Radiation damage formation in InP, InSb, GaAs, GaP, Ge, and Si due to fast ions

The basic physical mechanisms of damage formation in semiconductors due to swift heavy ion (SHI) irradiation are not yet fully understood. In the present paper damage evolution and the formation of ion tracks during SHI irradiation in InP, InSb, GaAs, GaP, and Ge are investigated for irradiation with Xe or Au ions having specific energies ranging from about 0.8 to 3 MeV/u. Based on these experimental results and those obtained by other authors for cluster-ion irradiation of InP, GaAs, Ge, and Si, extensive calculations were performed in the framework of the thermal spike model. As we published previously, the model was extended to correctly treat processes being specific to semiconductors. Additionally, the computer code was modified to perform calculations for cluster ions too. The calculated track radii are compared with those measured for the various irradiation conditions. Thereby, one unknown parameter in the calculations was determined by fitting one data point. With this procedure a very good agreement between calculated and measured track radii is obtained for InP, Ge, and Si irradiated under various conditions. This implies that the extended thermal spike model is well capable to explain track formation in SHI irradiated semiconductors. Furthermore, special experiments were performed, the results of which also support the thermal spike model of ion track formation and contradict competing mechanisms such as Coulomb explosion, shock waves, or lattice relaxation. Thus, visible (amorphous/heavily damaged) ion tracks occur if the electronic energy deposition per ion and unit length clearly exceeds the threshold value necessary for melting. This is possible for elemental ion irradiation of InP and InSb. In Ge and Si (and probably also in GaAs and GaP) the energy deposition necessary for melting is that high that it cannot be reached by elemental ion irradiation. Moreover, at least in InP and GaAs ion tracks can be formed also in a subthreshold irradiation regime if the material is predamaged. This suggests that the existence of point defects and clusters of point defects in the crystal lattice noticeably increases the electron-phonon coupling efficiency, resulting in a more efficient energy transfer to the lattice. Within the thermal spike model, this means that for a given electronic energy deposition in the predamaged crystal a higher temperature is reached than in the perfect one.

Figure 1

(Color online) Relative concentration of damage, $n_{da}$, taken at depth of 200 nm versus ion fluence $N_I$. The experimental data are for InSb, InP, GaAs, GaP, and Ge irradiated with 593 MeV Au at RT. In all cases the data points can be fitted by sigmoid curves, which are plotted in the figure.

Figure 2

Bright field PV-TEM images of virgin InP samples irradiated at RT with either (a) 375 MeV Xe up to the ion fluence of $8.4\times {10}^{11}{\text{cm}}^{-2}$ or (b) 573 MeV Au up to the ion fluence of $1\times {10}^{11}{\text{cm}}^{-2}$. In both cases the InP samples were tilted off the $⟨100⟩$ axis in the microscope.

Figure 3

Bright field PV-TEM images of a virgin InSb sample irradiated at RT with 573 MeV Au up to the ion fluence of $5\times {10}^{10}{\text{cm}}^{-2}$. Part (a) was obtained by TEM analysis along $⟨100⟩$ direction, while (b) stands for the sample tilted $17°$ off the axis in the microscope.

Figure 4

Bright field PV-TEM images of either virgin (a)–(c) or conventionally predamaged (d) GaAs samples irradiated with 573 MeV Au up to the ion fluences of (a)–(b) $1\times {10}^{12}{\text{cm}}^{-2}$ at RT, (c) $1\times {10}^{12}{\text{cm}}^{-2}$ at LNT, and (d) $5\times {10}^{10}{\text{cm}}^{-2}$ at RT. Parts (a), (c), and (d) were obtained by TEM analysis along $⟨100⟩$ direction, while (b) stands for the same sample as shown in (a) but tilted 30° off the axis in the microscope. The dark straight lines in (b) indicate the incidence direction of the Au projectiles. The dark arrows in all parts of the figure indicate some tracks.

Figure 5

Bright field PV-TEM image of GaP sample irradiated with 573 MeV Au up to the ion fluence of $1\times {10}^{12}{\text{cm}}^{-2}$ at RT.

Figure 6

(Color online) SIMS results on the intermixing in Bi/GaAs and Bi/InP layered structures due to 593 MeV Au irradiation ($5\times {10}^{13}{\text{cm}}^{-2}$ at RT). The upper and the bottom parts of the figure stand for Bi/GaAs and Bi/InP layered structures, respectively. The bottom part shows also the respective profiles of Bi for comparison. The small peaks at the surface (at $t=0$) are inevitably caused by the transitional character of the target sputtering at the very beginning.

Figure 7

(Color online) Experimental damage cross section, $S_{\text{expt}}$, versus the electronic energy loss, ${\epsilon }_e$, for ${\text{C}}_{20}$ or ${\text{C}}_{60}$ cluster ions. The data on InP are from Refs. 63, 68, the data on GaAs and Ge are from Refs. 67, 64, respectively, and the data on Si are from Refs. 65, 66.

Figure 8

(Color online) Dependence of the maximum atomic temperature at the ion track axis $\left(r=0\right)$ on the energy of fast Au ions. All results are for RT irradiations. The data for InP are from Ref. 51.

Figure 9

(Color online) Track radii calculated by our program HEAT $\left(R_{\text{calc}}\right)$ versus experimental (TEM) values of track radii $\left(R_{\text{expt}}\right)$ for RT irradiations of InP and Ge with various ion species and energies. The experimental data on the ${\text{C}}_{60}$ cluster-ion irradiations are from Refs. 63, 64, 68.

Figure 10

Bright field PV-TEM image of virgin InP sample irradiated at RT with 573 MeV Au up to the ion fluence of $1\times {10}^{11}{\text{cm}}^{-2}$. The TEM analysis was performed in the $⟨111⟩$ direction. The inset in the top right corner shows the corresponding diffraction pattern.