1. Introduction

 

Throughout the recorded history of alchemy and distillation, many interesting still designs have been suggested by the renowned philosophers, doctors, and chemists of their time. These scientists sought to improve the efficiency of the apparatuses used for essential oils, mineral acids, and alcohol with many creative designs that bear little resemblance to modern stills.

The earliest alembic stills are dated to around 1,850 BC from Cyprus. It is because of the long dedicated career of Professor Maria Rosaria Belgiorno that the ancient stills found there have undergone archaeological reconstruction and proven to work well for creating perfumes. An unusual still design appeared centuries later in 4 CE, where Zosimos of Panopolis provides an account of Mary the Jewess and her interesting dibikos and tribikos stills dating back to 1 CE. Her legacy lives on to present day in the bain-Marie double boiler bath. Her multi-receiver still designs will be the subject of a future research project on our part to observe ethanol distillation.

It is the period of roughly the 16th through 18th centuries in Europe that provide an abundance of esoteric still and condenser sketches. Giambattista della Porto provided designs inspired by mythology and animals. Condensers shaped as cones, zigzags, and right-angled keys appear in the scientific literature in France and Germany at this time. Alchemical recipes for ardent spirits and “waters of life” are compiled by the likes of John French in London. It is from this period that we have focused our research and recreations from manuscript drawings. It would seem that, over time, these ancient designs have been dismissed as conceptual thinking, too fanciful, or too complex. The purpose of this research project is to reconstruct these unusual stills and condensers using modern borosilicate glass, temperature monitoring, and analytical tools, and observe their performance in the context of consumable alcohol distillation.

2. Methods of Reconstruction

We (Mohawk Spirits Distillery, under the direction of Dr. Eric M. Stroud) have selected five air-cooled esoteric designs from the following scientists for our reconstruction study:

1. Hieronymus Brunschwig (c 1500) – “Twins” circumambulatory still

2. Adam Lonicer (c 1555) – Key shaped apparatus

3. Nicholas Le Fevre – (1662) – Apparatus with “zigzag” condenser and Moor’s head

4. Herman Boerhaave (1693) – Tall conical apparatus

5. Johann Conrad Barkhausen (c 1700) – Twisted helix apparatus

All apparatuses were reconstructed from sketches of their respective authors. Little to no dimensions are provided in the ancient sketches, therefore, reconstruction followed the geometry and scaling as closely as possible. As all of the reconstructed apparatuses are fully functional stills in the United States, they were assigned serial numbers and reported as an equipment amendment to the United States Alcohol and Tobacco Tax and Trade Bureau (TTB) in November 2020 under Mohawk Spirits Distillery, LLC. All apparatuses were constructed by our glassblowing partner, Q-Glass (Towaco, New Jersey). Borosilicate glass was used in all reconstructions.

A. Methods of Characterization

 

We intentionally performed our distillation runs very slowly with careful temperature control, with some lasting more than 24 hours. This time was given to ensure that the air-cooled condensers of these ancient designs were not overheated, and that distillate was collected as close to room temperature as possible.

Three characterization distillation runs were performed for each still. The first characterization run used a binary solution of ethanol and water (neutral spirits) at 40% absolute volume (%ABV). This run determined how well each still would separate the ethanol from water up to the theoretical azeotrope. Fractions collected periodically during the distillation run were analyzed for % ethanol by volume using an Anton-Paar Snap 41 digital alcohol hydrometer.

The second characterization run used a ternary solution of a 40% v/v aqueous solution of ethanol spiked with (1% v/v) trans-cinnamaldehyde (B. P. 248°C). This run provided data on how well the “tails” or high-boiling point congeners of a wash are separated by the stills. Since cinnamaldehyde is ready detected by taste and smell, each fraction was evaluated by organoleptic methods. Cinnamaldehyde has an odor threshold of 50-750ppb and a flavor threshold of around 1000ppb. The organoleptic score was applied as follows:

l 0 – Not detectable

l 1 – Faintly detectable, near limit of threshold

l 2 – Detectable and balanced with ethanol

l 3 – Strongly detectable, dominant

The third characterization run used a 10% ABV wash produced from New York Finger Lake grape marcs to produce a marc brandy (grappa) as a single-pass distillation. This run was only used to produce consumable samples for third parties to evaluate by taste and odor.

