Pulped natural

Where it is practiced

Brazil, first and foremost. Minas Gerais is its center: Cerrado Mineiro, Sul de Minas, Matas de Minas. Sera and colleagues published in 2025 a study of 175 farms belonging to the BSCA, the Brazilian specialty coffee association. It places in Minas Gerais 82% of the specialty coffee area it records. São Paulo carries about 10%, with the Mogiana paulista. The rest is split among Goiás, Paraná, Mato Grosso do Sul, Rio de Janeiro and Espírito Santo.

Southwest Bahia practices it too. It is there that Lima and colleagues conducted in 2005 their comparison of the three processing routes, on working farms.

The same 2025 study shows that the semi-dry route dominates among Brazilian specialty producers, alone or alongside other routes. Of the 175 farms, 39 practice only one route, dry or semi-dry. 52 practice at least two, 52 practice all three, and 32 could not be classified.

Outside Brazil, the same technical operation exists under the name honey, mainly in Central America. The differences in practice are described in the gray area.

Its spread dates from the 1990s. The birth date of the process is not established, see the gray area.

The principle

Pulped natural removes the skin and the pulp of the cherry, then dries the parchment with all its mucilage. Mucilage is the layer of sugary, sticky flesh stuck to the parchment. There is no tank, no fermentation water, no demucilager. The vocabulary of anaerobic therefore does not apply here: nothing is closed, the parchment dries in the open air. The process does two things at once.

The first is a mechanical ripeness sort. The washer-separator first sets aside, by density, the fruit that floats. The cherry huller (descascador, in Portuguese) then presses the fruit against a grid. Only the skin of ripe fruit gives way. That of green fruit resists, and the greens come out whole through another chute. Sera and colleagues present this sort as the reason for the process, in a country where harvest is by stripping the branch, all ripeness stages mixed.

The second is a change in drying kinetics. Freed of its skin and pulp, the parchment enters drying at about 45 to 50% moisture on a wet basis, that is, relative to the total mass. The mucilage, for its part, remains. Avallone and colleagues assayed its polysaccharides. The pectic substances make up about 30% of the alcohol-insoluble residue. The extracted pectins carry about 60% uronic acids and are strongly methyl-esterified. It is this material that sticks. The bean, for its part, is still alive. Kramer and colleagues followed its chemistry during drying. A first peak of GABA, an amino acid that accumulates when the cell is under stress, accompanies the expression of isocitrate lyase. This enzyme of the glyoxylate cycle marks a seed that is mobilizing its reserves, hence a germination metabolism under way. Two GABA peaks follow, this time in phase with maxima of expression of dehydrins. Dehydrins are proteins that respond to water shortage, here in the embryo and then in the endosperm. The way drying is run therefore acts on the internal chemistry of the bean, and not only on its surface. What becomes of the surface microbial flora during drying on the patio (terreiro, in Portuguese), on the other hand, none of the sources we were able to read measures for this route.

The sequence

1. Harvest. In Brazil, stripping of the branch by hand or by machine. The fruit arrives at all ripeness stages mixed, which makes sorting downstream obligatory.

2. Received the same day. The delay governs what follows. A cherry that waits ferments in the trailer, and the defect is already there before the machine.

3. Washing and separation by density, in the washer-separator. The floaters (boias, in Portuguese) go off to another circuit. Three things are set aside here: dry fruit, fruit pierced by the borer and malformed fruit.

4. Hulling. The ripe cherry bursts against the grid, the bean passes with its mucilage, the skin leaves by auger. The green fruit, firmer, resists and comes out whole. The setting of the gap governs everything. Too tight, it also crushes the greens and sends them into the lot.

5. Drying, immediately, without a pass through a tank. This is what distinguishes the process from the washed route, which ferments the parchment in a tank before washing it.

6. Spreading on the patio, the drying area. A technical guide of a Brazilian cooperative gives 7 liters per square meter for pulped natural (cereja descascado, in Portuguese), against 14 for natural coffee. The layer is half as thick because the mucilage glues the parchments together. This figure comes from a single source, see the gray area.

7. Stirring. The same guide recommends up to sixteen turnings per day on the first day, once the surface water has gone. Stirring breaks up the clumps before they form, and prevents a wet core from remaining in the middle of the pile.

8. Temperature control. Common practice caps the mass of parchment coffee at 40 °C. Cunha and colleagues ran all their drying trials under this ceiling. It is a working value, not a demonstrated damage threshold, see the gray area.

9. End of drying. The working target is 11 to 12% moisture, and the published trials stop between 11 and 13%. One goes down to there because less available water puts the lot out of reach of storage molds.

10. Mechanical variant. A full drying with hot air capped at 40 °C leads from 45-50% to 11-13% in 15 to 37.5 hours. Cunha and colleagues measured these durations on a pilot line. The same authors also tried a two-stage drying: hot air down to 30% moisture, then hot air assisted by microwaves. The whole then falls to about ten hours.

