my own normalization with preg_replace

Question

I would like to build a normalization functions for strings:

'abcd  0123' -> 'abcd0123'
'Déjà vue' -> 'Dejavue'    //some specific characters are converted
'&§%-chars' -> 'chars'     //unrecognized characters are dropped

I used following code:

function normalize(string $text): string {
    $patterns = array('/[éèë]/i', '/à/','/ç/','/[^a-z0-9]/i');
    $replacements = array('e','a','c');
    return preg_replace($patterns, $replacements, $text);
}

The last pattern has no replacement, so it should replace with an empty string (wanted).

However, this is nog having the desirable effect:

'abcd  0123' -> 'abcd0123'  //ok
'Déjà vue' -> 'Deejevue'    //want 'Dejavue'
'&§%-chars' -> 'chars'      //ok

How can I correct my code to get the desired result?