Making a Menger sponge in stained glass

stained glass Menger sponge viewed in overcast natural lighting

Over the course of six weeks in January and February 2019, I took an introduction to stained glass course from Olivia Vavroch at Vavroch Glass Studio in north Portland, OR.

I took the course out of a general curiosity about a medium I was basically wholly unfamiliar with; I expected to learn a little bit and to come away with something reasonably interesting to look at. In fact, I learned a great deal, and ended up making a piece of glass work that I'm absolutely in love with.

The process was interesting to me every step of the way; I'll provide here a walkthrough of the whole thing with photos I grabbed throughout.

small scraps of glass, with cutting tools, on a work table

Our first class session was spent getting familiar with the tools for glass scoring and breaking; here's a bunch of strips of glass, and smaller bits, that I made with the glass cutter and running pliers on the right.

The fundamental glass breaking process is using the glass cutter to create a small score along the desired break line and then using running pliers (or in some cases breaking pliers) to apply pressure to either side of the score to snap the glass into two halves along the newly-created weak fissure. A bit more on this later! graph paper sketches of three stained glass Menger sponge designs

Homework in the run up to session two: produce a couple of possible designs for a project. I latched onto a Menger sponge because, well, me, and was pleased to find that this fairly simple treatment met up with Olivia's suggestion of about 50 pieces total. This is 48, with 10 pieces for each of the three colored faces of the sponge and 18 pieces in the gradient around the perimeter. The colored version up top became my basic design blueprint, with a few lines/intersections refined in the uncolored sketch below it carried on into my actual working blueprint. The other one, bottom right, just didn't feel as interesting.

There are a few tricky bits with glass shapes this design took into account; the main thing is that sharp interior corners are a no-go because of the way glass breaks. It's very hard to get a sharp inside corner without setting yourself up for a piece of glass that is at some point abruptly *crack!* two pieces of glass instead!

So I broke up the three main square-donut faces of the Menger sponge form into eight separate pieces each, so that the internal pairs of triangles at the center wouldn't be meeting up at pieces with interior corners. I also arranged the points where the outer pieces meet up with the outer edge of the central sponge figure so as to avoid everything meeting at four way intersections, which wasn't strictly necessary but seemed like it might be a little more structurally sound, so e.g. where 31 and 32 in the top left meet the upper left edge of the central sponge figure it's in the middle of piece 2, rather than right at the intersection between pieces 1 and 2.

a cardstock blueprint of a stained glass design

The last bit of homework was taking that sketch and transferring it to a full-size layout to bring in and discuss. I scaled my piece up to 15 by 15 inches, worked out in pencil on a large sheet of sketchbook paper.

We then spent the start of session two discussing the designs and whether and how to tweak them. Mine was workable as is, thanks to the design considerations Olivia had laid out in the first session, so I got to work transferring it to the sturdier cardstock pictured above, using a simple graphite-from-a-number-2-pencil carbon transfer tracing process, and then went over the new tracing in thick sharpie marker and numbered the pieces.

The numbering becomes useful for organization purposes later on when you're working with dozens of small similar-ish bits of glass. Duplicating the pattern instead of just doing the original on cardstock is important so you still have a 1:1 working blueprint after cutting up the cardstock!

special scissors for cutting out a channel of material from the cardstock

Next I pieced out the cardstock design into individual elements, using a special pair of scissors that cut twice, about 1/16th of an inch apart, removing a small channel of material from the cut in the process. This leaves a little wiggle room between pieces, which is important later!

Olivia's take on these scissors is: they work but they're annoying to use! You can put a couple of razor blades together with a small gap instead, which is apparently less fiddly and annoying to work with. My take on these is: I didn't know better and they did the job, so, hey, good enough! I'll try the razors on my next piece.

cut up pieces of card stock

Here's the whole design, in 48 separate numbered pieces of cardstock.

several chunks of glass on a light table, greens and blues and purples

The other thing we did in session two: start picking out glass for the project. I wanted a cool (as in color temperature, but also as in rad I guess) gradient for the outer perimeter, to complement the warm reds and oranges and yellows of the Menger sponge itself, and so I ended up looking through a whole lot of glass to try to find appropriate subtle color steps and interesting textures.

