You can full_join a tibble of columns with only NAs:

library(tidyverse)

mtcars %>%
  as_tibble() %>% 
  full_join(tibble(mpg = NA, topspeed = NA)

# or if want to have all columns of another tibble
mtcars |> 
  select(cyl) |> 
  full_join(mtcars |> filter(FALSE))