Chocolate making, professionally

Conching: What It Does to Flavor and Texture

By Oliver BennettPublished October 15, 2026Updated October 15, 202614 min read

Conching is the long mixing, shearing and heating of refined chocolate until a dry or pasty mass becomes a smooth, flowing liquid and its aroma settles. It decides how well the chocolate pours and coats, and it shapes part of how it tastes. Makers argue about the hours because two things run on different clocks: in the studies that measured both, flow stopped improving after a few hours, while aroma compounds kept falling for as long as the machine ran hot. Conching is one step in the bean-to-bar process, between refining and tempering.

What conching does

Conching turns the dry or pasty mass that leaves the refiner into flowing chocolate, and it changes the aroma on the way. A 2023 review by Ditchfield and colleagues describes it as a stirring operation above 50 °C for a few hours that removes undesirable volatile compounds and develops flavor, viscosity and texture.

Physicists at the University of Edinburgh looked at the mechanics. In a 2019 paper, Blanco and colleagues name two main jobs for conching: breaking up aggregates of particles, and lowering the friction between them. Their results suggest that lecithin, usually described as a dispersant, may work mainly as a lubricant that lets the particles slide past each other.

Where the conche came from

The conche goes back to Rodolphe Lindt in Bern. According to the Historical Dictionary of Switzerland, Lindt founded his firm in 1879 and stirred his chocolate in a longitudinal machine with a shell-shaped trough that he designed himself. He also added pure cocoa butter to the mass, so it could be poured into molds instead of pressed, and the chocolate came out less brittle and melted more gently.

The dictionary gives no year for the machine. The PNAS paper by Blanco and colleagues dates the invention to 1879. The name comes from the trough: a 1980 patent explains that conching is so called from the shell-like form of the original machines.

The three phases: dry, plastic and liquid

Conching runs through a dry, a plastic (also called pasty) and a liquid phase, each named for the state of the mass. In a 2024 study at a 200 kg single-shaft conche, Guckenbiehl and colleagues describe the dry phase as the stage where refined powder firms into a plastic mass. The plastic phase carries the strongest shear forces and most of the change in texture and flavor. The liquid phase starts when cocoa butter, an emulsifier or both go in, and the paste turns into a smooth, liquid melt.

Older accounts, such as a 2008 review by Afoakwa and colleagues, split the same process in two: flake or powder becomes a paste, then the paste becomes a free-flowing liquid once cocoa butter and lecithin go in.

The dry phase shows up on the motor. In one industrial power curve, published in a 1993 patent from a large confectioner, power dropped steeply during roughly the first half hour of dry conching. At about that point cocoa butter and emulsifier were added and the motor was reversed for the wet phase.

A melanger skips the dry phase altogether. It refines and conches in the same drum, so its mass is fluid from the start.

What changes in the chocolate

Water comes out, fat spreads over the particles and the aroma shifts. What happens to acids depends on the machine.

Water

Moisture has to go early, because it makes the particles clump. A 2021 study by Urbańska and colleagues, citing earlier work by Bolenz, explains that during refining, water moves onto the surface of the crushed sugar and forms a layer of dissolved sugar that can glue particles together. It should be removed in the dry phase, and the finished chocolate should hold less than 1.0% water, preferably 0.4 to 0.6%. Too much water can also lead to sugar bloom.

A conche is a poor dryer, though. Bolenz and colleagues dried milk chocolate refiner flakes in an oven to below 0.6% water before conching, which let them drop the dry phase and replace a conventional 5-hour conche with a 20-minute liquefaction. Sensory tests found negligible differences.

Coating and flow

Conching makes chocolate flow by spreading fat over the particles, not by reshaping them. The flake that comes off a roller refiner is a dry powder of sugar and cocoa clumps with the fat locked inside, and why it arrives that way is covered in chocolate refining.

When Hoskin and Dimick looked at chocolate from four different conches under a scanning electron microscope in 1980, the sugar and cocoa particles showed little or no change in size or shape as conching went on. The surfaces changed instead. They smoothed as cocoa butter spread over them, and flattened further when lecithin was added. As the chocolate became more fluid, its plastic viscosity fell sharply and quickly.

Guckenbiehl’s 200 kg study measured how long that takes. Viscosity after 3 hours and after 6 hours was significantly lower than after 1 hour, but the 3-hour and 6-hour samples didn’t differ from each other. Conching at 80 °C coated the particles more efficiently than 60 °C and left a lower viscosity overall. Residue from the previous batch, which the authors call a pre-charge, also brought viscosity down quickly, but it diluted the aroma. A clean conche may let more acetic acid escape, they note, while stretching the time needed to reach good flow.

