Please note first of all before you start reading this older post that I have long since included an updated/expanded version in the Mouldmaking and casting section of my Methods pages which can be accessed above.
Following on from the account of the mouldmaking process in Part 1, this post deals with the making of a lightweight, hollow cast in fibreglass. A form of this size and girth would never normally be cast solid, except perhaps if it were intended for outside and cast in concrete. Also, since the surface will be added to and painted, fibreglass is more suited to this treatment than plaster.
To backtrack a little first, the mould comprised of two straightforward halves and each of these halves consisted of a thick layer of silicone rubber reproducing every surface detail of the prototype together with a rigid and separable Jesmonite jacket to hold the silicone in the right shape. It’s very important during demoulding that the jacket can be separated easily from the silicone first, so that this can be flexed and pealed away from the cast without damaging it. For this reason the silicone skin needs to be built up around features that would otherwise cause undercutting. In the photo below the ears, for example, are completely buried and I even had to be careful to fill out the area under the chin.
Here are the two mould halves on separation from the clay prototype. A little of the clay has stuck to the silicone and to the jacket (where a clay wall was added at the base) and this needs to be cleaned off first. With large forms this is easiest to do in the bath using a shower hose, warm water and detergent.
Natural clay is generally one of the easiest materials to clean off, though it can ocasionally be rather stubborn in which case an old toothbrush is useful. Deep detail needs to be especially checked. I’ve also found that absorbent ‘micro-fibre’ cloths are ideal for initial drying because there’s no danger of leaving snagged fibres on the mould.
A little more clean-up may be necessary, this time concerning the silicone itself. When a thickener is used to convert the silicone to a paste it makes it easily spreadable for a while but towards the end of its pot life it becomes increasingly sticky and rubbery. It is still usable but it often means that the outer surface of the silicone skin becomes quite bumpy, as below. This is the surface which will receive the fitting jacket, so either Jesmonite, plaster or fibreglass resin will reproduce this surface when applied to it. If there are particularly raised strokes such as the one in the centre of the photo these may catch when it comes later to reassembling the mould parts ready for casting. These should be sliced off using either scalpel or scissors, and can be done either at this point or earlier before the jacket is made. Another technique for smoothing out the silicone surface while it is still wet is to apply clingfilm to it and smooth that out with the hand, leaving the clingfilm there until the silicone has cured. The clingfilm can then be pealed off.
More importantly, the same check should be made on the silicone edges which will meet each other. Air pockets are almost inevitable when working with thickened silicone and these will be filled when the second, corresponding half is applied causing protrusions, illustrated below, which could also not quite ‘marry’ again.
Silicone rubber does shrink a little when it cures, but generally the linear shrinkage rate is so minimal that that this doesn’t noticeably affect the overall size of casts, at least for normal sculptural purposes. Condensation cure silicones (see ‘Vocabulary’, the moulding silicones most commonly used including this one) on the whole have a shrink rate of between 0.5 and 1%, whereas with addition cure silicones the rate is usually much smaller at between 0.1 and 0.2%. As I said, this doesn’t really have any impact on size, but especially with large skin moulds like this one it is at least one of the reasons why the edge of the silicone skin piece starts pulling away very slightly from its containment jacket, as illustrated below. If this contraction occurred to the same degree and in exactly the same places on both halves of the mould it may not pose a problem but of course this is rarely the case. It means that the two mould halves won’t line up perfectly, the seam on the cast will be more pronounced in places and there will be some more work filling and correcting this. One common solution I’ve seen is to cut dovetail shapes at an earlier stage around the edge of the silicone skin before the containment jacket is made over it. These will lock the silicone in place at least round the edge. I didn’t do that in this case, thinking that I could try another idea.
Here I’ve bent a large-headed pin into a fish-hook shape and pushed it through a strip of ‘gaffer’ or duct tape. The hook was embedded firmly in the silicone, pulled just enough to close the gap and then taped in place. The tape stuck well to the Jesmonite (about a week had passed since making the jacket, allowing it to dry out thoroughly). Since most of the casting work will be done in the open mould halves the hooks can stay in position until these shells are finished.
