13C-NMR,1H-NMR, and FT-Raman Study of Radiation-
Induced Modifications in Radiation Dosimetry Polymer Gels
M. LEPAGE,1A. K. WHITTAKER,2L. RINTOUL,3C. BALDOCK1
1Centre for Medical and Health Physics, Queensland University of Technology, GPO Box 2434, Brisbane, Queensland
2Centre for Magnetic Resonance, University of Queensland, Brisbane, Queensland 4072, Australia
3Centre for Instrumental and Developmental Chemistry, Queensland University of Technology, GPO Box 2434, Brisbane,
Queensland 4001, Australia
Received 29 December 1999; accepted 7 April 2000
investigate the properties of a polymer gel dosimeter post-irradiation. The polymer gel
(PAG) is composed of acrylamide, N,N?-methylene-bisacrylamide, gelatin, and water.
The formation of a polyacrylamide network within the gelatin matrix follows a dose
dependence nonlinearly correlated to the disappearance of the double bonds from the
dissolved monomers within the absorbed dose range of 0–50 Gy. The signal from the
gelatin remains constant with irradiation. We show that the NMR spin–spin relaxation
times (T2) of PAGs irradiated to up to 50 Gy measured in a NMR spectrometer and a
clinical magnetic resonance imaging scanner can be modeled using the spectroscopic
intensity of the growing polymer network. More specifically, we show that the nonlinear
T2dependence against dose can be understood in terms of the fraction of protons in
three different proton pools. © 2000 John Wiley & Sons, Inc. J Appl Polym Sci 79: 1572–1581,
1H- and13C-NMR spectroscopy and FT-Raman spectroscopy are used to
polymer gel; dosimetry; magnetic resonance imaging; NMR; FT-Raman
In clinical radiotherapy there is a requirement to
measure 3-dimensional (3-D) radiation dose dis-
tributions. This is particularly important in situ-
ations where high dose gradients exist and where
it is necessary to verify dose distributions calcu-
lated by 3-D treatment planning computers. Com-
mon dosimeters (i.e., ionization chambers, ther-
moluminescent diodes, radiographic films, etc.) do
not allow the measurement of 3-D dose distribu-
tions because of severe spatial limitations.1
In 1984 it was proposed to use NMR relaxation
measurements of the Fricke2ferrous sulfate do-
simeter.3Ionizing radiation causes ferrous (Fe2?)
ions to be converted to ferric ions (Fe3?) through
radiolysis of an aqueous gel system. Fe3?ions
exhibit a larger paramagnetic enhancement than
Fe2?ions, and the NMR spin–lattice (T1) and
spin–spin (T2) relaxation times of the dosimeter
are related to the concentration of Fe3?produced
and hence to the absorbed radiation dose. In this
radiation dosimetry system, ferrous sulfate solu-
tions were incorporated into a gel matrix in order
to spatially stabilize the absorbed dose signature.
Dose maps can be obtained using magnetic reso-
nance imaging (MRI). A limitation in ferrous sul-
fate dosimetry systems is the diffusion of ions in
Correspondence to: C. Baldock (email@example.com).
Contract grant sponsor: Queensland Cancer Fund.
Journal of Applied Polymer Science, Vol. 79, 1572–1581 (2001)
© 2000 John Wiley & Sons, Inc.
the dosimeter, resulting in a spatially unstable
relaxation time or rate distribution.4–10
Dosimetry based on the radiation-induced
chemical changes in polymers was suggested in
1954.11As a result, the change in viscosity of
polyacrylamide solutions, which was supposedly
proportional to the absorbed dose of radiation,
was proposed as a dosimeter in 1961.12More re-
cently, polyacrylamide gel (PAG) radiation dosim-
eters were proposed in which the copolymeriza-
tion of acrylamide (AA) and N,N?-methylene-bi-
sacrylamide (BIS) monomers is believed to take
place. The monomers and final products are re-
tained in a gel matrix of either agarose or gela-
tin.13,14In these radiologically tissue equivalent
polymer networks,15,16the crosslinking of the
monomers was suggested to increase during the
copolymerization process in proportion to the ab-
sorbed radiation dose, although the final products
were not identified. Nevertheless, the resulting
change in the T2could be monitored using NMR
relaxation measurements. The uses of PAG for
radiotherapy dosimetry applications were report-
ed.17–24In addition to MRI,13,23optical tech-
niques including FT-Raman,25laser attenua-
tion,26and X-ray computed tomography27can
also be used to monitor the changes in PAG do-
simeters upon irradiation.
Despite the great clinical potential of gel do-
simetry, it is probably fair to say that no clinical
radiotherapy treatment has been altered because
of the results of gel dosimetry. This is partly due
to a lack of understanding of the fundamental
processes taking place in these materials post-
irradiation. The chemical modifications occurring
in PAGs have not been ascertained, leaving un-
answered the fundamental questions concerning
the behavior of the changes in NMR relaxation
times after irradiation with ionizing radiation.
Consequently, there is a good deal of confusion in
the existing literature where heuristic conclu-
sions have been drawn without firm experimental
evidence. This lack of knowledge results in unac-
ceptably large uncertainties in the evaluation
process. It is therefore necessary to undertake
fundamental measurements on these materials to
be able to develop gel dosimeters that will be of
use in the field of radiation oncology.
