Introduction Highly sensitive semiconductor gas sensors hold great potential for applications in trace gas detection. Reliable detection of ppb-level NO2 is crucial for environmental monitoring, which however still remains a challenge. In this work, we adopted a one-pot encapsulation strategy for synthesizing indium-doped ZnO porous hollow cages. In(acac)3 was selected as the indium source due to its molecular diameter between the cavity size (11.6 Å) and the aperture size (3.4 Å) of ZIF-8. During the crystallization progress of ZIF-8, it was in situ trapped in the cavities and In(acac)3@ZIF-8 composites were formed (Scheme a, b). During calcination, metal ion species, composed of In(acac)3@ZIF-8, were oxidized, leading to the formation of In-doped ZnO porous hollow cages (Scheme c, d). In this case, ZIF-8 served as the host backbone for the uniform distribution of the indium source as well as the production of porous hollow structure. In doping led to significantly enhanced NO2-sensing performance, achieving a response of 3.7 to 10 ppb of NO2, a high sensitivity of 187.9 ppm-1, and a limit of detection as low as 0.2 ppb. The mechanism of enhanced NO2-sensing properties was discussed in the terms of the intrinsic excellent gas accessibility of porous hollow structure and electronic sensitization by In doping. Our findings could be extended to design other porous doped ZnO oxides for high performance gas sensors and other applications. Method In(acac)3@ZIF-8 composites were synthesized by a one-pot encapsulation method. route. Briefly, 810 mg of Zinc nitrate hexahydrate and corresponding amounts of Indium(III) acetylacetonate were dissolved into 40 mL methanol to form precursor solutions. 40 mL of methanol solutions containing 721.6 mg of 2-methylimidazole were added into above solutions under stirring. The resulting solutions were stirred for 20 min and aged for 24 h at room temperature. After precipitation, obtained ZIF-8 derivatives were washed with methanol via centrifugation. The precipitates were further dried at 60 °C for 24 h and calcined at 500 °C for 2 h to obtain pristine and In-doped ZnO powders. The obtained samples are hereafter denoted as In/ZnO-X, where X represents the nominal In/Zn atomic ratio (multiplied by 100) used during the synthesis, e.g., In/ZnO-0 for un-doped ZnO, In/ZnO-2 for In/Zn of 2.0 at.%, In/ZnO-10 for In/Zn of 10.0 at.%. Results and Conclusions Figure 1a depicts dynamic electrical resistance changes of pristine and In-doped ZnO sensors in the presence of NO2 from 0.2 to 10 ppm at 300 °C. The sensors showed typical n-type semiconductor response behavior, the increase of electrical resistance in the presence of NO2. In addition, the electrical resistance increased with the increase of NO2 concentration for all the sensors, and at each concentration the increment of electrical resistance for In/ ZnO-10 was always much higher than that for other sensors. Further investigations on the performance of In/ZnO-10 in the NO2 concentration below 0.2 ppm were conducted. Figure 1b shows the dynamic response and recovery curves for the In/ZnO-10 sensor, which were measured by alternately exposing the sensor to ambient air and air/NO2 mixture at 300°C. The In/ZnO-10 sensor was effective for the detection of NO2 at 10, 25, 50, and 100 ppb, and a high response of 3.7 to 10 ppb of NO2 gas was obtained. And the calculated limit of detection was as low as 0.2 ppb, 2 orders of magnitude lower than that (23.4 ppb) of pristine ZnO. Therefore, the In/ZnO-10 sensor could be suitable for the detection of ppb-level NO2 put forward by the WHO for monitoring environmental pollution. Figure 1c exhibits the response of pristine and In-doped ZnO sensors as a function of NO2 concentration in the range below 1 ppm at 300 °C. The response value increased linearly with the increasing NO2 concentration, and the slope (i.e., sensitivity) also became larger with increasing In content. A large sensitivity of 187.9 ppm-1 was obtained for the In/ ZnO-10 sensor, 117.4, 9.5, and 9.1 times larger than that for pristine ZnO (1.6 ppm-1), In/ZnO-2 (19.8 ppm-1) and In/ZnO-6 (20.7 ppm-1), respectively. The wide linear dependence and high sensitivity of the In/ZnO-10 sensor are favorable for calibration and determination of the NO2 concentration in practical applications. Further tests at higher concentrations above 1 ppm showed the slope decreased with the increasing NO2 concentration, indicating the sensors were near saturation (Figure 1d). Nevertheless, the In/ZnO-10 sensor still possessed the highest response value and sensitivity among the sensors in the high-concentration range.