Title: Prediction and identification of point defect fingerprints in the X-ray photoelectron spectra of TiN$_x$

Abstract: We investigate the effect of selected N and Ti point defects in $B$1 TiN on N 1s and Ti 2p$_{3/2}$ binding energies (BE) by experiments and ab initio calculations. X-ray photoelectron spectroscopy (XPS) measurements of Ti-deficient TiN films reveal additional N 1s spectral components at lower binding energies. Ab initio calculations predict that these components are caused by either Ti vacancies, which induce a N 1s BE shift of $-0.53$ eV in its first N neighbors, and/or N tetrahedral interstitials, which have their N 1s BE shifted by $-1.18$ eV and also shift BE of their first N neighbors by $-0.53$ eV. However, the {\it ab initio} calculations also reveal that the tetrahedral N interstitial is unstable at room temperature. We, therefore, unambiguously attribute the detected signal to Ti vacancies. Furthermore, the vacancy concentration in Ti-deficient TiN was quantified with XPS supported by ab initio calculations. The largest BE shifts of $-1.53$, $-1.80$ and $-2.28$ eV for Ti 2p$_{3/2}$ electrons are predicted for the Ti tetrahedral, split (10$\overline{1}$)-aligned and split (111)-aligned interstitial atoms, respectively, and we, therefore, propose XPS could detect them. Other defects such as N vacancy or N split (10$\overline{1}$)-aligned interstitial introduce smaller N 1s and Ti 2p$_{3/2}$ BE shifts and are unlikely to be detectable experimentally. Our work highlights the potential of ab initio-guided XPS measurements in detecting and quantifying point defects in $B$1 TiN.