11. Rest in tulha, the resting silo, then hulling of the parchment before shipment.

What it gives in the cup

What was measured in tasting. Lima and colleagues worked on farms in southwest Bahia, on the 2005 harvest. They compared there the same Catuaí Amarelo prepared as natural, as pulped natural and as fermented pulped (despolpado, in Portuguese). The design was fully randomized, with six replicates. They recorded five measures: color index, pH, electrical conductivity, potassium leaching and the prova de xícara. The latter is the tasting that ranks the lot on the Brazilian scale of bebidas. The natural gives the lowest pH and the highest values of conductivity and potassium leaching. The authors read there the most aggressive treatment for the cell membranes of the bean. The cup score drops for the natural on only one of the farms. Pulped natural and fermented pulped come out similar on color, conductivity, leaching and cup.

What this trial does not say. The published abstract does not indicate that the samples were presented blind, nor how many tasters worked. We could not read the full text. We therefore do not call this trial a blind trial, and we do not take up the detail of scores farm by farm.

What is measured, in green bean chemistry. Duarte, Pereira and Farah compared in 2010 the semi-dry route and the washed route on 17 Brazilian arabica cultivars and progenies. They assayed by liquid chromatography nine chlorogenic acids, caffeine, trigonelline and sucrose. The semi-dry route gives more sucrose (p = 0.02), fewer chlorogenic acids (p = 0.02) and less trigonelline (p < 0.01). No difference on caffeine.

Why sucrose matters. At roasting, it hydrolyzes into glucose and fructose, two reducing sugars. These are what react with amino acids in the Maillard reaction, and form most of the brown and aromatic compounds. The sucrose that remains caramelizes on its own, without amino acid. A bean richer in sucrose therefore starts with more raw material for both pathways. The 2010 study did not taste: it assayed. The link with a sweeter cup remains a hypothesis, not a result.

What is only claimed. The profile “low acidity, sweet flavor, honey or cane aroma” circulates in the Brazilian technical literature and among sellers. It is repeated from author to author. No blind trial comparing several origins establishes it as a general rule. One measurement even runs against the first half of this sentence. In Lima and colleagues, the pulped natural had a lower pH than the fermented pulped, hence a more acidic bean. The pH of the green bean and the acidity perceived in the cup are not the same measurement. But nothing in these figures supports a low acidity.

What goes wrong

The defect passes the machine. This is the risk specific to this process. In a washed route, the coffee still goes through the tank and then the washing channels, where light and damaged beans separate in the water. Here, after the huller, the parchment goes straight to drying. If the washer and the huller are poorly set, floaters, pierced beans and malformed beans enter the final lot. Nothing makes up for them afterwards. The quality of a pulped natural rests on the setting of the machines.

The crushed green. A huller set too tight makes the skin of immature fruit give way too. The ripeness sort, which is the reason for the process, then fails. Green, astringent beans are found in the cup.

The clump on the patio. Mucilage is a glue. A layer too thick, stirring too rare, and the parchments weld together. The core of the mass stays wet, and fermentation there becomes uncontrolled. This is the ardido bean of the Brazilians, the over-fermented of the English tables.

Drying that goes backward. This is the real sanitary danger. The temperature gap between day and night makes water condense, and the water activity of the lot rises. Water activity is the share of water actually available to microorganisms. Pardo and colleagues measured on green beans the growth of Aspergillus ochraceus and its production of ochratoxin A. Mycelial growth is at its maximum around 30 °C, between 0.95 and 0.99 water activity. Toxin production peaks at 20 °C and 0.99. At 0.80 water activity, they detect none.

Too dry. Silva and colleagues analyzed lots of pulped natural taken from 32 farms in southern Minas Gerais, on the 2001-2002 harvest. Most of the lots came out below the recommended water content. Drying beyond the target costs the seller weight, and that weight does not come back.

Overheating. Common practice caps the parchment mass at 40 °C. The tools that measure the damage exist and are used: the electrical conductivity of the exudate and the potassium leaching report on the integrity of the bean’s membranes. Lima and colleagues use them to rank the processing routes against one another. None of the sources we read establishes, however, that 40 °C is the point where the membranes give way.

Dependence on climate. The process assumes a harvest in the dry season. Under a humid sky, sugary parchment spread on the ground becomes a burden, not an asset. It dries slowly, it sticks, and every day that passes humid is a day handed to molds.

Benchmarks

• Moisture at the start of drying45 to 50% on a wet basis
• Target moisture at the end of drying11 to 12%, the published trials stopping between 11 and 13%
• Spreading on the patio7 liters per square meter, against 14 for the natural (single source)
• Maximum temperature of the parchment mass40 °C (working value, not a demonstrated threshold)
• Complete mechanical drying with hot air capped at 40 °C15 to 37.5 hours
• Maximum growth of Aspergillus ochraceusaround 30 °C, between 0.95 and 0.99 water activity
• Water activity with no ochratoxin A detected0.80
• Share of Minas Gerais in the specialty areas recorded in 202582%