Air Classification of Chickpea (Cicer arietinum L.) Flour: Proximate Composition and Protein Enrichment Outcomes

Chickpea (Cicer arietinum L.) flour is a promising ingredient for plant-based foods, and dry fractionation can yield fractions with distinct protein and starch contents. This study aimed to produce chickpea flour by milling and air classification and to determine the proximate composition of the resulting samples. Chickpeas purchased in Campinas, São Paulo, Brazil, were manually sorted, sanitized, dried at 120 °C for 20 min in a tray oven, and processed in an integrated milling and pneumatic-classification system (Netzsch Lab Compact Plus). Airflow was set to 100 m³ h⁻¹ and feed rate to 15 rpm, with milling wheel speeds of 2500, 3000, and 3500 rpm and classifier wheel speeds of 3000, 3500, and 4000 rpm. Six samples were evaluated: two operationally designated as starch-rich fractions (A3 and A4), three as protein-rich fractions (P3, P4, and P5), and whole flour (I), obtained after roasting and milling. Moisture, ash, protein, and lipid contents were determined by gravimetric, incineration, Kjeldahl, and Bligh–Dyer methods, respectively, and carbohydrates were calculated by difference. All analyses were performed in triplicate, and results are expressed as mean ± standard deviation. Moisture ranged from 3.41% to 4.09%, ash from 2.81% to 6.49%, protein from 18.03% to 23.60%, lipids from 6.87% to 7.63%, and carbohydrates from 59.37% to 68.79%. Sample A4 had the highest protein content (23.60 ± 0.79%), whereas P5 had the lowest (18.03 ± 0.72%); whole flour contained 20.86 ± 0.34%. The protein-rich samples showed no consistent enrichment relative to whole flour, indicating that the operational labels did not confirm the intended compositional separation. Ash contents in A3 and A4 (6.49% and 5.75%, respectively) were higher than in the other samples, which may be associated with the co-partitioning of mineral-dense cotyledonary cell walls or residual hull fragments into these coarse fractions. No inferential statistical tests were applied; therefore, numerical differences should not be interpreted as statistically significant. Although the six samples differed in proximate composition, the classification labels did not consistently correspond to higher protein concentration. Verification of the ash results, together with evaluation of particle-size distribution and fraction yield, is needed to better assess the performance of the dry-fractionation process.

Funding Agencies: FAPESP (grant numbers 2020/07015-7, 2022/11541-1, and 2024/02262-7).