Detection and Localization of Defects in Monocrystalline Silicon Solar Cell

Near-surface defects in solar cell wafer have undesirable influence upon device properties, as its efficiency and lifetime. When reverse-bias voltage is applied to the wafer, a magnitude of electric signals from defects can be measured electronically, but the localization of defects is difficult using classical optical far-field methods. Therefore, the paper introduces a novel combination of electric and optical methods showing promise of being useful in detection and localization of defects with resolution of 250 nm using near-field nondestructive characterization techniques. The results of mapped topography, local surface reflection, and local light to electric energy conversion measurement in areas with small defects strongly support the development and further evaluation of the technique.


Introduction
Although the concept of photovoltaic (PV) devices descends from the mid-19th century, its modern age began after 1950 [1].Solar cells fulfill two principal functions: photogeneration of charge carriers-electrons and holes-in a light-absorbing material, and separation of the charge carriers to a conductive contact that transmits the electric current [2].Their efficiency is limited by a number of factors, which include fundamental power losses (incomplete absorption of light or dissipation of a part of the photon energy as heat); losses caused by the reflection of light from the cell surface; and finally, a recombination of the electron-hole pairs in the substrate.
The basic methods for the characterization of silicon solar cells are generally electrical measurements [3][4][5].Electrical methods represent an integral measurement on the whole cell.Unfortunately, they do not enable to localize defects occurring in the structure.Local defects in the pn junction may be associated with structural imperfections (such as grain boundaries, dislocations, and scratches), impurities, higher concentrations of donors and acceptors, or both [6].Therefore, it is important not only to find most harmful defects, but also to understand their nature and identify the factors which affect adversely their formation and recombination properties.The used PV analytical tools are generally divided into two groups.
(i) Mapping techniques which allow the access to the areas of interest (usually the areas with high probability of defects).For materials analysis, lifetimemapping tools such as surface photo voltage (SPV) [7], microwave photo conductance decay (MW PCD) [8] are frequently applied.Other methods map spatial distribution of photocurrent induced by laser beam (LBIC) [9], or by electron beam-induced current (EB-IC) [10] over whole wafer.
(ii) Nonmapping techniques that are applied to provide better insight into the nature of recombination centers [11,12].
For this reason, and for more precise mapping, the use of local characterization methods seems to be very important.Owing to the diffraction, there are Rayleigh limitations of the resolution in traditional optical microscopy.A higher resolution can be achieved with confocal microscopy [13] or Scanning Near-field Microscopy (SNOM) [14].
The combination of SNOM with LBIC allows creating a strong characterization method of Near-field Optical Beam-Induced Photocurrent (NOB-IC) [15].This method provides a measurement of the current locally induced by optical near-field.The combination of high resolution of the microscope with locally induced light by sharpened optical fiber allows obtaining a resolution bellow the wavelength of used light.
Due to the fact that solar cells are optoelectronic devices based on photoelectric effect, it is natural and desirable testing them by using optical and optoelectronic methods.Nevertheless, almost all scientific groups studied one kind of these characteristics only-electrical or optical ones.Our previous effort has been focused on the investigation of solar cells [16,17], because their local properties are not well described yet.To elucidate slightly more this problem, elaborated combination of electric and localized optical measurement, which allows the detection and localization of defects in the solar cell wafer, and to compare experimental results and obtain higher resolution, is presented.

Experimental Methods and Material
The sample of monocrystalline silicon solar cell wafer with area of 10 cm × 10 cm has been tested.Most important part of the solar cell is its p-n junction.When reverse-bias voltage is applied, lower voltage breakdown of p-n junction occurs in defect sites.reversing the current shift in the homogeneous breakdown may be primarily formed by the current flux in local defects.The emission from defect could be considered as noise current.In areas of increased concentration of free charge carriers due to the small cross sections, there is a large current density which can lead to strong local heating and then to the local diffusion and heat breakdown.

Photoelectric Measurement.
To set a suitable reverse voltage, which leads to emission of radiation from defects, computer-controlled voltage source (VS), filter capacitor (C), and parallel load resistance (R L ) were used (Figure 1).The circuit was connected to the reverse state monocrystalline solar cell wafer (SC) with area of 10 cm × 10 cm.The electrical signal was detected at the load resistor R L = 5.17 Ω.
The obtained signal was amplified by preamplifier (PA) and amplifier (A).Plastic optical fiber (OF) with aperture of 200 μm has scanned over solar cell and collected a weak optical signal emitted from this place and sent it to the photomultiplier (P) and amplifier (A) (Figure 1).for the electric one.The voltage was measured by digital voltmeter (DV) and values were stored on PC.With this setup, the effective values of noise current versus reverse-bias voltage of electric and optical signals were measured.The bias voltage was continuously set from 0 to 25 V. Figure 2 represents the relation between defect noise current and reverse-bias voltage over whole solar cell wafer.By repeating this measurement, the noise signals appeared for the same values of bias voltages.
The corresponding optical application in upper arm provides localization of defects or imperfections.When the reverse-bias voltage was applied, any noise signal from defect site has not been observed up to first important current peak at U r = 6.8 V (peak 1 in Figure 2).With further increasing of the voltage, other near-surface defects appeared in different sites on the sample.When bias voltage reached a value of U r = 18.5 V, several very intensive spots, originated mainly in ill-cutting edges of solar cell, defects in p-n junction, or imperfections of structure, have been clearly localized (Figure 3), and a corresponding current signal was quite strong (peak 5 in Figure 2).For other values, the location of these sites could vary in function of defects nature.Above 23 V, the electric noise signals inside silicon wafer dominated over defect signals and interpretation of results was no more meaningful.