Various techniques can be employed to obtain
information on the chemical changes occurring at
a molecular level. For example, NMR spectros-
copy and FT-Raman have the potential to provide
quantitative information on the presence of dif-
ferent chemical groups in a molecule.
We measured the13C-NMR,1H-NMR, and FT-
Raman spectra of a PAG dosimeter with the view
to relating the macroscopic parameter T2to mi-
croscopic radiation-induced modifications in the
system. The polymer gel consists of two mono-
mers (AA and BIS) dissolved in an aqueous gela-
tin matrix. The measurements were performed on
different samples having absorbed doses up to 50
Gy. The results provide evidence for the polymer-
ization of the monomers that form a polyacryl-
amide network within the gelatin matrix. We
show that the T2of irradiated samples, measured
with a NMR spectrometer and with an MRI scan-
ner, can be modeled using the number of protons
of the growing polyacrylamide network derived
from FT-Raman spectroscopy. The latter calcula-
tions draw a direct link between the experimental
observations of a growing polymer network and
the associated changes in the T2recorded with
clinical MRI scanners.
NMR Spectroscopy and Relaxation Time
The NMR measurements were performed with a
Bruker MSL-300 spectrometer operating at a fre-
quency of 300.13 MHz for1H and 75.482 MHz for
13C. The spectrometer was equipped either with a
commercial 7- or 4-mm double air bearing, magic
angle spinning probe, or a static 7-mm double
resonance probe. Unless otherwise indicated, the
samples were spun at approximately 1.5 kHz to
provide partial averaging of the dipolar couplings,
as well as averaging over B-field distortions.28
The 90° pulse time for
spectrum width was 5 kHz, and 8000 data points
were acquired. The 90° pulse time used in the
high power proton-decoupled single pulse
spectra was 5.5 ?s. The spectrum width was 38
kHz, and 4000 data points were acquired.
The chemical shift scale of the1H-NMR spectra
followed the arbitrary setting of the H2O peak to
4.9 ppm while that of the13C spectra was refer-
enced to the low-field resonance of adamantane
(38.23 ppm). The samples were left to equilibrate
at room temperature (22°C), transferred in ZrO2
rotors, and immediately scanned.
The T2measurements were performed using
the Carr–Purcell–Meiboom–Gill sequence. A sin-
gle point was collected at the echo maximum for
1024 echoes in a pulse train. The time between
successive 180° pulses was varied from 1.2 to 6.0
1H-NMR was 6 ?s, the
MODIFICATIONS IN RADIATION DOSIMETRY PAG
ms to optimize the recording for our samples hav-
ing a range of T2values from 200 ms to 2 s,
respectively. In each case, the final intensity was
less than 5% of the original intensity. This condi-
tion ensured reliable exponential fits to the data.
The T2measurements were also undertaken
using the head coil of a Siemens Vision MRI scan-
ner operating at a frequency of 64 MHz. The
sample temperature was approximately 21°C. A
container holding the vials of gel was imaged with
a single slice in the coronal (horizontal) plane
using a modified conventional 32 echo spin–echo
pulse sequence. A repetition time of 4000 ms and
repeated echoes with an echo spacing of 50 ms
was used. The field of view was 256 ? 256 mm,
the slice thickness was 5 mm, and the pixel size
was 1.0 ? 1.0 mm. Four signal acquisitions were
used. Data were subsequently transferred to a
personal computer and processed to calculate T2
image “maps” using software written in-house.29
Regions of interest of similar area and number of
pixels were drawn on each vial in the map to
obtain the T2values.
A complete description of the FT-Raman spec-
trometer (2000 FT-Raman spectrometer system,
Perkin–Elmer, Beaconsfield, UK) used here was
reported previously.25Briefly, the beam from a
continuous wave Nd:yag laser emitting at 1064
nm is directed onto the samples and the scattered
radiation is detected with an InGaAs detector.
There were 250 acquisitions at a resolution of 8
cm?1, and a laser power of 400 mW recorded for
The samples were prepared in a glove box purged
with N2. Oxygen is known to have an inhibitory
effect on the free-radical chemistry of irradiated
liquids30and on free-radical polymerization.31
More specifically, O2was shown to rapidly react
with acrylamide radicals.32For this reason, the
atmosphere inside the glove box contained less
than 0.2 ppm of O2, as measured with an oxygen
meter (Quest Technologies). The samples for
NMR spectroscopy and FT-Raman spectroscopy
were kept in sealed “P6” glass vials (Amersham
International) prefilled with N2. The samples to
be imaged in the MRI scanner were kept in screw-
top glass vials. The samples were made of AA
(99?%, electrophoresis grade, Aldrich), gelatin
grade, Aldrich), BIS
(300 bloom, Aldrich), and H2O (deionized) or D2O
(99.9?% D atom, Aldrich). The details of the man-
ufacture of the gels were described previously.23
The polymer gel used in this study consisted of 3%
AA, 3% BIS, 5% gelatin, and 89% H2O by weight.
This formulation was previously given the acro-
nym PAG,23which is loosely used here, even for
our samples containing D2O instead of H2O.
