# Digital Accessioning Wins, Fails and Dragon Tales: Part 2 The story continues with Part 2 -- read on for the exciting conclusion! May 11, 2026 Julianna Barrera-Gomez *In* [*Part 1 of this story*](https://preservation.library.harvard.edu/news/2026/05/digital-accessioning-wins-fails-and-dragon-tales-part-1)*, we left off with DAP (Digital Accessioning Program) having been burned by the dual-headed dragon known as Degralescence, and further whapped by it’s moldy tail, leaving degraded disks damaged and a new mold problem to face. Having quested for knowledge from sages, DAP looks for a plan to escape the dragon’s reach and rescue the data. Will they prevail?* ## **Exiting the Fuzz** Moldy disks certainly weren’t part of the project plan, but we were determined to learn as much as we could from this. With permission from our partners at Harvard University Archives (HUA), and an understanding that their project deadline for processing this larger hybrid collection was fast approaching, I set about to find a path forward out of this fuzzy predicament. I scrubbed off the dirty drive heads using isopropyl alcohol and updated our workflow to make sure heads were cleaned more regularly, to prevent damage from any accretions. Taking advice from colleagues in Media Preservation Services, we decided to make the little TEAC disk drive we’d been working with DAP’s designated drive for moldy suspects. That way we can keep other drives reserved for non-mold work, to help prevent cross project contamination. ![Photo of mold on a disk, with an inset image of what that mold looks like zoomed in with a pocket microscope, where viewers can see the webby-looking hyphae.](/sites/g/files/omnuum11066/files/styles/hwp_1_1__360x360_scale/public/2026-05/Mold3-5PocketMicroscope.png?itok=yqVbjp59) Mold growth on a 3.5-in disk, with a photo taken while using a pocket microscope overlaid in bottom left. (Photos by DAP)Then, we learned some more facts about mold and why this was happening: - Mold is a fungus - Fungi have “evolved a high degree of metabolic versatility” that lets them grow on lots of surfaces \[1\] - Fungi create hyphae (the webby-looking stuff humans can see) that secrete enzymes onto colonized surfaces, breaking down polymers in the process \[2\] - Fungi love water and will bloom (create hyphae) in high humidity environments, if the relative humidity (RH) is above 50% (remember that these disks had come from a high humidity location) \[3\] - When magnetic media breaks down in high humidity (binder degradation), chemical reactions occur that create water and other byproducts that are great for mold feasting \[4\] So we needed to prevent mold blooms and sequester the affected media. First, I double checked our environment: the DAP Lab isn’t strictly a collections storage environment, but we do enjoy a stable temperature and (most importantly) a stable RH that keeps below the threshold for mold reblooming. To be extra safe, I’d bagged the project disks in sealed plastic bags with silica gels, to keep these dry and separate while we planned to revisit the project later in the fall semester. Next, we needed a way to better identify mold—how could we tell if that brown splotch is mold or just grime? Colleagues in Harvard Library Preservation Services (HLPS) introduced DAP to the exciting world of pocket microscopes. These small, affordable devices made examining contaminants much easier, allowing us to sort and triage disks. When viewed with the pocket microscope’s LED light, mold is suddenly visible as the raised organic filaments (hyphae) interconnected across the disk surface. After sorting, I’d identified 38 of the 75 disks that contained mold growth, and I’d documented lots of other contaminants too (like dust, fibers, hair, and miscellaneous gunk). ![Image shows three examples of what contaminants look like in a digital pocket microscope, displaying mold growth, scratches in the magnetic binder, and dust/dirt/debris.](/sites/g/files/omnuum11066/files/styles/hwp_1_1__960x960_scale/public/2026-05/DigitalMicroscopeExamples.png?itok=M2VAkjLN) Examples of damage found on 3.5-in floppies using a digital screen microscope. (Photos by DAP)## **A Counter Potion is Brewed** Having become very familiar with what mold and contaminants look like up close, we needed a way to remove these rooted growths without scratching the delicate disk surface. Online forums used by computer enthusiasts had lots of shared cleaning stories, but unresolved debates about cleaning methods and how safe they were. A helpful preservation colleague, Dr. Leontien Talbloom, had kindly shared presentations from others in the cultural heritage field that discussed washing small batches of dirty disks with soapy water.