Tjandraatmadja et al. 2010 — Sb, Cr, Co, Mo, Se, F, Cl, N, P in 156 Australian household and personal-care products (CSIRO Smart Water Fund report)

A 103-page CSIRO Water for a Healthy Country Flagship report prepared for the Victorian Smart Water Fund (Round 3, Project 5) characterising the concentration of nine priority elements — antimony (Sb), chloride (Cl), cobalt (Co), chromium (Cr, total), fluoride (F), molybdenum (Mo), selenium (Se), Total Kjeldahl Nitrogen (TKN) and total phosphorus (TP) — in 156 household and personal-care products purchased from Melbourne supermarkets in early 2008. The intent of the report is wastewater-management rather than consumer-product safety: the elemental concentrations are converted into weekly per-capita household loads using assumed product dosages and use frequencies (Roberts 2005), then compared against the anthropogenic excretion load to attribute domestic-wastewater pollution between human waste and household-product use. From a heavy-metals occurrence standpoint the analytical-chemistry contribution is the per-category concentration tables for the four panel metals tracked by Heavy Metal Index — antimony (an HMTc Category-1 analyte), chromium (HMTc Cr-VI is not speciated here; total Cr is reported), and the trace-element panel companions cobalt, molybdenum and selenium — measured on an as-purchased finished-product basis. Antimony was the least-frequently detected element (9.6 % of products, detection limit 5 µg/L); chromium was detected in 64 % of products at concentrations ranging from 0.002 to 6.2 mg/kg. Brand identities (private, market-leader, environmental-label) appear as descriptive columns in the source tables and are stripped here per the Part 12 brand firewall; only product-form-and-brand-class descriptors are retained.

Key numbers

Antimony (Sb) concentration in household products (Table 6, p. 22; units mg/kg of finished product, as-purchased basis; LOD 5 µg/L per Table 4 → ≈0.05 mg/kg for typical 10 % w/v dilution):

Sb was below the 5 µg/L detection limit in every brand of toilet cleaner, disinfectant, sunscreen, soap, hand wash, shampoo, hair colouring, deodorant, dishwashing liquid, dishwashing tablet, rinse aid, fabric softener, soaker, wool wash, and most cleaning and personal-care formulations. Overall detection frequency: 15 / 156 = 9.6 %. The highest single Sb concentration recorded was 0.82 mg/kg in a dishwasher powder; the second highest was 0.47 mg/kg in a laundry powder. Sb was detected in 46 % of washing powder brands and 20 % of laundry liquid brands.

Chromium (Cr, total) concentration in household products (Table 9, p. 28; units mg/kg of finished product, as-purchased basis; LOD 0.5 µg/L per Table 4 → ≈0.005 mg/kg for typical 10 % w/v dilution):

Cr was below detection in all brands of shaving foam, soap bars, toilet cleaners, and one brand of depilatory product. Overall Cr detection frequency: 64 % of the 156-product set. Highest single concentrations: 174 mg/kg in one in-cistern toilet freshener (this is the report’s largest single Cr datum and the highest weekly load product), 6.2 mg/kg in a market-leader hand wash, 3.9 mg/kg in a deodorant, 3.50 mg/kg in a fabric softener and a laundry liquid. The report does not speciate Cr-VI versus Cr-III — total Cr only — so the dataset does not feed the HMTc Cr-VI analyte directly.

Cobalt (Co) concentration in household products (Table 8, p. 26; units mg/kg of finished product; LOD 0.5 µg/L → ≈0.005 mg/kg in typical dilution):

Co was below detection in all brands of laundry powder, soaker, dishwasher detergents, hair colouring, oral care, and deodorants. Cobalt was detected most frequently in bathroom personal-care products: 67 % of conditioners, 73 % of shampoos, 80 % of hand washes, 100 % of body washes. Highest single concentration: 7.6 mg/kg in one dishwashing liquid (note the report median for that category is 0.01 mg/kg — the 7.6 datum is the maximum of one outlier brand). Maximum weekly loads ranged 0.8 mg/pe/wk (in-cistern freshener), 0.56 mg/pe/wk (dishwashing liquid), 0.49 mg/pe/wk (laundry liquid).