Below is the front half of the mould set up, ready for starting the fibreglass shell. The standard procedure for fibreglass lamination is to first apply a thin coat of polyester resin (see ‘Vocabularly’ for more general details, including the various types) which is usually thickened and known as the gel coat. The purpose of the gel coat, which is carefully brushed on, is to properly capture all the surface detail without air bubbles etc. and to act as a ground layer upon which further layers of resin and matting can be applied. Without this the matting structure would most likely show through the surface. The gel coat is also normally filled, i.e. pigmented, according to the surface needed on the cast.
However I decided to approach this slightly differently, rather than buying or mixing up a separate gel coat for the initial covering. The casts needed a white surface because they would be painted, so I started by brushing on an unthickened layer of GP (general purpose) polyester resin with an amount of white pigment mixed in.
Good digital kitchen weighing scales, measuring in increments of 1g are necessary for this, especially when working with small batches. Even then it is a somewhat rough measure, but at least I could ensure that I wasn’t adding too much. The initial surface coat is the most crucial part of the casting and if something happens to this due to an incorrect mix one may not discover that until all the rest of the work has been done and the cast is demoulded. I portioned 2g of white pigment per 25ml resin, and since 25ml of resin weighs 28g (the SG is 1.12) this was safely below the maximum. Resins will generally take up to 10% pigment by weight without becoming affected.
I thoroughly mixed the measured amount of pigment with a very small amount of resin (just a teaspoonfull) first in a separate cup and put this aside. This, rather than tipping the powder pigment directly into the catalysed batch, is known as wetting and ensures that pigment and resin are properly combined. In actual fact I’ve found that powder pigments generally combine more readily and smoothly with polyester resin than they do with water. I then measured out 25ml of resin in another cup, added catalyst and mixed this thoroughly before combining with the pigment. It is usually better to mix resin and catalyst thoroughly before adding any filler. Polyester resins need only a small amount of catalyst (normally between 1% and 3% by weight). The reason for this range is that larger volumes of mixed resin will become very hot while curing (the resin is exothermic) needing only the minimum amount of catalyst whereas a small batch .. such as 25ml .. may need more. In the case of the brand I am using, Tiranti’s ‘Multi-purpose’ polyester resin, the catalyst is supplied in small dropper bottles with directions for dosage. Once catalyst is mixed with the resin one has on average 15-20mins to work with it under normal temperatures. I found during the course of this assignment that I could comfortably work with just 25ml at a time.
Here, above and below, I first tipped the catalysed mix of resin and pigment into the mould and spread this around both by rocking and using a cheap paintbrush. This coat didn’t need to be thick, but it did need to cover well, evenly and right up to the mould edges. At least with this first coat, it is important to let it cure firmly enough before the next coating is applied. I waited a little more than an hour in this case. If this is done too soon the action of the brush could break up the thin coat and cause fissures on the outer surface of the cast.
For the second, thicker coat I mixed up the same proportion of resin and white pigment but added an ammount of Fillite, which is an industrially produced ash commonly used to bulk out resins or other media. After some testing I kept with a mix of roughly 6 teaspoonfulls of Fillite per 25ml resin, which gave a ‘smoothly spreadable sludge’. This is different to thickening the resin with a thixotropic additive, which would normally convert the resin to a non-slump, spreadable but gel-like paste. The Fillite is inert, meaning that although it bulks out the mixture physically it doesn’t chemically change the consistency of the resin itself. So in this case it needs to be coaxed back up the sides for a while, at least until the resin itself starts to cure and thicken.
But I chose to use this kind of mixture because I could pour it more easily into the nose and ear cavities while spreading the rest over other areas. The Fillite gives thickness and strength while keeping the weight down, but my other reason for using it was that it makes the resin easier to cut into in the event that alterations are later needed to the surface.
