Brewing R&D Specimen Report · No. 02 · May 2026

A Study of Iteration
Across Four Steeped Tea Formats

Methodology refinement traced through forty-eight documented iterations: temperature, ratio, steep duration, bloom protocol, filtration, cooldown curve, and water mineralization, each tuned in service of clarity and balance.

Format
Steeped · Cold
Specimens
4 distinct teas
Iterations
48 logged runs
Water Profiles
7 RO profiles
§ Abstract

Where the first specimen report dealt with matcha as object — eleven powders assayed against fixed protocol — this second report reverses the lens. Here the protocol itself is the specimen. Across roughly fifteen months at the bench and on the line, the team iterated through brewing variables for a small set of steeped tea bases: temperature, leaf ratio, steep duration, bloom geometry, filtration medium, cooldown curve, and the mineral profile of the brewing water.

The report logs each iteration in compact form, identifies the variable under adjustment, and records the resulting impression. Each of the four specimens converged on a locked production specification currently running on the cart. Vendor and origin identifiers have been redacted; tea-style descriptors are preserved as they are categorical, not proprietary.

§ I · Methodology

The methodology evolved.

Four cross-cutting axes carried across nearly every specimen. Their evolution forms the spine of the report: a moving frame against which each tea was tuned.

Water Program — Reverse Osmosis Remineralization

Values are mg of each compound added per 5-gallon jug of RO water (the working batch size at the bench), rounded to the nearest 50 mg; the current production specification (W₇) is described qualitatively only. Listed TDS is what the conductivity meter reads on the finished water; W₂ and W₃ share the same added-mineral recipe but sit ~10 mg/L apart because they were mixed on different RO source systems with different residual feed TDS. The late-stage shift from W₅ → W₇ is structural rather than incremental — silica is retired from the recipe and sodium bicarbonate is introduced as a second alkalinity source alongside potassium bicarbonate[1]. A side branch (W₂: silica added, bicarbonate lifted vs W₁) was carried in parallel for a small set of T03 trials before being retired in favor of the main production line.

i.
Filtration progression
Early runs used open mesh strainers, which permitted residual fine particulate to remain in suspension. Later runs adopted a two-stage protocol: mesh strainer for the bulk, then a fine cloth (nut-milk) filter before final cooldown. The cloth meaningfully reduces post-strain extraction without completely silencing it — a sought-after middle weight.
ii.
Cooldown curve
The standard sequence is a two-stage ice bath: first to ~110 °F to neutralize aggressive residual extraction, then a re-filter, then a cold ice bath to ~42 °F for service. Direct freezer chilling and pour-over-ice were both tested and both produced flatter cups: the team prefers the controlled descent.
iii.
Bloom & immersion
For larger stockpot batches the team standardized on a full-immersion bloom: tea is dispersed across the surface, water is poured to wet leaves evenly, heat is cut, and the pot is moved off-burner so the steep proceeds along a natural cooldown. Hybrid bloom variants (strainer dip then drop) are reserved for kyusu-scale work.
Investigative technique
The minute-interval sample-out
A general-purpose tool used during specimen development whenever the steep window is uncertain. Run a single brew at a fixed temperature and ratio, then draw small samples at one-minute intervals across a wide candidate range — typically four to ten minutes — and chill each separately. The endpoints are usually obvious in the cup: the early draws read thin and under-extracted, the late draws astringent and over-extracted. The defensible spec sits in the middle. The technique trades vessel complexity for time efficiency: one brew yields the full curve.
Methodology · limitations
What this report is, and is not.

Specimen Report No. 01 — the matcha usucha survey — was a controlled blind survey: two raters, a calibration anchor, replicate tastings, six sensory dimensions on a five-point scale. This document is a different kind of artifact. The data here is a field log: n = 1 per iteration, no blinding (the team developing the protocol is also the team tasting the cup), endpoints recorded in prose ("clarity," "depth," "muddied") rather than scored on a rubric, with temperatures written down in mixed °C and °F as they came off the probe. The Newton's-cooling parameters in the cooldown charts are empirical fits, not physical bath temperatures.

What this means in practice: the report is rigorous about what was done but cannot adjudicate what works best — those judgments rest on a single tongue, a single batch, a single day. The iteration log is published because the discipline of writing it down is itself a quality-control mechanism, and because anyone evaluating Iro should be able to see the path from kettle to cup.

§ II · The Specimens

The specimens.

Four teas, ordered by depth of iteration logged. Each specimen terminates in a locked production specification — what is currently running on the cart.

