A novel, water-based stripper offers a fast, environmentally friendly approach to high quality removal of photoresist and the carrying out of metal lift-off, claims Dirk Schumann from Bubbles & Beyond.

Although lasers,LEDs and transistors are significantly difference species, their manufacture involves many similar processing steps. This includes the coating of surfaces with photoresists, which can be subsequently stripped or used as the foundation for a metal lift-off process.

One metric for judging how good these resists are is the ease with which they can be removed after certain processing steps. Why is that important? Because the resist can change during subsequent processing – for example, thermally activated cross-linking may take place, making the subsequent stripping step more complex.

Another factor that can complicate resist removal is metal layer deposition. Subsequent processing must not damage this layer, ruling out the use of materials with aggressive stripping chemistry. And complicating matters further, microelectronic structures are getting increasingly complex: Their fabrication demands the use of fluids that remove resists out of graves or bridge patterns that cannot be directly accessed from the top of the structure.

On top of this list of device-related concerns, engineers must take into account the impact of the stripper on the environment, and whether it complies with an increasingly strict set of regulations. Most state-of-the-art resist removals use solvents, such as N-methyl-2-pyrrolidone (NMP), D-methyl sulfoxide (DMSO) and acetone; or even more aggressive mixtures based on ammonium hydroxide/hydrogen peroxide (SC1) or the harmful Piranha Clean. Plasma ash processes are also widely used in industry, and involve etching with gases.

Going green

At Bubbles and Beyond, a start-up from Leipzig, Germany, we are addressing the need for environmentally friendly, effective stripping chemicals with our development of novel, water-based stripping fluids, which we refer to as phase fluids. They work in a neutral pH range, are biodegradable, do not contain of aggressive ingredients, have a high water share and are free from N-Methyl-2-pyrrolidone (NMP), which is a reproductive toxicant.

Working in partnership with the Fraunhofer Center Nanoelectronic Technologies in Dresden, Germany, we have benchmarked the performance of our phase fluids – that have the commercial name lisoPUR – against a conventional photoresist alkaline stripping agent. These tests, involving the removal of photoresists from silicon or glass substrates, reveal that our phase fluid has a high degree of cleaning efficiency and does not cause significant contamination. Thanks to these strengths, our novel fluid is suitable for use in the compound semiconductor industry.

Our strippers are liquid-liquid-based complex fluids, which form dynamic structures and are an enhanced micro emulsion. They differ from classical emulsions, which feature dispersed spherical oil or water droplets, and are built up from dynamic and flexible plasmicells – these are globular shapes of fluid, shown in Figure 1.


Figure 1: Freeze fracture of a phase fluid makes it easy to identify the globular plasmicells

These plasmicells interact with each other, often changing their forms within milliseconds. This leads to the creation of a highly dynamic, molecular super-structure with a fractal dimension that alters shape between 1,000 and 8,000 times every second. Thanks to this behaviour, phase fluids exhibit a highly effective mode of action: They penetrate layers through the smallest openings to lift off material from the surface (see Figure 2). Initially, the fluids start to penetrate into the photoresist. Then, thanks to low surface tension, they start to creep the varnish and fragment the resists, before lifting off these fragments. It is then possible to remove all residues from the surface with a water rinse.


Figure 2: The working principle of the Bubbles & Beyond phase fluid lisoPUR. The layer (orange) is penetrated, fragmented and finally lifted off from the substrate (red)

Addressing contamination concerns

With any new form of stripping chemistry, there are concerns over contamination. To alleviate these fears we have carried out investigations using Fourier transform infrared spectroscopy. In addition, we have developed an adequate rinse and drying process after the phase fluid cleaning step. With this procedure, the resulting interface is suitable for subsequent standard cleaning sequences. Our rinse-off procedure also ensures a low addition of post-stripping defects.

Scrutinising surfaces with Fourier transform infrared spectroscopy provides evidence of full surface recovery after the application of a phase fluid (see Figure 3). It is clear from these spectra that the fluid leaves no residues after a rinse off with water. Regular cleaning rinses, such IPA, can also be used. Another characteristic of the phase fluids is revealed by these results: Adding water (or alcohol) to the phase fluid immediately stops its activity.


Figure 3: Spectra reveal surface recovery after processing. The upper trace shows the silicon sample prepared with a phase fluid (without rinse), while the middle and lower plots reveal  surface recovery after rinse off with deionised water and the spectra for the bare silicon reference, respectively

In addition to this study of molecular surface recovery, we have undertaken a ‘high scale’ assessment of remaining residues, searching for particles with a KLA-Tencor SP2 tool. This effort commenced by applying a fluid, via a single spin process, to a 300 mm silicon wafer.

Post-cleaning followed, using deionised water, followed by a shorter step with diluted, cold SC1 – a mixture of ammonium hydroxide and hydrogen peroxide. The KLA-Tencor tool, set to detect particles with sizes from 0.12 µm to 1 µm, determined that just 38 more defects were found after processing (see Figure 4).


Figure 4: A KLA Tencor tool enables particle measurement on a 300 mm silicon wafer subjected to a single spin process. The scans compare before and after phase fluid application (removal is followed by a deionised water rinse and SC1 short cleaning)

Although these first results are only on the laboratory scale, they are very encouraging, showcasing the promise that phase fluids have throughout the entire semiconductor industry.

By selecting the optimum conditions – such as the right temperature, ultrasonic treatment and drying procedure – it is possible to realise fast stripping of various photoresists, including positive and negative varnishes. For example, these harmless process fluids can remove a photoresist greater than 10 µm-thick in less than 5 minutes to leave a pristine surface.

We can alter the formulation of our phase fluids so that they can be tailored to a particular task. Through our work with the Fruanhofer Center Nanoelectronic Technologies, we have found that cleaning efficiency increases by heating the fluid, and also using steps such as agitation and ultrasonic treatments.

The opportunity to work with a water-based formulation, rather than harmful fluids, has caused quite a stir within the market. Trials have revealed that if our phase fluid is heated to 80°C, metal lift off takes just 10 minutes. To realise full compatibility with a semiconductor environment, the approach involves a deionised rinse process, followed by SC1 cleaning when necessary. Note that phase fluids can be used in either bath or single-wafer processing.

These results highlight the very optimistic outlook for phase fluids. Initially they’ll be used for photoresist removal and metal lift-off, but as time goes by, they will start to also make an impact in the cleaning of various pieces of equipment, such as catch cups and photomasks.

Further reading

www.intelligent-fluids.de