CoffeariumBeta
  1. Plant
  2. Farm
  3. Process
  4. Trade
  5. Roast
  6. Brew
  7. Cup
Post-harvest processing · Step 8

Grading and sorting of green coffee

How green coffee is separated by bean size, density and colour, and how defects are counted to give a lot its grade – screens in 64ths of an inch, gravity tables, optical sorters and defect equivalents.

Across the chain: Equipment

Source-checkedby Coffearium verification pass· Last source-checked 8 Oct 2026· Prepared with AI assistance· 14 sources (10 tier A)· 3 contested· Updated 10 Oct 2026
Reading depth
Contents
  • Green coffee is sized on perforated screens whose numbers are hole diameters in 64ths of an inch; round holes for flat beans, slotted holes for peaberries.4,3
  • Density is separated on vibrating, air-fed gravity tables; colour by optical sorters that view each bean and blow rejects out with air.11,10
  • Grades count defects in a fixed sample and convert them to "full defects" with an equivalence table; each system weights defects differently.1,3,5
  • For SCAA specialty grade, a sample may hold at most 5 full defects and no primary defects; the SCA’s archived standard allows zero category 1 and five or fewer category 2 defects.1,2,14

What grading and sorting do#

After hulling, green coffee is still a mixture: large and small beans, flat beans and peaberries, dense and light beans, sound and defective ones, and some stones, sticks and husk. During dry processing the beans are classified by size, shape, weight and colour, and recombined into the commercial preparation the buyer asks for.13 A preparation is a combination of bean sizes and shapes with a stated maximum number of defects.13

Two different things happen here. Sorting is physical: machines and people separate beans. Grading is a judgement on a sample: the beans are sieved and the defects counted, and the lot receives a grade or type.1,3,4 Grades can also depend on altitude or region, botanical variety, wet or dry preparation, and bean size, sometimes with shape and colour.7

Sizing: screens in 64ths of an inch#

Beans are sized on perforated metal screens. ISO sieves are defined by metric apertures; the traditional screen number is the nearest round-hole diameter (or slot width) in 64ths of an inch.4 Brazilian regulation writes it out directly: screen 14 is 14/64 inch.3 One step on the scale is therefore 1/64 inch, about 0.4 mm.

Flat beans and peaberries need different screens. A flat bean passes a round hole by its width; a peaberry is rounder and is separated on slotted (oblong) holes.4 The Brazilian standard sorts flat beans (chato) on round-hole screens 19, 18, 17, 16, 15, 14 and smaller, and peaberries (moca) on oblong-hole screens 13 down to 8; its laboratory series also includes round screens 13 and 10 for flat beans.3

012345678mmScreen 20 (round)7.94 mm4 Screen 19 (round)7.54 mm3 Screen 18 (round)7.14 mm3 Screen 17 (round)6.75 mm3 Screen 16 (round)6.35 mm3 Screen 15 (round)5.95 mm3 Screen 14 (round)5.56 mm3 Screen 13 (slotted)5.16 mm3 Screen 12 (slotted)4.76 mm3 Screen 11 (slotted)4.37 mm3 Screen 10 (slotted)3.97 mm3 Screen 9 (slotted)3.57 mm3 Screen 8 (slotted)3.18 mm3
Data table
mmSource
Screen 20 (round)7.94 mm4
Screen 19 (round)7.54 mm3
Screen 18 (round)7.14 mm3
Screen 17 (round)6.75 mm3
Screen 16 (round)6.35 mm3
Screen 15 (round)5.95 mm3
Screen 14 (round)5.56 mm3
Screen 13 (slotted)5.16 mm3
Screen 12 (slotted)4.76 mm3
Screen 11 (slotted)4.37 mm3
Screen 10 (slotted)3.97 mm3
Screen 9 (slotted)3.57 mm3
Screen 8 (slotted)3.18 mm3
Screen numbers and their nominal hole size in millimetres (number × 25.4 / 64), with the round-hole range for flat beans and the slotted range for peaberries used in Brazil. Calculated from the 1/64-inch definition; ISO sieves use the nearest metric aperture.3,4

For a size analysis, ISO 4150 takes 300 g of the laboratory sample, weighed to 0.1 g, and shakes the nest of sieves for 3 minutes at 110–130 cycles per minute.4 The screens most often used are 18, 17, 16, 15, 14 and 12.4 In the SCAA method reproduced by FAO, specialty and premium lots may deviate by at most 5% above or below the screen size stated.1

Size and altitude grades by origin#

Many producing countries name their export grades after size, altitude or both.

