CC BY
4STABILITY OF PHOTOLUMINESCENCE DECAY KINETICS AND PERFORMANCE OF POROUS SILICON BASED PHOTOELECTRIC STRUCTURES*
B. O. Semerjyan, Z. N. Adamian, Kh. S. Martirosyan
Department of Physics of Semiconductors and Microelectronics, Yerevan State University Alex Manoogian st., 1, Yerevan, 375025, Armenia Tel.: (374 10) 527133, e-mail: [email protected]
The time-resolved visible slow photoluminescence of porous silicon (PS) obtained by electrochemical & method have been measured within the temperature range 90-400 K. A study has been made of the g interrelationship between time-resolved photoluminescence characteristics (intensity, emission, spectra, © relaxation times and their temperature dependence) and aging (structural, dielectric properties, changing characteristics during storage).
It was shown that formation of antireflective and passivating PS layer increase the short-circuit current and solar cell efficay by 15-25 %.
Optoelectronic characteristics of studied structures are stabilized after air exposition during three months.
Introduction
The perspectives for alternative sources of energy — such as solar cells, converting solar energy directly to electricity, — are increasing every year. Two main directions can be seemed in the solar cell technology. The first one — races for the world record in efficiency, no matter how much it can cost. The second one — search for the optimal technology, not very expensive but ensuring efficiency of the cells high enough for most applications. Our work belongs to the second part.
It is well known that porous Si (PS) made on the surface of crystalline Si wafers drastically reduces the optical losses by acting as an antireflec-tion coating layer (ARC) [1—3]. PS can be obtained in an aqueous hydrofluoric solution by two methods: electrochemical etching and chemical etching. The chemical etching is self-regulating so that the simultaneous control of the PS parameters (thickness and porosity) is difficult to achieve. The electrochemical method permits one to control them through the use of an additional variable, the current density, which is a key formation parameter determining the PS porosity.
Objective: The present research was focused on study of sample "aging" influence on time resolved spectra of photoluminescence (PL) of PS based photosensitive structures (solar cells and photodiodes) and performance parameters.
Influence of aging of photosensitive PS-based structures on spectral features (short-circuit current, SCC, photoconductivity) of solar cell (SC)
and photodiode (PH) photo response was studied. For more significant data on SC photosensitivity they were also irradiated by modulated light imitating solar radiation provided with appropriate current spectrum and density.
Experimental materials
Consequently, the PS layer has to be developed after producing the junction, which is within the n+ region, where a phosphorus doping exists. In Si solar technology, p-n+ junctions are made by diffusion of phosphorus from the surface, thus creating the n+ surface layer on Si-based p-type substrate. Since the fabrication of the junction is a high temperature (900 °C) process, a PS layer prepared before that treatment would be completely oxidized.
Shallow p-n+ junctions were produced by the manufacturer on <100> oriented p-type (0.3^1 Ohm-cm) monocrystalline Si. The thickness (0.3^1.5 |im) the doping profile of the shallow n+ emitter were optimized for the solar energy conversion. PS < was formed in the n+ emitter under galvanostatic i conditions using a two electrode setup. The electrolyte was 30 % HF. The Si wafers were cut into ^ 1x1 cm2 pieces was isolated for the electrochemical '1 PS formation. The Si working electrode and Pt | counter electrode were connected to a potenciostat. | The as — prepared samples were removed from the g sample holder, rinsed with deionized water and stored & in ambient atmosphere in the dark. g
The surface microrelief of porous silicon was ™ a structure consisting of open blind surfaces and
International Scientific Journal for Alternative Energy and Ecology ISJAEE № 12(32) (2005) Международный научный журнал «Альтернативная энергетика и экология» АЭЭ № 12(32) (2005)
* Доклад на Второй конференции по возобновляемой энергетике «Энергия будущего» (Ереван, июнь 2005 г.). The report on the Second Renewable Energy Conference "Energy for Future" (Yerevan, 2005)
Статья поступила в редакцию 17.10.2005. The article has entered in publishing office 17.10.2005.
could be described by the sponge state model resulting from the deduction of the etched mass of silicon. That is the model of solid state containing pores in the form of branched net of as-if-drilled channels (Fig. 1).
® Fig. 1. Porous silicon surface microrelief
The shape of the macropores may be simplified and modeled as a surface of hemisphere, cylinder or conical surface of rotation. It is determined by the technology of their formation. The average diameter of the observed macropores on the specimens' surface was about 2 |im. The porosity of the silicon in the near-surface layer took on values from 0.3 to 0.6.
For measurement and registration of photoelectric characteristics of PS-based structures, solar cells (SCs) with and without antireflective coat and SCs with PS coat, experimental equipment for measuring and registration of spectral dependency of stationary photoconductivity of PS-based semiconductor sample was constructed.
The studied samples were placed in special glass or cryostats. The light of exiting radiation that is quartz halogen (or hydrogen), was directed by means of mirror condenser onto an input slit of the monochromator SPM-2. The radiation (0.35 |im) outgoing from monochromator slit was focused via quartz lens, and aluminum mirror was directed to examine sample on a copper disk of cryostat or sample holder. Diameter of the disk suspended on shock-absorber of mechanical modulator arresting the radiation from light source was chosen as large as possible.
