Szymczykowska et al. 2026 — Elemental composition of Japanese matcha powder and infusions

This study analyzed the macro- and microelemental profiles of two organic Japanese matcha types — daily matcha (DM, second-to-third harvest) and traditional matcha (TM, first-to-second harvest) — from the Uji region of Kyoto. Each type was analyzed in both powder and infusion form, with infusions prepared at 25, 70, 80, and 90 °C, all in triplicate. Quantification used ICP-OES (Thermo Fisher iCAP 7400) with yttrium internal standard and validation against NIST SRM 8414. Of the heavy metals tracked by Heavy Metal Index, the paper measured Pb, Cr (total), and Ni; it did not measure Cd, As (iAs or tAs), Hg (tHg or MeHg), Al, Sn, Sb, or U. Brewing temperature had no statistically significant effect on elemental content in either infusion type. Matcha type (DM vs TM) showed statistically significant differences only for Ca and Na (higher in DM) and K and Ni (higher in TM); no significant DM/TM differences were detected for any element in the powder. Nickel was detected in both infusion types but was below the detection threshold in the powders; the authors attribute this to matrix interferences in the digested powder samples rather than true absence. Lead concentrations in all tested products fell within the permissible range of EU Commission Regulation 2023/915 (0.050–0.9 mg/kg in plant products depending on category), according to the authors.

Key numbers

The paper presents quantitative data primarily through boxplot figures (Figures 1–4) and supplementary tables (S1–S4 in the MDPI online supplement). Exact per-cell numerical values are not reproduced in the body text. The numerical anchors available from the prose are summarized below.

Elemental ranking patterns (mean concentration order, all temperatures combined; from Discussion §4):

• Daily matcha infusion (mg/L): K > P > Mg > Ca > Na > Mn > Zn > Cu > Sr > Fe > Ni > Pb > Cr
• Traditional matcha infusion (mg/L): K > P > Mg > Ca > Mn > Na > Zn > Cu > Fe > Sr > Ni > Pb > Cr
• Daily matcha powder (mg/kg): K > Ca > P > Mg > Mn > Na > Fe > Zn > Cu > Sr > Cr > Ni (Ni below detection)
• Traditional matcha powder (mg/kg): K > P > Ca > Mg > Mn > Na > Fe > Zn > Cu > Sr > Cr > Ni (Ni below detection)

In both infusion types, Cr is the lowest-concentration element among all elements measured. In both powder types, Ni was below detection.

Heavy-metal absolute concentrations. The paper presents heavy-metal numerical values through Figures 1–4 boxplots (median, IQR, whiskers) and supplementary Tables S1–S4 (not in the body text). Author-reported intake-percentage statements in Discussion §4 (p. 11) provide a second numerical channel. Where the two channels disagree, the figure values are the more direct measurement and are flagged as such below.