№
ID
Specimen
Iter.
Water
Status
01
T01
Dark Roast Oolongcold milk-tea base
23iter
W₁ → W₇
Active
02
T02
Tie Guan Yincold, light floral oolong
12iter
RO → W₇
Active
03
T03
Four Season Oolongcold, bright floral
9iter
RO → W₃
Active
04
T04
Jasmine Greencold, aromatic green
4iter
RO → W₇
Active
T01
FormatCold, milk-tea base
TypeRoasted oolong
Iterations23 logged
StatusActive production

The most-iterated specimen in the file. Twenty-three iterations trace a path from a kyusu-scale baseline to a 6-liter stockpot specification, with the dominant axes of refinement shifting from ratio and time in early runs to filtration discipline and cooldown geometry in later ones. The 23rd iteration is the spec currently running on the cart.

# H₂O Spec Method Finding
01 RO 7g : 250g · 90 °C · 3:00kyusu, 50g creamer + 20g fructose-sucrose Initial bloom ~10 s, freezer chill ~2 h, shaken to serve. Smooth, rounded, smoky; minimal astringency; could carry less sweetness.
02 RO 120g : 3000g · 91 °C · 4:00first stockpot batch Direct steep, no bloom protocol formalized. Under-extracted at this ratio; rounded but light; ratio needs lifting.
03 RO 200g : 3000g · 91 °C · 7:00hybrid strainer-dip bloom Strainer pre-dip; geometry only partially soaked the leaves. Over-steep flattened depth; the longer dwell did not compensate for poor wetting.
04 RO 150g : 3000g · 91 °C · 4:30full-surface immersion bloom Tea dispersed by bowl; small stirs; heat off. Approaching iter. 01's depth at scale; smokiness slightly muted; ready to test 90 °C / 4:00.
05 RO 150g : 3000g · 90 °C · 4:00natural off-burner cooldown Pot off heat at start of steep; covered between stirs. Solid balance with 50g creamer / 15g fructose-sucrose. Bloom protocol locked.
06 RO 170g : 3000g · 90 °C · 4:00tuned for organic milk As iter. 05; ratio bumped to compensate for richer dairy. Balance holds in organic-milk context.
07 RO 117g : 2250g · 89 °C · 4:30scaling check Same protocol, smaller batch. Confirms ratio scales linearly under fixed protocol.
08 W₁ 170g : 3000g · 89 °C · 4:30first remineralized run Standard immersion bloom; ice-bath cooldown. Markedly greater clarity, brighter top end; more weight could be desirable. Test harder extraction or higher TDS.
09 W₁ 175g : 3000g · 89 °C · 4:30slow overnight cool As iter. 08; passive fridge cool overnight. Less clarity than iter. 08 — slow descent muddied the cup.
10 W₁ 200g : 3000g · 89 °C · 4:30aggressive ratio test Ice bath, ~70 min to 41 °F. 200g overbearing — ratio ceiling identified.
11 W₃ 180g : 3000g · 89 °C · 4:45higher TDS reading Ice bath to ~80 °F (~35 min), then fridge. Trending toward target weight at this TDS.
12 W₃ 180g : 1000g + 500g ice · 89 °C · 5:45pour-over-ice cooldown Strain directly through cloth onto ice. Pour-over-ice silenced post-strain extraction; cup read like generic black milk-tea. Method retired.
13 W₃ 90g : 1500g · 89 °C · 4:45cloth filter, fridge cool Strain at 83.3 °C through nut-milk cloth; into fridge. Body weight and top end balanced; cloth meaningfully but not fully silences residuals — exactly the desired middle weight. Locked profile candidate.
14 W₃ 90g : 1500g · 89 °C · 4:45partial ice bath Cloth strain, ice bath only to 135 °F (~4 min), then fridge. Evaporated milk swap reads cleanly; cup wants either fuller ice bath or direct fridge for clarity.
15 W₃ 95g : 1500g · 89 °C · 4:45ambient cooldown trace Cloth strain at 77.6 °C; logged passive descent (123.5 °F at 56 min, 92.5 °F at 118 min, 59 °F at 4:35). Ambient cooldown profiled for future SOPs.
16 W₃ 185g : 3000g · 89 °C · 4:45scaled to market run As locked protocol, scaled. First market-grade run at this water profile.
17 W₃ 185g : 3000g · 89 °C · 5:15extended steep test Ice bath to 50 °F. +30 s steep widens depth without astringency penalty.
18 W₃ 300g : 4864g · 89 °C · 5:15two-stage filter, full ice bath Mesh strainer first; ~8 lb ice bath; cloth-filter transfer. Two-stage filtration sequence formalized as SOP.
19 W₃ 3 × (300g : 4864g) · 89 °C · 5:15900g loose-leaf line run Triple-batch repeat of iter. 18 protocol. Confirms three-batch parallelization holds line consistency.
20 W₄ Production line runspecifics held internal Two-stage filter; ~8 lb ice bath; pitcher transfer. Production ratio settled at this density.
21 W₅ Production line runwinter water revision As iter. 20; new water profile. Mineral revision absorbed without flavor disruption.
22 W₆ Production line runspring revision As iter. 20; revised water profile. Higher Mg and baking soda introduced as a second alkalinity source; cup reads marginally rounder.
23 W₇ Production line runmulti-batch parallelization Doubled batch; current-generation SOP. Most recent line iteration.
Thermal profile · post-strain

