And here we can understand further about the influence of some parameter (cut velocity, average power and frequency) to quality of laser cutting. [14]
It is therefore necessary to investigate processing across the full rep. rate range of the laser to determine the optimal cutting regime for Flexible PCB. From the previous experiments single pass cutting is employed.
Figure 2.12: Maximum Cut Velocity as a Function of Rep. Rate and Average Power
In the figure above we see the results of tests where the maximum cutting velocity for the FPCB is determined for repetition rate ranges from 6-22kHz at maximum output power on the workpiece. This is found by running a sequence of cut passes at increasing velocity until the FPCB is no longer cleanly cut. The rep. rate range is broader than stated in the product specification because the researchers are operating at the extremities of the laser performance in order to identify prevailing trends. A number of these trends emerge from figure 7: firstly that the highest cutting velocity is achieved at 20 kHz and that it diminishes at lower rep. rates.
On the same graph the available power of the laser is plotted, and it can be seen that maximum is at 12 kHz. This shows that even though 20kHz achieves the fastest cutting, it
does so with 17% less power available than at 12kHz. Strongly suggesting that the most efficient cutting regime is to be found at the highest standard repetition rate of 20 kHz with the highest number of low energy, low irradiance pulses.
To corroborate the findings, maximum cut velocity is found for the full range of repetition rates but at a constant average power. This power is dictated by the lowest available maximum output across the frequency range – 11.8W at 4kHz. Through the use of an external Powerlase attenuator, the experiment is repeated, but each rep. rate is attenuated to this constant average power.
Figure 2.13: Maximum Cut Velocity vs. Rep. Rate at Constant Power
The above results serve to corroborate these findings; high rep. rate processing is the most efficient regime for cutting Flexible PCB. Consequently 20 kHz is chosen as the optimum setting.
The nature of the cut kerf is found to change with repetition rate and process velocity. This provides insight into the mechanism favouring the 20 kHz regime.
Figure 2.14: Kerf Entrance vs. Rep. Rate for Full Cuts
The figure above shows the entrance kerf width for all successful cuts over a range of rep. rates. As previously observed the maximum cut velocity is significantly lower at lower rep. rates - however the dimensions of the entrance kerf are significantly wider at lower repetition rates for a given velocity. This is illustrated in the pictures below.
Figure 2.15: Entrance Kerf for 4 & 20 kHz, Cut Velocity 50mm/s – Constant Magnification
There is evidence of a much wider heat affected zone in the 4 kHz sample – plus a significantly larger region exhibiting surface discolouration. This suggests that the reason for the wider kerf is due to lateral energy transfer during the cutting process. The more powerful pulses creating a larger melt region leading to a broad kerf and also resulting in greater sideways heat conduction. It is also likely that the more intense pulses generate more surface plasma creating unwanted effects such as re-irradiation of energy and defocusing of subsequent pulses - all reducing the efficiency of the cutting mechanism as energy is dissipated laterally.
Figure 2.16: Entrance and Exit Kerf Width with Increasing Cut Velocity at 20 kHz
A further interesting observation is that of taper with increasing velocity. A consistent trend as shown in figure 11 above is that as cut velocity increases taper increases - with the entrance kerf width broadly constant as the exit kerf decreases. This offers the end user a route to control taper should it be significant to their application. Reducing taper at the expense of cut velocity.
Figure 2.17: Sections of FPCB Kerfs at Increasing Cut Velocity
By sectioning samples of the Flexible PCB at various cut rates it is possible to ascertain the resulting cut quality but also develop greater understanding as to the mechanism of kerf formation, as shown in figure 12. It is clear by inspecting these sectioned images that the kerf erodes in a geometry close to that of the near Gaussian energy distribution of the incident laser. With the most intense central section of the beam penetrating further into the material than the less intense wings. As velocity is reduced the central part of the beam which machines the deepest, penetrates the substrate to achieve a cut. An equilibrium state is reached in which, for a given part of the workpiece, sufficient energy is introduced by overlapping pulses to machine through the material, creating a cut. As velocity slows further the exit kerf widens because the less intense edges of the beam then have sufficient overlapping pulses to ablate through the substrate. Therefore at high speed, a non-penetrating
‘V-shaped’ groove is eroded in the FPCB, as speed is reduced a tapered cut is formed, and as speed reduces further, taper diminishes.
It is to be noted that the cross section cuts appear of very high quality, with no evidence of delamination between the layers, no measurable recast or Heat Affected Zone
supports the hypothesis that a 532nm Qswitched laser represents an optimal choice for cutting FPCB in terms of speed, quality and process tolerance. It should also be noted that whilst successful cutting is shown at 140mm/s and has been achieved at higher velocities, the authors recommend a maximum cutting speed of 120mm/s for this laser and optical configuration. This allows for variations in substrate and set-up, giving sufficient margin for error to ensure consistent processing in a high volume 24/7 production environment.
A further observation of the 170mm/s section shows that it is very similar to a section of a blind-via drilled in FPCB. Laser cutting and laser drilling are mechanistically similar processes. This suggests that a 532nm Q-switched DPSSL may also be suitable for via drilling of Flexible PCBs.
As the cut velocity increases the surface discolouration increases around the entrance kerf. It is clear from figure 12 that this is not due to adhering surface dross or debris. The authors suggest that it is proportional to the taper of the kerf - because at higher cut velocities the exit kerf is smaller so the molten material will be ejected above the kerf rather than below it. Such material will settle on the surface as soot and fine particles, and the Copper Oxide may exhibit colours corresponding to its oxidation state Cu (II) Oxide – Black, Cu (III) Oxide – Red/Brown. As the cut velocity slows, the exit kerf will widen and a greater proportion of the molten material be ejected down through the kerf – reducing surface discolouration. This can be seen very clearly in the figure below.
Figure 2.18: Kerf Entrance and Exit Images at Increasing Cut Velocity – Constant Magnification
A constant kerf entrance width, an exit kerf width diminishing with increasing cut velocity and increasing surface discolouration can be seen. It may be possible to reduce or eliminate this effect either by employing a post-processing wash to remove the debris, or by applying an inert assist gas to the workpiece to inhibit soot deposition. The use of an extraction system would aid the removal of such debris and compliment an assist gas strategy.
Conversely if it proves significant to the application – a slower cutting velocity will reduce this effect.
Figure 2.19: Example of Intricate Profiles Possible
The figure above serves as an example of the intricate profiles it is possible to achieve at the settings described. The ‘Tail-Fin’ kerfs are 40μm wide upon a 120μm pitch, and the cut velocity was 120mm/s throughout. As can be seen, despite the intricacy of the part, there is no evidence of distortion or thermal damage. Illustrating the suitability of this type of laser for cutting Flexible Printed Circuit Board with heat sensitive embedded components.
Chapter3 Experimental
In all cases, experiments were carried out on blank PCB of composition Copper- Polyimide- Copper and Flexible PCB (cover layer include three parts: PI, ID, paper).
Figure 3.1: PCB was marked by laser.
Figure 3.2: another PCB was marked
Figure 3.3: Flexible product
Figure 3.4: Flexible PCB
Using green laser 532nm to mark and cut on PCB and cover layer. To cut cover layer used low M² Q switch DPSS laser, it has an available repetition rate range from 20- 60KHz, with cutting speed is 30- 400mm/s and average power range from 20- 50W.