Work slits of the monochromator were changed ¡5 from 1 mm (spectral range 0.2-0.5 |im) to 0.25 if mm (wavelength over 0.9 |im) which provides constant spectral slit width in 10-15 nm range.
5 When measuring SC spectral characteristics, 'il the energy flow on monochromator output is sev-| eral orders lower compared to even single solar I flow density. When measuring the spectral sensi-É tivity, there is certain dependence of diffusion
T
6 length of non-essential charge carrier — LB at g basal layer on the exposition level. Hence, when ™ measuring the SC spectral sensitivity, expositions
were obtained as closer to real conditions of SC performance. In regard to this, for obtaining more significant data on SCs' sensitivity they were simultaneously irradiated both by modulated flow of monochromator radiation and non-modulated flow imitating the solar radiation at given spectrum and flow density.
The experimental equipment for measurement and registration of time-resolved spectrum (kinetics) of PS glow based on pulse molecular nitrogen laser on electron-vibrational transitions (t = 8 ns, X = 337 nm) was constructed.
The peculiarity of the optical part is the special sample holder with a mirror, which enables direction of the light of exiting source through light filters by means of cylindrical quartz lens onto the sample, also luminescence focused on slit of grating monochromator SFD-2. As a light detector, we designed and fabricated a pulse photodetecting apparatus (PDA) with photoelectric multipliers (PM) with a known spectrum of photoresponse enclosed in its individual, rigid housing with windows for the light direction onto PM photocatode. The light from the monochromator (SFD-2) was collected through lens system and is directed to the end face of fiber-optic braids provided that a light conduit connected with PDA windows.
Fig. 2 shows the total reflectivity, R, as a function of wavelength for free samples with Q = 0.4, 0.9 and 1.4 C/cm2. A decrease of reflectivity below 10 % is achieved for Q = 0.95 C/cm2, between 400 and 950 nm. Further increase of Q does not cause any significant change of the reflectivity. The reflectivity spectra display a modulation due to multiple reflections in the PS layer, which confirms that the "n" (refraction coefficient) of the PS layer is different from that of the bulk Si.
X, nm
Fig. 2. Total reflectivity of bulk Si (Ref) and three PS layers formed with different charges (Q = 0.3; 0.9 and 1.4 C/cm2) on 1.2 ^m deep p-n+ junctions.
The electrochemical studies of PS formation on deep emitters of p-n+ junctions have shown that the PS layer thickness is approximately proportional to the charge Q passed during its formation, and that the Si/PS interface reaches the junction for Q in the order of 1.1 C/cm2 when PS is made by electrochemical etching.
We have carried out experiments on a set of five samples prepared and at 5 mA/cm2 constant current and passing the following Q: 0.3; 0.9; 1.1 and 1.4 C/cm2.
The reflection spectra of solar cells with macroporous antireflection coating (ARC) were also investigated It is revealed that the reflection spectra of such a solar cells have a minimum in significantly more wide wavelength range than in case of using of nanoporous ARC.
International Scientific Journal for Alternative Energy and Ecology ISJAEE №12(32) (2005)
Международный научный журнал «Альтернативная энергетика и экология» АЭЭ № 12(32) (2005) ' '
Солнечная энергетика
1000 900 800 700 ^ 600 | 500 ^ 400 300 200 100 0
0.6 0.8 X, |m
Fig. 3. Absolute spectral sensitivities of typical Si SC: 1 — Terminal theoretic values of S(X) at 0-surface recombination, with unlimited carrier's life time and diffusion length of the auxiliary charge carriers and 0 value of reflection coefficient; 2, 3 — ARC (PS) SC and non ARC SC with shallow p-n junction (thickness of doped n-layer, d > 0.4 |m); 4 — ARC (PS) SC with deep p-n junction (thickness of doped n-layer, d < 1.2 |m).
The I-V characteristics under white illumination of junctions with (PS) and without PS are shown in Fig. 4, as the plot of the short-circuit current ratio (lscPS/lscREF) and open-circuit voltage ratio (VocPS/VocREF) versus the charge Q.
0
200
400
800
1000
1.5
1 1
1
0.5 0
\ ^scPS ^scRef
к---V /V ' ocPS ' ' ocRef 1 1 1 ' 1
0.5 •£
£
0.3
0.5
0.7
1.3
1.5
1.7
0.9 1.1
Q, C/cm2
Fig. 4. Variation of the ratios of short-circuit current
(/SCPS//SCRef) and °pen circmt voltage (VOCPS/VocRef) of devices
without PS (Ref) and with PS formed with several indices of Q (1.2 mm deep p-n+ junctions)
The general results is an improvement of electronic properties of the cells with increased thickness of the PS layer up to Q = 0.8 C/cm2. For the Q >1.1 C/cm2, I-V characteristics strongly decrease. These results are consistent with the description of the PS formation on the emitter presented above in the electrochemical section, showing that the n+ emitter becomes almost completely porous at Q ~ 1.1 C/cm2, a value when the degradation of the p-n+ junction properties occurs.