Molybdenum (Mo) concentration in household products (Table 11, p. 32; units mg/kg of finished product; LOD 0.5 µg/L per Table 4):

Mo was detected in 29 % of the 156-product set. Below detection in all hair colouring brands. Detection prevalence ≥45 % only in toilet fresheners (86 %), recycled toilet paper (62.5 %), laundry powder concentrates (62 %), dishwasher products (50 %), and shampoos and deodorants (45 %). Maximum estimated weekly Mo load is 0.34 mg/pe/wk (laundry powder concentrate).

Selenium (Se) concentration in household products (Table 12, p. 34; units mg/kg of finished product; LOD 4 µg/L per Table 4 → ≈0.04 mg/kg in typical dilution):

Se detected in 18 % of products (30 / 156). Highest single concentration: 4.6 mg/kg in one anti-dandruff shampoo (selenium sulfide is a known anti-dandruff active). Maximum weekly Se loads: laundry liquid 0.15 mg/pe/wk, anti-dandruff shampoo 0.12 mg/pe/wk, laundry powder 0.09 mg/pe/wk.

Frequency-of-detection summary across the four HMI-tracked panel metals in the 156-product set (Section 4.1 narrative, p. 21): TP 97 % > TKN 84 % > Cl 79.5 % > Cr 64 % > Mo 29 % > F 22.7 % > Co < 30 % > Se 18 % > Sb 9.6 %. Of the HMI panel, antimony is the least-frequently detected; total chromium is the most-frequently detected.

Per-product weekly load attribution (Sections 4.1.1-4.1.7 and Appendix 2; units mg of element per person per week, mg/pe/wk, calculated from concentration × dose × frequency for a one-person household):

Comparative load vs human excretion (“anthropogenic load”) for a one-person household (Executive Summary, p. xiii-xiv):

• Antimony: product load varied 0 - 95× the anthropogenic load; median product load ≈ 9 % of excretion load; laundry and kitchen the dominant product sources.
• Chromium: products generated 50 % - 276× the anthropogenic load — i.e., for most product brand mixes household products exceeded human-waste contribution to wastewater Cr, often by ≥ one order of magnitude.
• Cobalt: product load < 0.25× to 13× the anthropogenic load (~25 % to 1300 %).
• Molybdenum: product load up to 97 % of anthropogenic load.
• Selenium: product load up to 60 % of anthropogenic load.

The framing for these comparisons is wastewater-management (the load arriving at Melbourne sewage plants), not consumer exposure. The percentages are not bioavailable-dose comparisons.

Methods (brief)

Sample acquisition. 156 unique household and personal-care product brands were purchased from major Melbourne supermarkets during February-March 2008. For each product category at least one private/supermarket-label brand, one market-leader brand, and (where available) one brand carrying an “environmentally friendly” / “natural” label were included; one sample per brand. The brand mix is enumerated in Sections 3.1.1 (cleaning), 3.1.2 (dishwashing), 3.1.3 (laundry), 3.1.4 (personal care). Brand identities appear in the underlying product-by-product appendix tables but are not reproduced here per Part 12.

Sample preparation and digestion. Liquid products were diluted 1:10 v/v in deionised water (or analysed in their original state for nutrient analysis). High-organic-content solids (e.g., toilet paper) were dry-ashed: a 5 g sample was ashed in a covered crucible in a muffle furnace at 450 °C for one hour or until mass stabilised. Ashed and diluted samples were then acid-digested with high-purity nitric acid (Fluka TraceSelect) and hydrochloric acid (BDH AnalaR) at 95 °C following APHA/AWWA method 3030E, with acid added in 5 mL increments; digests were transferred to a 100 mL volumetric flask and made to volume with deionised water. Reagent blanks were prepared for each 20-sample batch.