And holy crap, this glass! All this stunning glass! The biggest single surprise to me in this class was just how much character individual chunks of glass could have; the Vavroch Studios supply, in large and small sheets and workable scraps, is a small treasure trove, every sheet and scrap a bit different, with texture and opacity and mixed colors and varying thicknesses and surface treatments. You could lose a lot of time just fussing with colors and this-ing and that-ing the possibilities. If we hadn't had a class schedule to keep to I might have spent a couple more hours just mixing and matching.

various glass breaking tools and pieces of glass at a work table

Session three: time to start cutting our pieces. This took up the bulk of sessions three through five, in fact; the process of cutting, grinding, and finishing any given piece of glass is reasonably quick, probably ten minutes or less when you get into the groove of things, but when there's several dozen pieces total that adds up!

I'll talk about the process a bit more below, but here you can see what my glass-breaking workspace looked like most of the time: running pliers, breaking pliers, and above them on the ruler a pair of grozing pliers; glass cutter; a plastic waffle board to catch stray glass and a wooden cutting board that I preferred as my actual cutting surface; and then a bunch of cardstock piece patterns and the various chunks of glass I was cutting pieces out of, labeled in marker for reference.

small shards of orange glass scraps in a metal bucket

Getting your pieces out of a larger piece of glass is a subtractive process; ideally you can take the general outline of the piece off of the source glass while leaving as large a chunk of intact glass behind as possible for future use, but you're also gonna end up hucking off a bunch of little scraps. These shards range from teeny to maybe 2 inches on the long side for some of these triangles. The larger scraps could in theory be used for small pieces in another work that requires some fine details.

a triangular chunk of glass sitting on the bed of a rotary glass grinder

Once a piece has been cut to size, it goes to the grinder for shaping and edge texture; if your edges were exactly right from the first break (as the few straight-edged equilateral triangle pieces from my interior more or less were), you'd still run the piece briefly against the grinder edge just to take the smooth broken-glass sharpness off, both for safety and to give the adhesive copper foil we use later a better surface to attach to.

several shaped pieces of orange glass sitting on a full design blueprint

Here's my progress, eight pieces in; I've cut and grozed my forms but not yet ground them. You can see the black outline around each piece, traced with marker around the perimeter of the cardstock pieces I'd cut out earlier, with an eighth to a sixteenth of an inch of extra glass jutting out past the perimeter in most cases.

Which is too much extra glass! This was one of my rookie mistakes, which Olivia steered me right on shortly thereafter, saving me a bunch of work in the long run. There were a couple things going on here:

1. I was worried about breaking my glass incorrectly on curves. Glass breaks along a straight line really well in general; curves are trickier because you're breaking the glass by creating tension on both sides of a manufactured imperfection (the scored line created with the glass cutter), and curves can introduce a whammy factory where the glass might break nicely all along your curve but might decide to break in a straight line tangent to any point along that curve, leaving you with a piece that's missing a, uh, piece. Longer and steeper curves are a bigger risk for this; my curves mostly weren't really that dangerous.

2. I was intending to just fix it all up in the grinding stage; so what if I left a bunch of extra on? I'll grind it off! Which, technically, is a legitimate way to go, but grinding a lot of material takes longer and grinding a lot of material 48 times would take a *lot* longer.

So I had cut my pieces as rough outlines with safe straight lines, used grozing pliers (which take off little chunks of glass a millimeter or so at a time) to manually pare down the shapes to within an eighth of an inch or so of my outlines, and then planned to grind the rest away.

Olivia's take: the curves I had would mostly break fine, and by running a score line right along those curved edges and breaking them there I'd save myself a ton of time and hand-cramping effort (and potential grozing and grinding errors, which could ruin a piece just as much!) and make better, quicker progress.

Worst case scenario, a piece breaks wrong now and then. Time saved not wasting a lot of unnecessary labor on each piece would more than make up for the extra few minutes marking the piece's template onto another piece of glass and cutting again. This absolutely bore fruit and my working pace (and confidence with scoring and breaking the glass) picked up significantly afterward. In the end I think I had exactly one of my 48 pieces break in an unlucky direction, and I had another one cut within a couple minutes.

red, orange, and yellow glass pieces laid out in a Menger sponge image on the blueprint

Progress as of some point in the middle of session four. I have all the faces and interior triangular details of the Menger sponge design roughed out, here, and it's now clearer why having that numbered blueprint to work with is so useful. This will remain my main assembly working surface for the rest of the project.