The samples were irradiated at 22°C with ?
photons from60Co in a Gammacell 200 (Atomic
Energy Canada Limited). Typically, the samples
irradiated to 1 and 2 Gy were placed in a lead
attenuator and the dose rate was 0.06 Gy s?1.
Otherwise, the samples were placed in a polysty-
rene holder and the dose rate was 0.27 Gy s?1.
Both dose rates were calibrated with an ion cham-
ber.33A comparison between two gels irradiated
to 10 Gy at one or the other dose rate showed no
detectable difference in the T2values. Because
the T2evolves rapidly in the first few hours fol-
lowing irradiation, and more slowly thereaf-
ter,34,35the samples were left in a refrigerator for
at least 3 days prior to the NMR investigations.
Consider a target, containing initially n0reactant
molecules, that absorbs a differential dose D. In a
polymerization reaction, the rate of polymeriza-
tion is proportional to the number of reactant
molecules36and a first order process is expected.
The number of reactant molecules nDremaining
after the absorption of a dose D is
nD? A ? n0e?D/C
where A is a parameter introduced to take into
account the background intensity underlying a
considered feature in a spectrum. The overall
shape of the curve is an exponential decay with an
initial intensity A ? n0and the final intensity is
A. If we remove the background intensity from a
spectrum, the dose constant C can be redefined as
a half-dose, D1⁄2 ? 0.693C, which is simply the
dose required to obtain half the initial number of
With the present formulation, the water protons
accounted for about 91.9% of the total number of
protons. Hence, the1H-NMR spectra were highly
LEPAGE ET AL.
dominated by the water signal. Replacing H2O
with D2O lowered the water signal. It was then
possible to increase the receiver gain and thus
obtain a better signal to noise ratio for the weaker
peaks without extending the measurement time.
Figure 1 shows the13C-NMR spectra of a PAG
dosimeter irradiated at 0, 10, and 40 Gy. Figure 2
shows1H-NMR recorded at 300.13 MHz for PAGs
irradiated at 0, 10, and 50 Gy. The peaks were
assigned in comparison with spectra recorded
from separate aqueous solutions of the chemicals
diluted in H2O and D2O. A similar procedure was
used in an independent measurement of the1H
spectrum of an unirradiated gel and the peak
assignments were identical.37Some structures in
the1H spectrum of polyacrylamide may be influ-
enced by the solvent, the pH, and the tempera-
ture.38However, the assignments were in agree-
ment with published
amide.39,40 1H-NMR spectra were reported for AA
and BIS in solution, and the relative position of
the peaks was similar.41
In the13C spectra (Fig. 1), a doublet near 129
ppm was specifically assigned to the vinyl groups
of AA and BIS. Numerous peaks were found be-
tween 0 and 80 ppm and these could be attributed
to gelatin, which incidentally is composed of 20 or
more different monomeric species.42From the
separate measurements on individual aqueous
solutions of AA and BIS, we identified the follow-
ing structures between 166 and 183 ppm. The
signal from the carbonyl group of BIS was located
at 168.6 ppm while that of AA was at 170.8 ppm.
The spectrum from gelatin contained several
peaks located at 171, 174.5, 177.4, 181.6, and
184.7 ppm. In the polymer gel system including
AA, BIS, and gelatin, all these peaks were super-
imposed. In the same spectral region, the high
resolution13C of polyacrylamide from Ganapathy
et al. showed a single narrow peak assigned to the
carbonyl carbon at about 175 ppm.40
peak near 4.9 ppm was attributed to residual H2O
and HDO molecules. We note that the intensity of
this peak was 3.5% of the water peak when H2O
1H spectra of Figure 2 the dominant
The spectra for the samples having absorbed doses of 0,
10, and 50 Gy are shifted for clarity.
1H-NMR spectra at 300.13 MHz of a PAG.
The spectra for the samples having absorbed doses of 0,
10, and 40 Gy are shifted for clarity. The structures
ascribed to the CAC double bonds of AA and BIS de-
crease with the absorbed radiation dose.
13C-NMR spectra at 75.482 MHz of a PAG.
MODIFICATIONS IN RADIATION DOSIMETRY PAG
was used. At the foot of this peak, a small peak
located at 4.8 ppm was attributed to the central
CH2group of BIS. A second set of experiments
was performed on three samples containing H2O
having absorbed 0, 10, and 50 Gy. The samples
were spun at 6 kHz to prevent the spinning side
bands from overlapping the structures. The re-
sults are shown in Figure 2 for the chemical shift
range from 6.65 to 8.9 ppm. A doublet having
peaks located at 7.1 and 7.8 ppm was assigned to
the inequivalent protons of the NH2group of AA.
A peak centered at 8.7 ppm, attributed to the NH
groups of BIS, was superimposed on a broad sig-
nal attributed to the NH groups of gelatin. A
1H-NMR study of the amide side chain of a com-
mercial polyacrylamide sample revealed a triplet
centered near 7.7 ppm and a singlet at about 7
ppm that were assigned to the trans and cis pro-
ton of the amide group, respectively.38Two sharp
multiplets observed near 5.9 and 6.3 ppm were
assigned to the vinyl groups of AA and BIS. The
large number of peaks recorded between 0 and
4.65 ppm were attributed to gelatin. We observed
new peaks at 2.3 and 1.8 ppm from solutions of
AA in H2O irradiated at 50 Gy (not shown). The
latter, which may have been due to CH2groups of
a polyacrylamide network, were not seen in the
spectra of the irradiated polymer gel of Figure 2.