Active ice bath versus passive ambient.

The same strain temperature, two cooldown geometries. Passive descent (iter. 15) is logged at four points across four-and-a-half hours; active ice bath is anchored by iter. 14's intermediate (135 °F at 4 min) and iter. 10's endpoint (41 °F at ~70 min). Curves are Newton's-cooling fits through the logged anchors — the passive fit reproduces all four iter. 15 anchors within ~1 °F. n = 1 per condition; not replicated. Within these single traces, active crosses 100 °F in 8.7 min; passive in 99 min — roughly an order of magnitude slower. Fit parameter Tamb is empirical, not the physical bath temperature.

100 °F · WARM 41 °F · TARGET 8.7 min 99 min 0 60 120 180 240 200 160 120 80 40 MINUTES POST-STRAIN TEMPERATURE · °F Passive ambient · iter. 15 Active ice bath · iters. 10, 14
Passive ambient — measured (5 anchors) Active ice bath — interpolated from endpoints (open circle = inferred)
Findings — T01
  • The cloth-after-mesh two-stage filtration is the single highest-leverage upgrade in the file: it leaves enough residual mass for body without leaving over-extraction.
  • Active ice-bath cooldown appears to beat passive fridge descent for clarity in this iteration set; the candidate optimum sits at a controlled, fast curve to ~110 °F before final filter.
  • Ratio ceiling at this water profile sits in the 6–7% leaf-to-water by weight band; runs above this threshold read as overbearing.
  • Water mineralization shifts from W₁ → W₃ produced the most audible improvement in clarity; W₄ → W₇ refinements are felt as roundness, not contrast[3].
Locked spec
T01 · Dark Roast Oolong (cold)
Ratio
~6–7% w/wleaf : water
Temp
~88–90 °Cat pour-in
Time
~5:00–5:30full-immersion bloom, off-burner
Water
W₇RO + remineralization
Filter
Mesh → clothtwo-stage
Cooldown
Ice bathto ~42 °F
T02
FormatCold, light floral
TypeTie Guan Yin oolong, light floral
Iterations12 logged
StatusActive production

A delicate spec. The work here was largely about restraint — finding a leaf ratio that did not crowd the floral top notes and a steep window short enough to preserve clarity. Several mid-iteration runs over-extended ratio under the assumption that more leaf would amplify the floral character; the cup told a different story.