• Pb in infusion: Figure 1 (p. 5; y-axis mg/L) shows DM Pb roughly 0.001–0.025 mg/L (1–25 µg/L) and TM Pb roughly 0.001–0.040 mg/L (1–40 µg/L), both with broad IQRs. No statistically significant DM vs TM difference for Pb in infusion. The authors state infusion-temperature variation does not affect Pb content.
• Pb in powder: Figure 4 (p. 8; y-axis mg/kg) shows DM Pb roughly 0.5–1.5 mg/kg (median ~0.7 mg/kg) and TM Pb roughly 0.5–7 mg/kg (median ~5 mg/kg, whisker reaching ~7 mg/kg). The authors state in Discussion §4 (p. 11) that “the permissible concentration of lead in plant products ranges from 0.050 mg/kg to 0.9 mg/kg, depending on the type of product (vegetables, fruits, spices, etc.) [per EU Reg 2023/915], which means that the tested products are safe in this respect.” Caveat: Figure 4 visually suggests that TM matcha powder Pb concentrations exceed the upper bound of that EU 2023/915 plant-product range (0.9 mg/kg), and that even DM matcha powder Pb is in the upper part of that range. This is a paper-internal tension between the authors’ compliance assessment and their own Figure 4. The literal numerical values in Table S4 are needed to resolve which side is correct; do not rely on the authors’ compliance statement for matcha powder Pb without that confirmation.
• Cr in infusion: Figure 2 (DM, p. 6) shows Cr roughly 0.002–0.014 mg/L (2–14 µg/L) across the four brewing temperatures, with median ~0.008 mg/L. This is broadly consistent with the author-reported “approximately 2.5% of the 40 µg/day RI” per 100 mL cup (which back-derives to ~10 µg/L). No Cr speciation was performed; values represent total Cr.
• Cr in powder: Figure 4 (p. 8) shows Cr roughly 2.0–2.7 mg/kg (median ~2.4 mg/kg) for both DM and TM, with no statistically significant difference. Caveat: the authors’ Discussion §4 statement that “a 1.75 g teaspoon of matcha covers 0.50% of the 40 µg/day RI” back-derives to 0.114 mg/kg powder Cr, roughly 20× lower than Figure 4. This appears to be an arithmetic or transcription error in the Discussion text rather than a figure error: the figure value of ~2.4 mg/kg, applied to a 1.75 g serving, gives 4.2 µg Cr, or ~10.5% of the 40 µg/day RI — consistent with the figure but not the text. The figure-based value is the more direct measurement; Table S4 confirmation would settle this.
• Ni in infusion: Figure 1 (p. 5) shows DM Ni roughly 0.06–0.08 mg/L (60–80 µg/L) and TM Ni roughly 0.06–0.16 mg/L (median ~0.10 mg/L, 60–160 µg/L). Statistically significant TM > DM difference (Figure 1 asterisk). The authors state observed Ni concentrations correspond to approximately 0.9–1.1% of EFSA’s TDI of 13 µg/kg body weight/day per 100 mL serving for a 70 kg adult (910 µg/day TDI), which back-derives to ~8.2–10.0 µg Ni per 100 mL = ~82–100 µg/L. The text-derived value (~82–100 µg/L) sits at the lower end of the figure-derived range; combined, the central tendency for matcha-infusion Ni is approximately 80–100 µg/L for DM and approximately 100 µg/L (with substantially higher whisker) for TM. The authors note the absolute difference is biologically negligible relative to the TDI and that a water-matrix Ni contribution cannot be fully excluded.
• Ni in powder: Figure 4 (p. 8) shows Ni at the analytical floor (y-axis ~0.00010 mg/kg, with both DM and TM samples plotting on a single horizontal line indistinguishable from the detection limit). The authors describe this as below detection in both matcha types and propose that matrix-bound Ni in the solid phase causes signal suppression during ICP-OES rather than true absence; the soluble fraction extracted into infusion is more analytically accessible. They explicitly caveat that a water-matrix Ni contribution to the infusion values cannot be independently excluded.

Macro/microelement intake-fraction anchors (provided by the authors as context for non-heavy-metal nutritional contribution; included here for completeness, since the same data pages report all elements together):

• Mn: a 100 mL cup of matcha (1.75 g) provides “just under 40%” AI (2.3 mg/day, men) and “just under 50%” AI (1.8 mg/day, women); 1.75 g powder ~35% AI (men) / “just over 40%” AI (women). Largest single-nutrient contribution among elements measured.
• Mg: 100 mL infusion ~0.90% RDA (420 mg/day, men) and ~1.20% RDA (320 mg/day, women); 1.75 g powder ~0.70% / ~0.80%.
• Ca: 100 mL infusion <0.30% RDA (1000 mg/day); 1.75 g powder ~0.50% RDA.
• P: 100 mL infusion ~1.10% RDA (700 mg/day); 1.75 g powder <0.80% RDA.
• K: 100 mL infusion ~1.20% AI (3500 mg/day); 1.75 g powder ~0.50% AI.
• Na: 100 mL infusion ~0.10% AI (1500 mg/day); 1.75 g powder ~0.01% AI.
• Zn: 100 mL infusion <0.7% RDA (men, 11 mg/day) / <1% RDA (women, 8 mg/day); 1.75 g powder ~0.40% / ~0.50%.
• Cu: 100 mL infusion ~2% RDA (0.9 mg/day); 1.75 g powder ~3% RDA.
• Fe: 100 mL infusion ~0.10% RDA (men, 10 mg/day) / ~0.07% RDA (women, 18 mg/day); 1.75 g powder ~1.2% / ~0.66%.
• Sr: detected in both infusions and powders; no RDA, RI, or UI value established for strontium; no toxicity threshold reported.