B. Exclusions and Limitations

Our original research proposal specified the use gas chromatography and time-of-flight mass spectrometry to analyze each fraction from the distillation runs. Because of the 2020 Wuhan Coronavirus pandemic, our access to our partner instrument laboratory was lost for much of our project time. As of November 2020, laboratory access is slowly resuming, and we plan to follow up in this regard in 2021.

3. Barchusen Apparatus

Johann Conrad Barkhausen (Barchusen in the Dutch) was a chemistry teacher in Utrecht around the year 1694. He ascended to professor at the University of Utrecht in 1704, and held that post until his death in 1723. His apparatuses for distillation are very similar to Charas and other contemporary French chemists. Glassblowing and glass materials by this time were advanced, but the distillation designs were not particularly revolutionary amongst the chemists of this period. The twisted helix condenser used by Barchusen is remarkable because of its unusual shape (Fig. 1). It is one single long glass tube formed into this shape, and it is an air-cooled design.

clip_image004 clip_image006
Figure 1. Sketch of the Barchusen Apparatus, adapted from Forbes, 1948. Figure 2. The author with a modern-day borosilicate Barchusen Apparatus attached to a 3000mL boiler. The black rectangles on the tubing are liquid crystal temperature indicators.

A. Construction

A modern-day Barchusen apparatus was reconstructed based on the above sketch (Fig. 2). The bottom fitting is a 29/40 standard taper ground glass joint, adapting it to a standard 3000mL round bottom with a digitally-controlled heating mantle. The receiver is positioned below a dip point at the end of the condenser. In the ancient design, there are a total of 16 bends, but we reconstructed 14 due to the difficultly of the glassblowing. The first seven bends in the conduit are for the hot vapor to make its way to the apex, which we have termed the “reflux side”. These are followed by another seven bends for the distillate to fall by gravity, the “condenser side”. There is no mixing at intersection points because the conduit at each intersection is thermally insulated with glass standoffs (Fig. 3). There is no liquid coolant on this ancient column. Like the designs from Chagas, LeFevre, and Porto, we suspect this may have been operated outdoors, or in a cool cellar, or in a cold season.

clip_image008 clip_image010
Figure 3. Overlap of the path at intersections showing no junction. Figure 4. Liquid crystal temperature indicators along the vapor path

B. Dimensions

The reconstructed apparatus dimensions are as follows:

l Apparatus height = 30.5” (77.5cm) from base to apex of twist

l Loop width = 5.5” (14cm)

l Conduit diameter = 0.5” (1.3cm) outside diameter

l Total path length = 120″ (305cm)

C. Temperature monitoring and control

Adhesive liquid crystal temperature displays were applied along the conduit to monitor run temperature. These indicators (Hallcrest #4557D 0.5″ x 1.77″) provide a temperature indication of 60-90°C in 5°C increments. Three indicators were placed on the reflux side and three were place on the condenser side between the bends (Fig. 4). A 110VAC digitally controlled electrical heating and stirring mantle with a programmable set point was used for heating. A digital display on the mantle provided a real time reading in degrees Celsius.

 

D. Operation

The operation of the Barchusen requires careful boiler temperature control, a cool ambient temperature, and boiler volume scaled to the apparatus conduit diameter and total path length. Boilers with a volume of 2,000mL and 3,000mL worked well, however, 10,000mL proved too large to control for this particular recreation. In operation, the wash in the boiler is slowly brought to a gentle boil. Washes from 10% to 40% absolute volume ethanol will exhibit boiling at standard pressure and temperature conditions around 88’C. The room temperature was held at 20°C with air conditioning. As hot vapors enter the condenser, the conduit will reach thermal equilibrium along the reflux side within ten (10) minutes. During a distillation run, the point of where condensation occurs is readily visible as the front moves up the apparatus. Countercurrent hot vapors are moving upwards through a thin film “pipe” of condensate that is returning to the boiler. Each bend provides additional contact surface for intimate mixing of the vapor phase and liquid phase as well. Liquid crystal temperature indicators will display a 75°C point one at a time as the hot vapors climb the helix. Throughout our characterization runs, the reflux side of the apparatus was steadily held at 75°C from base to apex.

