Determination of New Radical Species in Ammonium Tartrate Dosimeters by CW- and Pulsed-EPR Techniques

Samples of ammonium tartrate irradiated with doses of about 0.1–1 kGy by different beams of ionizing radiation (60Co–γ, 19 MeV protons and 62 MeV per nucleon carbon ions) were studied by continuous-wave electron paramagnetic resonance (cw-EPR) and by pulse-EPR techniques. Careful analysis of the cw-EPR and of the echo-detected EPR spectra allowed the identification of a second radical in the system besides the already known radical formed at high temperature by an hydrogen elimination at C(2) position [M. Brustolon et al., Res. Chem. Int 4:359, 1996]. The spectrum of the radical is compatible with that of a radical obtained by hydroxyl elimination.

The amino acid L-alanine is today the most popular dosimeter for high-dose radiation; its prominent dosimetric features are the good tissue equivalence, the linearity of its dose-response over a wide range, the high stability of some of the radiation-induced free radicals. In an intercomparison trial between six independent EPR laboratories with the aim of increasing the accuracy of EPR dosimetry, Lalanine was chosen as radiation-sensitive material [4]. In part as a consequence of the dosimetric applications, this amino acid is one of the most exhaustively studied molecules in the field of radiation chemistry. Thanks to a number of EPR studies it has been found that several radicals contribute to its overall EPR spectrum, although one of the radiation-induced radical species, called the 'stable alanine radical' (SAR), dominates the spectrum ( [5] and references therein). The latter spectrum is complex and wide, both properties being a drawback for the dosimetric approach. The stability and the magnetic properties of the different radicals are in part different [6]. A further limit of the L-alanine dosimeter is its low sensitivity for low radiation doses, as those applied in medical radiation treatments.
Several other materials have been proposed and tested to overcome these problems (see, e.g., [7][8][9]). Suitable materials should show a high yield of radicals, good sensitivity for low-dose radiation, sharp spectral lines, a linear dose-response curve, thermal stability of the radicals at room temperature, and a narrow and simple EPR spectrum.
Ammonium tartrate (AT) is the most studied EPR dosimeter after L-alanine, as it fulfills most of these requirements [10]. The X-band EPR spectrum of irradiated AT has a width of about 1 mT, about 1/10 of that of the EPR spectrum of L-alanine. The powder spectrum is a single band.
The main radical produced by gamma irradiation has been studied by electronnuclear double resonance (ENDOR) and electron spin echo envelope modulation spectroscopy, and found to be Radical A (see Scheme 1) [11]. A comparison between the results of a quantitative cw-and a pulse-EPR study has given the total and the microscopic radical concentrations in AT crystals X-irradiated at different doses [12]. A series of AT single crystals and powder samples have been exposed to radiations with different linear energy transfer (LET), and studied by two types of pulse-EPR investigations: two-pulse electron spin echo decay by varying the microwave power and a double electron-electron resonance (DEER) study. The results have shown the capacity of pulse-EPR spectroscopy to discriminate the Scheme 1 Chemical structure of the stable radical obtained by gamma irradiation of AT (A) as determined by previous ENDOR and ESEEM studies [11], and the structure proposed of a new types of radicals (B) produced by OH elimination quality of various radiation beams such as c-ray photons, protons and thermal neutrons [3,13]. Recently, also a cw-technique making use of out-of-phase second harmonic acquisition has been used for radiation quality discrimination on irradiated ammonium tartrate pellets [14]. These studies have shown that AT crystals are suitable for obtaining information on dose and quality of ionizing radiations. However, in contrast with the deep insight into the radiation chemistry of L-alanine, very little is known about the formation of other radicals besides Radical A, despite some experimental evidences on the presence of more radicals. The first is reported in [11], and it is a small splitting of some ENDOR lines due to the presence of two A radicals in slightly different conformations. Another evidence is the presence of low-intensity side bands at the edge of the Radical A EPR spectrum.
The presence of other radicals besides Radical A is particularly relevant when studying the radical distribution in the AT crystal [3,12,13] by pulse-EPR experiments (Hahn echo decay and DEER). As these experiments are based on the spin relaxation properties of the radicals due to their dipolar interaction, the presence of radicals with different phase memory times can affect the results.
In this paper, we report a preliminary investigation by pulse-EPR on AT single crystals irradiated with 60 Co c-ray photons, 19.3 MeV protons and 62 MeV per nucleon carbon nuclei aimed at revealing the presence of radicals contributing to the EPR spectra, also related to the different irradiation.

