PHOTOVOLTAICS: Photoluminescence imaging speeds solar cell inspection


Laser-based photoluminescence imaging is a technique for rapid characterization of silicon samples by revealing a host of material and device properties. With acquisition times of typically a few seconds or even fractions of a second, megapixel luminescence images can be captured of silicon bricks, unprocessed and partially processed wafers, and fully processed cells. The quantitative luminescence imaging applications that have been developed over the last few years open up intriguing possibilities for inline process monitoring during silicon solar cell production.

Photovoltaics (PV) technology has grown into a substantial industry with annual revenues in 2009 estimated to be more than $40 billion, and it is expected to continue growing exponentially over years to come. The majority of PV manufacturing capacity is in wafer-based crystalline silicon solar cells; in 2008 the PV industry consumed 68% of silicon feedstock and has thus become the main consumer of silicon. So-called grid parity (the point at which an alternative means of generating electricity is equal in cost, or cheaper than grid power) is within reach in many countries but requires further improvements in manufacturing efficiency and yield to achieve necessary cost reduction.

Achieving high production yield at lower cost requires advanced process monitoring tools that allow inspection with high lateral spatial resolution of specific electronic and device parameters at various parts of the PV value chain at line speed on every sample. Sample throughput in PV manufacturing is up to 3600 wafers per hour. Existing measurement techniques, which in many cases were developed for much smaller sample throughput in microelectronics applications, are unable to fulfill such requirements—they are either too slow or cannot provide sufficient spatial information or both.

Click to Enlarge
FIGURE 1. In a photoluminescence imaging setup, the output from a high-power fiber-coupled infrared (IR) laser is expanded to homogeneously illuminate a silicon brick, wafer, or solar cell. While the sample is illuminated (red arrows), a sensitive IR camera takes a snapshot of the luminescence signal (blue arrows) emitted by the sample.

Photoluminescence (PL) imaging of silicon bricks, wafers, and solar cells has been developed over recent years. Compared to measurement techniques such as microwave photoconductance decay (μ-PCD) mapping, PL imaging offers vastly superior speed and spatial resolution, making it perfectly suited for process monitoring in production. Developed by researchers at the University of New South Wales (UNSW) in Sydney, the patented technology enables PL images to be measured with acquisition times less than 1 s.1 Only a few years after its introduction, research institutes and silicon wafer and solar cell manufacturers worldwide use commercial PL imaging equipment, which is offered by BT Imaging, a UNSW spin-off company.2

How it works

In PL imaging of silicon samples, the entire surface of the sample—typically 156 × 156 mm—is illuminated homogeneously. The external optical excitation stimulates luminescence emission from the sample, which is captured with an infrared (IR) camera. In other words, a photograph is taken of the luminescent sample (see Fig. 1). With both excitation and detection being optical signals, PL imaging is contactless and can be applied to a wide range of samples, including silicon bricks and wafers. The measured luminescence comes from spontaneous emission within the sample; hence, the PL intensity allows assessment of the bulk electronic quality inside the sample. So photoluminescence imaging can be likened to x-ray imaging of wafers and cells.

Click to EnlargeFIGURE 2. Bulk minority carrier lifetime in microseconds can be obtained from a PL image on a silicon brick. Variations in bulk lifetime due to crystal defects (dark line-shaped features) and due to contaminated regions (dark bands at the bottom and the top) are assessed in a matter of 30 s, where previously several hours were required.

Experimental challenges in PL imaging applications are: a) the required continuous optical power of typically 100 mW/cm2 over an area of at least 156 × 156 mm; b) the need for homogeneous illumination over that area; and c) the low luminescence quantum efficiency from silicon. For example, on some of the measurements on as-cut wafers discussed further on, the number of luminescence photons emitted by the sample per incident photon is as low as 10-8.

High-power near-IR fiber-coupled laser diodes are often used as light sources in PL imaging. They provide the required high total optical continuous-wave power of more than 50 W, are sufficiently monochromatic and can easily be collimated to allow separation by optical filtering of the weak luminescence signal from the orders-of-magnitude stronger reflected excitation signal, and they can be turned into uniform illumination sources over large areas using customized beam-shaping optics. Furthermore, they are cost-effective, which is particularly relevant for inline PL imaging in PV production.


