> tibble(
+   x=1:5,
+   y=list(c('a','b'),c('a'),c("u","x","z"),c('a','b'),c('a'))
+ )  ->jj
>   as.data.table(jj)[, .(x = x, y =  as.data.table(transpose(y)))]
   x y.V1 y.V2
1: 1    a    b
2: 2    a     
3: 3    u    x
4: 4    a    b
5: 5    a     
Warning messages:
1: Element 2 must be length 2, not 1 
2: Element 3 must be length 2, not 3 
3: Element 5 must be length 2, not 1

> library(tibble)
> library(data.table)
data.table 1.13.2 using 8 threads (see ?getDTthreads).  Latest news: r-datatable.com
> 
> xx <- tibble(
+   x=1:5,
+   y=list(c('a','b'),c('a'),c("u","x","z"),c('a','b'),c('a'))
+ )
> 
> 
> as.data.table(xx)[, .(x = x, y = as.data.table(transpose(y)))]
   x y.V1 y.V2 y.V3
1: 1    a    b <NA>
2: 2    a <NA> <NA>
3: 3    u    x    z
4: 4    a    b <NA>
5: 5    a <NA> <NA>
> packageVersion("data.table")
[1] '1.13.2'

pmt  %>% unnest(check) %>% group_by(tr_id) %>% mutate(rank=row_number()) %>% pivot_wider(c(-rank,-check),names_from = rank,values_from = check))

 user  system elapsed 
  2.027   0.025   2.053 

system.time({
  tmp_dt <- as.data.table(pmt);
  tmp_dt <- tmp_dt[, list(vals = unlist(check)),by=tr_id];
  tmp_dt[, index := 1:.N, by =tr_id];
  dcast(tmp_dt, tr_id ~ index, value.var = 'vals', fill = NA) })

   
user  system elapsed 
  0.185   0.005   0.189 

system.time(
  as.data.table(pmt)[, .(x= tr_id, y = as.data.table(transpose(check)))])

 user  system elapsed 
  0.086   0.006   0.093