Sign up to receive email updates, action alerts, health tips, promotions to support our work and more from EWG. You can opt-out at any time. [Privacy]


Greener School Cleaning Supplies

Research Links Air Quality to Cleaning Supplies

November 3, 2009

Greener School Cleaning Supplies: Study Methodology

School District Survey

We went straight to the source to determine which cleaning supplies are used in California public schools. We canvassed custodial and operations staff from many of the largest and most diverse school districts in the state, along with smaller districts in high-asthma areas. We asked for information on everything from window cleaners to floor finishes. We collected full or partial lists of cleaners in use, or approved for use, in 13 districts, including Bakersfield City, Fairfield-Suisun Unified, Fresno Unified, Jefferson Unified (Daly City), Los Angeles Unified, Oakland Unified, Sacramento City Unified, San Bernardino City Unified, San Francisco Unified, San Jose Unified, Stockton Unified, Visalia Unified, and West Contra Costa County Unified. Many school districts refused to provide this information without assurance that we would not reveal publicly which cleaning supplies they use.

EWG staff also reviewed the limited data on chemical composition and health and safety information available in the Material Safety Data Sheet (MSDS) that the manufacturers provide for each product.

Air Contamination Tests of Individual Cleaning Products

We conducted in-depth air quality testing of 21 representative cleaners used in schools in California to identify air contaminants emitted during product use. We selected the 21 products tested based on criteria that included a) product type; b) high volume and frequency of use; c) use in multiple districts; d) widely available brands; e) ingredients of concern listed in MSDS; f) total volatile organic compound data listed in MSDS; g) pine- or citrus-based cleaners. We tested multi-purpose cleaners, floor cleaners, glass cleaners, bathroom cleaners, disinfectants, floor treatments, a floor stripper, a graffiti remover, a metal cleaner, and an air freshener.

Tests were conducted by Air Quality Sciences (AQS, Marietta, GA;, an ISO 17025-accredited, 9001:2000-registered, and AIHA-accredited indoor air quality testing and research company. Cleaning products were tested in small stainless steel environmental chambers with volumes of 0.09 ± 0.005m3. Chambers designed and built by AQS meet construction specifications and performance requirements established by U.S. EPA guidelines and ASTM Standard D 5116-06, "Standard Guide for Small Scale Chamber Determination of Organic Emissions from Indoor Materials/Products," and ASTM Standard D 6670-01, "Standard Practice for Full-Scale Chamber Determination of VOCs from Indoor Materials/Products."

Application levels employed were consistent with those specified in the recent California Air Resources Board report on air quality concerns associated with cleaning supplies (Nazaroff 2006). For cleaners applied using spray bottles, approximately 11 g/m2 of product was tested; for liquid cleaners applied via pouring rather than spraying action, approximately 17 g/m2 of product was tested; for floor products, approximately 27 g/m2 of product was tested; for the air freshener, a single spray of product was tested; for the powder cleanser, 176 g/m2 of product was applied to the pre-wetted surface.

Products requiring dilution were diluted to levels appropriate for "medium" or "normal" cleaning strength, following manufacturers' instructions. An exception to this protocol was made for Simple Green, a product that is packaged in a spray bottle and is therefore commonly used at full-strength, rather than in the diluted form as suggested by the manufacturer. Another exception was made for Alpha HP Multi-Surface Cleaner: this product was diluted to be used as a non-certified disinfectant (dilution 1:64), rather than as a Green Seal-certified multi-purpose cleaner (dilution up to 1:256). Air contaminants produced by more dilute cleaning solutions may be estimated by dividing levels measured by the dilution ratio, then eliminating those chemicals with values below the detection limit. This method was used to estimate contaminants that would be released by Alpha HP at the Green Seal-certified, cleaner dilution level.

