In Figure 1 the transmittance spectra of the commercial conductive glasses are shown. As can be seen from this Figure, two aspects are relevant for these commercial conductive glasses: (a) T15 has a higher transmittance than T7, which is derived from the lower thickness. As known, T~ exp(-αd), where α is the absorption coefficient and d the thickness. On the other hand, the sheet resistance varies with the inverse of the thickness (R_sh=ρ/d).

Figure 1

A TCO with a sheet resistance of 7 (/sq. is thicker than a TCO of 15 (/sq. Therefore it will have a smaller transmittance, in correspondence with Figure 1, the thermal treatment does not substantially change the transmittance with respect to the as-grown samples. A band gap value of 3.76 eV was estimated from the derivative of these spectra. The surface morphology of T-7 and T-15 via SEM images is shown in Figure 2.

Figure 2

Larger grain sizes are observed in the as-grown T-7 sample, with respect to those of T-15. The measured thickness of the T-7 was higher than that of T-15 (0.38 and 0.35 µm, respectively). Larger grain sizes may explain lower resistivity of T-7 than T-15, due to the fact that a lower amount of grain boundaries. The grain sizes do not change substantially with the post-thermal annealing in Ar atmosphere, confirming the results of optical transmission shown in Figure 1.

Table 1 shows the influence of the post thermal annealing in Ar atmosphere on the electrical, optical and morphological properties of the T-7 and T-15. Both, Haacke’s figure of merit and the modified one give maximum values for the thermal annealed samples, with higher value corresponding to T15, which means that this TCO is optimized with respect to T7 in its optical, electrical and morphological properties.

Table 1

Commercial conducting glass	Rsh (Ω/sq)	Tave (%)	R (nm)	T10/Rsh (Ω-1)×10-3	(T10/Rsh)·1/R (Ω-1nm-1)×10-5	EG (eV)
T-7-asgrown	7.8	64.6	20.2	1.6	8	3.8
T-7 annealed	8.6	64.0	19.8	1.3	7	3.8
T-15-asgrown	15.0	75.9	17.2	4.2	24	3.9
T-15 annealed	14.9	74.0	9.2	4.0	40	3.9

From the results shown in Table I it concludes that the T-15 is more appropriate to be used as TCO in the processing of solar cells than the T-7, and therefore this TCO was chosen as a substrate in the following processes.

A buffer layer must have the following properties: (a) highly resistive to eliminate as much as possible the interdiffusion, through it, between the atoms of the adjacent layers; (b) of the smallest possible thickness to reduce its incidence on the series resistance of the device; and (c) a high transmittance. Taking into account the above, we have defined the following figure of merit for buffer layers:

( R s h   T a v e 10 ) / d (3)

The maximum value of this figure of merit optimizes the three mentioned aspects. Figure 3 shows the transmittance spectra of the SnO₂ buffer layers after the thermal annealing. Table 2 summarizes the values of the sheet resistance, average transmittance, thickness, resistivity and the corresponding figure of merit of as-grown and annealed SnO₂ samples deposited by PSP and magnetron sputtering techniques on the glass substrates.

Figure 3

Table 2

SnO2 deposition	Rsh (Ω/sq)×103	Tave (%)	d (nm)	ρ (Ω.cm)	Rsh·Tave10/d (Ω/nm)×106
PSP as-grown	1000	75.8	17	1.76	3.7
PSP annealed	1100	76.3	17	1.76	4.3
Sputt. as-grown	31.7	67.6	35	0.09	0.02
Sputt. annealed	70.0	64.4	31	0.2	0.02

The sheet resistance of the films deposited by PSP is higher than those deposited by sputtering in all cases. These results can be explained as follows: In the PSP method the formation of the compound SnO₂ is the result of the reaction between the Sn from the salt with the oxygen of the surrounding air in the growth chamber (the carrier gas used in these deposits is air). The reaction between a salt containing Sn and oxygen using air as a carrier gas in the PSP technique is well known, such that the SnO₂ samples grown by this technique are characterized by growing in an oxygen rich environment. A different situation occurs in the sputtering depositions, where the properties of the SnO₂ thin films depend on the stoichiometry of the target, the conditions of the environment inside the chamber and the deposit parameters.