Compare the orange pieces here to how they looked in the previous photo: I've given them time on the grinder and taken all the excess material off, smoothed up the edges, and the pieces sit fairly comfortably on the blueprint. The yellow pieces on the top face have not yet been ground, but they were cut more or less on the (sometimes curved) lines of the actual intended pieces and so you can see there's much less extra material outside the marker lines and what there is is basically smooth already.

the sponge glass pieces surrounded by a partial gradient of purples, blues, and greens

Progress as of the end of session four; I'd done my initial cutting and breaking of about half the outer gradient as well. Note the binder clip on the collection of cardstock pieces in the lower right: this is a handy way to keep a bunch of small bits of paper together. Olivia keeps a whole jar of the things on the work table.

a rough shard of glass and its cardstock pattern

A closer look at grinding before and after: here's a piece that has been broken but not ground, with its cardstock template held in place over it. You can see bits of glass sticking out; my cut and break lines for this piece were actually pretty good, right at the intended perimeter, but sometimes the glass will break a little unevenly across the width of the glass and leave little jaggy bits like you can see here.

the previous piece of glass after being ground into shape to match the pattern

After grinding, those jaggy bits are gone, and I have a nice flat edge hiding just behind my template piece. Perfect, or at least totally sufficient which given the margins of error and wiggle room in the final assembly process is more or less the same thing.

a broken piece of yellow glass with the number 8 written on it

Then again, sometimes you grind a piece so it's juuuust right, and then you set it down on the edge of a counter and miss and it lands on the floor and, hey! Time to cut another copy.

outline of the pattern for number 8 on a larger piece of yellow glass

So: grab another piece of glass (I had cut the original from a larger piece of scrap so I didn't have to go hunting for a replacement sheet) and outline the card stock piece again. If there's a flat edge on your piece, and a flat edge on the glass, great: line 'em up and save yourself a break!

close-up view of a glass cutting tool

Close-up of the glass cutting tool; that tiny wheel at the tip does all the work, the rest is just a housing for it to spin its sharp little edge along as you press it down against the glass and move across your intended break line. I'd dip the cutter tip in oil periodically while working to help it make a clean cut on the glass without getting snagged from friction.

large piece of yellow glass freshly cracked in two along a straight line

Here I'm using the running pliers to break off the initial chunk of glass from the larger sheet, along the top edge of the final piece.

The running pliers have a curved mouth, like matching frowns above and below, vs. the normal flat mouth of typical pliers. After you've scored your glass, you line up the mouth of the running pliers with the score line (you can see the center mark on the mouth of the pliers below, lined up with that break line through the glass) and then squeeze. The curve of the running pliers ends up applying downward pressure on both sides of the score line, inducing the glass to break where the pliers' mouth is, and that break will then tend to run (hey, get it?) along the score line to the far edge of the glass.

a yellow number 8 partly separated from a yellow sheet of glass

And thus! A small chunk of glass, with two straight edges already done, and a larger piece of leftover glass someone else can work with in the future.

a roughly complete replacement piece number 8

A couple more curvy score lines, and couple fairly clean breaks, and piece number 8 version 2.0 is ready for the grinder. You can see that little bit of jutting glass in the top right; that's a more prominent messy break than most pieces get, and it's two seconds on the grinder to get rid of.

the whole design laid out in roughly-shaped glass on the design blueprint

Milestone, somewhere during session five: all my pieces are cut and ground! The design is looking good and all the pieces fit together reasonably well.

closeup of some of the above rough layout

The fit between pieces isn't perfect; you can see that there's slight unevenness and gaps from one to the next, and not all of my curves match up flush from one piece to the next. That's okay! We'll get up to more stuff with copper foil and solder that'll, literally, fill in the gaps. (You can also see a whole lot of grinding crap accumulated on the surfaces of the glass. Stopping and washing stuff turns out to be a recurring bit of this process.)

a piece of dark turqouise glass, ready to be foiled

With the pieces all assembled, it's time to start foiling! This is the typical process for smaller stained glass works; there's also a lead channel process, which doesn't allow for as much fine detail work but is used for larger pieces (think picture windows and such) because it's much more structurally sound when the work has a lot of mass.

The copper foil process starts with a clean, properly ground piece of glass, like this one.

the rough edge of that turquoise piece

You can see how rough the edge is; the original broken edge was, well, glass-smooth, which would give the adhesive on the copper foil not much to cling to and would also tend to razor right through the copper at the top and bottom edges of the glass. Grinding solves both problems!

a close view of the copper foil, with a paper backing partly peeled away from the adhesive side

The copper foil is indeed just a thin thin film of copper with an adhesive attached to one side. The adhesive goes on the glass edge, and copper ends up facing outward.

copper foil wrapped partway around the edge of the turqouise piece

The foil goes around the whole perimeter of the glass piece, creating a circuit that clings to itself (which helps prevent it from peeling off later under the stress of gravity), and a small bit of foil overlaps out on to the horizontal faces of the glass piece, about a sixteenth of an inch usually.

copper foil wrapped around the whole piece

The foiled piece, with a perimeter of overlap around the face of the glass.

close up of a plastic bone-folder tool, a blunt wide knife shape

You have to burnish the foil onto the glass, not just casually tape it down; you work over the surface back and forth a bit with a tool to make sure the adhesive attaches well to the ground edge of the glass, and gets worked down into the sometimes fairly bumpy-textured details of the faces of the glass piece.

