Device Structure

Figure 4.4(a) shows the conventional solar cell structure for a reference in this study,

which was grown on a c-plane sapphire substrate, followed by a 3-μm-thick undoped GaN buffer layer and a typical p-i-n structure, consisting of a 100-nm-thick n-type GaN, a 100-nm-thick intrinsic In0.3Ga0.7N absorption layer, and a 100-nm-thick p-type GaN.

Figure 4.4(b) plots its built-in ( E

_in) and polarization-induced (

E

_P) electric fields within the In0.3Ga0.7N absorption layer versus doping concentrations of p- and n-GaN layers (NA

=N

D). The net electric field,

E

_in+

E

_P, is also plotted in the figure. Accordingly,

E

_in and

E

_P are in the opposite direction (

E

_inpoints from n- to p-type GaN layers

Fig. 4. 4 (a) Conventional solar cell structure used as a reference device, and (b) its built-in (E_in^{) and} polarization-induced ( E_P) electric fields within the In0.3Ga0.7N absorption layer versus doping concentrations of p- and n-GaN layers (NA =ND). (c) The proposed structure of polarization-induced doping n-i-p solar cell, and (d) the corresponding polarization-induced doping density and electric field along the growth direction of the device.

along the [0001] direction, while

E

_P of Ga-face polarity wurtzite crystal points along

the [ 0001 ] direction), and E_P is generally larger than E_in for the doping concentrations of NA, D

≤ 1×10

¹⁹ cm^-3. Therefore, the net electric field will move photogenerated carriers created by light absorption of intrinsic In0.3Ga0.7N layer in a detrimental way (i.e., electrons to p-GaN, while holes to n-GaN) as plotted in

Fig.

4.4(a), making it impossible for carrier collection and diminishing the photocurrent of

the device. By further increasing the doping concentrations to NA,D

>1×10

¹⁹ cm^-3

, E

_P decreases rapidly as the piezoelectric polarization sheet charges are screened by highly doped GaN regions on either side of the In0.3Ga0.7N absorption layer. After that,

E

_instarts dominating the net eclectic filed of the device, which renovates the potential profile

of

the In0.3Ga0.7N absorption layer for efficient carrier collections. However,

such highly impurity-based doping concentrations (>1×10

¹⁹ cm^-3

) are difficult to achieve especially in p-type GaN due to its inefficient thermal activation of dopant acceptors.

For the improved structure (Fig. 4.4(c)), the original 100-nm-thick p- and n-types GaN layers are replaced by the 100-nm-thick graded InxGa1-xN layers, where the indium compositions are linearly grading from x = 0% to x= 30% and back to 0% to construct the p- and n-type regions of the solar cell. As plotted in

Fig. 4.4(c), the

negative polarization charge field created by grading InxGa1-xN layer (GaNIn0.3Ga0.7N) attracts holes to realize p-type doping in the bottom of the device.

Meanwhile, as a result of attracting holes, negative charges (electrons) are hence supplied to the positive polarization charge filed induced by reversely grading InxGa1-xN layer (In0.3Ga0.7NGaN), resulting in the Fermi level close to the conduction band and defining an n-type conducting region in the top of the device.

While the light absorption layer of the improved device is kept the same as that of the conventional III-nitride p-i-n solar cell. Most importantly, in the absence of any impurity doping, such device’s built-in electric field across the In0.3Ga0.7N absorption layer actually is equivalent of its polarization-induced electric field (

E

_in=

E

_P), beneficial for the electric drifting and efficient collection of photogenerated carriers.

Figure 4.4(d) shows the polarization-induced doping density and electric filed along the

growth direction of the improved device. Accordingly, the distributed profiles of field-ionized carriers (red dash-line for holes; black dash-line for electrons) are symmetrical, and a high carrier density of p = n = 3×10¹⁸ cm^-3 is achievable by using

polarization-doping scheme. As the polarization charge is uniformly built into the unit cell, the polarization-doping profiles are precisely controlled and well defined within the graded InxGa1-xN regions, and that is difficult to achieve by random impurity doping due to the segregation or diffusion tendencies of acceptors and donors [4.13].

The magnitude of

E

_in(=

E

_P) within the In0.3Ga0.7N absorption layer is around 2×10⁷ V/m. As plotted in Fig. 4.4(b), the conventional III-nitride p-i-n device would achieve a similar value of electric field, however, the doping concentrations must be as high as

N

A,D ~10²⁰ cm^-3 to efficiently screen the

E

_P, and again, that is unrealistic for the MOCVD growth.

4.3 Results and Discussion

Fig. 4. 5 Calculated strain profile in the x (= y) and z direction of (a) impurity doping p-i-n and (b) polarization-induced doping n-i-p solar cells.

