Sweat Notebook
Previous knowledge and experimentation
Undestanding my sweat : General info | Facts table | Sweat rate | pH mapping
Collecting my sweat: General info | Improvement
BTB (ph reactive) experimentation: General info | Final biomaterials | Final textiles
Crystals experimentation: General info | Salt extraction | CuSO₄ crystals
+ Experimentation
I started by mordanting a 100% cotton t-shirt using rock alum (10–15% of the fabric's weight in alum, dissolved in hot water). After soaking and drying, I applied BTB dye in select areas.
To compare mordant types, I tested three cotton samples: one untreated, one mordanted with alum, and one with copper sulfate (no precise measurements used here). Both mordanted samples developed tiny crystallites during soaking, which stiffened the fabric. The alum left the fabric white/transparent, while the copper sulfate turned it blue, and later greenish after BTB dyeing. This shift may enhance color contrast, as copper sulfate intensifies blue–green tones, making pH reactions potentially more visible.
I gathered multiple samples of biomaterials, fabrics, and t-shirts and applied two types of sweat:
Fresh sweat (low salinity): barely caused color shifts, with only slow and subtle changes in some biomaterials.
Old sweat (higher salinity): produced more noticeable changes in textiles. Mordanted areas (both alum and copper sulfate) seemed to block liquid absorption, delaying or reducing the visibility of color change. No major difference was observed between the two mordants in this test.
The most striking color shifts appeared when applying a calcium bicarbonate solution to the BTB-dyed fabric, producing intense blue stains contrasting sharply with the base orange. These blue marks remained stable over time on both mordanted and non-mordanted textiles. In contrast, sweat triggered a transition from orange to a dull greenish tone, but this change faded quickly: revealing the ongoing challenge of achieving lasting reactivity with sweat as a source.
To unify the base color across textile and biomaterial samples, three jars were prepared: one with vinegar (shifting to yellow), one with bicarbonate (shifting to blue), and one with copper sulfate (resulting in a light green tone). This experiment confirmed that BTB’s color changes are reversible and controllable, offering a method to “reset” the starting hue for consistent results.
When copper sulfate was introduced into the biomaterial, it caused a translucent greenish hue and slightly clarified the sample. Over time, unexpected fractal-like crystal patterns formed on the surface, giving the material an intricate texture, though this also made it more rigid. This reveals a dual potential: both chromatic and structural transformation.
To improve color contrast in sweat reactivity, three BTB dye solutions with different concentrations were tested on textile samples. Results showed that increased BTB concentration enhanced contrast, up to a threshold. The middle concentration produced the clearest and most consistent color shift, establishing an optimal ratio for future dyeing.
Tests revealed large differences in how fabrics absorbed sweat: some repelled it, others absorbed and evaporated it rapidly. The most dramatic color reaction was found in a cotton fabric mordanted with copper sulfate and dyed with a medium BTB concentration, where sweat turned green-blue upon contact. However, the stain didn’t anchor deeply into the textile, raising questions about lasting visibility.
I cooked a batch of the gelatin–agar–BTB biomaterial using sweat previously used in copper crystallization. The result was not ideal, high density and foaming from residual copper sulfate disrupted the texture, making it unsuitable for direct application.
Instead of using thread, I submerged a piece of t-shirt fabric directly into a copper sulfate solution. The outcome was very effective: crystals formed well and adhered strongly to the cotton. I also tried using a less saturated solution, which still formed smaller, well-distributed crystals without excessive buildup or unwanted precipitates. This opens up possibilities for controlling the crystal scale and density on the garment.
To explore odor fixing, two airtight jars were prepared: one with natural cloves, the other with clove essential oil. Inside each, a sweat sample and two cotton pads (one with water, one with sweat) were sealed for several days. Afterward, all cottons had absorbed the strong clove scent. There was no significant difference between water and sweat-soaked pads, but the sweat sample itself retained the modified scent longer, suggesting potential for odor transformation.