There are a number of different types of fibreglass matting. One’s choice depends mainly on the undulations of the surface to be covered. As it comes fibreglass matting appears rather thick and stiff, but when ‘wetted’ with resin it becomes much more mouldable (the binder holding the strands in place dissolves allowing them to move around more). So even a heavy matting (450g per sq metre) can be coaxed into curves, but there is a definite limit. For this piece I bought a few sq metres of 300g chopped strand mat and did some sample test pieces mainly to find out how many layers I might need for strength. I cut the matting into small (about 10x5cm pieces, ordinary scissors work just fine) but it was difficult to mould these into the sharper details such as the eyes or around the chin. So in the end I used the 300g matting wherever possible but with multiple layers of fibreglass tissue (which is gossamer thin and very mouldable) on the more complicated parts. Below, the chopped strand matting on the left and fibreglass tissue to the right. The strands in the matting are not woven, merely compressed together.
For the test pieces, below, I followed all the intended steps; a thin pigmented coat, a second layer thickened with Fillite (with the aim of filling parts like the ears almost completely) and then just 2 layers of fibreglass matting fairly generously wetted with resin. The result was very strong so I decided that that the 2 layers would be sufficient for a form of this size provided that the seam was properly strengthened.
Any work involving polyester resin, but especially fibreglass lamination, can only be carried out inside if the room is equipped with a strong extractor fan .. and not even if one has all windows and doors open! In fact even the extractor fan may not be enough if the work is prolonged. The styrene vapour emitted by the resin is harmful if breathed in.
The best working procedure is, firstly in terms of preparation: to cut up a good stock of small pieces of matting (as I said I cut most pieces roughly 5x10cm but I had some larger ones for the flatter areas and either longer/thinner strips or small squares for more involved parts); to have a number of suitable brushes on hand together with at least 100 disposable plastic beakers; a bundle of wooden stirrers; a supply of paper towels and more than just a few pairs of latex gloves. I also had acetone, for cleaning up any uncured resin and for cleaning the brushes as I went. I had around 5 cheap natural bristle brushes which could be used and cleaned up in rotation. The acetone evaporates quickly so it will not usually affect the resin when the brush is used again.
First a batch of resin needs to be catalysed (as I said, I worked in 25ml batches all catalysed at 2%) and brushed generously onto a smallish area of the mould, a piece of matting laid on top and gently stippled down with the brush then quickly dabbing more resin on top. The stippling action is important (as opposed to stroking) because it forces resin into the fibres better. Also a stroking action is likely to slide the matting around too much. Once one piece has been thoroughly wetted the next piece can be applied, slightly overlapping, and so on until the resin remaining in the cup starts to get a little like treacle. If one’s established the right rhythm (and not mixed up a far too ambitious amount!) one should be left with just a brushful or two at this point and these can be smeared anywhere in the mould to make use of them. The mixture will turn from treacle to gel almost immediately at this point so one can’t lose any time in cleaning the brush now. Below, at the mould edge where the two halves will meet there’s no sense in trying to position the matting painstakingly against the line. It’s better to overlap a little, because this can be fairly easily trimmed down later and it will give a stronger, thicker edge.
For the second cast (I had to make two, and most of the photos here were taken during the first casting) I decided to make things a little easier by doing some extra filling of the nose, ears etc. before applying the regular matting. I mixed a fairly thick batch of resin plus Fillite, filled these hollows, and found that if I layered some fibreglass tissue over them I could do all at once without the mixture slumping.
It’s best to trim down this edge within a few days of laminating, preferably the next day, because although the resin will seem quite hard it’s still at its so-called green stage and filing it down will be relatively easy. I found that the type of file shown here, for filing ceramic tiles, worked very well. The edge of the fibreglass needs to be filed down exactly level with(or a touch under) the silicone edge and better if the edge is made to slope inwards slightly. The black gaffer tape is to prevent debris collecting between the silicone mould and the jacket since these should not be separated until later once the parts are joined.