The apparatus operates in simple full reflux until the hot vapors climb through seven bends and reach the peak. After the peak, the vapors and condensate are no longer working against gravity, and they enter the cooler side of the apparatus. In operation, it was important not to overheat the peak, else hot vapors will continue to travel through the condensing bends and exit to the receiver with little to no condensation. In careful operation, the condenser side of the apparatus is at room temperature, and the condensate is returned at or close to room temperature to the receiver flask. At any time, the slight room temperature change may also be adjusted to provide optimal cooling.

E. Characterization Data

During the distillation of a 40% ABV neutral spirits wash, the Barchusen apparatus produced a distillate of 92% ABV or greater (Fig. 5). There was a curious slight increase in purity at the end of the run, instead of an expected decrease heralding the “tails”.

clip_image012

Figure 5

Distillate return is a gradual, at about 80ml/hour given the scale of this still. In this run, a total of 948mL of ethanol was collected against a theoretical maximum of 1000mL, giving a recovery of 94.9% (Fig. 6).

clip_image014

Figure 6

The taste and odor notes from the cinnamaldehyde both appeared at the end of the distillation run. This indicates that the cinnamaldehyde was separated primarily by boiling point, then by solubility, as would be found in a second “spirits” run (Fig. 7).

clip_image016

Figure 7

 

 

4. Boerhaave Cone Apparatus

Dr. Herman Boerhaave was born at Voorhout (Netherlands) in 1668 and earned his doctorate at Leyden in 1690, teaching theoretical medicine, then practical medicine, then chemistry. He is one of the scientists who is credited with assigning the word “alcohol” for the spirits of wine. While an esteemed teacher, his contributions to distillation design are limited. There is one mention, however, that is of interest for craft distilling: a “conical condenser in the form of a sugar-loaf” (Fig. 8).

clip_image018 clip_image020
Figure 8. Sketch of the conical Boerhaave Condenser, adapted from Forbes, 1948. Figure 9. A conical Rosenhut condenser, adapted from Forbes, 1948.

A. Construction

A tall modern-day Boerhaave apparatus was reconstructed based on the above sketch (Fig. 10). In our modern-day construction of this unit, borosilicate glass was used. A 24/40 standard taper ground joint connected to a boiler flask, and a 14/22 standard taper ground glass joint at the peak of the cone was provided for sampling and temperature monitoring (Fig. 11). The receiver attachment is a 28/15 ground glass ball joint offset to about 90° away from the cone body. Boerhaave’s design appears to be an air cooled condenser, in a shape resembling a tall, elongated Rosenhut. The difference is that the Rosenhut (Fig. 9) would collect distillate from the interior walls at about midpoint, whereas in the Boerhaave design, the vapor must pass the cone peak and exit an exterior side arm .

clip_image022 clip_image024
Figure 10. The author with a reconstructed modern-day

Boerhaave Cone

Figure 11. Glass stand-offs to insulate the condensing sidearm from the hot cone walls.
clip_image026 clip_image028
Figure 12. Liquid crystal temperature indicator at base of the cone. Figure 13. The top of the cone, with a precision thermometer placed at the last (highest) vapor-liquid transition point.

The exterior small diameter side arm is where most of the condensate is collected. Because this is an air-cooled device, the side-arm must be well separated from the hot cone under reflux, which is hot. Glass stand-offs supports to the cone body accomplish this well (Fig. 11).

 

B. Dimensions

The approximate reconstructed cone dimensions are as follows:

l Cone base diameter = 4” (10.2cm)

l Tip diameter = 0.6” (1.5cm)

l Cone height from ST = 36” (91.4cm)

l Sidearm diameter = 0.4” (1cm)

l Sidearm separation distance = 2” to 2.5” (5cm to 6.4cm)

l Path length = 72” (183cm)

 

C. Temperature monitoring and control

Adhesive liquid crystal temperature displays were applied to the main cone body to monitor run temperature (Fig. 12). These indicators (Hallcrest #4557D 0.5″ x 1.77″) provide a temperature indication of 60-90°C in 5°C increments. Four indicators were placed on the cone, equally spaced apart. A 110VAC electrical heating with resistive set point control was used to control boiler temperature.