Statistically significant DM vs TM differences (from Results §3 and Figure 1):

• Infusions: K and Ni higher in TM; Ca and Na higher in DM. All other elements: no significant difference.
• Powders: no statistically significant differences between DM and TM for any element.

Brewing-temperature effect: Across 25, 70, 80, and 90 °C, no statistically significant differences in elemental content were observed in either DM or TM infusions for any element (Figures 2 and 3; Tables S2 and S3).

Comparison to prior matcha literature (from Discussion §4): The paper identifies Kolackova et al. 2020 (J. Food Compos. Anal. 92:103581) as the only prior matcha elemental-composition study. That study reported the matcha-powder elemental order as K > P > Ca > Mg > Na > Fe > Zn > Mn > Cu > Ni > Sr > Cr > Pb across ten matcha samples of varying origins; the present study’s powder rankings (above) diverge in the placement of Mn and Na but agree on the overall hierarchy of K/P/Ca/Mg at the top and Cr/Ni at the bottom. The authors attribute differences to cultivar, harvest time, processing, and sample origin.

Methods (brief)

Plant material: Two certified organic Japanese matcha products from the Uji region of Kyoto, both processed from Tencha leaves of Camellia sinensis (L.) Kuntze using identical shading, steaming, drying, and stone-grinding procedures; harvest timing was the primary distinguishing variable (DM second-to-third harvest, TM first-to-second harvest). Both certified by JONA (Japan Organic And Natural Foods Association) and AgroBioTest (Polish EU-aligned organic certifier).

Infusion preparation: 1.75 g matcha powder transferred to a conical flask, 100.0 mL distilled water added at the test temperature (25, 70, 80, or 90 °C), closed flask rotated at 180 rpm (Brunswick Excella E24, New Brunswick Scientific, Edison, NJ, USA) for 10 min, filtered to separate plant material, cooled to room temperature before analysis. Each infusion prepared and analyzed in triplicate.

Sample digestion: Microwave-assisted acid digestion using MARS 5 CEM system (CEM Corporation, Matthews, NC, USA). 0.8 mL sample volume transferred to clean polypropylene tubes, 2 mL of 65% HNO3 added, 30 min prereaction phase under clean hood, then 0.5 mL of unstabilized 30% H2O2 added. Transferred to Teflon vessels and digested at 180 °C for 35 min (15 min ramp-up, 20 min hold). Post-digestion: cooled, transferred to acid-cleaned 15 mL polypropylene tubes, 5-fold dilution. Yttrium internal standard added (0.5 mg/L final concentration), 1 mL of 1% Triton X-100 (Sigma, Kawasaki, Japan), diluted to 10 mL with 0.075% HNO3. Blank samples: 500 µL concentrated HNO3 diluted identically.

Quantification: ICP-OES on iCAP 7400 (Thermo Fisher Scientific Inc., Waltham, MA, USA), concentric nebulizer, cyclonic spray chamber, both radial and axial viewing modes. Reported analytical wavelengths: Zn 206.200 nm, Cr 205.560 nm, Mn 257.610 nm, Cu 224.700 nm, Fe 259.940 nm. Wavelengths for the other elements (Ca, K, Mg, Na, Ni, P, Pb, Sr) are not specified in the methods text. Method validation: NIST SRM 8414 (Bovine Muscle Powder) reference material; yttrium recovery 90–106%; R² values for all standard curves 0.998–1.000. Limit-of-detection values are referenced but not numerically reported in the text.

Elements measured (n=13): Ca, Cr, Cu, Fe, K, Mg, Mn, Na, Ni, P, Pb, Sr, Zn. The Heavy Metal Index analyte slate excludes most of these as macro/microelements; Pb, Cr (total only, no speciation), and Ni are the three heavy-metal analytes covered by the paper.

Not measured: Cd, As (tAs or iAs), Hg (tHg or MeHg), Al, Sn, Sb, U. Al is a notable absence given green tea’s known status as an Al hyperaccumulator (cf. ozturk2024-al-tea-infusion-teapot for Al in green/black/white tea infusions including matcha-relevant green tea data).

