Trap design and construction

The design for the spiral baffled traps with a cylindrical inlet was conceived by Professor Chris Todd of the Gatty Marine Lab prior to the beginning of this study. The addition of the conical entrance was formulated through group discussions that took place at the beginning of the study in 2003.

The traps were built from 50 ml polypropylene, specimen tubes (Greiner Cellstar, product no. 210270;www.greinerbioone.com). They have a 2.8 cm diameter and are 11.5 cm tall. A cylindrical skirt extends 1.5 cm from the tube wall

surrounding a conical base.

The trap design can be described in three sections. The bottom section is the collection chamber. This was built from a complete specimen tube. The conical bottom was cut away with a scalpel that was first heated in a Bunsen burner flame. The plastic melted away easily with very little disfigurement of the tube walls; however the hot blade was kept a safe distance of a few mm from the wall and the skirt so that the heat from the blade did not warp or texture the surface. Once the cone had been cut and pushed out of the tube the excess cone plastic could be carefully shaved with a sharp scalpel blade until it was flush with the internal walls of the tube. The threaded cap end of the tube trap then provides a watertight seal for closing the trap.

The second section of the trap was composed of two sections of baffles. These limited turbulent washout within the trap thereby encouraging the sinking of passive particles and improved retention of the killing solution. The lower part of that section consisted of full baffles. Tubes were cut 3 cm from the skirt end so that they included only the skirt, the cone and some of the tube wall. A lathe was used to produce a

clean, straight cut. A metal ruler was then gently scraped around the cut end to remove any remnant plastic and clean the edge. This was important so as to provide a neat joint with the adjacent section. The conical end of the part was cut down with a lathe to leave a 0.5 cm hole at the end of the cone flush with the skirt. Four bleed holes were created at equidistant points on the baffle with a thick paper clip wire (approximately 1 mm) that had been heated in a Bunsen flame. These were made as close to the skirt wall as possible whilst not disfiguring the wall. They allowed air to bubble up through the baffles whilst the trap was re-filled, ensuring a constant volume of killing solution amongst days. Pairs of 3 cm full baffles were aligned in sequence above the collection chamber. A third full baffle was included above these two parts, however this baffle aperture was only 1.7 cm, including only the skirt and cone. To accurately cut the part close to where the tube wall ended, and the skirt and cone began, the lathe cut slightly above the join and then a circular sander was used to grind the plastic to the desired point. The baffle and bleed holes were created in the same manner as the large full baffles.

The upper part of the baffled section was constructed of parts similar to the smaller full baffle. However, rather than creating a hole at the base of the cone, three quarters of the cone were removed with a hot scalpel blade leaving a triangular ‘flap’ or quarter baffle. The triangular baffle was cut over a piece of graph paper to aid the accurate repetition of the baffle shape. Four baffles were arranged in a progressively 90° offset pattern creating a clockwise descending spiral of quarter baffles.

The third section was the inlet. Multiple designs of inlet were deployed throughout this project. They were essentially constructed in one of two ways. The first and earliest design was the same as the 3 cm part used in the large baffles. The

cone was removed using the same method as the collection chamber. This left a 5 cm2 cylindrical inlet to the trap.

The second design type was a coned inlet. The end of a coned, skirtless 50 ml specimen tube (Greiner Cellstar, product no. 210261;www.greinerbioone.com)was cut 3 cm from the tip of the cone. The design of the inlet could be modified according to requirements by carefully grinding the tip with a circular sander to produce the desired inlet area. Traps of 0.25 cm2, 0.5 cm2, 1 cm2and 2 cm2aperture area were built and deployed throughout this project. Grinding was done after casting the traps in resin. Because the inlet must be produced with a degree of accuracy it made the coned end easier to handle. Henceforth, these coned traps are referred to as 0.25 cm2, 0.5 cm2, 1 cm2and 2 cm2traps.

The trap parts were taped together with sellotape to hold them in place. The sellotape was stretched because it was wrapped around each part to produce a tight fit. It was important at this point to ensure the trap does not bend. Rolling over a flat surface indicated whether this occurred and it was easy to identify any mis-aligned parts.

The trap was then set in a hard resin. Before this was possible it was necessary to construct a mould in which the resin would be poured around the trap. A 25 cm long section of 4 cm polypropylene waste pipe was used as a model trap to create the silicon mould. Two sets of three 1 cm wide shrink-wrap plastic collars (two layers each) with two 0.7 cm gaps between each of the three collars were positioned at the top and bottom of the model trap. The trio near the inlet end of the tubes was spaced 8 mm from the model trap inlet end; the lower set of collars began 11 mm from the cap end. For the 5 cm2trap moulds, the end of a skirted Greiner tube was attached to the inlet end of the trap mould. The Greiner tube was wrapped with sellotape 1.4 cm from

the start of its skirted end so that it would fit snugly inside the plastic pipe with the unwrapped cone section protruding. For the coned inlet mould the end of a skirtless Greiner tube was used in exactly the same manner. At the cap end of the model trap a Greiner tube was inserted into the waste pipe with the cap end protruding. Again tape was applied, this time 27mm from the capped end to seal the thinner centrifuge tube against the waste pipe.

1 kg of Flexil-S RTV-30C silicon and standard 5 % green catalyst (Jacobson Chemicals;www.jacobsonchemicals.co.uk) was used to produce a trap mould from the model trap. 500 g of liquid silicon were initially mixed and fully degassed in a vacuum pump. The trap model was laid in a 10 x 10 x 25 cm plastic tub. The tub was half filled with the liquid silicon and left to go-off around the model trap. A mould release agent (J-Wax, Jacobson Chemicals;www.jacobsonchemicals.co.uk) was liberally applied to the surface of the first half of the mould and then the second half of the mould was poured on top following the same protocol. Once the silicon had set the two halves were pulled apart and the model trap removed. The cap end of the mould was neatly cut away with a scalpel before the thicker part of the mould formed by the 4cm waste pipe began. Notches were cut into the mould down to the widest part of the trap mould. This would allow the resin to be poured in when the sellotaped trap parts were arranged in the mould.

To set the trap in resin the mould was sprayed with J-Wax and the taped trap parts were laid in the mould. The mould was taped together firmly between two wooden boards to ensure it stayed rigid and no leakage could occur between the mould halves. 165 g of epoxy (Robnor Resins Ltd., Product no. PX771C/NC;

no. HY1300GB) and degassed in a vacuum pump. The resin was poured into the mould around the trap parts and the mould was left overnight to set.