Prompt Gamma-Ray Activation Analysis For Food Supplements
Introduction
Prompt gamma-ray activation analysis (PGAA) is a valuable analytical method for a variety of fields including: geoscience, materials science,food science and medical applications. One such food science application is for the analysis of trace elements in foods, food supplements or food fortification. Every year, Americans spend upwards of 30 billion dollars on supplements. In order to ensure quality of these supplements, which are regulated by the FDA Dietary Supplement Health and Education Act of 1994 (DSHEA), external labs and tests can quantitatively ensure active ingredient concentrations. Boron is one such supplement that is commonly found in foods and is thought to play a role in bone growth and maintenance, wound healing, regulation of sex hormones, vitamin-d deficiency reduction, prevention of osteoarthritis and other important biological processes.
Boron quantification can be easily done using PGAA because it is a highly sensitive method for boron. It has been used to look at foods, steel, and a variety of other materials. Boron quantification can also be used for fluence monitoring, neutron self-shielding and scattering, and detector efficiency. TheoryPGAA looks at the emission of capture gamma rays in a nuclide of interest. PGAA is a type of activation analysis which is non-destructive and can look at samples on a wide range of scales. Following neutron irradiation, nuclei undergo radiative capture reactions and transform into a compound nucleus. The compound nucleus releases characteristic radiation, which can be measured to determine the original nuclei of interest. Another common type of activation analysis is neutron activation analysis (NAA), which looks at radiation, most commonly the decay gammas, coming from the radioactive daughter nuclei of certain elements.
Almost all elements emit capture gammas which causes a complicated energy spectrum to exist for not only the nuclide of interest but also from shielding materials and other materials inherent in the system. Consequently, the ability to measure these gamma rays depends heavily on the matrix of the element of interest. PGAA facilities are equipped with a neutron source with a well characterized flux. This can be accomplished using a neutron source, neutron generators or research reactors. In order to improve sensitivity, thermal or even cold neutrons are the preferred energy of interest since cross section obeys the inverse velocity law. Compared to NAA, prompt facilities deal with higher energy gamma rays. Decay gamma rays tend to be on the order of a few keV to 1 MeV while capture gammas are on the order of 1 to 10 MeV. This is one of the reasons that the detectors need to be well shielded to reduce damage to the detectors. This is also why the spectra are convoluted because at higher energies single and double escape peaks begin to muddy up the spectra. High-purity germanium detectors (HPGe) are often the detector of choice for PGAA because of the high resolution needed to evaluate the messy energy spectrum. In order to analyze peak data, results from the energy spectrum need to be compared to known data which has been extensively tabulated in the IAEA database.
In order to determine the elemental components of a sample, a standardization to a known must be completed. The first type of standardization is a relative method, where the gamma-ray peak count rates are standardized to the gamma-ray peak count rates of a standard sample. This method evaluates branching ratios and reaction rates of the sample with a standard under the same conditions as the sample. In NAA these relative standards are often added directly to the sample so the flux seen by these samples during irradiation is constant. However, the geometry of PGAA makes it such that the sample and standard must often be irradiated and counted separately so the variations in flux for each sample must be accounted, or normalized to. This is often completed using titanium or chromium flux foils. In order to determine the elemental mass of the sample, a ratio of the net peak area per live time per unit mass for the standard, also called the analytical sensitivity, must be compared to the net peak area per live time per unit mass for the sample.
A method of standardization that reduces error associated with flux variations and sample geometry is internal standardization. Varying amounts of internal standard are added to the sample to determine a calibration curve and the slope of this curve can be used to determine concentrations. A ratio of the analytical sensitivity of the sample and internal standard normalized to detection efficiencies is the commonly used k0 factor.
The PGAA facility at the University of Texas is located at one of the beam ports of a TRIGA Mark II research reactor with a thermal power of 1. 1MW. This facility was constructed in 1995 and was reconstructed in 2007 in order to reduce background interference. The updated schematic of the beamport is shown in figure 1. Following the reactor face, a six meter neutron guide which has a three hundred meter radius of curvature preferentially directs low velocity, i. e. low energy neutrons toward the sample of interest. A neutron counter is placed around this beam guide and is used to determine fluctuations in the neutron flux coming from the reactor, it should behave linearly with power. The neutron flux reaching the sample at reactor power of 950 kW was previously determined to be 5. 32x106cm2 s.
The HPGe detector is covered with a lead shielding cave to reduce background and to protect the detector. The detector is also covered with a Li2CO3 neutron absorber. This system is currently set up with a MAESTRO gamma spectrum acquisition and analysis software HV set to 0. 93 kV. Prior to determining a comparator method, an efficiency curve was taken using a Eu-152 source. This source was placed on top of the sample holder and was counted for 10 hours. A background, with the reactor off for two hours, was also performed and subtracted from the Eu-152 spectrum counts. Next, the linearity of the neutron counts with power were examined by performing neutron counts for one and two minutes with increasing reactor power. An energy calibration was done on the system using Co-60, Eu-152, Cs-137 and water sources. In order to qualitatively determine that the sample is in the beam, a layer of cadmium was added to the sample vial, as large as the proposed sample. The sample was then prepared by removing the top of two capsules and scoring the powder to the top of half of the capsule.
