The differences between the light and
heavy water, taking into account the density
difference (10%), were very small. We decided to
measure ®ve membranes in X-ray experiments: a dry
membrane as a reference sample, two vaporized
membranes and two membranes soaked in liquid,
applying in all the above cases both light and heavy
water.
For the X-ray measurements, the membranes were
covered with thin Mylar foil to avoid reduction in
the water content. For the small-angle X-ray
scattering (SAXS) experiment, which needed thick
samples, small pieces of the membrane were put in a
sample vessel, covered with two Mylar windows at
both sides.
The SAXS method considers scattering on whole
particles having a mean electron density different
from that of the medium and forming interface
boundaries with it. Such particles can be pores or
inclusions [9]. The total radiation intensity scattered
in the small-angle region, using a Kratky-type set for
measurements, can be expressed by the equation
I(K) 2
I
0
dR
I
0
dë
I
0
dt
I
0
dx
e
O(x)P(t)W (ë)D(R)m
2
(R)Ö(KR)
where I(K) is the scattered intensity as a function of
the wavevector, D(R) is a function of the size
distribution of particles depending on their linear
parameter R (in the case of spherical particles, R is
the radius of a sphere), m
2
(R) is an integral over
excess density in the medium, Ö(KR) is the particle
shape factor, and Q, P, and W are functions
eliminating experimental errors (collimation and
wavelength band).
From this equation it is possible to calculate the
function D(R) by the inverse Fourier transform
method [10].
The measurements were performed in a compact
small-angle Kratky-type vacuum chamber using Cu
Ká radiation. The measurement range corresponds
to particle sizes from 1 to 30 nm. Appropriate
measurements were carried out to exclude collima-
tion error.
SAXS curves are shown in Fig. 1 for all samples.
One can see that there are no real small-angle effects
for dry and vaporized samples (Fig. 1, curves a±c)
and a small but real effect for watered samples
(Fig. 1, curves d and e). In the present experiment
the effect is not large but we should take into
account the small difference between the atomic
densities of water and cellulose. For both watered
samples the function, D(R), was calculated using the
0261-8028 # 1997 Chapman & Hall
1945