The environmental chambers used for these measurements were glove boxes, sealed chambers with attached gloves that allow for manipulation of the cleaning products and materials within. Use of glove boxes allowed simulation of the mechanical action of cleaning. Spraying and scrubbing can result in increased volatilization of chemicals, improving detection of those chemicals contaminating the air during and after real-world cleaning activity. All products were applied to the model surface in a manner consistent with manufacturers’ directions and actual school use information. Products remained untouched on the surface for one minute, or longer when dictated by manufacturer instructions. Paper towels were used to simulate scrubbing, wiping, and mopping action.

During testing, supply air to the chamber was stripped of volatile compounds, particles, and other contaminants, so that background levels present in the empty chamber were below strict limits (less than 10 micrograms/m3 for total volatile compounds or total particles, and less than two micrograms/m3 formaldehyde or any individual compound). Standard test conditions were 23 oC, 50 percent relative humidity, and one air change per hour.

For each product, three sets of samples were collected in the first hour after application in the chamber to identify which volatile compounds were released. The period during the collection of the first set of samples included both the sitting phase, the period of mechanical action associated with each product, and the 10 minutes following. The second set of samples was collected between 10 and 30 minutes following product application, and the third set was collected between 30 minutes and one hour after application. In contrast, to characterize chlorine and ammonia release, a single set of samples was collected over the full first hour after application. Measurements for all contaminants were used only to identify compounds, rather than to provide information on the specific levels of released chemicals.

Analysis of Aldehydes – Emissions of selected aldehydes including formaldehyde were measured following ASTM D 5197 and U.S. EPA IP-6A measurement by high performance liquid chromatography (HPLC). Solid sorbent cartridges with 2,4-dinitrophenylhydrazine (DNPH) were used to collect formaldehyde and other low molecular weight carbonyl compounds in chamber air. The DNPH reagent in the cartridge reacted with collected carbonyl compounds to form the stable hydrazone derivatives retained by the cartridges. These derivatives were then eluted using HPLC-grade acetonitrile. The sample was analyzed using reverse-phase HPLC liquid chromatography with UV detection. The absorbances of the derivatives were measured at 360 nm wavelength. The mass responses of the resulting peaks were determined using multi-point calibration curves prepared from standard solutions of the hydrazone derivatives. This method does not detect contaminant levels below five micrograms/m3. A level of certainty associated with the identity of each compound is provided. All compounds were identified with a certainty of at least 80 percent; some compounds were identified with a certainty equal to or greater than 99 percent.

Analysis of Volatile Organic Compounds (VOCs) – Measurements of volatile compounds were made using gas chromatography with mass spectroscopic detection (GC/MS). Chamber air was collected onto a solid sorbent, which was then thermally desorbed into the GC/MS. The sorbent collection technique, separation, and detection analysis methodology follows U.S. EPA Method IP-1B and ASTM D 6196 and is generally applicable to C6 – C16 organic chemicals with boiling points ranging from 35 oC to 250 oC. This method does not detect contaminant levels below five micrograms/m3. A level of certainty associated with the identity of each compound is provided.

Analysis of Ammonia & Chlorine – Measurements of ammonia and chlorine were made by Bureau Veritas North America, Inc. (Novi, MI). Ammonia was analyzed based on NIOSH S-347, using a 12.5 L air collection volume. This method does not detect contaminant levels below 1.1 mg/m3 (or 1.6 parts per million). Chlorine was analyzed based on NIOSH 6011, using an 18.7 L air collection volume. This method does not detect contaminant levels below 0.53 mg/m3 (or 0.18 parts per million).

Quality Assurance & Quality Control – AQS is an ISO 9001 registered and ISO 17025 accredited testing firm with defined and executed internal and third party verification programs encompassing emissions test methods and low-level pollutant measurements. To assure low levels of background contamination, supply air purity is monitored on a weekly basis, using identical methodology to the chamber testing. Precision of measurements of volatile compounds and aldehydes is assessed via relative standard deviation from duplicate samples. Accuracy of volatile compound measurements is based on the recovery of toluene mass spiked onto absorbent material. Accuracy of aldehyde measurements is based on Workplace Analysis Proficiency Scheme formaldehyde proficiency test results.