D. Operation

The operation of the cone requires careful boiler temperature control , a cool ambient temperature, and boiler volume scaled to the cone size. We found that 10,000mL and 22,000mL boilers overwhelmed the cone with hot vapors and insufficient cooling, however, 2,000mL and 3,000mL boilers worked well. We posit a 5,000mL boiler for a 4″ base diameter cone would be the maximum size at our reconstruction scale.

In operation, the wash in the boiler was slowly brought to a gentle boil. The room temperature was held at 20°C with air conditioning. As hot vapors entered the cone, the cone body reached thermal equilibrium within an hour. Liquid crystal temperature indicators displayed a 75°C point one at a time as the hot vapors climbed the cone walls. Throughout our characterization runs, the cone temperature held steadily at 75°C from base to apex. During these runs, majority of the rising hot vapors condense on the cooler cone walls and were returned slowly as a thin film to the boiler. The thin film edge and condensate “tears” were easily observed in front of bright lighting or cast shadows.

The cone is operating in simple full reflux until the hot vapors reach the peak of the cone, which has a highly constricted diameter. A small amount of hot vapors will eventually travel past the peak and into the condensing sidearm, providing the liquid distillate. It is important not to overheat the peak and the sidearm, else hot vapors will exit the sidearm with little to no condensation. In careful operation, most of the sidearm is at room temperature, and the condensate is returned at close to room temperature to the receiver flask. At any time, the slight room temperature change may also be adjusted to provide optimal cooling.

 

E. Characterization Data

Following warm up and an initial ramp-up in ethanol concentration, the run produces fairly consistent alcohol purity (Fig. 14). There is a pronounced drop in purity, heralding the end of the run.

clip_image030

Figure 14

Distillate return is a slow process, at about 65ml/hour given the scale of this still. In this run, a total of 966mL of ethanol was collected against a theoretical maximum of 1000mL, giving a recovery of 96.6% (Fig. 15).

clip_image032

Figure 15

The cinnamaldehyde notes do not appear until the end of the run, where one would expect the “tails” (Fig. 16). Up to this point, the distillate only has characteristic ethanol odor and taste.

clip_image034

Figure 16

5. Brunschwyg “Twins” Circulatorium

Hieronymus Brunschwyg (1450-1513) was a native of Strassburg and studied medicine at Bologna, Padua, and Paris. He published two very important works on distillation: “The Small Book of Distillation” (1500), and the so-called “Big Book of Distillation” (1512), containing 79 illustrations of distillation apparatuses. It is in the Big Book that we find a curious sketch of the “Twins” or circumambulatory still (Fig. 17). This apparatus most have been noteworthy, because other scientists of the period, namely, Charas (Fig. 18), Barlet (Fig. 19), and Porto (Fig. 20) provide nearly identical sketches of it.

clip_image036 clip_image038
Figure 17. The Brunschwyg Twins, adapted from the “Big Book of Distillation”. Figure 18. The Charas Twins, adapted from Forbes, 1948.
clip_image040 clip_image042
Figure 19. The Barlet Twins, adapted from Forbes, 1948. Figure 20. The Porto Twins and inspirational Gemini, adapted from “De distillationibus”, 1609.

A. Construction

A tall modern-day Twins circulatorium apparatus was reconstructed based on the above sketches (Fig. 21). Both cucurbits are 1000mL in volume. A 24/40 standard taper ground joint was incorporated to each cucurbit body for liquid access and a boiler temperature probe. The receiver cucurbit has a ground glass stopcock to remove distillate periodically. A 14/22 standard taper ground glass joint was provided at the peak of each ambix for temperature monitoring of the vapor. The cross-connecting solens are constructed using 28/15 ball joints at one end, and a 14/22 standard taper joint at the opposite end (Fig. 22). Once assembled, the vessel is pressure-tight.

clip_image044 clip_image046
Figure 21. The reconstructed modern-day Brunschwyg Twins Circulatorium. The boiler cucurbit (C1) is at left in the heating mantle. The receiver cucurbit (C2) is at right supported by the O-Ring stand. Figure 22. Profile view of the Brunschwyg Twins Circulatorium, showing the two solens between the ambices.