Figure 4.5 depicts the strain profiles in the x (equal to y) and z directions for (a) the

impurity doping p-GaN/i-In0.3Ga0.7N/n-GaN and (b) polarization-induced doping n-i-p solar cells. The GaN layer is assumed to be relaxed, and the InxGa1-xN layers are pseudomorphically grown on it. The calculation detail of strained nitride layers is

described in Ref. [4.15]. The strain values are used to compute the piezoelectric polarization of the devices. As shown in Fig. 4.5(a), abrupt strain differences along both basal and z directions are observed at the GaN/In0.3Ga0.7N hetero-interfaces, where the polarization-induced surface charges start accumulating. Consequently, the energy band of In0.3Ga0.7N absorption layer is severely tilted, addressing the band discontinuities at GaN/In0.3Ga0.7N hetero-interfaces. In contrast, the strain difference between the In0.3Ga0.7N absorption layer and its contiguous layers is alleviated profoundly in the polarization-induced doping n-i-p device (Fig. 4.5(b)), and so do the accumulation of polarization-induced surface charges and the formation of piezoelectric field. The field-ionized carriers induced by grading InxGa1-xN layers therefore determines the electric filed (directing from n-type to p-type) across the In0.3Ga0.7N absorption layer, and that tilts the energy band in a favorite way for carrier collection, and simultaneously eliminates the band discontinuities at hetero-interfaces.

We now examine the physical effect of polarization-induced doping on the enhancement of carrier collection. Figure 4.6(a) shows calculated energy band diagrams of the impurity doping p-i-n (p = 5×10¹⁷ cm^-3, and n = 5×10¹⁸ cm^-3) and polarization-induced doping n-i-p solar cells under 1-sun illumination. A sharp conduction and valence band offsets can be seen at hetero-interfaces for the impurity doping p-i-n structure, hindering the extraction of photogenerated carriers created in the In0.3Ga0.7N absorption layer out of the device. Additionally, the band tilting of In0.3Ga0.7N absorption layer is still detrimental for carrier drifting even under the illumination of solar light. As a result, the current generated in such an impurity doping p-i-n structure mainly attributes to the absorption in the GaN layers. As the improved device using polarization-doping scheme, a smooth potential profile without introducing discontinuities in both conduction and valence bands is observed, entirely

Fig. 4. 6 Calculated (a) energy band diagrams, and (b) distributions of photogenerated carriers (red line for holes, and black line for electrons) for both impurity doping p-i-n (p = 5×10¹⁷ cm^-3, and n = 5×10¹⁸ cm^-3) and polarization-induced doping n-i-p solar cells under 1-sun illumination.

due to the essential contribution of graded InxGa1-xN regions. As previously discussed with reference to

Fig. 4.5(b), the adverse impact of piezoelectric field on the band

tilting of In0.3Ga0.7N absorption layer is also alleviated, and the extraction of photogenerated carriers created by In0.3Ga0.7N absorption layer outside the device hence becomes feasible. Therefore, as plotted in Fig. 4.6(b), most of photogenerated carriers in impurity doping p-i-n solar cell are accumulated and eventually vanished at hetero-interfaces, while our polarization-doping n-i-p structure contributes stable output of photocurrents.

Fig. 4. 7 Temperature-dependent (50—400K) J-V curves of (a) impurity doping p-i-n and (b) polarization-induced doping n-i-p solar cells. The power conversion efficiency (η) versus reciprocal temperature (1/T) was also inserted in the figures.

Figure 4.7 shows the temperature-dependent (50 — 400K) J-V curves for (a)

impurity doping p-i-n and (b) polarization-induced doping n-i-p solar cells. The power conversion efficiency (η) versus reciprocal temperature (1/T) was also inserted in the figures. The short-circuit current (JSC) of the impurity doping p-i-n solar cell decreases when temperature is lowered from 400 to 50K (Fig. 4.7(a)), which is explained by the freezeout of thermally activated impurity dopants. While the device’s open-circuit voltage (VOC) actually increases with the decreasing temperature, it is mainly attributed to the enlarged size of energy band-gap of In0.3Ga0.7N absorption layer. Consequently, an extremely low power conversion efficiency of η ~ 0.1% is generally obtained in impurity doping p-i-n solar cells under all temperatures. In comparison, the JSC values of polarization-induced doping n-i-p solar cell are essentially independent of temperature, which is an indicative feature of field ionization as polarization charges are originally atomic and do not need thermal energy to be activated. With temperature variations, the calculated JSC values are near identical to JSC ~ 4.3 mA/cm², more than one order of magnitude enhancement as compared to that of impurity doping p-i-n solar

cells. Again, the increasing of device’s VOC with decreasing temperature is primarily due to the enlarged energy band-gap of light absorption layer. Hence, a highest power conversion efficiency of η ~ 8.5% can be achieved on our polarization-induced doping n-i-p solar cell at the temperature of T = 50K.

