Improving Espresso with Paper Filters
Vis-à-vis Aeropress Filters meet Espresso Machines
In May of 2019, Scott Rao popularized an espresso technique for high extraction (>25% vs 18% to 22% range for a 3:1 shot) involving Aeropress filters (filtered espresso). At the time, I was deep into an exploration of the staccato espresso technique, and I didn’t want to add another variable and another step to my process. I put the idea on the list of things to try, and in December of 2019, I finally had some time. I didn’t want to employ this technique for only regular shots; I wanted to understand if it had the potential to improve staccato shots.
Scott found he could get a tasty shot with higher extraction due to a finer grind by employing Aeropress paper filters. The technique involved using a much finer grind and then putting a damp paper filter above and below the coffee puck. The top filter was supposed to mitigate issues with the shower by reducing channeling. The bottom filter would stop the finer grind from clogging the filter. Since a finer grind was used, a higher extraction was achievable. He then used a DE1+ to make a 3:1 (output:input) shot.
A Short History on Aeropress and Espresso
An espresso basket is made of metal to withstand the high pressure required for espresso to be produced. Drip coffee uses a paper filter because there isn’t a pressure requirement. French press has a filter, but the grounds are all fully submerged at the same time. When Aeropress came out, the technique seemed to be in-between a French press and an espresso machine.
The idea of cutting an Aeropress paper filter and putting it inside an espresso basket isn’t new to the world of coffee; Scott popularized the technique. Often in experimentation, multiple people come to the same conclusion independently. I’m not sure if this is the case, but as a researcher, I feel obliged to dig in and find what other works have been done in the topic. Looking back to 2011, the idea of using an Aeropress filter was introduced to help reduce Cafestol, but a further look at whether the espresso was good, another author found positive results.
In 2015, there was renewed interest in the topic, but it didn’t take off in popularity. Again, this was single Aeropress filter on the bottom of the basket. They found a faster flow rate and improved taste.
This guy actually did what Scott did with two filters a year before Scott published his result. So while it is easy to claim originality, the idea has been out there, but not quite popularized. Scott measured extraction which revealed the technique was able to extract at a higher rate. Then Barista Hustle had a nice write-up after Scott’s video went viral in the coffee community.
While these discussions are not peer-reviewed research, they point to the idea being in existence for some time. I was originally going to speak about the technique as the Rao Paper Filter (RPF) technique, but I decided RPF should refer only the specific technique Scott used rather than all paper filter techniques. However, after finding the 2018 video, RPF should be called Double Paper Filter (DPF) because it was only popularized by Scott but first discovered by someone else. For naming, I’d like to call all these techniques and variants Paper Filter in the Filter (PFF) as a shorthand because I did a few variants with staccato shots.
Paper Filter in the Filter (PFF) in Staccato Shots
I didn’t immediately try the paper filter method with staccato shots because I was already getting a pretty amazing shot. However, I added it to my list, and back in January of 2020, I started experimenting. I figured it had some potential, but I was surprised where the optimal placement was within the shot.
First, I had to find the optimal placement. I wanted to experiment so I started with the paper on the bottom. I figured I would try on the bottom, then the top, then both, and then get fun by adding it between layers. A quick review: my Staccato shots go Fine (<400um) on the bottom, Coarse (>500um) in the middle, and Mid (400um < Mid < 500um) on Top.
On the bottom showed some taste improvement and some extraction improvement, but I was already getting a big benefit from the staccato layering. The major finding was that putting the paper filter between the Fine layer and the Coarse layer made a big difference in taste. The filter on the top didn’t seem to affect the outcome in terms of taste. If anything, it had a negative affect on taste, so I focused on the filter on the bottom or above the fine layer.
The key is a damp filter: too wet causes the shot to be a little less rich. Too dry causes channeling because the process of the filter getting wet initially has some issues with surface tension.
I collected some paired data, but again, similar to pressure pulsing, I had intermixed other experiments and fell far short of being able to test for statistical significance. This is a qualitative study suggesting paper filters could improve staccato shots taste while maintaining or slightly decreasing extraction. The data for regular shots suggests a higher extraction and an improved taste.
The two metrics are: Final Score and Coffee Extraction.
