Characterization of Naturally Occurring Alpha Diketone Emissions and Exposures at a Coffee Roasting Facility and Associated Cafe
Abstract No:
1700
Abstract Type:
Student Poster
Authors:
H Echt1, C Simpson2
Institutions:
1University of Washington, Seattle, WA, 2University of Washington, School of Public Health, Seattle, WA
Presenter:
Hannah Echt
University of Washington
University of Washington
Faculty Advisor:
Christopher Simpson
University of Washington, School of Public Health
University of Washington, School of Public Health
Description:
Over the last 20 years, there has been growing concern about the detrimental effects on respiratory health associated with occupational exposure to alpha diketones, diacetyl and 2,3-pentanedione in particular. Diacetyl is a volatile organic compound that is widely used as an additive in the flavoring and food manufacturing industry. Fermentation and pyrolysis of food products also generate diacetyl exposures, such as in the manufacture of roasted coffee. Occupational exposure to diacetyl and other similar alpha-diketones (e.g. 2,3-pentanedione) has been associated with obliterative bronchiolitis, a rare, severe respiratory disease colloquially known as "popcorn lung." While NIOSH has proposed recommended and short-term exposure limits for diacetyl and 2,3-pentanedione, OSHA has not yet set standards for the regulation of worker exposure to diacetyl or other alpha-diketones. Previous studies conducted at large- and small-scale coffee roasters have shown that production workers have been exposed to diacetyl and 2,3-pentanedione at levels above NIOSH's REL of 5 ppb and 9.3 ppb, respectively. In 2013, MMWR released a report recognizing two cases of obliterative bronchiolitis in employees who worked in both flavored and unflavored coffee processing areas at a coffee-processing facility in Texas. There is evidence from previous studies that the highest personal exposures to and area emissions of diacetyl and 2,3-pentanedione are associated with grinding and packaging ground coffee. Prior to roasting, green coffee beans contain little to no alpha diketones. The roasting process produces is what produces alpha diketones, and grinding roasted beans increases the surface area for off gassing of these chemicals.
Situation/Problem:
There is a need for more research on how widespread this exposure risk is and what process elements are associated with the highest levels of exposures. Production practices are widely variable throughout the coffee industry and levels of diacetyl in smaller craft coffee roasters or coffee shops where roasting, grinding, and brewing take place are largely unknown. This research aims to fill the gap in current knowledge by examining exposures to diacetyl and 2,3-pentaendione among workers in a small-scale coffee production facility and associated cafe. The aims of this research were to determine which part of the production process is associated with the highest mean alpha diketone emissions by measuring emissions produced during roasting, grinding, packaging, and brewing, determine the extent to which direct reading measurements of CO, CO2, and total volatile organic compounds can serve as lower-cost surrogate indicators for diacetyl concentrations in air, and to conduct a limited survey to quantify the effect that the process variable of roast type (i.e. light, medium, and dark roast) has on alpha diketone emissions produced from grinding coffee.
Methods:
To address the first two aims of this research, integrated personal and area air sampling using modified OSHA methods 1013/1016 was conducted over four days at a single small, craft coffee roaster and its associated, onsite retail cafe. Personal air samples were collected on workers who operated in the coffee roasting, grinding, and packaging areas, but not on baristas in the cafe. Area air samples were collected in different work areas in the roaster and cafe, including behind cafe shelves, on the barista counter adjacent to the espresso grinder, above the cooling tray of the roaster, above the grinder, and adjacent to the facility's weigh/fill machine. Photoionization devices (PIDs) were deployed to measure continuous total VOC concentrations at the roasting, grinding, and packaging stations and were positioned next to the machines used in each production station. Similarly, Q-Traks were used to measure continuous CO and CO2 concentrations at the roasting, grinding, and packaging stations. Direct reading instruments and area air samples were col-located in each of the areas sampled. Daily task observation forms were completed for both the production area and the cafe.
A series of emissions experiments involving grinding roasted coffee were conducted for a variety of roast types, including dark roast, medium roast, French roast (i.e. medium-dark roast), and white coffee (coffee that is underroasted to produce a higher caffeine content). Short-term, half hour air samples following a modified OSHA method 1013 were conducted with air samples placed adjacent to the grinder's spout. a photoionization device was co-located to the air samples.
Analysis of alpha-diketone samples followed OSHA methods 1013/1016 with some modifications. Gas chromatography/mass-spectrometry was used to enhance the sensitivity of these OSHA methods. Diacetyl and 2,3-pentanedione concentrations by volume (ppb) were determined for each air sample.
For the data collected at the coffee roaster, time series analyses of CO, CO2, and total VOC emissions were measured at each sample location. 8-hr TWA diacetyl and 2,3-pentanedione personal exposures for each day of sampling were calculated and compared to NIOSH recommended guidelines. Average diacetyl and 2,3-pentanedione emissions were calculated for each sampling locaation. Simple linear regression models were used to determine which direct reading measurement was the best predictor of diacetyl emissions. For each of the different coffee roasts ground in the emissions experiments, emissions rates for diacetyl were calculated per pound of coffee ground.
A series of emissions experiments involving grinding roasted coffee were conducted for a variety of roast types, including dark roast, medium roast, French roast (i.e. medium-dark roast), and white coffee (coffee that is underroasted to produce a higher caffeine content). Short-term, half hour air samples following a modified OSHA method 1013 were conducted with air samples placed adjacent to the grinder's spout. a photoionization device was co-located to the air samples.
Analysis of alpha-diketone samples followed OSHA methods 1013/1016 with some modifications. Gas chromatography/mass-spectrometry was used to enhance the sensitivity of these OSHA methods. Diacetyl and 2,3-pentanedione concentrations by volume (ppb) were determined for each air sample.