A set of FT-Raman spectra recorded from PAG
samples having absorbed up to 50 Gy was re-
ported earlier.25However, we acquired a new set
of spectra in order to get a better signal to noise
ratio to analyze structures having low intensities.
Figure 3 shows a portion of the FT-Raman
spectra of the polymer gel between 2800 and 3150
cm?1for three different absorbed doses of 0, 10,
and 50 Gy. The 0-Gy spectrum contains two prom-
inent peaks with maxima at 3111 and 3039 cm?1,
as well as structures at 2995, 2980, 2940, and
2884 cm?1. The first two were ascribed to the
antisymmetric and symmetric vinyl CH stretch
from both BIS and AA. The peak at 2995 cm?1
was ascribed to the alkyl CH stretch of AA. The
intensity of the peaks at 2940, 2980, and 2884
cm?1, although different from zero in the 0-Gy
spectrum, increased with the absorbed dose. The
peaks at 2940 and 2884 cm?1resulted from the
excitation of the antisymmetric and symmetric
alkyl CH stretches from gelatin and, as is dis-
cussed below, from a growing polymer network.
The 2980 cm?1band was attributed to the CH
group part of a polymer chain.
Finally, we recall that two bands located at
1256 and 1285 cm?1were previously assigned to
the vinyl ?CH2vibrations of BIS and AA, respec-
tively.25The intensity of the vinyl ?CH2vibrations
was dependent on the absorbed dose and thus
FT-Raman imaging was proposed to obtain dose
distributions with high spatial resolution.25
Peak Intensities as Function of Absorbed Dose
The water content of the samples did not change
significantly with the absorbed dose. Assuming a
G value of 5 for the dissociation of water mole-
cules,43only 4.7 ? 10?7% of the H2O molecules
will be broken after a 50-Gy absorbed dose. Fur-
thermore, the signal from water vibrations in all
FT-Raman spectra was found to be constant over
the absorbed dose range studied. Therefore, the
intensity scales of the different
corded for samples having absorbed from 0 to 50
Gy were normalized to the residual water peak
1H spectra re-
cm?1of a PAG for three different absorbed doses of 0, 8,
and 50 Gy, which are shifted vertically for clarity. On
the 0-Gy spectrum, two prominent peaks are seen at
3111 and 3039 cm?1. They are ascribed to the antisym-
metric and symmetric vinyl CH stretch from both BIS
and AA. Although their intensity decreases with the
absorbed dose, the intensity of the 2980 and the 2884
cm?1peaks increases. They are specifically ascribed to
the antisymmetric and symmetric alkyl CH stretches of
a growing polyacrylamide network.
FT-Raman spectra between 2800 and 3150
LEPAGE ET AL.
intensity. By comparing the1H spectra at differ-
ent doses, it was apparent that the peaks as-
signed to gelatin (0–4.65 ppm) remained un-
changed upon irradiation while those ascribed to
the vinyl group of AA and BIS decreased to zero.
The13C spectra were therefore normalized to the
integral of the signal originating from the gelatin
peaks (0–80 ppm). It was then apparent that the
peaks from the vinyl groups of AA and BIS de-
Figure 4 shows the integral of peaks assigned
to various AA and BIS groups. The half-dose of
the monoexponential fits to the data are also
shown in the figure. More specifically, the curves
were fitted with the exponential decay function of
eq. (1). We used a built-in monoexponential fitting
function based on the Levenberg–Marquardt al-
gorithm from MicrocalTM
Software, Inc., Northampton, MA), which gives
the value and the absolute uncertainty of the
parameters that best describe the data.
The vinyl ?CH2vibrations of BIS and AA were
observed at 1256 and 1285 cm?1, respectively,
with FT-Raman spectroscopy.25Those bands de-
cayed exponentially with the absorbed dose, al-
though with a different half-dose (Fig. 4, curve a).
More specifically, the half-doses of the exponen-
tial decays for AA and BIS were 8.36 ? 0.90 and
4.52 ? 0.49 Gy, respectively. A similar observa-
tion confirming the faster disappearance of BIS
was reported by Nieto et al. for the chemically
induced copolymerization of AA and BIS in solu-
tion and was taken as evidence for the larger
reactivity of BIS.41
Figure 4 represents the intensity of the peaks
assigned to the vinyl groups of AA and BIS lo-
cated near 129 (curve a) and 5.9 ppm (curve b) in
doses were 7.44 ? 1.18 and 7.62 ? 0.71 Gy, re-
spectively. The spectroscopic signature of the vi-
nyl groups of AA and BIS are superimposed in our
13C- and1H-NMR spectra. By simply weighting
the half-doses for AA and BIS obtained from FT-
Raman by the total number of double bonds, we
obtained a half-dose of 6.52 ? 0.70 Gy. Consider-
ing that the errors quoted only include the uncer-
tainty from the fitting procedure, the agreement
between the estimated weighted value and the
experimentally determined values was very good.