The next step, below, was to protect the jacket edge (on both halves of the mould) with a greasing of Vaseline. This was to prevent any resin from sticking to the Jesmonite surface. It would be impossible to remove if allowed to harden completely on it. The Vaseline shouldn’t come too near the fibreglass edge because Vaseline can inhibit the proper curing of the resin applied as a seal.
The two halves need a sealant applied to both edges which will glue them together when the two mould halves are re-assembled. This needs to be made the same colour as the surface coat because it will show. Straight polyester resin would be too runny to stay enough on the edge so for this I had to buy some thixotropic paste (the same resin in a stiff, gel-like form) designed to be mixed in with the resin to thicken it up. The resin can be progressively added to the paste (better this way than the other way round) until the desired thickness is achieved and then the mixture can be catalysed as per normal, before adding the pigment. Once even, I used a brush to dab it thickly onto the edges, but equally a small hole could be cut on the corner of a polythene bag and the mixture piped on.
As I mentioned before, I dispensed with creating the usual mould natches to locate the two mould halves together because I felt that the edges were contoured enough to lock together correctly. As a guide though I’d traced the edge line of one of the mould halves (which was slightly smaller) on the other half. I placed one on the other securely and the weight of the mould itself pressed them together. There was no need in this case to clamp the halves together.
I waited a day before demoulding, to be safe. Catalysing at 2% often means that it’s safe to demould within 2 hours but full curing and hardness isn’t usually reached before 3 days/one week. I didn’t want to wait this long because I wanted to take advantage of the green stage for trimming and cleaning up. The correct procedure is to remove the rigid jackets first before peeling away the silicone parts of the mould. In any case this is the only possible way when casting polyester resin into silicone because the resin has a tight ‘grab’ on the surface.
Also, care needs to be taken to ease the silicone gradually from protruding parts such as ears. Whereas, given the strength of fibreglass and resin, it’s unlikely that the ears would break it’s also a precaution against damaging the mould. The seam line is marked by a thin, ragged wall where the surplus from the sealing resin was squeezed outwards.
I was surprised by the number of small imperfections, by which I mean the small, raised grains which peppered the surface. These were caused by minute enclosures of air when applying the first thickened layer of silicone on the prototype surface, which the polyester then fills but which are largely invisible to the eye on the silicone mould itself. I thought at the time (when applying the silicone skin) that I should have used a little less thickener for this coat because it was getting sticky too soon, but it was too late to change it. As with many things .. I’ll do it differently next time! The surface needs to be sanded a little anyway and these blemishes are easy to get rid of. I don’t know what caused the scratch-like indentation on the cheek in the photo above. It wasn’t on the silicone and I can only guess that it might have been caused by a hard particle of something in the resin surface coat which left a trail while brushing. Again, this is a simple thing to repair.
Before cleaning up the surface there was one last and very important step in the casting process; to strengthen the seam inside. So far the two halves are only being held together by a thin edge of resin with no fibreglass reinforcement crossing over and with an almost enclosed form like this it is not possible to apply any to the inside. It is at least possible however to pour resin into the hollow form to run along the seam (from the small opening left at the base). I mixed up some further batches in succession (adding 4 teaspoonfulls of Fillite this time to give some substance without affecting strength) and poured the first straight down to collect in the head area, rocking gently to and fro along the seam axis for a while. I poured the other two, each in turn, from the base opening to run down the side seams. As a precaution I clad the seam on the outside with plasticine, shown below, just in case there were any gaps in the sealing I couldn’t see. This was also the reason why I’d left the ragged seam edge as it was for the moment. For this technique it’s quite a good idea to leave a small amount of resin each time in the mixing beaker. When the resin in the beaker gels one knows it’s safe to stop rocking the mould.
So that’s it, basically. Below is the finished cast. Imperfections can be filed or sanded off easily, especially if this is done within a few days before the resin achieves full hardness. There will however be a Part 3, which deals with putting the fixing in place so that the sculpture can be mounted securely on a block.