Statistical analysis: MedCalc Statistical Software v20.218 (MedCalc Software Ltd., Ostend, Belgium); Microsoft Excel 2017. Normality assessed by Shapiro–Wilk test. Non-parametric tests applied (distribution non-normal, variance heterogeneous): Kruskal–Wallis for between-group comparisons, Spearman rank correlation for parameter relationships. Results reported as median and interquartile range. Significance threshold p ≤ 0.05. Boxplot generation: Python 3.12, Matplotlib 3.10.5, Seaborn 0.13.2.

Sample-scope caveat (stated by the authors in §5 Limitations): only two matcha products from a single producer in a single geographical region were analyzed. Results should not be considered representative of all Japanese matcha. Mineral bioavailability and physiological effects were not evaluated.

Implications

Certification: This paper provides Pb, Cr (total), and Ni occurrence data for high-quality organic Japanese matcha powder and infusions. Coverage of the HMTc 10-analyte slate is incomplete — Cd, iAs, tAs, MeHg, tHg, Al, and Sn are not measured — so the paper alone is insufficient as standalone evidence for matcha certification thresholds. For the matcha product category, where the whole-leaf suspension is consumed (rather than a filtered infusion as with other tea categories), powder-basis concentrations are the relevant exposure basis for most analytes. Two paper-internal tensions are load-bearing for any downstream use of this source: (1) the authors’ EU 2023/915 plant-product compliance claim for Pb does not appear consistent with Figure 4 TM-powder Pb values (visually reaching ~5–7 mg/kg, above the 0.9 mg/kg ceiling cited by the authors); (2) the authors’ Cr-powder intake-percentage statement (0.50% of 40 µg/day RI) is roughly 20× lower than Figure 4 (~2.4 mg/kg powder Cr). Both tensions require Table S4 numerical confirmation before this source’s powder-basis values are used for matcha threshold work. Infusion-basis Pb, Cr, and Ni values appear internally consistent between Figures 1–2 and the author-reported intake percentages.

Courses: Critical matcha-specific teaching point: because matcha is consumed as a whole-leaf suspension rather than a filtered infusion, consumers receive the full elemental load of the powder, not just the water-soluble fraction. This is the structural reason why matcha exposure modeling differs from other tea products, and why powder-basis concentrations rather than infusion-basis concentrations are the correct input for matcha dietary exposure estimates. The brewing-temperature finding (no significant effect across 25–90 °C) is attributed by the authors to the fine suspension nature of matcha — surface area is already maximized, so equilibration with water is rapid at all temperatures studied.

App: For consumer-facing exposure estimation, matcha is a distinct ingredient from steeped green tea. The consumed form (whole powder as suspension) means powder concentrations are the correct input for the app’s risk model. Figure-based numerical anchors usable from this paper: infusion Cr ~2–14 µg/L (median ~8 µg/L); powder Cr ~2.0–2.7 mg/kg (median ~2.4 mg/kg); infusion Ni ~60–80 µg/L DM and ~60–160 µg/L TM (TM > DM, statistically significant); powder Ni at analytical floor; infusion Pb ~1–25 µg/L DM and ~1–40 µg/L TM; powder Pb ~0.5–1.5 mg/kg DM and ~0.5–7 mg/kg TM. Values are visually estimated from boxplot figures and should be replaced by Table S1/S4 exact values when available. The Pb-powder range for TM, in particular, suggests matcha powder may contribute meaningfully more Pb exposure than the authors’ EU-compliance narrative implies. Cd, As (any speciation), Hg (any speciation), and Al values must come from other sources.

Microbiome: Not addressed in this paper.