This justified the normalization of the NMR spec-
tra a posteriori and showed the good agreement
between the two spectroscopic approaches.
The main features in the region of the FT-
Raman spectra located between 2850 and 3150
cm?1(Fig. 3) can be reasonably well fitted with 6
Gaussians centered at the locations mentioned
above. The intensity of the bands located at 3111,
3039, 2940, and 2884 cm?1are shown in Figure 5.
The first two bands were found to have respective
half-doses of 7.28 ? 1.62 (ACH2antisymmetrical
stretch, BIS and AA) and 6.63 ? 0.53 Gy (ACH2
symmetrical stretch, BIS and AA). On the other
hand, the bands located at 2940 and 2884 cm?1
were growing with the absorbed dose with respec-
tive half-doses of 6.20 ? 1.81 and 5.37 ? 2.15 Gy.
The relatively large uncertainty on the latter
half-dose value stemmed directly from the rela-
tively low signal intensity of the band located at
2884 cm?1(Fig. 3).
1H spectra, respectively. The half-
Spin–Spin Relaxation Time for Different Absorbed
Figure 6 shows the T2measured from PAG do-
simeters that absorbed up to 50 Gy, which were
measured with the 300.13-MHz NMR spectrome-
ter and 64-MHz clinical MRI scanner. A tentative
exponential fit to the data gave half-doses of 4.68
OCACH2vinyl (?CH2) vibrations of AA and BIS with
the1H- and13C-NMR signal from the same chemical
moiety. The PAGs are composed of 3% AA, 3% BIS, 5%
gelatin, and 89% H2O by weight in the FT-Raman
(curve a) and13C-NMR (curve b) results and 89% D2O
in the1H-NMR results (curve c). The half-doses from
the monoexponential fits are discussed in the text.
A comparison of the intensity of the
MODIFICATIONS IN RADIATION DOSIMETRY PAG
? 0.37 and 2.56 ? 0.06 Gy, respectively. The solid
lines were obtained from a model described below.
Also shown in Figure 6 are the commonly em-
ployed “dose response” [i.e., relaxation rate R2
(1/T2) against dose] graphs.
In the presence of a chemical initiator, AA and
BIS are well known to form a polyacrylamide
network in an aqueous solution.44–47Copolymer-
ization of the vinyl groups of the monomers (AA
and BIS) creates a COC chain with amide side
chains. BIS acts as a crosslinker that binds two
chains together. The creation of clusters of finite
size leads to an increased turbidity in polyacryl-
amide gels (PAGs).47
Our1H-NMR spectrum of an irradiated solu-
tion of AA (not shown) was very similar to the
spectrum of a polyacrylamide sample published
by Ganapathy et al.40This showed that radiation-
induced polymerization of AA was achievable
within the dose range used in this study. How-
ever, in a gelatin matrix, although the intensity of
the double bonds from AA and BIS decreases rap-
idly with dose, the characteristic signature of the
CH and CH2groups did not emerge from the
gelatin background intensity. If we assume that
copolymerization occurred (as evidenced below),
the polyacrylamide chains crosslinked with BIS
were likely to precipitate out of solution. As the
quantity of polymer increased with the absorbed
dose, the relaxation via interaction with the poly-
mer surface became more efficient and the T2of
this polymer gel dropped rapidly13(Fig. 6). The
line width of the NMR peaks of the polymer is
and relaxation rate (R2) for PAG samples in the 0–50
Gy range. The results were obtained at (E) 64 and (F)
300.13 MHz. (—) The results as obtained with a model
described in the text.
The nuclear spin–spin relaxation time (T2)
tions of AA, BIS, and a polyacrylamide network in a
PAG as a function of the absorbed dose. The vibrations
from the polyacrylamide network grow with the same
half-dose as the vibrations from the monomers decay.
The intensity of the COH stretching vibra-
LEPAGE ET AL.
likely to become so broad that the peak will be-
come very difficult to observe at room tempera-
ture, as was shown to be the case for the copoly-
merization of AA and BIS in solution.39It can
thus be expected that the signal from other chem-
ical groups in AA and BIS would “disappear” with
the same half-dose as the double bonds. Cross-
polarization (CP) experiments, on another hand,
are more sensitive to the presence of rigid poly-
mer chains. We measured the CP spectrum for
the samples having absorbed 0 and 50 Gy. The
latter showed the presence of a peak located at 41
ppm and a shoulder near 35 ppm. These peaks
correspond to CH and CH2groups of polyacryl-
The FT-Raman results of Figure 5 clearly show
the formation of alkyl CH2vibrations, which in-
dicates the formation of a saturated chain of car-
bon atoms. Within the uncertainties in the fits,
the half-dose for the appearance of the alkyl CH2
vibrations was the same as for the disappearance
of the double bonds. Furthermore, the value of the
half-dose was confirmed by both the1H- and13C-
NMR results (Fig. 4).
In Bansil and Gupta’s47Raman study a band
at 3045 cm?1was observed that grew as copoly-
merization of AA and BIS in water proceeded.
