Depth Profiling Of Small Molecule Ingress Into Planar and
Cylindrical Materials Using NRA and PIXE
Richard W. Smith, Gary Massingham and Anthony S. Clough
Department of Physics, University of Surrey, Guildford, Surrey, GU2 7XH, UK
Abstract. The use of a 3He ion micro-beam technique to study the ingress/diffusion of water into a planar fibre optic
grade glass and a cylindrical drug-release polymer is described. One-dimensional concentration profiles showing the
depth of water ingress were produced. The depth of penetration of water into the glass was measured by fitting a
gaussian function to the concentration profile. The ingress of water into the drug-release polymer was found to be
Fickian and a cylindrical diffusion model used to obtain a diffusion coefficient.
Scanning micro-beam techniques have been
routinely used at Surrey for the depth profiling of
molecular diffusion into materials. The diffusing
molecules are labelled with deuterium and the material
treated as necessary. To investigate water (H2O)
diffusion heavy water (D2O) is used instead. The
material is then cut to expose a cross-section (showing
the diffusion profile) and a 3He ion micro-beam
scanned over the surface. The diffused molecules are
located by detecting the high energy protons (~12-13
MeV) from the D(3He,p)4He nuclear reaction shown
below and correlating them with the position of the
beam.
INTRODUCTION
There are many instances when it is important to be
able to measure molecular ingress/diffusion into a
material. The oil industry for example is interested in
the diffusion of water into fibre optic pressure sensors.
These are used to measure the pressure in oil wells
where they can be exposed to temperatures exceeding
200 to 300°C and pressures of up to ~600 bar. Water
diffusion into the sensor can have undesirable effects
leading to unreliable operation. Inhibition of this
diffusion will probably require both careful selection
of the glass and the use of a hydrophobic coating.
Studies are presently being carried out to find a
suitable combination. Because the production and
coating of a pressure sensor (~100 µm diameter) is
difficult and expensive, planar samples will initially be
produced and tested. Once a suitable glass and coating
have been found an actual sensor will then be
produced and tested.
3
He + D → p + 4He + Q
(Q = 18.35 MeV)
Elements within the sample can be located by
detecting their characteristic X-rays in a similar way.
In principle this technique can be used to profile the
diffusion of any hydrocarbon into any material. But in
practice the technique is limited by the ability to cut
the sample cleanly without disturbing the diffusion
profile. To date this technique has been applied to
many studies including water diffusion into planar
drug-release polymers [2], water diffusion into fibre
optic pressure sensors [3,4] and hair products diffusing
into hair [4].
The pharmaceuticals industry are developing drugrelease polymers [1] to be implanted in the body to
treat a variety of diseases. The drug release mechanism
involves the ingress of body fluids (water) into the
polymer causing the drug to egress out of the polymer.
A knowledge of this ingress is necessary to predict the
drug release and enable polymers to be designed for
specific purposes.
CP680, Application of Accelerators in Research and Industry: 17th Int'l. Conference, edited by J. L. Duggan and I. L. Morgan
© 2003 American Institute of Physics 0-7354-0149-7/03/$20.00
478
Separate
software
windows
were
created
encompassing the silicon X-ray peak, the proton and
backscattered ion peaks. Counts in these windows
were correlated with respect to the position of the
scanning beam and stored as two-dimensional maps.
EXPERIMENTAL
Sample Preparation
A Schott 8330 glass sample, of dimensions 2 cm ×
0.5 cm with a thickness of ~0.51 mm, was treated by
immersion in heavy water at a temperature of 200°C
for 1000 hours at a pressure of 19 bar. After treatment
the sample was cleaved into two 1 cm × 0.5 cm pieces
exposing a cross-section. One piece was clamped
between two copper blocks on an aluminium sample
plate with the exposed cross-section uppermost and
protruding slightly above the surface.
RESULTS AND ANALYSIS
Figs. 1 and 2 show the distribution of silicon and
deuterium respectively throughout the treated glass.
Silicon being the main component of glass indicates its
position and observation of the deuterium map shows
how far into the glass the heavy water has penetrated.
A cylindrical drug-release polymer depot, ~2.25
mm in diameter, loaded with a known fraction by
weight of a proprietary drug was supplied by the
manufacturer along with a phosphate buffered saline
(PBS) solution formulated to mimic body fluids. Two
1.8 cm long polymer sections were placed in separate
jars filled with 45 ml PBS solution (90%) and 5 ml
heavy water (10%). These were warmed to a
temperature of 37°C for times of 1 hour and 4 hours.
After the specified times the sections were removed
and lightly dried to remove any excess solution. A
segment ~0.3 cm long was cut from the centre of each
section and mounted, cut face uppermost, between two
double-sided adhesive tape covered carbon blocks and
screwed to an aluminum sample plate. The whole
sample plate was then immersed in liquid N2 to freeze
the water diffusion profile.
FIGURE 1. Two-dimensional map showing the distribution
of silicon throughout the glass.
NRA and PIXE
Each sample plate was in turn attached to the liquid
N2 cooled copper stage in the scanning micro-beam
line target chamber of the Van de Graaff accelerator at
the University of Surrey. The glass sample was
bombarded normal to its surface for 10 minutes at a
beam current of 3-4 nA, by a focussed scanning 1.5
MeV 3He micro-beam of 10 µm diameter at the sample
surface. The beam was raster scanned over a square of
side 2.1 mm. The drug-release polymer samples were
bombarded normal to their surface for 15 minutes at a
beam current of 1 nA, by a focussed scanning 1.3 MeV
3
He micro-beam of 10 µm diameter at the sample
surface. The beam was raster scanned over a square of
side 2.8 mm. Particle induced X-rays from silicon in
the glass were detected by a Si(Li) X-ray detector.
