Trap density

Water as an impurity accelerates the oxidation rate. Figure 4 compares growth curves for Si02 under dry and steam conditions. Halogens can also be introduced to the oxidation process, thereby reducing sodium ion contamination. This improves dielectric breakdown strength, and reduces interface trap density (15). 

Concerning the nature of electronic traps for this class of ladder polymers, we would like to recall the experimental facts. On comparing the results of LPPP to those of poly(para-phenylene vinylene) (PPV) [38] it must be noted that the appearance of the maximum current at 167 K, for heating rates between 0.06 K/s and 0.25 K/s, can be attributed to monomolecular kinetics with non-retrapping traps [26]. In PPV the density of trap states is evaluated on the basis of a multiple trapping model [38], leading to a trap density which is comparable to the density of monomer units and very low mobilities of 10-8 cm2 V-1 s l. These values for PPV have to be compared to trap densities of 0.0002 and 0.00003 traps per monomer unit in the LPPP. As a consequence of the low trap densities, high mobility values of 0.1 cm2 V-1 s-1 for the LPPPs are obtained [39]. 

Figure 9-27. Experimental (dots) and theoretical (solid line) t/V characteristics of. a Ca/PPV/Ca electron-only device with a thickness, L, of 310 nm. The theoretical curve is obtained assuming an exponential trap distribution with a trap density of Nt=5-I()17 cm 1, a trap distribution parameter Tt 1500 K, and an equilibrium electron density n = L5-I011 cm"1. The dashed line gives the hole SLC according to Eq. (9.13). Reproduced from Ref. 85J.

Apart from fundamental constants and the liquid temperature, the variable parameters in the effective mobility equation are the quasi-free mobility, the trap density, and the binding energy in the trap. Figure 10.2, shows the variation of prff with e0 at T = 300 K for /tqf = 100 cm3v 1s 1 and nt = 1019cm-3. It is clear that the importance of the ballistic mobility (jl)l increases with the binding... 

In comparing the results of the quasi-ballistic model with experiment, generally pq[ = 100 cn v s-1 has been used (Mozumder, 1995a) except in a case such as isooctane (Itoh et al, 1989) where a lower Hall mobility has been determined when that value is used for the quasi-free mobility. There is no obvious reason that the quasi-free mobility should be the same in all liquids, and in fact values in the range 30-400 cmV -1 have been indicated (Berlin et al, 1978). However, in the indicated range, the computed mobility depends sensitively on the trap density and the binding energy, and not so much on the quasi-free mobility if the effective mobility is less than 10 crr v s-1. A partial theoretical justification of 100 cm2 v 1s 1 for the quasi-free mobility has been advanced by Davis and Brown (1975). Experimentally, it is the measured mobility in TMS, which is considered to be trap-free (vide supra). 

As for the trap density, a lower limit of 1018 cm-3 has been taken, based on the fall of trapped electron yield in hydrocarbon glasses at a dose 1020 eV/gm (Willard, 1975). An upper limit of trap density 1020 may be argued on the basis of Berlin and Schiller s (1987) finding that a quasi-free electron interacts with... 

Table 10.4 lists the values of trap density and binding energy obtained in the quasi-ballistic model for different hydrocarbon liquids by matching the calculated mobility with experimental determination at one temperature. The experimental data have been taken from Allen (1976) and Tabata et ah, (1991). In all cases, the computed activation energy slightly exceeds the experimental value, and typically for n-hexane, 0/Eac = 0.89. Some other details of calculation will be found in Mozumder (1995a). It is noteworthy that in low-mobility liquids ballistic motion predominates. Its effect on the mobility in n-hexane is 1.74 times greater than that of diffusive trap-controlled motion. As yet, there has been no calculation of the field dependence of electron mobility in the quasi-ballistic model. 

TABLE 10.4 Electron Mobility, Trap Density, Binding Energy, and Activation Energy in the Quasi-ballistic Model... 

In this way, the trap depth of the poly-N-vinylcarbazole was calculated as aE =0.56 eV and from the area of the 5°C peak the trap density was estimated to be of the order of 7 x 10l3 cm-3. 

VH =2x10 m /mol 1 and the solubility 0.005435 mol H,/m >/MPa. The molar volume of the material is 7.116xl0 m3/mol and this corresponds to a density iVt =8.46xl028atoms/nv. The parameters fi and a were set equal to 1. For the trap density Nr, we assumed that it increases with plastic strain according to the experimental results of Kumnick and Johnson15 which also indicate a trap binding energy of 60 kJ/mol. 

Chung, G. Y., et al., Effect of Nitric Oxide Annealing on the Interface Trap Densities Near the Band Edges in the 4H Polytype of Silicon Carbide, Applied Physics Letters, Vol. 76, No. 13, March 27, 2000, p. 1713. 

By optically creating carriers with a pulse of above band-gap illumination, then monitoring the subsequent current transient due to thermal detrapping, Hurtes et al (1978) and Fairman et al (1979) were able to apply the DLTS method to bulk, high-resistivity materials. This method, however, is unable to distinguish between electron and hole traps, and the calculation of trap densities is difficult. 

The TSC signal strength will be proportional to the appropriate trap density, as seen in Eq. (55). Unfortunately, the lifetime t (or xp for hole emission) also enters the equation, and this quantity will depend on other traps and recombination centers in the sample. If t could be separately determined, then the TSC method could be calibrated for trap density. Essentially the same considerations hold for the transient-current methods (PITS and OTCS), as seen in Eq. (65). A further complication enters if eni and epi are of comparable magnitude, and ejepi is unknown [cf. Eq. (83)]. 

The lifetime T and diffusion coefficient D of photoinjected electrons in DSC measured over five orders of magnitude of illumination intensity using IMVS and IMPS.56) fis proportional to the r m, indicating that the back reaction of electrons with I3 tnay be second order in electron density. On the other hand, D varied with C0 68, attributed to an exponential trap density distribution of the form Nt(E) <=< exp[ P(E - Ec)l(kBT) with 0.6. Since T and D vary with intensity in opposite senses, the calculated electron diffusion length L = (JD-z)m does not change linearly with the irradiance. 

The photorefractive gain, shown in Figure 2b, is predominantly determined by two factors the trap density, NE, and the relative conductivity factor, . The latter factor accounts for the minimum... 

---