This vibrational band was assigned to a CH2
stretching from a CACH2group from a BIS mol-
ecule part of a polymer chain, but having one
double bond unreacted. In the present study the
intensity of this band was below the limit of de-
tectability in Figure 3. Other researchers made a
direct correlation between the amount of cross-
linker (BIS) and the degree of crosslinking in a
PAG.50However, the observation that some BIS
molecules in polyacrylamide may have an unre-
acted double bond argues against such a direct
We obtained different G values (defined as the
number of molecules of material changed, or of
product formed, for each 100 eV of absorbed en-
ergy in the system43) from the NMR spectra and
FT-Raman spectra. To a first approximation, the
initial decrease of the double bonds intensity can
be considered linear from 0 to 8 Gy. We observed
that 50% of the double bond intensity from AA
and 27% of that of BIS remained after 8 Gy.
Knowing the initial chemical composition of the
polymer gel, we calculated the G values for the
consumption of the double bonds of AA to be
G(?db) ? 2.54 ? 105and for BIS to be G(?db)
? 3.42 ? 105. From those values, assuming every
double bond of AA and BIS participates in the
formation of an alkyl chain (i.e., copolymeriza-
tion), the G value for the formation of alkyl CH2
groups was calculated to be about G(OCH2O) ? 5
In a diluted solution, the initiation of polymer-
ization is thought to be caused by free radicals of
water. Therefore, if we make the crude assump-
tion that every radical from water [G (total radi-
cal) ? 10]43attacks a double bond from AA or BIS
to initiate a polymerization reaction, we find that
the degree of polymerization (i.e., the number of
repeating monomer units in a polymer chain31) is
of the order of 104, the same order of magnitude
expected for vinyl polymerization31and that ob-
tained for copolymerization of AA and BIS in so-
lution.51The simple degradation model we used
assumes a probability of reaction of a double bond
with a water radical that is independent of the
amount of polymer formed (i.e., independent of
the remaining number of double bonds). The
monoexponential decrease of the double bond in-
tensity with the absorbed dose may suggest that
this condition is realized in PAG dosimeters. This
may imply that the diffusion of the monomers in
the gelatin matrix is independent of the absorbed
radiation dose, within the range covered in this
study, in agreement with monomer diffusion mea-
surements on the same system.37
Copolymerization reactions of AA and BIS cre-
ate a solid polyacrylamide network within the
gelatin matrix. We made calculations using the
classic model of fast exchange52in which the mea-
sured T2is given by
protons in the mobile, polymer, and gelatin pools,
respectively; and T2,mob, T2,pol, and T2,gelaare
their respective intrinsic NMR spin–spin relax-
ation times. The fractions of protons in the first
two pools were evaluated using the initial concen-
trations of the solution and the curve for the pro-
duction of polymer CH2vibrations in Figure 5.
The mobile pool initially contained the protons
from water and the monomers, the latter protons
being transferred with a half-dose of 7.50 Gy in
the polymer pool. The fraction of gelatin protons
was estimated to be 4.35% and was kept constant
as suggested by Figures 1 and 2. Using a similar
equation, the T2,gelacould be determined experi-
mentally to be 37 ms from a sample containing 5%
gelatin and 95% H2O. In order to fit the data, the
T2,polwas set at 13 ? 1 ms and T2,mobwas set at
H, and fgela
are the fraction of
MODIFICATIONS IN RADIATION DOSIMETRY PAG
2 s for 300.13 MHz. For comparison, the results
obtained from the clinical MRI scanner operating
at 64 MHz were added to Figure 6. The T2,mobwas
set at 3 s, which agreed with the observation of an
increase for T2with decreasing field strength for a
polymer gel.13The results are presented as solid
lines in Figure 6, and it is apparent that the
relaxation properties of the PAG are well repre-
sented by the model.
On the basis of this simple model, it is not
strictly correct to fit a monoexponential decay to
the decrease in T2with the absorbed dose. How-
ever, the half-dose that was obtained from the
300-MHz measurements was 4.68 ? 0.37 Gy. In-
terestingly, this value is smaller than the half-
dose used to calculate the number of protons in
the polymer chain. Several authors simply as-
sumed that the observed changes in T2were
solely due to the “presence” of the solid formed
polymer that was thought to increase linearly
with the absorbed dose.13,35,50,53Although it may
be approximated to a linear function in the first
few Grays of radiation, the present work showed
that there was no linear relationship between the
increase in polymer protons and the decrease of
T2(or the increase of R2). Furthermore, the non-
linearity of T2(or R2) with the absorbed dose has
profound implications in the determination of the
absorbed dose from the polymer gels in applied
radiotherapy gel dosimetry.54,55
A recent article presented a study of the evolu-
tion of T2as a function of postirradiation time and
showed a continuous decrease over many days.35
Although the authors did not characterize the
processes involved or the products formed and
their affirmation was solely based on T2measure-
ments, they claimed a polymerization reaction
was proceeding throughout the time interval. We
point out that a continuous phase separation or a
rearrangement of the polymer giving it a tighter
structure could also have a similar effect. We will
address this issue in more detail in a forthcoming
article.56We gathered some evidence that the T2
relaxation is dominated by chemical exchange.57
In this regard, not all amide protons in the poly-
acrylamide chain may be easily accessed for ex-
change, especially in the formation of complex
We characterized the chemical changes occurring
in a PAG radiation dosimeter, consisting of AA,
BIS, gelatin, and water, using
NMR, and FT-Raman spectroscopy. We showed
that irradiation of AA and BIS in a gelatin matrix
leads to copolymerization of the monomers, with-
out the inclusion of gelatin molecules. Specifi-
cally, the double bonds from AA and BIS disap-
peared with a half-dose near 7.5 Gy while the
signal from gelatin remained constant over the
absorbed dose range up to 50 Gy. A polyacryl-
amide network grew with the same half-dose,
within experimental errors. This polyacrylamide
network within the gelatin matrix was character-
ized with CP experiments results identical to a
similar experiment performed on a polyacryl-
amide sample.40The T2values recorded at 64 and
300.13 MHz were fitted accurately by a simple
classical model. The number of protons belonging
to the polymer proton pool was directly derived
from the FT-Raman spectra. Based on this model,
new gel formulations covering a selected range of
T2values can be envisaged.
