Investigating the link between friction and astringency in coffee beverages
Kieran Nar and Eleanor Barrett
November 2024
Introduction
According to the British Coffee Association, coffee is the most popular drink worldwide, with approximately two billion cups consumed every day [1].
Coffee beverages come in many forms, from iced to hot, commodity to specialty, and options ranging from batch brews to concoctions like half caff, single shot, soy milk, caramel lattes. Though, regardless of personal preference, a wide array of factors must be considered before that first sip can be enjoyed. These factors can be broadly categorised into farming, roasting and brewing, with the latter often being the most controllable by the end consumer.
This work explores whether any correlations exist between the taste characteristics and tribological properties of different brewed coffees.
Background
As shown in Figure 1, a cup of filter coffee typically comprises of 98.5%-98.8% of water, and only 1.2%–1.5% of dissolved coffee matter [2].
Figure 1 – The typical composition of a brewed filter/batch coffee
The total dissolved solids (TDS) within a coffee sample can also be expressed as an extraction percentage (Ext%), by considering both the total beverage mass and coffee dose, as outlined in Equation 1 below:
Equation 1
Brewing control charts, developed by organisations such as the Specialty Coffee Association (SCAA), which has major divisions in the USA (SCAA) and Europe (SCAE), as well as the Norwegian Coffee Association (NCA), illustrate how TDS and Ext % correlate with taste. These charts are overlaid in Figure 2.
Figure 2 – NCAA, SCAA and SCAE brewing control charts.
In general, under-extracted, balanced and over-extracted coffees exhibit the taste qualities listed in Table 1 [3].
Table 1
The typical taste qualities exhibited by under-extracted, balanced and over-extracted coffees.
Ultimately, the combination and quantity of organic and chlorogenic acids extracted from a roasted, ground coffee will determine the flavour characteristics of that brewed beverage. However, beyond simply contributing to sourness alone, these acids have distinct sensory properties. Quinic and Lactic acids, in particular, contribute to the perceived astringency often associated with over-extracted coffees.
In similar works focussed on wine, astringency has been found to correlate with the tribological properties of the samples tested. This research aims to explore whether a similar relationship exists in coffee, specifically between Ext% and coefficient of friction (COF).
Materials and Methods
All tribology testing was performed using an MTM equipped with a low load beam (PCS Instruments). The MTM, pictured below, is a benchtop ball-on-disc tribometer capable of evaluating the frictional properties of a wide range of rolling and sliding contacts, whilst the low load beam enables loads of between 2N to 8N to be accurately applied during testing.
An MTM2 system.
Previous work by Wang et al. [4] was adapted for this study, leading to a test method comprising three consecutive stages: (1) dry sliding, (2) artificial saliva sliding and (3) artificial saliva + coffee sliding. The same PDMS ball and disc (50 Duro) were used throughout this study. An illustrated description of this test method can be seen in Figure 3.
Figure 3 – The three test stages performed in this study, specifically (a) dry sliding, (b) artificial saliva sliding and (c) artificial saliva + coffee sliding.
Prior to testing, PDMS ball and disc samples were ultrasonically cleaned in brewed coffee for five minutes at ambient temperature, then dried with compressed air. A Dry Mouth Oral Balance Gel (Biotene) was used as a substitute to human saliva and applied to the PDMS disc at 12 evenly distributed locations using the Grease Applicator (PCS Instruments).
The operating conditions employed during each stage of testing are detailed in Table 2.
Table 2
The operating conditions applied during each test stage.
Under-extracted, balanced and over-extracted samples were brewed from the same coffee. Details of the coffee used can be seen in Table 3.
Table 3
Details of the coffee used in this study, including those related to terroir, processing and roasting.
The brewing process was performed using a Hario Switch dripper via an immersion method, as illustrated in Figure 4. This brew method was chosen as it allowed TDS and Ext% to be primarily controlled by grind size and steep time, rather than pouring strategy, which a key variable during other brew methods, such as drip brewing.
Figure 4 – The three primary phases of the immersion brewing process performed to prepare each sample: (a) coffee addition, (b) steeping and (c) draw down.
Each coffee sample was brewed using a ratio of 12g of ground coffee to 200ml of water, specifically Ashbeck Still Natural Mineral Water (Tesco). As previously mentioned, and detailed in Table 4 below, grind size and steep time were varied to achieve different TDS and extraction levels. A Comandante C40 MK3 grinder, adjustable in equal increments of clicks (which correspond to the separation of its conical burrs), was used to grind each coffee immediately before brewing. The grind sizes selected were similar to those typically used for espresso, filter, and cafetiere brewing, ranging from relatively fine to relatively coarse, respectively.
Table 4
The parameters that were controlled to produce three different coffee samples.