  • Kenya grades internally as E, AA, AB, PB, C, TT and T, on bean size and density, and then splits these by liquor quality.7
  • Colombia reported five grades to the ICO – Premium, Supremo, Extra, Excelso and Caracol – each defined by screen size.8
  • Central America uses altitude bands: Strictly Hard Bean (SHB) in Guatemala and Strictly High Grown (SHG) in El Salvador are the highest bands; El Salvador’s lower ones are High Grown and Central Standard.7 “European preparation” (EP) allows fewer defects than “American preparation”.7 In the ICO’s 2018 compilation of national quality standards, Guatemala is recorded as not yet implementing quality standards for coffee, so these names are trade usage rather than regulation there.8
  • Brazil classifies by screen and by “type”, a count of defect equivalents in 300 g, from type 2 (cleanest) to type 8.3

The altitude limits and defect allowances behind SHB, SHG, EP and American preparation are given in one trade guide opened for this entry; no second source was found, so the numbers are left out here.

Density and colour sorting#

Gravity (density) tables. A gravity separator has a perforated deck that slopes gently in two directions and is shaken by an eccentric drive.11 Air blown up through the deck lifts and loosens the layer and brings the lighter particles to the top.11 The vibration walks the heavier particles towards the higher side of the deck, while the lighter ones move down towards the lower corner.11 An adjustable gate sets the feed rate, and movable dividers at the discharge edge set which weight ranges go to which outlet.11 The patent describes a general gravity separator and names coffee among the materials such devices have long separated;11 no coffee-specific trial of table settings was opened for this entry.

Optical sorters. In a gravity-fed sorter, individual coffee beans run down a steep channel or chute.10 Near the bottom, an optical sensor views each bean as it passes, against a diffusing background panel, and detects off-standard items.10 After a very short delay an ejector fires and blows the detected bean out of the stream.10 Commercial sorters work in monochromatic mode (one wavelength), bichromatic mode (two wavelengths) or both.9 Sensing can be partly or wholly outside the visible range; the patent notes that infrared sensing is common.10 High-speed optical sorting is used for coffee, beans, peanuts, peas and rice.9

1234567
  1. Hopper and vibratory feeder. 10
  2. Gravity chute. Beans slide down in single file.10
  3. Optical sensors judge each bean by its reflectance, in one or two wavelength bands, visible or infrared.9,10
  4. Background panel. 10
  5. Air ejector. Fires after a very short delay.10
  6. Accepts.
  7. Rejects.
How a chute-fed optical sorter works: beans slide single-file down a chute, are viewed against a background at one or two wavelength bands, and rejects are blown out by an air ejector after a short delay.9,10 Schematic, not to scale.

Ultraviolet light. Under UV light some green coffee beans show a blue fluorescence, and the trade has long used this to spot faulty beans.12 In a study from Brazil, a larger share of fully fluorescent beans in a natural specialty lot had a negative effect on the final cup score and on astringency.12

Counting defects#

A defect count turns a sample into one number. Each defect type has an equivalence: how many of it make one “full defect”. The weights differ by system.

  • SCAA, as reproduced by FAO. Primary defects count heavily: one full black, one full sour or one pod/cherry is one full defect.1 Secondary defects count lightly: five broken or chipped beans, five floaters or five shells make one full defect.1 Specialty grade allows at most 5 full defects and no primary defects, no quakers, and moisture between 9 and 13% (basis not stated); it must also show at least one distinctive attribute in the cup.1 Premium grade allows up to 8 full defects.1
  • Brazil (MAPA). Defects are intrinsic (bean defects) or extrinsic (foreign matter and husk). One black bean is one defect; two sour (ardido) beans, five immature (verde) beans or three shells make one.3 A large stick, stone or clod counts as five.3
  • ISO 10470. The ISO defect reference chart sorts defects into five categories and applies to any species, variety and processing.5 Its coefficients weigh each defect’s effect on mass loss and on sensory quality, for use in contracts or lot classification.5
  • ICO export floor. The ICO quality programme excludes Arabica with more than 86 defects per 300 g, and any coffee below 8% or above 12.5% moisture measured by ISO 6673.6