\[5\] This method had apparently been successful (see [her blog](https://medium.com/digital-preservation-at-cambridge-university/fuzzy-logic-cleaning-floppy-disk-with-our-conservation-department-part-one-c2d3e0320a2c) for references), but it posed a problem for DAP: disassembling disks to wash in a water bath would take more time than we had, and we were also concerned that introducing water to both the unstable binder and the mold might cause more degradation and/or mold outbreaks. What we needed was a faster method that helped mitigate the water and drying risk. This is where a blog post from the University of Michigan Library, written by Christina Min and titled “[Mold is Weird: Part 1](https://blogs.lib.umich.edu/bits-and-pieces/mold-weird-part-1)” shined like a beacon of light. In the post, Min shares a story of collaborating with trusted conservation colleagues and testing multiple methods for cleaning mold found on 5.25-in floppy disks.\[6\] A cleaning method that worked for them: disassembling the disk and swabbing with a 70:30 solution of isopropyl alcohol and distilled water, which gently removes the mold growth and has the added benefit of drying very quickly (thanks to the alcohol). HLPS’s conservation scientist also confirmed this cleaning solution is a proven method for gently wiping away mold from other delicate binder materials such as film and photographs.\[7\] This looked like our needed potion to counter that fuzzy tail. Geared up with cleaning and protective supplies from conservation colleagues, and time booked at the fume hood, I decided to try it. ## **Testing for Potency** Having sorted through the moldy disks to find levels of contamination, I identified 7 of these that were covered in more than 50% mold growth. I figured these would be good candidates for opening and attempting to clean. Luckily, we’d found a decades-old guide from the Canadian Conservation Institute on media care and cleaning, where the author had demonstrated how to safely wedge open the plastic welds of a 3.5-in floppy disk for data recovery (I used a blunt metal spatula).\[8\] ![A montage view of the process of opening and cleaning an entire disk.](/sites/g/files/omnuum11066/files/styles/hwp_1_1__960x960_scale/public/2026-05/CleaningEntireDisk.png?itok=ksWEKmtx) Cleaning a disk with extensive mold by breaking open the disk welds and removing the inner magnetic media for fuller access while cleaning. Note the protective polyester sleeve has evidence of a liquid spill. (Photos by DAP)Opening the disk up, I saw a likely accelerant: the inner polyester fabric lining, which was meant to help clear dust and keep the magnetic media clean, had evidence of brown liquid stains on both sides. I flashed back to a day in high school yearbook, when a fellow student’s sugary drink had spilled on our table and had saturated a few floppy disks I’d been using to save my copy drafts. Back then, I’d quickly washed out the disks in the sink and left them hanging in the school’s dark room over the weekend, and they magically worked again the next week. I wondered how those disks would look today. The disk I opened had clearly lived a life, and as I opened up others, I mused at the materiality of floppy disks. These contaminants documented the dusty, dirty, sometimes splashy places the disks had been kept as they faithfully saved data encoded into tiny magnetic particles embedded into the coating on this delicate inner surface. It felt like the disks were owed a rescue attempt. ## **The Deadline Looms** So now we knew how to clean these, and to my delight I was able to extract data from the disks, even those with deep scratches. But given that the project deadline was fast approaching, how could I speed this up? Did we need to break open 31 more disks, if these had a smaller mold growth area? Turns out, of the 7 disks that were opened, only two had evidence of deeper soiling on the polyester sheets inside. The rest looked pretty normal, and even spot checking with the pocket microscope, it didn’t look like this fabric was any more dirty than other non-moldy disks I’d looked at. While I’d been able to put the opened disks into a sanitized reusable case to insert them in the reader, I wondered if that was totally necessary—as Min had documented in her work, mold spores on the dirty fabric lining may still create interference with reads, but using a “clean” case was no guarantee the disk would read without multiple tries. Low on time and facing this pile of disks, we made the decision to test a new “spot cleaning” method, in which I would hold open the disk shutter, spin the disk hub to check for small mold spots on the surface that the heads will touch, and then quickly swab these away using the alcohol-water solution. ![Montage view of cleaning a disk through the disk reader access port, and using a pocket microscope to view the cleaning process.](/sites/g/files/omnuum11066/files/styles/hwp_1_1__960x960_scale/public/2026-05/SpotCleaningDisk.png?itok=GJUAsNjo) Second