Metal analysis instrumentation. Antimony, cobalt, chromium, molybdenum, and selenium were quantified by inductively coupled plasma mass spectrometry on a Thermo X-series ICP-MS. The other element panel members (chloride at WSL115 ion chromatography in an external NATA-accredited lab; fluoride by APHA 4500-F ion-selective electrode; total Kjeldahl nitrogen by APHA 4500-N Org B at a NATA lab; total phosphorus by APHA 4500 at the CSIRO Highett laboratory) are reported separately and are not the HMI-relevant analytes. The report cites a parallel ICP-AES instrument (Varian Liberty Series II) but does not state explicitly which panel metals were apportioned between ICP-MS and ICP-AES; the text in Section 3.2.1 names ICP-MS specifically for Sb, Co, Cr, Mo and Se, and that is the assignment carried here. The report does not name a certified-reference material, recovery percentage, or interlaboratory comparison for the Sb/Co/Cr/Mo/Se panel; calibration is documented only via method-detection-limit values (Sb 5 µg/L, Co 0.5 µg/L, Cr 0.5 µg/L, Mo 0.5 µg/L, Se 4 µg/L, all in the prepared digest).

Concentration calculation. Element concentration in the finished product (mg/kg) was computed as Conc_product = Conc_measured × V_sample / M_product × 10³, where Conc_measured is the digest concentration (mg/L), V_sample is the volume analysed (L) and M_product is the mass of product tested (g). Geometric mean and median across all brand samples in each product category were the central-tendency statistics reported in Tables 6-15.

Load calculation. Per-capita weekly mass load was estimated as M_week = f × dose × Conc_product, where f is the product-use frequency per week and dose is the manufacturer-recommended per-use dose (kg). Frequency and dose assumptions for a one-person household are tabulated in Table 5 (e.g., shampoo 5 g × 5.3 /wk; laundry powder 126 g × 3 /wk; dishwasher powder 60 g × 1.8 /wk; toothpaste 3 g × 14 /wk; deodorant 0.2 g × 7 /wk). Loads are presented in Tables 28-36 of Appendix 2.

Implications for HMI / HMTc

• Antimony in detergents is real but low-frequency. The 9.6 % detection frequency across 156 mixed household and personal-care products, with a concentration ceiling of 0.82 mg/kg in dishwasher powder and 0.47 mg/kg in laundry powder, is consistent with the source’s own characterisation of Sb as “a minor contaminant in household products” (Section 4.1.1, p.22). Per-capita weekly Sb load is under 0.18 mg/pe/wk even for the highest-Sb laundry-powder brand. The data feed the Sb panel cell on laundry-detergents, automatic-dishwasher-detergents, dish-soaps-manual (below detection in all 6 manual liquid brands tested), toilet-bowl-cleaners, floor-cleaners-hard-surface (Sb detected in one of four floor cleaners at 0.07 mg/kg), fabric-softeners (below detection in all 7 brands), toothpaste (one of nine brands at 0.15 mg/kg).
• Chromium pattern is dominated by one in-cistern outlier. The 174 mg/kg Cr datum in one in-cistern toilet freshener — three orders of magnitude above the next-highest household-product Cr value — drives the 35.8 mg/pe/wk weekly Cr load attributed to in-cistern fresheners. The provenance of the 174 mg/kg figure is plausibly a chromate dye in the cistern-water-staining formulation (the report does not state this explicitly). For HMI synthesis the in-cistern category should be evaluated separately from the rest of the chromium panel because of this outlier, and the absence of Cr-VI speciation means the Cr-VI HMTc cell cannot be informed by this dataset.
• Cobalt occurrence in personal-care wash products. Co was detected in 67 % of conditioners, 73 % of shampoos, 80 % of hand washes, and 100 % of body washes — the most frequently-Co-detecting categories in the panel. Concentrations were low (≤7.6 mg/kg) but the prevalence-across-brands pattern is consistent with Co being a trace impurity of dye, surfactant, or fragrance components rather than a category outlier. The dataset informs the Co panel cell on shampoo-adult, conditioner-adult-rinse-out, hand-soap, body-wash-shower-gel.
• Selenium-sulfide anti-dandruff shampoos. One brand of anti-dandruff shampoo recorded 4.6 mg/kg Se, by far the highest single Se value in the dataset. Selenium sulfide is a well-known anti-dandruff active ingredient (1-2.5 %); 4.6 mg/kg is two orders of magnitude below a labelled selenium-sulfide active concentration, so this brand presumably uses an alternative anti-dandruff active and the 4.6 mg/kg represents a trace impurity rather than a labelled Se ingredient. The dataset feeds shampoo-adult (anti-dandruff sub-segment) Se panel cell.
• Molybdenum trace-level signal in laundry powder and cistern products. Mo was concentrated in laundry powder concentrates (62 % detection, median 0.56 mg/kg, max 1.35 mg/kg) and toilet fresheners (86 % detection in cistern category). Loads remain below 0.34 mg/pe/wk for any single product. Feeds laundry-detergents and household-specialty-cleaners-other Mo panel cells.
• Brand-class signal (private vs market-leader vs environmental-label). The report’s most useful brand-class observation is that environmental-label brands generally carried lower phosphorus than the category mean but no consistent across-the-board pattern emerged for the trace metals. For Sb, Cr, Co, Mo, Se there is no environmental-label-versus-conventional advantage reliable across categories. This finding constrains the consumer-substitution lever for trace-metal load reduction and is the kind of evidence HMTc Category-13 (Cosmetics) market-ratcheting decisions need to reference.