Reference values: screen sizes#

Nominal hole sizes are calculated from the definition (screen number × 1/64 inch).3,4 ISO sieves use the nearest metric aperture, so a real sieve can differ slightly.4

Screen Nominal hole (in) Nominal hole (mm) Hole shape and use (Brazil) Trade use
20 20/64 7.94 – –
19 19/64 7.54 Round, flat beans3 –
18 18/64 7.14 Round, flat beans3 Commonly used ISO sieve4
17 17/64 6.75 Round, flat beans3 Commonly used ISO sieve4
16 16/64 6.35 Round, flat beans3 Commonly used ISO sieve4
15 15/64 5.95 Round, flat beans3 NYBOT: at least 50% of the sample on 15 or larger6
14 14/64 5.56 Round, flat beans3 NYBOT: at most 5% below 146
13 13/64 5.16 Slotted, peaberries; also round for flat beans in the laboratory series3 –
12 12/64 4.76 Slotted, peaberries3 Among the commonly used ISO sieves4
11 11/64 4.37 Slotted, peaberries3 –
10 10/64 3.97 Slotted, peaberries; also round for flat beans in the laboratory series3 –
9 9/64 3.57 Slotted, peaberries3 –
8 8/64 3.18 Slotted, peaberries3 –

For a slotted screen the number refers to the slot width; the slot dimensions are specified in Annex B of ISO 4150, which was not available for this entry.4

Reference values: defect equivalents#

Number of defects that make one full defect (or, where stated, the full defects that one item counts for).

Defect SCAA (FAO 2000) Brazil (MAPA 2003)
Full black 1 → 1, primary1 1 → 13
Full sour (ardido) 1 → 1, primary1 2 → 13
Pod / dried cherry (coco) 1 → 1, primary1 1 → 13
Partial black 2–3 → 11 –
Partial sour 2–3 → 11 –
Immature (verde) – (quakers barred in specialty)1 5 → 13
Insect damage (brocado) 2–5 → 11 2–5 → 13
Poorly filled (mal granado) – 5 → 13
Shell (concha) 5 → 11 3 → 13
Broken / chipped (quebrado) 5 → 11 5 → 13
Floater 5 → 11 –
Water damage 2–5 → 11 –
Parchment (marinheiro) 2–3 → 11 2 → 13
Hull / husk 2–3 → 11 Large husk 1 → 1; small husk 2–3 → 13
Large stone or stick 2 → 1, primary1 1 → 5 (stick, stone or clod)3
Medium stone or stick 5 → 1, primary1 1 → 2 (regular size)3
Small stone or stick 1 → 11 –
Sample 300 g hulled coffee1 300 g3

In the FAO table every figure is the number of occurrences equal to one full defect, so two large stones or five medium stones make one full defect; in Brazil a single large stick, stone or clod counts as five defects. The table reproduces both sources as published.1,3 The current SCA handbook was not opened for this entry, so its values are not shown.

Where defects come from#

Grading standards define defects mostly by what they look like. In the Brazilian regulation only the sour bean is defined with a cause.3

  • Black (preto). A bean or piece of bean of opaque black colour.3 It is a primary defect in the SCAA system.1 The sources opened here do not assign a cause.
  • Sour (ardido). A bean or piece of bean with a brown colour in various shades, which the regulation attributes to fermentation processes.3 Also primary in the SCAA system.1 Which processing faults (over-fermentation, delayed processing) cause it belongs to fermentation and harvesting; no source opened here tests them.
  • Immature (verde). An immature bean, with silverskin stuck to it and a closed ventral groove.3 Quakers – beans that roast pale – are barred from SCAA specialty grade.1
  • Insect-damaged (brocado). A bean damaged by the coffee berry borer, with one or more holes.3
  • Poorly filled (mal granado). A bean with incomplete formation and little mass.3
  • Shell (concha). A shell-shaped piece formed when interlocked beans separate.3
  • Broken (quebrado). A piece of bean of variable shape or size.3 Breakage in the huller is covered under hulling and dry milling.
  • Floater. A secondary defect in the SCAA table.1