B. Dimensions

 

The reconstructed Brunschwyg Twins apparatus dimensions are as follows:

l Cucurbit volumes: 1000mL

l Solen path lengths: 6” (15.2cm) joint to joint, 10” (25.4cm) overall

l Cucurbit to ambix heights: 16” (40.6cm) base to 14/22 joint

l Ambix diameters at equator: 4.7” (12cm)

 

C. Temperature monitoring and control

A 1000mL electrical heating and stirring mantle with a digital setpoint control was used to heat the wash. Precision glass mercury thermometers were installed at the top of each ambix to monitor vapor temperature.

D. Operation

The theory of operation is straightforward: One cucurbit is a boiler and holds the wash (C1), and the other cucurbit is the receiver (C2). The hot vapors leaving C1 have two condensation paths: The ambix above C1 which returns condensate into the cool C2 receiver, and the solen leading from C1 to the cooler C2 ambix. Most of the distillate is observed flowing in downward the C1-C2 solen. The bulbs above each pot allow rising vapors to quickly expand through a neck into a cooler space. The vapors condense on the ambix walls, forming “tears”, which flow downward by gravity to the solen.

E. Characterization Data

In operation. The C2 ambix vapor temperature remains close to room temperature even when the boiler is operating at 90°C. The C1 ambix vapor temperature is approximately half of the boiler temperature when distilling (Fig. 23).

clip_image048

Figure 23

Distillate return rapidly increased when the boiler setpoint was moved from 87°C to 89°C, and then leveled off at 6mL/hr into the “tails” fraction for the final 7 hours with a decreasing alcohol purity. All of the expected 400mL of ethanol was recovered (Fig. 24).

clip_image050

Figure 24

The Brunschwyg Twins still performs like a pot still. The “tails” containing cinnamaldehyde became more pronounced at the end of the run, but the cinnamaldehyde was also detectable during the “hearts” middle fraction between the 4th to 6th hour of run (Fig. 25).

clip_image052

Figure 25

 

 

6. Lonicer Apparatus

Adam Lonicer (1528-1586) was the town physician in Frankfort, Germany. He studied medicine and mathematics and taught at the University of Marburg circa 1553. He authored the works Naturalis historie opus novum (1551), and Herbarium (1555), the latter of which contained a special appendix about distillation. A strange apparatus with a right-angled key shape is provided, which Lonicer intended for essential oil extraction (Fig. 26). It is very similar to one later sketched by Porto in his third chapter of De Destillatione (1609, Fig. 27). It also closely resembles an apparatus given by Andreas Libavius in his work, Syntagma (1611) that gives credit to Lonicer (Fig. 28).

 

clip_image054 clip_image056
Figure 26. Lonicer’s Still and Condenser, adapted from Forbes, 1948. Figure 27. Porto’s Still Condenser, adapted from Forbes, 1948.
clip_image058 clip_image060
Figure 28. Libavius’ Still and Condenser, adapted from Forbes, 1948. Figure 29. The author with a modern-day Lonicer condenser, attached to a 5000mL boiler. The liquid crystal temperature indicators are visible on the crook and the angled top.

 

A. Construction

Our reconstruction of the Lonicer apparatus was limited only to the second key-shaped condenser in the ancient sketches (Fig. 29). We interpreted the sketches to show that the first unit is a reflux column and is hot along its entire path. The condenser is the second key-shaped unit positioned on top of a “thumper”. We decided that, given careful temperature control, only one apparatus was required to understand how this design functioned.

We followed the same 90° crook and the angled top as indicated by Lonicer. Countercurrent vapor-liquid contact is provided by the crook. The angled top is where most of the condensation occurs, sending the condensate down a straight conduit that is at room temperature. The base of our reconstruction used a standard taper 29/42 ground glass joint for ease of attachment to other laboratory glassware. Because of the height and diameter of our reconstructed unit, a 5,000mL boiler was selected with excellent results. The outlet is simply a drip point, and a received flask is positioned under it.

 

 

B. Dimensions

l Path length = 84″ (213cm)

l Conduit diameter = 1″ I.D. (2.5cm)

l Crook dimensions = 6.5” wide (16.5cm), 10.25” (26cm) tall

 

C. Temperature monitoring and control

Adhesive liquid crystal temperature displays were applied to the Lonicer apparatus to monitor run temperature. These indicators (Hallcrest #4557D 0.5″ x 1.77″) provide a temperature indication of 60-90°C in 5°C increments. Two indicators were placed on the crook (top and bottom) and one indicator was placed on the angled condenser. The boiler was heated using a 5000mL electrically-controlled mantle.