# H₂O Spec Method Finding
01 RO 57.3g : 3000g · 91 °C · 4:45five teabags · low-temp open steep Bags suspended in strainer; 20 s wet then drop; occasional stir. Well-balanced and floral; ratio could rise but current leaves room.
02 RO 45.3g : 1500g · 90 °C · 4:30four bags · smaller pot 3L pot; surface temp 78 °C at strain. Good balance with mixed dairy (whole + cream + creamer).
03 RO 90.7g : 3000g · 90 °C · 4:30eight bags · standardized stockpot 7L stockpot; freeze-cool after strain. Good balance and clarity with milk. Anchor for further refinement.
04 W₃ 90.7g : 3000g · 90 °C · 4:30first remineralized run Ice bath to 70 °F, then fridge. Profile holds under remin water; cooldown geometry tested.
05 W₃ 56.6g : 1500g · 90 °C · 4:40five bags · ratio audit Cloth filter; ice bath to 42 °F. Lower ratio is necessary for this tea's intrinsic lightness — extra leaf muddies clarity.
06 W₃ 68.4g : 1500g · 90 °C · 4:30six bags Cloth filter; ice bath, ambient log (108.9 °F at 14 min, 49.6 °F at 67 min). Cooldown trace logged for future SOPs.
07 W₃ 180.4g : 3000g · 90 °C · 4:3016 bags · ratio ceiling test Ice bath from start of brew. Clarity diminished; possible over-extraction from residual bag handling. Omit residual squeeze in future.
08 W₄ 158g : 3000g · 90 °C · 4:3014 bags · scaled run Ice bath. First market-scale run at moderated ratio.
09 W₄ Production line runtemp drop introduced Ice bath to 70 °F only. Shifted to lower brew temp to preserve top notes.
10 W₅ Production line runrefine cooldown Cloth filter; ice bath to 100 °F (~8 min); refrigerate. Stockpot ice bath to 100 °F before fridge gives the cleanest finish.
11 W₆ Production line runtuned scale-up As iter. 10, larger volume. Spec extrapolates cleanly to mid-batch volume.
12 W₇ Production line runcontinued line run As iter. 10–11; market run notation pending. Locked spec held across consecutive market runs.
Findings — T02
  • Floral character is preserved by resisting ratio increases — extra leaf muddied the cup rather than amplifying the florals.
  • A 4 °C drop in brew temp (90 → 86 °C) widened clarity at the same time without losing aroma.
  • Two-stage cooldown (ice bath to 100 °F, then fridge) is the cleanest finish for delicate florals.
  • Skipping the residual-squeeze on bag transfer protects against drift over long runs.
Locked spec
T02 · Tie Guan Yin (cold)
Ratio
~5–6% w/wleaf : water
Temp
~85–87 °Cat pour-in
Time
~4:15–4:45strainer pre-wet, drop at 20 s
Water
W₇RO + remineralization
Filter
Clothno residual squeeze
Cooldown
Ice bath → fridgeto ~100 °F, then ~42 °F
T03
FormatCold, bright floral
TypeSi Ji Chun oolong
Iterations9 logged
StatusActive production

A natural counterweight to T01 — bright, floral, refreshing — and the only specimen in the file where an alternate water profile (W₂) was explored in parallel. The arc is similar: kyusu baseline → stockpot scale-up → two-stage filtration → market run.

# H₂O Spec Method Finding
01 RO 7g : 250g · 90 °C · 4:00kyusu reference Single kyusu pour, no specified bloom. In milk context, reads coconutty / flan-like; in isolation, honeyed and refreshing. Ratio may be light.
02 W₃ 170g : 3000g · 89 °C · 4:45first stockpot at scale Full-immersion bloom; off-burner; surface temp 81.1 °C at 4:30. Ice bath to 40.5 °F in ~1:45. Profile reference established.
03 W₃ 140g : 1500g + 500g ice · 89 °C · 4:45pour-over-ice variant Hot brew strained over ice. Pour-over-ice flatness echoed T01 iter. 12 — method retired.
04 W₃ 90g : 1500g · 89 °C · 4:45extended cooldown profiling Cloth filter; ice bath to 41 °F (~1:45); ambient log to 46.4 °F at 69 min. Cooldown trace logged. 1200g output yield from 1500g input under cloth filter.
05 W₂ 105g : 1500g · 89 °C · 4:45side-experiment water profile Logged temp curve: 83.9 °C at 1', 75.9 °C at 4', 67.6 °C post-pour. W₂ trial; mineral profile diverges from production line. Not adopted.
06 W₃ Production line runscaled to market run Ice bath to 42 °F. First market-scale run on W₃.
07 W₃ Production line runtuned ratio Ice bath to 42 °F. Production density settles in.
08 W₃ Production line runtwo-stage filter formalized Mesh → ice bath to 110 °F → cloth → ice bath to 42 °F. SOP filtration sequence locked across specimens.
09 W₃ Production line runcontinued line run As iter. 08, smaller volume. Spec scales cleanly downward.
Thermal profile · brew → strain → serve

A single iteration's full arc.

Iter. 5 is the only run in the file with logged in-pot temperatures during the steep — three measured anchors (83.9 °C at 1', 75.9 °C at 4', 67.6 °C post-pour at 4:45). The brew-phase fit (k ≈ 0.046 / min) is anchored to those three points and cools slowly under a covered, off-burner pot; the ice-bath segment (k ≈ 0.10 / min, roughly twice as aggressive) is a Newton's-cooling extrapolation from the strain temperature — intermediate ice-bath points are interpolated, not separately logged in iter 5. n = 1; not replicated. Tamb is a fitted parameter rather than the physical pot or bath temperature. Strain happens at the dotted line.