The fill factor (not shown here) is slightly affected by the PS formation up to 1.1 C/cm2, but drastically reduced for Q = 1.4 C/cm2.
The main parameter characterizing luminescence kinetics is the radiation relaxation timei. Spectral dependence of characteristic time relaxation of PL glow was detected (Fig. 5). Influence of sample "aging" on time-resolved glow spectra was registered. PL glow was registered within 0.21.2 |im wave length.
1200
600 X, nm
Fig. 5. Spectral dependence of pulse PL intensity pick on PS layer at different air expositions (porosity 50 %, thickness 0.2 ^m): ♦ — after 4 hours; ■ — after 3 weeks; ▲ — after 12 weeks
The results obtained showed shifting of pick of PS PL to the short-wave range ("blue shifting") parallel to increase of its intensity in room temperature conditions.
The main parameter of luminescence kinetics is the relaxation time of radiation intensity dropped e folds — t . The spectral dependence of typical relaxation time of PL (PL life time) was studied (Fig. 6). Decrease of decay time of PS PL evidences inhomogeneity of widening of PL spectrum. Parallel to increase of photon quant energy of PS, PL spectrum there was recorded a decrease of relaxation time of the radiation. Dependence of PL decay kinetics on simulating radiation intensity was also studied. As a result, in addition to intensity enhancement, the PL spectrum was narrowed and the time relaxation was decreased.
1000
100
10
1.2 1.3 1.4 1.5 1
1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7
E, eV
Fig. 6. Dependence of PL decay lifetime constant on PS-induced light energy in different air expositions: ♦ — in 4 hours; ■ — in 3 weeks; ▲ — in 12 weeks after PS layer formation
PS layer formation increases the SCC and SC efficiency by 15-25 %. Efficiency of CSs covered with antireflective and passivating PS (AM-1.5) is about 11 %.
The role of the surface passivation by PS-layer was further evidenced by UV nitrogen laser pulses. A significant increase of the characteristic decay time is observable after anodization when the sample was excited by UV laser radiation, which is absorbed mainly at the surface region. The signal amplitude decreases e-folds in about 20 ns and 90 ns for non-etched and etched surfaces, respectively. It is concluded that after PS coating, the
International Scientific Journal for Alternative Energy and Ecology ISJAEE № 12(32) (2005) Международный научный журнал «Альтернативная энергетика и экология» АЭЭ №12(32) (2005)
2
1
carrier recombination at the surface becomes an order slower.
The photovoltaic parameters of these solar cells as well as their long-term variation are studied too.
Conclusion
I
In different experiments the thickness and the cu refractive index of PS layer have been detected. S The values of these parameters for the coat-
's ings were found to be 90-120 nm and 1.7—2.1, -5 respectively. These values are very close to the y optimal antireflection coating for SC. The main g results are followings:
1. Formation of PS layer thinner than 0.2 |im § always improves SC parameters determined from ® I-V measurements.
2. Refractive index of PS is always less than the refractive index of bulk silicon, its value increases along with increase of the HF content in the electrolyte.
3 . PS layer thickness is nearly proportional to the electrical charge transmitted to the electrolyte during the etching.
4. We have demonstrated two different applications of PS layers formed on already manufactured SC. Short anodical treatment provides anti-reflective coatings and, additionally, passivates the surface of the cells. Microporous layers, which are formed on Si surfaces after relatively long anodi-cal etching, cause surface roughening that enhances the optical confinement in the cells and increasing the efficacy.
Optoelectronic characteristics of studied structures are stabilized after air exposition during 3 months.
A study has been made of the interrelationship between time-resolved photoluminescence characteristics (intensity, emission, spectra, relaxation times, and their temperature dependence) and aging (structural, dielectric properties, changing characteristics during storage). The time-resolved visible slow photoluminescence of PS obtained by electrochemical method have been measured within the temperature range 90-400 K.
Time resolved spectra of PS PL were excited by pulse molecular nitrogenous laser with electron-vibrating transitions (tp = 8 ns; X = 337 nm) PL was registered in 0.2-1.2 |im wave length range.
It was shown that formation of antireflective and passivating PS layer increased the short-circuit current and SC efficacy by 15-25%. The efficacy of the latter structure measured basis of oxidized and non-oxidized (subjected to air exposed "aging") PS in photodiode regimen. Quant efficacy was registered as 0.7 (wave length 0.4-0.7 |m). As a result of PS "aging" the spectral characteristic of photodiode shifted towards shortwave zone up to 0.4 | m.
Reference
1. Canham L. T. // Appl. Phys. Lett. 1990. Vol. 57. P. 1046.
2. Aroutiounian V. M. // Reports NAS RA. 1995. Vol.95, No. 4. P. 229.
3. Aleksanyan A. G., Yeremyan A. S., Aroutiounian V. M. // J. Contemp. Phys. 2000. Vol. 35, No. 4. P. 187.
International Scientific Journal for Alternative Energy and Ecology ISJAEE №12(32) (2005) 7Q
Международный научный журнал «Альтернативная энергетика и экология» АЭЭ №12(32) (2005)
