04 Sep 2008
As the march of the fibre laser continues unabated, Bill O'Neill looks at a new generation of MOPA-based sources that are set to challenge DPSS Q-switched technologies.
The meteoric rise of the high-power industrial fibre laser has seen the number of applications using this technology increase significantly. Today, any self-respecting laser scientist is likely to have a number of fibre laser systems in order to satisfy a wide range of materials processing requirements.
The recent advances in design, performance, cost reduction and brightness of modern ytterbium (Yb)-fibre lasers have forced users to rethink their processing options when it comes to marking and machining materials with 1064 nm light.
In addition to singlemode continuous-wave (CW) and quasi-CW Yb laser architectures with power levels from watts to tens of kilowatts, a new entrant on the market is a versatile pulsed master oscillator power amplifier (MOPA) Yb fibre laser. A 20 W MOPA high-brightness Yb fibre laser can offer a range of pulse parameters from 20 to 200 ns FWHM, peak powers approaching 2 GW cm–2, and pulse repetition rates from CW up to 500 kHz. This level of performance opens up a significantly broader spectrum of material processing applications ranging from high-speed marking to micromachining.
Making the most of a MOPA
The majority of industrial laser users are keen to obtain the best possible combination of power, brightness and temporal control. Conventional Q-switched DPSS lasers will provide the user with a range of wavelengths from the ultraviolet to the infrared, repetition rates typically less than 200 kHz, and low M2 values typically less than 1.2. Temporal control is often limited by the physical response of the active medium, with pulse widths of the order of tens of nanoseconds being typical depending on the pulse repetition rate.
DPSS lasers generally have a restricted range of pulse energies, which falls as the repetition rate is increased. This limitation restricts the processing options and reduces productivity at high repetition rates. The complexity of modern DPSS systems invariably leads to high product costs, which affects the financial returns that one can make when offering materials processing solutions.
A MOPA-based fibre laser (see figure 1) takes the output from a directly modulated seed laser and passes it into a Yb-based fibre amplifier. The MOPA configuration provides increased pulse energy and peak power and greater control of pulse parameters that can be set at different stages.
The system considered in this instance has the following specifications: peak emission wavelength 1064±10 nm; CW to 500 kHz; maximum pulse energy 0.8 mJ; peak power 15 kW; pulsing capability (25 preset waveforms for optimized peak pulse power at specified repetition rates); collimated output beam diameter 3.1 mm; and an integrated Faraday isolator. The beam intensity spatial profile is near Gaussian with M2<2. Output pulse stability is typically ∼1% (1 s) with a beam divergence of <1.6 mrad. Using single-emitter telecom laser diodes with an individual mean-time-to-failure of greater than 450 000 hours as the seed source also guarantees high reliability and extremely long lifetimes.
The MOPA approach creates a versatile laser with a performance exceeding that offered by typical DPSS systems such as Nd:YAG and YVO4. Fine processing controls are enabled by enhanced peak powers and pulse energies. The average power output can be maintained at full repetition rates unlike DPSS systems that tend to roll off with increasing frequency.
A directly modulated laser diode gives the user the ability to select a range of pulse waveforms that can be optimized for peak pulse power at a specified repetition rate. This feature is a distinct benefit since great flexibility of pulse forms and lengths can be achieved over the whole operational frequency range.
The majority of materials processing applications have a productivity that is related to power, with most applications having energy density or peak power thresholds. Figure 2 shows a typical selection of pre-set pulse waveforms. At higher frequencies the MOPA system is able to modify the pulse envelope in order to achieve high peak-power output. Processing with these output pulse waveforms provides much greater control than traditional Q-switched systems, which tend to produce flattened pulse profiles at higher pulse repetition rates. The MOPA approach offers a variety of options for processing a host of materials that can provide superior processing results.
General materials processing
Laser marking is a high-volume application, with service providers needing to process a wide range of materials. Marking of metals can be achieved by vaporization, annealing, oxidation, or surface melting. Marking of polymers involves vaporization, carbonization, foaming, bleaching or modification of polymerization characteristics.
Each technique requires particular pulse parameters very often severely testing current DPSS system technologies. Marking is a significant business with capability and cost being paramount. As spot separation determines marking visibility, high repetition rates produce clearer marks at higher speeds, which in turn requires high pulse energies at high repetition rates.
Thin-film ablation applications are increasing, particularly with the rapid growth of the TFT display and solar cell industries. Processing performance depends on the coating used and requires a broad range of pulse energies and high repetition rates to comply with strict productivity requirements.
The provision of variable waveforms from the MOPA architecture gives the user more careful control of pulse energies and width. Applications that require low pulse energies to avoid substrate damage, e.g. molybdenum on glass, can now be processed very quickly with high repetition rates.
Silicon processing
Silicon (Si) exhibits a room temperature absorption coefficient of around 0.54 cm–1 at a wavelength of 1 µm. Many researchers have cited this property as the reason why 1 µm sources should be avoided.
Whilst this is true at room temperature, the optical properties of molten Si are very different. Upon melting, the transmission of the 1 µm line is reduced to almost zero and stays at zero for as long as the melting temperature is maintained. When the heat source is removed, a low transmission level can be maintained for several hundred nanoseconds.
Given the highly dynamic response of Si under intense illumination and the subsequent reduction in transmission, it is possible to influence the machining response of Si at a wavelength of 1 µm given sufficient control over the pulsewidth envelope. The ability to deliver high peak power pulses at 1 µm can have a significant effect on resultant characteristics of Si processing.
The usual laser of choice is the Nd:YAG DPSS laser operating at 532, 355, or 266 nm. Previous work has shown that the absorption length in Si is of the order of 1 µm at 532 nm and around 200 µm at 1064 nm. These are room temperature values and as such do not relate to the conditions offered at the melting point and above. However, the transmission of Si can be reduced to a level such that the absorption length is of the order of 0.07 mm. It is therefore possible to control the nature of the interaction by carefully selecting the basic pulse profile and power density. Figure 4 shows an array of micro pits machined in Si.
One can clearly see the very limited recast and clean surface morphology (see figure 5). Si is removed through explosive vaporization and compares very well with those pits achieved using DPSS lasers operating at much shorter wavelengths.
The future
The availability of a MOPA-based Yb-fibre laser has provided considerable choice when it comes to selecting pulse profiles, pulse repetition rates and incident power density. Early processing trials for marking, thin-film ablation (see figure 6) and machining show considerable promise.
The markets for Si processing applications are set to increase dramatically as the global production of solar cells ramps up. MOPA-based systems are likely to gain increased market share in this field as power outputs are set to increase to around 100 W and beyond.
There are many processing operations that could be carried out with this new low-cost high-efficiency laser system. Results from early stage investigations discussed here will no doubt be supported by further examples as MOPA fibre laser configurations are used in a wider range of industrial laser processing operations.
• Bill O'Neill is the director of the Centre for Industrial Photonics, Institute for Manufacturing, University of Cambridge, UK. For more information, contact wo207@eng.cam.ac.uk.
• This article originally appeared in the September 2008 issue of Optics & Laser Europe magazine.
View pdf of article
© 2024 SPIE Europe |
|