The authors acknowledge P. J. Murry and P. M. Ja-
yasekera for help provided in the manufacture of the
gels. We thank D. J. T. Hill and J.-P. Jay-Ge ´rin for
useful discussions. We also thank Southern X-ray Clin-
ics, Wesley Hospital, Brisbane, for access to their MRI
1. Schreiner, L. J., Ed. Proceedings of the 1st Inter-
national Workshop on Radiation Therapy Gel Do-
simetry; Canadian Organization of Medical Physi-
cists: Edmonton, Canada, 1999.
2. Fricke, H.; Morse, S. Am J Roent Radium Ther
Nucl Med 1927, 18, 430.
3. Gore, J. C.; Kang, Y. S.; Schulz, R. J. Phys Med Biol
1984, 29, 1189.
4. Schulz, R. J.; deGuzman, A. F.; Nguyen, D. B.;
Gore, J. C. Phys Med Biol 1990, 35, 1611.
5. Olsson, L. E.; Westrin, B. A.; Fransson, A.; Nordell,
B. Phys Med Biol 1992, 37, 2243.
6. Balcolm, B. J.; Lees. T. J.; Sharp, A. R.; Kulkarni,
N. S.; Wagner, G. S. Phys Med Biol 1995, 40, 1665.
7. Baldock, C.; Harris, P. J.; Piercy, A. R.; Patval, S.;
Prior, D. N.; Keevil, S. F.; Summers, P. Med Phys
1995, 22, 1540.
8. Harris, P. J.; Piercy, A. R.; Baldock, C. Phys Med
Biol 1996, 41, 1745.
9. Kron, T.; Jonas, D.; Pope, J. M. Magn Reson Imag-
ing 1997, 15, 211.
10. Pedersen, T. V.; Olsen, D. R.; Skretting, A. Phys
Med Biol 1997, 42, 1575.
11. Alexander, P.; Charlesby, A.; Ross, M. Proc R Soc A
1954, 223, 392.
12. Boni, A. L. Radiat Res 1961, 14, 374.
LEPAGE ET AL.
13. Maryan ˜ski, M. J.; Gore, J. C.; Kennan, R. P.; Download full-text
Schulz, R. J. Magn Reson Imaging 1993, 11, 253.
14. Maryan ˜ski, M. J.; Schulz, R. J.; Ibbott, G. S.;
Gatenbury, J. C.; Xie, J.; Horton, D.; Gore, J. C.
Phys Med Biol 1994, 39, 1437.
15. Keall, P.; Baldock, C. Med Phys 1998, 25, A194.
16. Keall, P.; Baldock, C. Australas Phys Eng Sci Med
1999, 22, 85.
17. Maryan ˜ski, M. J.; Ibbot, G. S.; Eastman, P.; Schulz,
R. J.; Gore, J. C. Med Phys 1996, 23, 699.
18. Audet, C.; Duzenli, C.; Kwa, W.; Tsang, V.;
Mackay, A. Med Phys 1996, 23, 803.
19. Ibbott, G. S.; Bova, F. J.; Maryan ˜ski, M. J.; Zhang,
Y.; Holcomb, S. D.; Avison, R. G.; Meeks, S. L. Med
Phys 1996, 23, 1070.
20. Ibbott, G. S.; Maryan ˜ski, M. J.; Eastman, P.; Hol-
comb, S. D.; Zhang, Y.; Avison, R. G.; Sanders,
M. S.; Gore, J. C. Int J Radiat Oncol Biol Phys
1997, 38, 1097.
21. Baldock, C.; Hasler, C. D.; Keevil, S. F.; Greener,
A. G.; Billingham, N. C.; Burford, R. P. Med Phys
1996, 23, 1490.
22. Baldock, C.; Hindocha, N.; Billingham, N. C.; Bur-
ford, R. P.; Keevil, S. F.; Procter, E. Med Phys 1996,
23. Baldock, C.; Burford, R. P.; Billingham, N. C.; Wag-
ner, G. S.; Patval, S.; Badawi, R. D.; Keevil, S. F.
Phys Med Biol 1998, 43, 695.