MDEF CHALLENGE
To move forward with a realistic and reproducible design, I decided to work directly with commercially available white T-shirts rather than unknown fabric swatches. This ensures that whatever works in my tests can be directly applied in the final garment. I selected four distinct T-shirts from Primark with varying fiber compositions to compare their responsiveness to BTB dye and sweat interaction.
1
Long sleeve basic tee
100% Cotton
2
Sleeveless sports tee
100% Polyester
3
Slim fit short sleeve tee (Ken)
56% Viscose, 40% Nylon, 4% Elastane
4
Basic short sleeve tee
52% Cotton, 40% Polyester, 8% Elastane
Each fabric sample was cut into smaller swatches and divided into three versions based on mordant treatment:
A = No mordant
B = Mordanted with alum
C = Mordanted with copper sulfate
Based on literature recommendations, I used 10% w/w mordant relative to the fabric mass. The total mass of samples for the alum batch was 17g, and 21g for the copper sulfate batch.
Alum (B)
17g
300 mL
1.7g
Copper Sulfate (C)
21g
300 mL
2.1g
All samples were soaked in hot water solutions and left to mordant before proceeding to dyeing.
All mordanted and non-mordanted swatches were immersed in a single BTB dye bath to ensure consistency. Each sample was labeled with its number and letter (e.g., 3C), and gently stirred during immersion to ensure homogeneous dye distribution. The samples soaked overnight (~12 hours). After dyeing, they exhibited varying shades of yellow to orange, indicating BTB binding, and were left to fully dry before reactivity testing
From each main dyed sample, I cut two additional sub-swatches:
One was submerged in water (pH 7,5)
The other in vinegar (pH 2)
Both were then dried and tested with sweat. The vinegar-soaked samples remained bright yellow and showed no visible color change upon contact with sweatl due to residual acidity (afeter drying were okey). However, the water-soaked samples returned to full reactivity faster, demonstrating that thorough drying is key to restoring pH sensitivity.
To evaluate BTB’s performance, I tested each sample against six different liquids of increasing pH. All pH levels were measured using paper strips.
Vinegar
2
Vinegar + Water
5
Alcohol
6
Tap Water
7.5
My Sweat (stored)
8.5
Sodium Bicarbonate Sol.
10
It’s worth noting that my sweat had a notably high pH, especially compared to the usual skin-level values (~5–7), likely due to protein-rich apocrine content or changes during storage.
Each sample had six small dots marked where one drop of each liquid would be applied using a syringe. This consistent layout allowed for accurate visual comparison and mapping of reactivity across fiber types and mordants.
Sweat
Caused green/blue color shift in most samples, but faded over time
Bicarbonate Solution
Created intense, lasting blue stains, especially in non-mordanted samples
Fabric #2
Most interesting visual shifts (faint white halos after drying)
Fabric #3
Least reactive overall
Mordant Comparison
No clear winner yet; both left minor residual color even after washing
The temporality of the stains was key, most liquids triggered short-lived reactions that faded quickly, especially on yellow-based fabrics. The first four liquids left no visible trace, while bicarbonate caused strong, lasting blue marks, confirming BTB’s stronger reactivity in basic conditions. Sweat did shift colors but left minimal residue.
Among fabrics, sample 2 stood out for its transition from white to blue with a subtle halo effect after drying, offering a more visually striking and poetic result than the flatter green transitions seen in samples 1 and 4.
As halos overlapped, inner zones turned white, revealing the original textile, making it temporary but expressive. In contrast, sample 4 stains appeared and vanished with no trace, creating a matte, flat transition that lacked aesthetic strength. Faint residue remained but lacked the elegance of sample 2’s vanishing marks.
When rinsed in plain water and left to dry, samples 1, 3, and 4 maintained their initial reactivity. Sample 1 uniquely showed patches of white upon drying, which could signal changes in BTB saturation or local pH shifts. Interestingly, a slow transition to green-yellow hues was observed across most samples as they aged.
Rinsing samples with diluted vinegar restored their yellow starting tone, demonstrating how the BTB remains chemically reversible. Once completely dry, these fabrics regained their pH responsiveness, confirming the need for dry conditions to re-activate after acid treatment.
