1. Introduction

Laser direct-write (LDW) techniques are an increasingly attractive alternative to traditional lithographic technologies for many surface micropatterning applications. The former offer significant advantages in terms of processing speed and simplicity, and allow for the relatively simple combination of multiple dissimilar materials within a single microdevice.

A rapidly growing family of additive LDW techniques are the so-called Laser-Induced Forward Transfer (LIFT) methods [1, 2, 3]. For LIFT, a thin film (the donor) typically of the order of 100 nm - 1 µm thick, is coated onto a transparent substrate (the carrier). The coated carrier (the target) is brought into close contact (typically less than 10 µm separation) with the substrate to be deposited onto (the receiver). Film material is transferred by irradiating the carrier-film interface through the carrier, with a focussed or demagnified laser pulse, with the resultant melting and ablation of the film propelling a section of film to the receiver [1].

 

Fig. 1. Schematic of LIFT and deposits of Cr on Si under optimal printing conditions from a 30nm (left) [4] and 80nm donor.

Download Full Size | PDF

Broadly speaking, LIFT techniques fall into 3 categories, distinguished by the composition of the donor film. In the first category, the process is performed using a single donor layer, consisting entirely of the material to be deposited; various metals [4, 6], biomaterials [5], and oxides [7], amongst other materials have been patterned this way. The maximum achievable resolution is currently around 300 nm [4]. There are essentially two ways in which forward-transfer can be achieved from a single-layer film: either the film melts through and sub-laser spot size droplets form on the free surface [4, 6] or ablation of the film at the constrained interface propels material to the receiver [5]. For donor films that are strong or brittle, e.g. metals, glasses, ceramics etc., ablation-driven LIFT can be a violent process that typically results in disintegration of the donor film during lift-off from the carrier [5]. The two transfer regimes are shown schematically in Fig. 1 alongside SEM micrographs of Cr depositions under optimal printing coniditons from a 30nm (left) [4] and 80nm (right) donor. Evidently there is an inherent disadvantage to this method in that the donor film must act as its own propellant, either by melting or ablating, resulting in significant damage during transfer.

The second form of LIFT eliminates the need to directly irradiate the donor film by inserting a sacrificial layer of a second material, often referred to as a Dynamic Release Layer (DRL), not intended for deposition, between the donor and the carrier [3]. The potential of a range of materials, including metals [8], hydrogenated silicon [9] and polymers [10, 11, 12], to be used for the sacrificial layer has received significant study in recent years. Particular success has been achieved in transferring ’soft’ materials, e.g. biomaterials in solution [8] and polymers [10]. Recently the deposition of all the components required for an organic LED, in a single deposition event from a layered donor film, was demonstrated using a designed triazene polymer as the sacrificial layer [12]. Potential complications of this technique are the possibility of contamination of the deposited donor material with residual propellant and the need for the donor film deposition process to be compatible with the propellant, for example it would be impossible to grow a single-crystal film on a low-dissociation temperature polymer propellant.

The final LIFT variant commonly encountered in the literature involves preparing the donor material in particulate form and mixing it with a sacrificial matrix material. This technique is known as Matrix-Assisted Pulsed Laser Evaporation-Direct Write (MAPLE-DW) [2]. MAPLE-DW has been demonstrated to be suitable for the micro-manufacture of various devices including electronic components and chemical sensors [2] and biomaterials [13]. The main challenge with MAPLE-DW is the necessity to use donor material in a particulate form. This can mean that deposited material properties can be significantly different from those of the bulk; for example, the resistivity of Ag lines deposited by MAPLE-DW was found to be approximately 2 orders of magnitude higher than similar lines deposited by LIFT [2]. Such discrepancies arise from the unavoidable porous nature of MAPLE-DW deposits.

2. Ballistic laser assisted solid transfer (BLAST)

Despite the great progress made using LIFT techniques, there is still no way to directly print solid segments from ’hard’ films, i.e. metals, glasses, crystals, ceramics etc., essentially intact after transfer to maintain the as-deposited properties of the donor. Such a capability would be ideal for the LIFT of, for example, single-crystal, oriented, or single domain donor films, or for donors containing pre-made structures. Recent results using a triazene polymer DRL have shown promise in this area [12]; however, there are the potential issues of residual-polymer contamination and difficulty in depositing certain donor materials on the heat-sensitive polymer.