After brewing and measuring the TDS of each coffee sample, they were divided into three glass containers (35ml in each) for triplicate testing and stored in a water bath at 70°C. This ensured that each sample could be added to the MTM at 65°C (stage 3) to closely match their preferred consumption temperature.
Results and Discussion
Coffee samples with varying TDS and extraction percentages were produced, as shown in Table 4; however, an upper extraction limit was reached when using the immersion brewing method outlined in Section 3. This extraction limit is a key advantage of immersion brewing, as the lack of agitation – from pouring, for example – inherently helps prevent the development of bitter taste characteristics.
The friction results from each timed step (stages 1, 2 and 3) can be seen in Figure 5.
Figure 5 – The COF results of each coffee sample across all test stages.
During PDMS ball on disc dry sliding (stage 1), high levels of boundary friction were exhibited.
Upon the addition of artificial saliva to each PDMS disc, coefficient of friction was reduced significantly, initially to below 0.1. These results were all similar and in good agreement with those reported by Wang et al. [4], thus highlighting the similarity in lubricating performance between the substitute and human saliva. However, it can also be seen that COF gradually increased over this test stage, likely due to the displacement of artificial saliva from the contact. In reality, mucins within human saliva bind with oral tissue surfaces in the mouth to form a polymer brush-type layer that provides sustained lubrication [5]. It is expected that using a Grease Scoop (PCS Instruments) would have helped maintain a greater supply of artificial saliva within the contact during testing.
As shown in Figure 5, upon adding each sample to the MTM in Stage 3, under-extracted, balanced and over-extracted coffees exhibited different frictional responses; with log curves fitted to aid their qualitative comparison. Each fitted curve was similar in shape but varied in magnitude. Balanced coffee samples incurred the least amount of friction, followed by under-extracted and over-extracted samples. Moreover, the largest differences in COF were observed during the first 400 seconds, after which under-extracted and over-extracted coffee samples plateaued to similar COF values for remainder of each test.
These COF results corresponded with the levels of oral harshness expected from each sample. More specifically, over-extracted, bitter and astringent coffee samples, which are generally considered to be the most unpleasant to consume, were found to have the highest friction responses. Under-extracted samples also yielded greater COF results than balanced coffee samples. As outlined in Table 1, these samples typically exhibit acidic taste characteristics, further supporting the correlation between friction and perceived oral harshness.
In beverages with higher TDS yields, such as espresso, where TDS values range from 8 to 12, these differences in friction are expected to be more pronounced, as has previously been observed for wine, where typical TDS values range between 1.5 and 3.
Additionally, TDS measurements provide a good indication of viscosity. However, no correlation between TDS, rheology and friction could be discerned in Figure 5.
Conclusions
In this investigation, the acidic and astringent taste qualities that dominate in under- and over- extracted brewed coffee beverages were found to correspond with their frictional responses.
Under-extracted, balanced and over-extracted coffees were produced via an immersion brewing process, resulting in samples with varying TDS and Ext% values. Triplicate testing was performed using an MTM system equipped with a low load beam. A test method comprising of three consecutive stages – (1) dry sliding, (2) artificial saliva sliding and (3) artificial saliva + coffee sliding – was performed.
The results from Stage 3 showed a correlation between taste and friction, relative to a balanced, ideal coffee. Elevated COF responses were linked to the perceived oral harshness experienced when consuming coffee beverages with pronounced acidic or astringent qualities.
This study demonstrated that a link between taste and tribology exists in coffee, and highlighted the potential of the MTM for food and beverage research.
References
[1] Coffee consumption (2021) British Coffee Association. Available at: https://britishcoffeeassociation.org/coffee-consumption/ (Accessed: 15 November 2024).
[2] (2023) Coffee TDS measurement with SmartRef digital refractometer. Available at: https://www.my-smartref.com/blogs/smartref-news/coffee-tds-measurement-smartref-coffee-refractometer (Accessed: 15 November 2024).
[3] (2021) Coffee extraction and how to taste it. Available at: https://www.baristahustle.com/coffee-extraction-and-how-to-taste-it/ (Accessed: 15 November 2024).
[4] Wang, S., Olarte Mantilla, S.M., Smith, P.A., Stokes, J.R. and Smyth, H.E. (2020), “Astringency Sub-Qualities Drying and Pucker are Driven by Tannin and pH—Insights from Sensory and Tribology of a Model Wine System,” Food Hydrocolloids, 109, 106109.
[5] Weston, A. et al. (2024) ‘Thirst and the influence of ionic concentration; an investigation into the effect on salivary lubrication and the role of MUC5B’, Surfaces and Interfaces, 54, p. 105183. doi:10.1016/j.surfin.2024.105183.