2.1. Experiments regarding the PCDTBT:CdSe NPL blending ratio
The detector was evaluated while using PCDTBT:CdSe NPL hybrid active layers with various blending ratios. The amount of PCDTBT used was fixed at 5 mg while the blending ratio was adjusted by adding suitable amounts of CdSe NPLs to give ratios of 2:1, 1:1, 1:2, and 1:3. This was done to confirm optimal PCDTBT:CdSe NPL blending ratio to use in the active layer. First, the absorbance of the film state active layer and the CsI(Tl) scintillator emission spectrums according to the PCDTBT:CdSe NPL blending ratio are shown in Figure 6 (a). The absorption of the PCDTBT:CdSe NPL film tends to increase as the amount of CdSe NPLs increases. Figure 6 (b) shows the obtained X-ray detector parameters such as the collected current density (CCD) and sensitivity according to the PCDTBT:CdSe NPL blending ratio. Unlike the trend noticed with the absorbance, the highest sensitivity of 0.0566 mA/Gy∙cm2 was found with a 1:1 blending ratio. When the ratio of PCDTBT increases in the hybrid blend, the generated charge transfer decrease due to decreases in the amount of CdSe NPLs that can act as acceptors. In contrast, if the CdSe NPLs ratio becomes too high, trap sites that cause generated charge recombination are created. As such, the 1:1 blending ratio was found to be the optimum PCDTBT:CdSe NPL blending ratio.
To understand the differing trends between the detector’s absorbance and X-ray parameters, a comparison experiment was conducted that looked at defect density and RS. Figure 7 (a) shows a J-V curve recorded under dark conditions that can be used to calculate trap density, this done while representing the X and Y axes on a log scale and using the following defect density equation:
(4)
where Ndefects is the defect density, εr is the complex dielectric constant of the PCDTBT:CdSe NPL, ε0 is the permittivity of free space (8.854 X 10-4 F/m), q is the elementary charge (1.602 X 10-19 C), and L is the thickness of the PCDTBT:CdSe NPL film, which was measured using a surface profiler (KLA-Tencor Alpha-step AS-500). Also, VTFL is defined as the trap-filled limit voltage, this represents the voltage at the point where the trap-limited space-charge limit current (SCLC) and the trap-filled limit region overlap on the log scale J-V curve recorded under dark conditions. The extracted defect density and RS are shown in figure 7 (b). Since the defect density gives an indication of how many trap sites that cause charge recombination there are in the PCDTBT:CdSe NPL active layer, any reduction in defect density improves the stability of the detector [34]. When comparing the defect density according to the PCDTBT:CdSe NPL blending ratio, it was most significantly reduced at a ratio of 1:1 with a trap density of 1.02 x 1017 cm-3. In addition, RS, which affects the current transfer of the detector, was also the lowest at a 1:1 blending ratio (RS = 81 kΩ). Therefore, the hybrid PCDTBT:CdSe NPL blending ratio of 1:1 gives the lowest defect density and RS making it the optimal blending ratio.
2.2 Experiments regarding CdSe NPLs ligand exchange
To ensure the best charge transport conditions in the active layer, the optimal PCDTBT:CdSe NPL blending ratio of 1:1 was fixed for all further experiments. In our first experiment the oleic acid (OA) ligand on the CdSe NPL surface was exchanged with pyridine ligand. The OA ligand is a kind of long-chain ligand that helps with dispersion in a solution, but has the disadvantage of interfering with charge transport between NPL particles [35]. In contrast, as short-chain ligands such as pyridine, 3-mecaptopropionic acid (MPA), and 1,2-ethylenedithiol (EDT) shorten the distance between CdSe NPL particles, they thereby increase interactions with organic material and improve charge transport properties [36, 37]. Among these short-chain ligands, we chose to apply pyridine ligand in our experiments due to its good electrical conductivity. However, if the proportion of pyridine in the NPL surface is increased too much, the distance between the NPL particles becomes too small leading to aggregation. Considering these characteristics, experiments were performed to change the ratio of ligands by adjusting the ligand exchange time, thermogravimetric analysis (TGA) measurements were then performed to confirm the appropriate ratio of ligands. To proceed with the ligand exchange experiments, CdSe NPLs with OA ligand dispersed in toluene was mixed with pyridine in a 5:2 ratio, before being heated to 30℃ where it was maintained overnight. After that, the mixed solution was added to methanol in a 1:1 ratio and centrifugation was performed. At this point various ligand exchange times were used to produce different ligand ratios: 0, 6, 12, 18 and 24 hours. TGA measurements were then performed on the CdSe NPLs with OA ligand as well as on the CdSe NPLs after pyridine ligand exchange, the results of two conditions (0 and 12 hours) for comparison are shown in figure 8 (a).
In the CdSe NPLs with OA ligand sample, weight loss of the OA ligands starts around the boiling point at 286℃ and continues until 500℃. The total weight loss was equal to 26.36% of the starting weight, indicating the CdSe NPLs were covered by only OA ligands. The TGA curve for the CdSe NPLs with exchanged pyridine ligands has more than one step. The weight loss was 7.22% from the pyridine boiling point at 115℃ to the intermediate plateau temperature at 362.63℃ while the total weight loss below 362.63℃ was 19.53%. We estimated the fraction of CdSe NPL surface sites covered by pyridine and OA ligands [38]. The surface coverage of pyridine ligands was calculated to be 50.74% while the surface coverage of OA ligands was 38.43% after the12-hours ligand exchange time. All surface coverages according to the pyridine ligand exchange time are shown in Figure 8 (b).