In this work, we present initial results from a further complementary LIFT technique for the deposition of solid sections of ’hard’ donor films, which do not display significant evidence of melting or ablation damage during transfer. This technique we have termed Ballistic Laser-Assisted Solid Transfer (BLAST). BLAST does not require a DRL material and so is potentially applicable for the forward transfer deposition of any ’hard’ donor.

It is known that LIFT has a well-defined fluence threshold, J_T, dependent on the specifics of the donor film (e.g. thickness, absorption, and thermal properties) and patterning laser (e.g. wavelength and pulse duration), below which no forward transfer occurs [4]. However, it is also known that evaporation of small amounts of material from a laser-irradiated surface can occur below the fluence threshold for significant bulk modification [14]. Whilst the amount of evaporation is small enough to be considered insignificant for micromachining, in a LIFT environment any evaporated material would be trapped between donor film and carrier.

The idea behind BLAST (see Fig. 2(a)) is to delaminate the donor from the carrier prior to transfer using a pulse with fluence J<J_T. This means that, to achieve transfer, only the shear strength of the donor must be overcome and not also the stiction of the film to the carrier. The result is a gentler transfer process and the possibility of less damage during transfer.

 

Fig. 2. Schematics of BLAST (a) and forward-transfer with a shaped pulse (b). Inset to (a): SEM micrograph of a sample deposit with 10, 310mJ/cm ² pulses from the 80nm donor.

Download Full Size | PDF

3. Experimental

A Ti:sapphire laser (800nm, 110fs, 250Hz rep rate) was used for BLAST experiments. The use of ultrashort pulses for BLAST is considered to be preferable so that the laser-induced damaged area in the donor is minimised. Laser pulses of ≈4mm diameter were centrally incident on an ≈450µm circular aperture, resulting in an approximately uniform circular beam. An image of the aperture was relayed to the target resulting in an ≈12µm diameter cicrular spot at the carrier-film interface, as measured by the laser damaged area. A high-speed, computer controlled shutter (Uniblitz LS3) was used to control exposure of the target to the laser.

Cr films 80 and 160nm thick were evaporated onto fused silica for BLAST targets. Si wafers were used as receivers and target-receiver separation was kept around 2µm by means of Mylar spacers. All results presented here were performed under vacuum at ≈0.1mbar, although BLAST has also been observed to work in an ambient atmosphere. Samples were mounted on a 2-axis, computer-controlled translation stage.

4. Results and discussion

The threshold fluence for single pulse forward transfer from both donors was measured to be around J_T≈390mJ/cm ², in very good agreement with the value, measured in a previous study, for the onset of significant ablation and phase explosion in a thin (30nm) Cr donor film [4]. Delamination of both donors without transfer could be observed as a bulging of the films following irradiation with a single pulse with J>0.75J_T.

The inset to Fig. 2(a) shows an SEM micrograph of a typical deposit obtained from the 80nm donor film using 10 pulses with fluence of J=310mJ/cm ². The well-defined edges and lack of surrounding debris clearly indicated forward transfer in solid phase. Note that fluence was below that measured previously for the onset of melting in a Cr film [4].

Deposits like that shown in Fig. 2(a) demonstrated the ability of multiple-pulse BLAST to forward transfer donor material in solid phase, however the shapes of such deposits varied considerably. To obtain reproducible deposits, it was necessary to directly machine a weakened area into the donor film that could preferentially shear to release the pellet. The easiest way to achieve this was to adjust the carrier-film interface position such that, instead of coinciding with the best image of the aperture, the interface was positioned in the Fresnel (near-field) diffraction regime. In this way, an outer ring could be added to the laser spatial profile which defined the weak region during the pre-transfer evaporation stage. Figure 2(b) demonstrates the potential benefits of forward-transfer using shaped pulses.

 

Fig. 3. Cross-section (a) and line-plots (b) of the calculated intensity profile with distance from the best image plane (z=0).

Download Full Size | PDF

To calculate theoretically the beam profile used when performing BLAST with Fresnel-diffraction spatial profiles, a simple model using fast Fourier transforms (FFT) was constructed. Assuming beam profile at the best image plane, f, then the beam profile at a perpendicular distance z from the best image plane, I(z), can be calculated using FFTs as

where iFFT denotes an inverse FFT and H(z) is the transfer function

where F is the associated frequency scale of the FFT.

