Proton therapy efficiency can be described as the ratio between tumour and non-tumour dose, while the tumour receives the planned dose. This efficiency is limited by the energy deposition property of the proton. To enhance the efficiency beyond this physical limit, targeted nuclear reactions during proton therapy could be exploited. For this purpose the 11B(p,3a) reaction has
been researched. The three alpha particles created have a high linear energy transfer (LET) and cause an increase in energy deposited at the reaction site. This reaction has a high cross section for low proton energies and protons have a low energy in the tumour area. Hence, proton 11B capture will more frequently occur in the tumour area, which can increase the tumour to non-tumour dose. Proton boron capture therapy (PBCT) has been studied, using Monte Carlo simulation and by conducting experiments, which suggested a 90 and 50% increase in energy deposited respectively. However, these results were contradicted by other research, stating the 11B(p,3a) reaction could
not significantly increase energy deposited during proton therapy. This discussion formed the foundation of this thesis. Two Monte Carlo methods, MCNP6 and Geant4, were used to investigate the reproducibility of the results obtained in previous research. First, MCNP6 simulations were done, which did not result in significant dose increase for proton therapy. To investigate the reproducibility of MCNP6 a second Monte Carlo method, Geant4, was used. Geant4 simulations produced similar results to MCNP6, equally not reproducing the promising results of previous research. Neither Monte Carlo method included all possible reaction cross sections for alpha particle creation during PBCT. Therefore, a simple Boltzmann model was made to simulate proton and alpha particle transport, including these missing reaction cross sections. To simulate the effect of below 1 MeV proton boron capture, a non-dynamic Boltzmann model was used, which resulted in a non significant increase in alpha production during PBCT. Afterwards, an alpha-proton-alpha avalanche reaction was investigated. For the effects of this reaction to be investigated a dynamic Boltzmann model was used, which resulted in no significant increase in alpha production either. To investigate if the 11B(p,3a) reaction could cause the increase in cell death found by experimental work, the cell killing potential of each alpha particle created was calculated analytically. Using the alpha production rate from the non-dynamic model, this calculation resulted in ten cells killed per alpha particle created. The alpha particles created during PBCT have a track length similar to the size of one cell, therefore it is unlikely for an alpha particle to kill ten cell. None of the obtained results, provide evidence that the 11B(p,3a) reaction will enhance the efficiency of proton therapy. Thus, we hypothesize that the experimental results could be related to boron having a radio sensitizing effect on cells. To test this irradiating boronated cells with gamma rays is proposed as the next step for further research.

Neutron capture therapy (NCT) is a binary form of radiotherapy that utilizes the high cross section of some nonradioactive elements, in particular boron-10 and gadolinium-157, for thermal neutron capture. The energetic charged particles (electrons and alpha particles) released in the capture reaction cause a high localized deposition. For almost a century, NCT has been regarded as the holy grail of targeted radiotherapy, but its superiority over other available treatments has not been demonstrated so far. Limitations included amongst others a poor and non-selective biodistribution of Boron carrier molecules, the limited availability of suitable neutron sources, a lack of tools for patient selection and individualized dosimetry, and a shortage of properly conducted clinical trials.

To obtain radionuclides for medical use from an isotope generator, mother and daughter nuclides need to be separated. In contrast to adsorption-based generators where the mother nuclide is retained on a static column, in liquid generators both mother and daughter nuclides are in solution and need to be separated via extraction. Microfluidic techniques are promising for the extraction process because mass transfer is very efficient and in theory the laminar two-phase flow in the microchannel can be easily separated at the end.

The generator of interest in this thesis is the W-188/Re-188 generator. However, some work also concerns the chemical similar Mo-99/Tc-99m generator and the Lu-177m/Lu-177 generator.

The start objective was to find a less water miscible replacement for the commonly applied organic phase methyl ethyl ketone (MEK) in the W-188/Re-188 generator, since with this system precipitation was observed previously in the microchannel. Several organic phases have been tested for precipitation, but also for their extraction efficiencies and abilities to separate rhenium from tungsten. The best results were observed for 0.2 M Aliquat 336 in 1,3-diisopropylbenzene, an organic phase that shows no precipitation, has suitable wetting behaviour in the microchannel and an even higher extraction efficiency than MEK.

Another issue addressed in this thesis is the incomplete phase separation at the Y-splitter (end) of the channel. Either some aqueous phase leaves through the organic outlet or some organic phase leaves through the aqueous outlet. One possible reason for the leakage is the uniform surface of both outlets, and therefore the preferred wetting of the hydrophilic glass wall of the microchannel with one phase. Coating one side of the microchannel hydrophobically showed that the leakage can be stopped. The effects of radiation on the coating have been investigated using an external gamma source and a threshold of 5 kGy before deterioration was found.

An even easier method to achieve complete phase separation is the use of a membrane separator.
Here, two phases are combined in simple 0.5 mm tubing in droplet fashion before being separated by a membrane. With the above mentioned organic phase, an extraction efficiency of 95% was reached within 5.3 minutes contact time. This is considerably longer than the typical contact time in the microchannel (in the order of seconds), but the total handling time was reduced from over 3 hours (with a microchannel) to 16 minutes by using a membrane separator. For lutetium, an extraction efficiency of 99% was reached after 2 minutes contact time, reducing the total handling time by a factor of 2.5 compared to the conventional method.