For the data collected at the coffee roaster, time series analyses of CO, CO2, and total VOC emissions were measured at each sample location. 8-hr TWA diacetyl and 2,3-pentanedione personal exposures for each day of sampling were calculated and compared to NIOSH recommended guidelines. Average diacetyl and 2,3-pentanedione emissions were calculated for each sampling locaation. Simple linear regression models were used to determine which direct reading measurement was the best predictor of diacetyl emissions. For each of the different coffee roasts ground in the emissions experiments, emissions rates for diacetyl were calculated per pound of coffee ground.
Results / Conclusions:
Five of seven personal full-shift air samples exceeded the NIOSH REL for diacetyl. The highest concentration of diacetyl measured as an average across a full-shift was 21.0 ppb. One of seven personal full-shift air samples exceeded the NIOSH REL for 2,3-pentanedione with a concentration of 11.4 ppb. On days where workers were doing more grinding and packaging ground beans, personal exposures to diacetyl and 2,3-pentanedione were higher compared to days where workers were doing less grinding and packaged whole and/or ground beans. Diacetyl and 2,3-pentanedione area emissions were highest at the weigh/fill machine (median: 172 ppb, 5th - 95th percentile = 10.6 - 315 ppb), followed by the grinder (median: 57.3 ppb, 5th - 95th percentile = 7.0 - 105 ppb), the roaster (median: 7.5 ppb, 5th - 95th percentile = 1.2 - 39.1), the barista counter (median: 3.1 ppb, 5th - 95th percentile = 0.28 - 9.1 ppb), and the background area (median: 2.0, 5th - 95th percentile = 0.56 - 2.9 ppb). Area measurements of diacetyl and 2,3-pentanedione collected near the baristas and in the cafe were consistently below NIOSH recommended guidelines, indicating that exposure of cafe workers and customers at this facility were below levels of concern. Similar to previous studies, these data suggest that processing tasks associated with ground coffee are the biggest contributors to workers' exposure to alpha diketones.
Simple linear regression models were created to determine the extent to which direct reading measurements (i.e. total volatile organic compounds (TVOCs), CO, and CO2) could serve as lower cost indicators for diacetyl emissions, as sampling for these compounds is expensive and time-consuming. The model predicting diacetyl emissions from co-located TVOC emissions (measured by PIDs) had both the highest R-squared value (R-squared = 0.95) and the lowest mean squared prediction error. Moreover, when this relationship was broken down by sampling location, this relationship was the same for the two areas with the highest emissions - the weigh/fill station and the grinder.
Samples at this facility were purposely taken during the facility's busy season of production and on days that they set aside for roasting. The workers are not doing this level of production every day of the week or every week of the year, and so these results represent a worst-case emissions and exposures scenario. Time and budget constraints resulted in only 4 full sampling days at this single facility. This resulted in small sample sizes. Production processes vary throughout the coffee industry, so it is possible that this facility is not representative of other coffee processing facilities.
Results of this sampling campaign were reported and discussed with the facility's owner. We are currently designing local exhaust ventilation to reduce alpha-diketone emissions produced during grinding and packaging beans. The final product will be gifted to the facility to help reduce their exposures.
Simple linear regression models were created to determine the extent to which direct reading measurements (i.e. total volatile organic compounds (TVOCs), CO, and CO2) could serve as lower cost indicators for diacetyl emissions, as sampling for these compounds is expensive and time-consuming. The model predicting diacetyl emissions from co-located TVOC emissions (measured by PIDs) had both the highest R-squared value (R-squared = 0.95) and the lowest mean squared prediction error. Moreover, when this relationship was broken down by sampling location, this relationship was the same for the two areas with the highest emissions - the weigh/fill station and the grinder.
Samples at this facility were purposely taken during the facility's busy season of production and on days that they set aside for roasting. The workers are not doing this level of production every day of the week or every week of the year, and so these results represent a worst-case emissions and exposures scenario. Time and budget constraints resulted in only 4 full sampling days at this single facility. This resulted in small sample sizes. Production processes vary throughout the coffee industry, so it is possible that this facility is not representative of other coffee processing facilities.
Results of this sampling campaign were reported and discussed with the facility's owner. We are currently designing local exhaust ventilation to reduce alpha-diketone emissions produced during grinding and packaging beans. The final product will be gifted to the facility to help reduce their exposures.
Primary Topic:
Sampling and Analysis
Secondary Topics:
IH Profession
Risk Assessment and Management
Co-Authors
Please add your co-authors below. Co-authors are listed for professional courtesy and will not be communicated with regarding the decision notification or any on-site logistics, if accepted. Only the primary presenter listed is expected to attend and present the content on-site.
Acknowledgements and References
List any additional people who worked on the project or provided guidance and support along with details on the role they played in the research. (Please include first name, last name, organization, city, state and country).
Marc Beaudreau (UW DEOHS FRCG, Seattle, WA, US) provided technical support with sampling equipment and during emissions experiments.
Nancy Beaudet (UW DEOHS FRCG, Seattle, WA, US) provided support and guidance during field sampling.
Marty Cohen (UW DEOHS FRCG, Seattle, WA, US) provided support and guidance during emissions experiments.
Mariah Dittmore (UW DEOHS, Seattle, WA, US) provided support with data collection and entry.
Mae Coker (UW DEOHS, Seattle, WA, US) provided support with data collection and entry.
I'd like to thank the Coffee Roasting and Processing Facility Owner and Staff for their participation and willingness to collaborate on this research project.
Washington State's Department of Labor and Industries' SHIP Grant Award No. 2018YH00366 provided funding for this project.
NIOSH Training Grant No. T42OH008433-15 provided funding for my education.