Near-Field Measurements.
In the second method (nearfield NOBIC experiment) [18,19], a very small area of silicon solar cell surface (approx.150 nm in diameter) has been excited by green laser diode (λ = 532 nm) light transmitted through a nanometer-sized (70 nm) aperture in the Agcoated sharpened single mode fiber probe (Figure 4).
The excitation light was amplitude modulated by the light chopper at frequency of 300 Hz.The input power coupled into the optical fiber probe was 3 mW, and output power from the fiber probe varied between 10 nW and 100 nW, when detected by remote detector.Consequently, the detected photo-induced current varied in the range 100-300 PA.The localized photo-induced current across the layer of solar cell was then detected as a function of the tip position above the sample surface mounted on an x-y-z piezo and was scanned (the scanning step of 50 nm) related to the probe tip.During the scan the tip-sample distance is kept constant at (5 ± 1) nm using an optical shear force feedback control.Thanks to this setup, the xy current distribution map of solar cell has been obtained.The photo-induced current signal has been detected by a lock-in nanovoltmeter while the solar cell was reverse-biased or unbiased [20].
The accuracy of this method depends primarily on the light spot size and on the scanning step of the piezo driver, which are inversely proportional.By long step the accuracy of the method is low, but whole measurement process accelerates because of reduced number of measured points.Therefore, it was very important to choose an optimal ratio scanning step/spot size.
The topography of the sample with a pyramidal texture is shown in Figure 5(a).The electrical response signals, corresponding reflectance, and topography are demonstrated by dependence on spatial coordinate in Figure 5(b).Black curve corresponds to the profile of the sample surface.Blue one corresponds to the electrical response signal and purple one represents a local surface reflectance (in one scanning line).
Relative electrical response mapped by color scale onto original topography of the sample is shown in Figure 6.The pyramidal structure form Figure 5 has been etched so to decrease the electrical response on the tops of texture, which allow obtaining higher electric efficiency of the cell.This new mesa structure of the samples is presently the object of intensive study.

Conclusions
The novel combination of methods for samples local electric detection and optical localization with micro-and nanoscale resolution for the study of monocrystalline silicon solar cell wafer is presented.applying the reverse-bias voltage, several intensity spots, originated mainly in ill-cutting edges of solar cell, defects in p-n junction, or imperfections of structure, have been clearly localized (Figure 3), and noise current signal peak was quite strong (peak 5 in Figure 2).Above 23 V, the electric noise signals inside silicon wafer dominated over defect signals and interpretation of results was no more meaningful.A combination of NOBIC and reflectivity measurement with the resolution of about 250 nm has also been established.After calibration of the setup, the accuracy of the combined reflection measurements is better than 5%.At short circuit condition, the NOBIC photocurrent of this cell dominates over the variation of the reflectivity.Based on correlations with aperture-SNOM, the sites corresponding to largest and smallest reflections have been assigned.The photocurrent is the smallest on top of protruding peaks which have a greater local reflectivity, and is the largest in the valleys with the smallest reflectivity.Using this correlation we have found, that the photocurrent for applying a forward voltage decreases inhomogeneously at different locations (Figures 5 and 6).The measurement has shown smaller relative fall of the photocurrent for the illumination of valleys in comparison with the peaks in the structure.
Proposed characterization method based on scanning probe microscopy technique SNOM allows nondestructive and noncontact sample study (defects in p-n junction, structure imperfections, and local photoelectric measurements).A maximum of optically excited photocurrent is indicator of local conversion efficiency due to local light constant energy excitation, and number of imperfections is a quality indicator for solar cell lifetime and only precise testing can help to determine a nature of defects.At present time, it is quite difficult find a correlation between defect nature and its appearance.Therefore, for further improvement of monocrystalline silicon solar cells efficiency, more intensive mapping and nonmapping measurements of optical and electric properties are challenged.

Figure 1 :Figure 2 :
Figure 1: Scheme of experimental setup for the measurement of effective values of noise current versus reverse-bias voltage of electric (lower arm) and optical (upper arm) signals.

Figure 3 :
Figure 3: Localization of defects imperfection areas of solar cell wafer using reverse-bias light emission from different wafer sites (U r = 18.5 V).

Figure 4 :
Figure 4: SNOM experimental setup for the measurement of effective values of electric response, optical properties-reflectance, and topography of solar cell in the near-field.

Figure 5 :
Figure 5: Local topography of solar cell sample with pyramidal structure (a), and corresponding scan of electrical response, reflectivity and topography (b).