You don't seem to understand what you are doing. You index the reference genome with bwa. You align your fastq to the indexed genome with bwa. The .sai file is an intermediate file, it (or the two .sai files from a paired end experiment) are converted into a .sam file

Samtools doesn't do any of that.

Though piping that .sam file straight into samtools for compression into .bam format is a smart thing to do.

Also note that indexing the reference genome with bwa, and indexing a .bam file with samtools are two entirely different things. You might not even need the .bam file index.

Before you use samtools you need to map your reads first. Then you get a sam/bam file based on your aligning tool. Only after that you use that bam/sam file to index it.

To index it, first you sort the bam/sam files.

samtools sort sample.bam sample_sorted

sample.bam here is your input file and sample_sorted here is the output file in the bam format too. Notice you don't mention the .bam extension in the output file name.

Then, you index it using the following command.

samtools index sample_sorted.bam