For each product tested, EWG has provided a list of the air contaminants detected in the Cleaning Supplies database, along with information on health concerns associated with these chemicals and with the limited number of ingredients mentioned by product manufacturers.

Air Contaminants Measured During Classroom Cleaning

For a complete, quantitative assessment of the contaminants released during typical classroom cleaning activities, a room-sized environmental chamber was fitted with classroom appliances and cleaned using three product types: a floor cleaner, a window cleaner, and a general-purpose cleaner for desks, whiteboards, and shelves. The classroom was cleaned twice -- first with conventional cleaning products, and then with certified green cleaning products -- allowing assessment of the relative amounts of pollution produced by conventional vs. green cleaning.

Product Type: Conventional Cleaning: Green Cleaning:
Floor Cleaner
  • Twister
  • (Champion Chemical)
  • Marauder
  • (Buckeye International)
Window Cleaner
  • 21 Glass Concentrate
  • (Waxie)
  • Glance Non-Ammoniated Glass & Multi-Purpose Cleaner
  • (JohnsonDiversey)
General-Purpose Cleaner
  • Simple Green
  • (Sunshine Makers)
  • Alpha HP Multi-Surface Cleaner
  • (JohnsonDiversey)

As the environmental chamber was smaller than a standard classroom, the amounts of floor, windows, desks, whiteboards, and shelves were scaled appropriately:

Measurement: Standard Classroom: Environmental Chamber:

Room Volume (m3)

231 31.2

Floor Area (m2)

89.2 12.05

Window Area (m2)

4.46 >0.60

Desk Top Area (m2)

12.3 1.66

Whiteboard Area (m2)

9.9 1.34

Bookcase/Shelving Area (m2)

7.81 1.05

Cleaning of the model classroom environment was approached with great care. The process began when a laboratory technician, wearing a Tyvek suit, latex gloves, and eye protection, entered the chamber with all cleaning supplies, including cleaning products, rope mops for the floors, and paper towels for the windows and other surfaces. The chamber door was immediately closed and the cleaning phase commenced.

Cleaning activities were designed to simulate typical or "medium soil" cleaning requirements, using product application levels consistent with recent research by the California Air Resources Board, as described above (Nazaroff 2006). This involved leaving products undisturbed on surfaces for brief time periods in some cases. The number of spray applications was optimized to achieve the desired product applications. Specific cleaning protocols used are detailed below.

Window Cleaning: six sprays of the product were applied to a 0.60 m2 piece of glass representing a window surface. The surface was left wet for one minute, then wiped with paper towels until clean.

Whiteboard Cleaning: 14 sprays of the product were applied to a 1.34 m2 section of powder-coated aluminum representing a whiteboard surface. The surface was left wet for one minute, then wiped with paper towels until clean.

Bookcase Cleaning: 11 sprays of the product were applied to a 1.05 m2 section of powder-coated cold rolled steel representing a shelving surface. The surface was left wet for one minute, then wiped with paper towels until clean.

Desk Top Cleaning: 18 sprays of the product were applied to a 12.3 m2 section of powder-coated cold rolled steel representing a desk top area. The surface was left wet for one minute, then wiped with paper towels until clean.

Floor Cleaning: The 12.05 m2 area of stainless steel flooring was separated into quadrants, and each quadrant was cleaned. A rope mop was submerged into the cleaning solution, tilted to drain off the excess solution, and spread over a quadrant using six mop strokes (back and forth). The mop was wrung and then used to soak up the remaining solution in 24 half-strokes with additional wringing after each four to six strokes. Once all four quadrants were cleaned, a final mopping of the floor was performed using a wrung mop to complete the dry-mopping phase.

Once the cleaning protocol was completed, the technician exited the chamber, removing all supplies. Testing resumed according to the sampling test protocol detailed below.