Experimental
AT pellets were irradiated using three different beams: 60 Co c photons, protons with 19.3 MeV initial energy, and carbon nuclei with 62 MeV per nucleon initial energy. Irradiations with 60 Co c photons, 1.25 MeV mean energy, were carried out with the irradiator IGS-3 located in the Dipartimento di Ingegneria Nucleare, Università di Palermo. The AT powders were irradiated at various doses between 0.1 and 5 kGy. Irradiation with protons and with carbon ions (nuclei) were performed at the accelerator facilities of the Laboratori Nazionali del Sud-Istituto Nazionale di Fisica Nucleare (LNS-INFN) at Catania, with doses varying between 0.1 and 5 kGy.
The cw-EPR spectra were recorded with an ECS Bruker instrument equipped with a cylindrical cavity. Pulse-EPR measurements were performed with an X-band Bruker ELEXYS spectrometer, equipped with a dielectric resonator and a nitrogen gas-flow cryostat for low temperature measurement. The powder samples were placed inside 3-mm ID quartz EPR tubes. The pulse experiments were performed using the standard 2p-pulse sequences: p-s-2p 1 -s-echo for the electron spin echo (ESE). The power of the pulse was varied, whereas the length was kept fixed at 28 ns. The echo-detected EPR (ED-EPR) spectra were obtained by recording the echo intensity as function of the magnetic field.
All the simulated EPR spectra have been obtained by an appropriate use of the simulation program, Easyspin [15].

