Fast radio bursts ( FRBs ) are usually suggested to be associated with mergers of compact binaries consisting of white dwarfs ( WDs ) , neutron stars ( NSs ) , or black holes ( BHs ) .
We test these models by fitting the observational distributions in both redshift and isotropic energy of 22 Parkes FRBs , where , as usual , the rates of compact binary mergers ( CBMs ) are connected with cosmic star formation rates by a power-law distributed time delay .
It is found that the observational distributions can well be produced by the CBM model with a characteristic delay time from several ten to several hundred Myr and an energy function index 1.2 \lesssim \gamma \lesssim 1.7 , where a tentative fixed spectral index \beta = 0.8 is adopted for all FRBs .
Correspondingly , the local event rate of FRBs is constrained to ( 3 - 6 ) \times 10 ^ { 4 } f _ { b } ^ { -1 } ( \mathcal { T } / 270 s ) ^ { -1 } ( \mathcal { A } / 2 \pi ) ^ % { -1 } ~ { } Gpc ^ { -3 } yr ^ { -1 } for an adopted minimum FRB energy of E _ { \min } = 3 \times 10 ^ { 39 } erg , where f _ { b } is the beaming factor of the radiation , \mathcal { T } is the duration of each pointing observation , and \mathcal { A } is the sky area of the survey .
This event rate , about an order of magnitude higher than the rates of NS-NS/NS-BH mergers , indicates that the most promising origin of FRBs in the CBM scenario could be mergers of WD-WD binaries .
Here a massive WD could be produced since no FRB was found to be associated with a type Ia supernova .
Alternatively , if actually all FRBs can repeat on a timescale much longer than the period of current observations , then they could also originate from a young active NS that forms from relatively rare NS-NS mergers and accretion-induced collapses of WD-WD binaries .