Quantitative measurements of aldehydes and volatile organic compounds were made covering five specific time periods, beginning with the cleaning procedure and extending for four subsequent hours:

  • During cleaning (~38 minutes), and for 10 minutes following cleaning
  • From 10 minutes to 30 minutes following cleaning
  • From 30 minutes to one hour following cleaning
  • During the second hour following cleaning
  • During the fourth hour following cleaning


Quantitative measurements of ammonia and chlorine were made using air collected during cleaning and for an hour following cleaning, as well as for the second hour following cleaning.

Background air samples were collected prior to application of the cleaners, but with the test materials (mop, bucket, and paper towels) in the chamber, to verify that cleaning materials were not responsible for the air contaminants detected.

Measurements of aldehydes, volatile organic compounds, ammonia, and chlorine proceeded as outlined above. However, in this case aldehyde measurements were quantifiable at or above a level of 0.1 micrograms based on a standard air collection volume of 45 L. Volatile organic compound measurements were quantifiable at a level at or above of 0.04 micrograms based on a standard air collection volume of 18 L. Detailed results are presented in an appendix.

Contaminants' Health Effects

EWG assessed health effects of air contaminants from definitive government, industry and academic sources, consistent with practices established under current green certification standards.

Air contaminants were linked to health effects in a manner largely consistent with current green certification standards. The Association of Occupational and Environmental Clinics (AOEC) has established a well-respected list of asthmagens, chemicals known to cause the development of asthma in previously healthy individuals (AOEC 2009). Green Seal-certified products must not contain ingredients identified as sensitizer-induced asthmagens by AOEC (Green Seal 2008). Those asthmagens not specifically identified as sensitizers by AOEC are not specifically excluded by this criterion.

Chemicals identified by EWG as linked to cancer include those listed as known, probable, reasonably anticipated, or possible human carcinogens by the International Agency for Research on Cancer (IARC; Groups 1, 2A, and 2B), National Toxicology Program (Groups 1 and 2), EPA Integrated Risk Information System (weight-of-evidence classifications A, B1, B2, C, carcinogenic, likely to be carcinogenic, and suggestive evidence of carcinogenicity or carcinogen potential), or by Occupational Safety and Health Administration (as carcinogens under 29 CFR 1910.1003(a)(1)). Green Seal-certified products must not contain these carcinogenic ingredients, or ingredients known to produce or release these compounds (Green Seal 2008). EcoLogo-certified products cannot be formulated or manufactured with any carcinogens listed by IARC in Groups 1, 2A, or 2B (EcoLogo 2008).

Reproductive toxins identified in this report are those chemicals listed as such (including developmental, female, and male toxins) by the State of California under the Safe Drinking Water and Toxic Enforcement Act of 1986 (California Code of Regulations, Title 22, Division 2, Subdivision 1, Chapter 3, Sections 1200, et. seq., also known as Proposition 65). Green Seal-certified products must not contain these reproductive toxins, or ingredients known to produce or release these compounds (Green Seal 2008).

Hormone disrupting chemicals identified in this report are those chemicals listed as priorities for research on these properties by the European Union in Appendix 9 of Towards the Establishment of a Priority List of Substances for Further Evaluation of Their Role in Endocrine Disruption (European Commission DG ENV 2000). EcoLogo-certified products cannot be formulated or manufactured with any potential hormone disruptors listed in Appendix 1 of this report, which lacks a handful of chemicals listed in Appendix 9 (EcoLogo 2008).

Neurotoxins identified in this report are those chemicals known to be neurotoxic to humans in a recent review by Grandjean and Landrigan (2006).

A variety of other health and toxicity concerns tied to cleaner chemicals are drawn from EWG’s Skin Deep database. Originally created to highlight concerns associated with ingredients in personal care products, Skin Deep is a core database of chemical hazards, regulatory status, and study availability created by pooling the data of more than 50 databases and sources from government agencies, industry panels, academic institutions, or other credible bodies (EWG 2009). Chemicals in the database receive an overall toxicity score from 0 to 10, with a score of 0 signifying few known health concerns, and 10 signifying many known health concerns.