moldboxMy current color dilemma has me making more and more color samples for my glasswork. As I mentioned in my reactivities post I do color samples in two ways. For regular kilnwork I make shallow clear glass trays with superglue, fill the tray cells with measured frit quantities, and wind up with color sample charts that are immensely useful when figuring out color-matching and what reacts with which.

The set I’m working on now, though, are for pate de verre, and they look at depth and texture as well as color. I’m really getting into layering Bullseye powders and frits into a mold to obtain subtle glows and textures. Putting a dark red transparent behind a half inch of BE translucent white gives a different effect than putting it behind a half-and-half mix of BE crystal clear and BE light peach cream. (In fact, I’m finding the frit tint colors are much purer using BE crystal clear than BE clear.)

To do that my samples need to match the depth they’ll be in the final piece and (preferably) be cooking against the same type of investment. These are bas-relief pieces that max about an inch thick and they’re curved, not flat. So I’ve built some sample molds that I’ll slip into the kiln alongside the pate de verre to fire on the same schedule.

moldboxpouringSo…here’s how they’re made. I start with ice cube and baking trays, and fill them with silicon to make permanent “negative” models. In the picture at the top of the page I’ve surrounded an ice cube tray with an adjustable investment frame made from $5 worth of scrap acrylic. It consists of four rectangles of quarter-inch acrylic with two thicker pieces glued on in an “L” shape. The Ls allow me to quickly slide the 4 pieces together to make a correctly sized box for my mold, then use spring clamps to hold them together.

This is from a mold box design used by glass artist Linda Ethier; I’ve just translated it to acrylic to allow me to see what I’m doing.

I’ll make my reusable sample model with TAP Plastic’s RTV silicon mix–it’s a two-part mix that’s messy, messy, messy (wear gloves). I mix 10 parts (by weight) of this really sticky, goopy resin with one part catalyst, mix well (really well) until it’s of uniform color and then pour into the mold.

moldboxpullingsiliconIf I’m doing this with a positive that requires exact duplication I’ll deaerate the mix–bubbles can cause all kinds of distortion–by poking pinholes in the bottom of the mixing bowl and letting the silicon trickle into the mold. (In this case I don’t care about bubbles so I just poured (and as you’ll see, I got bubbles!)

I pour the mix until it fills the ice cube tray and overflows a bit–the cured silicon is easy to trim later. When it cures in 5 hours, I have a flexible, reusable model that allows me to invest molds as I need and remove the model without damaging anything.

See how easily it comes out? And how flexible it is?

moldboxsiliconNow that I have the model, I notch one edge (which helps me identify samples later on when they come out of the kiln. Then I make a 5-sided box of roofing tar paper (bottom and sides) to contain the model for investing. Most casters make a simple cylinder of the tar paper and dam the bottom edges with clay to contain leaks, but I’ve had more success slitting and folding the tar paper underneath the mold. I never, ever have a leak and it’s a lot faster to construct (and wastes less clay).

I’m using the acrylic mold frame again because the tar paper is flexible and can bow under the weight of the investment–the acrylic keeps the mold nice and even all around, preventing hot and cold spots when it fires.
Here’s what the finished molds look like. I use the same investment I’ll be using with the final piece–it makes a difference–and in this case that’s R&R910, a very nice sometimes reusable mix that takes nice detail and leaves a good satin finish unless it’s messed with.

Do much more than lightly brush the surface and I find it leaves all kinds of mold mix sand lightly embedded in the surface, but that’s another story.

moldboxinvestedOnce I’ve made a few molds, I’ll fill them with the colored glass powders I’m testing for pate de verre.

BTW, what I’m doing is very much like what the Higuchis do for their incredible pate de verre and it’s what most pate de verre artists do, so please don’t get the idea I’m claiming this is anything special.

In fact I got the idea for using ice cube trays from watching the Higuchis’ MasterClass video from Corning. If you haven’t seen it, order it–it’s a nice overview of what they do although I wish it had a LOT more detail.

I weigh out a total of 25g of powder for each compartment, mix them thoroughly (I dump them into an empty frit jar and shake like crazy for about 45 seconds) and then really PACK them into a compartment. 25g is level with the top; if tightly packed the powder will shrink to about half the height after firing. I generally figure 50% shrinkage for any powderwork.

moldboxpowderfilledOnce I have a few of these made, I fire them to the same schedule I’m going to use with the real work. Since the amount of heatwork can affect the final color (and sometimes texture) of the glass, I prefer to duplicate the real conditions as much as possible in the test.

In this case I fired these with WindWoman. Although she’s not going to need as much annealing or cooldown time as Emerging1, the actual forming time (2 hours at 1550 degreesF) is the same, so this should be pretty accurate.

moldboxfusedIn the photo above, I’m trying different proportions of BE crystal clear and BE Light Peach Cream, and then looking at what happens when you mix 20% color with 80% of 50-50 cream and crystal mix, evenly mixed together. Here’s the mold just out of the kiln–you can see the loss in volume.

Here’s what the samples look like from this test.

moldboxcubesI’ve knocked off any “prickers” caused by the powder melting down the edges of the mold, but that’s about the only cleanup I’ve done. The finish is a nice satin except for some small pieces of sand embedded in the glass where the mold was probably disturbed. (You can’t mess with the surface at ALL). The colors look pretty true on my computer.

Note what the Light Peach Cream is doing in combination with BE Pimento Red, Marigold Yellow and Orange Opal. Bullseye’s reactivity charts used to note that Light Peach Cream contains selenium, which will react with glass containing copper or lead. They’ve gone to some nice dark earth colors, but absolutely NOT the colors you’d expect…

What interests me most in these samples are the honeyed tones that the Light Peach Cream and Crystal Clear take on. LPC is a very warm cream that doesn’t look all that golden to me in its pure form (the cube at top right), but when you start to dilute it it really goes yellow-brown. I’ll probably use much less LPC than I’d intended (only 20% or less), thanks to these tests. (In fact, I’m thinking I need to add some pink to this, maybe Erbium pink–sigh–more tests)