The machines match these visual definitions. A gravity table separates particles by weight,11 and a colour sorter rejects beans whose reflectance at one or two wavelengths falls outside the standard.9,10 How efficiently each machine removes each defect type in coffee was not measured in the sources opened. In seeds generally, surface discoloration caused by fungi is detectable in the visible range, whereas internal attributes are detectable in the near infrared.9

137456892
  1. Adjustable feed gate. 11
  2. Adjustable dividers set the fractions.11
  3. Deck slopes down to the front …
  4. … and from the high side to the low side.
  5. Air rises through the perforated deck and loosens the layer; light beans ride on top.11
  6. Eccentric drive. Vibration walks heavy beans uphill.11
  7. Heavy beans: to the high side.
  8. Light beans: to the low corner.
  9. Middle fraction.
A gravity table: a perforated deck sloping two ways, air rising through it and vibration walking heavy beans uphill while light beans slide to the low corner.11 Schematic, not to scale.

Size, density and cup quality#

Trade grades rank larger screens higher.8 The evidence on cup quality is mixed. In a Mexican study of three Catimor varieties, flat beans on screen 13 scored almost seven points higher than peaberries of the same screen in the Colombia variety.13 But small beans (screens 13 and 14) not separated by shape scored equal to or higher than bigger beans, and the effect of size and shape depended on the variety.13 In this study, size alone did not predict cup quality.

In the Brazilian UV study, more fully fluorescent beans had a negative effect on the final score and astringency, and partially fluorescent beans had a negative effect on score, astringency, sweetness, acidity, bitterness and aftertaste.12

Measurement at grading#

  • Sample. 300 g for sieving (ISO 4150) and for Brazilian classification; 300 g or 350 g for the SCAA/SCA defect count, depending on the source.1,2,3,4,14
  • Moisture. ISO 6673 is the reference method named in the ICO export rule for green coffee; accepted range 8–12.5%.6 SCAA specialty grade asks for 9–13% (basis not stated).1
  • Water activity. The SCA’s archived (heritage) green coffee standards set no moisture range; for specialty grade they set a water-activity limit instead: the beans must measure lower than 0.70 aw.14 See moisture and water activity for the safe ranges.
  • Bulk density. No grading standard opened for this entry sets a limit.
  • Visual examination. ISO 10470 provides the reference chart for identifying defects.5

Open questions#

  • Coffee-specific settings for gravity tables (deck slopes, air, vibration, recirculation of the middle fraction) have not been published in the sources found.
  • The causes of black and sour beans, and the share each processing step contributes, need primary studies linked to this entry.
  • How well colour sorters remove stinkers and mouldy beans, and at what loss of sound beans, is not reported in the sources opened.
  • Whether density predicts cup quality better than screen size has not been tested head to head in the sources opened.

Key studies#

Luna González et al., 2019 – Bean size, shape and cup quality. Three Catimor varieties grown in Mexico were separated by screen and by shape (flat, peaberry, mixed) and cupped. Shape and size changed the scores, but not always in favour of large beans: small, unseparated beans often cupped as well as or better than larger ones, and the effect depended on variety.

International Journal of Food Properties 22, 758–767 · doi:10.1080/10942912.2019.1603997

Sarmento et al., 2023 – Fluorescent beans under UV light. A natural specialty Arabica lot was split into treatments with increasing shares of beans that fluoresce under ultraviolet light, which were then analysed and cupped. Fluorescent beans had a negative effect on the final score and several sensory attributes, which supports the trade’s long-standing use of UV lamps to find hidden defects.

Research, Society and Development 12(8) · doi:10.33448/rsd-v12i8.42930

Pasikatan & Dowell, 2001 – Optical sorting of seeds. Reviews how optical sorters detect insect- and fungus-damaged seeds, from monochromatic and bichromatic visible sorting to near-infrared and fluorescence methods. It names coffee among the commodities sorted at high speed.