Wiki pages updated on ingest

• matcha-powder
• tea
• lead
• chromium
• nickel
• matcha
• tea-infusions

Verification notes

Merge-enhance pass 2026-05-20 (Claude Opus 4.7, autonomous ingest cycle; existing page updated: 2026-05-13 predated the 2026-05-14 schema cutoff):

• Stripped invalid ingredient slug [[ingredients/green-tea]]. No wiki/ingredients/green-tea.md exists in the current taxonomy; green-tea is an alias on the umbrella [[ingredients/tea]] page (and exists separately as a matrices-vocabulary token). Per CLAUDE.md Part 10 (5-paper threshold for sub-variant creation), the ingredient array now uses the umbrella [[ingredients/tea]] alongside the matcha-specific [[ingredients/matcha-powder]]. The green-tea pedagogy is preserved verbatim in the body prose.
• Stripped invalid product slug [[products/matcha-beverages]]. No wiki/products/matcha-beverages.md exists. The two existing routing destinations are [[products/matcha]] (the matcha product category) and [[products/tea-infusions]] (the broader umbrella for brewed tea drinks). Both are retained; the invented matcha-beverages slug is dropped.
• Corrected matrices token matcha-infusion → tea-infusion. No matcha-infusion token exists in the matrices controlled vocabulary visible in the taxonomy snapshot or in any other tea/matcha source page. All other tea-infusion source pages (ozturk2024, brzezicha-cirocka2016, jurowski2023, li2013, kazeminia2023) use tea-infusion. Conformed for cross-source consistency. matcha-powder retained since it is in the matrices vocabulary.
• Corrected ## Wiki pages updated on ingest list. Prior version listed [[ingredients/matcha]] (no such page) and [[ingredients/green-tea]] (no such page). Replaced with [[ingredients/matcha-powder]] and [[ingredients/tea]]. Added [[products/matcha]] (was missing from list despite being declared in frontmatter).
• Corrected truncated raw_path. Prior revision ended at “…Pot.pdf” (truncated mid-filename); restored to the full filename “…Potential Role as a Functional Food in Metabolic Health.pdf”.
• Expanded Key numbers with back-derived absolute concentrations. The paper reports heavy-metal content primarily as percentages of European RDA/RI/AI/TDI values, with absolute values in supplementary tables not reproduced in the text. Computed approximate µg/L (infusion) and mg/kg (powder) values for Cr and Ni from the author-stated percentages, labeled as back-derived. Documented explicitly that Pb absolute values are not in the text and require Table S1 / S4 access.
• Added Methods detail. Prior revision noted ICP-OES, MARS 5 digester, NIST SRM 8414, and “MARS 5 CEM 180 °C”. The merged version preserves all of those (Part 12 Exception 2 covers the vendor names: PerkinElmer-equivalent Thermo Fisher iCAP 7400 ICP-OES, CEM MARS 5, NIST SRM 8414, MedCalc, Brunswick Excella E24, Sigma Triton X-100, GL Sciences) and adds: the full digestion protocol (0.8 mL sample, 2 mL HNO3, 30 min prereaction, 0.5 mL H2O2, 180 °C/35 min, 5-fold dilution), the analytical wavelengths reported for Zn/Cr/Mn/Cu/Fe (with explicit note that wavelengths for the other 8 elements are not stated), the validated yttrium recovery range (90–106%), the calibration R² range (0.998–1.000), and the statistical analysis package and methods.
• Documented the Ni-in-powder non-detection caveat. The paper presents a methodological hypothesis (matrix-bound Ni in solid phase causing signal suppression during ICP-OES, vs water-extractable Ni mobile in infusion) for why Ni was detected in infusions but not powders. The paper also explicitly states this interpretation has caveats: a water-matrix Ni contribution cannot be fully excluded. Both points are now in the merged page rather than implied.
• Added the Limitations §5 sample-scope caveat to the Methods section. The prior revision noted “single region and single producer” in passing; the merged version surfaces the authors’ own §5 Limitations statement that results should not be extrapolated to all Japanese matcha, and that mineral bioavailability and physiological effects were not evaluated.
• Tightened the Pb statement. Prior text: “all tested products are safe with respect to Pb under EU Regulation 2023/915, which sets permissible Pb in plant products at 0.050–0.9 mg/kg depending on product type.” The merged version preserves this paraphrase and labels it explicitly as the authors’ assessment (rather than a wiki-side claim), per Part 2 firewall discipline. It also notes that exact Pb values from supplementary tables would be required for quantitative pooling.
• Tightened Implications language for Part 2 firewall compliance. Prior Certification block was already disciplined (“incomplete analyte coverage limits its use as a standalone HMT&C evidence source”); the merged version preserves this framing and adds the explicit caveat that Cr-in-powder is a useful occurrence anchor “if confirmed against Table S4” rather than presenting back-derived estimates as established values.
• Cross-linked to ozturk2024-al-tea-infusion-teapot in Methods as the Al-coverage complement, since the present paper does not measure Al and Al is a known concern for tea-leaf matrices.
• Preserved cite_key, raw_handle (manual-fetch-kimi per the existing-page pattern; the autonomous cycle’s other 2026-05-20 merge-enhances of pre-2026-05-14 pages used the same value), license, evidence_tier, doi, sample_n, jurisdictions, source_type, publication. No brand-by-brand contamination listings appear in the source; no Part 12 brand-firewall actions were needed beyond Exception 2 instrument-vendor allowances already in place. No new ingredient, product, regulation, or metal pages proposed.