In contrast, for sample 2, both water and vinegar rinses reduced its reactivity compared to its untreated state. Previously visible halo effects and vibrant reactions were no longer present, indicating that washing procedures might strip or neutralize BTB residues more effectively in this fabric.
Sample 4 showed slightly improved stain permanence with longer dye immersion but could not avoid eventual fading. In samples 3 and 4, overdosed sulfated mordanting without drying before dyeing caused significant loss of reactivity, either by oversaturation or blocking absorption pathways.
For sample 2, the new test with high-concentration copper sulfate followed by drying before dyeing is still under evaluation. However, preliminary signs are more promising, suggesting drying may prevent the over-mordanted loss seen in previous trials.
New samples dyed with very high BTB concentrations developed an unusual reddish tone compared to the usual yellows and greens. Prolonged immersion made the dyeing uneven, with patches of color forming due to pooling and absorption differences, limiting visual consistency.
When testing the reactivity of these reddish sample 2 swatches, the typical halo effect disappeared. Instead, sweat formed irregular greenish-brown blotches with little visual appeal, indicating that excessive dye concentration eliminates the soft, poetic response originally desired.
One sample 2 dyed with medium-high BTB concentration displayed volatile blue patches without halos but with an attractive, bright appearance. This suggests that lower dye intensity and shorter immersion times may provide the best visual results for this specific fabric, even if color permanence remains limited.
To expand material testing beyond cotton, I purchased three new sport T-shirts from Decathlon made mostly of polyester or polyamide. Polyester blends tend to show more interesting chromatic effects and dry differently, potentially offering stronger visual contrasts.
5
62% Polyester, 35% Polyamide, 3% Elastane
6
90% Polyester, 10% Elastane
7
100% Polyester
All four selected T-shirts (Samples 2, 4, 5, and 6) were mordanted with alum (alumbre de roca) using the recommended ratio based on fabric weight. This step prepares the fabric to better absorb the BTB dye. After mordanting, the shirts were left to dry in the sun.
For the dyeing, I prepared a large basin with BTB diluted first in alcohol, then topped with water. Each T-shirt was dyed one by one: fully submerged, gently stirred for five minutes, then wrung and hung to dry. As each shirt entered the bath, the solution gradually shifted from deep orange to yellow, because of the lower concentration of BTB.
All dyed shirts achieved the intended soft yellow-whitish tone. Sample 4, due to its cotton content, absorbed the dye more intensely and resulted in a slightly deeper yellow hue. Unlike previous experiments, this time the color was perfectly uniform across all pieces.
To create copper sulfate crystals only on the outer sleeve, I inserted a plastic bottle inside the sleeve and tied it in place. The sleeve was then dipped into a warm, saturated (but not over-saturated) CuSO₄ solution. A jar was placed inside the bottle to keep it submerged. This ensured that crystals formed only on the outer surface.
As an expressive element, I used a syringe to apply my signature biomaterial onto the front of one T-shirt, drawing the phrase Sweat Matters. While still a draft design, I plan to keep this title in at least one final piece. This experiment explores typography, tactility, and sweat-reactive textures.
To avoid biomaterial bleeding through to the back, I placed hard plastic trays inside the shirts to stretch and stabilize the fabric. I also tested smaller fabric swatches with biomaterial applied on top to evaluate reactivity and adhesion between the textile and the gel.
The first attempt to grow copper sulfate crystals on T-shirt 4 failed. Despite submerging only the sleeve, the liquid spread throughout the fabric, leaving most of the shirt wet. Crystallization appeared only as a crusty residue at the edge of the wet area. This might be due to insufficient saturation or absorption preventing local concentration.
Copper sulfate solution was directly applied to small swatches of T-shirt 2 to test localized crystallization. The result was a rough texture with tiny crystals and stiffened fabric. Reactivity to sweat was reduced, possibly because the fabric was already tinted blue, making pH-based color shifts less visible.
For the second test, both sleeves of T-shirt 5 were used: one submerged in copper sulfate, the other in BTB. Again, the liquid spread beyond the sleeve and into the body of the shirt, making it difficult to isolate crystal formation or control the dyed area.