Protons from the D(3He,p)4He reaction and
backscattered ions from the carbon blocks (used to
normalise the drug-release polymer data) were
detected by two silicon surface barrier detectors.
FIGURE 2. Two-dimensional map showing the distribution
of deuterium throughout the glass.
479
A one-dimensional profile was produced from the
deuterium map by summing the counts in lines
perpendicular to the direction of diffusion (the x
direction in Fig. 2). This resulted in the profile shown
in Fig. 3 where the two peaks correspond to heavy
water penetration into both sides of the glass. The
peaks outer edges slope due to both the resolution of
the beam and non-uniformities in the sample surface
and can be approximately described by a gaussian
function. The two peaks are also symmetric about the
centre suggesting that the peak inner edges slope for a
similar reason.
Concentration (arbitrary units)
120
100
FIGURE 4. Normalised two-dimensional map showing the
distribution of deuterium throughout the polymer immersed
for 1 hour.
80
60
40
20
0
-0.1
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Thickness (mm)
FIGURE 3.
One-dimensional profile showing the
penetration of heavy water into the glass.
The heavy water penetration was measured using the
FWHM calculated by fitting a gaussian function [5] to
the peaks. The FWHMs were 0.070 ± 0.002 mm for
the left peak and 0.066 ± 0.004 mm for the right peak.
Both the FWHM and concentration (peak height) are
greater for the left peak. This effect has been observed
in other studies and is thought to be due to the position
of the sample during treatment i.e. one side facing
downwards and hence being exposed less to the
solution.
FIGURE 5. Normalised two-dimensional map showing the
distribution of deuterium throughout the polymer immersed
for 4 hours.
One-dimensional radial profiles were produced by
summing the normalised counts in different radial
annuli and dividing the sum by the corresponding
circumference. This resulted in the profiles shown in
Figs. 6 and 7 where a radial distance of 0 mm
corresponds to the centre of the polymer and ~1.1 mm
corresponds to the surface of the polymer. Using a
least squares fit diffusion curves were fitted to the
profiles using the solutions of Fick’s second law for
Fickian diffusion into cylindrical matrices [6]. For the
1 hour profile the solution most suitable for small
times was used and for the 4 hour profile the solution
most suitable for large times was used. Assuming a
constant diffusion coefficient, surface concentration
and radius with time – all requirements of the model –
the resulting diffusion coefficients were found to be
Figs. 4 and 5 show the distribution of deuterium
throughout the drug-release polymers soaked for 1
hour and 4 hours respectively. To account for beam
current fluctuations these two-dimensional maps were
normalised using the sum of the number of
backscattered ions from an area of the carbon block.
The location of deuterium indicates how far into the
polymer heavy water has diffused.
480
(9.5 ± 3.0) × 10-5 mm2 min-1 for the polymer soaked
for 1 hour and (9.0 ± 3.0) × 10-5 mm2 min-1 for the
polymer soaked for 4 hours. The average diffusion
coefficient was (9.3 ± 2.1) × 10-5 mm2 min-1.
using the same technique as described for the
cylindrical drug-release polymer.
The diffusion of heavy water into the cylindrical
drug-release polymer was found to be Fickian with a
diffusion coefficient of (9.3 ± 2.1) × 10-5 mm2 min-1.
These results are part of a much larger study
investigating water diffusion over times up to 1 week
for polymers loaded with different amounts of drug.
Concentration (arbitrary units)
2.5
Experimental data
2
Fit
1.5
ACKNOWLEDGMENTS
1
The authors are grateful for access to the EPSRC
ion beam centre accelerator facility at the University of
Surrey and the assistance of its staff. The assistance of
H. Rutt from the University of Southampton, UK who
treated the glass sample is also acknowledged.
0.5
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
Radial distance (mm)
FIGURE 6.
Normalised one-dimensional radial
concentration profile of heavy water in the polymer
immersed for 1 hour.
REFERENCES
Concentration (arbitrary units)
4
1. Fan, L. T., and Singh, S. K., Controlled release: a
quatitative treatment – Polymers properties and
applications 13, Berlin: Springer Verlag, 1998.
Experimental data
3
Fit
1. Smith, R. W., Massingham, G., and Clough, A. S.,
“Investigation of drug-release polymers using nuclear
reaction analysis and particle induced X-ray emission”
accepted for publication in Proc. of 10th Int. Conf. on
Radiation Measurements and Applications, Nucl. Instr.
and Meth. A.
2
1
0
0
0.2
0.4
0.6
0.8
1
1.2
2. Hollands, R., Rutt, H., Clough, A. S., Peel, R., and
Smith, R., Nucl. Instr. and Meth. B 174, 519-525 (2001).
1.4
Radial distance (mm)
Normalised one-dimensional radial
FIGURE 7.
concentration profile of heavy water in the polymer
immersed for 4 hours.
4. Smith, R. W., and Clough, A. S., “Depth profiling of
diffusion into cylindrical matrices using a scanning
micro-beam” in Proc. of 7th Int. Conf. on Accelerators in
Appl. Res. and Technol., edited by B. Sealy et al., Nucl.
Instr. and Meth. B 188, 2002, pp126-129.
DISCUSSION AND CONCLUSIONS
5. Gnoll, G. F., Radiation detection and measurement, New
York: John Wiley & Sons, 1989, pp. 76-79.
Water was found to penetrate the glass to a depth
of the order 10-1 mm. This study is ongoing and will
involve the analysis of a variety of different glasses
and coatings. When a suitable combination has been
found a fibre optic sample will be produced and tested
6. Crank, J., The Mathematics of Diffusion, Oxford:
Clarendon Press, 1975, pp. 73-74.
481