Deep level transient spectroscopy (DLTS) is used extensively to study defects in semiconductors. We demonstrate that great care should be exercised in interpreting activation energies extracted from DLTS as ionization energies. We show how first-principles calculations of thermodynamic transition levels, temperature effects of ionization energies, and nonradiative capture coefficients can be used to accurately determine actual activation energies that can be directly compared with DLTS. Our analysis is illustrated with hybrid functional calculations for two important defects in GaN that have similar thermodynamic transition levels, and shows that the activation energy extracted from DLTS includes a capture barrier that is temperature dependent, unique to each defect, and in some cases large in comparison to the ionization energy. By calculating quantities that can be directly compared with experiment, first-principles calculations thus offer powerful leverage in identifying the microscopic origin of defects detected in DLTS.PACS numbers: 72.20.Jv, 84.37.+q Point defects and impurities are present in all semiconductors. They can act as recombination centers that lower the efficiency of optoelectronic devices, or as carrier traps in electronic devices such as transistors. Microscopic identification of the detrimental defects is crucial in order to mitigate their impact. Deep level transient spectroscopy (DLTS) is a powerful technique for determining the properties of defects; from an analysis of electrical measurements on a pn junction or Schottky diode, properties such as the position of the defect level within the band gap, electrical nature (donor or acceptor), density, and carrier capture cross section of specific defects can be obtained. [1][2][3] Translating this wealth of information to a microscopic identification of a given defect requires comparison with theoretical or computational models, and first-principles calculations based on density functional theory (DFT) have proven very helpful. [4][5][6][7][8] One of the key quantities measured in DLTS is the activation energy for carrier emission from a defect, ∆E a . Defect identification is often based on comparing ∆E a with values of the defect ionization energy ∆E i determined from zero-temperature first-principles calculations. However, the underlying theory of DLTS 2,3 makes clear that ∆E a and ∆E i are distinct, and the use of ∆E i can affect the correct identification of a defect.In the present study we describe a first-principles approach to explicitly determine the activation energies measured in DLTS. Recent advances have enabled the quantitative prediction of defect levels in the band gap 9 and the ability to accurately describe nonradiative carrier capture. 10 We will show that ∆E a can significantly differ from ∆E i for some defects, demonstrating the need to explicitly calculate the activation energy in order to correctly identify defects detected by DLTS. Our analysis is general, but will be illustrated with examples of deep defects in GaN...