D. Operation

 

Once boiling begins, hot vapors rise up the first straight section and reach the first 90° bend in the crook. At this point, reflux of the condensate is observable, and the liquid-vapor front moves through the crook quickly to the second 90° bend. The next straight section slows this progression, and all the while, there is visible return of the condensate to the boiler by gravity. The most difficult control point is keeping the hot vapor edge at the peak of the angled top, not allowing it to pass through to the long downward arm. The boiler temperature control is not quick enough to adjust for this. Adding or subtracting a cooling fans and aluminum foil wrap helped to control this.

Despite a steady, controlled heat input to the boiler, it was difficult to produce a steady output from this apparatus, as the hot vapors would surge through the angled top at times, giving a lower purity in the distillate.

E. Characterization Data

During the distillation of a 40% ABV neutral spirits wash, the Lonicer apparatus produced an average purity of 90% ABV ethanol (Fig. 30).

clip_image062

Figure 30

The distillate return was not quite linear, but regression gives approximately 59ml/hr (Fig. 31):

clip_image064

Figure 31

 

We have included the data from our grappa distillation run showing where we made our cuts for best odor and taste (Fig. 32, 33). The foreshots and heads fractions had a pronounced ethyl acetate note, which disappeared at an unexpected slight drop in the boiler temperature near hour 7. This is where the beginning of the hearts fraction was taken. Another unexplained slight increase in alcohol purity occurred in the tails, similar to the Barchusen.

clip_image066

 

Figure 32

 

clip_image068

Figure 33

7. LeFèvre Apparatus

 

Nicasius le Febure, also known as Nicolas le Fèvre, was born in the Ardenns in the early 17th century. He was a Calvinist and held appointments as a scientist in the French court as the “demonstratur for Vallot” and later in the English court as the laboratory manager at St. James, London. In his treatise, Traite de Chimie (1646, 1660), he sought to improve the design of distillation apparatuses. It is in this work where we find his curious zigzag condenser (Fig. 34). Another French chemist, medical doctor, and fellow Calvinist, Moyse Charas, also gives a similar zigzag air cooled design in his Pharmacopeè Royale Galenique et Chymique (1676-1753, Fig. 35).

 

clip_image070 clip_image072
Figure 34. LeFèvre’s apparatus, adapted from Forbes, 1948. Figure 35. Charas’ apparatus, adapted from Forbes, 1948.
clip_image074 clip_image076
Fig. 36. The author with the reconstructed zigzag condenser atop a 22L boiler, with Moor’s head ambix, and receiver flasks. This apparatus stands over 8 feet tall. Fig 37. A closer view of the receiver parrot and flask, below the Moor’s head.

A. Construction

Our reconstruction of the LeFèvre apparatus used the same number of zigzags as the ancient sketch, which totaled three (Fig. 36). We elected to use 90′ angles in the bends, whereas Le Fèvre and Charas designs are slightly more acute. A careful study of both ancient sketches confirmed that the central rod is used for support and not for transiting vapor or condensate. We used solid glass rods for our support. The top and bottom of our reconstruction uses standard taper 24/40 ground glass joints for ease of attachment to other laboratory glassware. Because of the height of our reconstructed unit, a 22,000mL boiler was selected with excellent results.

 

We modified a 1,000mL ProGlass Moor’s head ambix for the top of the zigzag. This ambix was modified to include 24/40 standard taper ground glass joints at the base and top. The ambix outlet is a long exterior solen arm terminating with a drip tip. The drip tip was positioned with a bushing 1,000ml graduated funnel to monitor volume (Fig. 37).

 

B. Dimensions

l Zigzag height = 58″ (147cm)

l Tubing diameter = 0.75″ (1.9cm)

l Path length = 85″ (216cm)

l Solen arm length = 8″ (20.3cm)

l Ambix diameter = 6.5″ (16.5cm)

 

C. Temperature monitoring and control

Adhesive liquid crystal temperature displays were applied to the zigzag to monitor run temperature. These indicators (Hallcrest #4557D 0.5″ x 1.77″) provide a temperature indication of 60-90°C in 5°C increments. Four indicators were placed on the zigzag equally spaced along the path length. The boiler was heated using a dual element mantle (Glas-Col TM118) with 500VA rheostats. The boiler has three additional necks that aid in filling and draining. A precision glass thermometer was secured into one neck to monitor the wash temperature.