The biomaterial adhered well to the fabric, though it caused slight shrinkage at the borders. Interestingly, while the biomaterial was uniformly green when applied, it shifted in color based on the pH of each fabric’s dye, reflecting strong reactivity. Acidic and basic drops caused visible, though slower, color changes compared to the fabric, creating layered chromatic effects.
After dyeing, T-shirt 5 showed very weak reactivity to pH and was discarded. T-shirts 2 and 6 showed promising initial results with similar contrast and visual responses to sweat.
Sweat was applied multiple times on T-shirts 2 and 6, with intervals for drying. Initial stains faded after 30 minutes, leaving a bluish halo and lighter center. Reapplying sweat revived the color reaction, especially around the edges.
The double-layer texture of T-shirt 6 created more dynamic visual effects. It retained contrast slightly better than T-shirt 2 over repeated tests, showing enhanced reactivity and a more vibrant aesthetic after sweat exposure.
To test stain removal, the sleeve of T-shirt 6 was submerged in a vinegar-water solution. The color lifted well, leaving behind a subtle yellow-orange halo. Future tests will evaluate whether reactivity to pH is restored after this process.
During the artifact presentation, I wore the final T-shirt (number 6), marked with “Sweat Matters” in biomaterial. Since I couldn’t sweat naturally, I injected my own sweat into the fabric using a syringe. The color shift was less striking than expected, likely due to over-saturation or forceful application. Smaller sweat drops in tests had shown better results. I’ll repeat tests by actually wearing the shirt and sweating naturally, possibly using a tighter-fitting version to increase fabric contact.
++ Experimentation
After drying, the BTB-dyed sleeve showed a dirty, brownish edge and patchy orange gradient near the shoulder. As seen before, oversaturation with BTB caused harsh, non-uniform colors. Over time, the dark tones faded slightly and became lighter greens and blues, but still not the desired clean effect.
On the other sleeve, copper sulfate formed a rough, fractal texture. Visually interesting, but it stiffened the fabric and, as expected, completely blocked BTB reactivity.
Between the three finalists, T-shirt 7 (100% polyester, with sleeves) emerged as the best. It reacted similarly to T-shirt 2 but had the correct shape. T-shirt 6 was discarded due to its dual-texture surface, which reduced visual contrast.
Different dye concentrations were applied to areas of T-shirt 7 to evaluate optimal reactivity. Higher BTB levels produced striking blue contrast. Lighter areas showed weaker response, confirming that strong initial staining is beneficial.
When testing pH 7–8 liquids across the dyed zones, the blue faded or failed to form in lower-concentration areas. Stronger BTB zones gave much more vivid responses, emphasizing the importance of embracing an orangey base tone for full contrast.
Acidic liquid erased previous stains perfectly. However, reactivity in those same spots was weaker afterward, especially if not fully dried, highlighting again the importance of drying before testing reactivity again.
The color response across the pH spectrum was incredible. I finally have a clean, vibrant canvas. Now the only open question is whether my real sweat will consistently be above pH 7 to trigger blue staining in real-time scenarios.
I also tried restoring the shirt by placing it directly into the BTB dye bath. Stains disappeared and reactivity increased, although faint orange tonal traces from past markings remained, adding visual texture.
Before entering the sauna, I ingested an oral rehydration solution containing: glucose (20 g), sodium chloride (3.5 g), potassium chloride (1.5 g), and trisodium citrate (2.9 g) in water. After sweating and collecting my sweat, I measured the pH and found it to be surprisingly acidic, around pH 5.
I repeated the exact same sauna routine, but this time without ingesting the electrolyte solution. The sweat pH came out higher, around neutral to slightly basic (~pH 7). Conclusion: the supplement likely acidifies the sweat. I also tracked both sweat samples over time, and both gradually shifted toward green on the pH strips, meaning sweat becomes more basic when stored, which is a problem since the T-shirt is meant to react in real-time.
I tested both sweat samples using a water quality meter to measure conductivity or other properties but didn’t find any meaningful differences or insights beyond pH.