Aroma: what leaves, what stays, what forms

Aroma changes by evaporation, by moving compounds between the cocoa butter, the cocoa particles and the sugar, and by forming a few. That is how Guckenbiehl and colleagues summarize the earlier work, and their own measurements show how large the losses can be. At elevated temperature, odorant concentrations in the fat phase fell by up to 78%. Between 1 and 6 hours, conching forward at 80 °C with a pre-charge, tetramethylpyrazine fell by 70.7%, 2-phenylethanol by 47.1% and benzaldehyde by 52.5%. At 60 °C without a pre-charge, the same three fell by 10.6%, 7.3% and 42.0%.

Some compounds go up. Comparing dark chocolate before and after conching, a 2002 study by Counet and colleagues found that no new key odorant was formed, but levels of branched pyrazines, Furaneol and 2-phenyl-5-methyl-2-hexenal rose significantly. Most Strecker aldehydes were lost by evaporation.

A good part of what leaves is unwanted. In a 2013 study by Owusu and colleagues, unconched dark chocolate was described as musty and smoky, with fruit and banana notes. After 6 or 8 hours at 80 °C the musty and smoky off-flavors had dropped, and so had the fruit. Tasters rated the unconched samples higher for banana.

Browning chemistry plays a small role in dark chocolate. Clark and colleagues, citing Hoskin and Dimick, note that even traditional conching doesn’t reach the temperatures Maillard reactions need. For milk chocolate the question is still debated.

Acids: where the studies disagree

Acetic acid falls in a large conche, but in smaller machines and model systems it has been seen to rise or hang on. At 200 kg, Guckenbiehl’s team measured it falling through the plastic phase in each program they reported: by 42.3% between 1 and 6 hours conching forward at 80 °C with a pre-charge, 40.1% at 60 °C conching backward, 31.9% at 60 °C forward with a pre-charge, and 22.7% between 2 and 6 hours without a pre-charge. Most of the drop came early and then leveled off. Reversing the direction and conching at 80 °C both helped clear excess acetic acid from the fat phase.

The same study also caught the opposite. Without a pre-charge at 60 °C, acetic acid in the particle phase had risen after 6 hours, both in total and as a share, a result the authors called unexpected.

A companion study on model systems by the same group points to a cause. Under conching conditions, total acetic acid rose by 13.8 to 56.9% in every model that contained crystalline sugar or cocoa particles, which indicates that free acetic acid forms during conching. The 200 kg study adds a hypothesis about timing: acetic acid sits mainly in the cocoa particles, so it probably evaporates best early on, before fat has coated them. If that holds, it would also explain why acid can linger in a melanger, where the mass is coated in fat from the start.

Whether the leftover acid matters on the tongue is a separate question. A 1980 thesis by Hoskin, as summarized by Clark and colleagues, found that volatile fatty acid levels tracked rotation speed rather than conching time, and that trained sensory panels couldn’t tell apart samples with measurably different levels of acetic acid.

How long and how hot

No single time or temperature is right. Published programs run from a 20-minute liquefaction to a few days in a melanger, and the sources tie the time to temperature, recipe, machine and batch size. Higher temperature shortens conching, as a 2008 review by Afoakwa and colleagues puts it. The same review gives typical industrial values drawn from older textbooks, with cooked-flavor changes in milk chocolate above 70 °C. They are in the table below as typical published values.

Machine design has moved the hours too. A 1994 patent from a large confectioner puts early longitudinal conches at 24 to 36 hours and the vertical mixer conches that followed at about 5 to 8 hours.

Types of conche

Conches differ mainly in how they move the mass and how quickly they get through the dry phase.

The longitudinal conche is the original design. The 1980 patent from a grinding-equipment company describes a tank with a granite bed, where a roller pushes the chocolate to and fro, and a shaped end that throws the chocolate back over the roller into the machine.

Rotary conches replaced the long trough with mixing shafts. The 1994 confectioner’s patent describes vertical mixers that used plow and shear blades to tumble the dry refinings and then work energy into the mass once it turned plastic. It also describes horizontal designs with three shafts carrying kneading and shearing blades, the center shaft turning against the two outer ones, with reversible direction. That “double-overthrow” action keeps fresh material moving into the high-shear zones, and its purpose was a shorter dry phase.

Single-shaft batch conches are built large. One maker’s page lists batches of 6,000 to 9,000 kg and gives no conching time.

Direction matters as much as design. In Guckenbiehl’s 200 kg conche, running backward worked the mass with the flatter side of the shear element and put more shear and mechanical stress on it, while running forward helped mixing and let moisture out.

