Why do I get significantly different accuracy result when I train the same model with 2 different methods in R (rpart and train)

Question

When I use the train() function from the caret package as follows:

model1 = train(class~., data = training_data, method = "rpart")

and t