The production of a standard industrial solar cell includes the casting of feedstock into silicon ingots, cutting of ingots into bricks and wafers, and finally the processing of wafers into solar cells. The key quality parameter of the finished solar cell is its solar energy conversion efficiency. The latter depends strongly on the electronic material quality, which is given by the so-called excess minority carrier lifetime, and by device parameters such as the series resistance. Special PL imaging applications to assess these parameters have been developed with applicability across the entire PV value chain. Often these applications involve measuring and analyzing two or more images taken under different excitation and/or detection conditions.

Click to Enlarge
FIGURE 4. Analyzing PL images taken under different illumination and bias conditions shows series resistance (Ω/cm2) information. The pattern of features with enhanced series resistance is caused by the metal belt in the firing furnace.

Bulk lifetime measurements on silicon bricks: Application of PL imaging to silicon bricks prior to wafer cutting allows assessment of the electronic material quality at an early production stage. A PL image taken on the side facet of a 24 × 15.6 cm silicon brick can be calibrated into a color scale that represents the bulk minority carrier lifetime in microseconds (see Fig. 2).3 Measurement time for this image was 30 s. The same measurement with equivalent spatial resolution using conventional μ-PCD mapping would take several hours. A further advantage of PL imaging is that it allows measurement of the so-called "bulk minority carrier lifetime," a more relevant physical quantity in practice than the "effective minority carrier lifetime," which is measured by μ-PCD mapping. The bulk lifetime information from PL can predict wafer quality and provide feedback on feedstock quality as well as offering a fast process monitoring tool that allows process optimization and process excursions to be detected early. An example of process excursions in production is the formation of slip lines in monocrystalline silicon ingots, which can be reliably detected by PL imaging on silicon ingots and slabs without the need for any surface preparation.4

As-cut wafer characterization: Wafer manufacturing and cell manufacturing are separate parts of the PV value chain, generally performed in different factories by different companies. Photoluminescence imaging can be used for both outgoing quality control (wafer makers) and incoming quality control (cell makers). An as-cut multicrystalline silicon wafer can be imaged without any surface preparation (see Fig. 3). The BT Imaging iLS-W1 used for that measurement is an automated inline PL imaging system with a throughput of up to 2400 wafers per hour, capable for the first time of measuring as-cut wafer PL images with less than 1 s exposure time. The red color overlay marks the position of specific types of defects related to structural defects within the silicon crystal lattice—defects that have a dominant impact on cell efficiency. The position of those defects was obtained from the PL image by image-processing algorithms trained to identify specific spatial features. Systematic PL imaging inspection of silicon wafers and automatic assessment of metrics derived using image-analysis methods allow cell performance to be predicted at the start of cell manufacturing—a unique opportunity enabled for the first time by fast PL imaging.5,6,7 Analyzing PL on silicon wafers from bottom to top of a silicon brick also enables the 3D distribution of defects within a brick to be measured, valuable information for gaining understanding and better control of the crystallization process.

Series resistance imaging: A strong correlation exists between the local PL signal and the local voltage within a solar cell. This allows quantitative assessment of the local series resistance in a fully processed solar cell.8 Photoluminescence-based series resistance imaging is based on measuring a number of PL images with different combinations of illumination intensity and bias conditions. It provides the local series resistance with high resolution in a matter of seconds. Previously the only widely used technique to measure such variations in solar cells was with a product called Corescan, a measurement technique that typically requires 40 min per cell and destroys the sample in the process.9 An example of a series resistance image from PL imaging shows series resistance patterns that match the shape of the belt in the firing furnace, providing immediate clues on the origin of the problem (see Fig. 4). With measurement times on the order of seconds, inline Rs imaging is conceivable, allowing small process excursions to be identified and rectified earlier than previously possible.

Click to Enlarge
FIGURE 3. In a photoluminescence image of a 156 × 156 mm as-cut multicrystalline silicon wafer, a red color overlay displays the result of automatic image-processing analysis for a specific defect type. Metrics derived from such image analysis allow correlation with and prediction of cell performance based on measurements on the starting material. The image was taken using a BT Imaging iLS-W1 inline inspection system.

Photoluminescence imaging of silicon bricks, wafers, and solar cells has seen rapid development since its introduction in 2005. The application of PL imaging is not limited to the specific applications mentioned here. A large number of additional applications have now been demonstrated and that number is growing even more rapidly with the availability and penetration of commercial PL imaging systems offered by BT Imaging. The company recently launched specialized equipment for inline quality control and process monitoring. With its combination of speed and high spatial resolution, PL imaging is on the verge of being adopted for inline monitoring in PV manufacturing. It can also be expected to find applications in other arenas, such as the semiconductor and LED industries.