Wiki pages this source may touch

• antimony (Sb in laundry powder, dishwasher powder, toothpaste, floor cleaner, toilet paper, 9.6 % detection frequency across 156 Australian household products)
• chromium (total Cr in 64 % of 156 household products; one in-cistern toilet freshener at 174 mg/kg; not speciated)
• cobalt (Co in personal-care wash products at 67-100 % detection prevalence in shampoos, conditioners, hand washes, body washes)
• molybdenum (Mo in laundry powder concentrate and toilet fresheners)
• laundry-detergents (Sb 8/24 brands 0.04-0.47 mg/kg; total Cr 13/13 powder + 7/11 liquid; Mo 8/13 powder; Se 5/13 powder + 4/11 liquid)
• laundry-detergent-boosters-additives (soakers: 2 brands, total Cr 0.19-0.23 mg/kg, Mo and Se trace)
• fabric-softeners (Sb below detection in all 7 brands; Cr 4/7 brands max 3.50 mg/kg)
• automatic-dishwasher-detergents (Sb 2/5 powder brands 0.82 and 0.10 mg/kg; Cr 4/4 tablets + 4/5 powders; Mo 2/4 tablets + 2/4 powders)
• dish-soaps-manual (Sb below detection in all 6 brands; Cr 3/6; Co 3/6; Mo 2/6)
• rinse-aids (Sb below detection in the one brand tested)
• toilet-bowl-cleaners (Cr 7/7 fresheners 0.26-174 mg/kg; Mo 6/7; in-cistern category outlier)
• floor-cleaners-hard-surface (Sb 1/4 brands 0.07 mg/kg; Cr 3/4)
• all-purpose-cleaners (Cr in 1-2 of 2-4 general-purpose brands; Mo 2/4 toilet cleaner; Se 1/2 multi-purpose)
• household-specialty-cleaners-other (in-cistern fresheners: Cr 174 mg/kg outlier driving 35.8 mg/pe/wk Cr load)
• shampoo-adult (Cr 8/10 brands; Co 8/11; Mo 5/11; Se 3/11 incl 4.6 mg/kg anti-dandruff)
• conditioner-adult-rinse-out (Co 2/3 brands 0.28-0.39 mg/kg; Cr 2/3 brands 0.88-1.3 mg/kg)
• body-wash-shower-gel (Co 4/4 brands 0.35-4.9 mg/kg; Cr 3/4)
• hand-soap (Co 4/5; Cr 4/5 brands up to 6.2 mg/kg)
• bathing-soaps (Se 4/6 soap bars 0.28-0.39 mg/kg; Cr below detection in all 6)
• deodorants (Cr 3/9; Mo 4/9; Se 3/9 brands 0.12-0.62 mg/kg)
• sun-suntan-products (Cr 6/7 SPF30+ sunscreens 0.01-0.11 mg/kg; Co 1/7)
• body-hand-leave-on-skin-care (Cr 1/5 body lotion brands)
• face-neck-leave-on-skin-care (Cr 3/4 facial cream/lotion brands 0.002-0.35 mg/kg)
• facial-cleansers (Cr 2/2 brands 0.5-0.7 mg/kg)
• shaving-cream-gel-foam (Cr below detection in all 5 brands)
• toothpaste (Sb 1/9 brands 0.15 mg/kg; Cr 8/9 brands 0.30-1.21 mg/kg)
• mouthwash-oral-rinse (Cr 4/5 brands 0.01-0.7 mg/kg)
• permanent-hair-dyes-tints (Cr 3/6 brands 0.14-0.22 mg/kg; Mo below detection in all 6)
• hair-dye-rinse-off (same 6 brands as permanent dyes; the report does not differentiate the rinse-off and permanent sub-segments)
• hair-removal-products (depilatory cream 1 brand tested; Cr below detection)
• facial-cleansers (facial cleanser + facial scrub lumped under “Face cleanser” in source Tables 7, 9; Cr 2/2 at 0.5-0.7 mg/kg; the facial scrub brand is included in the 156-product count but its data are not separately tabulated)