100 °F · WARM STRAIN · 4:45 0 5 10 15 20 25 200 160 120 80 40 MINUTES FROM BREW START TEMPERATURE · °F in-pot brew ice-bath cooldown
Iter. 5 — three measured kettle/strain anchors; ice-bath segment interpolated
Findings — T03
  • Both T01 and T03 confirmed independently that pour-over-ice cooldown silences depth — the method is retired across the file.
  • Ratio settles around 7% leaf-to-water for line operations, slightly above T02 and consistent with T01's ceiling.
  • The W₂ side experiment was carried in parallel for this specimen only; mainstream production water (W₁ → W₇) is the trend line for everything else.
Locked spec
T03 · Four Season Oolong (cold)
Ratio
~7–8% w/wleaf : water
Temp
~88–90 °Cat pour-in
Time
~4:30–5:00full-immersion bloom, off-burner
Water
W₃ (side line)candidate to migrate to W₇
Filter
Mesh → clothtwo-stage with intermediate ice bath
Cooldown
Ice bathto 110 °F, re-strain, then to 42 °F
T04
FormatCold, aromatic green
TypeJasmine-scented green tea
Iterations4 logged
StatusActive production

A short, focused arc. The temperature setting (82 °C) and the 3:30 steep window were both established quickly; the remaining iterations refined ratio for line consistency and adopted the standardized two-stage filtration once it had been formalized on T01 / T03.

# H₂O Spec Method Finding
01 RO 115g : 1875g · 82 °C · 3:30baseline Direct steep; no formalized bloom. Temperature and time anchor established for greens.
02 RO 230g : 3250g · 82 °C · 3:30scaled batch Mesh → ice bath to 110 °F → cloth → ice bath to 42 °F. Two-stage filter sequence applied.
03 W₄ Production line runtuned ratio Two-stage filter; ice baths. Ratio holds clarity for the aromatic profile.
04 W₇ Production line rundenser line run Two-stage filter; ice baths. Most recent line iteration.
Findings — T04
  • Greens require the lowest brew temperature in the file (82 °C); the jasmine top notes are temperature-sensitive in a way the oolongs are not.
  • The 3:30 steep window holds clarity even at higher ratios — atypical for the file, where lifting ratio usually penalizes top end.
  • Filtration SOP transferred over from T01 / T03 without modification.
Locked spec
T04 · Jasmine Green (cold)
Ratio
~8–9% w/wleaf : water
Temp
~81–83 °Cat pour-in
Time
~3:15–3:45direct steep
Water
W₇RO + remineralization
Filter
Mesh → clothtwo-stage with intermediate ice bath
Cooldown
Ice bathto 110 °F, re-strain, then to 42 °F
§ III · Locked Protocols

The locked protocols.

The specifications currently running across active production. Each is the terminus of a documented iteration trail; each is reviewable against the iteration log when a question arises at the cart.

ID Specimen Ratio (% w/w) Temp Time Water Cooldown
T01 Dark Roast Oolong (cold)milk-tea base ~6–7% ~88–90 °C ~5:00–5:30 W₇ Mesh → cloth → ice bath → 42 °F
T02 Tie Guan Yin (cold)floral ~5–6% ~85–87 °C ~4:15–4:45 W₇ Cloth → ice bath → fridge
T03 Four Season Oolong (cold)floral ~7–8% ~88–90 °C ~4:30–5:00 W₃ → W₇* Mesh → cloth → ice bath → 42 °F
T04 Jasmine Green (cold)aromatic green ~8–9% ~81–83 °C ~3:15–3:45 W₇ Mesh → cloth → ice bath → 42 °F

*T03 currently runs on the W₃ side line; migration to W₇ pending the next batch turnover.

§ Coda — open work

Two threads remain open. The first is whether T03 should migrate from its legacy W₃ side line onto the production W₇ water — a small experiment, but one with consequences for inventory simplicity at the cart.

The second is structural. Across all four locked specs, the shared variables — full-immersion bloom, off-burner steep, two-stage filter, controlled ice-bath descent — are now consistent enough to formalize as a single house brewing protocol, with per-tea parameters sitting on top. The next document in this series will likely be that SOP.

§ IV · References

Selected literature.

The chemistry frame behind the iteration log. The findings reported above are this site's own field log; the references below support the underlying mechanisms invoked along the way. Inline numeric markers in the prose link back to the entry that backs the specific mechanism claim. Where work is from the coffee literature, the analogy to tea is noted.

