Methods

The particle size distributions of each grind were determined according to ASTM C136/C136M-14 Standard Test Method for Sieve Analysis for Fine and Coarse Aggregates procedure. Samples of 100.0 g of coffee grinds were added to a sieve stack including sieve sizes #20 (0.841 mm mesh opening), #40 (0.420 mm mesh opening), #100 mesh (0.149 mm mesh opening), and pan, to generate grain size distributions for each coffee used in this study.

The cold brewing process was carried out at room temperature (ranging from 21 °C to 25 °C over the experimental period) adapted from a home-brewing recipe suggested by The New York Times’ Cooking website⁹. A sample of 35.0 g of coffee was placed in 350 mL of carbon-filtered municipal water. The coffee was contained in T-Sac™ tea filter bag (size 4) and placed in a 32-ounce Mason jar fitted with a screw-top lid. The filter bag was used to reduce grind loss during sampling and ensure grinds remained submerged during steeping. The coffee/water mixture was sampled every 15 minutes for the first hour, then every 30 minutes until hour 7, and then once an hour until hour 12, a final sample was taken at 24 hours. Samples collected after the first hour were diluted (1:4) with DI water and filtered using HT Tuffryn (Pall) 25 mm diameter, 0.2 μm pore size membranes. Fresh water was added to replace the volume sampled to maintain constant volume. This introduced a small dilution effect in the resulting solution. Additionally, even with the closed system, there was inevitably evaporation over the 24 hour testing period. Coffee received from Kona Joe Coffee was not processed in any way prior to use, to best match home-brewing conditions. Data presented are an average of triplicate experiments analyzed in duplicates (n = 6).

Hot brew extraction was conducted using the same coffee to water ratio as was used in the cold brew method. The water was heated to 98 °C and added to coffee grounds in a traditional French press carafe. The water and grounds were allowed to sit for 6 minutes before the filter was depressed and the coffee decanted. Since additional experiments showed that longer mixing times did not result in additional caffeine or 3-CGA extraction, 6-minute extraction times were used for all hot brew experiments. Two samples were taken from each hot brew and each experiment was performed in triplicate (n = 6).

Caffeine and 3-CGA were measured in both standard solutions and coffee extracts using an adapted methodology reported in GL Sciences Technical Note No. 67⁵³. An Agilent 1200 Series high performance liquid chromatography system (HPLC) was fitted with a Supelco 5 µm column (15 cm × 4.6 cm) (Supleco, Bellefonte, PA) run at 40.0 °C with a mobile phase mixture of 75% mobile phase A and 25% mobile phase B (A: 95% 2.0 mM phosphoric acid and 5% methanol; B: 95% methanol and 5% 2.0 mM phosphoric acid). The flow rate was 1.0 mL/min with an injection volume of 10.0 µL. Caffeine and 3-CGA were detected using a diode array detector at 280 nm and 325 nm respectively.

The pH of each brewed coffee sample was measured with a Mettler Toledo FiveEasy^TM F20 benchtop pH/mV meter.

Two-tailed t-test and ANOVA were employed for determination of similarities in equilibrium concentrations of 3-CGA and caffeine with consideration of the roast, grind size, and brewing method. The output of the statistical analysis is included in the supporting information.

All data generated or analyzed during this study are included in this published article (and its Supplementary Information files).