24. Oldham, M.; Baustert, I.; Lord, C.; Smith, T. A. D.;
McJury, M.; Leach, M. O.; Warrington, A. P.; Webb,
S. Phys Med Biol 1998, 43, 1113.
25. Baldock, C.; Rintoul, L.; Keevil, S. F.; Pope, J. M.;
George, G. A. Phys Med Biol 1998, 43, 3617.
26. Gore, J. C.; Ranade, M.; Maryan ˜ski, M. J.; Schulz,
R. J. Phys Med Biol 1996, 41, 2695.
27. Hilts, M.; Audet, C.; Duzenli, C. Med Phys 1999,
28. Schmidt-Rohr, K.; Spiess, H. W. Multidimensional
Solid-State NMR and Polymers; Academic: Lon-
29. Murry, P. J.; Baldock, C. Australas Phys Eng Sci
Med 2000, 23, 44.
30. Fo ¨ldia ´k, G., Ed. Radiation Chemistry of Hydrocar-
bons; Elsevier: Amsterdam, 1981.
31. Flory, P. J. Principles of Polymer Chemistry; Cor-
nell University Press: Ithaca, NY, 1953.
32. Chambers, K. W.; Collinson, E.; Dainton, F. S.;
Seddon, W. A.; Wilkinson, F. Trans Faraday Soc
1967, 63, 1699.
33. Baldock, C.; Fitchew, R.; Murry, P. J.; Murry, M.;
Back, P.; Kron, T. Med Phys 1999, 26, 1130.
34. Baldock, C.; Lepage, M.; Rintoul, L.; Murry, P. J.;
Whittaker, A. K. In Proceedings of the 1st Interna-
tional Workshop on Radiation Therapy Gel Dosim-
etry; Schreiner, L. J., Ed.; Canadian Organization
of Medical Physicists: Edmonton, Canada, 1999.
35. McJury, M.; Oldham, M.; Leach, M. O.; Webb, S.
Phys Med Biol 1999, 44, 1863.
36. Chapiro, A. Radiation Chemistry of Polymeric Sys-
tems; Interscience Publishers: New York, 1962.
37. Haraldsson, P.; Nyde ´n, M.; Ba ¨ck, S. Å. J.; Olsson,
L. E. In Proceedings of the 1st International Work-
shop on Radiation Therapy Gel Dosimetry; Schre-
iner, L. J., Ed.; Canadian Organization of Medical
Physicists: Edmonton, Canada, 1999.
38. Miyoshi, T.; Takegoshi, K.; Hikichi, K. Polym J
1994, 26, 485.
39. Rajamohanan, P. R.; Ganapathy, S.; Ray, S. S.;
Badiger, M. V.; Mashelkar, R. A. Macromolecules
1995, 28, 2533.
40. Ganapathy, S.; Ray, S. S.; Rajamohanan, P. R.;
Mashelkar, R. A. J Chem Phys 1995, 103, 6783.
41. Nieto, J. L.; Baselga, J.; Herna ´ndez-Fuentes, I.;
Llorente, M. A.; Pie ´rola, I. F. Eur Polym J 1987, 23,
42. Eastoe, J. E.; Leach, A. A. In The Science and
Technology of Gelatin; Ward, A. G., Courts, A.,
Eds.; Academic: London, 1977; Chapter 3.
43. O’Donnell, J. H.; Sangster, D. F. Principles of Ra-
diation Chemistry; Edward Arnold: London, 1970.
44. Richards, E. G.; Lecanidou, R. In Electrophoresis
and Isoelectric Focusing in Polyacrylamide Gels;
Allen, R. C., Maurer, H. R., Eds.; de Gruyter: Ber-
45. Gelfi, C.; Righetti, P. G. Electrophoresis 1981, 2,
46. Cohen, Y.; Ramon, O.; Kopelman, I. J.; Mizrahi, S.
J. Polym Sci B Polym Phys 1992, 30, 1055.
47. Bansil, R.; Gupta, M. K. Ferroelectrics 1980, 30, 63.
48. Inoue, Y.; Fukutomi, T.; Chu ˆjo ˆ, R. Polym J 1983,
49. Hikichi, K.; Ikura, M.; Yasuda, M. Polym J 1988,
50. Maryan ˜ski, M. J.; Audet, C.; Gore, J. C. Phys Med
Biol 1997, 42, 303.
51. Baselga, J.; Llorente, M. A.; Herna ´ndez-Fuentes,
I.; Pie ´rola, I. F. Eur Polym J 1989, 25, 471.
52. Callaghan, P. T. Principles of Nuclear Magnetic
Resonance Microscopy; Clarendon Press: Oxford,
53. Audet, C. PhD Dissertation, McGill University,
54. Baldock, C.; Murry, P.; Kron, T. Phys Med Biol
1999, 44, N243.
55. Baldock, C.; Lepage, M.; Jayasekera, M.; Ba ¨ck,
S. Å. J., unpublished results.
56. Lepage, M.; Whittaker, A. K.; Rintoul, L.; Ba ¨ck,
S. Å. J.; Baldock, C., unpublished results.
57. Lepage, M.; Whittaker, A. K.; Baldock, C., unpub-
MODIFICATIONS IN RADIATION DOSIMETRY PAG