In our experimental conditions the deposition is made under Ar atmosphere in the chamber and therefore the oxygen content in the film depends on the stoichiometry of the target and the loss of this element in the deposition process. Reactive sputtering with a flow of oxygen is usually introduced in the chamber to control of oxygen content in the sample [20,21]. Oxygen vacancies give the n-type conductivity character to the SnO₂ compound, because the vacancies of oxygen cause that the tin yields its electrons to the crystalline structure. An increase in the number of oxygen vacancies determines an increase in conductivity. In general, thin films with a higher oxygen content should be obtained when the PSP technique is used, respect to the sputtering ones and therefore more resistive samples, as shown in table II. Following the same reasoning, the thermal annealing should not have significant effect on the electrical properties of films deposited by PSP, while in the case of the sample deposited by sputtering the same thermal annealing in oxygen atmosphere is more effective and therefore an increase in the shunt resistance and resistivity is obtained. Concluding, resistive SnO₂ films with thickness of 17 nm are obtained by the PSP technique and less resistive with thicknesses of the order of 30 nm by sputtering, which determines that the former better fulfill the properties of buffer layers than the latter, as shows the figure of merit. In the case of films deposited by PSP on a glass substrate, the post-thermal annealing does not change the properties of the films with respect to as-grown one.

Higher average transmittance is obtained in the case of the films deposited by PSP with respect to those deposited by sputtering, due to the thicknesses of the layers and the calculated values of band-gap were approximately the same (4.0 eV). The band-gap varies from 3.54 eV to 4.20 eV according to the values reporting in literature conditions [23,24].

From preliminaries XPS measurements (no shown in this work) the O/Sn ratio value was estimated for layer deposited by PSP and sputtering. The O/Sn ratio value closer to 2 (1.9) for the sample deposited by PSP and 1.65 by sputtering, reveals a more stoichiometry in the case of sample deposited by PSP in comparison with the sputtering ones. It is known that completely stoichiometric SnO₂ is an insulator or at most an ionic conductor. However, in practice this material is never stoichiometric and invariably presents defects, fundamentally oxygen vacancy.

In order to compare the properties of the TCO/SnO₂ bilayer with those described above in the case of commercial TCO, we have first calculated the equivalent resistance of the bilayers. From the electrical point of view, these bilayers form a parallel resistance system, in which the total resistance of the system is given by the following relationship:

1 R T = 1 R T C O + 1 R b u f f e r ⇒ R T = R T C O R b u f f e r R T C O + R b u f f e r (4)

Figure 4 shows the transmittance spectra of the TCO/SnO₂ structures, where the buffer layers have been deposited by PSP and sputtering techniques. In both cases the T15 and the buffer layers were subjected to the thermal treatments explained above.

Figure 4

Comparing Figures 1 and 4 it is observed that the deposit of the buffer layers, both by PSP and by sputtering on the TCO, does not alter its transmittance, as a consequence of the thicknesses of the buffer layers, significantly lower than the TCO ones. A slight decrease in the thickness of the layers under thermal treatment was observed, probably due to a greater compactness of the grains and a loss of material under this treatment.

The morphological properties obtained from the AFM technique are shown in Figure 5. With the thermal annealing, the film particle of the sample deposited by PSP increases while an opposite process occurs for the sample deposited by sputtering, and the surface roughness was similar for both structures. The post-thermal treatment has a greater effect in reducing the roughness of the structure in the case of the SnO₂ layer deposited by sputtering, respecting that deposited by PSP.

Figure 5

Finally, the calculated values of R_T according with the equation (4) correspond to the measured ones. Table 3 summarizes the values of the shunt resistance, average transmittance, thickness, RMS roughness and Haacke’s modified figure of merit of the TCO/buffer structure.

Table 3

T-15/SnO2 conditions	Rsh (Ω/sq)	Tave (%)	d(TCO+ SnO2) (nm)	R (nm)	(T10/Rsh)·1/R (Ω-1nm-1)×10-5
SnO2 by PSP	14.1	72.9	376	2.1	107
SnO2 by Sputtering	15.3	72.5	385	2.8	90

Figure 6 shows the top view SEM micrographs of the surfaces of the TCO/SnO₂ bi-layers both with post- thermal annealing. A better uniformity is observed in the layer deposited by sputtering on T15 with a better contrast, according with the higher conductivity. Therefore, a better morphological property is obtained when the SnO2 layer is deposited on the conducting glass substrate T15 by sputtering technique.

Figure 6