1. T. Trupke et al., "Photoluminescence imaging of silicon wafers," Appl. Phys. Lett., 89, p. 044107 (2006).
3. T. Trupke et al., "Photoluminescence imaging on silicon bricks," 24th European Photovoltaic Solar Energy Conf., Hamburg, Germany (September 2009).
4. W. McMillan et al., "Slip line detection on mono ingots and wafers at production throughput using photoluminescence," 25th EPVSC, Valencia, Spain (September 2010).
5. M. Demant et al., "Analysis of luminescence images applying pattern recognition techniques," 25th EPVSC, Valencia, Spain (September 2010).
6. J. Haunschild et al., "Comparing luminescence imaging with illuminated lock-in thermography and carrier density imaging for inline inspection of silicon solar cells," 24th European Photovoltaic Solar Energy Conf., Hamburg, Germany (September 2009).
7. W. McMillan et al., "In-line monitoring of electrical wafer quality using photoluminescence imaging," 25th EPVSC, Valencia, Spain (September 2010).
8. T. Trupke et al., "Spatially resolved series resistance of silicon solar cells obtained from luminescence imaging," Appl. Phys. Lett., 90, p. 093506 (2007).
9. A.S.H.v.d. Heide et al., "Mapping of Contact Resistance and Locating Shunts on Solar Cells Using Resistance Analysis by Mapping of Potential (Ramp) Techniques," 16th EPVSC, Glasgow, UK (2000).

Editor's note: Find out more about the market for photovoltaic equipment and market outlook at the 2011 Lasers & Photonics Marketplace Seminar in January. For information or to register, visit

Thorsten Trupke is a founder and CTO, and Wayne McMillan is VP, sales & marketing, at BT Imaging Pty Ltd., 2-12 Foveaux Street, Surry Hills, NSW, 2010, Australia; e-mail:;

Tell us what you think about this article. Send an e-mail to

More Laser Focus World Current Issue Articles
More Laser Focus World Archives Issue Articles

Most Popular Articles


Handheld Spectrometers

Spectroscopy is a powerful and versatile tool that traditionally often required a large and bulky instrument. The combination of compact optics and modern pa...

Fracking, climate change, and lasers:  new tools to reduce fugitive methane emissions

This webcast, sponsored by Hamamatsu Corporation, covers recent developments and field deployments of compact quantum-cascade-laser (QCL)-based methane senso...

Opportunities in the Mid-IR

The technology for exploiting the mid-IR is developing rapidly:  it includes quantum-cascade lasers and other sources, spectroscopic instruments of many...

Fiber Optic Sensors – Fundamentals, Principles and Applications

In this webcast, sponsored by Nufern, we focus on optical fiber sensing technology.  Fundamental concepts will be presented first, followed by the under...
White Papers

Narrow-line fiber-coupled modules for DPAL pumping

A new series of fiber coupled diode laser modules optimized for DPAL pumping is presented, featur...

Accurate LED Source Modeling Using TracePro

Modern optical modeling programs allow product design engineers to create, analyze, and optimize ...

Optical Isolators Improve Engraving Performance of Pulsed Fiber Lasers

The deleterious effects of back reflections on pulsed fiber lasers used in marking and engraving ...
Technical Digests

WAVELENGTH-SWEPT LASERS: Dispersion-tuned fiber laser sweeps over a 140 nm range for OCT

By eliminating the use of mechanical tunable filters and instead tuning by intensity-modulation i...

Keeping pace with developments in Raman spectroscopy for molecular and nanoparticle research

For demanding or custom spectroscopy solutions, care must be taken in selecting and integrating a...

HIGH-POWER FIBER LASERS: Working in the kilowatt regime

High-power materials-processing fiber lasers are available in an increasing variety of forms, as ...

Click here to have your products listed in the Laser Focus World Buyers Guide.


Phantom ir300

The Phantom ir300 provides extended spectral response beyond visible light spectrum up ...

Miro Airborne

Miro Airborne is a high-speed camera designed for airborne applications.

Phantom Miro Family

The Phantom Miro family are small, lightweight digital high-speed cameras.


Photonics Bretagne

Offers a cluster composed of research centers, schools and companies all in the field o...

Raw Communications

Provider of marketing services in the fiber optic data communications industry includin...

XiO Photonics B V

Offers strong competence in integrated optical products for visible light applications....

Social Activity

Copyright © 2007-2014. PennWell Corporation, Tulsa, OK. All Rights Reserved.PRIVACY POLICY | TERMS AND CONDITIONS