D. Operation

 

The operation of the LeFèvre apparatus, like all of the other ancient reconstructions, requires careful boiler temperature control, a cool ambient temperature, and boiler volume scaled to the apparatus conduit diameter and total path length. Since this was our largest apparatus reconstructed, our selection of 22,000mL boiler was well-founded. In operation, the wash in the boiler is slowly brought to a gentle boil. The room temperature was held at 20°C with air conditioning. As hot vapors entered the zigzag, each section reached thermal equilibrium along the reflux side within ten (10) minutes. This heating pattern was very similar to the Barchusen apparatus. During a distillation run, the point of where condensation occurs is readily visible as the front moves up the apparatus. Hot vapors are moving upwards through a thin film “pipe” of condensate that is returning to the boiler. The 90°C angle at each zigzag provides additional contact surface for intimate mixing of the vapor phase and liquid phase as well. Liquid crystal temperature indicators displayed a 75°C point one at a time as the hot vapors climb the zigzag up to the ambix. Throughout our characterization run, the reflux side of the apparatus was steadily held at 75°C from the boiler to the base of the ambix.

The apparatus operates in simple full reflux until the countercurrent hot vapors and condensate climb through all zigzags against gravity and reach the ambix. Once within the cooler ambix, condensation occurs in the rapidly increasing volume and the condensate falls from the walls towards the solen and into the receiver. In operation, it was important not to overheat the zigzag, else hot vapors will continue to travel through the ambix and exit the solen to the receiver with little to no condensation. In careful operation, the ambix walls are slightly warmer than room temperature and the condensate is returned at or close to room temperature to the receiver flask. At any time, the slight room temperature change may also be adjusted to provide optimal cooling.

 

E. Characterization Data

During the distillation of a 40% ABV neutral spirits wash, the LeFèvre apparatus produced an average purity of 91.4% ABV ethanol (Fig. 38).

 

clip_image078

Figure 38

 

During this run, 2,619mL of ethanol was recovered from total ethanol of 2,800mL, resulting in a 93.6% recovery (Fig. 39).

clip_image080

Figure 39

 

 

8. Results

The following table summarizes the alcohol purity for the distillation of binary solutions of 40%v/v aqueous ethanol with no other additives (Fig. 40):

Apparatus Highest Alcohol Purity Achieved, First Pass Path Length % Ethanol Recovery via Binary Distillation Boiler Volume Boiler Setpoint for Distillate to Appear
Barchusen 93.11% 305cm 94.9% 2,000mL 85°C
Boerhaave 92.78% 183cm 96.6% 3,000mL 85°C
Le Fevre 92.44% 216cm 93.6% 22,000mL 88°C
Lonicer 92.22% 213cm 66.6% 5,000mL 88°C
Brunschwyg 85.23% 25.4cm 100% 1,000mL 85°C

Figure 40

All of the tall column air-cooled apparatuses achieved over 90%v/v purity on the first pass distillation. Their performance was reminiscent of rectification observed in a packed or tray column, although not quite reaching the theoretical water-ethanol azeotrope maximum. The behavior of the Brunschwyg apparatus was similar to a “pot” or alembic still, and the alcohol purity is noticeably lower. The poor recovery of the Lonicer is attributed to the difficulty in controlling the position of the hot vapor front during the run. Changing the boiler setpoint does not provide timely control of this.

9. Considerations for Improvements

 

The addition of pure copper sponge or wire in the hot vapor path of any of these stills would aid in the removal of volatile sulfides from the wash. Similarly, these reconstructions may be constructed entirely out of copper metal.

The addition of a cooling jacket to the key transition areas of the still would improve distillate collection. These areas include the peak of the Barchusen, the sidearm of the Boerhaave, the solens between the Brunschwyg Twins, and the angled top arm of the Lonicer.

