I use this function to encode all the xml entities and also all the &something; that are not defined in xml like ™
You can also decode what you encode with my decode function.
My function works a little like the htmlentities.
You can also add other string to the array if you want to exclude them from the encoding.

<?php
function xml_entity_decode($text, $charset = 'Windows-1252'){
    // Double decode, so if the value was &amp;trade; it will become Trademark
    $text = html_entity_decode($text, ENT_COMPAT, $charset);
    $text = html_entity_decode($text, ENT_COMPAT, $charset);
    return $text;
}

function xml_entities($text, $charset = 'Windows-1252'){
     // Debug and Test
    // $text = "test &amp; &trade; &amp;trade; abc &reg; &amp;reg; &#45;";
    
    // First we encode html characters that are also invalid in xml
    $text = htmlentities($text, ENT_COMPAT, $charset, false);
    
    // XML character entity array from Wiki
    // Note: &apos; is useless in UTF-8 or in UTF-16
    $arr_xml_special_char = array("&quot;","&amp;","&apos;","&lt;","&gt;");
    
    // Building the regex string to exclude all strings with xml special char
    $arr_xml_special_char_regex = "(?";
    foreach($arr_xml_special_char as $key => $value){
        $arr_xml_special_char_regex .= "(?!$value)";
    }
    $arr_xml_special_char_regex .= ")";
    
    // Scan the array for &something_not_xml; syntax
    $pattern = "/$arr_xml_special_char_regex&([a-zA-Z0-9]+;)/";
    
    // Replace the &something_not_xml; with &amp;something_not_xml;
    $replacement = '&amp;${1}';
    return preg_replace($pattern, $replacement, $text);
}
?>