How can I empose the ntree parameter into the train() function of caret package?

I am using the following function to do cross-validation with the random forest algorithm on my dataset. However, ntree raises an error, saying that it is not used in the function. Even though I have seen that usage as a recommendation comment before in one of the threads regarding this issue, it did not work at me. Here is my code:

cv_rf_class1 <- train(y_train_u ~ ., x_train_u , 
                      method ="cforest", 
                      trControl = trainControl(method = "cv", 
                                               number = 10, 
                                               verboseIter = TRUE),  
                                               ntree = 100))

If I cannot change the ntree parameter, it uses 500 trees as default in the function and it raises another error for me (subscript out of bounds), so I cannot make it work for my problem. How can I fix this issue in order to make my function work?