I mostly used this plastic bone folder for my burnishing, as I was able to manage it fairly quickly and consistently.

a rectangular razor blade with tape around the non-sharp side

But you can also burnish with a razor blade, which has the advantage of being more flexible than the bone folder which with practice allows quicker and more precise work. Olivia prefers the razor method, but warned us that you have to be sure not to accidentally cut your foil while you burnish.

a small bleeding cut on my index finger because someone let use a razor blade

You should also avoid cutting yourself.

several pieces of cut glass, some windex, and a cloth

Several pieces waiting to be foiled. Before applying adhesive copper foil, you also want to clean each piece well to get all the accumulated grit and such off, all the leavings of the grinding process. You want the foil sticking to the glass itself, not sticking to particulate resting temporarily on that surface.

several pieces of cut glass with copper foil attached

Several pieces succesfully foiled.

the whole stained glass design with foil on all the pieces, laid out on the design blueprint

Milestone: the entire work, done up in copper foiling and laid out on my working blueprint. I started foiling at the tail end of session five and got feedback from Olivia that it was going correctly, and then took everything home and did the bulk of it in the week interim before the final class.

close up of a few foiled pieces

The foils looks really nice, and is also going to disappear entirely! It's a bit of a shame, really. But we're not using it for its looks, we're using it for its solder-friendly chemistry.

foiled glass with small tacks of solder at intersections

And so: session six, final class, time to put the whole thing together. Soldering is the big step here, and something I'd never done before, so I was a bit nervous going in.

Solder won't stick to glass. Solder will stick just fine to copper. Copper with adhesive will stick to glass. You see where this is going. The assembled piece will have solder running along all of the edges where two copper-foiled pieces of glass meet, and the copper attached to the glass pieces will hold everything in place.

close up of some solder tack points

First step: tacking everything in place. I made small solder points at various intersections to make sure the whole pattern was fixed roughly in place and wouldn't move around on me later as I did a more thorough soldering job.

We used 60/40 lead/tin solder, at I think about 410 degrees on the soldering iron (not pictured at all because it's a two handed job); the soldering process turned out to be surprisingly straightfoward, applying a little bit of liquid flux (an acid that helps the solder flow well) to the copper at each working area, laying the hot head of the iron on that spot, and then pressing a bit of solid soldering wire against the head to melt it. The bead of now-molten solder then flows right around the head of the iron and onto the spot where it's resting, creating a small bead/pool of metal material that will then immediately cool and solidify once the iron is pulled away.

stained glass design full soldered

After tacking, I came back with more flux and more solder and ran a bead of molten solder along the length of every edge of the design, creating a solid web of connective material. Where there were gaps between foiled edges of pieces I hadn't matched curves on perfectly while grinding, the solder melted down between to fill up the empty space.

I took to soldering quickly enough; I will need more practice to get anything like artful at it, but the basic process was less fiddly than I expected. Getting a relatively even line of material wasn't too difficult; probably my biggest challenges were pacing the rate of solder for the pieces where I needed to actually fill a gap between two ill-fitting curves rather than just lay solder on top of a nice flush joint, and remembering to pull the solder wire away first rather than the hot iron away first.

(If you pull away the iron instead of the solder, the solder wire you're melting immediately sets up and now you've just got a spool of wire attached to your window! Get that iron back in there and melt it off again, and hope you didn't leave a big mess of extra solder pooling up.)

close up of soldered piece, with spattery bits

Close-up of some solder lines. Not terrible, not perfect; good enough for a first go! You can see a lot of little beads of spattered solder here, too; I was probably using a little too much flux a lot of the time, and that can sputter and spit a little when it meets molten metal, and leave these tiny lead/tin pyroclasts laying around. Fortunately, since solder doesn't stick to glass at all, it cleans right off!

cleaning the soldered piece with a round scrub brush

And so we clean, after soldering one side and flipping it over and soldering the other. Flux is necessary for soldering but it's also a corrosive acid, so getting it all off after is important to avoid ending up with ugly etching on your glass surfaces. I scrubbed both sides of the piece with a mix of ammonia and soap, using this large scrub brush and then a toothbrush, really working all the crannies.

scrubbing the piece with steel wool

After that, I scrubbed all of the solder work with steel wool. Solder will react with oxygen over a fairly short period of time, building up a patina of oxidized material that (a) looks pretty ugly in the long run and (b) interferes with a good surface treatment in the short run even when it doesn't look ugly yet.