Phase 2 audit application 2026-05-20 (fresh-context Agent subagent, verdict REVISE):

• Applied ⚠️ Cr-powder figure-vs-text discrepancy finding. Auditor flagged that Figure 4 (p. 8) shows Cr powder ~2.0–2.7 mg/kg, ~20× higher than the back-derived 0.114 mg/kg from the authors’ “0.50% of 40 µg/day RI” statement (p. 11). Verified independently against the PDF: Figure 4 second-row right plot has y-axis labeled “Cr [mg/kg]” with values 2.0–2.7 across both DM and TM boxplots. Reverse-checking the figure value: 2.4 mg/kg × 1.75 g = 4.2 µg Cr per teaspoon = 10.5% of the 40 µg/day RI, not 0.50%. The authors’ Discussion §4 statement is arithmetically inconsistent with their own Figure 4. The Key numbers section now reports the figure-based value as primary and flags the text-derived back-derivation as inconsistent.
• Applied ⚠️ Pb-powder figure-vs-author-claim caveat finding. Auditor flagged that Figure 4 shows TM matcha powder Pb values reaching ~5–7 mg/kg, which exceeds the 0.9 mg/kg upper bound of EU Reg 2023/915 plant-product limits the authors themselves cite. Verified independently against the PDF: Figure 4 fourth-row middle plot has y-axis labeled “Pb [mg/kg]” with TM box spanning ~0.5 to ~7, median ~5 mg/kg; DM box spans ~0.5 to ~1.5, median ~0.7 mg/kg. The authors’ compliance claim (“the tested products are safe in this respect”) is not supported by the figure for TM powder and is in the upper part of the cited range even for DM powder. The Key numbers section now reports the figure values, flags the tension explicitly, and notes that Table S4 numerical confirmation is required before either side of the contradiction is relied on. The Implications section is updated to surface both tensions as load-bearing for downstream use of this source.
• Applied ⚠️ Mn anchor paraphrase tightening. Auditor flagged that the wiki’s ~40%/~50% paraphrase loses the paper’s “just under 40%” / “just under 50%” hedges. Verified against PDF p. 10: paper text reads “just under 40% AI coverage for men and just under 50% for women” / “approximately 35% AI coverage for men and just over 40% AI coverage for women”. Restored the “just under” / “just over” qualifiers to preserve the paper’s actual phrasing.
• Added figure-based numerical anchors to Key numbers and App block. Now that the Cr-powder and Pb-powder figure-vs-text contradictions required reading Figures 1–4 directly, also added visually-estimated infusion Pb (DM ~1–25 µg/L; TM ~1–40 µg/L), infusion Cr (DM ~2–14 µg/L median ~8), powder Ni (at analytical floor), and infusion Ni (DM ~60–80 µg/L; TM ~60–160 µg/L median ~100) ranges with explicit “visually estimated from boxplot” caveat. These give downstream synthesis a baseline before Table S1/S4 numerical confirmation is available.
• Did not change frontmatter, slug vocabulary, methods detail, or firewall posture. Auditor confirmed ✅ on Checks 2 (slugs), 3 (speciation/methods), 4 (brand firewall), and 5 (wiki/HMTc firewall) — no edits needed in those areas.
• No findings rejected as false positives. All three ⚠️ findings were verified against the PDF and applied.

Page history

The five most recent substantive edits to this page. The full version history lives in git; when DOI minting comes online (see schema docs), each entry below will also link to a version-pinned DataCite DOI.