method: Spot cleaning a moldy disk by cleaning through the shutter access hole using cotton swabs dipped in 70:30 alcohol and water solution, and manually spinning the disk to clean the surface that the disk reader touches. (Photos by DAP)This method was much faster, as it allowed me to quickly wipe mold away, confirm that it was cleaned off (with the pocket microscope) and it dried in seconds, without risking wetting the mold or the polyester protective sheet inside. Testing the first 5 disks, I was excited to get successful reads. Our predicament had a solution! I pushed on through the pile, made progress, and was able to get these previously inaccessible digital files to our HUA partners so that they could proceed with their archival processing goals. They considered the project a success, and so did we in DAP. ## **Mold Revisited: The Connections Between Us** Mold is indeed weird, but I’ve emerged with a newfound respect for this organism that’s found a way to secretly populate something as unappetizing as old computer media. **To help spread the word about moldy media and how to clean them, DAP created a guide on our wiki to share with our project partners:** [Identifying & cleaning media contaminants](https://harvardwiki.atlassian.net/wiki/x/I4AGM) I also acknowledge that DAP has benefited from the generous time and expertise of teammates in HLPS and other library conservation labs, colleagues at other universities, the shared advice of experienced computer enthusiast on online forums, and of course our HUA partners who had faith in DAP and let us test these cleaning methods on their collections. In a way, I kind of visualize these information-sharing connections like the interconnected hyphae I saw in my pocket microscope—spreading across voids, seeking and sharing information with the broader mycelial network. With this collective help, DAP broke through a literal mold barrier, allowing much of the magnetic data on this project’s disks to escape in one last read. While Degralescence will have the final victory with the physical disks, we salvaged quite a lot of the data—files, metadata and file fragments that archivists will be able to evaluate as they decide what is important to select for collection stewarding. No doubt some of these files will be arranged and described within their collections for wider research, and those digital collections may find their way into our [Digital Preservation Service](https://harvardwiki.atlassian.net/wiki/x/I4AGM), where new digital preservation foes will emerge. But those are stories for another day, as our battle-tested [Digital Preservation Services](https://harvardwiki.atlassian.net/wiki/spaces/digitalpreservation/overview) team stands vigilant and ready to ward foes off and continue to save the data. *Julianna Barrera-Gomez is the Digital Accessions Specialist heading up the centralized Digital Accessions Program (DAP) offered by Harvard Library’s Digital Preservation Services. When not learning the hard way how to clean and read disks, she endeavors to help colleagues safely steward collections to avoid many different digital preservation dangers.* ## Notes: \[1\] Wikipedia contributors. (2026, April 8). Fungus. In Wikipedia, The Free Encyclopedia. Retrieved 14:32, May 7, 2026, from [https://en.wikipedia.org/w/index.php?title=Fungus&oldid=1347754242](https://en.wikipedia.org/w/index.php?title=Fungus&oldid=1347754242). \[2\] Wikipedia contributors. (2026, February 23). Mycelium. In *Wikipedia, The Free Encyclopedia*. Retrieved 14:34, May 7, 2026, from [https://en.wikipedia.org/w/index.php?title=Mycelium&oldid=1340090403](https://en.wikipedia.org/w/index.php?title=Mycelium&oldid=1340090403). \[3\] “Managing Mold Contamination,” Harvard Library Preservation Services (HLPS), Accessed October 6, 2025, . \[4\] Bogart, John Van (1995) Magnetic Tape Storage and Handling: A Guide for Libraries and Archives . \[5\] Talbloom, Leontien (2025) “Fuzzy Logic: Cleaning Floppy Disk with our Conservation Department – Part One.” Medium: Digital Preservation at the University of Cambridge Libraries and Archives. Accessed September 29, 2025. \[6\] Min, Christina (2018) “Mold is Weird Part 1 & Part 2.” Bits and Pieces, Michigan Library Blogs. Accessed October 22, 2025. \[Online\]. Available: . \[7\] “PMG Mold Remediation - 6.3.1 Alcohols,” American Institute for Conservation (AIC) Wiki. Accessed: Mar. 06, 2026. \[Online\]. Available: [https://www.conservation-wiki.com/wiki/PMG\_Mold\_Remediation#Alcohols](https://www.conservation-wiki.com/wiki/PMG_Mold_Remediation#Alcohols). \[8\] Iraci, Joe. (2002) "Disaster Recovery of Modern Information Carriers: Compact Discs, Magnetic Tapes, and Magnetic Disks." CCI Technical Bulletin 25. . See also:- [ Digital Preservation ](/hlps-tags/digital-preservation) Share on:- [ Facebook ](#) - [ Twitter ](#) - [ Linkedin ](#)