We develop practical codes for achieving the capacity-energy function in simultaneously transmitting information and energy over a single noisy line. In particular, two polar coding techniques are developed for discrete memoryless channels. One technique involves concatenating nonlinear mappings with linear polar codes, whereas the other involves using randomized rounding to achieve the capacity-energy function. Further, capacity-energy achieving polar coding techniques for the additive white Gaussian noise (AWGN) channel with peak power constraint are also provided.