Introduction In the past several years there has been a growing interest in the development of radiopharmaceuticals labeled with metallic radionuclides (Anderson et al. 1999). Of particular interest is the positron emitter Cu-64 (t½ = 12.7 h) for molecular imaging of small molecules as well as peptides and antibodies (Smith 2004). This has led us to the recent implementation of a solid target production facility using commercially available target irradiation station and chemistry modules. Routine production of Cu-64 was achieved with an average production yield of 0.32 mCi/μAh, however purification of Cu-64 has proven to be problematic; with several metallic contaminants compromising subsequent radiolabeling. We report in this work, the step by step procedure which led us to the successful production of low metal contaminant 64Cu with high specific activity and high labeling efficiency. Material and Methods Detailed implementation of our solid target was reported earlier (Poniger et al. 2012). A Nirta Solid Target from IBA was coupled to our 18/9 cyclotron using a 2-meter external beam line. A pneumatic solid target transfer system (STTS) designed by TEMA was use to deliver the irradiated target disks to a dedicated hotcell. Modules from IBA (Pinctada metal) were used for electroplating 64Ni onto a Ag disk and for acid dissolution and purification of the irradiated target. Typical irradiation parameters were 14.9 MeV at 35 μA for 5–6 hours with 64Ni plating’s ranging from 10–60 μm thickness at 6–12 mm. Radionuclidic purities were evaluated by gamma spectroscopy and traces of metallic impurities were determined by ICP-MS or ICP-AES. Labeling efficiency was evaluated by measuring the amount of 64Cu uptake per 20 μg of scFv-cage. Results and Conclusion Initial 64Cu purifications following the manufacturers recommended method resulted in high levels of Cu, Fe and Zn metal contaminants (see TABLE 1, ID 1). Note that little Ag contamination is observed nevertheless the 64Ni is plated directly on a Ag disk. After several productions, visual inspection of the module quickly revealed that the heater block used for heating the back of the Ag target disk was heavily corroded. Replacing the copper heater block with a PEEK heater block drastically reduced the levels of Cu and Fe contaminants. Unfortunately unusually high levels of Zn were still observed regardless of the stringent conditions and ultrapure reagents used during the processing (see TABLE 1, ID 5). In our quest for answers, ICP-MS analysis of the 64Ni plating solution as well as critical stock reagents such as Milli-Q water (18 MΩ cm−1) and 30% HCl TraceSelect Ultra (Sigma) was performed (see TABLE 1, ID 2,3,4). The results were surprising, with high level of Zn found not only in the 64Ni plating solution, but as well in the HCl TraceSelect Ultra. It was hypothesized that the Pinctada’s glass bottles (Kay, 2004) used to store the reagents, especially concentrated acidic solutions were the source of Zn contamination and all glass bottles were replaced by LDPE or PFA types. Our hypothesis was confirmed by subsequent ICP-MS analysis of fresh samples of HCl TraceSelect Ultra and the 64Ni plating solution prepared/stored in plastic containers (see TABLE 1, ID 6,7). We also confirmed by ICP-MS analysis that no contamination occurred when performing a non-radioactive dissolution/purification sequence on the Pinctada module using a blank PTFE target disk in conjunction with the change to plastic reagent storage bottles (see TABLE 1, ID 8). Initially the purification protocol was modified as described by Ometakova et al., 2012 to help reduce the co-elution of Zn contaminants with the 64Cu from the AG1-X8 resin. This change resulted in a significant amount of 64Cu eluting from the resin during the resin washing steps, so that protocol was abandoned and the protocol as described by Thieme et al., 2012 was adopted. By modifying the AG1-X8 resin washing protocol to this new method and eluting the 64Cu from with 0.1M HCl rather than Milli-Q water (see TABLE 1, ID 9), we were able to further reduce metal contaminants, especially Zn. During the course of these experiments, the true specific activity of 64Cu increased from as low as 12 mCi/μmol of Cu (n = 2, TABLE 1, ID 1) to 649 mCi/μmol of Cu (n = 7, TABLE 1, ID 5) and finally to 4412 mCi/μmol of Cu (n = 3, TABLE 1, ID 9). In the same time, the effective specific activity increased from 0.03 ± 0.02 mCi per 20 μg of scFv-cage, to 3.7 ± 0.3 mCi per 20 g of scFv-cage with 64Cu. In conclusion, a significant reduction in Cu, Fe and Zn contaminants was achieved when processing 64Cu using the Pinctada module: i) after replacement of the Cu heater block; ii) after elimination of glass reagent storage containers from the Pinctada module and procedures during preparation of the 64Ni plating solution and iii) after implementation of a new purification protocol (Thieme et al. 2012). Introduction of a 6M HCl wash-up cycle of the module prior to the dissolution procedure was also effective. However in recent 64Cu productions slightly elevated Ag levels have been observed and are under investigation (see TABLE 1, ID 9).