Evaluation of the carcinogenicity of gallium arsenide

Abstract

Gallium arsenide (GaAs) is an important semiconductor material. In 2-year inhalation studies, GaAs increased the incidence of lung tumors in female rats, but not in male rats or male and female mice. Alveolar proteinosis followed by chronic active inflammation was the predominant non-neoplastic pulmonary findings. IARC classified GaAs as carcinogenic to humans (group 1) based on the assumption that As and Ga ions are bioavailable. The European Chemical Agency Risk Assessment Committee concluded that GaAs should be classified into Carcinogenicity Category 1B (presumed to have carcinogenic potential for humans; ECHA). We evaluate whether these classifications are justified. Physico-chemical properties of GaAs particles and the degree of mechanical treatment are critical in this evaluation. The available data on mode of action (MOA), genotoxicity and bioavailability do not support the contribution of As or Ga ions to the lung tumors in female rats. Most toxicological studies utilized small particles produced by strong mechanical treatment, destroying the crystalline structure. The resulting amorphous GaAs is not relevant to crystalline GaAs at production and processing sites. The likely tumorigenic MOA is lung toxicity related to particulate-induced inflammation and increased proliferation. It is concluded that there is no evidence for a primary carcinogenic effect of GaAs.

• Alveolar proteinosis
• bioavailability
• chronic lung inflammation
• gallium arsenide
• semiconductor
• solubility
• threshold of arsenic carcinogenicity

Declaration of interests

The authors declare that this article was produced as part of consulting for the company Freiberger Compound Materials (FCM), Freiberg, Germany and the Working Group IMAT of the VDMA, Frankfurt, Germany.

The affiliation of the authors is as shown on the cover page and includes university and independent toxicology consultants. REACh ChemConsult GmbH, CinTox and Schenk Industry Consulting are consulting firms that provide scientific analysis and advice for both private and public organizations in areas that include (eco)toxicology and risk assessment; New York Medical College and University of Nebraska Medical Center are providers of medical education. Drs S. Cohen and G. Williams were, however, engaged as private consultants. The preparation of this review was supported by the company Freiberger Compound Materials (FCM), Freiberg, Germany and the Working Group Innovative Materials (IMAT) of the Verband Deutscher Maschinen- und Anlagenbau e.V. Frankfurt/Main.

Freiberger Compound Materials is producing and marketing GaAs and is lead company in the European REACH Regulation of gallium arsenide. The Working Group Innovative Materials (IMAT) was formed to initiate, support and coordinate research on the health effects of new and existing materials, such as semiconductors, in the area of the optoelectronic industry. Each of the authors was a consultant to FCM/the Working Group IMAT. The FCM company/IMAT working group was not given the opportunity to review and comment on the paper. The analysis and conclusions drawn are the exclusive professional judgement of the authors and do not necessarily represent the views of the sponsors.

Notes

¹Mean ± Standard Deviation. Calculated from data and statistical errors as provided in the study by means of maximum likelyhood estimation.

²In fact NTP evaluation of clearence rates can be refined with this finding yielding similar or better Χ² values.

³This fact allows comparison of the chamber control steady-state arsenic concentration in the blood with the dietary ingestion of As: Assuming 40–50% bioavailability of the As from nutrition, a daily food consumption of 70 g/kg bw and a blood volume of 60 ml/kg bw leads to a deposition rate of 0.24–0.30 µg/g per day. The clearence rate can be estimated from the F344 erythrocyte half-life of 66.3 ± 1.6 days (Derelanko, 1987). The resulting steady-state arsenic concentration is about 23–28 µg/g. This seems in line with the NTP findings.