Final score is the average of a scorecard of 7 metrics (Sharp, Rich, Syrup, Sweet, Sour, Bitter, and Aftertaste). These scores were subjective, of course, but they were calibrated to my tastes and helped me improve my shots. There is some variation in the scores. My aim was to be consistent for each metric, but some times the granularity was difficult.
Total Dissolved Solids (TDS) is measured using a refractometer, and this number combined with the output weight of the shot and the input weight of the coffee is used to determine the percentage of coffee extracted into the cup.
Data Summary:
The paper filters lead to a shorter shot that was better tasting, but less coffee was extracted. It seemed to cause the shot to blond earlier. This might have interfered with the measurement because it caused the shot to end sooner than the normal ratio. I’m aiming for the mouthfeel not necessarily a specific weight (usually a 1:1 is the right rato).
I enjoyed the taste so much that I added it to my usual staccato routine as well as my non-stacacto routine (Comparison video). I suspect the placement is due to the filter having an effect on what extracts are coming from the coarse and mid layers. I think the filter might take away from the oils, but it’s hard to confirm with just TDS. I’ve tried a filter on the bottom as well as above the fine, but it took away from the flavor.
To have a better look, I have compared two shots with and without the paper filter layer. Both of these shots are regular 18g shots in a 20g VST basket. The roast was a little darker so the basket was full at the 18g point.
The PFF shot comes out faster which allows one to get a finer grind. I theorize that it also helps reduce some of the channeling issues caused by the distribution of hole sizes across the filter.
Then things got weird.
I replaced the old metal shower screen in my Kim Express. I cleaned it well, and I figured I would keep it around. I tried putting it after the coarse layer, and it seemed to improve the taste. I did this test on a 7g VST filter overpacked to 15g, so the metal filter completely covered the puck below it because of the filter shape. I didn’t quite have an explanation. I’ve had trouble messing around with the metal filter that compared to the Aeropresso filters because the shower screen is not as wide as the whole filter causing some side channeling.
More Data: A Deeper Look at PFF Coffee Extraction
I collect data on every shot of espresso, so I pulled my last 6 months of data to help understand the differences. I have 281 shots where I collected extraction from either a Brix refractometer or Atago Digital refractometer (the difference between them is not statistically significant based on previous data). This data is not paired, but one can see overall how does PFF extraction does across a larger sample size.
I also collect two refractometer data points for each shot using two cups: Total Dissolved Solids (TDS) at the 1:1 shot and TDS at the 3:1. I don’t use a scale under my machine because there isn’t enough room, so there is a slight variance in my 1:1 ratio and my 3:1 ratio. I drink the 1:1, and I throw away the rest. As a result, I have 562 data points to look at: woohoo!
Let’s focus on the 1:1 shot because the majority of the extraction occurs there:
The data shows how I’ve increased the length of my shot in part due to PFF. For Staccato shots, it seems PFF takes away from extraction but filters out the less desirable components of the extraction.
An additional confuser is pre-infusion because PFF has driven me to use a longer pre-infusion. I initially did pre-infusion for 10 seconds, but I have pushed that to 30s in the past two weeks.
A trend is difficult to surmise, and a separate experiment would be beneficial. Add it to the list!
Curious question: If a long pre-infusion (at ~2 bars) gets such high extraction, why pull at 9 bars aside from getting the shot to come out faster?
One main conclusion I’ve found from my data related to PFF is that a 3:1 shot is unnecessary as most of the coffee extraction is occurring at the 1:1 or even towards the 1.5:1. I’m not convinced pulling longer is improving taste, so at that point, just water down the 1:1 shot to 3:1.
Additionally, these PFF experiments have had three on my shots:
- I incorporated a paper filter for my most of my shots both regular and staccato.
- I increased pre-infusion times in my ritual from 10 seconds to between 20 and 30 seconds.
- I used the paper filters as way to image the extraction of the shot (something for a future article).
If you like, follow me on Twitter and YouTube where I post videos of espresso shots on different machines and espresso related stuff. You can also find me on LinkedIn.
Further readings of mine:
Staccato Espresso: Leveling Up Espresso
Coffee Solubility in Espresso: An Initial Study
Espresso Simulation: First Steps in Computer Models
Pressure Pulsing for Better Espresso
The Tale of the Stolen Espresso Machine
Affordable Coffee Grinders: a Comparison