The Truth Nutra boron sample was weighed and added to the teflon sample vial. This sample is composed primarily of rice flour and contains some stearic acid and calcium borogluconate. The cellulose cap was not processed during sample prep. The sample was imaged using a NeutronOptics neutron camera in order to determine self-attenuation and processed using PHD Capture 1. 14. 0. 1 software. The sample was then irradiated for 2 hours while counting and with the neutron counter. The boron comparator, Inorganic Ventures, MSB-100PPM, was added to the cleaned teflon vial up until the same point as the sample. Then, the boron comparator was also irradiated for 3 hours while counting while the neutron counter was also counting. Lastly, a background with only the teflon sample holder was irradiated and counted with the neutron counter. GENIE Acquisition and Analysis software was then used to determine net counts for the Boron peak in the sample, comparator and background. These values were normalized to the neutron counts and live time as well. Results and AnalysisBefore running the samples, the neutron counter, Ortec 871, was calibrated and the efficiency curve was taken. The neutron counts were linear, as shown in figure 2, when reactor power was increased and one, two and three minute counts were taken. A phone timer was used to complete these counts.
Prior to determining a comparator method, an efficiency curve using Eu-152, using Source: DX 695, was completed. The source was measured to be 396 kBq on July 1st, 1994. Equation 1, an exponential decay equation was used to determine that at the time of use the sources activity was 115 kBq.
Absolute efficiency looks at the ratio of the total number of counts recorded by the detector at a given photopeak to the number of photons emitted by the sample at that given energy. Efficiency versus energy is plotted in figure 3 and highlights that for HPGe detectors efficiency decreases exponentially with energy. If the comparator is a different element, such as for a k0 method or an internal comparator method, the gamma-ray peak energy efficiencies would need to be compared for these two when determining the mass of the sample. This does give insight into how the efficiency of the detector is behaving despite not using these actual values. The efficiency curve could also be helpful with spectral analysis if ultimately deciding to further interpret the supplement content. Next, neutron images were taken in order to verify that the sample was covered by the beam. The cadmium sliver was placed in the sample holder to simulate the height of the sample. First, the cadmium is irradiated and undergoes a capture reaction. The neutrons impinge upon the scintillator screen shown in figure 4. The scintillator then converts the neutrons into visible light for the camera to interpret. The mirror is in place to ensure the neutrons that are not attenuated do not directly hit the camera. A similar image was taken of the sample itself to qualitatively determine self-attenuation.
The standard reference material chosen has a boron concentration of 100 ppm while the sample material has a concentration of boron of approximately 7500 ppm. Very few high concentration boron reference materials exist so this reference was the most appropriate option. However, self-shielding effects are significant with increasing concentrations of an analyte of interest. Boron atoms are significantly affected by self-shielding especially among different matrices. The reference material is a liquid composed primarily of water with some nitric acid, so hydrogen nitrogen and oxygen. While the sample is primarily rice flour with some stearic acid and calcium, so carbon, hydrogen and oxygen.
Further study would need to be completed to quantitatively determine this effect. Placing a neutron counter behind and in front of the sample, and then the standard, could be a method used to normalize this effect. The sample was prepared and weighed using 0. 66152±0. 00056 grams of the supplement powder from two supplement pills. The comparator was prepared and weighed using 1. 10914±0. 00093 grams of the liquid standard which was the same height as the sample in the vial. There is an error associated with this since it was difficult to be certain that this same height was reached for sample and comparator. This could be mitigated in further study by filling each to the top of the teflon vial instead of to the lip of the vial.
The mass of boron in the standard was 0. 111±0. 0008 mg and the counts were determined from the ROIs. The mass of boron in the supplement was 5. 65±0. 0056 mg when normalized to the neutron count and 5. 83 ±0. 0056 mg when normalized to the live time. Equation 3 was used for error propagation when variables were multiplied or divided. z=z*(xx)2 +(xx)2 equation 3
These values were compared to the averaged mass of 0. 383±0. 0003 g per pill. With our sample weight of 0. 66152±0. 00056 grams this would imply that there were 1. 727±0. 002 pills. This value should be less than two since the samples were scored at the top. According to the 3 mg of boron per supplement with a 10% uncertainty, there was an expected mass of boron of 5. 182±0. 518 mg. Figure 7 shows the expected and measured values and their associated uncertainties. The boron content of the supplement normalized to the counts are within the expected range. This is the more reliable value because neutron flux can change during time, even at constant power. Supplement content transparency makes it such that the 10% uncertainty on the label is arbitrary. A variety of studies have shown the large variations in labeled active ingredients in supplements. For instance, a study in Australia looking at supplement contents found an variations in precision from -13% to +61% on average. A study looking at multivitamins in the United States show that supplements are often “loaded” with extra active ingredients to ensure that there has not been degradation once it is consumed.
The confidence in the boron concentrations could be improved with a variety of further studies. First, self-attenuation was ignored here and could be studied to increase confidence. Incorporating the standard homogeneously into the sample could improve many of the associated uncertainties and matrix effects. Different matrices slow down the Li-7 nucleus to a different degree which ultimately determines the energy offset, or doppler broadening.