I tested a new sample of polyester submerged in a more saturated copper sulfate solution. However, no real crystallization occurred on the fabric itself, the crystals only formed at the bottom of the jar. The textile became stiffer and rougher, but no solid crystals adhered to the surface.
I inspected stained samples under the microscope to identify weird accumulations. They were not bacterial clusters, as I initially suspected, but rather random fibers, dust, and non-living debris, gross, but not alive.
I observed the polyester fabric under the microscope. The weave appeared highly regular and synthetic. The sample with copper sulfate crystallization had a slightly blue tint, but overall texture was just more rigid.
I wore the over-dyed final T-shirt (sample 7) for a few hours. The result: zero visible change on the fabric, but my armpits turned yellow from the BTB bleeding into the skin, not the desired outcome :(
To reduce this skin-staining effect, I washed the T-shirt with tap water. It turned greenish due to the water’s pH, then slowly dried back to a dirty yellow tone. Some previous stains remained faintly visible.
While rinsing the shirt several times, I noticed the wash water turned increasingly blue, indicating rising pH levels. I then prepared two solutions: (A) tap water + vinegar, and (B) tap water + baking soda, to test if pH-altered rinsing would affect future reactivity.
The swatch washed in the baking soda solution (B) stopped reacting completely. In contrast, both the shirt washed multiple times with plain water and the swatch rinsed with vinegar (A) remained reactive, both to acids and bases.
I attempted to clean the shirt with pure alcohol to see if previous stains could be removed. It worked reasonably well, some staining faded, and reactivity was preserved. I also tried soaking it in vinegar-water for hours to reset its color and test future responsiveness.
The vinegar-soaked shirt became visibly lighter and cleaner, though yellow stains, especially around the armpits, remained. Meanwhile, T-shirts 7A and 7B were mordanted with different concentrations: 7A with a super-concentrated alum solution (10x recommended, later rinsed), and 7B with 0.5x the standard. 7A was dyed with a medium-low BTB concentration, and 7B with a medium-high one. 7B showed a deeper yellow tone after drying.
The vinegar-treated shirt showed a curious effect: sweat drops initially repelled but eventually burst and absorbed slowly, triggering mild reactions. Both 7A and 7B responded well to stored sweat, with 7B showing stronger blue and contrast due to its more intense dye.
Over time, the vivid blue faded on both shirts, resulting in similar soft marks. When applying acidic liquid, both shirts reacted similarly, fading to paler tones, leaving a whitish trace where sweat had been.
To test with fresh sweat, I exercised and applied it directly to the yellow T-shirts. As expected, there was no visible reaction. Measuring my sweat’s pH revealed a value below 5, too acidic to trigger a color change.
I partially dyed both shirts blue by soaking areas in a sodium bicarbonate bath. Applying fresh sweat on these blue zones showed a more visible effect, some drops intensified the blue while others erased it, turning the fabric yellowish again. Rubbing the fabric helped partially remove the blue as well.
To prepare for live trials, I fully soaked both 7A and 7B in a sodium bicarbonate solution, turning them completely blue. This setup will allow testing how my fresh, acidic sweat interacts with the alkaline-dyed fabric during real-time wear.
After soaking both shirts 7A and 7B in a bicarbonate solution, their intense blue tone quickly faded as they dried. I wore one to the gym and noticed improved reactivity compared to yellow-dyed versions. Blue halos appeared around the collar and underarms, but as sweat increased, the halos vanished, leaving the shirt even whiter. Although the dye didn’t transfer to my skin, it did rub off onto my workout towel, showing poor fixation.
The shirt that hadn’t been worn faded even more. After just a few days, its blue had almost completely disappeared, turning pale and dull, confirming the impermanence of this light blue tone.
To intensify the blue, I re-dyed both shirts. I first dissolved BTB in alcohol, added water, and adjusted the pH with stronger bases: one bath used washing soda (sodium carbonate, pH ~11), and the other a small amount of caustic soda (sodium hydroxide, pH ~13–14). Both are highly basic substances and should be handled with caution due to skin irritation risks.hh
As expected, the caustic soda shirt developed the most intense blue, even slightly purplish. After drying, both shirts retained stronger blue tones than in the previous bicarbonate trials. They didn’t fade to white as quickly, showing more durability.
