[en] Compared to the standard Cu(In,Ga)Se${}_2$ (CIGSe) solar cells with a 2-3 μm thick absorber, ultrathin CIGSe SCs with less than 500 nm absorber thickness have the advantages of high-volume efficiency and less raw materials consumption. However, the reduced thickness of CIGSe causes insufficient light absorption and hinders the achievement of high efficiency for the ultrathin solar cells. Light trapping nanoparticles (NPs) can increase light absorption in solar cells. In addition, if the opaque Mo back contact was replaced with a transparent conductive material, the NPs can trap the light incident from the front and rear side simultaneously, which results in a bifacial semi-transparent ultrathin (BSTUT) CIGSe solar cell. In${}_2$O${}_3$: Sn (ITO) has good conductivity and relative high transparency, which fits the requirements of the back contact for BSTUT CIGSe solar cells. This thesis optimizes the front and rear photovoltaic (PV) performance of the ITO-based BSTUT CIGSe solar cells under three aspects, namely Na doping, back contact interface modification and light management. Firstly, four different Na doping methods are compared to decide about the optimal strategy for ultrathin CIGSe on ITO substrates. The four methods are Na diffusion from soda-lime glass (SLG), a NaF precursor, NaF post-deposition treatment (PDT), and a NaF precursor combined with PDT. When comparing the PV performance of the samples with different Na incorporation methods, the solar cells with NaF PDT doping exhibit the maximum enhancement compared to the reference (Na-free solar cell). In addition, the NaF PDT dose is optimized in detail and the resulting samples are characterized with multiple methods to explore the working mechanism of NaF PDT and the potential efficiency of BSTUT CIGSe solar cells. The NaF PDT mainly increases the doping density N${}_A$ in the CIGSe absorber and enlarges the contact potential difference V${}_D$ at the CIGSe/CdS interface. The NaF PDT can also increase the recombination velocity S${}_b$ and reduce the effective back barrier E${}_{CIGSe/ITO}$ at the CIGSe/ITO interface. Combining those two effects, the NaF PDT levels up the open circuit voltage V${}_{oc}$ of the solar cells, even though the short circuit current density j${}_{sc}$ is slightly decreased. We also verify this working mechanism of NaF PDT via SCAPS simulation. The average optimal efficiency Eff of the solar cells is 12.1% with 622 mV V${}_{oc}$, 29.6 mA/cm${}^2$ j${}_{sc}$, and 65.6% fill factor FF. Secondly, SiO${}_2$ point contacts are integrated at the CIGSe/ITO interface to modify the S${}_b$ at the back interface. For comparison, we use Mo back contacts as references for the SiO${}_2$ passivation effects. Consistent with our previous work, the point contacts increase the PV performance of the Mo-based solar cells. However, SiO${}_2$ passivation deteriorates the V${}_{oc}$ of our ITO-based ultrathin CIGSe solar cells. SCAPS simulations suggest that the barrier height E${}_h$ at the CIGSe/ITO interface decides about the effect of passivation (decreasing S${}_b$) of SiO${}_2$ for the ultrathin CIGSe SCs. According to the simulations, a decreasing S${}_b$ increases the effective barrier height E${}_{h,e}$ when E${}_h$ > 0.17 eV (Schottky-like contact), which means passivation is detrimental for the V${}_{oc}$ of the ultrathin CIGSe solar cells. The CIGSe/ITO interface is a Schottky-like contact, so the SiO${}_2$ point contacts decrease the performance of the ITO-based BSTUT CIGSe solar cells. On the contrary, a decreasing S${}_b$ increases the collection efficiency of photogenerated carriers when E${}_h$< 0.17 eV (quasi-Ohmic contact), so passivation benefits the solar cells. The decreased S${}_b$ increases the E${}_{h,e}$ slightly, but the overall E${}_h$ is small for the solar cells with a quasi-Ohmic contact. The improved collection efficiency of the photogenerated carriers dominates the passivation effects and benefits the Eff of the solar cells. The CIGSe/Mo interface reveals a quasi-Ohmic back contact, so the passivation increases the Eff of Mo-based ultrathin CIGSe SCs. Thirdly, the front and rear efficiency of BSTUT CIGSe solar cells are optimized using different substrates (SLG and alkali-free pgo glass), ITO thicknesses (100-400 nm) and various NaF PDT doses (0-8 mg). SLG-based solar cells show better front PV performance due to the extra incorporation of Na in the CIGSe co-evaporation process. However, solar cells on pgo glass show higher efficiency under rear illumination because alkali-free glass has a higher transparency than SLG, especially in the long wavelength range. The thicker ITO increases both the front and rear V${}_{oc}$ of the solar cells due to the Burstein-Moss shift in the ITO layer, which decreases the valence band offset ΔE${}_v$ at the CIGSe/ITO interface. However, the rear Eff is evened for solar cells on different thicknesses of ITO because thicker ITO also induces more sever parasitic absorption and leads to a lower rear j${}_{sc}$. For BSTUT CIGSe solar cells with different NaF PDT doses, the rear PV performance trend is similar to the one under front illumination. The solar cell with the optimal conditions (300 nm ITO, 4 mg NaF PDT) achieves 11.8% front Eff and 6.4% rear Eff. Fourthly, SiO${}_2$ nanoparticles (NPs) are inserted at the CIGSe/ITO interface to enhance the overall light absorption of BSTUT CIGSe solar cells. The NPs induce waveguide modes and enhance front and rear absorption in the ultrathin CIGSe layer. The NPs also induce jet-like forward scattering, which further increases the collection efficiency of photogenerated carriers under the rear illumination. Compared to the references, the front j${}_{sc}$ increases by 4.1-4.4 mA/cm${}^2$ and the rear j${}_{sc}$ by 6.4-7.4 mA/cm${}^2$ for the BSTUT CIGSe solar cells with SiO${}_2$ NPs. The front and rear V${}_{oc}$ gain of the solar cells with NPs can be quantitatively estimated by the relation between j${}_{sc}$ and V${}_{oc}$, which means the passivation effects of the SiO${}_2$ NPs are trivial compared to the dominating light trapping effects. Compared to state-of-the-art Mo-based ultrathin CIGSe solar cells (15%), our BSTUT CIGSe solar cells still have room for performance improvement, especially in V${}_{oc}$. The record V${}_{oc}$ is 733 mV for Mo-based and 635 mV for our ITO-based ultrathin CIGSe SCs. For the j${}_{sc}$, however, the record-high j${}_{sc}$ is 26.4 mA/cm${}^2$ for Mo-based and 31.1 mA/cm${}^2$ (front illumination) for our ITO-based SCs with light trapping SiO${}_2$ NPs, which shows an advantage of the BSTUT CIGSe solar cells. The bifacial Eff is 15.0% from summing up 100% front and 30% rear Eff of our best solar cell, which is close to the 15.2% record of the Mo-based ultrathin CIGSe solar cells. The findings in this thesis can help exploit solar energy with higher efficiency and lower fabrication cost. A summary and outlook will be presented at the end about how the efficiency of BSTUT CIGSe SCs could be further optimized.