Reexamination of decoy-state quantum key distribution with biased bases

In order to improve the key rate of the decoy-state method, we need to jointly study yields of different bases. Given the delicate fact that pulses of the same preparation state can have different counting rates if they are measured in different bases, for example, those vacuum pulses and those single-photon pulses, existing results of decoy-state quantum key distribution using biased bases are actually flawed by assuming that they are equal. We fix this flaw through using the idea that yields of pulses prepared in different bases are the same provided that they are prepared in the same state and also measured in the same basis, for example, those single-photon pulses prepared in different bases but measured in the same basis. Based on this, we present correct formulas for the decoy-state method using biased bases Taking the effects of statistical fluctuations into account, we then numerically study the key rates of different protocols with all parameters being fully optimized Our result confirms the prior conclusion that the decoy-state method using biased bases can have an advantage over the symmetric protocol with unbiased bases. We obtain high key rates of our four-intensity protocol without using any vacuum source.