If you're using awk, you don't need grep. awk '/regex/ { print }' filenames is equivalent to grep "regex" filenames

I'm abusing awk's record-separator here, RS, so that it considers each < character to be a "newline". So it will split the two iframes apart itself, and "iframe src=" will always show up at the beginning of the 'line' if present, which I test with a regex /^iframe src=/. If a match is found, it rips the URL out of parameter $2 with a gsub, tests if we've printed it already, and if not, prints it.

Code:

find . -name "*php*" -or -name "*htm*" |
        xargs awk -v RS='<' '
# For all records beginning with 'iframe src':
/^iframe src=/ {
        # get rid of src= and "
        gsub(/(src=)|"/, "", $2);
        # Print only if we haven't seen it before
        if(!X[$2]++) print $2
}'