  1. 01
    Hendon, C. H.; Colonna-Dashwood, L.; Colonna-Dashwood, M. "The Role of Dissolved Cations in Coffee Extraction." J. Agric. Food Chem. 2014, 62 (21), 4947–4950. DOI: 10.1021/jf501687c
    Foundational to modern coffee-water remineralization. Density functional theory binding energies between Na⁺, Mg²⁺, Ca²⁺ and coffee acids, caffeine, and eugenol — the molecular basis for choosing CaCl₂, MgSO₄, and bicarbonate ratios in the W₁–W₇ recipes. Applied here by analogy: tea and coffee both involve polyphenol/alkaloid extraction from a plant matrix, so the cation-coordination argument carries over, but it is not direct tea evidence.
  2. 02
    Bai, F.; Chen, G.; Niu, H.; et al. "The types of brewing water affect tea infusion flavor by changing the tea mineral dissolution." Food Chemistry: X 2023, 18, 100681. DOI: 10.1016/j.fochx.2023.100681
    The tea-specific counterpart to Hendon. Compares five water types brewing green tea; Ca and Fe ions most strongly modulate volatile aroma compounds. Low-mineral, neutral-pH water produced the highest aroma intensity. Direct support for keeping the mineral program at low TDS and for building water from RO rather than from tap; the report's specific choice to retire silica is not addressed in Bai but is consistent with the broader "low-mineral water preserves aroma" finding.
  3. 03
    Yan, Z.; Zhong, Y.; Duan, Y.; Chen, Q.; Li, F. "Effect of Water Hardness on Catechin and Caffeine Content in Green Tea Infusions." Molecules 2021, 26 (12), 3485. DOI: 10.3390/molecules26123485
    As water hardness rises, total catechin yield falls — driven by alkaline-catalyzed auto-oxidation of EGC/EGCG. Astringent EGCG and caffeine concentrations climb with [Ca²⁺]. Supports the report's empirical finding that the W₁ → W₃ mineralization shift produced the most audible clarity gain, while later refinements (W₄ → W₇) read as roundness rather than contrast.
  4. 04
    Spiro, M.; Siddique, S. "Kinetics and Equilibria of Tea Infusion: Analysis and Partition Constants of Theaflavins, Thearubigins, and Caffeine in Koonsong Broken Pekoe." J. Sci. Food Agric. 1981, 32 (10), 1027–1032. DOI: 10.1002/jsfa.2740321012
    Foundational tea-infusion equilibrium paper. Two-phase (leaf / aqueous) partition model, with partition constants and enthalpy changes measured at 79.5 °C and 94.0 °C for theaflavins, thearubigins, and caffeine. The mathematical framing underneath "early draws thin / late draws astringent" — different compounds have different equilibrium partitioning and different rate constants in subsequent parts of the same Spiro series, so extracted composition shifts over the steep window.
  5. 05
    Sánchez-López, J. A.; Yener, S.; Smrke, S.; Märk, T. D.; Bonn, G.; Zimmermann, R.; Biasioli, F.; Yeretzian, C. "Extraction kinetics of tea aroma compounds as a function of brewing temperature, leaf size and water hardness." Flavour Fragr. J. 2020, 35 (4), 365–375. DOI: 10.1002/ffj.3571
    The closest match to this report's tuned variables — same three knobs tracked over time via PTR-ToF-MS. Temperature dominates aroma yield; leaf size dominates early-stage aroma profile; water hardness has the smallest effect on aroma in the tested range. The last finding is a useful counter-balance to the W₁ → W₇ narrative — much of the perceived improvement here may sit downstream of variables other than mineralization.
  6. 06
    Lian, G.; Astill, C. "Computer simulation of the hydrodynamics of teabag infusion." Food Bioprod. Process. 2002, 80 (3), 155–162. DOI: 10.1205/096030802760309179
    CFD model of teabag infusion under static vs. mechanically-agitated conditions. Static infusion is dominated by buoyancy-driven natural convection around the leaf bed; agitation produces forced flow that exceeds the buoyancy circulation, giving faster and more homogeneous extraction. The kyusu (small, static) vs. stockpot (large, stirred) extraction-rate difference observed in this report is the same physics scaled up — the paper's findings apply by extension, not by direct test of these vessels.

← Back to all R&D reports

Iro Tea · Internal R&D · Specimen Report No. 02 伊呂 · iro May 2026