Dr. Christopher Hendon discusses the impact of water chemistry, freezing coffee, mathematical models, and artificial intelligence on cup quality.
Dr. Christopher Hendon’s current employer, the University of Oregon, touts him as “Dr. Coffee”. He’s an expert in coffee science, and his research into how water chemistry and grind temperature affect flavour have been used and discussed on stage at the World Barista Championship (WBC) many times.
But everyone starts somewhere, and Christopher tells BeanScene when coffee first appeared on his scientific radar, he wasn’t aware of its complexities.
“I was the typical customer who thought they knew something about coffee but really didn’t know anything,” Christopher says. “A lot of people have had some interaction with the science of coffee, even if it’s as rudimentary as knowing there’s a molecule in it that makes them feel energetic. But coffee’s actually pretty complicated, which makes it a playground for scientists.”
While he’s a Melbourne native, Christopher’s interest in coffee didn’t flourish until studying at the University of Bath in the United Kingdom. There, he befriended Maxwell Colonna-Dashwood of the local Colonna & Small’s café.
“I enjoyed learning about the coffee industry and Maxwell was interested in science, so we started to share our experiences and knowledge,” Christopher says. “We decided to combine the two topics and explore water chemistry. It was something that we could easily test and taste the difference of in the shop.”
Maxwell took what they learnt from these early studies and applied it in the 2014 WBC. Christopher also published their results in the 2014 journal article “The role of dissolved cations in coffee extraction”, and the two co-authored a book titled Water for Coffee in 2015.
“In a nutshell, coffee has finite acids which you can taste and are acted upon by various different minerals that exist in water. If you’re looking to taste all of these acids, you’re going to want to have a really low mineral composition. However, you probably will not enjoy the full force of less-developed coffee and will require some mineral composition to help moderate the flavour in the cup,” Christopher says.
“Water composition actually has quite a profound impact on coffee. It determines majorly what happens during the extraction, and afterwards, how you’re going to perceive that cup, those flavours, contained within the liquid.”
This means the same coffee brewed in regions with disparate levels of water hardness can have different tastes, which can be a positive and a negative.
“There’s lots of variables associated with water. For instance, roasters use it when they’re determining their own coffee’s quality, but there’s going to be a misalignment between what they and the end customers taste if a different water has been used to brew it,” Christopher says.
“On the other hand, one of the beauties of the whole industry is that, independent of where it’s roasted, there’s still a local terroir aspect to coffee, in the sense that where you brew it is going to impart some of the regional flavours into that cup.”
Now that Christopher understood the role of water composition, he looked into how to optimise the grinding process.
“Our experiments initially started with how grinding coarser or finer affects cup quality. Controlling the shape and size of the fine particles is extremely important for determining the cup quality, so we looked into ways to change their size and consistency,” Christopher says.
“We noticed that if roasted coffee was cooled below 0°C, when that coffee was passed through a grinder, the diameter of the fines became a more homogenous, predictable, unified size.”
He says this – combined with the benefits of prolonging the coffee’s freshness – helped pique the industry’s interest in frozen coffee. This paper was published in Scientific Reports (2016), “The effect of bean origin and temperature on grinding roasted coffee”.
With a better understanding of the role of water composition and grind size and temperature, Christopher set out to build a mathematical model that baristas could use to reproduce the same shot time after time while reducing wastage. Christopher will present this model, alongside the rest of his work, when he tours Australia for Toby’s Estate Coffee Roasters’ Knowledge Talks series in November 2019.
“Everyone in coffee has tasted a lot of espresso shots they really dislike. But when you find a tasty recipe, you would like the confidence that you can reproduce that exact shot hundreds of times,” Christopher says.
“Our goal with the model is to allow flavour reproducibility, not homogenisation. To do that, we need to understand what could give rise to changes in flavour for a given shot of espresso or filter coffee, some of the obvious ones being changes in water temperature, chemistry, pressure, grind profile, and dose, for example. These are the intuitive parameters that baristas navigate every day.”
One of the key values that needed further experimentation was extraction time. To figure this out, Christopher ran the model and compared it to espresso shots brewed in a lab setting by Studio Ali Manager Michael Cameron.
During this experimentation, Christopher was able to identify that a degree of microscopic channelling – water finding the path of least resistance through the puck – would occur when too fine a grind was used, causing different sections of the puck to under- or over-extract. Due to this, switching to a slightly coarser grind at a slightly lower dosage led to extremely high extractions, running in much quicker time. Christopher says this can potentially reduce a café’s coffee wastage by up to 25 per cent.
“The critical finding in this is – with certain parameters set – the barista can use a low dose of coffee, around 15 grams ground relatively fine, and the shot can run as fast as seven seconds, but still taste phenomenal, and the extraction yields are extremely high with very low shot-to-shot variability,” Christopher says.
He acknowledges that reducing coffee usage in consuming countries has potential negative implications for producers. However, he assures BeanScene that he also runs projects aimed at maximising value of coffee waste products at origin, and that increasing reproducibility in a café only serves to strengthen the industry and consumption as a result.
While the mathematical model intends to help baristas identify and reproduce the ideal espresso recipe post-brewing, Christopher and his team have now turned their attention to predicting the flavours of green and roasted beans.
“It would be powerful to be able to predict whether a particular Ethiopian coffee may have that rare pineapple flavour. Moreover, is this flavour coming from a molecule that we could detect prior to roast?” Christopher asks.
“Once we have that, we’ll be able to more accurately understand the emergence of flavours, without having to roast the coffee. Buyers could use this to look for green coffees that contain good potential, and then it’s up to the roaster to not ruin it.”
To do this, Christopher’s team is employing machine learning, a subset of artificial intelligence (AI), using large databases of molecules to train a computer to “taste” and predict the flavours of chemically complex green coffee.
“We need AI for this because, from a chemical perspective, it’s not clear what piece of the molecule – the structure, composition, or our receptors – gives rise to the flavour or smell,” Christopher says.
“It’s difficult to predict the potential for this technology in the coffee industry. I can imagine a system that would be able to perceive what is in the cup, and then be able to self-dial in its own coffee based on flavour profiles made using chemical analysis in real time. So there’s potential for automation, but, at least in our studies, the human in the café is not as much of a variable as people think.”
Christopher will tour Australia for Knowledge Talks from 7 to 12 November. Click HERE for details.
This article appears in FULL in the October 2019 edition of BeanScene. Subscribe HERE.
For more information, visit www.tobysestate.com.au