Organic light-emitting diodes (OLEDs) have been widely applicable in the flat-panel display products, such as mobile phones and television. They also possess more outstanding application advantages in the area of wearable display products in the future and have a great potential to serve as a key competitor in the next-generation solid-state lighting technology. For the commercial OLED products, phosphorescent materials that contain the rare metals are essential. This is because that phosphorescent OLEDs can give more efficient radiation, as the result of harvesting both singlet and triplet excitons for radiative utilization, in contrast to traditional fluorescent emitters. However, because of the limited resources of the rare metals, e.g., indium, on the earth, the phosphorescent materials are rather expensive, and their sustainable supply are also questionable. Besides, the long-term stability of blue phosphorescent emitters are also challenging. To further lower the fabrication cost of materials and devices, numerous research efforts have been devoted to developing rare-metal-free, purely organic light-emitting materials, which could be more low-cost and still exhibit excellent electroluminescence performances. In this review, we firstly introduce the evolution of OLED materials and devices. The OLED materials could be roughly classified into three generations. For the 1st generation traditional fluorescent emitters, three quarters of electro-generated excitons are wasted as heat release, due to the spin forbidden rule of triplets. This suggests that the OLEDs employing traditional fluorescent emitters could only achieve a maximum internal quantum efficiency of 25%. For the 2nd generation phosphorescent emitters, the spin orbital coupling due to heavy atom effect results in the mixing of singlet and triplet energy levels, and make triplet radiative transition feasible. Thereby, phosphorescent OLEDs can obtain a maximum internal quantum efficiency close to 100%. Recently, the 3rd generation hyperfluorescence emitters with the thermally activated delayed fluorescence (TADF) character could also make the devices thereof give an internal quantum efficiency close to 100%. This paves a road to realize low-cost and high-performance OLEDs based on purely organic semiconductors. We also outline the evolution of the OLED device configuration. For single-, double- and multi-layer device configuration of conventional OLEDs, their performances are mainly dependent on the emission layer (EML). The current methodology to develop high-performance OLEDs is to maximize the light emission ability of EML. Based on the understanding of the excitonic property of organic semiconductors, we show that efficient EML-free planar pn heterojunction light-emitting diodes (pn-OLEDs) should be realizable, given that the planar heterojunction property is well modulated. This indicates a way to directly construct efficient fluorescent OLEDs without an EML from the perspective of planar heterojunction composed of p type and n type charge transport materials. As following, the progress of OLEDs based on the layer-to-layer charge transfer mechanism using purely organic semiconductors is introduced. Main discussion focuses on pn-OLEDs, including that how this concept was proposed and the progress of relevant materials and devices. We also discuss the challenges of developing high performance pn-OLEDs and their potential applications in other optoelectronic areas, such as organic light-emitting transistors and electrically-pumped organic laser.