Figure 3(a) shows a cross-section of how the calculated intensity distribution varies with z for 0≤z≤50µm (λ=800nm). Figure 3(b) shows plots of the intensity distribution for indicated values of z. As can be seen, for certain values of z, e.g. z=2,14,32µm, the beam displays an outer ring and central intense peak of comparable intensity, whilst the intensity across the rest of the laser spot is lower. We believe that beam profiles near the best image (e.g. at z≈2µm) are optimal for forward transfer as the irradiance is approximately constant across the centre of the beam and only very slightly higher (≤5%) in the ring region. The threshold for ultrashort laser damage is typically very well defined so it is desirable that no part of the beam is dominant in intensity as this would cause damage. It should be noted that the diameter of the ring region is typically around 10 - 12µm for z≤5µm.

4.1. LIFT and BLAST with shaped pulses

Single-pulse LIFT and multiple pulse BLAST experiments were performed using Fresnel profiles. An SEM micrograph of a deposit obtained from the 80nm donor with a single shaped pulse (z≈5µm) is shown in Fig. 4(a). Exact values of J were difficult to determine given the non-uniform intensity profile, but within the ring J_ring≈360mJ/cm ² and J_cent≈340mJ/cm ² in the centre. The outer ring of the laser caused melting of the boundary region, resulting in a reduction in the force required to expel the deposit. The benefits of the Fresnel profile for LIFT are apparent in the improved smoothness and uniformity of the deposit, and the reproducible circular shape (c.f. Fig. 1). Visible damage was restricted to the boundary region.

 

Fig. 4. SEM micrographs of LIFT (a), and BLAST from the 80nm (b) and 160nm Cr donor (c).

Download Full Size | PDF

Figure 4(b) shows an SEM micrograph of a deposit fabricated using 10 shaped pulses from the 80nm film, with z≈5µm, J_ring≈320mJ/cm ², and J_cent≈300mJ/cm ². As with the previous BLAST deposits, the lack of surrounding splatter and well-defined edges indicated transfer in solid phase. Comparison with the inset to Fig. 2(a) clearly demonstrates that shaped pulses were preferable for BLAST studies in terms of reducing damage and reproducible shapes.

An SEM micrograph of a BLAST deposit produced from the 160 nm Cr donor using 9 pulses with J_ring≈350mJ/cm ², J_cent≈330mJ/cm ², and z≈5µm is shown in Fig. 4(c). Evidently the deposit was again transferred in solid phase, but the thicker donor film, and hence greater force required for shearing, resulted in pellets with slightly less regular edges. Higher fluence was required, however the thicker donor also provided increased capacity for heat to diffuse away from the irradiated region, lessening damage to the pellet compared with similar deposits from the thinner film (c.f. Fig. 4(b)).

A strict interplay between number of pulses and laser fluence was observed, and by adjusting the fluence accordingly it was possible to achieve solid transfer with between 2 and ≈30 pulses from both donor films. However, fluence was the key parameter to control the temperature reached by the donor. It was found that solid pellet transfer was only possible with fluences between 0.75J_T and 0.95J_T, with the optimum fluence primarily dependent on the donor thickness. This placed a limit on the number of pulses to obtain best deposits of ≈5-10 pulses. Beam shaping had little effect on the optimum fluence for solid transfer or number of pulses required and only affected the reproducibility of the deposit shape.

5. Conclusion

We have demonstrated that, using multiple sub-threshold femtosecond pulses for forward transfer deposition of Cr films, solid pellets of ≈10µm diameter could be transferred. It was observed that the initial pulses delaminated the donor from the carrier before forward transfer resulting in a gentler lift-off process. It was found that the optimum fluence and number of pulses for solid pellet transfer were primarily determined by the film thickness.

Pulses exhibiting a Fresnel diffraction profile were used to improve the reproducibility of pellet shape by defining a weakened region into the film prior to transfer. The laser profile was observed to have little effect on optimum fluence and number of pulses.

The results indicate that BLAST has the potential for the forward transfer of a range of delicate or heat-sensitive, solid-state donor films, where existing techniques are not applicable.