4.4 Conclusions

In summary, we propose a new type of III-nitride n-i-p solar cell not formed by impurity doping. The approach of polarization-induced doping is particularly useful for III-nitride solar cells with high indium contents in the InxGa1-xN absorption layer as the band discontinuity at hetero-interfaces is considerably alleviated. Most importantly, as polarization charges are atomic in origin and do not need thermal energy to be activated, the short-circuit current of the proposed polarization-induced doping n-i-p solar cell are essentially independent of temperature and contribute to a high power conversion efficiency, and that offers a great potential for the next generation of photovoltaic cells.

References

[4.1] E. Fred Schubert, Light-Emitting Diodes (Cambridge University Press; 2 edition, 2006).

[4.2] F. Bernardini and V. Fiorentini, “Nonlinear macroscopic polarization in III-V nitride alloys,” Phys. Rev. B, 64(8), 085207 (2001).

[4.3] V. Fiorentini, F. Bernardini, and O. Ambacher, “Evidence for nonlinear macroscopic polarization in III-V nitride alloy heterostructures,” Appl. Phys. Lett., 80(7), 1204–1206 (2002).

[4.4] A. Nakajima, Y. Sumida, M.H. Dhyani, H. Kawai, and E.M.S. Narayanan,

“GaN-based super heterojunction field effect transistors using the polarization junction concept,” IEEE Electron Device Lett., 32(4), 542-544 (2011).

[4.5] S. Helkman, S. Keller, Y. Wu, J. S. Speck, S. P. DenBaars, and U. K. Mishra,

“Polarization effects in AlGaN/GaN and GaN/AlGaN/GaN heterostructures,” J.

Appl. Phys., 93(12), 10114 (2003).

[4.6] J. P. Ibbetson, P. T. Fini, K. D. Ness, S. P. DenBaars, J. S. Speck, and U. K.

Mishrab, “Polarization effects, surface states, and the source of electrons in AlGaN/GaN heterostructure field effect transistors,” Appl. Phys. Lett., 77(2), 250 (2000).

[4.7] I. Ho and G. B. Stringfellow, “Solid phase immiscibility in GaInN,” Appl. Phys.

Lett., 69(18), 2701 (1996).

[4.8] S. Y. Bae, Y. H. Song, S. R. Jeon, D. M. Kim, Y. D. Jho, and D. S. Lee,

“Pseudomorphic thick InGaN growth with a grading interlayer by metal organic chemical vapor deposition for InGaN/GaN p–i–n solar cells,” J CRYST GROWTH, 387, 23-28 (2014).

[4.9] G. F. Brown, J. W. Ager, W. Walukiewicz, J. Wu, “Finite element simulations of compositionally graded InGaN solar cells,” SOL ENERG MAT SOL C., 94(3), 478-483 (2010).

[4.10] W. Walukiewicz, J. W. Ager, K. M. Yu, Z. Liliental-Weber, J. Wu, S. X. Li, R.

E. Jones, and J. D. Denlinger, “Structure and electronic properties of InN and In-rich group III-nitride alloys,” J. Phys. D, 39(5), R83 (2006).

[4.11] J. Simon, V. Protasenko, C. Lian, H. Xing, and D. Jena, “Polarization-induced hole doping in wide-band-gap uniaxial semiconductor heterostructures,” Science, 327(5961), 60-64 (2010).

[4.12] D. Jena, S. Heikman, D. Green, D. Buttari, R. Coffie, H. Xing, S. Keller, S.

DenBaars, J. S. Speck, U. K. Mishra, and I. Smorchkova, “Realization of wide electron slabs by polarization bulk doping in graded III–V nitride semiconductor alloys,” Appl. Phys. Lett., 81(23), 4395 (2002).

[4.13] O. Mayrock, H. J. Wunsche, and F. Henneberger, “Polarization charge screening and indium surface segregation in (In,Ga)N/GaN single and multiple quantum wells,” Phys. Rev. B, 62, 16870 (2000).

[4.14] B. Y. Ber, Y. A. Kudriavtsev, A. V. Merkulov, S. V. Novikov, D. E. Lacklison, J. W. Orton, T. S. Cheng, and C. T. Foxon, “Secondary ion mass spectroscopy investigations of magnesium and carbon doped gallium nitride films grown by molecular beam epitaxy,” Semicond. Sci. Technol., 13(1), 71-74 (1998).

[4.15] A. E. Romanov, T. J. Baker, S. Nakamura, J. S. Speck, and ERATO/JST UCSB Group, “Strain-induced polarization in wurtzite III-nitride semipolar layers,”

J. Appl. Phys., 100(2), 023622 (2006).

Chapter 5 Current Matching Using CdSe Quantum

在文檔中 新穎高效率太陽能電池之研究 (頁 108-119)