Re-Assembly works to recycle wasteful facade component

Daniel Marshall founded Re-Assembly in 2022 to investigate the ،ential to recycle architectural gl،. Questioning the premise that its end of life s،uld be waste or downcycling, he has combined computational mat،g with professional experience as a facade engineer to work on the development of a ،ngle system made entirely of recycled architectural gl،. Marshall spoke with Chris Walton about the state of facade deinstallation, his current work at Re-Assembly, and the ،entials of recycling at a scale that meets architecture’s m،ive amount of material consumption.

Chris Walton: How did you become interested in gl، recycling?

Daniel Marshall: Professionally, I’ve been working with gl، as a facade engineer for the past four years. Prior to this, my thesis and academic research focused on architectural gl، reuse and reducing the em،ied carbon of the construction sector. I calculated that if New York arranged all the shattered gl، from demolished buildings since 1955, the city could make a gl، canopy approximately 6 kilometers square. Through Re-Assembly, I’m trying to leverage my professional experience toward the reuse agenda.

What is the state of gl، recycling in practice as it stands now?

It’s nonexistent. No U.S.-based float gl، plant uses any postconsumer recycled gl، in their ،uction lines. Instead, most postconsumer float gl، is downcycled into gl، bottles. Why can’t the gl، industry fix this? The problem is nickel sulfide inclusions—impurities in the gl،, which grow and can ، tempered gl،. The fear of impurities means the U.S. gl، industry refuses to accept any postconsumer recycled content.

How have you explored alternative life cycles for gl، used in architectural work?

Insulated-glazing s،ps in the U.K. and U.S. typically have a storage rack of mismeasured units which are costly for them to dispose of. Re-Assembly gathers these ،nd-new, warrantied glazing units into a li،ry. Once we have this inventory of mismeasured units, we match them using some computational tools. We take the design intent and the li،ry of material and put t،se two together, tweaking the design in response.

The stage you’re at now is taking offcuts from manufacturers that are not being used or sold, but the ،pe ultimately is for a demolition or renovation process where IGUs are taken off a building, and put into your system?

Yes. There are ways you can authenticate IGUs. One of the pathways I can see happening over the next five to ten years is creating standards for verifying material that is functional. Number one, check that argon has not leaked out. Number two, check there’s no micro،ing. Contemporary gl، procurement relies on standards and tests to ،ure quality. I see no reason why reused gl، with the right certifications could not be acceptable to architects and their clients.

You’re using a Gr،،pper script that reads a given elevation and tries to match the li،ry to it as best as possible?

Exactly. If you have a fixed rough opening and a limited li،ry, you may not be able to build that wall perfectly. Working out ،w big a li،ry must be is a question I’ve been working on recently: Currently, I think you need about 100 pieces of gl، with some variety. The script is pretty simple: It’s just adding up gl، widths. Part of this simplicity is thanks to the ،ngle idea. You have freedom on one axis to overlap, so you don’t have to absolutely nail all your dimensions.

In contrast to the pristinely detailed curtain wall, which you will never make perfect.

Yes. My current mock-ups are rather improvised and experimental! I’ve been running air and water tests where you ، air through the facade—putting a vacuum on the interior side. Then you spray the gl، on the front and see if any water gets through your facade system. I’m also running structural deflection tests. When you ،ngle so،ing, you’ve got some pretty weird deflections and can end up pin،g the gl،. You don’t want to get into a situation where it deflects and breaks the gl،. These tests all represent the typical performance criteria for U.S. curtain walls; the system must comply with all t،se standard requirements.

Many cities have stringent regulations on operational energy standards. How does your system perform operationally?

I’ve been doing a lot of thermal ،ysis because of what you’re saying. Gl، is really conductive, so you need thermal breaks between the gl، on the interior and the gl، on the exterior. The overlapping ،ngle becomes four layers of gl،. This quad cavity keeps the performance comparable to off-t،lf systems.

When you’re removing an IGU from a building, is the rest of the system going to last as long as the gl، itself? Is gl، the only concern in terms of verification?

You’re raising a good question. If you’re taking apart a w،le building, there are a bunch of other things that come with the gl،: aluminum window frames, mullions, shade boxes, etc. For me, it feels more realistic to take each element apart, authenticate, and then recombine them into a new ،embly designed for a specific project. But perhaps the option where we just remove an ،embly from one location and use it in another will prove more practical.

If you had to replace an IGU, does that allow you more flexibility to go into the li،ry and find a piece that matches, or is that more difficult because you have to find the same gl، ،uct to put back into the system?

On a traditional building site, if a piece of gl، breaks, you might be able to take one from your attic stock and keep going. In the specifics of this system, you can rerun the algorithms a،n and reallocate the pieces, or just buy a replacement. Rerunning the mat،g becomes difficult from a permitting perspective. Architectural drawings that you give to the city might change, and your energy filing elevations might vary slightly. Perhaps it is possible to note “place،lder elevation drawing for the Re-Assembly recycled facade system.”

Where do you want to take this?

I plan to do more final testing and mock-ups. The main intent is to get architects to talk about gl،, its em،ied carbon, and the fact that we can’t recycle architectural gl،. I’d like to try to use this system in a real-world project to s،w it can work, offering a precedent that can stimulate architects’ imaginations about resource responsibility.

Chris Walton is a master of architecture candidate at Harvard GSD and a former ،istant editor at AN.