Fast and continuous conching has been studied mainly for milk chocolate, by the same group behind the pre-dried flake work and its earlier concept for fast continuous conching. In a high-shear mixer, only very high throughput raised viscosity slightly, and sensory differences stayed negligible.

A refiner-conche refines and conches in one vessel; how it works is in chocolate refining. In a 2022 case study at a mid-sized Peruvian company, Valverde-Ayllon and colleagues sampled a 20-hour refiner-conche process every 2 hours. After 12 hours the physical properties stopped changing, and a trained panel found no significant difference between 12 and 20 hours.

Melangers sit at the other end of the scale. A melanger used as a conche refines and conches at the same time. Clark and colleagues point out that melangers are typically run at ambient temperature, while conches are generally heated to up to 80 °C. They conclude that research on conching can still inform melanger profiles, and expect one measurable difference: more alcohols retained in the melanger because it runs cooler. How warm a melanger actually gets from friction is covered in the bean-to-bar process.

How to tell when conching is done

A batch is done when its flow has stopped improving, its moisture is on target and it tastes the way you want; the clock alone can’t tell you. Flow is the easiest of these to measure, and both the 200 kg study and the refiner-conche case found a plateau: between 3 and 6 hours in one, after 12 hours in the other. A viscosity reading or a simple flow test at set intervals shows when yours arrives.

Moisture is the second check. The target cited earlier is below 1.0%, preferably 0.4 to 0.6%, and since a conche dries slowly, a wet batch can need time that its flow no longer does.

On industrial machines the motor carries information too. The 1993 confectioner’s patent follows motor power through conching because the power curve tracks the consistency of the mass, and it proposes controlling the conche at constant power so that total conching time falls once unnecessary conching is cut out.

Taste is the maker’s call. The bean-to-bar maker Dandelion Chocolate wrote in a 2015 blog post that refined but unconched chocolate was “smooth in texture but terrible in flavor”, and that the longer a chocolate stays in a melanger, the mellower it tends to be. Flavor development doesn’t always coincide with the nibs reaching the right particle size, they added, which is why they wanted to split refining and conching across two machines.

Can you over-conch?

Longer and hotter conching keeps removing aroma after flow has stopped improving, but the published work sets no point at which chocolate counts as over-conched, and some panels taste no difference at all. Guckenbiehl’s team found that higher temperatures speed up the fall in viscosity and the fall in value-adding odorants together, so good flow and good aroma call for a compromise.

Panels have noticed the loss in some studies. In Urbańska’s milk chocolate mass, conched in a food processor for 1 to 3 hours, longer conching lowered the intensity of chocolate aroma and the panel’s scores followed. The cocoa trade’s own quality guide assumes the same: its tasting-sample protocol keeps ground cocoa mass below 55 °C so that the bean’s flavor isn’t “stripped or reduced as it would be in conching”. That is a rule for preparing tasting samples, but it shows how the industry thinks about heat and time.

Other panels found nothing. In a milk chocolate with freeze-dried blueberry, de Freitas and colleagues found that a 6-hour conche was already enough, with no significant difference in descriptive analysis or consumer acceptance across 6 to 72 hours apart from aroma liking. The refiner-conche panel and Bolenz’s 20-minute milk chocolate point the same way.

Extra hours past the flow plateau cost aroma compounds and buy little texture. Whether anyone tastes the loss depends on the chocolate, the program and the panel.

Where conching sits in the line

Conching comes after refining and before tempering. The wider sequence is laid out in how chocolate is made, and the machines a factory needs, in order, are in the chocolate production line.

Against Salmonella, conching is only a weak kill step, so it can’t stand in for a validated roast; the data are in chocolate food safety. If you are buying a conche, the questions to put to every vendor are in the equipment RFQ checklist.

Sources and open questions

The 5 kg study by Albak and Tekin (2016) has internal inconsistencies, so its figures appear here only next to other studies. Its conclusion reports a 57% fall in volatiles by mass, but its own totals run from 37,606 to 30,197 ppb, about a fifth lower, after rising again in the hot pasty phase.

Blanco and colleagues’ PNAS paper used a model system (milk crumb powder, sunflower oil and lecithin in a planetary mixer), not a real conche, and was part-funded by a chocolate company. The time and temperature ranges in Afoakwa’s review are drawn from textbooks by Beckett, Minifie, Awua and Whitefield. Hoskin’s 1980 thesis is cited through Clark and colleagues.

Read in abstract or title only: Hoskin and Dimick 1980, Counet 2002, Guckenbiehl 2024b, Bolenz 2005 and 2007, Valverde-Ayllon 2022 and de Freitas 2025.

A defined over-conching threshold, if one exists, may be in the 2022 review by Augusto and Bolini, whose full text wasn’t available.