Predictive Monte Carlo ion implantation simulator from sub-keV to above 10 MeV

Abstract

In this paper is reported a general and accurate binary-collision-approximation- (BCA-)based Monte Carlo ion implantation model for implants into crystalline silicon. The combination of an improved semiempirical electronic stopping power model and Ziegler-Biersack-Littmark universal potential enables us to simulate a wide variety of implant species with only two different electronic stopping parameters for different implant species. With the model parameters fixed for a given implant species, excellent agreement is found with experimental secondary ion mass spectroscopy data for the energy range from sub-keV to above 10 MeV, and for different implant directions including random equivalent orientation, 〈100〉, 〈111〉, and 〈110〉 channeling directions. When compared with other BCA-based Monte Carlo simulators, it is demonstrated that more accurate results can be obtained for ultralow energy and very high energy implants. Furthermore, it is shown that, while the existing ion implantation simulators with the electronic stopping power based on the effective charge theory fail to predict the long tails of the deeply channeled implant species (such as Al), our model can predict these long tails successfully. Finally, an efficient damage model is also presented, which requires only one additional free parameter to accurately account for the damage accumulation and dechanneling effect. For high dose implants, substantial speed improvement over MARLOWE-based Monte Carlo simulators is observed.

Partitioning to elastic and inelastic processes of the energy deposited by low energy ions in silicon detectors

“…This was first recognized by Hobler et al [24]. It was used in their IMSIL code and further improved by Tian [25]. In these works, the model uses for the local part x loc , the formula of Oen and Robinson [18] and for the nonlocal part x nloc -the LSS model.…”

“…2, we used the (dE/dx) L results of LSS calculated with (1) as the basic data with which we compare dE/dx of other models by plotting (for E 0 < 100 keV) the ratio R L = (dE/dx)/(dE/dx) L as function of Z 1 for Z 2 = 14. It is done for (i) the results of Land and Brennan (LB) [26] where the improved local model of [17] was used, (ii) the data of SRIM-03 [10] and (iii) the values of c of Tian [25]. One can see the strong oscillatory dependence on Z 1 of dE/dx of [26] and SRIM-03.…”

“…Such results are the implantation profiles, both for amorphous and crystalline phases. According to [25] and using (dE/dx) L of Eq. (1),…”

Characterization of amorphous and crystalline silicon nanoclusters in ultra thin silica layers

“…In order to determine the screening effect of the surrounding matrix on the clusters, the universal "screening length" [24] was calculated for Er-oxide, Pd, Pd-oxide, and SiO 2 . The universal screening length is expressed as…”

“…This discovery resulted in a demand for using Ge in solid state electronics devices [24]. Ge has favourable properties utilized in several optic and electronic devices, such as photo detectors and single transistors [25,26].…”