Verification notes

• Brand-firewall handling (Part 12). The CSIRO report’s underlying appendix tables (Tables 19-26, pp. 70-85) list every analysed product by brand name alongside its per-element concentration. None of those brand-named cells is reproduced in this wiki page; concentrations are summarised by product category and by brand-class (private / market-leader / environmental-label) descriptor only. The “Leader / Enviro / Private” columns in the body tables above are the report’s own brand-class descriptors and do not name brands.
• Cr speciation gap for HMTc. The HMTc Category-13 (Cosmetics) and Category-1 (Heavy Metal Tested & Certified Cleaning) Cr panel cells require Cr-VI specifically. The Tjandraatmadja dataset reports only total Cr by ICP-MS; Cr-VI is not separated. The HMTc Cr-VI cells therefore cannot draw on this dataset directly; the data inform the umbrella metals/chromium page and the per-product total-Cr cells (which can be informative for context but are not the binding HMTc analyte).
• In-cistern Cr 174 mg/kg outlier. Section 4.1.4 reports concentrations in toilet fresheners ranged 0.26 to 174 mg/kg. The 174 mg/kg figure is the largest single Cr datum in the report and drives both the 35.8 mg/pe/wk weekly Cr load attribution (Figure 6) and the broader cistern-category Cr signal. The report does not identify whether this brand is a chromate-pigment cistern-staining product (the typical formulation explanation for very high Cr in cistern fresheners) or a different chemistry. For HMI synthesis the in-cistern brand should be treated as a category-specific datum, not pooled with general toilet-cleaner Cr values.
• Selenium-sulfide anti-dandruff datum (4.6 mg/kg). The single 4.6 mg/kg Se reading in an anti-dandruff shampoo is consistent with a low-residual selenium-sulfide active ingredient or trace impurity from a different anti-dandruff active. The report does not name the active ingredient or the brand and does not differentiate anti-dandruff from general-purpose shampoo subcategories in Table 12. The Se panel cell on shampoo-adult should note this outlier’s likely active-ingredient origin.
• No CRM / recovery / blank-correction data reported for the panel metals. The CSIRO report tabulates instrument detection limits (Table 4) but does not document spike-recovery percentages, certified reference material results, or method recovery against an inter-laboratory comparison for Sb / Co / Cr / Mo / Se. This is a B-tier (industry/government report with documented method but minimal QA reporting) rather than A-tier (peer-reviewed paper with CRM and inter-lab QA) limitation. The dataset is structurally a single-laboratory single-time-point characterisation.
• Frontmatter scope choice. The products: array enumerates the product slugs the source’s category-level data measure directly. Toilet paper has no current products/ slug and is therefore listed only in matrices: as toilet-paper. The matrices: field carries household-product-stream descriptors (laundry-detergent, dish-detergent, hard-surface-cleaner, fabric-softener, cosmetic-personal-care, shampoo, toothpaste, deodorant, sunscreen, toilet-paper, domestic-wastewater) consistent with corpus precedent for similar household-product papers (see ecetoc1992-tr045-ni-co-cr-consumer-products and abulude2007-soaps-detergents-akure-nigeria).