10. Conclusions

 

The careful reconstruction of five ancient distillation apparatus has proven that these forgotten and esoteric designs are capable of producing high purity ethanol on a single distillation run using only air cooling. The designs by Barchusen, LeFèvre, Boerhaave, and Lonicer were proven to produce 90% ABV ethanol or greater from the first distillation of grape marc washes and binary ethanol-water solutions. The use of clear borosilicate glass, electronic temperature control, and very slow boiling over a long period allowed the observation of liquid-vapor interactions in all apparatuses.

 

The curious angles and bends in the Barchusen, LeFèvre, and Lonicer designs act as plates of separation and enrich the vapor phase. The highest purity observed in this research project was 93.11% ABV ethanol from the Barchusen apparatus. Interestingly, a simple cone shape will also provide as high as 92.78% ABV ethanol, as demonstrated by the Boerhaave apparatus. The Brunschwyg Twins Circulatorium, despite its interesting design, performed much like a typical alembic and we did not find it exceptional.

These ancient designs are not without challenges. The ambient temperature must be cool throughout the distillation run, and air conditioning is a requirement during the summer months. Because the purity of alcohol distillate from these stills is always very high throughout the entire run, “cuts” must be made by taste and smell. Wide temperature swings or large changes in proof are not observed until very late in the distillation run. Sizing the boiler and carefully controlling the boiler rate are paramount. It is very easy to overheat these designs and sent hot vapor through the entire apparatus with aggressive boiling.

We believe our research is the first successful reconstruction in glass of such designs in glass in centuries. We encourage craft distillers to reconstruct and scale up these easy designs in copper for the production of boutique craft spirits and further differentiation in the production of craft spirits.

11. Acknowledgements

 

Mohawk Spirits Distillery would like to thank the American Distilling Institute and President Erik Owens for their grant award that made this research possible. We would like to extend special thanks to our glassblowing partner, Q-Glass (Towaco, New Jersey) for their help with apparatus construction. Finally, we thank our legal partner, Hops & Vine Consultants (www.hopsandvine.net, New York), for their assistance with the TTB equipment registration of all of the stills described in this report.

 

 

 

 

 

12. References

Belgiorno, Maria Rosaria. “Behind Distillation. A research born after the discovery in Cyprus of 2000 BC alembics”. 2017. ISBN 978-9963-2448-1-2.

Boerhaave, H., “Elementa chemiae” (Leiden, 1731/1732, 2 vols.).
Boerhaave, H., “Elements de chimie” (paris, 1754, 6 vols.).
Brunschwygk, Hier, “Liber de arte distillandi de simplicibus oder Buch der rechten Kunst zu Distillieren die eintzigen dinge” (Strassburg, 1500, Johann Gruninger, the so-called “Small Book of Distillation).
Brunschwygk, Hier, “Liber de arte Distillandi de Compositis; Das Buch der waren Kunst zu distillieren die Composita and simplicia und das Buch thesaurus pauper” (Strassburg 23 Februar 1512, the so-called “Big Book of Distillation). (Other editions: Strassburg 1519, 1531; Francfort 1553, 1594, 1598, and several adaptions by Ulstad, Ryff, Uffenbach).

Burdock, G.A. (ed.). Fenaroli’s Handbook of Flavor Ingredients. 4th ed.Boca Raton, FL 2002, p. 293

Forbes, R. J. E. “Short History of the Art of Distillation”. 1948. J Brill, Leiden, Holland.

French, John. “The Art of Distillation an Alchemical Manuscript Being Certain Select Treatises on Alchemy and Hermetic Medicine” (London, 1650). Republished by Theophania Publishing, ISBN 9871770830059.

Lemery, N., “Cours de cymie contenant la maniere de faire les operations en usage dans la medicine par une method facile avec des instructions et raisonnements sur chaque operation our l’instruction de ceux qui veulent s’appliquer a cette science” (Paris, 1675, 1744, and 21 more editions).
Meilgaard MC (1975) Flavor chemistry of beer part II: flavor and threshold of 239 aromavolatiles. MBAA Technol Q 12:151–168

Mohawk Spirits Distillery, online (February 6, 2020): http://www.mohawkspirits.com

Rasmussen, Seth C. “The Quest for Aqua Vitae. The History and Chemistry of Alcohol from Antiquity to the Middle Ages”. 2014. Springer ISBN 978-3-319-06301-0.