I tested both with slightly acidic and basic liquids. The shirt dyed with sodium carbonate reacted similarly to previous halos, now slightly more vivid. The one dyed with caustic soda had a new behavior: fresh drops triggered sharp, intense blue stains immediately, working with both acid and base inputs.
However, over time, these sharp blue stains still faded. Unlike the previous halo-style fades, the caustic soda version didn’t leave a soft trace, but lost its intensity completely, raising questions about how these reactions age visually.
Visualising my sweat
I wore the caustic soda-dyed shirt (7A) during a full commute, from home to Sant Adrià del Besòs, walking, train changes, and a packed, stuffy Renfe. I started sweating mildly. Small blue dots appeared on the chest, marking early sweat points. In the armpits, distinct halos formed, four on one side and two larger ones on the other.
While the back was hard to track visually, wearing a backpack led to a noticeable sweat zone forming where the fabric pressed against my skin. The armpit halos grew, and some areas shifted to a greenish tone, likely from acid traces in fresh sweat. I ended the first test upon arrival and captured images of both the front and back to document the stain pattern.
This test proved very promising. Finally, I had visible, reactive sweat marks without major dye transfer to my skin, just minor blue traces under the arms and on the back. Interestingly, despite not being visibly stained during wear, my back later tinted another shirt blue, suggesting my skin had absorbed and later released the BTB.
For the second test, I wore the shirt dyed only with carbonates to a gym workout. The session included a hike uphill and strength training. I sweated heavily, and unlike the first shirt, I saw white dots forming on the chest, these grew into larger faded patches as the session continued.
Sweat running from my neck left a large white area at the collar. The back became quite pale from repeated contact with the towel during machine workouts, and the towel turned blue, again proving dye transfer remains a limitation.
I documented both shirts using my phone’s scanning mode, which enhanced color contrast. The images resembled heat maps, showing clearly how sweat emerged, spread, and shifted over time, a real cartography of perspiration.
I re-tested the fabric samples from the MDEF Challenge using the latest, deeper blue dye tone to check for any new contrasts. Despite variations across fiber types, the most visually striking and poetic results still came from the 100% polyester fabric.
After each wear, the T-shirts were rinsed with water to remove sweat stains. They released highly pigmented blue water, suggesting dye bleeding was still active. Once wrung and dried under sunlight, they returned to a clean white base, ready for re-dyeing.
To avoid the risks of caustic soda, I moved forward using only sodium carbonate (washing soda) as the base to shift BTB to blue. With a medium BTB concentration and 25g of sodium carbonate, the results yielded intense and even contrasts across the fabric.
Post-dyeing, curious patterns emerged. Some purple stains formed—possibly due to uneven dye deposits or chemical imbalances. During drying, the wettest areas retained strong blue tones, while drier areas faded to a duller white-blue.
Wearing one of the shirts during a cardio session (running and stair climbing in the sun), I noticed that most sweat came from my face and forehead. As it dripped down, it pooled around the collar of the shirt, forming a distinct sweat halo pattern at the neck.
The front showed clusters of blue drops around the neckline from facial sweat, while the back revealed scattered, isolated sweat marks, offering a subtle but expressive visual mapping.
I wore another dyed T-shirt to a crowded techno party at Pacha, Barcelona. The heat and movement created the perfect test environment.
By the end of the night, the shirt displayed complex staining: splashes across the front, likely from a mix of sweat and spilled drinks, and large yellow-toned regions, especially around the collar, where acidic sweat created a dramatic reaction. The back had expanded halo patterns showing heavier sweat accumulation.
To test performance setup, I used the heat cannon that will be part of my MDEFest exhibition. The goal was to mimic a mini-sauna effect that could trigger sweat in real time and demonstrate the reactivity of the shirt.
Despite the cold ambient temperature, the heat cannon did raise my body temperature slightly. However, it couldn’t replicate the full-body sweating effect of a real sauna. The shirt revealed only small patches of sweat on the back and a larger reactive zone on the front belly area, where I had been seated and sweat had pooled.
Last updated