Even the couple of hours that passed between when I started soldering and when I finished up and got to this cleaning stage is already long enough for some of that oxidized build-up to have accrued. So, time to scrub!

a bottle of blue liquid labeled Copper Patina, and a small jar with a brush and some liquid in it

Now that I've gotten rid of the natural patina on the solder, it's time to add a decorative one to take its place. There are multiple chemical patinas you can use on solder; Olivia provided a gunmetal grey option and a copper option. I liked the look of the foiled pieces so much that I decided to go with copper.

close up view of soldered piece with copper patina applied

This stuff goes on like magic; you brush it on and the metal immediately fades from neutral grey solder color to a copper sheen. After a lot of slow, deliberate stages to this glass working process, having a step that happens more or less instantly is a nice little treat.

After brushing the patina on, I then worked it over gently but thoroughly with a cotton rag, to make sure the patina reached every little surface of the exposed solder so my copper finish will be nice and consistent and no exposed bit of solder will get grimy from natural oxidation down the line.

view of the whole stained glass piece with lead channel frame and copper patina on all the soldered lines

And so: the assembled piece, nearly done. This is the back face, incidentally; stained glass in principle goes both ways, though because some kinds of glass have actual depth to their coloration you can have inconsistent surface appearances on either side of a piece of glass. Here you can see some milky white elements in some of the orange pieces in particular, which I don't dislike but preferred to group on the back side of the piece to keep the colors a little more solid on the front.

During the soldering process I also added a frame of lead channel (or rather, for efficiency reasons, Olivia did while talking me through it) around the perimeter of the piece. The lead channel is sturdy, which gives the whole thing some extra stuctural integrity; it's a flexible material that Olivia and I stretched out straight after which she mitered the ends of the four 15" long pieces to make nice corners.

I then tacked solder onto the four corners, and ran solder up the last little bit along those outer edge pieces to the lead frame so everything would be well-attached. Olivia also walked me through soldering rings at the top left and right of the frame, to have something to securely hang the piece from.

a bottle of milky off-white liquid labeled Glass Polish

Nothing left now but, literally, polish. I poured/shook on this glass polish slurry and gave the whole thing a going-over with another cloth to get a nice coat of wax on everything.

close up of some green glass with a waxy cloudy build up on it

Give it a minute, let the wax set up.

rubbing the glass with a cloth

And then buff!

that same close up of green glass, all clear and shiny now

And hey, nice shiny glass!

the completed, polished project in overhead lighting, showing a variety of colors and textures and with specular lighting bouncing off the surface of the glass in places.

And that's about 20 hours of work done, creating a piece of art in a new medium that has left me absolutely charmed, well beyond my expectations.

the completed project viewed through a light table, casting light through the assembled glass; the colors are very different than the previous picture, with a constant luminance and a differnet character to the color and texture of the various chunks of glass.

And putting light through it is a whole final revelation, a huge payoff to all the work. This is a view through a light table in Olivia's studio, taken right at the end of the final class session. It was a rainy Portland late afternoon in early February, and the natural light was already fairly gone, but even this artificial lighting takes the piece from a nice looking arrangement of glass and solder in the photo above to...this.

I really hadn't known what I'd get from this class. I'm floored again and again by how striking this material looks, how much nuance and character and organic variation glass specimens can have, how visually rich the juxtaposition of these colors and textures can be.

And, in the end, I'm struck by how approachable the medium was. I had the advantage of having access to a well-stocked workshop and Olivia's excellent instruction in a small class setting, which made the whole process painless (aside from stabbing myself with that razor), but I've come away from these six weeks feeling absolutely confident that I could do this on my own. The class has demystified what seemed from a distance almost arcane.

I may end up setting up a small studio space for this stuff in my basement; the core toolset is small (glass cutter, a few pliers, a soldering kit, a grinder, basic eye and hand protection, some miscellaneous chemicals and cleaning tools, a work sink, and a table to cover in shards of glass) and relatively inexpensive.

But personal studio aspirations aside: if you're at all interested in working with glass, I heartily recommend finding a local course or connecting with a local stained glass artist. If you're in the Portland area, check out Vavroch. You may be as surprised as I was by how doable and how satisfying it is.

Josh Millard, Feb. 12 2019

(my art site; my twitter; my instagram)