Characterization of Naturally Occurring Alpha-Diketone Emissions and Exposures at a Coffee Roasting Facility and Associated Retail Café

Abstract

Background: Alpha-diketones such as diacetyl and 2,3-pentanedione have been used as artificial flavorings in a variety of industries and are produced naturally when food products such as coffee beans are roasted. Exposure to these compounds has been associated with bronchiolitis obliterans, a rare and severe respiratory disease. In the current paper, we (i) evaluate which steps in the coffee production process are associated with the highest alpha-diketone emissions at a small craft coffee roaster and associated café, (ii) determine the extent to which direct-reading measurements of CO, CO2, and total volatile organic compounds (VOCs) can serve as lower-cost surrogate indicators for diacetyl concentrations, and (iii) conduct a limited emissions study to quantify the effect that the process variable of roast type has on diacetyl emissions from grinding beans.

Methods: Exposure and area concentration data for diacetyl and 2,3-pentanedione were collected over 4 days of sampling at a single coffee roaster and associated café. Additional measurements of café patrons' exposure to diacetyl were collected in seven other craft roastery/cafes in Seattle, WA. For the emissions experiments, integrated area air samples for diacetyl were collected using sorbent tubes over 30-min intervals for each roast type with the sorbent tubes positioned next to a grinder placed in an exposure chamber. Sorbent tubes were analyzed for alpha-diketones using gas chromatography-mass spectrometry. A photoionization detector (PID) was also used to measure continuous total VOC concentrations at the coffee roastery, and during each grinding experiment.

Results: Diacetyl concentrations in five of the seven personal samples from the craft roastery were above the United States National Institute of Occupational Safety and Health (NIOSH) Recommended Exposure Limit (REL) of 5 ppb as an 8-h time-weighted average (TWA)-and one of the seven personal samples exceeded the NIOSH REL for 2,3-pentanedione-9.3 ppb as an 8-h TWA. Median diacetyl and 2,3-pentanedione emissions were highest at the bagging machine followed by the grinder, roaster, barista, and background areas. The arithmetic mean diacetyl concentrations from the seven personal samples collected from café patrons was 3.96 ppb, suggesting that diacetyl exposure poses a negligible health risk to café patrons. Correlations between diacetyl and total VOCs, CO, and CO2 showed that diacetyl was well correlated with total VOCs, but poorly correlated with CO and CO2. Based on our limited emissions study, French roast was associated with the highest mass emission factor of diacetyl.

Conclusions: Results from the exposure assessment study indicated that coffee production workers at this facility had elevated exposures to diacetyl and 2,3-pentanedione compared to recommended guidelines, whereas baristas and café patrons received lower exposures. Area sampling showed that the areas with the highest alpha-diketone emissions were the grinder and the bagging machine, which are both areas associated with tasks involving ground roasted coffee. Future research could focus on designing and evaluating effective engineering controls, in the form of local exhaust ventilation, with the goal of reducing alpha diketone exposures, as well as conducting similar studies at other small-scale craft coffee roasters and cafés to better understand the variability in these emissions and exposures within these types of facilities.

Keywords: 3-pentanedione; alpha-diketone; coffee grinding; coffee industry; coffee roasting; diacetyl; exposure monitoring 2; personal exposure.

Figure 1.

Box plot comparing diacetyl and 2,3-pentanedione concentrations by sampling location. The centerline in each box represents the median value, the lower and upper borders of the box represent the 75th and the 25th percentiles, and the whiskers extend to the largest or smallest value, at most 1.5 × IQR (interquartile range) beyond the box borders.

Figure 2.

Scatterplot showing a correlation between total volatile organic compounds and diacetyl air concentrations measured at all sampling locations.

Figure 3.

Example time-series analysis plot of the continuous concentration of TVOCs inside the chamber during one trial of the range-finding experiment. After 30 min passed from the start of grinding, over 90% of the emissions generated from grinding had evacuated the chamber

Figure 4.

Box plot comparing diacetyl mass emission factors for grinding of each roast type. The centerline in each box represents the median value, the lower and upper borders of the box represent the 75th and the 25th percentiles, and the whiskers extend to the largest or smallest value at most 1.5 × IQR (interquartile range) beyond the box borders. For each roast type, each experiment consisted of three trials.