Applied Spectroscopy Reviews 36(4), 399–416 · USDA-ARS full text

Contested
  • Sample size for the SCAA/SCA defect count: FAO’s 2000 reproduction of the SCAA standard uses 300 g of hulled coffee1, while a 2024 ICO paper describing the SCA definition gives 0 to 5 defects per 350 g2 and the SCA’s archived green coffee standards require a 350 g sample.14 Size analysis under ISO 4150 and Brazilian classification also use 300 g.4,3 The current SCA handbook was not opened for this entry.
  • Which defects are "primary": the SCAA table reproduced by FAO lists full black, full sour, pod/cherry and large and medium stones and sticks as primary1; other systems do not split defects this way – Brazil divides them into intrinsic and extrinsic,3 and ISO 10470 uses five categories.5
  • Bean size and cup quality: trade grades rank larger screens higher8, but in a Mexican study of three Catimors small beans (screens 13 and 14) not separated by shape scored equal to or higher than bigger beans.13
Share or cite this entry

Sources

  1. AMutua, J. (2000). Post Harvest Handling and Processing of Coffee in African Countries, Annex 7: Green Coffee Classification and Grading. Food and Agriculture Organization of the United Nations.
  2. BInternational Coffee Organization (2024). Existing definitions of specialty coffee (FGSC 03/24). International Coffee Organization.
  3. AMinistério da Agricultura, Pecuária e Abastecimento (MAPA), Brazil (2003). Instrução Normativa nº 8, de 11 de junho de 2003 — Regulamento Técnico de Identidade e de Qualidade para a Classificação do Café Beneficiado Grão Cru. MAPA (copy hosted by Sindicafé-MG).
  4. AInternational Organization for Standardization (ISO/TC 34/SC 15) (2011). ISO 4150:2011 Green coffee or raw coffee — Size analysis — Manual and machine sieving. ISO.
  5. AInternational Organization for Standardization (ISO/TC 34/SC 15) (2004). ISO 10470:2004 Green coffee — Defect reference chart. ISO.
  6. AWinston, E.; Op de Laak, J.; Marsh, T.; Lempke, H.; Chapman, K. (2005). Arabica coffee manual for Lao-PDR, Chapter 8: Quality and export standards. FAO Regional Office for Asia and the Pacific, Bangkok.
  7. BInternational Trade Centre (ITC) (2011). The Coffee Exporter's Guide, Third Edition. International Trade Centre (UNCTAD/WTO), Geneva.
  8. BInternational Coffee Organization (2018). National Quality Standards (ICC 122-12). International Coffee Council, 122nd Session.
  9. APasikatan, M. C.; Dowell, F. E. (2001). Sorting Systems Based on Optical Methods for Detecting and Removing Seeds Infested Internally by Insects or Fungi: A Review. Applied Spectroscopy Reviews 36(4), 399–416.
  10. ALong, W. C. (ESM International Inc; later Satake USA Inc) (1996). Sorting machine (European Patent EP0517950B1). European Patent Office.
  11. AKulseth, L. G. (Forsbergs Inc) (1975). Gravity separator (US Patent 3,888,352). United States Patent Office.
  12. BSarmento, A. B. I.; Paiva, L. C.; Martins, J. P.; Campos, K. A.; Silva Neto, O. G.; Santos, D. A. (2023). Quality of Coffea arabica L. beans that emit fluorescence when subjected to ultraviolet light. Research, Society and Development 12(8). doi:10.33448/rsd-v12i8.42930 · other copies
  13. ALuna González, A.; Macías López, A.; Taboada Gaytán, O. R.; Morales Ramos, V. (2019). Cup quality attributes of Catimors as affected by size and shape of coffee bean (Coffea arabica L.). International Journal of Food Properties 22, 758–767. doi:10.1080/10942912.2019.1603997 · other copies
  14. ASpecialty Coffee Association Heritage Coffee Standards (archived), Green Coffee Standards. Specialty Coffee Association.

Tier A · authoritative · B · trade and expert · C · general. Supporting quotes for every source are in the data version of this entry. If a publisher’s site refuses access, try the doi link or “other copies”.

Suggest a change

On Grading and sorting of green coffee. Tell us what you would change and why. A source that supports it helps most.