• Jurisdiction scope. Sampling was Melbourne, Victoria, Australia, in early 2008. The brand mix reflects the Australian retail market at that date and may not generalise to other markets. The dataset feeds the AU jurisdiction signal on each touched product category.
• No DOI assigned. CSIRO Water for a Healthy Country Flagship reports in the 1835-095X series do not carry DOIs; no_doi_assigned: true is set in frontmatter.
• License. Copyright Commonwealth of Australia 2009 (cover-verso copyright notice). The report is freely distributed by CSIRO via the National Research Flagships programme but is not under an open-content licence; downstream reuse limited to citation and category-level summary as practised here.
• Audit subagent (2026-06-03, Claude general-purpose) — three ⚠️ concerns applied; verdict REVISE; no ❌ definite errors.
  • (1) Source internal discrepancy on Sb in floor cleaner, toilet paper, and toothpaste. Auditor noted that p.22 narrative (“one brand of floor cleaner (0.016 mg/kg), one brand of toilet paper (0.069 mg/kg) and one brand of toothpaste (0.147 mg/kg)”) disagrees with Table 6 on the same page (toilet paper 0.02, floor cleaner 0.07, toothpaste 0.15). The toothpaste values round consistently (0.147 ≈ 0.15). The floor cleaner / toilet paper values are mutually transposed between narrative and table: narrative-floor-cleaner 0.016 rounds to 0.02 (matching Table 6’s toilet paper row), and narrative-toilet-paper 0.069 rounds to 0.07 (matching Table 6’s floor cleaner row). This is an internal inconsistency in the source report — either the narrative or Table 6 swapped the floor-cleaner/toilet-paper labels. The wiki page cites Table 6 values (0.02 for toilet paper, 0.07 for floor cleaner) because that is the published table the source’s downstream analyses (Figure 3 weekly loads) appear to rely on. Flagged here so downstream synthesis can apply a sensitivity check against the narrative interpretation if needed.
  • (2) exfoliants-scrubs slug removed from products: frontmatter and from the “Wiki pages this source may touch” section. Verified against source p.18 (§3.1.4): “1 brand of facial cleanser, 1 brand of facial scrub and 1 brand of depilatory cream” are listed as separate products in the methodology. Verified against Tables 7, 9 across the analyte panel: the source’s data tables list a “Face cleanser” row only (n=2/2 for Cr), which is the facial cleanser + facial scrub lumped. No per-analyte data for the facial scrub alone are reported. Since the routing layer would fan an exfoliants-scrubs slug to a destination with no source-specific data to populate, the slug was removed; the facial-scrub brand’s data are routed through facial-cleansers (the lumped “Face cleanser” row) instead, with a note in the wiki-pages-touched list documenting the lump.
  • (3) Implications-section language tightened. The phrase “places Australian household-product Sb at the upper-trace-impurity level rather than a deliberate-additive level” was interpretive characterisation slightly ahead of the source’s own framing. Rewritten to track the source’s own Section 4.1.1 conclusion that Sb is “a minor contaminant in household products”. Also removed the “(P97/P45 pooling)” parenthetical from the Sb-implications bullet because P97/P45 is HMTc-side workflow methodology that should not appear on a source page (Part 2 firewall hygiene). No data values changed.

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.