Results and Discussion
The EPR spectra of AT irradiated at room temperature with radiation beams of 60 Co c-ray photons, 19.3 MeV protons and 62 MeV per nucleon carbon ions are all identical (Fig. 1). The overall extension of the spectrum is about 5.2 mT. It shows a central intense band and two lower-intensity components on both sides with some resolved structure. The central line is well simulated by the hyperfine experimental tensors attributed to Radical A, determined by ENDOR in [11]. This dominant radical is centered on the C(2) carbon atom, and it is obtained by loss of a proton. The undamaged AT molecule is given by two structurally identical halves -OOC-CHOH-, but no inversion center is present, and therefore the two carbon atoms C(2) and C(3) are not equivalent. The ENDOR study showed that in Radical A the unpaired electron is localized on C(2) (Scheme 1). The hyperfine tensors used in the simulation are due to three protons a, b and c in Scheme 1, i.e., to two b-protons (H a and H b in Scheme 1) with respect to the unpaired electron, and one c-proton (H c ). The hyperfine tensor of the b-proton H b is small, due to the conformation of the radical with H b roughly in the nodal plane of the p z orbital of C(2) [11], where the unpaired electron is located. The sidebands show that other radicals besides Radical A are present, with larger hyperfine tensors. Let us assume that the sidebands are due to another radical, Radical B. In this paper, we try a preliminary determination of the nature of Radical B, by using cw-and pulse-EPR to disentangle the spectra.
The cw-EPR spectra shown in Fig. 1 were all recorded at a relatively low microwave power (10 -5 -10 -6 W). The samples were irradiated by different beams; nevertheless, they exhibit the same lineshape, meaning that the secondary species are produced with the same yields with respect to Radical A at the considered irradiation doses. By increasing the mw power the lineshape of the spectrum changes, as the relative intensity of the wings is enhanced with respect to the central band, showing different relaxation properties of the two radicals. The saturation level of Radical B is at higher mw power than Radical A, indicating faster spin relaxation times. This can be attributed to the larger anisotropy of the hyperfine tensors of Radical B, as discussed later. We made use of these findings to isolate the Radical B spectrum: a weighted difference between the previous low-power spectra and those acquired in the range 10 -3 -10 -4 W quenches the intensity of the central band, leaving a distorted spectrum that can be attributed to Radical B.
Pulse-EPR spectroscopy supplies another approach to separate the spectra of the two species, on the basis of the different rates of dephasing of their spin packets, by recording the echo-detected EPR spectra (ED-EPR). The main spin relaxation mechanism for Radical A is instantaneous diffusion (ID) [12], and we can safely presume that it is also the case for Radical B. The previous relaxation mechanism depends on the concentration of the spin excited by the pulse (fraction of the total concentration) and on the tilt angle induced by the microwave pulses on the onresonance magnetization [3]. At optimal conditions (highest echo intensity) the ID relaxation rate is faster for the species with larger concentrations. When this relaxation mechanism is dominant, for sufficiently long delays the most concentrated fast relaxing species can disappear below the noise level leaving only the contribution due to slower relaxing diluted species. This has been shown clearly in [16], where in some ED-EPR spectra the lines of the AT Radical A bearing a 13 C were more intense than those of the same radical with 12 C.
In Fig. 2 the ED-EPR spectra of AT powder sample irradiated with c-ray beam are reported as a function of the delay time, and of the microwave power. One can see the dramatic decay of the Radical A spectrum on increasing the delay time and the pulse power.
By following the lineshape change of the ED-EPR on varying the interpulse delay (see Fig. 2, left) we are confident that at the longest delays the contribution to the spectrum from species A is negligible, so that we are left with the spectrum of species B.
It is worth comparing the spectrum obtained by the difference of the cw-EPR spectra at high and low mw power with the ED-EPR one at high power. The two spectra are shown in Fig. 3, where the ED-EPR spectrum has been differentiated for an easier comparison.
The comparison between the cw difference spectrum and the ED-EPR spectra shows that the spectral profiles are very similar. Their differences can be attributed in part to the low spectral resolution of the ED-EPR spectra (around 0.2 mT). We note that ED-EPR are slightly wider as at the edge of the wings in the ED-EPR spectrum 13 C bands are visible, with a splitting above 5 mT, as already observed in [16]. This is a valuable result, by considering that we have obtained the two types of spectra with completely different EPR methods. It is worth noting that in both experiments we have taken advantage of the difference in spin relaxation rates between the two radicals; in cw-EPR the relaxation rates are faster for Radical B, whereas in ED-EPR the dephasing of spin packets is faster for Radical A. Actually, the faster relaxation in the first case can be attributed to the vibronic modulation of very anisotropic hyperfine tensors of the individual radicals B, whereas the faster spin dephasing in the second case is due to the pulse modulation of radical-radical interactions, in greater number for the more concentrated Radical A.
In [17] Sagstuen et al. investigated the primary radicals formed by irradiating the Na, K-salt of tartaric acid at low temperature, following their transformation to secondary stable radicals up to room temperature. Two distinct radicals were observed to grow during temperature annealing, bringing to two carbon centered radicals. One of them has the same structure as AT Radical A, whereas the other one has the following structure: -OOC-CH(OH)-CH-COO -(let us call it Radical BS). The hyperfine tensors of this radical were determined by Samskog et al. [18]. On the other hand, by room temperature irradiation of the same compound, Suzuki and Abe [19,20] observed the radical produced by the cleavage of the interior C-C bond, -OOC-CHOH (radical C).
On the hypothesis that our Radical B is the same as Radical BS studied in Rochelle salt by Samskog [18] and also observed by Sagstuen [17], we simulated the spectrum in Fig. 3 by using the hyperfine tensors of Radical BS (-78 -28 -50), (39 31 20) MHz corresponding to CH a and CH b (see Scheme 1, right). By best fit, we obtained for c proton (COH c ) the hyperfine tensor (24 16 6) MHz. The width and the main features of the spectrum are nicely reproduced. Radical C is not compatible with the experimental spectra obtained in Fig. 3, as the sum of the largest principal values relative to the two protons of the structure can account for a spectrum width of 4.2 mT, at most.
The relative percentage of the two species in the cw-EPR spectrum was obtained by measuring the contribution of the Radical B spectrum (spectrum black in Fig. 3) to the cw spectra, reported in Fig. 1. We calculated that the relative percentage of the Radical B to the total spectrum is about 35 %.

Conclusions
By cw-and pulse-EPR experiments we have shown that in the EPR spectra of ammonium tartrate irradiated with different beams two main secondary radicals are formed, radicals A and B. The EPR spectrum of the Radical A is well known; it is very narrow and used for EPR dosimetry. The second radical has been detected for the first time in this paper, it has a much wider EPR spectrum, and we attributed it to the radical NH 4 OOC-_ CH-CHOH-COONH 4 . Its percentage weight has been estimated to be 35 % of the total EPR signal. The presence of this radical should be taken into account for the use of ammonium tartrate as a dosimeter. The present study highlights the need of a thorough ENDOR investigation, to obtain information on the structure of the radical, and in particular on the localization of the unpaired electron, either on C(2) or on C(3), as the two halves of the ammonium tartrate molecule are structurally different in the crystal.