The TGA measurements determined the OA to pyridine ligand coverage ratio according to the ligand exchange time and evaluated the detector with PCDTBT:CdSe NPL covered OA and pyridine ligand. Figure 9 (a) shows an active layer AFM image of the film’s state according to the pyridine ligand exchange time. The average surface roughness (Rq) increases rapidly as the pyridine ligand exchange time increases from the 6-hour exchange time and continues to increase over time since the distance between the particles becomes closer and closer while agglomeration occurs. As a photo and X-ray detector, RS, defect density, CCD, and sensitivity were evaluated and shown in Figure 9 (b). The RS trend of the detector changed because RS is affected by the defects present in the active layer. Therefore, the defect density according to the ligand exchange time was calculated and compared to the RS trend. As the ligand exchange time increased from 0-, 6-, 12-, 18-, and 24-hour, the RS recorded were 81.4, 74.8, 32.6, 42.8, and 51.4 kΩ, while the defect densities recorded were 10.2, 9.96, 9.62, 9.71, and 9.76 x 1016 cm−3, respectively. The defect density and RS trends were similar with the lowest RS and defect density coming after the 12-hour ligand exchange time. Since the electrical properties were also improved under the same conditions for the best RS and defect density, the CCD and sensitivity were also highest at 0.62935 µA/cm2 and 0.1444 mA/Gy∙cm2 after the 12-hour ligand exchange time. As such, the 12-hour ligand exchange time was identified as optimal and the spin-coating experiments were conducted using active layers that were fabricated using a 12-hour ligand exchange time.
2.3 Experiments regarding the spin-coating conditions of the PCDTBT:CdSe NPL hybrid active layer
Finally, experiments were conducted to find the optimal spin-coating conditions to use when fabricating the PCDTBT:CdSe NPL hybrid active layer. The thickness of the active layer was adjusted by setting the spin rate of the spin-coater to one of 1000, 1300, 1600, 1900, or 2200 rpm, the previous active layers were formed using a 1300 rpm spin rate. The thickness of the active layers were measured using a surface profiler (KLA-Tencor Alpha-step AS-500). The carrier mobility of the detector was also calculated to determine the charge transport characteristics according to the spin-coating condition of the active layers. The carrier mobility was calculated using the space charge limited current (SCLC) model, with a slope of 2 in the log scale J-V curves under dark conditions, and using the following Mott-Gurney equation [39, 40]:
(5)
where, J is the current density, L is the thickness of the PCDTBT:CdSe NPL active layer and Va is the applied voltage in the SCLC region, ε0 is the free-space permittivity constant of a vacuum (8.85 x 10-12 F/m) and εr is the relative permittivity of the applied material. The relative permittivity of the blended PCDTBT:CdSe NPL material was determined using the permittivity equation for blended materials [41] where the relative permittivity was calculated according to the spin rate. Figure 10 (a) shows the log-scale X, Y graph used to extract the mobility and defect densities at spin rates of 1300 and 1900 rpm. The defect density calculation is presented in section 2.1, while the mobility, defect density, thickness and detection sensitivity according to the spin rate are shown in Figure 10 (b). As the spin rate increases, the thickness tends to decrease from 90, 88, 84, 81, to 76 nm, respectively. The measured thickness were then applied to the mobility and defect density calculations, the results showed mobility tends to increase up to 1900 rpm before then decreasing as the rpm is further increased with respective calculated values of 3.92, 4.39, 5.83, 7.60, and 6.83 x 10-6 cm2/V∙s. Conversely, the defect density decreases up to 1900 rpm and then increases after with respective calculated values of 10.89, 9.64, 9.62, 9.31, and 9.40 x 1017 cm-3. Comparing the 1300 rpm pristine spin rate used with the 1900-rpm rate, the mobility improved by 73.12% while the defect density was reduced by 3.42%. These results suggest that adjusting the spin-coating condition of the PCDTBT:CdSe NPL hybrid active layer can significantly improve the charge transport, which makes using the optimal thickness important to realizing a high-performance detector. In addition, the PCDTBT:CdSe NPL hybrid detector was found to also achieve the highest sensitivity of 0.20 mA/Gy∙cm2 at a spin rate of 1900 rpm, giving a 39.13% improvement compared to the pristine spin rate of 1300 rpm.
In order to confirm the performance of an X-ray detector with a PCDTBT:CdSe NPL hybrid active layer, the linearity of the sensitivity was measured according to the applied voltage and absorbed dose during X-ray irradiation. The voltage applied to the PCDTBT:CdSe NPL hybrid detector was set in the range from -0.2 to -2.0 V, as shown in Figure 11 (a). It was found the CCD and sensitivity of the detector with the PCDTBT:CdSe NPL hybrid active layer tends to saturate at -1.4 V. The results of experiments conducted to look at the CCD-DCD as the absorbed dose of radiation was changed in the range 0.55 ~ 5.46 mGy are shown in Figure 11 (b), these were performed using an applied voltage of -1.4 V. As shown in Equation (3), the two values (CCD and DCD) should have a linear relationship, the coefficient of determination (R2) of the linear relationship for our detector was 0.994, while the slope of the fitted linear line was similar to the measured sensitivity of 0.20 mA/Gy∙cm2.