The core ruling is stated plainly in Shulchan Arukh, Yoreh De'ah 274:7: a Sefer Torah that has been vocalized with nikud is invalid, and this remains so even if the nikud is later removed, and the same is true of a scroll with verse punctuation marked in ink.
Arukh HaShulchan, Yoreh De'ah 274:14-15 situates this within the broader requirement of 'ketivah temimah,' a complete and proper writing in which the letters, their forms, and the open and closed parshiyot must follow the received tradition exactly, and anything written otherwise is to be put away (geniza).
Mishneh Torah, Tefillin, Mezuzah and the Torah Scroll 7:8 likewise instructs the scribe to be careful with the large, small, and dotted letters and irregular letter-forms exactly as transmitted from scribe to scribe, and Tur, Yoreh De'ah 274 similarly requires copying precisely from an exemplar without deviation in the letters, spacing, or forms, reflecting the same demand for an unadorned, traditionally transmitted text.
The reason behind the prohibition is explained by Ramban and echoed by Rabbeinu Bahya, Bamidbar 11:15 and Ketem Paz on Zohar 104b:2: because nikud fixes a single reading and meaning for each word, vocalizing the Torah would close off the many layers of meaning inherent in the unvocalized text, which is precisely why the Torah was given without nikud so that it could be expounded in all its depths; this is also the basis given in Beit Yosef, Yoreh De'ah 274:15 citing the Ramban's responsum and Rabbeinu Yerucham, who root the disqualification in the principle that the text must stand only as it was given at Sinai, with the oral tradition of vocalization and cantillation transmitted separately, as reflected also in Nedarim 37b:5-6 describing punctuation and vocalization as components of the oral transmission accompanying the written text.
Rav Eliezer Melamed (Peninei Halakhah, Likkutim I 3:14) draws out the practical halachic conclusion from this body of sources: a Sefer Torah written with nikud or cantillation marks is invalid, because only the letters and the tagin belong to the written text, while the nikud and ta'amim belong to the orally transmitted tradition and may not be written in the scroll itself, so that the reader must instead learn the correct vocalization by heart.
Kitzur Shulchan Arukh Yalkut Yosef, Orach Chaim 36:4 extends this concern to practical scribal shortcuts, ruling that even a mere scratch mark made with a fingernail beside the letter shin to indicate its pronunciation (rather than true nikud in ink) requires that the Sefer Torah be corrected and the mark erased.
Binyan Tziyon 93:2-3 addresses a related practical case of a chazzan who scratched marks resembling cantillation signs into the parchment to ease his reading, including over sacred names, and affirms that such scrolls require correction, invoking the Shulchan Arukh's ruling (citing Rabbeinu Yerucham) that a vocalized scroll is invalid even once the vocalization is removed, while discussing whether such correction can actually restore validity.
Rav Menashe Klein (Mishneh Halachos 12:199) rules that a Sefer Torah written with nikud is invalid even once the nikud is removed, though nikud added after the writing may be erased so that the scroll becomes valid again, and he similarly forbids marking the scroll with cantillation signs since they alter the reading.
Addressing the specific modern practice of placing a transparent overlay marked with nikud and cantillation atop the scroll to assist an inexperienced reader, Rav Menashe Klein (Mishneh Halachos 11:242) finds this seemingly valid in principle but rules against doing so out of concern for the appearance of impropriety, preferring that established alternatives be followed, while Rav Ovadia Yosef (Maayan Omer 1:7:79) likewise takes up this question of a nylon overlay with nikud and te'amim, initially stating that it should not be done, and examines whether reading through such a transparent overlay is even considered reading from the written text at all.
Rav Betzalel Stern (B'tzel HaChochma 4:106) brings Rashi's and Tosafot's discussions on the significance of dotting in the Torah's text, including the tradition that Moshe would explain any dotted letter as one he had not erased, underscoring that dots and vocalization marks are treated as meaningful departures from the plain letters of the scroll.
Background sources on the sheet further reinforce the premium placed on precise, traditionally transmitted letter-forms and crowns in a Sefer Torah, such as the requirement of three crownlets on seven specific letters in Menachot 29b:6-7, and the broader demand in Shabbat 133b:4-5 that a Sefer Torah be written beautifully as an expression of 'zeh Keli ve'anvehu,' both of which reflect the same concern that the written text conform exactly to the received form rather than be altered for convenience.